PROGRESS IN PLANT NUTRITION:

PLENARY LECTURES OF THE XIV INTERNATIONAL

PLANT NUTRITION COLLOQUIUM
Progress in Plant Nutrition:
Plenary Lectures of the XIV
International Plant Nutrition
Colloquium
Food security and sustainability of agro-ecosystems through basic and applied research

Edited by

W.J. HORST
A.BURKERT
N. CLAASSEN
H. FLESSA
W.B. FROMMER
H.GOLDBACH
W.MERBACH
H.-W.OLFS
V. ROMHELD
B. SATTELMACHER
U. SCHMIDHALTER
M.K.SCHENK
and

N.v. WIREN

Reprinted from Plant and Soil, Volume 247, No.1 (2002).

SPRINGER-SCIENCE+BUSINESS MEDIA, B.V.

Library of Congress Cataloging-in-Pub1ication Data

A C.I.P. Cata10gue record for this book is avai1ab1e from the library of Congress.

ISBN 978-90-481-6191-1 ISBN 978-94-017-2789-1 (eBook)

DOI 10.1007/978-94-017-2789-1

Printed an acid-free paper

AH Rights reserved
© 2002 Springer Science+Business Media Dordrecht
Originally published by K1uwer Academic Publishers in 2002
Softcover reprint of the hardcover 1st edition 2002
No part of the material protected by this copyright notice
may be reproduced or utilized in any form or by any means,
electronic or mechanical, inc1uding photocopying,
recording Of by any information storage and retrieva1
system, without written permission from the copyright owner.
Contents

PROGRESS IN PLANT NUTRITION:

PLENARY LECTURES OF THE XIV INTERNATIONAL
PLANT NUTRITION COLLOQUIUM

Food security and sustainability of agro-ecosystems through basic and applied research

Preface vii

Plenary lectures

1. Plant nutrition research: Priorities to meet human needs for food in sustainable ways
I. Cakmak 3

2. Access and excess problems in plant nutrition

M. van Noordwijk and G. Cadisch 25
Symposium: Genetics and molecular biology of plant nutrition

3. Molecular mechanisms of potassium and sodium uptake in plants

P. Maser, M. Gierth and J.1. Schroeder 43

Symposium: Nutrient functions

4. A proposed role for copper ions in cell wall loosening

S.c. Fry, J.G. Miller and J.C. Dumville 57
Symposium: The role of the apoplast in plant mineral nutrition

5. The functions of cell wall polysaccharides in composition and architecture revealed through
mutations
N.C. Carpita and M.C. McCann 71

Symposium: Mineral nutrition: plant quality and plant health

6. The impact of mineral nutrients in food crops on global human health

RM. Welch 83

Symposium: Salinity and plant - soil - water relations

7. Avenues for increasing salt tolerance of crops, and the role of physiologically based selection
traits
R Munns, S. Husain, AR Rivelli, RA James, AG. Condon, M.P. Lindsay, E.S. Lagudah,
D.P. Schachtman and RA Hare 93

Symposium: Mineral element toxicity and resistance

8. Mechanisms of metal resistance in plants: aluminum and heavy metals

L.V. Kochian, N.S. Pence, D.L.D. Letham, M.A. Pineros, J.v. Magalhaes, O.A. Hoekenga and
D.F. Garvin 109
Symposium: Nutrient acquisition: mechanisms and modelling

9. Use of modellinfl; to understand nutrient acquisition by plants

G.J.D. Kirk 123

Symposium: Soil organisms/plant interactions

10. Malate plays a central role in plant nutrition

J. Schulze, M. Tesfaye, R.H.M.G. Litjens, B. Bucciarelli, G. Trepp, S. Miller, D. Samac, D. Allan
and C.P. Vance 133

Symposium: Fertilizer use with regard to optimum yield and environment

11. Precision agriculture: a challenge for crop nutrition management

P.e. Robert 143

Symposium: Plant nutrition and sustainable development

12. Plant nutrient management for enhanced productivity in intensive grain production systems of
the United States and Asia
A. Dobermann and K.G. Cassman 153

Symposium: Nutrient dynamics in natural and agricultural ecosystems - Processes and modelling

13. Nitrogen saturation in temperate forest ecosystems: current theory, remaining questions and
recent advances
J.D. Aber 179
 Plant and Soil 247: vii, 2002.
vii
© 2002 Kluwer Academic Publishers.

Preface

Plant Nutrition deals with the functions and dynamics of mineral elements in plants, soils and ecosystems with the
objective to enhance plant growth and crop yields, to improve plant quality and health, to optimise soil fertility, to
reduce the need for external production inputs, and to reduce environmental constraints due to mineral and organic
fertilisation. Plant nutrition is an applied science contributing not only in the long term but also in the short term to
the performance of plants/plant communities in agro- and natural ecosystems and their sustainability.
Plant Nutrition is a very diverse discipline stretching from Molecular Biology to field application. The Inter-
national Plant Nutrition Colloquium is the international platform for Plant Nutrition gathering scientist from all
fields, in order to present and discuss the present state of knowledge and the substantial contribution that Plant
Nutrition has to make and can make to food production and sustainability not only of agro- but also of forest
and natural ecosystems. The world's population is expected to reach 8 billion by 2025 and most of this growth
in population will occur in developing countries. To feed the world with such a marked increase in population,
a great improvement in food production must be achieved particularly in these countries. To meet this challenge,
present agricultural productivity must be increased on the cultivated land. However, in many developing countries,
particularly in Africa, reduced soil fertility is a major threat both to food production and to ecosystem viability.
Food production is not only a quantitative challenge. Improving the nutrient status of plants provides a further
valuable means of enhancing food quality and is of extreme benefit to the health of both plants and humans. In
contrast to other parts of the world, in the industrialised countries in particular, and in East Asia, the excessive use
of fertilisers has led to the build-up of soil nutrient stocks resulting in increased nutrient fluxes into the environment.
The efficient use of fertilisers aimed at both optimising yield and minimising environmental pollution is, therefore,
a particularly important issue.
Meeting the challenge of sustainable plant nutrient management requires not only the application of current
knowledge but also the widening of knowledge through basic and applied research. Entering the 3rd millennium
there seems justification in our optimism in achieving this goal because of the availability and systematic use of
new methodological tools, such as molecular biological techniques, modelling, and ecosystem engineering. Such
techniques will be particularly helpful in the breeding of various plant genotypes of elevated capacity in the use of
soil and fertiliser nutrients, as well as those adapted to adverse soil conditions, and with increased micronutrient
density. The techniques should also allow the implementation of sustainable nutrient management for agro- and
forest-ecosystems.
The contributions in this book provide an excellent current state-of-the-art in the different areas of Plant
Nutrition research and its contribution to food security and agro-ecosystem sustainability while maintaining and
enhancing environmental quality.

Walter J. Horst
Hannover, 19 September 2002
PLENARY LECTURES

Opening Plenar Lecture:

I. Cakmak
Closing Plenar Lecture:
M. van Noordwijk
... Plant and Soil 247: 3-24, 2002,
. , © 2002 Kluwer Academic Publishers.
3

Plant nutrition research: Priorities to meet human needs for food in

sustainable ways

Ismail Cakmak
Faculty of Engineering and Natural Sciences, Sabanci University, 81474 Tuzla, 1stanbul, Turkey

Key words: food security, macronutrients, micronutrients, mineral fertilization, molecular biology, nutrient
efficiency, nutrient management

Abstract
The world population is expanding rapidly and will likely be 10 billion by the year 2050. Limited availability of
additional arable land and water resources, and the declining trend in crop yields globally make food security a
major challenge in the 21st century. According to the projections, food production on presently used land must be
doubled in the next two decades to meet food demand of the growing world population. To achieve the required
massive increase in food production, large enhancements in application of fertilizers and improvements of soil
fertility are indispensable approaches. Presently, in many developing countries, poor soil fertility, low levels of
available mineral nutrients in soil, improper nutrient management, along with the lack of plant genotypes having
high tolerance to nutrient deficiencies or toxicities are major constraints contributing to food insecurity, malnu-
trition (i.e., micronutrient deficiencies) and ecosystem degradation. Plant nutrition research provides invaluable
information highly useful in elimination of these constraints, and thus, sustaining food security and well-being
of humans without harming the environment. The fact that at least 60% of cultivated soils have growth-limiting
problems with mineral-nutrient deficiencies and toxicities, and about 50% of the world population suffers from
micronutrient deficiencies make plant nutrition research a major promising area in meeting the global demand for
sufficient food production with enhanced nutritional value in this millennium. Integration of plant nutrition research
with plant genetics and molecular biology is indispensable in developing plant genotypes with high genetic ability
to adapt to nutrient deficient and toxic soil conditions and to allocate more micronutrients into edible plant products
such as cereal grains.

Introduction ticularly widespread in Asia; about 70% of the under-

The world's population is expected to grow from 6
billion to around 10 billion by 2050. This is a me-
dian projection representing the addition of an extra
4 billion people onto the present population (Byrnes
and Bumb, 1998). Virtually all of the anticipated
increase in world population will occur in Africa,
Latin America and Asia. Presently, these regions have
already serious problems concerning food production,
access to food, water scarcity and nutritional dis-
orders/malnutrition. Pinstrup-Andersen et ai. (1999)
reported that more than 800 million people living in
the developing countries are undernourished caused
by inadequate food availability. The problem is par-
* FAX No.: 90-216-4839550; E-mail: cakmak@sabanciuniv.edu
nourished people globally live in Asia, predominantly
in India and China. Sub-Saharan Africa with about
200 million undernourished people is another crit-
ical region of the world suffering from malnutrition.
According to the FAO projections, the number of un-
dernourished people will decline to only 680 million
by 2010, and this decrease will mainly take place
in Asia, while in Sub-Saharan Africa food insecurity
will likely continue with increasing numbers of people
affected (FAO, 1996, 1999; Rosegrant et aI., 2001).
Among the undernourished people children under
5 years of age are particularly at risk. Globally, 30%
of children are underweight, and low birth weight be-
comes an important factor in child malnutrition and
premature death (Pinstrup-Andersen, 1999; UNDP,
1999). Micronutrient deficiencies such as iron (Fe),
4
3.0
zinc (Zn) , iodine (I) and vitamin A deficiencies are
critical issues in the developing countries, and result
t
'C
01967-82
]i 2.5 .1982-94
in severe impairments of human health and develop- >-
.1995-2020
ment such as impairments of physical growth, immune
system, cognitive development and enhancements in
.
1ii
I!!
0
2.0

.5 1.5
anemia and maternal mortality (Pinstrup-Andersen, J:.
i
1999; Welch and Graham, 2000). Micronutrient de- e
C!l
1.0
ficiencies are also widespread in well-developed in- 1ii
::I 0.5
dustrialized countries. More than 3 billion people c
C
cc 0.0
globally suffer from Fe and Zn deficiencies (Graham
Developing World Developed
et aI., 2001). Heavy and monotonous consumption countries countries
of cereal-based foods with low concentrations and re- Figure 1. Annual growth rates in cereal yields during 1967-82,
duced bioavailability of Fe and Zn has been considered 1982-94, and 1995-2020 (redrawn from Pinstrup-Andersen et al.,
a major reason for the widespread deficiencies of Zn 1999).
and Fe in developing countries (Graham et al., 2001;
Welch and Graham, 1999). Table 1. Expected ratios of crop production and
To feed a world with huge increases in popula- cultivated land in 2020 relative to those in 1990
tion and to sustain the well-being of humans, a large (from Evenson, 1999)
increase in food production must be achieved. The Crop Production Land cultivated
expected increases in world population will result in
a serious pressure on the existing agricultural land Wheat 1.58 1.06
via urbanization and intensification of crop produc- Maize 1.56 1.13
tion (Alexandratos, 1995). The global food projections Rice 1.66 1.07
model (The International Model for Policy Analysis of Other grains 1.48 1.09
Soybeans 1.77 1.14
Commodities, and Trade, IMPACT) developed by the
Roots/tubers 3.28 1.15
International Food Policy Research Institute (IFPRI)
indicates that to meet the food demand of the world Mean 1.89 1.11
population in 2020, annual cereal production needs to
increase by 40%, from 1773 billion tonnes in 1993 to
nearly 2500 billion tonnes in 2020 (Rosegrant et al.,
1999, 2001). Of this increase in total cereal demand, ulation pressure on agriculture is already very high.
85% will come from the developing countries. Others For example, in Sub-Saharan Africa the annual growth
estimates that the total world cereal production must of total cereal production dropped from 1.9% in the
rise to 2.7 billion tons (Dyson, 1999) or to nearly 4 1970s to 0.66% in the 1990s. These decreases were
billion tones (Borlaug and Dowswell, 1993) by 2025 attributed to falling cereal prices, decreases in use of
to match food requirements of the world population. fertilizers, growing water shortages and impairments
The projected increase in food production must be in soil fertility and management practices (Gruhn et
accomplished on the existing cultivated areas because al., 2000; Pinstrup-Andersen et aI., 1999; Rosegrant et
the expansion of new land is limited due to environ- aI., 2001). The decreasing trend in cereal production
mental concerns, urbanization and increasing water must be reversed; otherwise very serious food deficits
scarcity. Evenson (1999) reported that global plant- will occur resulting in a significant threat to human
based food production will increase by more than 60% nutrition and health.
by 2020 while the area used for crop production will Soil productivity is decreasing globally due to
expand by only around 10% (Table 1). However, re- enhanced soil degradation in the form of erosion, nu-
cent trends indicate that the growth rate (based on ton trient depletion, water scarcity, acidity, salinisation,
ha -1) in crop production has begun to decline in the depletion of organic matter and poor drainage. Nearly
last 10 years, and possibly, cannot keep pace with the 40% of the agricultural land has been affected by soil
projected increase in global food demand in the fol- degradation, particularly in Sub-Saharan Africa and
lowing decades (Fig. 1; Mifflin, 2000; Rosegrantet al., Central America (Table 2; Scherr, 1999). Enhanced
2001). This situation is especially dramatic in many soil degradation along with slowdown in cereal pro-
developing countries in Africa and Asia where pop- duction contributes to food insecurity in developing
 5
Table 2. Global estimates of soil degradation in agri- 80 1200
cultural land (from Scherr, 1999)
-0- Fertilizer Use
"iii"
"iii" -+- Cereal Production c
Region Agricultural land c g
g 60 950 c
Total Degraded Percent c ~
~ Ic
(million hectares) I 40 700 0
Africa 187 121 65 3: :g
:::l :::I
"0
Asia 536 206 38
~ !!

....
II..
South America 142 64 45 1: 20 450 a;
Central America 38 28 74 I!!
~
North America 236 63 26
Europe 287 72 25 o +-----,.----r--.--~----r-__r-~-,......... 200
Oceania 49 8 16 1~ 1~ 1~ 1m 1~ 1~ 1~ 1~ 1~

Years
World 1475 562 38 Figure 2. Growth in fertilizer use and cereal production in develop-
ing countries during 1961-1995 (redrawn from FAO, 1996).

increase in rice production since 1966 (from 240 to

countries. To improve food security for developing
countries, investments (including education, clean wa-
ter, irrigation, rural roads and agricultural research)
are projected to cost 579 billion US dollars between
1997 to 2020. Among the investments agricultural
research activities will account for 120 billion US dol-
lars (Rosegrant et al., 2001). According to Rosegrant
et al. (2001), the projected investments for improving
food security between 1997 and 2020 are feasible and
represent less than 5 % of the government expenditures
in the developing world over the next 20 years.
Fertilizer roles in meeting global food demand
To meet the challenge of 'food security', agricultural
production must increase on the existing land, and
therefore crop production must be intensified per unit
of agricultural land. Mineral nutrients are the ma-
jor contributor to enhancing crop production, and in
maintaining soil productivity and preventing soil de-
gradation. Generally, improving the nutritional status
of plants by applying fertilizers and maintaining soil
fertility has been the critical step in the doubling of
food production both in developed and developing
countries since the beginning of 'Green Revolution'
(Loneragan, 1997). As presented in Figure 2, increases
in cereal production in the past 40 years were well
associated with corresponding increases in fertilizer
consumption in developing countries (FAO, 2000).
According to Tillman (1999) the doubling of food pro-
duction during the past 40 years has been associated
with about 6.9-fold increase in N fertilization, 3.5-fold
increase in P fertilization and only 1.1-fold increase
in cultivated land area. In Asia, there was a 2-fold
483 million tons), while the area used for rice produc-
tion increased by only 13%, and fertilizer consumption
per hectare of land has increased from 20 to 145 kg
(Hossain and Singh, 2000). According to Hossain and
Singh (2000), this increase in fertilizer consumption is
not enough to sustain food security in the Asian coun-
tries, and should be increased to over 250 kg ha- 1 of
NPK.
Nearly all increases in projected food requirements
in the next decades will the result of enhancements
in yield per unit area and intensive use of agricul-
tural land. To increase yield capacity of crop plants
and to ensure global food demand in 2020, fertilizer
use should increase from 144 million tons in 1990 to
208 million tons in 2020 (Bumb and Baanante, 1996;
FAO, 2000). Possibly, this projected increase in fer-
tilizer consumption by 2020 will not be adequate to
meet both food production requirements and nutrient
depletion from soil resulting from nutrient removal by
harvesting crops. Byrnes and Bumb (1998) estimated
that fertilizer consumption should increase up to 300
million tons to match required demands of food pro-
duction and nutrient removal from soils by 2020. In
view of this estimate, in future, developing countries
should develop new policies and take new measures to
provide more support to farmers regarding the supply
of fertilizers.
Environmental concerns
Enhanced use of fertilizers raises concerns due to
adverse effects on the environment in terms of eu-
trophication of surface waters, pollution of drinking
water and gaseous emission causing global warming.
6
Therefore, nutrient use efficiency and improved soil
management become an important challenge, particu-
larly for Nand P fertilizers. When the application of
N fertilizers is not properly managed and is realized
at excessive levels, losses of N from agricultural lands
can occur by N03 leaching, NH3 volatilisation or by
N oxides emission. Leaching and runoff of N03 into
ground water and surface waters is a major environ-
mental problem in developed countries, particularly in
Europe (Howarth, 1998; Kroeze and Seitzinger, 1998).
Pollution of groundwater with N03 impairs the quality
of drinking water and causes various harmful effects
on human health. Contamination of lakes and rivers
with N03 stimulates algal growth and depletion of 02
resulting in an increasing risk of fish deaths on a large
scale, especially in coastal areas (Maene, 2000; Mo-
sier et aI., 2001). Large areas on the North Atlantic
Ocean and North Sea in Europe have been reported to
suffer from eutrophication caused by excessive use of
Non agricultural land (Howarth, 1998).
There is a very close relationship between applica-
tion rates of N fertilizers and emission of nitrous oxide
(N20) (Erickson et aI., 2000). Nitrous oxide is one of
the most important greenhouse gases affecting global
warming and increasing ozone destruction. About
0.5% (Veldkamp and Keller, 1997) or 1.5% (Smith
et aI., 1997) of fertilizer N applied is lost from soil
as gaseous emission. Several management strategies
have been developed to control and minimize N losses
such as use of N fertilizers with enzyme inhibitors (ur-
ease and nitrification inhibitors) and controlled-release
N fertilizers, timing and placement of fertilization, and
soil and plant analysis to define rates of N application.
For example, use of nitrification inhibitors is effective
to reduce N20 emission under field conditions (McTa-
ggart et aI., 1997; Mosier et aI., 1994) and the timing
of N application plays an important role in controlling
N losses as shown in flooded rice (Humphreys, 1988).
Like N03, P is also involved in eutrophication
of water through its runoff from soils into rivers and
lakes. Solubility of P in soils is very low, and therefore
P cannot be leached as easily as N03. Due to re-
peated applications of P-containing fertilizers together
with manure, P concentrations in topsoil increases,
leading to saturation of soil binding sites with P and
thus facilitating movement of P into surface waters.
Holford et ai. (1997) showed that P could be leached
when 17-38% of the sorption capacity of soils was
saturated. There is increasing evidence showing that
the major source of P in surface water showing eu-
trophication is the application of compost, manure
and sewage (Eghball and Gilley, 1999; Reynolds and
Davies, 2001). Phosphorus from manure application
moves much deeper in soil than the P from fertilizer
(Eghball et aI., 1996). Higher movement of P from
manure is attributed to existence of greater amount of
organic compounds in manure, facilitating P solubility
and movement in the soil profile.
Generally fertilizer-dependent environmental prob-
lems generally take place under specific soil con-
ditions or in poorly managed agricultural areas, in
particular in developed countries that employ high
fertilizer application rates. In contrast, in many de-
veloping countries with tremendous requirement for
food, continuous nutrient depletion and low usage of
mineral fertilizers are the concerns, not the environ-
mental pollution (Gruhn et aI., 2000). Nearly 40% of
all globally cropped land shows degradation (Table 2),
with nutrient depletion being a particular form of soil
degradation (Scherr, 1999). Nutrient depletion occurs
in many developing countries (e.g., in Sub-Saharan
Africa). In Sub-Saharan Africa, due to increasing pres-
sure on the cultivable land, farmers use the existing
land extensively without addition of adequate amounts
of fertilizers or consideration of proper soil manage-
ment practices. Annual use of mineral nutrients per
ha of arable land in 1996 has averaged around 9 kg
in Sub-Saharan Africa and 98 kg in the world (Gruhn
et aI., 2000). When the nutrients removed by plants
at harvest are not replenished by fertilization or by
proper soil management practices (i.e., by crop rota-
tion, use of plant residues and manure), soils become
depleted of mineral nutrients. With the cumulative
removal of nutrients marked decreases occur in soil
productivity. Bumb and Baanante (1996) reported that
nutrient removal exceeds nutrient replenishment by a
factor 3-4 in Sub-Saharan Africa. About 86% of the
countries in Africa show a net annual nutrient deple-
tion greater than 30 kg of NPK per hectare (Henao and
Baanante, 1999). In many developing countries net
nutrient depletion varies between 60 and 100 kg ha- 1
per year, resulting in serious decline in soil fertility.
It is estimated that in Africa, 1.5 billion US dol-
lars per year is needed to minimize nutrient depletion
and related decreases in soil productivity by applying
mineral fertilizers (Henao and Baananate, 1999).
Obviously, enhancing and maintaining soil fertil-
ity are very critical issues in meeting food security
in developing countries. Impaired soil fertility by
continuous cropping with low supply of mineral nu-
trients is considered a major threat not only to food
production but also to ecosystem viability (Pinstrup-

Andersen, 1999; Tillman, 1999). Reduced soil fertility
and crop production results in increased pressure to
bring more land into crop production at the expense
of forests and marginal lands. Such areas are gener-
ally poor in fertility and sensitive to rapid degradation
when cultivated. Conversion of forests into agricul-
tural uses also contributes to global warming. Mosier
(1998) reported that conversion of forests to agricul-
tural use promotes the turnover of soil-C and soil-N,
and thus increases in C loss and N20 emission from
soil. These consequences suggest that in developing
countries there is an urgent need for application of
adequate amounts of mineral nutrients and proper nu-
trient management systems, such a combination of
mineral fertilizers with organic materials, including
legumes in the rotation systems, timing and place-
ment of nutrient applications and balanced supply of
mineral nutrients. The contributions of these manage-
ment systems to improvements in soil productivity are
discussed below.
Nutrient management
Manure applications
Improvements of soil fertility and efficient use of min-
eral nutrients are of high ecological and economic
importance. Use of basic and practical research in-
formation provided through plant nutrition studies will
playa decisive role in establishing efficient and ecolo-
gically based nutrient management systems. Combin-
ation of mineral fertilizers with different types of or-
ganic materials, including legumes in rotational crop-
ping systems and recycling of crop residues are widely
recommended strategies to improve soil fertility and
enhance nutrient use efficiency.
Applying and promoting these strategies is of par-
ticularly importance in long-term continuous cropping
systems. The rice-wheat cropping system is the most
widespread cropping system in Southern and Eastern
Asia. Around 12.5 million hectares of land in Pakistan,
India, Nepal and Bangladesh and 10 million hectares
in China are under the rice-wheat cropping system
(Yadav, 1998; Yadav et aI., 2000a,b). In India, 33%
of the rice and 42% of the wheat production occurs
under this cropping system. Possibly, as a result of
long-term continuous rice-wheat cropping, yields of
both rice and wheat began to decline. Application of
green and farmyard manures has been shown to be an
important management strategy to reverse and sustain
7
Table 3. Grain yields averaged over years and loc-
ations in long-term rice-wheat system fertilized
with NPK alone or with different organic materials
in India (from Yadav et al., 2000b)
Treatment Yield (t ha- i )
Rice Wheat
Control 1967 1065
50%NPK 3180 2321
100%NPK 4551 3506
50% NPK + farmyard 4242 3629
50% NPK + crop residues 4007 3468
50% NPK + green manures 4465 3532
high crop productivity in rice-wheat cropping systems
in Asia. This effect of manures becomes more distinct
when they are applied together with mineral fertilizers.
In a long-term field experiment at six different loc-
ations using a rice-wheat rotation system, Yadav et
ai. (2000a) demonstrated that incorporation of green
manure resulted in additional increases in grain yield
of rice and prevented yield decline in wheat. Green
manures or farmyard manure applied along with a re-
duced rate of NPK applications were able to reduce
the mineral fertilizers used as much as 50% (Table 3).
Furthermore, also in continuous sorghum cropping
systems without rotation, grain yield decreased stead-
ily, and increases in yield and its maintenance could be
achieved when mineral fertilizers were combined with
manure applications (Bekunda et aI., 1997).
Green and farmyard manures have been shown to
improve the solubility and uptake of P from spar-
ingly soluble P compounds in soil and enhance the
utilisation of P from fertilizers. Organic compounds
released during the decomposition of manures in-
creases the availability of P from soil or fertilizers
(Iyamuremye and Dick, 1996). In India, the integ-
rated supply of P fertilizers along with cattle manure
under a wheat-soybean cropping system over 5 years
greatly improved wheat and soybean yield (Reddy et
aI., 1999). At an identical rate of P supply from fertil-
izer or manure, yield increases were much greater with
manure-P than with the fertilizer-Po This positive effect
of manure was ascribed to mobilization of native soil P
and improved physico-chemical properties of the soil
by manure use (Reddy et aI., 1999). Organic manures
are also highly beneficial in acid soils and contribute
to alleviation of adverse effects of Al toxicity on crop
production by reducing levels of exchangeable Al and
increasing concentrations of soluble P (see below).
8
Crop residues
Soils under continuous cultivation are depleted of
nutrients, and soil physical properties are impaired be-
cause organic matter is exhausted. In many countries
in Africa and Asia very little or no crop residue is
left in the field, much being used for feeding animals
or for fuel. Crop residues represent a good substrate
providing C and N for the activity of microorganisms
in the soil. Many of these organisms are involved in the
formation and stabilisation of soil aggregates. As plant
residues contain relatively large amounts of mineral
nutrients, their retention in the field also contributes
to the mineral content of soils and leads to savings
in the need to apply mineral fertilizers (Timsina and
Connor, 2001). In West Africa Sahelian region, use of
crop residues in the field resulted in marked increases
(up to 73%) in total dry matter production of differ-
ent cereals. The enhancing effect of crop residues on
cereal growth was ascribed to improved P availabil-
ity and by the protection of seedlings against wind
erosion provided by the crop residues (Buerkert et al.,
2000). In a recent field experiment in West Africa, the
effect of a composted material, prepared from crop
residues, animal manure and household refuses, was
studied with respect to soil properties and grain yield
of sorghum (Quedraogo et al., 2001). Application of
10 tons of compost per hectare resulted in up to 3-
fold increases in grain yield depending on the location.
The compost application was also effective in over-
coming the marked decreases resulting from delayed
sowing. The results reported by Buerkert et al. (2000)
and Quedraogo et al. (2001) indicate the importance of
crop residues and compost applications in preventing
soil degradation and maintaining crop productivity.
Such applications, when adopted on large scale areas
in Africa, could increase yields and maintain soil
productivity.
Legumes
In view of their well-documented role in reducing
the need for fertilizer N and in their ability to im-
prove soil fertility, there is an increasing trend to
include legumes in the crop rotation systems. Inclu-
sion of legumes in intercropping and crop rotation
system provides a number of valuable benefits includ-
ing enhanced soil physical structure, improved soil
fertility, reduced need for mineral N fertilizers, and
improved use efficiency of N and other mineral nutri-
ents. Such crop rotations also reduce the risk to crops
from crop pests and suppresses weed growth (Becker
and Johnson, 1999; Van Kessel and Hartley, 2000).
The combination of relatively low amount of fertilizers
with crop rotations including legume crops signific-
antly enhances crop production (Boddey et al., 1997;
Dakora and Keya, 1997). Continuous cropping sys-
tems without rotation with legumes were seen as a key
factor involved in declining crop production and soil
fertility in developing world, particularly in Africa.
In southern Africa, groundnut is increasingly con-
sidered an important grain legume in improving soil
fertility and sustaining crop production in the maize-
based cropping system. In field conditions in Zimb-
abwe, grain yield of maize was studied under con-
tinuous maize and groundnut-maize rotation systems
with and without NPK supply (Waddington and Karig-
windi, 2001). The inclusion of groundnut in the rota-
tion system almost doubled the grain yield of maize
in plots without application of inorganic fertilizers.
When inorganic fertilizers were applied, groundnut ro-
tation was still effective in increasing the grain yield.
In the northern Guinea Savanna of Ghana, the bene-
ficial effects on maize yield of rotating maize with
legumes could be demonstrated in spite of adequate
N fertilization and optimization of soil physical prop-
erties. Therefore, the positive effects of legumes were
attributed to elimination of allelopathic effects (Horst
and Hardter, 1994). Obviously, there are several causal
factors other than N nutrition and soil physical prop-
erties contributing to the yield improvement of crops
succeeding legumes. This area is a further potential
area of investigation in future.
The beneficial effects of legumes on crop pro-
ductivity vary between different types of cropping
systems. The most widespread cropping systems used
in Africa include the single cropping system, the
crop rotations and the intercropping of legumes and
cereals. The contributions of legumes in these crop-
ping systems with maize were studied in the West
African Savanna. The results demonstrated that maize
yields following the sole-cropped legumes were 2-fold
higher compared to the 'maize after maize' system
(Dakora and Keya, 1997). Maize yields were also less
after maize-legume intercropped system when com-
pared to the sole-cropped legume. Based on these
results, possibly, the intercropping of legumes and
cereals is less effective in increasing yield than the
crop rotation system involving cereal production after
the sole-cropped legumes system (Dakora and Kaya,
1997).

Inclusion of legumes in cropping systems enhances
soil N levels and reduces decline in soil fertility asso-
ciated with intensive cultivation. In field experiments
in Australia, cotton grown following non-legume ro-
tation crops (i.e., wheat) required an optimum N fer-
tilizer application of 179 kg N ha -1, while following
the grain or green manure legumes only 90 and 52 kg
N ha- 1were sufficient, respectively (Rochester et aI.,
2001). Rochester et al. (2000) also showed that includ-
ing legumes in rotation system positively affected soil
quality by improving soil physical properties. When
compared to continuous sorghum cropping system,
a legume rotation system was effective in increasing
the pH of the rhizosphere (Alvey et aI. , 2001). The
increase in pH was considered an important change
for plants allowing them to cope with acidity-induced
nutrient deficiency and AI-toxicity.
The application of rotation systems including
legumes in developing countries appears to be a fun-
damental strategy to improve soil fertility and crop
production. However, the soils in developing countries
generally also suffer from acidity (AI toxicity), P de-
ficiency or water stress, which are major constraints
to nodulation and N2 fixation (Hungria and Vargas,
2000). Therefore, future research activities should fo-
cus on development of new legume genotypes having a
high tolerance to the mentioned edaphic stress factors.
Balanced nutrient supply
Cultivation of high yielding cultivars under continu-
ous monoculture or via intensive cropping systems
without sufficient fertilization leads to depletion of nu-
trients in soils. Consequently, over time high yields
are no longer sustainable and make these cropping
systems uneconomical. Long-term intensive cropping
systems also disturb the balance of mineral nutrients
that exist in soils. In India, the deficit between re-
moved and added nutrients has been estimated to range
between 4 and 5 million tons in wheat and rice cultiv-
ated areas only (Singh, 1998). The nutrients depleted
should be replenished to sustain high yields under
continuous cropping systems. Currently, the mineral
nutritional problems occurring in this cropping system
are not confined to only N, P or K. This is a possible
reason for declining yields of rice and wheat in the last
years in several Asian countries, despite application of
high level of NPK (Tandon, 1995, 1998; Hossain and
Singh, 2000). Deficiencies of Zn and S are particular
examples occurring in the rice-cropping systems. Un-
9
0.0
Nutrients added
Figure 3. Wheat grain yield based on long-term multi location
experiments (based on Tandon, 1995).
less these mineral deficiencies are not corrected, high
yields cannot be ensured, and responses to increasing
supplies ofN, P and K will become very small or zero
depending on the soil type. Based on a large num-
ber of long-term field experiments at many locations,
Tandon (1995) summarized the effects of balanced nu-
trient supply on grain yield of wheat (Figure 3). Each
added nutrient other than N greatly increased grain
yield over the N application alone. Obviously, continu-
ous use of N alone leads to severe depletion of other
nutrients, with a corresponding decrease in the grain
yield. The substantial increases in grain yield by Zn
application in addition to the NPK application indicate
critical importance of these nutrients in crop produc-
tion in Asian countries. Likewise, S deficiency is also a
common nutritional disorder under intensive cropping
system. Long-term field experiments in Bangladesh
demonstrated high requirement for S in maintaining
high yields in rice production (Tandon, 1995). Applic-
ations of S fertilizers markedly improved grain yield:
depending on the season increases in rice grain yield
by S fertilization varied between 19 and 40%.
An adequate and balanced supply of mineral nutri-
ents in acid soils can be ineffective in maintaining high
yield when the soil pH is not elevated to a certain level
by addition of lime. In acid soils, crop production can
be markedly decreased to very low levels or even to
zero levels under long-term supply of N only or only
NPK. However, application of NPK together with lime
in acid soils can result in sustainable high yield levels
(Gruhn et aI., 2000; Tandon, 1995).
Aluminium toxicity
Acidity is a critical yield-limiting problem in many
soils. About 40% of cultivated soils globally have
10
acidity problem leading to significant decreases in
crop production despite adequate supply of mineral
nutrients such as N, P and K (Herrera-Estrella, 1999;
von Uexkiill and Mutuert, 1995). In acid soils major
constraints to plant growth are toxicities of hydrogen
(H+), aluminium (AI) and manganese (Mn) and de-
ficiencies of P, calcium (Ca) and magnesium (Mg).
Among these constraints Al toxicity is the most im-
portant yield-limiting factor (Marschner, 1991). As
mentioned above, liming represents an effective man-
agement strategy in overcoming or minimizing soil
acidity and related Al toxicity. A continuous acidi-
fication without liming may result in deterioration of
soils that can make even the acid-tolerant genotypes
useless.
By raising pH up to above 5.5 through lime (i.e.,
CaC03) applications soluble and exchangeable Al are
precipitated as hydroxy-AI species. Generally, lime is
added at 1.65 ton ha- l of CaC03-equivalent per mil-
liequivalent of exchangeable Al per 15 cm soil depth
(De Pauw, 1994). There is, however, a risk of over-
liming in soils by causing occurrence of deficiencies of
P and micronutrients. For amelioration of Al toxicity
and related P deficiency in acid soils application of
organic materials in form of crop residues, compost
and green manure have been considered a practical
and potentially cheaper practice over lime applications
(Haynes and Mokolobate, 2001). Incorporation of or-
ganic materials can also lower the requirement for lime
application. During the decomposition of organic ma-
terials in soils several organic compounds are released
which complex phytotoxic monomeric Al species in
soil solution and render them into non-toxic forms.
The most important organic compounds involved in
Al detoxification in soil solution are low molecu-
lar weight organic acids and humic and fulvic acids
(Haynes and Mokolobate, 2001; Hue and Amien,
1989). The organic compounds adsorb onto Al and
Fe oxides and thereby prevent sorption of P on these
oxides, a process which improves the availability of P
to plant roots (Haynes and Mokolobate, 2001; Hue et
al., 1994). Alleviation of phytotoxic effects of Al by
organic materials has also been attributed to increases
in soil pH (Noble et a!., 1996, Wong et a!., 1999).
Apparently, incorporation of organic matter into acid
soils should be widely adopted at the farm level. How-
ever, there are some concerns about whether adequate
amounts of animal manures are available to improve
soil fertility on all affected cultivated lands due to in-
sufficient numbers of animals (e.g., in West Africa;
Williams et al., 1995).
Development of new genotypes with high Al tol-
erance might be an important and sustainable strategy
to cope with soil acidity related constraints in soils.
As both liming and incorporation of organic materials
have some limitations to affect subsoil acidity, a com-
bination of genotypes having high Al tolerance with
liming and organic matter addition would be the ideal
solution to the AI-toxicity problem in soils. Plant spe-
cies and genotypes of a given species greatly differ in
their tolerance to Al toxicity indicating the existence
of a high genetic potential to develop AI-tolerant gen-
otypes (Aniol, 1991; Carver and Ownby, 1995; de la
Fuente and Herrera-Estrella, 1999). Two major plant
mechanisms have been described as being involved
in tolerance to Al toxicity: exclusion of Al from root
apex and internal detoxification of Al (Horst, 1995;
Kochian, 1995; Ma et a!., 2001). Recent evidence
suggests that organic acids with their high chelating
ability play a fundamental role in Al tolerance both
externally and internally. Root exudation of certain
organic acids is closely correlated with the level of
Al tolerance, for example malate in wheat (Ryan et
a!., 1995), citrate in maize (Pellet et al., 1995), ox-
alate in buckwheat (Ma et a!., 1997) and citrate in
soybean (Figure 4; Yang et al., 2000). The exudation
of organic acids from roots in response to Al toxicity
appears to be a highly promising adaptive mechan-
ism to allow plants to tolerate acid soils. Therefore,
development of novel plant genotypes with a high ge-
netic ability to exude organic acids from their roots
when grown on AI-toxic soils is a high priority for
research. Exudation of organic acids from roots under
acidic soil conditions can also be helpful in solubil-
isation of P from sparingly soluble hydrated oxides
of Al and Fe (Marschner, 1995). Supporting these
suggestions, very recently, excellent molecular evid-
ence has been presented showing that the expression
of a citrate synthase gene, isolated from Pseudomonas
aeruginosa, in tobacco resulted in a large increase in
citrate efflux and, accordingly, in a high tolerance to
both Al toxicity (de la Fuente et al., 1997) and P-
deficiency (Lopez-Bucio et a!., 2000) (see below for
further details).
Phosphorus deficiency
Phosphorus deficiency is also a widespread nutritional
problem greatly affecting crop production. It is es-
timated that 5.7 billion hectares of land (equivalent to
about 67% of the total farmland used worldwide) con-

~O~------------------------------~
I±I Shishio (SENSITIVE)
., _ .Suzunari (RESISTANT)
E:cco
I:
o~
300
~~
~ ~ 200
~.,
U III
"'-
U 0
't: E 100
u-C
!:!
o nutrients and delivers more mineral nutrients, par-
0-6 6-12 12-18 18·24
Treatment duration (h)
Figure 4. Effect of aluminum (50 /Lillol) treatment on citrate exuda-
tion over a 24-h period at 6-h intervals in soybean cultivars differing
in their resistance to aluminum toxicity (redrawn frOill Yang et al.,
2000).
tain too low levels of plant available P, which is lim-
iting crop production in many world regions (Batjes,
1997). Availability of P to plant roots is limited both
in acidic and alkaline soils, mainly, due to formation
of sparingly soluble phosphate compounds with Al
and Fe in acidic and Ca in alkaline soils (Marschner,
1995). Every year large amounts of P fertilizers, up to
35 million tons ofP20s (Isherwood, 2000) are applied
to soils for crop production globally, and only 10-20%
of the applied P fertilizers can be absorbed by plants
(Holford, 1997). The remaining is rapidly transformed
into unavailable P forms, which are not readily ab-
sorbed by plant roots. Development of plant genotypes
having a high capacity to use of both native soil-P and
added fertilizer-P is, therefore, very important. Some
plant species have evolved several adaptive mechan-
isms to improve their ability to cope with soils having
low levels of available P. Enhancements of root system
to better exploration the soil for P is a typical response
of plants to P deficiency, particularly by increasing the
production and elongation of root hairs (Gahoonia and
Nielsen, 1998) and by the formation of mycorrhizal
association (Dodd, 2000; Marschner, 1995).
Root hairs greatly contribute to the ability of roots
to take up mineral nutrients from the soil. Plants pos-
sessing a high root hair response to limited available
soil-P conditions are able to acquire/use P in larger
amounts than plants that have a poorer ability to pro-
liferate root hairs (Bates and Lynch, 2000). In a study
with 32_p labelled soil, Gahoonia and Nielsen (1998)
demonstrated that 70% of root hairs grown into the
labelled soil contributed up to 63% of the total P up-
take by barley. This is convincing evidence of the
11
particular role of root hairs in P acquisition from soils.
Therefore, increasing attention should be paid to the
research aiming at developing genotypes that have a
high genetic ability to produce extensive root hairs
under P-deficient soil conditions.
Mycorrhizal colonization of roots enhances the
ability of roots to explore the soil for P. This occurs
through the action of the symbiotic soil fungal my-
celliums (hyphaes) infecting roots. This association
results in greater exploration of the soil for available
ticularly P, to plant roots (Dodd, 2000; George and
Marschner, 1996; Marschner, 1998). The most widely
distributed type of mycorrhizal fungi is the arbuscular
mycorrhizae (AM). It is estimated that the extent of
fungus mycelium may be in the range of 10-100 m
per cm root or per gram of soil under field condi-
tions in P-poor soil (McGonigle and Miller, 1999).
Therefore, the contribution of mycorrhizal association
to mineral nutrition of host plants is much higher in
soils with poor availability of mineral nutrients than
in soils rich in nutrients. In pot experiments, mycor-
rhizal colonization contributed between 70 and 80%
of the total P uptake and 50 and 60% of the total Zn
and Cu uptake in white clover plants (Li et aI., 1991).
Benefits of AM are not only confined to mineral nu-
trition of host plants. Mycorrhizal fungi also increase
plant tolerance to root pathogens, water stress and
heavy metals, and are involved in improvement of N-
fixation by rhizobia and in physical structural stability
of soils (Dodd, 2000; George and Marschner, 1996;
Marschner, 1995). Due to such diverse beneficial ef-
fects mycorrhizas have a high potential in sustaining
soil fertility and restoration of agricultural lands that
are both physically and chemically degraded. Soils
adversely affected by physical and chemical factors
are widespread in developing countries (Table 2), and
can be improved through inoculation with selected
AM populations. There are a few examples showing
the importance of AM inoculation in low-input crop
production systems. On acid soils in Latin America,
several field experiments were performed to demon-
strate the role of AM inoculation on the growth of
cassava. The results showed that AM improved cas-
sava yields by 20-25% on the average (Sieverding,
1991, cited in Dodd, 2000).
A further particular adaptive response of plants to
P deficiency is the release of organic acids and en-
zymes from roots into the rhizosphere. Root exudation
of organic acids (i.e., citrate and malate) when cul-
tivated under P deficiency has been shown for rape
12
(Hoffland et aI., 1989) and for white lupin (Dinkelaker
et aI., 1989, 1992; Johnson et aI., 1995). Organic
acids released from roots of rape plants were able to
solubilize P from Ca-phosphates (Dinkelaker et aI.,
1989; Hoffland et aI., 1989) and contribute to P nu-
trition of plants. Calcicoles plants adapted to alkaline
soils released larger amounts of organic acids from
roots than the species with poor adaptation ability to
alkaline soils (calcifuges) (Tyler and Strom, 1995).
Studies dealing with genotypic variation in release of
organic acids within a species and its relation to P
deficiency tolerance are very few (Gahoonia et aI.,
2000; Gaume et al. 2001), and therefore, more ex-
tensive research is needed in this area. Recently, trans-
genic tobacco plants have been developed with high
citrate-overproducing capacity, and these plants had
a higher tolerance to P deficiency under alkaline soil
conditions compared to their wild-type counterparts
(Lopez-Bucio et aI., 2000). This work emphasizes the
importance of organic acids in the adaptation of plants
to P-limited soils.
Iron and zinc deficiencies in plants
Micronutrient deficiencies in plants are becoming in-
creasingly important globally. Intensive cultivation of
high yielding cultivars with heavy applications ofN, P
and K fertilizers leads to the occurrence of micronu-
trient deficiencies in many countries. Among the mi-
cronutrient deficiencies reported to occur worldwide,
Zn and Fe deficiencies are particularly good examples.
In plants, Fe deficiency often occurs in calcareous
soils where chemical availability of Fe to plant roots
is extremely low. It is estimated that Fe deficiency is
widespread occurring in about 30% of the cultivated
soils on the world, and results in large decreases in
crop production and quality (Chen and Barak, 1982;
Vose, 1982). Correction of Fe deficiency is not always
easy through the use of Fe fertilizers because of their
extremely poor solubility and the requirement that re-
peated applications during the cropping season are
needed to correct the deficiency. Therefore, remedi-
ation of Fe deficiency chlorosis by fertilizers is a costly
and time-consuming management. Plants evolved on
Fe-deficient calcareous soils have developed adaptive
mechanisms to overcome or minimize the effects of
Fe deficiency stress. Marschner and his colleagues
have identified two different types of adaptive root re-
sponses to Fe deficiency. The first strategy exists in all
plant families other than graminaceous family, and is
characterized by the mechanisms involving acidifica-
tion of rhizosphere, activation of a membrane-bound
ferric reductase enzyme and the release of reducing
substances from roots (Marschner, 1995; Marschner
and Romheld, 1994; Marschner et aI., 1986). These
mechanisms are highly inducible in response to Fe de-
ficiency; they improve solubilisation and uptake of Fe
from sparingly soluble Fe compounds in soil.
The other strategy (Strategy II) is confined only
to graminaceous species, and characterized by the re-
lease of the mugineic acid family phytosiderophores
(MAs) to chelate Fe in rhizosphere. The resulting
Fe(III)-MAs are taken up into root cells by an indu-
cible specific transporter in the root cell plasma mem-
brane (Marschner and Romheld, 1994; Romheld and
Marschner, 1986). The efficiency of the mechanisms
occurring in response to Fe deficiency greatly differ
among and within plant species, and this explains the
reason why a large genotypic variation exists within
and among plant species in their sensitivity to Fe de-
ficiency (Kawai et aI., 1988; Marschner et aI., 1986;
Rengel, 2001; Wei et aI., 1997). For example, among
cereal species rice has high sensitivity to Fe deficiency
and very poor ability to release MAs (Kanazawa et aI.,
1994; Marschner and Romheld, 1990). As discussed
in detail below, very recently it has been demonstrated
that genetically engineered rice plants to release more
MAs from their roots was associated with enhanced
growth and increased grain yield under Fe-deficient
soil conditions (Takahashi et aI., 2001).
Currently, the genetic variation for tolerance to Fe
deficiency within Strategy I and II plants is being ex-
ploited by using classical and modem breeding tools
to develop novel genotypes with high genetic ability
to induce adaptive root responses to Fe deficiency.
Therefore, in screening of genotypes for tolerance
to Fe deficiency, root mechanisms are increasingly
considered to be a reliable selection parameters (Ell-
sworth et aI., 1998; Jolley et aI., 1996; Romera et
aI., 1991). Studies concerning breeding for tolerance
to Fe deficiency are becoming numerous with the re-
cent developments in molecular marker technology
(Fairbanks, 2000; Lin et aI., 2000). The past studies
concerning the characterization of Fe deficiency toler-
ance in plants mostly focussed on physiological and
biochemical characterization of root mechanisms. In
recent years there is an increasing trend to character-
ize localisation, identification and isolation of genes
determining tolerance to Fe deficiency (see below).
Like Fe deficiency, Zn deficiency is widespread
throughout the world and occurs in nearly all coun-
 13
200
!!!
G)
~
t~
G) III 150
u. I:
~ 0
c,.-
ZC
c8
N!:. 100
-:-.
o G) Detection of
1:':':
0"
._ ::I Zn
ii~
E .- 50
::I
III
I:
0
(,,)

0
1993 1994 1995 1996 1997 1998 1999 2000
Years
Figure 5. Consumption of Zn-containing NP and NPK fertilizers in Turkey (Cakmak et al., 1999a; and unpublished results of the Turkish
Fertilizer Producer Association, 2001).

tries. Based on analysis of 298 soil samples from
different countries Zn deficiency has been found to be
the most widespread micronutrient deficiency world-
wide (Sillanpaa, 1990; Sillanpaa and Vlek, 1985).
According to Graham and Welch (1996), nearly 50%
of the soils cultivated for cereal production globally
have low levels of plant available Zn. In Turkey, Zn
deficiency is the most widespread micronutrient de-
ficiency: the half of all cultivated soils (28 million
hectares) has Zn deficiency (Eyupoglu et aI., 1993).
In Central Anatolia, the major wheat growing area of
Turkey, increases in grain yield by soil application of
Zn varied between 5 and 554% depending on location
and plant available concentration of Zn in soils (Cak-
mak et aI., 1996a). Large increases in grain yield by
Zn applications were also demonstrated in Australia
(Graham et aI., 1992) and India (Tandon, 1995, 1998).
The critical importance of Zn deficiency for wheat
production in Turkey has been shown first in 1994 in
the framework of a large-scale project supported by
NATO-Science for Stability Program (Cakmak et aI.,
1999a; Kalayci et aI., 1999). Before 1994 there was
no Zn-containing mix (NPK) fertilizer in Turkey. The
spectacular increases in grain yield by Zn fertiliza-
tion evoked a growing interest in the project results
among farmers and fertilizer companies. In 1995 the
first Zn-containing NPK fertilizer was produced in the
amount of 2000 tons, and this amount has increased
progressively and reached 189000 tons in 2000 (Fig-
ure 5; Cakmak et ai. 1999a, and unpublished results
of the Turkish Fertilizer Producer Association). This
is an excellent example showing the importance of
diagnosis of micronutrient status of plants and soils in
sustaining high yields in crops. In the future, special
attention should be paid to the micronutrient status
of plants and soils to ensure further increases in crop
production and sustained soil fertility, especially in
developing countries.
Cereal species and genotypes of a given cer-
eal species differ greatly in their response to
Zn deficiency and Zn fertilization. Tolerance to
Zn deficiency was found to decline in the or-
der rye>triticale>barley>bread wheat>oat >durum
wheat (Cakmaket aI., 1998; Ekiz et aI., 1998). Among
wheat species, durum wheats are particularly sensitive
to Zn deficiency, and this high sensitivity is possibly
related to a poor capacity of durum wheat to synthesis
and release MAs from roots (Cakmak et aI., 1996b,
1998; Rengel et aI., 1998). Under Zn-deficient con-
ditions, concentrations of MAs in both root extracts
and exudates were much lower in durum than in bread
wheat cultivars (Table 4).
Differences in tolerance to Zn deficiency between
sorghum, wheat and corn correlated well with the
amounts of MAs released from roots under Zn defi-
ciency (Hopkins et aI., 1998). In the case of bread
wheat cultivars differing markedly in tolerance to Zn
deficiency, the release rate of MAs from roots was,
however, not related well to the differences in toler-
ance to Zn deficiency (Cakmak et aI., 1998; Erenoglu
et aI., 1996). Apparently, besides release of MAs,
other mechanisms are also involved in tolerance to Zn
deficiency tolerance, such as root uptake and root-to-
14
Table 4. Effect of Zn supply (+Zn = 1 /LM, -Zn = 0) on the rate of phytosiderophore (PS)
release from roots and DMA (2' -deoxymugineic acid) concentration in roots of Zn-efficient
bread wheat cultivar Kirac and Zn-inefficient durum wheat cultivar Kiziltan grown for 14
days in nutrient solution (from Cakmak et al., 1996b)

Cultivars PS release from roots PS concentration in root

(/Lmol g-l root dry wt) (/Lmol DMA g-l root fresh wt)
-Zn +Zn -Zn +Zn

Kirac (Zn-efficient) 8.7 0.8 2.76 0.15

Kiziltan (Zn-inefficient) 0.8 0.9 0.65 0.31

shoot transport rates of Zn and internal Zn utilisation
(Cakmak et al., 1999a; Rengel, 1999,2001).
In view of global widespread occurrence of Zn
deficiency, development and release of Zn deficiency
tolerant genotypes is a relevant topic. Existence of
substantial genotypic variation in tolerance to Zn defi-
ciency can be exploited to enhance tolerance. Little
information is available on how tolerance is inher-
ited. In future, an intensive research activity should
be carried out on this subject. Identification of mo-
lecular markers for the genes affecting Zn deficiency
tolerance and their mapping within the chromosomes
would greatly stimulate breeding activities for Zn de-
ficiency tolerance. By using wheat-rye translocation,
substitution and addition lines in pot experiments it
has been shown that the rye chromosomes 1R, 2R
and 7R most likely carry genes enhancing tolerance
to Zn deficiency (Cakmak et aI., 1997; Schlegel et
al., 1998), with genes on IRS and 7RS being most
effective (Schlegel and Cakmak, 1997).
Recently, differences in tolerance to Zn deficiency
were studied in wild and primitive wheats, as these
species are increasingly used as a genetic material
for improving modem wheats for different traits. For
example, Aegilops tauschii (DD), the source of the
D genome in bread wheat (BBAADD), and Triticum
monococcum (AA) are being used in such programs.
Interestingly, many of the wild wheats and Aegilops
species that exhibited very high tolerance to Zn defi-
ciency originated from Turkey (Harlan, 1981; Nesbitt
and Samuel, 1998). Since Turkey is among the coun-
tries having the most severe Zn-deficient soils in the
world (Cakmak et., 1996a, 1999a; SillanpiUl, 1990),
it is mostly likely that these wild wheats and wild
relatives of wheat possess an exceptional ability to tol-
erate Zn deficiency in soils. They can be exploited in
breeding wheats for high tolerance to Zn deficiency.
This suggestion is supported by our recent results:
addition of the whole D genome from Ae. tauschii
(DD) or whole A genome from Tr. monococcum (AA)
to durum wheat (BBAA) very significantly enhanced
the tolerance of the synthetic wheats to Zn deficiency
(Cakmak et aI., 1999b). It seems likely that the genes
contributing to high tolerance to Zn deficiency toler-
ance are located on the A and D genomes. Future
research should focus on exploitation of the selected
wild and primitive wheats and Aegilops species to im-
prove modem wheats in terms of increased tolerance
to Zn deficiency.
Enhancing iron and zinc concentrations in seeds
As is the case with plants and soils, Fe and Zn
deficiencies are also the most widespread micronutri-
ent deficiencies in humans. These deficiencies affect
more than 3 billion people worldwide causing serious
health and productivity problems for various popula-
tion groups, especially among resource-poor women,
infants and children (Graham et al., 2001; Welch and
Graham, 1999). These deficiencies are particularly
widespread in developing countries where diets are
rich in cereal-based foods with low concentration of
bioavailabable Zn and Fe. Major health consequences
of Zn and Fe deficiencies include retardation of
growth, anemia and impaired immune functions, di-
minished intellectual development and retarded sexual
maturation (Shrimpton, 1993; Yip, 1994). The prob-
lem with Zn and Fe deficiency has been intensified
with the increased cultivation of high-yielding cul-
tivars of the 'Green Revolution' cereals whose grains
contain low concentrations of Zn and Fe, that are rich
in compounds which limit the bioavailability of Zn and
Fe to humans (Welch and Graham, 1999). Increasing
the concentrations of Zn and Fe in cereal grains is a
high priority research task, and will greatly contrib-
ute to the alleviation of micronutrient deficiencies in

human populations worldwide. Enhancements in con-
centration of micronutrients, especially Zn, also result
in several positive consequences for crop production,
such as improvements in seedling vigour, pathogen
resistance, competition against weeds and, finally, an
enhanced yields (Graham and Rengel, 1993; Rengel
and Graham, 1995).
One important strategy to increase micronutrient
concentrations in grains is fertilization of plants via
15
Table 5. Effects of different Zn application
methods on Zn concentrations in whole shoots
sampled at the beginning of stem elongation
stage and in mature grains of the bread wheat
cultivar Gerek-79 and the durum wheat cultivar
Kunduru-1149 grown on a Zn-deficient cal-
careous soil in Central Anatolia (from Yilmaz
et al., 1997)
Zinc Zinc concentration
application (mg Zn kg -1 dry weight)

soils or foliar applications. Depending on the plant methods Gerek-79 Kunduru-1149

species, soil application of Zn to soil can increase Whole shoots

Zn concentration of plants by as much as 2-3-fold Control 11 10
(Rengel et al., 1999). However, even with very high Soil 21 20
Zn fertilization rates, the Zn concentration in wheat Seed 14 12
grain does not show correspondingly high increases Leaf 68 47
from Zn fertilization. Furthermore, the same applies Soil+leaf 82 52
to Fe as well. Possibly, because of highly limited Seed+leaf 93 59
phloem mobility of Fe, soil or foliar applications of LSD (5%) 15 16
Fe remain ineffective in increasing Fe concentrations
in grains (Gupta, 1991). Interestingly, as has been Grain
shown in field experiments in Central Anatolia and Control 9 12
Australia, Zn deficiency in wheat can easily be cor- Soil 17 19
rected, and yield maximized by broadcast application Seed 11 10
Leaf 30 20
of Zn fertilizers; however, broadcast application of Zn
Soil+leaf 34 35
is not very effective in increasing Zn concentrations in
Seed+leaf 34 25
grains up to desired levels to meet human requirements
(Graham et al., 1992; Yilmaz et al., 1997). Under LSD (5%) 6 6
Zn-deficient soil conditions, the highest concentra-
tions of Zn in grain could be obtained by a combined
soil and foliar application of Zn (Table 5). In most
cases, Zn and Fe concentrations in grain do not in-
crease to the desired levels, even by using high rates
of fertilizer applications (Rengel et al., 1999). There trations in cereal grains. For example, Peterson et al.
is also no direct economic motivation for farmers to (1986) analyzed grains of 27 wheat cultivars from 24
improve the nutritional quality of grains alone by fer- countries grown at six different locations, and showed
tilization. However, as indicated above, if this type of that grain concentrations of Fe ranged from 26 to 69
fertilizer improves crop yields as well as nutritional and that of Zn from 20 to 53 mg kg-I. However, in
quality, farmers would be likely to adopt such prac- that study the environmental factors exerted greater
tices. In addition, most of the research conducted on influence on the genotypic variation found for mi-
the micronutrient nutrition of plants deals with cor- cronutrient concentrations than the genetic factors. In
recting deficiencies and with improving grain yield. the framework of the Consultative Group on Interna-
By contrast, research on enhancing micronutrient con- tional Agricultural Research (CGIAR) Micronutrient
centrations in grain or other edible parts of plants is Project (Bouis et al., 2000), using large number of
very limited. More research activities are, therefore, genotypes, a significant variation was demonstrated
needed aiming at improving the bioavailable levels of in grain concentrations of Zn and Fe as reported by
micronutrients in cereal grains. Gregorio et al. (2000) for rice, Beebe et al. (2000)
An alternative approach to increasing Fe and Zn for bean, Monasterio and Graham (2000) for wheat
concentrations in grain is to exploit the genetic vari- and Banziger and Long (2000) for maize. Presently,
ation in grain concentrations of these micronutrients the variation in micronutrient concentration found is
within plant genome. Highly significant genetic differ- being exploited in the breeding programs conducted at
ences have been reported for both Fe and Zn concen- several CGIAR centres.
16
Table 6. Concentration and content (totaI amount) of zinc in seeds
of wild diploid (T. boeoticum) tetraploid (T. dicoccoides) and mod-
em tetraploid (T. durum) and hexaploid (T. aestivum) wheats (from
Cakmak: et aI., 2000)
Species n Concentration Content
(mg kg-I) (/Lg seed-I)
Mean Range Mean
Wild wheats
T. boeoticum 12 89 45-177 1.3
T. dicoccoides 19 91 20-159 3.1
Modern wheats
T. durum 11 31 18-50 1.6
T. aestivum 16 27 15-61 1.2
n - number of accessions analysed.
Despite the important genetic vanatlOns among
crop genotypes, average values for Zn and Fe concen-
trations in cereal grains are still low, especially when
compared with legume grains which show higher con-
centrations of Zn and Fe as well as a wider range of
concentrations (Welch and Graham, 1999). As an al-
ternative to cultivated wheats, wild wheats have been
recommended as an important source of genetic ma-
terial for enhancing micronutrient concentrations in
grains (Cakmak et al., 2000). In screening of several
wild diploid (Triticum boeoticum) and wild tetraploid
(Triticum dicoccoides) wheats, an impressive vari-
ation was found in Zn concentrations (Table 6). The
variation found for Zn within wild wheats was much
greater than the variation within the cultivated wheats.
Certain accessions of T. dicoccoides also had substan-
tial amounts of Zn per seed, and this was not related
to seed size or seed weight (no concentration effects),
and could not be shown for other mineral nutrients.
These results suggest that the wild wheats, in partic-
ular T. dicoccoides, represents as a valuable source
of genetic diversity for increasing the Zn concentra-
tion in grains of cultivated wheats. The studies with
wheat-dicoccoides substution lines revealed that the
genes determining high density of Zn in grains are
located on the T. dicoccoides chromosomes 6A and 6B
(Cakmak et aI., 2000). Little is, however, known about
the mechanisms responsible for the high accumulation
of Zn in T. dicoccoides grain. Research investigating
the mechanisms of Zn accumulation in the grain of
T. dicoccoides is particularly important and urgently
needed.
Perspectives on the molecular biology of plant
nutrition
Contribution of plant nutrition to food security is not
confined only to the maintenance of soil fertility and
identification and correction of nutrient deficiencies or
Range
toxicities by applying mineral fertilizers and organic
materials to soils. Selection and characterisation of
0.7-3.0 plant genotypes with enhanced genetic ability to toler-
0.5-5.3 ate deficient or excess levels of mineral nutrients and
adapt to adverse soil physical and chemical conditions,
including soil acidity, salinity and water deficiency,
0.6-2.7 are important research areas for plant nutritionists. The
0.6-3.1 information obtained from this area during the last 20
years has substantially contributed to increased world
food production (Gruhn et aI., 2000; Loneragan, 1997;
Marschner, 1995). Linking the knowledge obtained
from plant-nutrition research to molecular biology-
based research will result in major progress in the
development of genotypes with elevated capacity to
adapt to adverse soil conditions. Presently, a number
of genes have been isolated and cloned which are in-
volved in root exudation of nutrient-mobilizing or ion-
detoxifying organic compounds (i.e., organic acids
and phytosiderophores) (Lopez-Bucio et aI., 2001; Ma
et aI., 2001; Richardson et aI., 2001; Takahashi et
aI., 2001) and uptake, transport and accumulation of
mineral nutrients such as, N03 -, NH4 +, H2P04 - ,
K+ and some micronutrients (Crawford and Glass,
1998; Goto et aI., 1999; Hirsch and Susman, 1999;
Raghothama, 1999,2000; von Wiren et aI., 2000a).
Aluminium toxicity and phosphorus deficiency
Several examples are given below showing the im-
portance of molecular approaches to improving plant
growth in soils having mineral nutrient problems. Suc-
cessful attempts have been made in the past 5 years
to develop transgenic plants that produce and release
large amounts of organic acids. Organic acids are
key compounds involved in the adaptive mechanisms
used by plants to tolerate AI-toxic and P-deficient
soil conditions (Kochian, 1995; Lopez-Bucio et aI.,
2000; Marschner, 1995). To test whether or not in-
creased biosynthesis and root exudation of citrate
improves tolerance of plants to Al toxicity and P
deficiency, Luis Herrera-Estrella's research group de-
veloped transgenic tobacco plants overexpressing a
citrate synthase gene from Pseudomonas aeruginosa
(de la Fuente et aI., 1997; Lopez-Bucio et aI., 2000).
Overexpression of citrate synthase enzyme in tobacco

plants enhanced citrate concentration in root tissue by
as much as lO-fold greater than the control plants.
The transgenic plants also increased citrate efflux from
roots by as much as 4-fold over the control plants.
These increases are associated with elevated toler-
ance to Al toxicity demonstrated by less Al inhibition
of both root growth and root hair formation (de la
Fuente et aI., 1997). Recently, the same research
group demonstrated that the transgenic tobacco plants
were also highly tolerant to low P supply in an al-
kaline soil. The increased tolerance of these transgenic
plants to P-deficient soil conditions was attributed to
enhanced solubilisation of sparingly soluble calcium
phosphates in soil by the increased citrate efflux from
roots (Lopez-Bucio et al., 2000). Similar results were
reported by Koyama et aI. (2000) using transgenic
Arabidopsis thaliana overexpressing a mitochondrial
citrate synthase enzyme from carrot (Koyama et aI.,
2000). In the latter experiment the transgenic Ara-
bidopsis plants having an elevated citrate efflux from
roots had a greater capacity to absorb P from soil
containing AI-phosphates. The above studies convin-
cingly demonstrate that enhanced synthesis and root
exudation of organic acids is a decisive step in de-
veloping high tolerance to both P deficiency and Al
toxicity in soils. However, in a very recent study, Del-
haize et aI. (2001) using the same transgenic tobacco
lines as well as additional transgenic lines, with even
with much greater expression of citrate synthetase pro-
tein than found by Lopez-Bucio et al. (2000) (up to
lOO-fold greater level) did not have either increased
citrate concentrations in their roots or increased cit-
rate efflux from their roots when compared with the
control plants. Consequently, increased Al tolerance
reported by de la Fuente et aI. (1997) was not con-
firmed by Delhaize et aI. (2001). These controversial
results were attributed to a possible sensitivity of the
Pseudomonas aeruginosa gene to different environ-
mental conditions. Irrespective of such controversial
results, manipulation of both biosynthesis and efflux
of organic acids to the external solution by roots would
be a key genetic engineering approach in developing
novel genotypes for sustainable crop production on
soils rich in Ca and Al phosphates.
In addition to the secretion of organic acids by
roots, the secretion of the enzyme phytase represents
an important target for engineering plants for growth
on P-limited soils. In cultivated soils most of the total
P exists in the form of organic P (Marschner, 1995).
Phytate represents an important proportion of the total
organic P in soils (Delal, 1977). To use this source
17
of P, some plants are able to activate extracellular en-
zymes including both acid phosphatases and phytases.
It is believed that plants are not able to efficiently use
P from phytate, due to limited ability of plant roots
to hydrolyze phytate in the rhizosphere (Hayes et aI.,
1999). Recently, Richardson et aI. (2001) generated
transgenic Arabidopsis plants expressing a phytase
gene from Aspergillus niger, and these plants could
secrete a substantial amount of this phytase enzyme
into their growing medium. Consequently, the trans-
genic plants showed a greater ability to use P from
phytate and grew much better over the control plants.
Richardson et aI. (2001) suggested that the phytase
activity around plant roots is a significant factor in
utilizing of P from phytate in soils. Therefore, de-
veloping plants that overexpress extracellular phytases
and other phosphatases is an important challenge in
future.
In improvement of uptake and use efficiency of P
in plants, Pi transporters might play a critical role.
With the development of P deficiency stress an ex-
tensive expression of Pi transporter proteins occurs in
root cells, particularly in cells staying in close con-
tact with soil solution, i.e., epidermal cells and root
hair cells (Daram et aI., 1998; Raghothama, 2000;
Smith, 2001). Such preferential expression and loc-
alisation of Pi transporter proteins in roots indicate
a direct role of Pi transporters in utilisation of soil
and fertilizer P. Studies are needed to demonstrate the
role of Pi transporters in growth and P uptake of crop
plants when grown in soils limited in P supply. In cell
culture experiments, expression of a high-affinity Pi
transporter gene of Arabidopsis thaliana in tobacco
cultured cells enhanced biomass production and P
uptake capacity of transgenic cells under P-limited
conditions (Mitsukawa et aI., 1997). Like Pi trans-
porters, many other mineral nutrient transporters have
been isolated and functionally characterized in plants,
such as nitrate, ammonium and sulphate transporters.
Also these transporters are preferentially expressed in
root hairs under deficiency of the corresponding nu-
trients (Hell and Hillebrand, 2001; Smith, 2001; Von
Wiren et aI., 2000a,b; Williams and Miller, 2001). A
major future challenge is to improve crop plants with
elevated levels of nutrient transporters to contribute to
the nutrient uptake of plants when grown under soil
conditions with limited supply of mineral nutrients.
Micronutrient deficiencies
In view of the fact that over 3 billion people suffer
18
from micronutrient deficiencies (Graham et aI., 2001),
a considerable amount of research in the 21st century
should be devoted to the development of genotypes for
enhanced uptake and accumulation of micronutrients
in edible plant parts. In this regard, genetic engineer-
ing provides excellent possibilities to enhance plant
capacity for acquisition of micronutrients from soils.
An increased uptake of Zn or Fe might be related to the
level of expression of the corresponding transporter
proteins located at the plasma membranes of root cells.
Iron deficiency-induced expression of transporter pro-
teins are known and characterized in yeast (Stearman
et aI., 1996) and Arabidopsis thaliana (Eide et aI.,
1996). The transporter protein encoded by the gene
IRTI is expressed in roots in response to Fe deficiency
and is involved in mediating Fe uptake in Arabidopsis.
Also the uptake of Zn is affected by the expression of
transporter proteins. In yeast and Arabidopsis thaliana
three genes, ZIP 1, ZIP2 and ZIP3, encoding Zn trans-
porters, were isolated and functionally characterized
(Grotz et aI., 1998; Zhao and Eide, 1996a,b). These
Zn transporter proteins are expressed in roots under
Zn-limited conditions, and involved in high- and low-
affinity Zn uptake system. The Fe and Zn transporter
genes can be engineered in crop plants to enhance
their capacity for uptake and accumulation of Zn and
Fe. Such genetically modified plants might contribute
greatly to improving both plant and human nutrition
on a global scale. However, the present progress in this
area is limited and, therefore, an extensive research
is needed in the future, particularly on the roles of
such transporters in translocation of micronutrients to
edible parts of plants, e.g., grains and seeds.
In Strategy I plants (i.e., dicots and non-
graminaceous monocots) reduction of Fe(III) to Fe(II)
by a ferric reductase enzyme is an obligatory step in
uptake of Fe. Therefore, genetic manipulation of this
enzyme can be an efficient way to improve Fe nutrition
of plants. There is a little work on the molecular char-
acterisation of the inducible plasma membrane ferric
chelate reductase enzyme in crop plants. Robinson et
aI. (1997, 1999) isolated the FR02 gene from Ara-
bidopsis thaliana. This gene is induced in Fe-deficient
roots in response to Fe deficiency, and encodes for
the expression of the ferric chelate reductase protein
within the plasma membrane. In yeast, two Fe (III)
reductase genes, FREI and FRE2, have been isolated
and cloned (Georgatsou and Alexandraki, 1994; Oki et
aI. 1999). Samuelsen et aI. (1998) studied the role of
these FRE genes in the Fe nutrition of plants by gener-
ating transgenic tobacco plants containing the FREl,
the FRE2 or both genes. The transgenic lines having
the FRE2 gene, but not the FREI gene, showed higher
tolerance to Fe deficiency as a result of greater root
reductase activity that caused higher chlorophyll and
leaf Fe concentrations in these transformants. These
results together with the results of Robinson et aI.
(1997, 1999) suggested that the FRE2 and FR02 genes
represent promising candidates to manipulate Strategy
I plants to enhance their tolerance to Fe deficiency.
In Strategy II plants ferric reduction is not an ob-
ligatory step; these plants take up Fe in the form
of Fe(I1I)-MAs complex (see above). Among the
graminaceous species, barley is the most tolerant to
Fe deficiency and releases larger amount of MAs
from roots, while rice is the most susceptible one
and accordingly secretes little MAs (Romheld and
Marschner, 1990; Takagi et aI., 1984). During the
past 5 years, Mori and his research group have pub-
lished extensively in this area. They have isolated,
cloned and characterized some of the crucial genes in-
volved in the biosynthesis of MAs, e.g., nicotianamine
synthase (NAS) and nicotianamine aminotransferase
(NAAT) (Higuchi et aI., 1999, 2001; Takahashi et
aI., 1999). Kanazawa et aI. (1994) demonstrated that
the activity of NAAT very closely correlated with the
differences between barley and rice in Fe deficiency
chlorosis and the release rate of MAs from roots. Re-
cently, Takahashi et aI. (2001) generated transgenic
rice plants by transferring the NAAT genes from bar-
ley into rice. They demonstrated that the transgenic
rice plants with overexpression of NAAT genes de-
veloped a higher tolerance to Fe deficiency in an
alkaline soil than the control (nontransformed) rice
plants. The higher tolerance of transgenic rice plants to
Fe deficiency was closely associated with the greater
levels of NAAT activity, root secretion of MAs and
shoot and grain yield. This excellent work by Mori's
research group represents a breakthrough in the plant
nutrition research, and will result in profound impacts
on developing Fe deficiency-tolerant cereals for large
areas of Fe-deficient calcareous soils, and may greatly
contributing to food security in the world.
The Fe concentration in grain of the transgenic rice
developed by Takahashi et aI. (2001) was not stud-
ied, and it would be interesting to know the levels
of bioavailable Fe in the grain of the transgenic rice
with respect to their possible contribution to improv-
ing the Fe nutrition of people. High Fe concentrations
in grains can be achieved by manipUlating or adding
the genes responsible for the synthesis of the Fe stor-
age proteins (i.e., ferritin), as has been successfully

demonstrated by Goto et al. (1999). Goto et al. (1999)
were able to transform rice plants using soybean fer-
ritin genes. Ferritin is a major Fe storage protein in
plants. The ferritin genes were expressed in the en-
dosperm of rice by including an endosperm promoter
gene. This resulted in about a 3-fold increase in Fe
concentration in the grain when compared to the un-
transformed rice. They speculated that the Fe content
in a meal-size portion of the transgenic rice would
be sufficient to provide 30-50% of the daily adult
Fe requirement. However, the proposed contribution
of the Fe present in transgenic rice grain to human
nutrition is questioned, because the bioavailability of
ferritin-Fe from plant sources to humans is possibly
very low (Graham et al., 2001). Nevertheless, the work
of Goto et ai. (1999) opens an exciting research area
to transform various food crops using the same or sim-
ilar molecular approach and other potential Fe storage
proteins or metabolites.
At present, there is very little adaptive research
performed under field conditions such as studying the
role of the genes affecting synthesis of MAs, organic
acids, and storage proteins, in agricultural soils hav-
ing limited availability or excessive levels of mineral
nutrients. This will be an important research priority
of plant nutrition in the future. In a recent study, it
was shown that the transgenic plants overexpressing
ferritin and accumulating high levels of Fe in tissue
under controlled environmental conditions were not
superior to the non-transformed plants regarding the
accumulation of Fe in tissue, when grown in dif-
ferent soil conditions in greenhouse (Vansuyt et aI.,
2000). This result indicates that integrating molecular
biology-based research with the knowledge provided
by plant nutrition research is essential to better under-
stand and characterize nutrient uptake and transport in
transgenic plants.
Therefore, integrated research between molecu-
lar biology and plant nutrition will advance research
activities concerning the development of new plant
genotypes with (i) greater efficiency for nutrient ac-
quisition from soils or efficient utilization of nutrients
within plants, (ii) enhanced adaptation ability to acid
or metal-contaminated soils, and (iii) increased seed
density of micronutrients and proteins. Progress in
these research areas will greatly contribute to the
alleviation of food needs and will improve the nutri-
tionally well-being and health of humans.
19
Conclusions
Plant nutrition-based research activities are indispens-
able in meeting food security needs in the 21st century
(Gruhn et aI., 2000; Loneragen, 1997; Marschner,
1995). The fact that at least 60% of the presently cul-
tivated soils globally have several mineral problems,
like toxicities of AI, Mn and Na, and deficiencies of
N, P, K, S, Fe and Zn, makes plant nutrition-based re-
search a major promising research area needed to meet
the demand for massive increases in food production
required for the growing world population. One of the
high priority objectives of plant-nutrition research will
be ensuring a long-term sustainable nutrient manage-
ment system for crop production, and developing more
efficient mineral nutrient uptake by crop plants and im-
proving intra and intercellular use of nutrients without
detrimentally affecting the environment. Creation of a
synergy between plant nutrition and plant molecular
genetics is required to assure a rapid progress towards
alleviating food insecurity issues in the 21st century.
Acknowledgements
This work is dedicated to the memory of the late
Prof. Dr. Drs. h. c. Horst Marschner. I thank Prof.
Dr. Ross M. Welch (Cornell University) and Prof. Dr.
Zed Rengel (Western Australia University) for valu-
able comments on the manuscript and correction of the
English text.
References
Alexandratos N 1995 World Agriculture Towards 2010: An FAO
Study. John Wiley and Sons, Chichester, West Sussex.
Alvey S, Bagayoko M, Neumann G, Buerkert A 2001 CerealJlegume
rotations affect chemical properties and biological activities in
two West African soils. Plant Soil 231, 45-54.
Aniol A M 1991 Genetics of acid tolerant plant. In Plant-Soil
Interactions at Low pH. Eds. R J Wright, V C Baligar and
R P Murrmann. pp. 1007-1017. Kluwer Academic Publisher,
Dordrecht.
Banziger M and Long J 2000. The potential for increasing the iron
and zinc density of maize through plant-breeding. Food Nutr.
Bull. 21, 397-400.
Bates T R and Lynch J P 2000 The efficiency of Arabidopsis thali-
ana (Brassicaceae) root hairs in phosphorus acquisition. Am. J.
Bot. 87, 964-970.
Batjes N H 1997 A world data set of derived soil properties by FAO-
UNESCO soil unit for global modelling. Soil Use Manage. 13,
9-16.
Becker M and Johnson D E 1999 The role of legume fallows in
intensified upland rice-based systems of West Africa. Nutr. Cycl.
Agroecosyst. 53,71-81.
20
Beebe S, Gonzalez A V and Rengifo J 2000 Research on trace
minerals in the common bean. Food Nutr. Bull. 21,387-391.
Bekunda M A, Bationo A and Ssali H 1997 Soil fertility manage-
ment in Africa: a review of selected research trials. In Replen-
ishing Soil in Africa. pp. 63-79. SSSA Special Books, No. 51,
Madison, WI.
Boddey R M, De Moraes S A, Alves B J R and Urquiaga U 1997
The contribution of biological nitrogen fixation for sustainable
agricultural systems in the tropics. Soil BioI. Biochem. 29, 787-
799.
Borlaug N E and Dowswell C R 1993 Fertilizer: To nourish infertile
soil that feeds a fertile population that crowds a fragile world.
Fertil. News 387,11-20.
Bouis H, Graham R D and Welch R M 2000 The Consultative Group
on International Agricultural Research (CGlAR) Micronutrients
Project: justification and objectives. Food Nutr. Bull. 21, 374-
381.
Buerkert A, Bationo A and Dosso K 2000 Mechanism of residue
mulch-induced cereal growth increases in West Africa. Soil Sci.
Soc. Am. J. 64,346-358.
Bumb BLand Baanante C A 1996 The role of fertilizer in sustain-
ing food security and protecting the environment to 2020, 2020
Vision Discussion Paper 17. International Food Policy Research
Institute, Washington, DC.
Byrnes B H and Bumb B L 1998 Population growth, food produc-
tion and nutrient requirements. In Journal of Crop Production.
Ed. Z. Rengel. pp. 1-27. The Haworth Press, New York.
Cakmak I, Yilmaz A, Ekiz H, Torun B, Erenog1u B. and Braun H J
1996a Zinc deficiency as a critical nutritional problem in wheat
production in Central Anatolia. Plant Soil 180, 165-172.
Cakmak I, Sari N, Marschner H, Ekiz H, Kalayci M, Yilmaz A and
Braun H J 1996b Phytosiderophore release in bread and durum
wheat genotypes differing in zinc efficiency. Plant Soil 180, 183-
189.
Cakmak I, Derici R, Torun B, Tolay I, Braun H J and Schlegel R
1997 Role of rye chromosomes in improvement of zinc efficiency
in wheat and triticale. Plant Soil 196, 249-253.
Cakmak I, Torun B, Erenoglu B, Oztiirk L, Marschner H, Kalayci
M, Ekiz H and Yilmaz A 1998 Morphological and physiological
differences in cereals in response to zinc deficiency. Euphytica
100,349-357.
Cakmak I, Kalayci M, Ekiz H, Braun H J and Yilmaz A 1999a Zinc
deficiency as a practical problem in plant and human nutrition in
Turkey: A NATO-Science for Stability Project. Field Crops Res.
60, 175-188.
Cakmak I, Cakmak 0, Eker S, Ozdemir A, Watanabe N and Braun
H J 1999b Expression of high zinc efficiency of Aegi/ops tauschii
and Triticum monococcum in synthetic hexaploid wheats. Plant
Soil 215, 203-209.
Cakmak I, Ozkan H, Braun H J, Welch R M and Romheld V 2000
Zinc and iron concentrations in seeds of wild, primitive and
modern wheats. Food Nutr. Bull. 21,401-403.
Carver B F and Ownby J D 1995 Acid soil tolerance in wheat. Adv.
Agron. 54,117-173.
Chen Y and Barak P 1982 Iron nutrition of plants in calcareous soils.
Adv. Agron. 35, 217-240.
Crawford N M and Glass ADM 1998 Molecular and physiological
aspects of nitrate uptake in plants. Trends Plant Sci. 10,389-395.
Dakora F D and Keya S 0 1997 Contribution of legume nitrogen
fixation to sustainable agriculture in Sub-Saharan Africa. Soil
BioI. Biochem. 29, 809-817.
Dalal R C 1977 Soil organic phosphorus. Adv. Agron. 29, 83-117.
Daram P, Brunner S, Amrhein N and Bucher M 1998 Functional
analysis and cell-specific expression of a phosphate transporter
from tomato. Planta 206, 225-233.
De la Fuente M J, Ramirez-Rodriguez V, Cabrera-Ponce J L and
Herrera-Estrella L 1997 Aluminum tolerance in transgenic plants
by alteration of citrate synthesis. Science 276, 1566-1568.
De 1a Fuente J M and Herrera-Estrella L 1999 Advances in the
understanding of aluminium toxicity and the development of
aluminium-tolerant transgenic plants Adv. Agron. 66, 103-120.
De Pauw E F 1994 The management of acid soils in Africa. Outlook
Agric. 23, 11-16.
Delhaize E, Hebb D M and Ryan P R 2001 Expression of a
Pseudomonas aeruginosa citrate synthase gene in tobacco is not
associated with either enhanced citrate accumulation or efflux.
Plant Physiol. 125,2059-2067.
Dinkelaker B, Romheld V and Marschner H 1989 Citric acid ex-
cretion and precipitation of calcium citrate in the rhizosphere of
white lupin (Lupinus albus L.). Plant Cell Environ. 12, 285-292.
Dodd J C 2000 The role of arbuscular mycorrhizal fungi in agro-
and natural ecosystems. Outlook Agricult. 29, 55-62.
Dyson T 1999 World food trends and prospects to 2025. Proc. Natl.
Acad. Sci. USA 96, 5929-5936.
Eghball B and Gilley J E 1999 Phosphorus and nitrogen in runoff
following beef cattle manure or compost application. J. Environ.
qual. 28, 1201-1210.
Eghball B, Binford G D and Baltensperger D D 1996 Phosphorus
movement and adsorption in a soil receiving long-term manure
and fertilizer application. J. Environ. Qual. 25, 1339-1343.
Eide D, Broderius M, Fett J, and Guerinot M 1996 A novel iron-
regulated metal transporter from plants identified by functional
expression in yeast. Proc. Natl. Acad. Sci. USA 93, 5624-5628.
Ellsworth J W, Jolley V D, Nualand D S and Blaylock A D 1998 Use
of hydrogen release or a combination of hydrogen release and
iron reduction for selecting iron-efficient dry bean and soybean
cultivars. J. Plant Nutr. 12, 2639-2651.
Ekiz H, Bagci S A, Kiral A S, Eker S, Giiltekin I, Alkan A and
Cakmak I 1998 Effects of zinc fertilization and irrigation on
grain yield and zinc concentration of various cereals grown in
zinc-deficient calcareous soils. J. Plant Nutr. 21, 2245-2256.
Erenoglu B, Cakmak I, Marschner H, Romheld V, Eker S, Daghan
H, Kalayci M and Ekiz H 1996 Phytosiderophore release does
not correlate well with zinc efficiency in different bread wheat
genotypes. J. Plant Nutr. 19, 1569-1580.
Erickson H, Keller M and Davidson E A 2001 Nitrogen oxide
fluxes and nitrogen cycling during postagricultural succession
and forest fertilization in the humid tropics. Ecosystems 4,
67-84.
Evenson R E 1999 Global and local implications of biotechnology
and climate change for future food supplies. Proc. Natl. Acad.
Sci. USA 96, 5921-5928.
Eyupoglu F, Kurucu N, Sanisa U 1993 Status of Plant Available
Micronutrients in Turkish Soils. In Annual Report. Report No.
R-118. Soil and Fertilizer Research Institute, Ankara, pp. 25-32.
FAO 1996 Food, Agriculture and Food Security: Developments
Since the World Food Conference and Prospects, World Food
Summit Technical Background Document No. 1 Rome.
FAO 1999 The State of Food Insecurity in the World. FAO, Rome.
FAO 2000 Fertilizer requirements in 2015 and 2030. FAO, Rome.
Fairbanks D J 2000 Development of genetic resistance to iron-
deficiency chlorosis in soybean. J. Plant Nutr. 23, 1903-1913.
Gahoonia T S and Nielsen EN 1998 Direct evidence on participa-
tion of root hairs in phosphorus (32 p) uptake from soil. Plant Soil
198, 147-152.

Gahoonia T S, Asmar F, Giese H, Gissel-Nielsen and Nielsen N
E 2000 Root-released organic acids and phosphorus uptake of
two barley cultivars in laboratory and field experiments. Eur. 1.
Agron.12,281-289.
Gaume A, Machler F, De Leon C, Narro L and Frossard E 2001
Low-P tolerance by maize (Zea mays L.) genotypes: signific-
ance of root growth, and organic acids and acid phosphatase root
exudation. Plant Soil 228, 253-264.
Georgatsou E and Alexandraki D 1994 Two distinctly regulated
genes are required for ferric reduction, the first step of iron
uptake in Sacchoromyces cerevisiae. Mol. Cell BioI. 14, 3065-
3073.
George E and Marschner H 1994 Nutrient and water uptake by roots
of forest trees. Z. Pflanzenerniihr. Bodenkd. 159, 11-21.
Goto F, Yoshihara T, Shigemoto N, Toki S, Takaiwa F 1999 Iron
fortification of rice seed by the soybean ferritin gene. Nat.
Biotechnol. 17, 282-286.
Graham R D and Welch R M 1996 Breeding for staple-food crops
with high micronutrient density. Agricultural Strategies for Mi-
cronutrients. Working Paper No.3. International Food Policy
Research Institute, Washington, DC.
Graham R D and Rengel Z 1993 Genotypic variation in zinc uptake
and utilization by plants. In Zinc in Soils and Plants, Ed. A D
Robson. pp. 107-118. Kluwer Academic Publishers, Dordrecht.
Graham R D, Ascher J S and Hynes S C 1992 Selecting zinc-
efficient cereal genotypes for soils of low zinc status. Plant Soil
146,241-250.
Graham R D, Senadhira D, Beebe S E, Iglesias C and Ortiz-
Monasterio I 1999 Breeding for micronutrients density in edible
portions of staple food crops: conventional approaches. Field
Crops Res. 60, 57-80.
Graham R D, Welch R M and Bouis H E 2001 Addressing micronu-
trient malnutrition through enhancing the nutritional quality of
staple foods: Principles, perspectives and knowledge gaps. Adv.
Agron. 70, 77-142.
Gregory G B, Senadhira D, Htut H and Graham R D 2000 Breeding
for trace mineral density in rice. Food Nutr. Bull. 21, 382-386.
Grotz N, Fox T, Connolly E, Park W, Guerinot ML and Eide D
1998 Identification of a family of zinc transporter genes from
Arabidopsis that respond to zinc deficiency. Proc. Natl. Acad.
Sci. USA 12,7220--7224.
Gruhn P, Goletti F and Yudelman M 2000 Integrated nutrient
management, soil fertility, and sustainable agriCUlture: Current
issues and future challenges. Food, Agriculture, and the Envir-
onment Discussion Paper 32. International Food Policy Research
Institute Washington, DC.
Gupta U C 1991 Iron status of crops in Prince Edward Island and
effect of soil pH on plant iron concentration. Can. J. Soil Sci. 71,
197-202.
Harlan J R 1981 The early history of wheat: earliest traces to the
sack of Rome. In Wheat Science-Today and Tomorrow. Eds. L T
Evans and J Peacock. Cambridge University Press, Cambridge.
Hayes J E, Richardson A E and Simpson R J 1999 Phytase and acid
phosphatase activities in extracts from roots of temperate pasture
grass and legume seedlings Aust. J. Plant Physiol. 26, 801-809.
Haynes R J and Mokolobate M S 2001 Amelioration of Al toxicity
and P deficiency in acid soils by additions of organic residues: a
critical review of the phenomenon and the mechanisms involved.
Nutr. Cycl. Agroecosyst. 59,47--63.
Hell R and Hillebrand H 2001 Plant concepts for mineral acquisition
and allocation. Curr. Opin. Biotechnol. 12,161-168.
Henao J and Baanante C 1999 Nutrient depletion in the agricultural
soils of Africa. 2020 Vision Brief 62. International Food Policy
Research Institute, Washington, DC.
21
Herrera-Estrella, L 1999 Transgenic plants for tropical regions:
Some considerations about their development and their transfer
to the small farmer. Proc. Natl. Acad. Sci. USA 96,5978-5981.
Higuchi K, Suzuki K, Nakanishi Y, Yamaguchi H, Nishizawa N K
and Mori S 1999 Cloning ofnicotianarnine synthase genes, novel
genes involved in the biosynthesis of phytosiderophores. Plant
Physiol. 119,471-479.
Higuchi K, Watanabe S., Takahashi M, Kawasaki S, Nakanishi
H, Nishizawa N K and Mori S 2001 Nicotianarnine synthase
gene expression differs in barley and rice under Fe-deficient
conditions. PlantJ. 25, 159-167.
Hirsch R E and Susman M R 1999 Improving nutrient capture from
soil by the genetic manipulation of crop. Trends Biotechnol. 17,
356-361.
Hoffland E, Findenegg GRand Nelemans J A 1989 Solubilisation
of rock phosphate by rape. 2. Local root exudation of organic
acids as a response to P starvation. Plant Soil 113, 161-165.
Holford I C R 1997 Soil phosphorus, its measurements and its
uptake by plants. Aust. J. Soil. Res. 35, 227-239.
Hopkins B G, Whituey D A, Lamond R E and Jolley V D 1997
Phytosiderophore release by sorghum, wheat and corn under zinc
deficiency. J. Plant Nutr. 21,2623-2637.
Horst W J 1995 The role of apoplast in aluminium toxicity and res-
istance in higher plants: a review. Z. Pflanzenerniihr. Bodenkd.
158,419-428.
Horst W J and Hardter R 1994 Rotation of maize with cowpea
improves yield and nutrient use of maize compared to maize
monocropping in an alfisol in the northern Guinea Savanna of
Ghana. Plant Soil 160, 171-183.
Hossain M and Singh V P 2000 Fertilizer use in Asian agriculture:
implications for sustaining food security and the environment.
Nutr. Cycl. Agroecosyst. 57, 155-169.
Howarth R W 1998 An assessment of human influences on fluxes
of nitrogen from the terrestrial landscape to the estuaries and
continental shelves of the North Atlantic Ocean. Nutr. Cycl.
Agroecosyst. 52, 213-223.
Hue N V and Amien I 1989 Aluminium detoxification with green
manures. Commun. Soil Sci. Plant Anal. 20,1499-1511.
Hue N V, Ikawa H and Silva J A 1994 Increasing plant available
phosphorus in an Ultisol with a yard-waste compost. Commun.
Soil Sci. Plant Anal. 25, 3291-3303.
Humphreys E, Chalk P M, Muirhead W A, Melhuish F M and White
R J G 1987 Effects of time of application of urea application on
combine-sown Calrose rice in south-east Australia. m. Fertilizer
nitrogen recovery, efficiency of fertilization and soil nitrogen
supply. Aust. J. Agric. Res. 38, 129-138.
Hungria M and Vargas MAT 2000 Environmental factors affecting
N2 fixation in grain legumes in the tropics, with an emphasis on
Brazil. Field Crops Res. 65, 151-164.
Isherwood K F 2000 The state of the fertilizer industry: past,
present and future. Annual General Meeting. 68th !FA Annual
Conference 22-25 May 2000, Oslo, Norway.
Iyamuremye E and Dick R P 1996 Organic amendments and
phosphorus sorption by soils. Adv. Agron. 56,139-185.
Jolley V D, Fairbanks D J, Stevens W B, Terry R E and Orf J H
1992 Using root iron-reduction capacity for genotype evaluation
of iron efficiency in soybean. J. Plant Nutr. 15,1679-1690.
Jolley V D, Cook K A, Hansen N C and Stevens W B 1996 Plant
physiological responses for genotypic evaluation of iron effi-
ciency in Strategy I and Strategy II plants: A review. J. Plant
Nutr. 19, 1241-1255.
Kalayci M, Torun B, Eker S, Aydin M, Ozturk L and Cakmak I
1999 Grain yield, zinc efficiency and zinc concentration of wheat
22
cultivars grown in a zinc-deficient calcareous soil in field and
greenhouse. Field Crops Res. 63, 87-98.
Kanazawa K, Higuchi K, Nishizawa N K, Fushiya S, Chino M,
and Mori S 1994 Nicotianamine aminotransferase activities are
correlated to the phytosiderophore secretions under Fe-deficient
conditions in Gramineae. J. Exp. Bot. 45, 1903-1906.
Kawai S, Takagi S and Sato Y 1988 Mugineic acid-family phyto-
siderophores in root-secretions of barley, com and sorghum
varieties. J. Plant Nutr. 11,633-642.
Kochian L V 1995 Cellular mechanisms of aluminium toxicity and
resistance in plants. Annu. Rev. Plant Physiol. Plant Mol. BioI.
46, 237-260.
Koyama H, Kawamura A, Kihara T, Hara T, Takita E and Shibata
D 2000 Overexpression of mitochondrial citrate synthase in Ara-
bidopsis thaliana improved growth on a phosphorus-limited soil.
Plant Cell Physiol. 9, 1030-1037.
Kroeze C and Seitzinger S P 1998 Nitrogen inputs to rivers, estu-
aries and continental shelves and related nitrous oxide emissions
in 1990 and 2050: a global model. Nutr. Cycl. Agroecosyst. 52,
145-212.
Li X L, Marschner H and George E 1991 Acquisition of phos-
phorus and copper by VA-mycorrhizal hyphae and root-to-shoot
transport in white clover. Plant Soil 136, 49-57.
Lin S F, Grandt D, Cianzio S and Shoemaker R 2000 Molecular
characterisation of iron deficiency chlorosis in soybean. J. Plant
Nutr.23,1929-1939.
Loneragen J P 1997 Plant nutrition in 20th and perspectives for the
21st century. Plant Soil 196, 163-174.
Lopez-Bucio J, Nieto-Jacobo M F, Ramirez-Rodriguez V and
Herrera-Estrella L 2000a Organic acid metabolism in plants:
from adaptive physiology to transgenic varieties for cultivation
in extreme soils. Plant Sci. 160, 1-13.
Lopez-Bucio J, Martinez de la Vega 0, Guevara-Garcia A and
Herrera-Estrella L 2000b Enhanced phosphorus uptake in trans-
genic tobacco plants that overproduce citrate. Nat. Biotechnol.
18,450-453.
Ma J F, Zheng S J and Matsumoto H 1997 Specific secretion of citric
acid induced by Al stress in Cassia tora L. Plant Cell Physiol. 38,
1019-1025.
Ma J F, Ryan P R and Delhaize E 2001 Aluminium tolerance in
plants and the complexing role of organic acids. Trends Plant
Sci. 6, 273-278.
Maene L M 2000 Efficient fertilizer use and its role in increasing
food production and protecting the environment. 6th AFA Inter-
national Annual Conference 31 J anuary-3 February 2000, Cairo,
Egypt.
Marschner H 1991 Mechanisms of adaptation of plants to acid soils.
Plant Soil 134, 1-20.
Marschner H 1995 Mineral Nutrition of Higher Plants. Academic
Press. London.
Marschner H 1998 Soil-root interface: biological and biochemical
process. In Soil Chemistry and Ecosystem Health. pp. 191-231.
SSSA Special Publication No 52. Madison, WI.
Marschner H and Riimheld V 1994 Strategies of plants for acquisi-
tion of iron. Plant Soil 165, 261-274.
Marschner H, Riimheld V and Kissel M 1986 Different strategies in
higher plants in mobilization and uptake of iron. J. Plant Nutr. 9,
695-713.
McTaggart I P, Clayton H, Parker J, Swan L and Smith K A 1997
Nitrous oxide emissions from grassland and spring barley, fol-
lowing N fertilizer application with and without nitrification
inhibitors. BioI. Fertil. Soils 25, 261-268.
McGonigle T P and Miller M H 1999 Winter survival of extraradical
hyphae and spores of arbuscular mycorrhizal fungi in the field.
Appl. Soil Ecol. 12,41-50.
Mitsukowa N, Okumura S, Shirano Y, Sato S, Kato T, Harashima S
and Shibata D 1997 Overexpression of an Arabidopsis thaliana
high-affinity phosphate transporter gene in tobacco cultured cells
enhances cell growth under phosphate-limited conditions. Proc.
Natl. Acad. Sci. USA 94, 7098-7102.
Monasterio I and Graham, R D 2000 Breeding for trace mineral in
wheat. Food Nutr. Bull. 21, 392-396.
Mori S 1999 Iron acquisition by plants. Curro Opin. Plant BioI. 2,
250-253.
Mosier A R 1998 Soil processes and global change. BioI. Fertil.
Soils 27, 221-229.
Mosier A R, Bronson K F, Freney J Rand Keerthisinghe D G 1994
Use of nitrification inhibitors to reduce nitrous oxide emission
from urea fertilized soils. In CH4 and N20: Global Emissions
and Controls from Rice Fields and Other Agricultural and Indus-
trial Sources. Eds. KMinami, A Mosier and RSass. pp. 197-207.
Yokendo Publishers, Tokyo.
Mosier A R, Bleken M A, Chaiwanakupt P, Ellis E C, Freney J R,
Howarth R B, Matson P A, Minami K, Naylor R, Weeks K Nand
Zhu Z L 2001 Policy implications of human-accelerated nitrogen
cycling. Biogeochem. 52,281-320.
Noble A D, Zenneck I and Randall P J 1996 Leaf litter ash alkalinity
and neutralisation of soil acidity. Plant Soil 179,293-302.
Nesbitt M and Samuel D 1995 Wheat domestication: archaeobotan-
ical evidence. Science 279, 1433-1433.
Oki H, Yamaguchi H, Nakanishi Hand Mori S 1999 Introduction
of the reconstructed yeast ferric reductase gene, refrel, into
tobacco. Plant Soil 215, 211-220.
Pellet D M, Grunes D L, Kochian L V 1995 Organic acid exudation
as an aluminium-tolerance mechanism in maize (Zea mays L.).
Planta 196,788-795.
Peterson C J, Johnson V A and Mattern P J 1986 Influence of cul-
tivar and environment on mineral and protein concentrations of
wheat flour, bran and grain. Cereal Chern. 63, 118-186.
Pinstrup-Andersen P 1999 Selected aspects of the future global food
situation. In 25th International Fertilizer Industry Association
Enlarged Council Meeting, Rome, Italy.
Pinstrup-Andersen P, Pandya-Lorch R and Rosegrant M. W 1999
World food propects: Critical issues for the early twenty-first
century. 2020 Vision Food Policy Report, International Food
Policy Research Institute, Washington, DC.
Quedraogo E, Mando A and Zombre N P 2001 Use of compost to
improve soil properties and crop productivity under low input
agricultural system in West Africa. Agric. Ecosyst. Environ. 84,
259-266.
Raghothama K G 1999 Phosphate acquisition. Annu. Rev. Plant
Physiol. Plant Mol. BioI. 50, 665-693.
Raghothama K G 2000 Phosphate transport and signalling. Curro
Opin. Plant BioI. 3, 182-187.
Reddy D D, Rao AS, Reddy K S and Takkar P N 1999 Yield sus-
tainability and phosphorus utilisation in soybean-wheat system
on vertisols in response to integrated use of manure and fertilizer
phosphorus. Field Crops Res. 62, 181-190.
Rengel Z 1999 Physiological mechanisms underlying differen-
tial nutrient efficiency of crop genotypes. In Mineral Nutrition
of Crops: Fundamental Mechanisms and Implications. Ed. Z.
Rengel. pp. 227-265. The Haworth Press, New York.
Rengel Z 200 I Genotypic differences in micronutrient use effi-
ciency in crops. Commun. Soil Sci. Plant Anal. 32, 1163-1186.

Rengel Z and Graham R D 1995 Importance of seed Zn content for
wheat growth on Zn-deficient soil. I. Vegetative growth. Plant
Soil 173, 259-266.
Rengel Z, Romheld V, Marschner H. 1998. Uptake of zinc iron by
wheat genotypes differing in tolerance to zinc deficiency. J. Plant
Physiol. 152,433-438.
Reynolds C S and Davies P S 2001 Sources and bioavailability of
phosphorus fractions in freshwaters: a British perspective. BioI.
Rev. 76, 27-64.
Richardson A E, Hadobas P A and Hayes J E 2001 Extracellular
secretion of Aspergillus phytase from Arabidopsis roots enables
plants to obtain phosphorus from phytate. Plant J. 25, 641-649.
Robinson N J, Sadjuga and Groom Q J 1997 The froh gene fam-
ily from Arabidopsis thaliana: putative iron-chelate reductases.
Plant Soil 196, 245-248.
Robinson N J, Procter C M, Connolly ELand Guerinot M L 1999
A ferric-chelate reductase for iron uptake from soils. Nature 397,
694-697.
Rochester I J, Peoples M B, Hulugalle N R, Gault R R and Con-
stable G A 2000 Using legumes to enhance nitrogen fertility and
improve soil condition in cotton cropping systems. Field Crops
Res. 70, 27-4l.
Romera F J, Alcantara and Guardia M D dela 1991 Characterisation
of the tolerance to iron chlorosis in different peach rootstocks
grown in nutrient solution. II. Iron-stress response mechanism.
Plant Soil 130, 120-124.
Romheld V and Marschner H 1986 Evidence for a specific up-
take system for iron phytosiderophores in roots of grasses. Plant
Physiol. 80, 175-180.
Romheld V and Marschner H 1990 Genotypical differences among
graminaceous species in release of phytosiderophores and uptake
of iron phytosiderophores. Plant Soil 123, 147-153.
Rosegrant M W, Leach N and Gerpacio R V 1999 Alternative future
for world cereal and meat consumption. Proc. Nutr. Soc. 58, 219-
234.
Rosegrant M W, Paisner M S, Meijer Sand Witcover J 2001 2020
Global Food Outlook; Trends, Alternatives, and Choices. A 2020
Vision for Food, Agriculture, and the Environment Initiative.
International Food Policy Research Institute, Washington, DC.
Ryan P R, Delhaize E and Randall P J 1995 Characterisation of AI-
stimulated efflux of malate from the apices of AI-tolerant wheat
roots. Planta 196, 103-110.
Samuelsen A I, Martin R C, Mok D W S and Mok M C 1998 Expres-
sion of the yeast FRE genes in transgenic tobacco. Plant Physiol.
118,51-58.
Scherr S J 1999 Soil degradation, a threat to developing-country
food security by 2020? Food, Agriculture, and the Environ-
mental Discussion Paper 27. International Food Policy Research
Institute. Washington, DC.
Schlegel R and Cakmak I 1997 Physical mapping of rye genes de-
termining micronutritional efficiency in wheat. In Plant Nutrition
for Sustainable Food Production and Environment. Eds. T Ando,
K Fujita, T Mae, H Matsumoto, S Mori and J Sekiya. pp. 287-
288. Developments in Plant and Soil Sciences, Vol. 78. Kluwer
Academic Publishers, Dordrecht.
Schlegel R, Cakmak I, Torun B, Eker S, Tolay I, Ekiz H, Kalayci M
and Braun H J 1998 Screening for zinc efficiency among wheat
relatives and their utilisation for an alien gene transfer. Euphytica
100, 281-286.
Shrimpton R 1993 Zinc deficiency: Is it widespread but under-
recognized? SCN News, 9,23-27.
Sillanpaa M 1990 Micronutrients assessment at the country level:
An international study. FAO Soils Bulletin 63. Food and Agri-
culture Organisation of the United Nations, Rome.
23
Sillanpaa M and Vlek P L G 1985 Micronutrients and the agroe-
cology of tropical and Mediterranean regions. Fertil. Res. 7,
151-167.
Singh G B 1998 Fertilizer use and sustainability of Indian agricul-
ture. In Nutrient Management for Sustainable Crop Production
in Asia. Eds. A E Johnston and J K Syers. pp. 279-289. CAB
International.
Smith F W 2001 Sulphur and phosphorus transport systems in
plants. Plant Soil 232, 109-118.
Smith K A, McTaggart I P and Tsuruta H 1997. Emissions of N20
and NO associated with nitrogen fertilization in intensive agri-
culture, and the potential for mitigation. Soil Use Manage. 13,
296-304.
Stearman R, Yaun D S, Yamaguchi-Iwai Y, Klausner R D and
Dancis A 1996 A permease-oxidase complex involved in high-
affinity iron uptake in yeast. Science 271, 1552-1557.
Takagi S, Nomoto K and Takemoto T 1984 Physiological aspect
of mugineic acid, a possible phytosiderophore of graminaceous
plants. J. Plant Nutr. 7,469-477.
Takahashi M, Yamaguchi H, Nakanishi H, Shioiri T, Nishizawa N
K and Mori S 1999 Cloning two genes for nicotianamine amino-
transferase, a critical enzyme in iron acquisition (strategy II) in
graminaceous plants. Plant Physiol. 121,947-956.
Takahashi M, Nakanishi H, Kawasaki S, Nishizawa N K and Mori
S 2001 Enhanced tolerance of rice to low iron availability in al-
kaline soils using barley nicotianamine aminotransferase genes.
Nat. Biotechnol. 19,466-469.
Tandon H L S 1995 Major nutritional constraints to crop production
and the soil fertility management strategies in different agrocli-
matic regions of Asia. In Proceedings of the International Potash
Institute Colloquium on Potassium in Asia: Balanced Fertiliza-
tion to Increase and Sustain Agricultural Production. pp. 43-72.
Chiang Mai, Thailand, 21-24 February, 1995. International
Potash Institute, Basel.
Tandon H L S 1998 Use of external inputs and the state of efficiency
of plant nutrient supplies in irrigated cropping systems in Ut-
tar Pradesh, India. In Proceedings of the IFPRIIFAO Workshop
on Soil Fertility, Plant Nutrient Management, and Sustainable
Agriculture: The future Through 2020. Eds. P Gruhn, F Go-
letti, and R N Roy. International Food Policy Research Institute,
Washington, DC and Food and Agriculture Organization, Rome.
Tillman D 1999 Global environmental impacts of agricultural ex-
pansion: the need for sustainable and efficient practices. Proc.
Natl. Acad. Sci USA 96, 5995-6000.
Timsina J and Connor D J 2001 Productivity and management of
rice-wheat cropping systems: issues and challenges. Field Crops
Res. 69, 93-132.
Tyler G and Strom L 1995 Differing organic acid exudation patterns
explain calcifuge and acidifuge behaviour of plants. Ann. Bot.
75,75-78.
UNDP 1999 Human Development Report. Oxford University Press,
Cary, NC.
Van Kessel C and Hartley C 2000 Agricultural management of grain
legumes: has it led to an increase in nitrogen fixation? Field
Crops Res. 65, 165-18l.
Vansuyt G, Mench M, Briat J F 2000 Soil-dependent variability
of leaf iron accumulation in transgenic tobacco overexpressing
ferritin. Plant Physiol. Biochem. 6, 499-506.
Veldkamp E and Keller M 1997 Fertilizer-induced nitric oxide
emissions from agricultural soils. Nutr. Cycl. Agroecosyst. 48,
69-77.
Von Uexkiill H R and Mutuert W 1995 Global extent, development
and economic impact of acid soils. In Plant-Soil Interactions at
low pH: Principles and Management. Eds. R A Date, N J Grun-
24
don, G E Rayment and M E Probert pp, 5-19. Kluwer Academic
Publishers, Dordrecht
von Wiren N, Gazzarrini S, Gojon A and Frommer W B 2000a The
molecular physiology of ammonium uptake and retrieval. Curro
Opin. Plant BioI. 3,254-261.
Von Wiren N, Lauter F R, Ninnemann 0, Gillisen B, Walch-Liu
P, Engels C, Jost W and Frommer W B 2000b Differential
regulation of three functional ammonium transporter genes by
nitrogen in root hairs and by light in leaves of tomato. Plant 1.
21, 167-175.
Vose P B 1982 Iron nutrition in plants: a world overview. J. Plant
Nutr. 5, 238-249.
Waddington S R and Karigwindi J 2001 Productivity and profit-
ability of maize + groundnut rotations compared with continu-
ous maize on smallholder farms in Zimbabwe. Exp. Agric. 37,
83-98.
Wei L C, Loeppert R H and Ocumpaugh W R 1997 Iron deficiency
stress response in Fe deficiency resistant and susceptable subter-
ranean clover: importance of induced H+ release J. Exp. Bot 48,
329-346.
Welch R M 1995 Micronutrient nutrition of plants. Crit Rev. Plant
Sci. 14, 49-82.
Welch R M 1999 Importance of seed mineral nutrient reserves in
crop growth and development In Mineral Nutrition of Crops:
Fundamental Mechanisms and Implications. Ed. Z. Rengel. pp.
205-226. Haworth Press, New York.
Welch R M and Graham R D 1999 A new paradigm for world
agriculture: meeting human needs. Productive, sustainable, nu-
tritious. Field Crops Res. 60, 1-10.
Welch R M and Graham R D 2000 A new paradigm for world
agriculture: productive, sustainable, nutritious, healthful food
systems. Food Nutr. Bull. 21, 361-366.
Williams L E and Miller A J 2001 Transporters responsible for the
uptake and partitioning of nitrogenous solutes. Annu. Rev. Plant
Physiol. Plant Mol. BioI. 52, 659-688.
Williams T 0, Powell J M and Fernandez-Rivera S 1995 Soil fer-
tility maintenance and food crop production in semi-arid West
Africa: Is reliance on manure a sustainable strategy? Outlook
Agric. 24,43-47.
Wong M T C, Gibbs P, Nortcliff S and Swift R S 1999 Measurement
of the acid neutralising capacity of agroforestry tree promising
added to tropical soils. J. Agric. Sci. 134, 269-276.
Yadav R L 1998 Factor productivity trends in a rice-wheat cropping
system under long-term use of chemical fertilizers. Exp. Agric.
34, 1-18.
Yadav R L, Dwivedi B S and Pandey P S 2000a Rice-wheat cropping
system: assessment of sustainability under green manuring and
chemical fertilizer inputs. Field Crops Res. 65, 15-30.
Yadav R L, Dwivedi B S, Prasad K, Tomar 0 K, Shurpali N J and
Pandey P S 2000b Yield trends, and changes in soil organic-
C and available NPK in a long-term rice-wheat system under
integrated use of manures and fertilizers. Field Crops Res. 68,
219-246.
Yang Z M, Sivagum M, Horst W J and Matsumoto H 2000 Alu-
minium tolerance is achieved by exudation of citric acid from
roots of soybean (Glycine max). Physiol. Plant 110,72-77.
Yilmaz A, Ekiz H, Tomn B, Giiltekin I, Karanlik S, Bagci S A
and Cakmak I 1997 Effect of different zinc application methods
on grain yield and zinc concentration in wheat grown on zinc-
deficient calcareous soils in Central Anatolia. J. Plant Nutr. 20,
461-471.
Yip R 1994 Iron deficiency: contemporary scientific issues and
international programmatic approaches. J. Nutr. 124, 1479-1490.
Zhao H and Eide D 1996a The yeast ZRTl gene encodes the zinc
transporter protein of a high-affinity uptake system induced by
zinc limitation. ProC. Natl. Acad. Sci. USA 93, 2454-2458.
Zhao H and Eide D 1996b The ZRT2 gene encodes the low affinity
zinc transporter in Saccharomyces cerevisiae. J. BioI. Chern. 271,
23203-23210.
 ... Plant and Soil 247: 25-40, 2002.
25
" © 2002 Kluwer Academic Publishers.

Access and excess problems in plant nutrition

Meine van Noordwijkl & Georg Cadisch2

1International Centre for Research in Agroforestry (ICRAF), Bogor, Indonesia, P.O.Box 161, Bogor, 16001, In-
donesia 2Department of Agricultural Sciences, Imperial College (RYe), University of London, RYe, Kent, TN25
5AB, UK

Key words: agroforestry, competition, complementarity, rhizosphere modification, safety net, simulation model,
WaNuLCAS

Abstract
As plant nutrition issues are redefined by society, new applications emerge for a basic understanding of nutrient use
efficiency in soil-plant processes to avoid excess on rich soils as commonly found in the temperate zone and make
the best of it under access-limited conditions common in the tropics. The main challenge of plant nutrition may be
to increase the width of the domain between the access and excess frontiers, rather than to define a single 'economic
optimum' point. Two approaches are discussed to widen this domain: the technical paradigm of precision farming
and the ecological analogue approach based on filter functions and complementarity of components in mixed plant
systems. Current understanding of plant nutrition, largely focused on monocultural situations, needs to be aug-
mented by the interactions that occur in more complex systems, including agroforestry and intercropping as these
may form part of the answer in both the excess and shortage type of situation. Simulations with the WaNuLCAS
model to explore the concepts of a 'safety-net' for mobile nutrients by deep rooted plants suggested a limited but
real opportunity to intercept nutrients on their way out of the system and thus increase nutrient use-efficiency at
the system level. The impacts of rhizosphere modification to mobilize nutrients in mixed-species systems were
shown to depend on the degree of synlocation of roots of the various plant components, as well as on the long-term
replenishment of the nutrient resources accessed. In conclusion, the concepts and tools to help farmers navigate
between the scylla of access and the charibdis of excess problems in plant nutrition certainly exist, but their use
requires an appreciation of the site-specific interactions and various levels of internal regulation, rather than a
reliance alone on genetic modification of plants aimed at transferring specific mechanisms out of context.

Introduction: threats to sustainability
Sustainability of farming and thus of agro-ecosystems
depends on the ability of farmers to overcome current
and future threats to a continuation of their enterprise
in some form or other. These threats can derive from
loss of on-site productivity, from 'angry neighbours'
who no longer accept the lateral flows through air
or water of elements and pesticides coming from the
farm, from 'worried customers' who do not trust the
quality of the products or don't agree with the produc-
tion conditions, or from 'regulatory bureaucrats' in a
policy frame that tries to control the activity of farm-
ers. Plant nutrition research has traditionally focused
on the first of these threats, and has helped to develop
plant and soil management schemes that provide for
adequate nutrient supply to the current crop without
unduly mining soil resources. A build-up of soil nu-
trient stocks, some not directly available to common
crops, was seen as an unavoidable side-effect of im-
proved crop nutrition and use efficiencies of 30-40%
for fertilizer N characterize the main grain production
systems of the world (Doberman and Cassman, 2001).
The apparent success of this type of plant nutrition in
the intensively used agricultural lands of the temperate
zone, as well as in specific areas in the tropics, has
given rise to increased fluxes into the environment and
thus to the angry neighbour, worried customer and
regulatory bureaucrat type of threat to sustainability
of current farming styles. In the Netherlands, the coun-
try with the highest nitrogen surplus per ha (Smaling
et aI., 1999), new legislation requires farmers to keep
26
decreased tolerance for
consequences of excess
uncontrollable / .
losscs ~"...
. ....• ~~~
troll/;erB bcttcrcontrol
"';~r,"~ ~:xC<ss or ~
tr<mtier"I"~~' over supply
A.~6~~tri~ bc~ter intemal.~·e~u-
~ ta'A'~~(j9::> ~ latlOn and buttenng
• ~ ~~~:~~~~ 1,'
Ul'eatl'l' ~
;(/ cllccti\'c saictyncts and
tiltcrs.for,lateral t10ws.
"'., lIutrlollls
~/
reslhence
Figure 1. Domain for balanced nutrition ('OK') between frontier A
that separates it from access limitations and frontier B that separates
it from excess problems; various factors squeeze or increase the
width of the OK domain.
track of a farm level nutrient budget and to pay for
surpluses that exceed a permissible level; as payments
are based on the net outcome at farm level it provides
incentives for a more efficient use and recycling of
nutrients in innovative ways (Neeteson, 2001).
To support the 'agility' of farmers to respond to
these threats, efforts to increase nutrient use efficiency
at field, farm and regional scale are needed. Where
part of the initial success of growth-stimulating plant
nutrition was based on a largely empirical approach,
exploration of the upper limits and management of
the excess situation need a more precise quantitat-
ive understanding of the soil-plant-environment rela-
tionship in heterogeneous fields (Cassman and Plant,
1992; Doberman and Cassman, 2001; Van Noordwijk,
1999).
In contrast to the problems of excess nutrition, in
substantial parts of the tropics the lack of adequate
supply of plant nutrients remains a major constraint to
crop growth (Buresh et aI., 1997). Lack of knowledge
of plant nutrition, however, is not the major determin-
ant of this situation. Fertiliser sources are generally
used wherever they are financially feasible for the
types of crops farmers grow, and relative prices of
nutrient sources, labour and harvestable products ex-
plain most of the situations where soil fertility remains
the main constraint to agricultural production for sub-
sistence crops (Laegreid et aI., 2001). Real impact on
farmer's livelihood options in these circumstances is
likely to come from economical and policy changes
rather than from technical options for plant nutrition
per se (Izac and Sanchez, 2001).
Threats to sustainable farming globally thus con-
sist of on-farm concerns for the nutrient balance (Ac-
cess problems), angry neighbours and worried or dis-
satisfied customers (Excess problems) and regulations
and policies that are supposed to guide between the
scylla and charibdis of access and excess problems,
but that not necessarily give the right signals and in-
centives to farmers navigating their boats in between.
Research can help to define and shift the two main
frontiers, but should also focus on how to increase the
width of the domain in between (Figure 1) not least to
reduce economic and environmental risks. Research
and development efforts on frontier A of restricted ac-
cess may easily lead to overshoot across frontier B of
unacceptable excess problems if there is only a narrow
zone in between. Over-use of fertilizer in parts of Asia
on horticultural crops and lowland rice (Dobermann
and Cassman, 2001) exists, for a variety of reasons
(Wang et aI., 2001) close to upland systems where fer-
tilizer access is still problematic. Similarly, regulation
efforts aimed at frontier B and controlling the neg-
ative environmental effects of agriculture, can bring
farms and the plants that form the basis of the produc-
tion across frontier A, into serious access problems.
Factors that determine the width of the 'SAFE' zone
(balanced nutritional environment domain) may need
further attention and instead of a single 'economically
optimum fertilization rate', we may need to focus on
the parameter domain where, given the variability of
weather and other uncertainties, environmentally safe
production is possible at acceptable returns to invest-
ment and labour. Global challenges in plant nutrition
are thus formed by problems of limited access, prob-
lems of excess and the links between these two. In this
paper we will explore the nature of the two frontiers
and the domain in between the scales of a single root
and society as a whole.
Excess and access problems in plant nutrition
from root to societal scale
Plant nutrition research can focus on genes, cells,
roots, whole plants, cropped fields, farms, the human
food chain and/or society and land use. While at each
level a type of efficiency (generally output per unit in-
put, measured over specified system boundaries and
for a given time frame) can be defined, these efficien-
cies are not directly linked. Efficiency is not a scale
independent property - efficiency of a system can be
obtained despite relative inefficiency of subsystems,

while efficiency of a system does not guarantee that
it contributes to the efficiency of macro-systems. For
example, cassava is a very effective and efficient nu-
trient scavenger that can acquire nutrients from poor
soils and provide reasonable crop yields under low in-
put and low-labour conditions, making it a crop of last
resort on depleted soils in the (sub)humid tropics. The
nutrient content of the tubers that are harvested is sub-
stantial, while the price per unit dry weight is low. So
the 'added value' by the plant relative to the nutrients
depleted from the soil is low, and if we calculate the
'nutrient replacement cost' we may find that up to 60%
of the price a farmer gets for the tubers would have to
be used for buying fertilizer to balance the budgets of
N, P and K (ignoring the other nutrients exported with
the harvest) (Van Noordwijk, 1999; Whitmore et aI.,
2000). So the plant-level nutrient uptake efficiency, the
human labour use efficiency in the production process,
and the short-term farm level nutrient use efficiency
(allowing for output without external input) are not
matched by a financial nutrient replacement efficiency
and hence the longer term farm level nutrient budgets
in cassava production areas show a clear negative trend
(Hairiah et aI., 2000). Part of cassava tubers are in-
dustrially processed and shipped around the world as
tapioca for use as fodder in intensive animal produc-
tion in the 'developed' world. This trade contributes to
the nutrient excess problems in animal waste disposal,
thus the problems of 'excess' are linked to those of
'access' by this crop (Bouwman and Booij, 1998). If
our aim is balanced nutrition at all relevant scales, we
need to pay specific attention to the connection points
between scales, and at each system level increase the
domain between frontiers A (access) and B (excess).
Although 'access' and 'excess' domains can be
recognized from root to societal scale (Table 1), the
definition of the frontiers, and therefore ofthe 'SAFE'
range, differs and suggest, for example, that inter-
ventions aimed at the frontier of root access to nu-
trients will not necessarily contribute to improving
access to food of adequate quality by all people of the
world. Where current molecular biology and biotech-
nological approaches focus on the access frontier at
(sub )cellular (Kochian, 2001) or plant (Cakmak, 2001;
Graham and Welch, 2001) level, it may be an under-
statement to say that challenges remain to a holistic
understanding of plant nutrition in its societal context.
Determinants of the 'SAFE' range at root and plant
level are primarily based on the quality of the plant
as a regulator, preventing access-mechanisms to gen-
erate internal excess of any of the nutrients or other
27
solutes in the rhizosphere, and selectively increasing
access by a numerical or functional response of the
root system (with its symbionts). Although demand-
driven regulation of plant nutrition was recognized
in the 1940's, explicit description of 'plant nutrient
demand' rather than concentration-dependent mech-
anisms as driver of uptake remains an exception (De
Willigen and Van Noordwijk, 1987). The responses of
plants in heterogeneous environments may differ qual-
itatively (Hairiah et aI., 1993; contrasted to: Collet and
Horst, 2001) from that in homogeneous systems and
thus important aspects of the response of whole plants
to their complex environments tends to be missed in
the homogenization that is generally perceived to be
necessary for valid experiments.
At field scale the boundaries between excess and
access problems are primarily determined by the vari-
ability of both nutrient supply and demand between
the plants that are managed as a single unit (typic-
ally a field). Variability in inherent soil supply or
demand within a field will tend to shift the yield-
response-to-fertilization curve to the right and the
nutrient-excess-response-to-fertilization curve to the
left, decreasing the 'safe' range from both sides (e.g.
increasing fertilizer demand and nutrient losses) and
confirming the challenges of productive and envir-
onmentally friendly agriculture (Van Noordwijk and
Wadman, 1992; Whitmore and Van Noordwijk, 1995).
Precision farming (Robert, 2001; Pierce and Novak,
1999) can potentially increase the width of the 'safe'
range, by reducing the size of the management unit.
A challenge for 'precision farming' however remains
where the variability is temporal rather than spa-
tial, e.g. due to poor predictability of mineralization
rates. Spatial variability within fields can lead to non-
homogeneous fertilization practices as soon as zones
that differ in nutrient supply or crop demand can be
recognized on the basis of differences in elevation,
landscape genesis, distance to drains (De Vos, 1997,
EI-Sadek and Feyen, 2001), erosion and sedimenta-
tion zones on slopes (Van Noordwijk et aI., 1998),
heterogeneity caused by slash-and-burn land clear-
ing and subsequent erosion/deposition (Ketterings,
1999; Rodenburg, 1999), animal defecation (including
birds perching on trees; Belsky, 1994) or shading by
hedgerows and field boundary vegetation (Mette et aI.,
2001). Specific agronomic practices further increase
spatial variability, e.g. by nutrient transfer from grass
strips to tree lines in orchards, tropical hedgerow in-
tercropping, or by injection of slurry or fertilizer into
the soil (Laegreid et aI., 2001). Where these practices
28
Table 1. Determinants of the frontiers between the access problem, balanced nutrition and excess problem domains at different scales
in plant nutrition

Scale Access sufficiency frontier Safe range Excess avoidance frontier

Root Zero-sink diffusion delivery Effective demand-based reg- Incomplete tolerance to con-
to root, despite all rhizosphere ulation of uptake and rhizo- centration build-up if mass
modifications, is less than de- sphere modification flow delivery to root surface
mand exceeds Cupta

Plant Compensatory root develop-
ment unable to reduce de-
mand per unit root to available
supply conditions
Field - crop season Patches with smallest supply
per unit demand cause unac-
ceptable yield shortfalls
Internal buffer and elastic re-
sponse in root development
and function can achieve ad-
equate supply of all nutrients
required at any phase of plant
development
Spatial variability in sup-
ply matches that of demand
within the management unit
(field or part thereof)
Internal storage of excess ele-
ments in harmless forms to
reduce accumulation around
roots
Patches with highest supply
per unit demand cause lateral
flows exceeding tolerance of
neighbours

Field - long term Total input is less than Balancing the nutrient budget Total input exceeds
use+unavoidable losses at field scale use+tolerable losses

Farm Nutrient export in farm
produce and unavoidable
losses occur at less than
(risk-adjusted, discounted)
replacement costs of nutrients
Human food chain Food of adequate quality and
nutrient ratios is not available
to, or affordable by part of
human population
Internal waste recycling and
efficient application strategies
increases domain where mar-
ginal benefits exceed marginal
costs of nutrient use for prof-
itable crops
Sufficient choice for balanced
diets at affordable consumer
prices
Uncontrollable inputs derived
from lateral flows exceed
use+tolerable losses; toler-
able losses smaller than un-
avoidable ones
Contaminant levels and nutri-
ent ratios in staple+additional
food supply exceed safety
limits

Society Limited stocks of exploitable
nutrients; lack of economic
incentives for nutrient recapit-
alization and balancing nutri-
ent budgets on farms
Adequate regulation and eco- Lack of economic incentives
nomic incentives for nutrient for recycling and use of nu-
chain managers trient rich waste products in
whole nutrient (food) chain

aCupt = uptake concentration or nutrient uptake flux divided by water intake flux.

increase heterogeneity within reach of the root system
of a single plant, they may improve the efficiency of
nutrient uptake (as in classical fertilizer band place-
ment practices) and thus widen the safe range. Where
heterogeneity is created on scales beyond the reach
(below the rooting zone or laterally beyond reach) of
a single plant, this will narrow the 'safe' range (Van
Noordwijk et aI., 1993). Spatial variability can al-
ways lead to two types of response by the farmer:
1) compensate for the patterns and try to achieve a
more homogeneous situation, or 2) accept the differ-
ences and adapt to them, by differentiating the crops
(including multi-species approaches), abandoning part
of the land and/or focus efforts on the parts with the
highest returns to investment of labour or inputs (Van
Noordwijk et aI., 1994, 1998). Technical precision ag-
riculture is largely based on the first approach, while
much of the management decisions by small farmers
living in heterogeneous environments are based on the
second approach. Improvements are likely to come
from a better integration of temporal variability into
the spatially explicit management practices (Alphen
and Stoorvogel, 2000).
Human nutrition and societal concerns over access
and excess problems in plant nutrition are controlled
by social and economic processes, for which technical
and ecological opportunities only form pre-conditions.
The access and excess frontiers are thus often based on
judgmental terms ('adequate', 'tolerable'), e.g. envir-
onmental risks and costs are often difficult to quantify

and food quality is not easy defined apart from limits
on pesticides and heavy metals, and involve different
stakeholders. Negotiations rather than an 'objective'
cost-benefit analysis of the various options therefore
drive the definition of the 'safe' range. Technical in-
creases in the width of the ' safe' range, however, can
improve the chances that the political negotiation pro-
cess can lead to mutually acceptable solutions for plant
nutrition, as is the case for other natural resource man-
agement issues (Van Noordwijk et aI., 2001). Where
the merits of technical solutions depend on climatic
and soil conditions, research on extrapolation domains
of 'success' and 'failures' can play an important role
in informing the policy debate.
Technical control and ecological analogue
approaches
Two paradigms are currently used to decrease the
conflicts between access and excess problems: the
technical control paradigm of precision farming and
hydroponics, and the ecological analogue paradigm of
increased buffering and resilience. The two paradigms
are not mutually exclusive, but they suggest different
interventions into the status quo. Precision farming
(Robert, 2001) differentiates farm operations includ-
ing fertilization within a field and thus creates smaller
management units to get a better tuning of supply to
demand. The terminology of precision farming ori-
ginated in the large fields of the US, imposed on a
landscape without respect for underlying soil variabil-
ity and thus highly heterogeneous, and is based on new
information processing opportunities in mechanized
farming. There is growing recognition that the small
scale tropical farmer, on fields developed organically
in the landscape has in fact also been making such
site-specific management decisions that characterize
precision farming (Sinclair and Walker, 1998; Berkes
et aI., 2000). Unfortunately, plant nutrition research
for the tropics has largely resulted in blanket fertilizer
recommendation schemes for the tropics (Dobermann
and Cassman, 2001) and ill-adapted innovations (Fuji-
saka, 1994) rather than supporting the decisions farm-
ers have to make in the real, heterogeneous world, and
this probably contributes to the low overall nutrient
use efficiencies of farming. The efficiency of fertilizer
and organic input use varies with method of applica-
tion (Cassman et al., 1998, Laegreid et al., 2001), but
farmer incentives to the use of techniques that increase
29
fertilizer use efficiency critically depend on fertilizer
price (Van Noordwijk and Scholten, 1994).
Doberman and Cassman (2002) define site-specific
nutrient management more broadly as the dynamic,
location-specific management of nutrients in a partic-
ular cropping season to optimize the congruence of
supply and demand of nutrients according to their dif-
ferences in cycling through soil-plant systems. This
definition accounts for (i) regional and seasonal dif-
ferences in yield potential and crop nutrient demand,
(ii) between- as well as within-field spatial variab-
ility in inherent nutrient supply, (iii) within-season
dynamics of soil N sup-ply and crop N demand, (iv)
location-specific cropping systems and crop manage-
ment practices and (v) location specific objectives
(tolerance for leaching losses, prices for products and
inputs). Precision farming as currently practiced in
temperate agriculture is still far from reaching its full
potential (Doberman and Cassman, 2002).
In hydroponics the low predictability of variable
soils is replaced in a more radical way by a system
that allows for better technical control of supply and
monitoring of demand in systems with a low buffer
capacity and rapid response to management interven-
tions (Heinen and De Willigen, 1995; Silberbush and
Ben-Asher, 2001; Van Os, 1999; Van Noordwijk,
1990). Both of these technical approaches (precision
farming and hydroponics) aim to reduce variability in
space and time and tend to minimize the use of organic
inputs (e.g. manures, plant residues, compost, etc.) be-
cause of their lower predictability of composition and
nutrient supply dynamics (although there have been
recent advances in databases and decision support sys-
tems that facilitate the use of organic materials e.g.
Palm et aI., 2001).
The ecological analogue approach, in contrast,
accepts variability of supply and demand and the ab-
sence of full synchrony between nutrient supply and
plant demand as facts of life and tries to reduce their
consequences (Ewel, 1999; Joffre et aI., 1999; Ong
and Leakey, 1999; Trenbath, 1999). Specifically for
nutrients, this may involve the use of 'safetynets' and
filter strips to mop up leftover nutrients leaving the
system (Van Noordwijk and Garrity, 1995). The terms
'safetynets' and 'filter' are used here in a generic sense
of anything that can intercept a vertical or lateral re-
source flow. These efforts may require an increase in
(manageable) complexity by intercropping and agro-
forestry and an increase in the internal (organic) buffer
and soil nutrient capital (Vandermeer et al., 1998; Van
Noordwijk and Ong, 1999).
30
The use of a biological safetynet in the form of fil-
ter strips at field boundaries has been suggested in the
temperate zone as one of the options for combining
environmental standards and economically feasible
production, returning some complexity to landscapes
dominated by monocultures. Some of the principles
of more complex agro-ecosystems of the tropics may
indeed be of value in the temperate, intensive agricul-
tural world as well (Vandermeer et aI., 1998; Sparovek
et aI., 2002), as they may offer options to filter ex-
cess nutrients before they reach the neighbours' water
or air, and can provide the visual attractiveness that
consumers appreciate. The possible functionality of
riparian filter strips may, however, be easily overestim-
ated, and requires specification for each type of filter
function.
Current understanding of plant nutrition is dom-
inated by experiments with and theories for mono-
cultures and is more geared to the technical control
than to the ecological analogue approach. Options for
combining external (inorganic) nutrient use with local
(organic) sources remain under-utilized because cur-
rent knowledge is too crude (Vanlauwe et al., 2001).
In multi-species systems the potential combinations
and benefits of such a combined inorganic-organic ap-
proach are increased due to the variations in spatial
and temporal availability of plant residues.
Although the basic equations governing nutrient
uptake in monocultures have been in use for several
decades at the level of roots and root systems (Nye
and Tinker 1974; De Willigen et aI., 2000), the spe-
cific effects of plant rhizospere modification (as a way
of shifting the 'access' domain) are still a focus of
research (Kirk, 2001; Claassen, 2001). Applications
to intraspecific competition or longer-term effects of
such modification are still relatively scarce. Rhizo-
sphere modification effects on P may carry over to
a subsequent crop, depending on the binding to soil
particles and the biological availability of the act-
ive compound (Hocking and Randall, 2001; Gransee,
2001; Lu et aI., 2001).
In the remaining part of this review we will ex-
plore how complex systems can playa role in issues of
'access' and 'excess' by considering impacts ofrhizo-
sphere modification by components of the system, and
safetynets for leaching nutrients.
Reducing nutrient access and excess problems in
complex agro-ecosystems
Complexity and residence time
Complexity in agro-ecosystems can consist of mul-
tiple components interacting spatially (at the same
time), temporally (at the same place) or both. Tem-
poral interactions are, almost by definition, one-way
effects where the earlier events influence the later
ones, while spatial interactions are often mutual, al-
though the interaction strength can differ. The lowest
form of complexity is thus found in rotational or se-
quential systems, and these have persisted into the
most intensive agricultural production systems of the
world, if only for reasons of pest and disease control
(Desaeger, pers. com.). Crop rotations are (also) com-
mon in the tropics and although agronomically they
are relatively well researched, their impacts on nutri-
ent access and excess cannot be easily separated from
other soil-borne effects. Overall, legume-rich fallows
can be expected to improve N supply to subsequent
crops, but the rate at which this supply can be accessed
depends on the rate of mineralization and hence on
the 'quality' of the litter (Palm et al. 2001). As vari-
ous organic and inorganic pools can be modified in
parallel, opinions on the parameters that provide the
best predictor of 'fallow effects' differ between au-
thors and crucially depend on the specific objective to
be achieved (Cadisch and Giller, 2000). For example
a plant residue with a low C:N ratio, low lignin and
polyphenol content may be considered 'high' quality
in terms of a fast nitrogen release but will be con-
sidered of 'low' quality for the purpose of mulch for
soil erosion protection.
A basic concept in nutrient access and excess is-
sues is that of mean 'residence time' in the rooted
soil layers, indicating the time nutrients remain avail-
able under given climate and soil conditions for uptake
by plants, before they are lost to deep ground water,
surface water or to the atmosphere. In most temper-
ate zone agricultural systems the residence time of
nitrogen will be close to one growing season, with
most of the leaching occurring in winter. In the trop-
ics, the residence time of nitrogen decreases rapidly
with increasing rainfall rates. De Willigen and Van
Noordwijk (1989) provided examples of how N use ef-
ficiency by a maize crop in the humid tropics depends
on synlocation and synchrony, relative to rooting depth
and leaching rate. For P residence times are generally
more than one growing season and thus carryover ef-

fects from one crop to the next one are likely to be
strong. Can we manage nutrient residence time in the
system? Slow release fertilizers have been developed
but are not yet frequently being used as multiple small
fertilizer application can achieve similar results. The
regulation of nutrient release with organic inputs is
determined by their quality, e.g. high C:N and/or
high ligninlpolyphenol content will increased resid-
ence time, but such effects are more difficult to control
to ensure at the same time sufficient nutrient availab-
ility for current demand (Cadisch and Giller, 2001).
Increasing residence time of nutrients in systems can
also be achieved by reductions of mobility of nutrients
by reduced soil water content e.g. via an increased
crop water use (Suproyogo et aI., 2002).
Competition and complementarity in simultaneous
systems
The correct interpretation of the balance of posit-
ive and negative interactions in simultaneous systems
remains a major challenge. Positive effects of inter-
cropping in wheat/maize and wheat/soybean can be
attributed (Zhang et aI., 2001) to better growth of the
border rows, due to increased light levels (2/3 of the
overall effect) and better soil exploration (113 of the
overall effect).
Root distribution of the component species has
often been interpreted as direct indication of competi-
tion and complementarity in mixed cropping systems,
induding agroforestry (Van Noordwijk et al., 1996;
Huxley, 1999; Wahid, 2001). The term competition
can refer to a process of acquisition of a shared re-
source, or to the consequences this has for the growth
and productivity of the competing plants. For a farmer
competition only matters if the competing plants differ
in value (per unit resource acquired) and thus intraspe-
cific competition in a densely packed monoculture is
not a concern, but competition with a weed is. Simil-
arly, competition is a grave concern (Sanchez, 1995)
in agroforestry system where trees of low direct use
value prove to be effectively competing with annual
crops valued by the farmer, while the interactions in
a complex homegarden where all plants have value,
is not seen as a problem (Van Noordwijk and Ong,
1999). For the plants involved, however, these value
judgements do not matter, and an investigation can
focus on the process of resource acquisition.
While low root length densities may be sufficient
for uptake of mobile nutrients such as nitrate in a
situation without competition (or one with intraspe-
31
cific competition in a monoculture of identical plants),
the relative competitive strength of plants in a situ-
ation with shared access to a soil layer is proportional
to their respective root length densities, making high
root length densities functional from the perspective
of an individual plant (Van Noordwijk, 1983; Robin-
son, 2001). A functional interpretation of root length
densities thus is essentially different for a mixed plant
situation than it is for the monocultures on which
most plant physiological and agronomic research is
focussed.
In a quantitative description of the way acquisition
is affected by the presence of competition, we can
distinguish:
• resource depletion by the other plants reduces
availability (in a way that can be directly off-set
by externally increasing resource supply),
• reductions of potential uptake per unit root length
by the presence of other roots or 'interference'; as
potential uptake is a non-linear function of total
root length density, caused by a reduction of soil
cylinder depleted per average root with increasing
total root length,
• reductions of mobility of nutrients by reduced soil
water content.
The first and third of these effects can, in special
circumstances, be off-set or made into positive effects,
by plants that increase the availability of nutrients in
their rhizosphere and or that increase soil water con-
tent in nutrient-rich soil layers through the process of
'hydraulic equilibration' (Smith et aI., 1999; Burgess
et aI., 2001). Although the second aspect of competi-
tion (,interference') would still remain, it is possible
that the overall effect is positive, depending on the
degree of 'synlocation' of the various types of roots. A
specific from of synlocation can be imposed if roots of
more than one plant species follow a limited number of
cracks or macropores derived from old tree root chan-
nels (Van Noordwijk et aI., 1991, 1993). The impacts
of rhizosphere modification by any component of a
mixed-species system depend on the degree of 'syn-
location' of the roots of the different species (Figure 2;
Van Noordwijk et aI., 1999).
We used a process-based model of agroforestry
(WaNuLCAS) in a case study to address the fol-
lowing question: Can we predict where and when
rhizosphere modification by one species can be be-
neficial for accompanying plants? The WaNuLCAS
(water, nutrient and light capture in agroforestry sys-
tems) model (Van Noordwijk and Lusiana, 1999,
2000; http://www.icraj.cgiar.org/sea/AgroModels) de-
32
only tree benefits ...........................................1ree&crop both benefiL
IWIU! mediJllII hieh
syn!ocation
Figure 2. Conceptual scheme of different degrees of synlocation
of roots of different plants and its consequences for the benefits of
rhizosphere modification by one of them.
scribes plant-plant interaction in above- and below-
ground aspects. Competition and complementarity in
use of nitrogen, phosphorus and water can be eval-
uated in the model for any combination of trees,
crops, planting times, organic and inorganic input
regimes, provided that basic properties of the root
system and aboveground growth are known for the
given soil and climate. The WaNuLCAS model in-
corporates key aspects of space (4 soil layers in 4
lateral zones), time (daily time steps, simulations up
to 25 years or beyond), complexity (1-3 tree species
can be grown simultaneously, crop calendar can be
specified for each zone separately) and management
(fertilization, organic inputs from outside or inside
the system, aboveground tree management by pruning,
manipulation of root distribution) of agroforestry. In-
teractions in P uptake in the model consist of a number
of separate terms that can be studied in isolation or in
combination in the model, in ways that are not easily
reproduced in real world experiments (Table 2).
For the current exploration of rhizosphere modific-
ation and safetynet functions a default parametrization
was used (Table 3) that reflects the BMSF site in
Lampung (Sumatra, Indonesia) where WaNuLCAS
validation tests were carried out focusing on deep N
uptake (Rowe, 1999; Suprayogo, 2000).
The simulation results presented in Figure 3 ap-
plied to a setting where the default settings predicted
approximately equal crop performance in three zones
adjacent to the hedgerow trees, as positive (via im-
proved N supply) and negative (via water supply)
interaction effects where in balance. If rhizosphere
modification by the trees is simulated as a 'medium'
intensity (the technical details are specified in the cap-
tion to Figure 3), it would not affect crop growth in
zones 3 and 4 (furthest away from tree), but could
either increase or decrease crop growth in zone 2
0.3
T Rhizosphere effect
N
'E Strong ~ Synlocation
~ ~High
•
medium~
•
:2 default
Q) 0.25
'50.
medium~
~
'(i
:E Strong None
0.2 - / - - - - , - - - - - - , - - - - - , - - - -
1 2 3 4
Zone, increasing distance to tree
Figure 3. Predictions by the WaNuLCAS model of the im-
pacts of rhizosphere P mobilization by hedgerows trees on the
yields of adjacent maize in three zones with increasing dis-
tance to the hedgerow, in dependence of the degree of 'syn-
location' of maize and hedgerow tree roots ; the medium and
strong T_Rhizosphere effect refer to parameter values of 10
and 100 cm2 day-I , respectively, that detennine the fraction
of the immobile P pool in the soil that can be mobilized per
day (1 - 1/ (1 + T _NutMob*rrT_Lrv(T_RootDiam+T_Rhizosphere
Diam)2/4», with a root and rhizosphere diameter of 1 mm each and
a tree root length density T_Lrv per zone as specified in Table 3).
(next to tree), depending on the degree of synloca-
tion; for strong synlocation and a medium rhizosphere
modification a slight positive effect would be expected
in zone 2. If an (unrealistically) strong rhizosphere
effect is simulated, the results show an expected in-
crease of crop yields across all zones in case of strong
synlocation, a neutral effect for partial synlocation
and negative effects in all zones in the absence of
synlocation,
Thus, we conclude that under high synlocation
both species can benefit from the increased availab-
ility of nutrients in their root environment, but the net
effect on the species modifying the rhizosphere may
be positive, neutral or negative depending on the de-
gree the companion crop was limited by the mobilised
resource and is itself a more effective competitor for
the resource limiting the mobiliser crop. The 'syn-
location' parameter that appears to be crucial for the
outcome of the interaction is potentially measurable
(Van Noordwijk et al. 1993), but no data exist as
yet. The problem of 'rhizosphere' modification and
synlocation may partly be overcome by an effective
mycorrhizal symbiosis that can scavenge in the rhizo-
sphere of neighbouring roots or even directly connect
them.
The impacts on the competing plants differ
between the initial phase of near-exponential growth,
where a setback in current growth affects future
 33
Table 2. Interaction effects on P nutrition between plants A and B, eitber grown simultaneously (with partially overlapping root
zones) or sequentially on tbe same site

Effect of plant A Simultaneous Sequential

Effect on potential P Determinants of Effect on potential P Determinants of
uptake by plant B impact on B uptake by plant B impact on B

Uptake interference Relative root lengtb 0

densities of A and B
Reduce P P demand A, root 0/- Recharge from less
concentration in soil density A, P buffering mobile P pools
solution by uptake soil
Increase P stocks by 0 ... + P harvest index 0 ... + P harvest index
residue recycling A, effective A, effective
decomposition rate decomposition rate
residues A residues A

(Potentially) increase (0 ... ++) Root synlocation, (0 ... +) Half-life time

P concentration by P immobile soil P mobilization effect,
mobilization reserves, mobilization immobile soil P
strength A reserves, mobilization
strengtb A

(Potentially) increase (0 ... ++) Root synlocation, (0 ... +) Half-life time

P concentration by organic Preserves, mineralization
mineralization of Porg phosphatase effect, organic P
effectiveness plant A reserves, phosphatase
effectiveness plant A

Reduce P mobility by Rainfall, soil type 0

soil water uptake
(Potentially) increase (+) Deep roots A, rainfall, 0
P mobility by soil type
hydraulic lift
(Potentially) increase (0 ... ++) Root synlocation, (0 ... +) Half-life time sorption
P mobility by reduced rhizopshere effect A change rhizosphere
P sorption effect A

Overall effect - ... + - ... +

growth as well, and the closed-crop stage of linear
growth, where only current growth rates are affected.
This means that negative early effects of competition,
e.g. before positive rhizosphere modification effects
become operational, can normally not be compensated
by positive effects later on in the growth period. The
interaction of all these positive and negative effects
acting at different spatial and temporal scales, may re-
quire the use of simulation models that can keep track
of cumulative effects as shown above.
Sajetynets in sequential and simultaneous systems
Concept
The concept of deep nutrient uptake by trees has been
discussed for more than a century (Huxley, 1999),
but only applies to soils that have appreciable nutrient
stocks derived from weathering or subsurface lateral
flows. For nutrients of low mobility, long time frames
apply and a low efficiency of uptake on a yearly time
scale can still lead to appreciable depletion over the
lifetime of a perennial crop or tree. For mobile nutri-
ents, however, the residence time of nutrients in the
subsoil is limited, and interception by trees or other
deep-rooted vegetation has to occur while nutrients are
on their way out of the reach of the root system or the
soil system. For such situations the term 'safetynet'
has been coined. Key questions on the way safetynets
and filters function in natural resource management
are:
34
- How effective are different types of filters for in-
tercepting flows as can be expected in different
rainfall regimes for nutrients differently sorbed
and buffered by the soil?
- How does the filter or safety net efficiency depend
on the 'mesh size' as determined by root length
density and thickness of the soil layer involved?
- How quickly will filters saturate under high in-
flows?
- How fast can the filters regenerate between events?
- Do filters have a direct value and can they be
treated as a separate 'land use practice'?
- How effectively can intercepted nutrients be-
reused in the systems?
A first theoretical analysis of safetynets by Van
Noordwijk (1989) showed that a limited window of
opportunities exist in a sequential system for a deep-
rooted fallow to intercept nutrients leached beyond the
shallow crop root zone in cropping years. As argued by
Sommer et aI. (2001), nitrate sorption in subsoil can
slow down nitrate movement to increase the chances
of recapture by subsequent fallow vegetation.. Sub-
soil nitrate sorption is common in acid tropical soils
(Suprayogo et aI., 2002, Wong et aI. 1990) although
its effectiveness is in decreasing leaching is limited
(Suprayogo et aI., 2002).
For nutrients of higher mobility a more perman-
ent presence of a safetynet is required. Cadisch et
aI. (1997) and Rowe et aI. (1999) have explored
how such a safetynet function may depend on tree
root length density in the layer underneath the crop
root zone and provided evidence from 15N placement
experiments that uptake from deeper layers can be
substantial, provided aboveground 'demand' exists.
Again, a simulation model that looks at the interac-
tions between leaching rates, aboveground demand,
N2 fixation and the possibilities for uptake from vari-
ous parts of the soil profile is needed to move beyond
qualitative statements. Rowe et aI. (1999) calculated
for hedgerow intercropping experiments in Lampung
(Indonesia) that the non-N2 fixing tree Peltophorum
dasyrrachis recycled 42 kg N ha- I year- 1 by uptake
from below the crop root zone, while the N2-fixing
Gliricidia sepium recycled only 21 kg N ha- 1 year-I.
Safetynet efficiency: Effect of rainfall
In a series of WaNuLCAS simulations (Table 3) that
excluded P limitations on crop or tree growth, a
gradual shift from water to N limited growth condi-
tions was obtained (Figure 4) by applying multipliers
on the daily rainfall records for Lampung. The pre-
dicted maize yields, with or without trees, was highest
for an annual rainfall of two-thirds of the actual record
(2318 mm year-I), with a rapid reduction for lower
rainfall values and a gradual decline for higher rainfall
(Figure 4A). The presence of a hedgerow tree shifted
the water limitation curve to the right (Figure 4B),
showing that even at an annual rainfall of 2300 mm
on 5-10% of the days in the cropping season water
can be the growth limiting factor in the presence of
hedgerows; it also shifted the nitrogen limitation curve
to the left, indicating a positive effect on N supply of
the (non-N2 fixing) tree Peltophorum dasyrrachis. N
leaching is predicted to be reduced by the presence
of hedgerow trees at all rainfall rates above 1000 mm
year-I, and the total filter efficiency increases (Fig-
ures 4C and D). Filter effectiveness for the same tree
and crop parameters decreases non-linearly with in-
creasing rainfall, as the residence time of solutes in the
deeper soil layers decreases non-linearly with a larger
surplus of rainfall over evapotranspiration. Below an
annual rainfall of 1000 mm year- 1 the amount of wa-
ter and N leaching into and out of the subsoil layer 4
(the deepest soil layer considered, e.g. 0.5-1 m depth,
see Table 3) becomes negligible and the calculation
of the filter function looses its meaning. The absolute
increase of filter efficiency due to the presence of trees
increases with decreasing rainfall from 9% at 1.4 times
the rainfall (3260 mm year-I) to 19% at 50% of the
default rainfall (1160 mm year- I ), for the default crop
and tree root length densities.
Safetynet efficiency: Effect of root length density and
distribution
A further WaNuLCAS model exploration of these
effects at the default rainfall showed a strong depend-
ence of filter functions in layer 4 (the deepest soil
layer considered) on the tree root length density in this
layer (Figure 5). Filter functions for situations with
hedgerow trees start to increase below the crop-only
situation (maize is supposed to have a few roots in
layer 4) if tree root length density exceeds a value of
0.001 cm cm- 3 , and reaches a maximum when tree
root length density in this layer becomes 1 cm cm- 3
(Figure 5B, note logarithmic scale). Tree biomass be-
nefits from more roots in layer 4, but saturates at a
tree root length density in layer 4 of 0.03 cm cm- 3
(Figure 5C), as this apparently is sufficient to carry the
tree through the dry season. In the predicted impacts
of the hedgerow on the maize crop (Figure 5D) the
 35
Table 3. WaNuLCAS parameter settings used as default for the simulations of Figures 3--6; detailed definitions and other default parameters
are specified in van Noordwijk and Lusiana (1999).

Parameter Value

AF_Zone(Znl) - width of zone that includes tree O.5m

AF_Zone(Zn2) - width of second zone 1m
AF_Zone(Zn3) - width of third zone 1m

AF_ZoneTot - total width of the simulated area 4m

AF_DepthLayl .. 4 - depth of the four soil layers 0.05,0.15,0.3 and 0.5 m, respectively

RCCLrvm1 (Crop root length density layerl) 6.67 cm cm- 3

RCCLrvm2 (Crop root length density layer2) 3 cmcm- 3
RCCLrvm3 (Crop root length density layer3) 0.1284 cm cm- 3
RCCLrvm4 (Crop root length density layer4) 0.0064 cm cm- 3

RCTLrv1..4Znl (Tree root length density layerl, Zn 1...4) 2, 1,0.5,0.25 cm cm- 3 , respectively
RCTLrvl..4Zn2 (Tree root length density layerl, Zn 1...4) 1.5,0.75,0.4,0.2
RCTLrvl..4Zn3 (Tree root length density layerl, Zn 1...4) 1.3, 0.45, 0.3, 0.15
RCTLrvl..4Zn4 (Tree root length density layerl, Zn 1...4) 1.2, 0.37, 0.15, 0.1

Cropping season 4 cropping seasons (2 per year), started on December 6th , followed
by next crop in 30th March
CaYlantYear 0, 1, 1,2, respectively
Ca_PlantDoY 340, 90, 340, 90, respectively

Tree planting Trees were planted 5 months before planting 1st crop
T_PlantY o
T_PlantDoY 205
T _ GioResInit, fuitial growth reserves in tree 0.01 g m- 2
T_CanBiomInit, Initial tree canopy biomass 0.1 g m- 2
T_WoodBiomInit, Initial tree wood biomass 0.1 g m- 2
T _WoodHInit, Initial tree height Om
Pruning Pruning prior to every crop planting
TYrunPlant? 1
Soil Texture:
Percentage of clay 30
Percentage of silt 10
Percentage of sand 60

Slope 10%
AF_SlopeSoilHoriz 10

Fertilizer No fertilizer added

CaYertAppYear, first year if fertilizer application 100

Mc2_SomhlitType: 3
Mc2_CorgInitMeh 1
Mc2_CrefMeth3 2

model predicts that a strong difference between crop
zones exists for tree root length densities in layer 4
up to 0.1 cm cm- 3 , and that higher tree root length
densities in layer 4 actually benefit the crop, despite a
higher biomass and thus stronger direct competition.
The positive (safetynet functions) and negative
(competition for water and N) impacts of simultan-
eous tree roots on maize yield was further analysed by
separating relative tree root distribution from absolute
root length density. Results for these calculations show
(Figure 6) that negative overall effects of the tree can
be expected for trees that have all their roots in the
topsoil, and for trees with only 0-10% of their roots
in the subsoil, at low overall tree root length. These
same relative tree root distributions at higher total root
length (i.e. higher absolute root lengths in both top
and subsoil) can have a moderate positive effect on
maize yield, while tree root systems with 20% or more
of their roots in the subsoil were consistently positive
for the crop, the higher the total root length, the more
positive the impact on maize.
36

A B
1.0

..
0)
E
"",

-g,4
--A- Maize only :: 0 0.8
water
Maize+Trees
~
·tf • •
6
tl 0

.~
s;:. 3
.lil
• <1='"
",-
I:
0.6
nitrogen
0 04
E '" .~
.-!:
"iii
::J
I:
~ 0.2
(;
•
I:
~ .iii'
~ 0.0

.. •
300 C 1.0 D
"",
..c:
Cl 250

•
0.8
-'"
'"0
•
I:
~-
200 .tj
..2 0.6

•
I:

..
Cl
.a
...iii
I:
:E 150

'"" u:: 04
•
..9l 100 "iii
Z "0
•
•
"iii I- 0.2
::J 50
I:
I:
~
0 0.0
0 1000 2000 3000 4000 0 1000 2000 3000 4000
Annual rainfall, mm Annual rainfall, mm

Figure 4. (A) Model predictions of maize yield (two crops per year; for a standard parametrization, see text) as a function of annual rainfall
(obtained by using multipliers on the 1993 Lampung daily record - see arrow), with or without the presence of regularly pruned hedgerows of
non-N-fixing trees; (B) The fraction of the growing periods that either N or water is the main limiting factor(causing at least 10% reduction in
growth on a given day and being the strongest current limitation); (C) Annual loss of N by leaching and lateral sub-surface flow; (D) Overall
filter efficiency for N, defined as uptake/(uptake+losses).

"<t
0.35

0.30
A

• .. Maize + Trees
"<t
0.20
B

••
....
.. .. •...
iii -:6.- Maize only ~ 0.15

•
~ 0.25 .!l!
I: I:
0.20
:e
0 0

A2
t; 0.10
I: c:
-t.l!!
0.15
.2
;;::
010 • ~
i;: 0.05
....... •• A
Maize +Trees
Maizeonly
Z 0.05 Z

0.00 000
1000 2000 3000 4000 0.0001 0.001 0.01 0.1 1 10
Annual rainfall, mm T4 Root length density, em em- 3
C D
3.6
24
.D· -El:: ~:.:&r~.:.~

...• ....••...
".r:"' ., 34
22 .. E) . . D. .. Li .. <;
"'
.r: 3.2

. ·····i'i:..•.0 ·
~ 20 Cl ...... E) .b .0··0
~ 3.0
'Ii 1.8
II> ,; 2.8
"'
E 1.6 lil
.s;:.
..... <)
.......•.
:a
0 2.6 ·n Zone4

.•.
0)

14 .1::1 . ,6. Zone3

24
"'
0)
!!! ~ .0- Zone2
I- 1.2 2.2 AvgField
1.0 2.0
0.0001 0.001 0.01 0.1 1 10 0.0001 0.001 0.01 0.1 1 10
Root length density T4, em em- 3 Root length density T4, em em- 3
Figure 5. Filter function (=uptake/(uptake+leaching) for nitrogen of the fourth soil layer, (A) as a function of annual rainfall and (B) as function
of tree root length density in layer 4 (N.B. logarithmic scale). (C) and (D) show tree biomass and maize yields in the same simulations. For
input parameters see Table 3.
 37

A 8
1.5 3.1

~%OftreerootintoPsOil:
2.9
.
;" 1.4 ~
~
. 2.7 --A-- 0
~

'" 2.5 --+-

.".. 2.3
___ 4020
~1.3 ~
___ 60
'0
___ 80
~1.2 E
a 2.1
.
~90
~ :0
'0; ~--T-95 •
::;;1.1 ~ 1.9 ~ ... . -0- no trees
t-
1.7 --+- 100

1.0 +--,---..,.--.....----,-~...; 1.5

O~ 02 0.4 OB OB 1.0 0.0 0.2 0.4 0.6 O.B 1.0
Tree root length per unit soli area, Tree root length per unit soil area,
relative to default paramaters relative to default parameters

Figure 6. Predicted maize yield (A) and tree biomass (B) for the default rainfall situation of Figure 4 (2318 mm year-I), when relative
distribution of tree roots with depth as well as total amount of tree roots are varied independently. Whereas the 'default' tree root system had
21.5% of its roots in the top layer, a series of data was made that had 0-100% of its roots in the top layer and the remainder allocated to the
deeper layers in proportion to the root length densities of the default case (the relative distribution over the four zones with increasing distance
to the tree was not modified). For each of these root distributions, the total amount of roots was varied from 0.1-1 times the default, while
maintaining the relative values.

A remarkable feature of these results is that at
default values for total root length, the tree root sys-
tems with 60% of their roots below the topsoil led to
(slightly) higher maize yields, than those with more
(up to 100%) in the subsoil, while at lower total root
system size the 100% in subsoil (0% in topsoil) was
better for the maize. Although this effect is much too
subtle to be recognized in any field data, it seems
counter-intuitive. On detailed analysis the differences
in crop N uptake between trees with 0 and 40% of their
roots in the topsoil arise during dry spells in the crop-
ping season when the sparse crop roots in the deeper
soil layers have slightly more N available in situations
where the trees forage partly in the topsoil.
Sensitivity analysis of the model thus shows that
the tree root length density below the main crop root
zone may have complex and partially unexpected ef-
fects on crop performance in situations where negative
effects via competition for water and positive effects
via improved N supply vary in intensity during the
growing season. Positive effects of the trees via im-
proved N supply would increase with time, and the
current results averaged over 2 years (4 crops) are only
a first indication of the overall impact of including
hedgerow trees into the maize production system.
Apparently the tree root length density required for
a safety-net function can be quite low, as it depends
on the residence time of the mobile solute in the layer
where the filter function is supposed to occur. For real
trees the known root-length densities indicate that only
a partial safety-net function can be expected to occur,
as root length densities in the subsoil are normally less
than 1 cm cm- 3 , but above 0.001 cm cm- 3 .
Maximum efficiency attainable for a safetynet is
less than 1 and depends on 'plant demand' and sup-
ply in other soil layers, residence time of the solutes
and the relative extraction efficiency. For a different
set of parameters Cadisch et al. (1997) obtained filter
efficiencies up to 90% for a tree root length density of
about 1 cm cm -3. If less N is available from other
layers and tree demand is higher, higher safetynet
efficiencies can be obtained in the model outcomes.
Overall, we conclude that filter functions for nitro-
gen on its way to leach out of the soil profile are related
to rainfall, tree N demand and tree root length density
in a strongly non-linear fashion. In formulating and
using WaNuLCAS as a mechanistic model for inter-
actions in more complex agro-ecosystems, we quickly
realize that parameters that can greatly influence the
model outcome are insufficiently known. These para-
meters refer to the longer term replenishment of the
nutrient pools accessed by deep-rooted trees or by
plants with specific rhizosphere effects, as well as to
the spatial correlation of roots or mycorrhizal hyphae
of interacting plant species. For the performance of
the system as a whole, however, management of the
aboveground biomass and its effect on nutrient de-
mand dominates the opportunities to make use of
more complex agro-ecosystems to achieve a higher
overall nutrient-use efficiency. The simulation model
provides a framework for combining knowledge on
component behaviour to system-level assessment of
plant nutrition.
38
Challenges to make it work
Complexity of agro-ecosystems is no guarantee for
avoiding excess nutrition problems. The relicts of a
previously functional complexity in the landscape in
the form of the hedgerows of North Germany (Mette et
aI., 2001) can reduce nutrient use efficiency at system
level if farmers ignore their role in recycling nutrients
and fertilize the partially shaded parts of their fields
as much as the rest. However, a better understand-
ing of the nutrient stocks and flows in such systems
may lead to a recognition of new functions of woody
landscape elements as filter strips. Filters are only
effective, however, if they remain undersaturated. If
the biomass from filters trips is not regularly removed
by harvesting, filter functions can only be expected
for N and only through denitrification (or ammonia
volatilization). Safetynets only function if there is
unfulfilled demand - they are thus not compatible
with yield maximization of all system components,
as supply is fluctuating with weather conditions, es-
pecially in systems with substantial organic reserves.
The ecological analogue thus has clear limitations if
close-to-maximum crop yields are expected. Where
regular cutting or pruning of filter strips is needed
to maintain their functionality, the biomass so ob-
tained must have sufficient direct value to justify the
labour use for the farmer. Overall, the technical con-
trol paradigm of reducing nutrient excess problems via
unrecognized spatial availability in supply may offer a
more direct approach to the parts of the world where
nutrient excess problems are urgent, but filters may
help in addition to that.
In conclusion, the concepts and tools to help farm-
ers navigate between the scylla of access and the
charibdis of excess problems in plant nutrition cer-
tainly exist, but their use requires an appreciation of
the site-specific interactions and various levels of in-
ternal regulation, rather than a reliance on genetic
modification of plants aimed at transferring specific
mechanisms out of context.
Acknowledgements
Thanks are due to the organizers of the 2001 interna-
tional plant nutrition colloquium for their invitation
to present this review and these views and obtain
feedback from many of the participants to that meet-
ing. Two anonymous reviewers helped to improve an
earlier version of the text.
References
Alphen BJ and Stoorvogel 11 2000 A functional approach to soil
characterization in support of precision agriculture. Soil Sci. Soc.
Am. J. 64, 1706-1713.
Belsky AJ 1994 Influence of trees on savannah productivity: test
of shade, nutrients and tree-grass competition. Ecology 75, 922-
932.
Berkes F, Colding J, and Folke C 2000 Rediscovery of traditional
ecological knowledge as adaptive management. Ecological Ap-
plications. 10, 1251-1262.
Bouwman AF and Booij H 1998 Global use and trade of feedstuffs
and consequences for the nitrogen cycle. Nutrient Cycling in
Agroecosystems 52, 261-267.
Buresh RJ, Sanchez PA and Calhoun F (eds.) 1997 Replenishing
Soil Fertility in Africa. Special pUblication No.5!. Soil Science
Society of America, Madison, Wisconsin, USA.
Burgess SSO, Adams MA, TUrner NC, White DA, Ong CK 2001
Tree roots: conduits for deep recharge of soil water. Oecologia,
126, 158-165.
Cadisch G, Rowe E and van Noordwijk M 1997 Nutrient harvesting
- the tree-root safety net. Agroforestry Forum 8(2), 31-33.
Cadisch G and KE Giller 2000 Soil organic matter management:
The role of residue quality in carbon sequestration and nitro-
gen supply. In Sustainable Management of Soil Organic Mat-
ter. BRees, B Ball, C Campbell and C Watson (eds.), CAB
International, Wallingford: 97-111.
Cakmak I, 2001 Plant nutrition research: priorities to meet hu-
man needs for food in sustainable ways. In WJ Horst, Schenk
MK, Buerkert A et al. (eds) Plant Nutrition - Food Security and
Sustainability of Agroecosystems. Kluwer Academic Publishers,
Dordrecht, the Netherlands. pp. 4-7.
Cassman KG and Plant RE 1992 A model to predict crop response
to applied fertilizer nutrients in heterogeneous fields. Fertilizer
Research 31, 151-163.
Cassman KG, Peng S, Olk DC, Ladha JK, Reichardt W, Dobermann
A and Singh U 1998 Opportunities for increased nitrogen use
efficiency from improved resource management in irrigated rice
systems. Field crops research 56, 7-39.
Claassen N, Steingrobe B and Syring KM 2001 A mechanistic
model to describe the effect of complexing root exudates on
transport and uptake of soil nutrients. In WJ Horst, Schenk
MK, Buerkert A et al. (eds) Plant Nutrition - Food Security and
Sustainability of Agroecosystems. Kluwer Academic Publishers,
Dordrecht, the Netherlands. pp. 600--601.
Collet L and Horst WJ 2001 Effects of heterogeneous Al and P
supply on root growth and screening of maize cultivars differ-
ing in Al resistance. In WJ Horst, Schenk, MK, Buerkert A et
al. (eds) Plant Nutrition - Food Security and Sustainability of
Agroecosystems. Kluwer Academic Publishers, Dordrecht, the
Netherlands. pp. 520--521.
De Vos J A 1997 Water flow and nutrient transport in a layered
silt loam soil. PhD Thesis, Wageningen Agricultural University,
Wageningen, the Netherlands.
De Willigen P and Van Noordwijk M 1987 Roots for plant pro-
duction and nutrient use efficiency, Doctoral thesis Agricultural
University Wageningen. 282 pp.
De Willigen P and Van Noordwijk M 1989 Rooting depth, syn-
chronization, synlocalization and N-use efficiency under humid
tropical conditions. In J van der Heide (ed.) Nutrient Manage-
ment for Food Crop Production in Tropical Farming Systems.
Institute for Soil Fertility, Haren, p 145-156.
De Willigen, P Nielsen, Claassen, NE and Castrignano AM 2000
Modelling Water and Nutrient Uptake. In Smit AL, Bengough

AG, Engels C, Van Noordwijk M, Pellerin S and Van der Geijn
(Eds.) 2000 Root Methods, a Handbook. Springer Verlag, Berlin,
pp. 509-543.
Dobermann A and Cassman KG 2001 Challenges for sustaining
productivity gains and environmental quality in intensive grain
production systems of Asia and the United States. In WJ Horst,
Schenk MK, Buerkert A et al. (eds) Plant Nutrition - Food Se-
curity and Sustainability of Agroecosystems. Kluwer Academic
Publishers, Dordrecht, the Netherlands. pp. 966-967.
EI-Sadek A and Feyen J 2001 Effects of subsurface drainage design
on nitrate leaching and crop yield. In WJ Horst, Schenk MK,
Buerkert A et al. (eds) Plant Nutrition - Food Security and
Sustainability of Agroecosystems. Kluwer Academic Publishers,
Dordrecht, the Netherlands. pp. 934--935.
Ewe! JJ 1999 Natural systems as models for the design of sustain-
able systems ofland use. Agroforestry Systems 45, 1-21.
Fujisaka S 1994 Learning from six reasons why farmers do not
adopt innovations intended to improve sustainability of upland
agriculture. Agricultural Systems 46, 409-425.
Graham RD and Welch RM 2001 Micronutrient interactions in hu-
mans: setting goals for plant breeders and agronomists. In WJ
Horst, Schenk MK, Buerkert A et al. (eds) Plant Nutrition -
Food Security and Sustainability of Agroecosystems. Kluwer
Academic Publishers, Dordrecht, the Netherlands. pp. 286-287.
Gransee A 2001 Effects of root exudates on nutrient availability
in the rhizosphere. In WJ Horst, Schenk MK, Buerkert A et
al. (eds) Plant Nutrition - Food Security and Sustainability of
Agroecosystems. Kluwer Academic Publishers, Dordrecht, the
Netherlands. pp. 626-627.
Hairiah K, Van Noordwijk M, Stulen I, Meijboom FW and Kuiper
PJC 1993 P nutrition effects on aluminium avoidance of Mucuna
pruriens var. utilis. Experimental and Environmental Botany 33,
75-83.
Hairiah K, van Noordwijk M and Cadisch G 2000 Crop yield, C and
N balance of three types of cropping systems on an Ultisol in N
Lampung. Netherlands Journal of Agricultural Science 48, 3-17.
Heinen M and De Willigen P 1995 Dynamics of water and nutri-
ents in closed, recirculating sand bed systems: modelling water
flow, nutrient transport, and root water and nutrient uptake. Acta
Hortic. 401, 501-508.
Hocking PJ and Randall PJ 2001 Better growth and phosphorus nu-
trition of sorghum and wheat following organic acid secreting
crops. In WJ Horst, Schenk MK, Buerkert A et al. (eds) Plant
Nutrition - Food Security and Sustainability of Agroecosystems.
Kluwer Academic Publishers, Dordrecht, the Netherlands. pp.
548-549.
Huxley PA 1999 Tropical Agroforestry. Blackwell Science, Oxford
(UK).
Izac AMN and Sanchez PA 2001 Towards a natural resource man-
agement paradigm for international agriculture: the example of
agroforestry research. Agricultural Systems 69, 5-25.
Joffre R, Rambal S and Ratte JP 1999 The dehesa system of
southern Spain and Portugal as a natural ecosystem mimic.
Agroforestry Systems 45,57-79.
Ketterings QM 1999 Fire as a land management tool in Sepunggur,
Sumatra, Indonesia: can farmers do without it? PhD Thesis, Ohio
State University, Ohio, USA.
Kirk GJD 200 I Use of modelling to understand plant nutrient ac-
quisition. In WJ Horst, Schenk, MK, Buerkert A et al. (eds) Plant
Nutrition - Food Security and Sustainability of Agroecosystems.
Kluwer Academic Publishers, Dordrecht, the Netherlands. pp.
528-529.
Kochian LV 2001 Aluminium and heavy metal toxicity and resist-
ance - lessons to be learnt from similarities and differences. In
39
WJ Horst, Schenk MK, Buerkert A et al. (eds) Plant Nutrition
- Food Security and Sustainability of Agroecosystems. Kluwer
Academic Publishers, Dordrecht, the Netherlands. pp. 442-443.
Laegreid M, Bockman OC and Kaarstad 0 2001 Agriculture, Fertil-
izers and the Environment. CAB International, Wallingford, UK,
294 pp.
Lu W, Zhang F and Cao Y 2001 Mobilization of soil phosphorus
by low-molecular-weight organic acids. In WJ Horst, Schenk
MK, Buerkert A et al. (eds) Plant Nutrition - Food Security and
Sustainability of Agroecosystems. Kluwer Academic Publishers,
Dordrecht, the Netherlands. pp. 554--555.
Mette R, Uckert G and Sattelmacher B 2001 A holistic approach
towards sustainable land use management in the hedgerow -
field crop interface. In WJ Horst, Schenk MK, Buerkert A et
al. (eds) Plant Nutrition - Food Security and Sustainability of
Agroecosystems. Kluwer Academic Publishers, Dordrecht, the
Netherlands. pp. 1026-1027.
Neeteson JJ 2001 Nitrogen and phosphorus management on Dutch
dairy farms: legislation and strategies to meet the regulations.
Biology and Fertility of Soilsin press.
Nye PH and Tinker PB 1977 Solute Movement in the Soil-Root
System. Blackwell, Oxford, 342 pp.
Ong CK and Leakey RRB 1999 Why tree-crop interactions in agro-
forestry appear at odds with tree-grass interactions in tropical
savannahs. Agroforestry Systems 45, 109-129.
Palm CA, Gachego C et al 2001 Organic inputs for soil fertility man-
agement in tropical agroecosystems: Application of an organic
resource database. Agriculture, Ecosystems and Environment 83:
27-42.
Pierce FJ and Novak P, 1999. Aspects of precision agriculture. Adv.
Agronomy 67, 1-85.
Robert PC 2001 Precision agriculture: a challenge for crop nu-
trition management. In WJ Horst, Schenk MK, Buerkert A et
al. (eds) Plant Nutrition - Food Security and Sustainability of
Agroecosystems. Kluwer Academic Publishers, Dordrecht, the
Netherlands. pp. 692-693.
Robinson 2001 Root proliferation, nitrate inflow and their carbon
costs during nitrogen capture by competing plants in patchy soil.
Plant Soil 232, 41-50.
Rodenburg J 1999 Causes and impacts of spatial variability of soil
fertility in the first cropping year after slash-and-burn of second-
ary forest (Indonesia). M.Sc Thesis. Soil Science, Wageningen
University, the Netherlands.
Rowe EC, Hairiah K, Giller KE, Van Noordwijk M and Cadisch G
1999 Testing the safety-net role of hedgerow tree roots by 15N
placement at depths. Agroforestry Systems 43, 81-93.
Rowe EC 1999 The Safety-net Role of Tree Roots in Hedgerow
Intercropping Systems. PhD Thesis, Department of Biological
Sciences, Wye College, University of London (UK).
Sanchez PA 1995 Science in agroforestry. Agroforestry Systems 30,
5-55.Silberbush M and Ben-Asher J 2001 Simulation study of
nutrient uptake by plants from soilless cultures as affected by
salinity buildup and transpiration. Plant Soil 233, 59-69.
Sinclair, FL and Walker D 1998 Acquiring qualitative knowledge
about complex agroecosystems. Part 1: Representation as natural
language. Agricultural Systems 56(3),341-363.
Smaling EMA, Oenema 0 and Fresco LO (Eds.) 1999 Nutrient Dis-
equilibria in Agro-ecosystems: Concepts and Case Studies, CAB
International, Wallingford.
Smith DM, Jackson NA, Roberts JM, Ong CK 1999 Reverse flow of
sap in tree roots and downward siphoning of water by Grevillea
robusta. Functional Ecology 13, 256-264.
Sommer R, de Sa TDA, Vielhauer K, Vlek PLG and FoIster H 2001
Water and nutrient balance under slash-and-burn agriculture in
40
the Eastern Amazon, Brasil - the role of a deep rooting fallow ve-
getation. In WJ Horst, Schenk MK, Buerkert A et a!. (eds) Plant
Nutrition - Food Security and Sustainability of Agroecosystems.
Kluwer Academic Publishers, Dordrecht, the Netherlands. pp.
1014-1015.
Sparovek, G., Ranieri, S.B.L., Gassner, A, De Maria, I.e., Schnug,
E., Santos, RF dos, Joubert, A (2002) A conceptual frame-
work for the definition of the optimal width of riparian forests.
Agricultural, Ecosystems & Environment (in press).
Suprayogo D 2000. Testing the Safety-net Hypothesis in Hedgerow
Intercropping Water Balance and Mineral-N Leaching in the
Humid Tropics. PhD Thesis, Department of Biology, Imperial
College of Science, Techonology and Medicine, University of
London (UK).
Suprayogo D, Hairiah' K, van Noordwijk' M, Cadisch G 2002 The
inherent 'safety-net' of Ultisols: measuring and modelling re-
tarded leaching of mineral nitrogen. European J. Soil Science 53,
1-10.
Trenbath BR 1999 Multispecies cropping systems in India. Agro-
forestry Systems 45, 81-107.
Vandermeer J, Van Noordwijk M, Ong C, Anderson J and Perfecto
Y 1998 Global change and multi-species agroecosystems: con-
cepts and issues. Agriculture, Ecosystems and Environment 67,
1-22.
Vanlauwe B, Sanginga N and Merckx R 2001 Alley cropping with
Senna siamea in South-western Nigeria: II. Dry matter, total N,
and urea-derived N dynamics of the Senna and maize roots. Plant
Soil 231, 201-210.
Vanlauwe, B., J. Diels, et al. 2001. Direct interactions between
N fertilizer and organic matter: evidence from trials with 15N
labelled fertilizer. In: B. Vanlauwe and R Merckx, (Eds.) Bio-
logical Nutrient Management. CAB International, Wallingford
(UK).
Van Os EA 1999 Closed soilless growing systems: a sustainable
solution for Dutch greenhouse horticulture. Water Sci. Techno!'
39,105-112.
Van Noordwijk 1983 Functional interpretation of root densities in
the field for nutrient and water uptake. Wurzeliikologie und ihre
Nutzanwendung. pp. 207-226.
Van Noordwijk M 1989 Rooting depth in cropping systems in the
humid tropics in relation to nutrient use efficiency. In J. van der
Heide (ed.) Nutrient Management for Food Crop Production in
Tropical Farming Systems. Institute for Soil Fertility, Haren. p
129-144.
Van Noordwijk M 1990 Synchronization of supply and demand is
necessary to increase efficiency of nutrient use in soilless horti-
culture.In ML Van Beusichem (ed.) Plant Nutrition - Physiology
and Applications. Kluwer Academic Pub!., p. 525-531.
Van Noordwijk M 1999 Nutrient cycling in ecosystems versus nutri-
ent budgets of agricultural systems. In EMA Smaling, OOenema
and LO Fresco (Eds.) Nutrient Disequilibria in Agro-ecosystems:
Concepts and Case Studies, CAB International, Wallingford. pp.
1-26.
Van Noordwijk M and Garrity DP 1995 Nutrient use efficiency in
agroforestry systems. In Potassium in Asia: Balanced Fertiliz-
ation to Increase and Sustain Agricultural Production. Interna-
tional Potash Institute, Basel, Switzerland. pp. 245-279.
Van Noordwijk M and Lusiana B 1999 WaNuLCAS, a model
of water, nutrient and light capture in agroforestry systems.
Agroforestry Systems 43, 217-242.
Van Noordwijk M and Lusiana B 2000 WaNuLCAS 2.0, Back-
ground on a model of water nutrient and light capture in
agroforestry systems. International Centre for Research in Agro-
forestry (ICRAF), Bogor, Indonesia. 186 pp.
Van Noordwijk M and Ong CK 1999 Can the ecosystem mimic hy-
potheses be applied to farms in African savannahs? Agroforestry
Systems 45, 131-158.
Van Noordwijk M and Scholten JHM 1994 Effects offertilizer price
on feasibility of efficiency improvement: case study for an urea
injector for lowland rice. Fertilizer Research 39, 1-9.
Van Noordwijk M and Wadman W 1992 Effects of spatial variab-
ility of nitrogen supply on environmentally acceptable nitrogen
fertilizer application rates to arable crops. Netherlands Journal of
Agricultural Science 40, 51-72.
Van Noordwijk M, Widianto, Heinen M and Hairiah K 1991 Old
tree root channels in acid soils in the humid tropics: import-
ant for crop root penetration, water infiltration and nitrogen
management. Plant Soil 134, 37-44.
Van Noordwijk M, Brouwer G and Harmanny K 1993 Concepts
and methods for studying interactions of roots and soil structure.
Geoderma 56, 351-375.
Van Noordwijk M, Dijksterhuis G and Van Keulen H 1994 Risk
management in crop production and fertilizer use with uncertain
rainfall: how many eggs in which baskets. Netherlands Journal
of Agricultural Science 42, 249-269.
Van Noordwijk M, Lawson G, Groot JJR, Hairiah K 1996 Root
distribution in relation to nutrients and competition. In CK Ong
and PA Huxley (Eds.). Tree-Crop Interactions - a Physiological
Approach. CAB International, Wallingford. pp. 319-364.
Van Noordwijk M, Van Roode M, McCallie EL and Lusiana B
1998 Erosion and sedimentation as multiscale, fractal processes:
implications for models, experiments and the real world. In F
Penning de Vries, FAgus and J Kerr (Eds.) Soil Erosion at Mul-
tiple Scales, Principles and Methods for Assessing Causes and
Impacts. CAB International, Wallingford. pp. 223-253.
Van Noordwijk M, Radersma Sand Lusiana B 1999 Modelling root
architecture and phosphorus uptake in agroforestry. Agroforestry
Forum 9(4),28-30.
Van Noordwijk M, Tomich TP and Verbist B 2001 Negotiation sup-
port models for integrated natural resource management in forest
margins and in landscapes with trees: can they solve local as well
as global problems? Conservation Ecology ,5(2): 21.
Wahid PA 2001 Radioisotope studies of root activity and root-level
interactions in tree-based production systems: a review. Applied
Radiation and Isotopes 54, 715-736.
Wang GH, Dobermann A, Witt C, Sun QZ, Fu RX, 2001. Perform-
ance of site-specific nutrient management for irrigated rice in
southeast China. Agron. J. 93: 869-878.
Whitmore APM and Van Noordwijk M 1995 Bridging the gap
between environmentally acceptable and economically desirable
nutrient supply. In DM Glen, MP Greaves and HM Anderson
(eds.) Ecology and Integrated Farming Systems, John Wiley &
Sons. pp. 271-288.
Whitmore APM, Cadisch G, Toomsan B, Limpinuntana V, Van
Noordwijk M and Pumomosidhi P, 2000 An analysis of the eco-
nomic values of novel cropping systems in N.E. Thailand and S.
Sumatra. Neth. J. Agricultural Science 48, 105-114.
Wong MTF, Hughes, T and Rowell D L 1990 Retarded leaching of
nitrate in acid soils from the tropics: measurement of the effective
anion exchange capacity. Journal of Soil Science 41, 655-663.
Zhang FS, Li L and Sun JH 2001 Contribution of above- and be-
low ground interactions to intercropping. In W.J. Horst, Schenk,
M.K., Buerkert A et al. (eds) Plant Nutrition - Food Security and
Sustainability of Agroecosystems. Kluwer Academic Publishers,
Dordrecht, the Netherlands. pp. 978-979.
Genetics and molecular biology of plant nutrition
 Plant and Soil 247: 43-54, 2002.
© 2002 Kluwer Academic Publishers.
43

Molecular mechanisms of potassium and sodium uptake in plants

Pascal Maser, Markus Gierth & Julian I. Schroeder!

Cell and Developmental Biology Section and Center for Molecular Genetics, Division of Biology, University of
California, San Diego, 9500 Gilman Drive, La Jolla, CA92093-0116, USA. 1 Corresponding author*

Key words: Arabidopsis thaliana, ion transport, potassium, salt sensitivity, sodium

Abstract
Potassium (K+) is an essential nutrient and the most abundant cation in plants, whereas the closely related ion
sodium (Na+) is toxic to most plants at high millimolar concentrations. K+ deficiency and Na+ toxicity are both
major constraints to crop production worldwide. K+ counteracts Na+ stress, while Na+, in tum, can to a certain
degree alleviate K+ deficiency. Elucidation of the molecular mechanisms of K+ and Na+ transport is pivotal to
the understanding - and eventually engineering - of plant K+ nutrition and Na+ sensitivity. Here we provide an
overview on plant K+ transporters with particular emphasis on root K+ and Na+ uptake. Plant K+ -permeable
cation transporters comprise seven families: Shaker-type K+ channels, 'two-pore' K+ channels, cyclic-nucleotide-
gated channels, putative K+ IH+ antiporters, KUPIHAKlKT transporters, HKT transporters, and LCT1. Candidate
genes for Na+ transport are the KUPIHAKlKTs, HKTs, CNGCs, and LCT]. Expression in heterologous systems,
localization in plants, and genetic disruption in plants will provide insight into the roles of transporter genes in K+
nutrition and Na+ toxicity.

Abbreviations: AKT - Arabidopsis K+ transporter; CNGC - cyclic-nucleotide-gated channel; HAK - high-affinity

K+ transporter; HKT - high-affinity K+ transporter; KCO - K+ channel outward-rectifier; KEA - K+ exchange
antiporter; KT - K+ transporter; KUP - K+ uptake permease; LCT -low-affinity cation transporter; SKOR - stelar
K+ outward-rectifier; Trk - K+ transporter

Introduction is toxic at high millimolar concentrations. Soil sa-

Antagonisms and synergisms between K+ and Na+
Potassium (K+) is an important macronutrient and the
most abundant cation in higher plants. K+ is essential
for enzyme activation, protein synthesis and photo-
synthesis, and it mediates osmoregulation during cell
expansion, stomatal movements and tropisms. Fur-
thermore, K+ is necessary for phloem solute transport
and for the maintenance of cation: anion balance in the
cytosol as well as in the vacuole. K+ supply from soil
can be rate-limiting for agricultural production. The
closely related cation sodium (Na+), in contrast, does
not generally fulfill physiological functions. Certain
halophytes require Na+ for growth, but for the ma-
jority of plants - including most crop plants - Na+
* FAX No: +1-858-534-7108.
E-mail: julian@biomail.ucsd.edu
linization due to irrigation is becoming increasingly
detrimental to agriculture (Flowers, 1999).
Na+ toxicity in plants is multifactorial, includ-
ing osmotic stress (Tarczynski et aI., 1993), inhibi-
tion of vital enzymes (Munns, 1993; Murguia et aI.,
1995), and competition with K+ in general. There
are substantial variations in salt sensitivity among
glycophytes, between different species and between
different cultivars or ecotypes of the same species
(Greenway and Munns, 1980), illustrating that Na+
resistance is a quantitative trait. Among grain crops,
for instance, barley (Hordeum vulgare) is the most
salt tolerant. Seedlings of the barley cultivar Califor-
nia Mariout showed no growth retardation at 400 mM
NaCl while other H. vulgare genotypes were severely
affected at lower salt concentrations (Epstein et aI.,
1980). Growth of Arabidopsis thaliana is severely
impaired at Na+ concentrations above about 100 mM.
44
Na+ can to a certain degree replace K+, partic-
ularly in its osmotic functions in the vacuole. Thus,
under K+ starvation addition ofNa+ may actually pro-
mote plant growth. The grain yield of rice grown at 25
p,M K+ more than doubled in the presence of 43 mM
Na+ (Mengel and Kirkby, 1982). Addition of Na+
also increased dry weight of Arabidopsis at 10 p,M K+
(Maathuis and Sanders, 1993). Na+ can exert benefi-
cial effects on plant growth even when K+ supply is
not limiting. In sugar beet, replacement of 5 mM K+
in the nutrient substrate with 2.5 mM of each K+ and
Na+ increased the plant dry weight and the sucrose
content in the storage root (Marschner et aI., 1981).
Furthermore, Na+ may improve the plant's water bal-
ance. Stomata of sugar beet leaves closed more rapidly
in response to drought stress when the plants were sup-
plied with Na+ plus K+ compared to K+ only (Hampe
and Marschner, 1982). Na+ -induced stomatal closure
in the halophyte Aster tripolium has been proposed as
a mechanism to minimize shoot Na+ accumulation by
reduction of the transpiration rate (Very et aI., 1998).
The extent to which Na+ can replace K+ varies
between different plant species, different cultivars of
the same species, and even between different leaves
of the same plant, younger leaves relying more on K+
than older ones (Lindhauer et aI., 1990). Interestingly,
halophytes require less K+ for growth than glyco-
phytes (Marschner, 1995), indicating a link between
the ability to substitute K+ with Na+ and salt toler-
ance. Potassium substitution and growth stimulation
by Na+ are of great interest for agronomic production,
for the improvement of both crop plants and fertilizers.
Transport processes determine [K+ ]:[Na+] balance
Plants are generally divided into Na+ -includers and
Na+ -excluders, with the majority of halophytes be-
longing to the Na+ -includers (Marschner, 1995).
These plants take up large amounts of Na+, which
is then sequestered into vacuoles of shoot and root
cells. Glycophytes are predominantly Na+ -excluders,
exhibiting an inverse relationship between Na+ uptake
and salt tolerance (Greenway and Munns, 1980). Na+
exclusion mechanisms can include Na+/H+ antiport
at the plasma membrane of wheat roots (Allen et aI.,
1995). Interestingly, cytosolic Na+ concentrations are
similar in glycophytes and halophytes. Maintenance of
a high cytosolic [K+]:[Na+] ratio is therefore critical
for the function of cells in general (Rubio et aI., 1995;
Zhu et aI., 1998).
Classification of plants as either Na+ -includers or
Na+ -excluders is somewhat artificial since the two
strategies are not mutually exclusive; reduced net Na+
uptake by roots and increased Na+ sequestration in
the shoots might work in parallel. Overexpression of
the vacuolar Na+/H+ exchanger AtNHX1 markedly
decreased the Na+ sensitivity of the glycophyte Ar-
abidopsis thaliana (Apse et aI., 1999), and the Ara-
bidopsis SOS1 (salt overly sensitive) locus has recently
been found to encode a putative Na+/H+ exchanger
(Shi et al., 2000). Interplay between H+ -ATPases
and Na+ /H+ exchangers in the plasma membrane and
the tonoplast is crucial for Na+ extrusion and Na+
sequestration, respectively (Blumwald et aI., 2000).
Transport processes are central to plant K+ nutri-
tion and Na+ toxicity since accumulation of K+ and
Na+ from soil and their distribution throughout di-
verse plant tissues is mediated by transporter proteins.
Elucidation and dissection of the molecular mechan-
isms by which root and other plant cells take up K+
and Na+ is therefore essential for understanding plant
growth and the functional basis of cation-uptake as-
sociated stresses such as K+ deficiency and salinity
stress. Based on the hypothesis that Na+ enters roots
through K+ and Ca2+ transporters we provide here
an overview on plant K+ -permeable cation transport-
ers with particular emphasis on root K+ and Na+
transport.
K+ transporters and their roles in root K+ and
Na+ uptake
The accumulation of K+ by the plant root symplast
imposes a substantial energetic cost and requires spe-
cialized transport systems (Gierth et aI., 1998). Trans-
membrane movement of K+ is catalyzed by channel
and transporter proteins, and energized by the negative
membrane potential of plant cells. In addition, trans-
porters may use H+ and in some cases Na+ gradients
or ATP hydrolysis to energize K+ translocation. Since
the molecular identification of initial plant K+ channel
and transporter cDNAs (Anderson et aI., 1992; Sen-
tenac et aI., 1992; Schachtman and Schroeder, 1994),
a number of plant genes encoding K+ transporters
have been cloned and characterized by expression in
S. cerevisiae, Xenopus oocytes and E. coli. Trans-
porter proteins frequently resist experimental investig-
ation due to their large size and hydrophobic nature,
yet the same characteristics render them more eas-
ily detectable in silico. The recent completion of the

CNGC
KCO
AKT
Figure 1. Plant K+ -penneable cation transporters. There are at least seven families of K+ -penneable transporters in plants that have been
characterized to date (see text for details). The predicted structures of K+ channels, CNGCs and HKTI transporters are supported by exper-
imental evidence (Doyle et al., 1998; Henn et al., 1995; Kato et al., 2001; Lopes et al., 2001; Sokolova et al., 2001; Uozumi et al., 1998).
Structure prediction for KUPIHAKlKTs, LCT and KEAs was perfonned with the HMMTOP program (Tusnady and Simon, 1998); whether
these proteins contain loop structures at their outer surface remains to be investigated.
Arabidopsis genome sequencing project offered the
opportunity to make an inventory of all of a plant's
putative transporter proteins (Ward, 2001). A genome-
wide survey revealed six major families of Arabidop-
sis cation transporters that are, with varying selectiv-
ity, permeable to K+: Shaker-type K+ channels (nine
genes), 'two-pore' K+ channels (six genes), cyclic-
nucleotide-gated channels (20 genes), putative K+ IH+
antiporters (six genes), KUPIHAKlKT transporters
(13 genes) and HKT transporters (one gene) (Maser
et al., 2001). An additional K+ -permeable transporter,
LCT 1, has been identified in wheat (Schachtman et aI.,
1997) (Figure 1).
Shaker-type K+ channels
The high selectivity of potassium channels for K+
over other monovalent cations (Heginbotham and
MacKinnon, 1993) is given by four pore-forming P-
loops at the outer surface of the membrane protein
(Doyle et aI., 1998). Shaker-type K+ channels (Tem-
pel et al., 1987) are tetramers and in animals and
plants have one P-Ioop and six transmembrane (TM)
domains per subunit (Uozumi et al., 1998; Sokolova
et aI., 2001). TM domains 5 and 6 surround the P-
loop, while TM domain four functions as a voltage
sensor: It contains charged residues and consequently
moves in response to voltage change (Glauner et al.,
1999), thereby controlling channel gating. Shaker-
type channels are subdivided into depolarization-
activated 'outward-rectifiers' and hyperpolarization-
activated 'inward-rectifiers'. Inward-rectifying K+
channels that are expressed in roots are candidate
45
HKT KUP/HAKlKT
LCT1
KEA
Cytoplasm
genes for K+ uptake (Maathuis and Sanders, 1995;
Schroeder et al., 1994). The high selectivity of inward-
rectifying channels for K+ indicates that their contri-
bution to Na+ uptake may be negligible. Less selective
outward-rectifying K+ channels have been implicated
in Na+ influx upon salt-induced depolarization of root
cells (Maathuis and Sanders, 1995; Schachtman et
aI., 1991). However, how substantial amounts of Na+
could be taken up by K+ efflux channels remains un-
known; further research would be needed to test this
model.
The first K+ channels cloned from plants were
the Shaker-type KATl (Anderson et al., 1992) and
AKTl (Sentenac et aI., 1992). Further channels from
Arabidopsis were identified subsequently (Ache et
al., 2000; Cao et al., 1995; Gaymard et aI., 1998;
Ketchum and Slayman, 1996), and a genome-wide
survey revealed a total of nine genes for Shaker-type
K+ channels (Maser et aI., 2001). K+ channels have
been extensively studied in the context of stomatal
movements (Schroeder, 1988; Schroeder et aI., 1984;
Thiel et aI., 1992; reviewed in Schroeder et al., 2001).
So far two channels have been shown to be involved
in Arabidopsis K+ nutrition: AKTl and SKOR. SKOR
(Stelar K+ Outward Rectifier) is expressed in stelar
tissues of the root. Plants disrupted in SKOR exhib-
ited lower K+ contents in the xylem sap and in shoot
tissues (Gaymard et al., 1998), indicating that SKOR
mediates K+ release into the xylem sap.
AKTl is a hyperpolarization-activated Shaker-type
K+ channel (BertI et aI., 1994) that is expressed in
roots (Lagarde et aI., 1996). Arabidopsis lines car-
rying a T-DNA insertion in AKTl were impaired in
46
Table 1. Nomenclature of Arabidopsis KUPIHAKlKT potassium
growth at low K+ concentrations, took up less 86Rb+
transporters. The Arabidopsis genome encodes for 13 different
than control plants, and their root cells lacked inward- KUPIHAKlKT proteins, which are lacking uniform naming and
rectifying K+ currents (Hirsch et aI., 1998). The numbering. Note that AtKUP5 and AtHAK5 are two different pro-
finding that growth of l::!.aktl plants was impaired teins, AtKUP3 is the same as AtKT4, and AtKUP4 is AtKT3.
Protein accession numbers are from the Arabidopsis Genome Ini-
at 10 /LM K+ (Hirsch et al., 1998) suggested that tiative (AGI) database (AtDB) and from GenBank (National Center
AKTI mediates root K+ uptake from external con- for Biotechnology Information). References: (1) Quintero and Blatt,
centrations in the micromolar range and demonstrated 1997; (2) Kim et al., 1998; (3) Fu et al., 1998; (4) Rigas et al., 2001;
that a channel can contribute to 'high-affinity' K+ (5) Rubio et al., 2000
uptake provided the membrane potential is negative
Name Synonyms AGIAcc. GenBankID Refs.
enough. Interestingly, l::!.aktl plants showed no growth
phenotype in the absence of NHt, and K+ permeab- AtKUP1 KT1 At2g30070 AAB88901 1,2,3
ility of l::!.aktl root cells was inhibited by NHt in a AtKUP2 KT2 At2g40540 AAC49845 1,2
dose-dependent manner (Spalding et aI., 1999). These AtKUP3 KT4 At3g02050 AAF19432 1,2
AtKUP4 KT3, TRH1 At4g23640 CAC16137 1,2,4
data suggest that non-AKTl pathways for K+ ab-
AtKUP5 KT5 At4g33530 CAB79319 1
sorption are inhibited by ammonium and that AKTI
AtHAK5 At4g13420 AAF36490 5
may function as a backup K+ transporter at low K+
AtHAK6 At1g70300 AAC18809 5
concentrations, when other transporters are disabled.
AtHAK7 AtSg09400 CAC05466 5
Non-AKTl-mediated depolarization of root cells by
AtHAK8 AtSg14880 CAC01887 5
K+ was stimulated by the addition of 2 mM Na+, AtKUP9 At4g19960 CAB78996
indicating that alternative pathways include Na+/K+ AtKUPlO At1g31120 AAD21693
symport (Spalding et aI., 1999). AtKUPll At2g35060 AAC12845
AKTl homologs have recently been identified AtKUP12 At1g60160 AAC24049
from tomato, potato (Hartje et aI., 2000), wheat
(Triticum aestivum) (Buschmann et al., 2000) and
ice plant (Mesembryanthemum crystallinum) (Su et
al., 2001). Expression of TaAKTl in wheat roots
reported. Whether KCOs contribute to root K+ or Na+
was induced upon K+ withdrawal (Buschmann et al.,
transport remains to be investigated.
2000). McAKTl (=Mktl) from the halophyteM. crys-
tallinum was shown to be root-specific and it localized Cyclic-nucleotide-gated channels
to the plasma membrane (Su et aI., 2001). McAKTl
mRNA and McAKTl protein were down-regulated in Cyclic-nucleotide-gated channels (CNGCs) are pre-
response to salt stress (Su et al., 2001). dominantly studied in the context of olfactory and
visual signal transduction in animals (Zagotta and
'Two-pore' K+ channels Siegelbaum, 1996). CNGCs are gated by intracel-
lular cGMP or cAMP (Broillet and Firestein, 1999)
'Two-pore' K+ channels were first identified in yeast and their activity is further modulated by calcium
(Ketchum et aI., 1995) and were later found in various and calmodulin (Kleene, 1999). Structurally related to
organisms including humans (Fink et al., 1996; Lesage Shaker-type K+ channels (Henn et aI., 1995), CNGCs
et al., 1996). 'Two-pore' K+ channels consist of sub- exhibit much broader substrate specificities including
units with two P-Ioops each. The functional channel is Ca2+ permeability (Heginbotham et al., 1992). Their
likely to be a dimer with one pore, with a similar struc- permeabilities to K+ and Na+ are almost the same
ture to a tetramer of one-P-Ioop subunits (Lopes et (Gamel and Torre, 2000). Plant CNGCs are therefore
al., 2001). The first 'two-pore' K+ channel described candidate genes for Na+ uptake.
from plants is Arabidopsis KCOl (Czempinski et aI., Plant CNGC homologs have been identified from
1997). When expressed in insect cells KCO 1 exhibited barley (Schuurink et aI., 1998), tobacco (Arazi et
K+ -selective currents upon depolarization, indicative aI., 1999) and Arabidopsis (Kohler et aI., 1999;
of an outward-rectifying K+ channel (Czempinski et Leng et al., 1999). Arabidopsis has a total of 20
aI., 1997). KCOI is expressed in seedlings, leaves and AtCNGC genes which can be divided into four groups
flowers (Czempinski et al., 1997). There are an ad- (Maser et aI., 2001). Interestingly, genetic disruption
ditional five KCO genes in Arabidopsis (Maser et al., of AtCNGCl caused Pb 2+ resistance in Arabidop-
2001), but their expression patterns have not yet been sis (Sunkar et al., 2000), while overexpression of a
 47

Table 2. Candidate genes for root Na+ uptake. their substrate specificities and expression pat-
terns. Candidate genes for root Na+ uptake are found in the HKTl and KUPIHAKlKT families,
LeTl, and among cyclic-nucleotide-gated channels. Na+ transport by AtKUPI and AtHAK5 is
proposed based on inhibition of K+ transport by Na+. Na+ transport by plant CNGCs has not
yet been investigated, but animal CNGCs were shown to transport Na+. Abbreviations: Ta-
Triticum aestivum; Hv - Hordeum vulgare; At - Arabidopsis thaliana; Os - Oryza sativa; n.d.
- not determined). References: (1) Schachtman and Schroeder, 1994; (2) Wang et al., 1998; (3)
Uozumi et al., 2000; (4) Horle et al., 2001; (5) Santa-Maria et al., 1997; (6) Rubio et al., 2000;
(7) Kim et aI., 1998; (8) Fu and Luan, 1998; (9) Rigas et aI., 2001; (10) Schachtman et aI., 1997;
(11) Kohler et aI., 1999; (12) Qu et al., 2000.

Protein K+ Na+ low [K+j Localization Ref.

transport transport induced

HKTl + + + Root cortex, shoot vasculature 1,2

HvHKTl n.d. n.d. + n.d. 2
AtHKTl + Mainly root 3
OsHKTl + + Root 4
OsHKT2 + + + Root 4

HvHAKI + + + Root 5,6

AtKUPI + (+) Shoot 7,8
AtKUP2 + n.d. Root, shoot 7
AtKUP3 n.d. n.d. + Root, shoot 7
AtKUP4 + n.d. n.d. Root, shoot 9
AtHAK5 + (+) + (shoot) Root, shoot6

LCTl + + n.d. Root, leaves 10

CNGCs + (+) n.d. n.d. 11, 12

tobacco homolog NtCBP4 (calmodulin-binding pro-
tein) enhanced Pb 2 + accumulation of tobacco plants
and rendered them Pb 2 + hypersensitive (Arazi et aI.,
1999). NtCBP4 has been localized to the plasma mem-
brane of tobacco cells (Arazi et aI., 1999). These find-
ings indicate that plant CNGCs are involved in cation
uptake (Arazi et aI., 2000) and that CNGCs should
be included to the list of potential Na+ uptake trans-
porters in plants. Recently, cAMP and cGMP were
shown to inhibit nonselective, voltage-independent
cation currents when added to the cytosolic side of
Arabidopsis root protoplast patches (Maathuis and
Sanders, 2001). Furthermore, membrane-permeable
cAMP and cGMP analogs inhibited root Na+ up-
take and increased the salt tolerance in Arabidopsis
(Maathuis and Sanders, 2001).
K+m+ antiporters
The genome of Arabidopsis harbors SIX genes en-
coding putative K+ IH+ exchangers of the CPA2
(cation:proton antiporter) family (Maser et al., 2001).
The genes were named KEAl (K+ exchange anti-
porter; W. Yao, N. Hadjeb and G. A. Berkowitz,
GenBank accession AF003382) to KEA6, but their
function remains to be investigated. Bacterial K+ IH+
antiporters are involved in defense against toxic elec-
trophiles by mediating acidification of the cytosol
(Munro et aI., 1991). In plants, K+IH+ antiporters
might play important roles in K+ homeostasis by load-
ing K+ into vacuoles or other acidic compartments
(Sze et aI., 1992). Localization of KEAs in Arabidop-
sis and functional characterization in heterologous ex-
pression systems will be necessary to determine their
physiological roles.
KUPIHAKlKT potassium transporters
K+ !!ptake transporters from bacteria named KUP
(Schleyer and Bakker, 1993) and high-~ffinity K+
transporters from fungi named HAK (Banuelos et aI.,
1995; Haro et al., 1999) comprise a family of K+
transporters that is also present in plants. Genes encod-
ing members of this family were cloned from barley
48

0.1
1MCHAK211 AtKUP21 ~~..
.----_J(~·~m~:t..!1I.11*~F~1
1AtKUP11
1AtKUP111
1AtKUP1 0 1 toO

l·mUMlfJ
l.m:t!1tlll toO

1AtKUP91

IIIIM31
b''4I6''1'
1AtHAKS1
I AtHAK7I
I AtKUPSI

IVfHAK11
I AtKUP12I l.mU-nlf'
Figure 2. Plant KUPIHAKlKT potassium transporters. There are 13 members of the KUPIHAKlKT family in Arabidopsis (Table 1). In addition,
full-length KUPIHAKlKT sequences are available from barley (H. vulgare), ice plant (M. crystallinum), poplar (Populus tremulaxtremuloides),
reed (Phragmites communis), sea grass (Cymodocea nodosa), bean (Viciafaba), and rice (0. sativa). Proteins from dicots are printed in black,
while those from monocots are printed in white on black background. In case of overlapping KT and KUP names (see Table 1), KUP names are
given for clarity. The tree was constructed with ClustalX (Thompson et al., 1997) and plotted with Treeview (Page, 1996). Numbers indicate
branch occurrence (in percent) during 1000 bootstrap runs. GenBank accessions: HvHAKl, AAC39315; HvHAK2, AAF36491; PtKUPl,
CAC39168; McHAKl, AAK53758; McHAK2, AAK53759; McHAK3, AAK53760; PcHAKl, BAB32442; CnHAKl, CAD20318; CnHAK2,
CAD20319; VfHAKl, CAD20577; OsHAK2, BAB67929; OsHAK3, CAD20995; OsHAK7, CAD20992; OsHAK9, CAD20999; OsHAKlO,
CAD20993; OsHAKll, CAD21001; OsHAK12, CAD21002; OsHAK14, CAD21004; OsHAK15, CAD21005; OsHAK17, CAD20996.

and Arabidopsis and named HAK (Rubio et aI., 2000;
Santa-Maria et aI., 1997), KT (Quintero and Blatt,
1997), or KUP (Fu and Luan, 1998; Kim et aI., 1998).
The Te system for classification of transporter pro-
teins refers to this family as KUPs (Saier, 1999). Table
1 summarizes the current nomenclature of Arabidopsis
KUPIHAKlKT genes.
KUPIHAKlKT genes are found in EST collec-
tions from maize, rice, soybean, tomato, cotton fiber,
onion, poplar, glycine,. and ice plant, and they ap-
pear to comprise multigene families in all those plant
species. A phylogenetic tree of KUPIHAKJKT trans-
porters reveals that individual proteins from monocots
(Liliopsida) and dicots (Magnoliopsida) intermingle
(Figure 2). Thus the expansion of the KUPIHAKlKT
gene family appears to have taken place before the sep-
aration of monocots and dicots. However, the phylo-
genetic tree also shows more recent gene duplication
events such as AtKUP10 and AtKUP11, or OsHAK11
and OsHAK12 (Figure 2, top left). Arabidopsis has 13
KUPIHAKlKT genes (Table 1, Figure 2). This variety
may indicate functional redundancies; however, there
is also evidence for functional diversification among
individual KUPIHAKlKT genes. Genetic disruption of
AtKUP4 (AtKT3) resulted in a 'tiny root-hair' (trh)
phenotype (Rigas et aI., 2001), while a semi-dominant
point mutation in AtKUP2 causes a short hypocotyl
and small leaves (Elumalai et aI., 2002).
The KUPIHAKJKT family comprises several can-
didates for root K+ and Na+ uptake transporters
(Table 2). AtKUP3 (Kim et aI., 1998), AtKUP4 (Rigas
et aI., 2001), AtHAK5 (Rubio et aI., 2000) and bar-
ley HvHAKl (Santa-Maria et aI., 1997) are expressed
in roots, and expression of AtKUP3 and HvHAKl is
induced by K+ starvation (Santa-Maria et aI., 1997;
Kim et aI., 1998). HvHAK1, AtKUP1, AtKUP4 and
AtHAK5 were shown to exhibit high-affinity K+ up-
take in S. cerevisiae (Fu and Luan, 1998; Rigas et

aI., 2001; Rubio et aI., 2000; Santa-Maria et aI.,
1997) and in transgenic Arabidopsis cells (Kim et aI.,
1998). Interestingly, HvHAKl mediated low-affinity
Na+ transport in addition to high-affinity K+ transport
when expressed in yeast (Santa-Maria et aI., 1997).
K+ transport in S. cerevisiae by AtKUP1 (Fu and
Luan, 1998) and AtHAK5 (Rubio et aI., 2000) was
competitively inhibited by Na+. These data indicate
a possible role for KUPIHAKlKT transporters in salt
stress (Table 2).
HKT transporters
HKT, yeast Trk and bacterial KtrB transporters are all
members of one superfamily. Trk transporters are the
main high-affinity K+ uptake systems in S. cerevisiae
(Gaber et aI., 1988) and S. pombe (Soldatenkov et
aI., 1995). TrkIHKT transporters have four P-Ioop-
like domains in a single polypeptide chain (Durell and
Guy, 1999), an architecture that is also found in many
bacterial cation transport systems: TrkH K+ transport-
ers (Schlosser et al., 1995; Nakamura et aI., 1998),
KtrB subunits of Na+/K+ transporters (Kawano et aI.,
2000; Tholema et aI., 1999), NtpJ subunits of Na+-
ATPases, and KdpA subunits of K+ -ATPases (Durell
et al., 2000; Laimins et aI., 1978). The first plant Trk-
like transporter, HKTl, was identified from wheat and
mediated high-affinity K+ uptake in yeast (Schacht-
man and Schroeder, 1994). HKTl homologs were
cloned from Arabidopsis (Uozumi et aI., 2000), Eu-
calyptus (Fairbairn et aI., 2000) and rice (Horie et aI.,
2001). In contrast to KUPIHAKlKTs, HKTl genes do
not constitute multigene families in plants; AtHKTl
from Arabidopsis and HKT1 from a diploid wheat pro-
genitor line are both single-copy genes (Schachtman
and Schroeder, 1994; Uozumi et aI., 2000), while
Eucalyptus and rice each have two HKT paralogues
(Fairbairn et aI., 2000; Horie et aI., 2001).
Wheat HKT1 was found to function as a Na+/K+
symporter when expressed in S. cerevisiae and in
Xenopus oocytes (Rubio et al., 1995). Coupling of
K+ to Na+ translocation could provide a means to
energize potassium absorption from low [K+] (see
Introduction). The structure-function relationship of
HKTI was investigated by developing a non-biased
genetic screen for selecting mutant transporters. By
functional selection in a sodium-hypersensitive and
K+ uptake deficient yeast strain several point muta-
tions in HKT1 have been identified that reduce Na+
transport without affecting K+ transport (Rubio et aI.,
1995, 1999). The mutations producing the strongest
49
reduction in Na+ translocation were located in the
predicted P-Ioops of HKTl (Gassmann et aI., 1996;
Rubio et aI., 1995, 1999). Furthermore, Na+ coupling
by HKT1 was reduced by site-directed mutagenesis
in the fourth predicted P-Ioop (Diatloff et aI., 1998).
These data provide further evidence for a role of the P-
loops in cation selectivity of TrkIHKT transporters and
for the evolutionary relationship between HKT trans-
porters and K+ channels (Durell et aI., 1999). The pre-
dicted topology based on the structure of K+ channels
has recently been confirmed experimentally for Ara-
bidopsis AtHKTl (Kato et aI., 2001). These findings
correlate with the hypothesis that HKT transporters
may have channel modes of ion transport (Gassmann
et aI., 1996).
Wheat HKTl is expressed in the root cortex
(Schachtman and Schroeder, 1994) and it functions as
a low affinity Na+ transporter under low [K+]:[Na+]
ratios (Rubio et aI., 1995). K+ starvation induces
HKTl expression in wheat and barley (Wang et aI.,
1998) as well as inward Na+ currents in wheat
root cortex cells (Buschmann et aI., 2000). Consist-
ent with these findings HKTl is expressed in wheat
root cortex cells (Schachtman and Schroeder, 1994).
Moreover, in a comparison of the salt-tolerant rice cul-
tivar Pokkali with a salt-sensitive one (IR-29), Na+
tolerance in Pokkali correlated with reduced Na+ up-
take and decreased expression of HKTl during salt
stress (Golldack et aI., 1997). HKTl is therefore a
candidate gene for root Na+ transport and has been
proposed to be a determinant of salt sensitivity in
plants (Rubio et aI., 1995) (Table 2).
Surprisingly, Arabidopsis AtHKTl did not com-
plement a Trkl,2 deletion when expressed in S.
cerevisiae, nor did it mediate K+ -dependent currents
in Xenopus oocytes (Uozumi et aI., 2000). However,
expression of AtHKTl caused large inward Na+ cur-
rents in oocytes and Na+ hypersensitivity in yeast,
which suggested a role for AtHKTl in salt sensitiv-
ity (Maser et aI., 2002; Uozumi et aI., 2000). Direct
evidence for the involvement of AtHKT1 in Na+ up-
take and salt sensitivity in Arabidopsis has recently
emerged from a screen for suppressor mutations of
the Arabidopsis sos3 mutant (Liu and Zhu, 1998):
disruption of AtHKTl suppressed the ~odium-Qverly­
~ensitive phenotype (Rus et aI., 2001), and sos31athktl
double mutant seedlings took up less Na+ than either
sos3 or wild-type plants (Rus et aI., 2001). The
physiological role of AtHKTl in Arabidopsis remains
to be elucidated.
50
Na+ transport was also observed for EcHKTl and
EcHKT2 from Eucalyptus (Fairbairn et aI., 2000; Liu
et aI., 2001). Interestingly, EcHKT-mediated currents
in oocytes were enhanced by hypotonic bath solu-
tions, indicating that these transporters may function
as osmosensors in Eucalyptus (Uu et aI., 2001).
LeTl
LCTI is a low-affinity cation transporter from wheat
that was cloned in a screen for complementation of
a S. cerevisiae trkl,2 disruption (Schachtman et al.,
1997). LCTl was characterized in yeast and found
to transport K+ ,Na+ (Schachtman et aI., 1997), as
well as Ca2+ and Cd2+ (Clemens et aI., 1998). Ex-
pression of LeTl in yeast caused Na+ hypersensitivity
(Amtmann et aI., 2001). Importantly, LCTl-mediated
Na+ transport was inhibited by Ca2+ (Amtmann et aI.,
2001; Schachtman et aI., 1997). Ca2+ has long been
known to counteract Na+ toxicity in plants (LaHaye
and Epstein, 1969), and a majority of root Na+ uptake
has been proposed to occur via voltage-independent,
Ca2+ -sensitive cation channels (Allen et aI., 1995;
Davenport and Tester, 2000; Amtmann and Sanders,
1999; Tyerman and Skerrett, 1999; White, 1999).
Voltage-independent Na+ -inward currents have also
been recorded in root protoplasts from maize (Roberts
and Tester, 1997) and Arabidopsis (P. Buschmann and
1. I. Schroeder, unpublished). However, no LCTl-
like proteins are encoded in the Arabidopsis genome.
The LeTl cDNA contains repetitive sequences and
is therefore difficult to sequence (Schachtman et al.,
1997). Therefore, shotgun genome sequencing ap-
proaches are more likely to miss such genes. Several
types of transporters are likely to contribute to Na+
currents in roots.
Conclusions
Arabidopsis as a model for crop plants?
Arabidopsis thaliana has been an indispensable tool
for the identification of genes involved in K+ nutrition
(Gaymard et aI., 1998; Hirsch et aI., 1998) and Na+
stress (Liu and Zhu, 1998; Zhu, 2000). The avail-
ability of the complete genome sequence facilitates
identification of candidate K+ and Na+ transporter
genes, and disruption by T-DNA or transposon inser-
tion permits functional analysis of individual genes
in planta. However, comparison of cation transport-
ers revealed differences in the genetic make-up and
transcriptional regulation between Arabidopsis and
well-characterized crop plants.
Several differences in gene regulation have been
observed between Arabidopsis and crop plants such
as wheat and barley. Expression of TaAKTl is
up-regulated in wheat roots upon K+ starvation
(Buschmann et aI., 2000), while no such effect was
observed for AKTl in Brassica (Lagarde et aI., 1996).
HvHAKl is induced in barley roots when the plants
are grown at low K+ concentrations (Santa-Maria et
aI., 1997), but its closest relative from Arabidopsis,
AtHAK5 (see Figure 2), showed a different regula-
tion: expression of AtHAK5 is slightly down-regulated
in the roots but up-regulated in the shoots under K+
starvation (Rubio et aI., 2000), whilst AtKUP3 is K+
starvation induced in roots (Kim et aI., 1998). Fur-
thermore, HvHAKl is expressed exclusively in roots
(Santa-Maria et aI., 1997) while AtHAK5 is expressed
in roots and to a lesser extent also in shoots (Ru-
bio et al., 2000). Finally, wheat HKTl is a K+ ,Na+
co-transporter (Rubio et aI., 1995) induced by K+
starvation in wheat and barley (Wang et aI., 1998),
while AtHKTl from Arabidopsis appeared to be a
Na+ transporter and expressed independently of K+
levels (Uozumi et aI., 2000). These findings indicate
that functional prediction of transporter genes across
plant species is not feasible based on sequence sim-
ilarity alone, but must be supported by determination
of expression patterns and substrate specificities. Dif-
ferences in gene regulation have also been observed
for stresses other than K+ starvation (Clemens et aI.,
1999; Ha et aI., 1999).
Multiple pathways for K+ and Na+ uptake
The presence of mUltiple pathways for uptake of
K+ and Na+ has long been predicted (Epstein and
Rains, 1965; Epstein et at, 1963), and several K+
and Na+ inward currents from root cells have been
studied electrophysiologically (Amtmann et aI., 1997,
1999; Buschmann et aI., 2000; Davenport and Tester,
2000; Demidchik and Tester, 2002; Findlay et aI.,
1994; Gassmann and Schroeder, 1994; Maathuis and
Sanders, 1993, 1995; Roberts and Tester, 1995, 1997;
Tyerman et aI., 1997; Vogelzang and Prins, 1994;
White and Lemtiri-Chlieh, 1995). The task of attrib-
uting known ion currents to corresponding transporter
genes appears to be particularly difficult for Na+ since
most likely several transporters contribute to Na+ up-
take and no plant gene has so far been proven to
mediate Na+ uptake in planta. Possible mechanisms

for Na+ entry into roots include permeation of Na+
through K+ and Ca2+ transporters, use of Na+ trans-
port to energize K+ uptake, and also Na+ selective up-
take. Candidate genes for root Na+ uptake are found in
several K+ transporter families: (1) HKT transporters,
(2) KUPIHAKlKT transporters, (3) cyclic-nucleotide-
gated channels, and (4) LeT1 (Table 2). Analysis of
their substrate specificities in expression systems, de-
termination of their localization in planta, and analysis
of Arabidopsis knock-out or overexpressor lines will
be necessary to elucidate the roles and likely redund-
ancies of those candidate genes in Na+ uptake and salt
stress.
Acknowledgements
We wish to thank G. Allen for critical reading of the
manuscript. This work is supported by a United States
Department of Agriculture grant (2000-01558) and
a National Science Foundation genome grant (DBI-
0077378) to JIS, Swiss National Science Foundation
and Human Frontier Science Program grants to PM,
and a Feodor Lynen Fellowship from the Alexander
von Humboldt Foundation to MG.
References
Ache p. Becker D. Ivashikina N, Dietrich P, Roelfsema M R and
Hedrich R 2000 GORK, a delayed outward rectifier expressed
in guard cells of arabidopsis thaliana, is a K(+ )-selective, K( +)-
sensing ion channel. FEBS Lett. 486, 93-98.
Allen G J, Wyn Jones R G and Leigh R A 1995 Sodium transport
measured in plasma membrane vesicles isolated from wheat gen-
otypes with differing K+ lNa+ discrimination traits. Plant Cell
Environ. 18, 105-115.
Amtmann A and Sanders D 1999 Mechanisms of Na+ uptake by
plant cells. Adv. Bot. Res. 29, 75-112.
Amtmann A, Laurie S, Leigh R and Sanders D 1997 Multiple in-
ward channels provide flexibility in Na+ /K+ discrimination at
the plasma membrane of barley suspension culture cells. J. Exp.
Bot. 48, 481-497.
Amtmann A, Jelitto T C and Sanders D 1999 K+ -Selective inward-
rectifying channels and apoplastic pH in barley roots. Plant
Physiol. 120,331-338.
Amtmann A, Fischer M, Marsh E L, Stefanovic A, Sanders D and
Schachtman D P 2001 The wheat cDNA LCT! generates hyper-
sensitivity to sodium in a salt-sensitive yeast strain. Plant Physiol
126, 1061-1071.
Anderson J A, Huprikar S S, Kochian L V, Lucas W J and Gaber
R F 1992 Functional expression of a probable Arabidopsis thali-
ana potassium channel in Saccharomyces cerevisiae. Proc. Natl.
Acad. Sci. USA 89, 3736-3740.
Apse M P, Aharon G S, Snedden W A and Blumwald E 1999 Salt
tolerance conferred by overexpression of a vacuolar Na+ IH+
antiport in Arabidopsis. Science 285, 1256-1258.
51
Arazi T, Sunkar R, Kaplan B and Fromm H 1999 A tobacco plasma
membrane calmodulin-binding transporter confers Ni2+ toler-
ance and Pb 2+ hypersensitivity in transgenic plants. Plant J. 20,
171-182.
Arazi T, Kaplan B, Sunkar R and Fromm H 2000 Cyclic-nucleotide-
and Ca2+ Icalmodulin-regulated channels in plants: targets for
manipulating heavy-metal tolerance, and possible physiological
roles. Biochem. Soc. Trans. 28,471-475.
Banuelos M A, Klein R D, Alexanderbowman S J and Rodriguez-
Navarro A 1995 A potassium transporter of the yeast Schwan-
niomyces occidentalis homologous to the Kup system of Es-
cherichia coli has a high concentrative capacity. EMBO J. 14,
3021-3027.
Berti A, Anderson J A, Siayman C L, Sentenac H and Gaber R F
1994 Inward and outward rectifying potassium currents in Sac-
charomyces cerevisiae mediated by endogenous and heterelog-
ously expressed ion channels. Folia Microbiol. 39,507-509.
Blumwald E, Aharon G S and Apse M P 2000 Sodium transport in
plant cells. Biochim. Biophys. Acta 1465, 140-151.
Broillet M C and Firestein S 1999 Cyclic nucleotide-gated channels.
Molecular mechanisms of activation. Ann. N Y Acad. Sci. 868,
730-740.
Buschmann P H, Vaidyanathan R, Gassmann W and Schroeder J
12000 Enhancement of Na(+) uptake currents, time-dependent
inward-rectifying K(+) channel currents, and K( +) channel tran-
scripts by K(+) starvation in wheat root cells. Plant Physiol122,
1387-1397.
Cao Y W, Ward J M, Kelly W B, Ichida A M, Gaber R F, Anderson
J A, Uozumi N, Schroeder J I and Crawford N M 1995 Multiple
genes, tissue specificity, and expression-dependent modulation
contribute to the functional diversity of potassium channels in
Arabidopsis thaliana. Plant Physiol. 109, 1093-1106.
Clemens S, Antosiewicz D M, Ward J M, Schachtrnan D P and
Schroeder J I 1998 The plant cDNA LCT! mediates the uptake
of calcium and cadmium in yeast. Proc. Natl. Acad. Sci. USA 95,
12043-12048.
Clemens S, Kim E J, Neumann D and Schroeder J 11999 Tolerance
to toxic metals by a gene family of phytochelatin synthases from
plants and yeast. EMBO J. 18,3325-3333.
Czempinski K, Zimmermann S, Ehrhardt T and Muller-Rober B
1997 New structure and function in plant K+ channels: KC01,
an outward rectifier with a steep Ca2+ dependency. EMBO J. 16,
2565-2575.
Davenport R J and Tester M 2000 A weakly voltage-dependent,
nonselective cation channel mediates toxic sodium influx in
wheat. Plant Physiol. 122, 823-834.
Demidchik V and Tester M 2002 Sodium fluxes through nonselect-
ive cation channels in the plasma membrane of protoplasts from
Arabidopsis roots. Plant Physiol. 128, 379-387.
Diatloff E, Kumar R and Schachtrnan D P 1998 Site directed muta-
genesis reduces the Na+ affinity of HKTI, an Na+ energized
high affinity K+ transporter. FEBS Lett. 432, 31-36.
Doyle D A, Morais Cabral J, Pfuetzner R A, Kuo A, Gulbis J M,
Cohen S L, Chait B T and MacKinnon R 1998 The structure of
the potassium channel: molecular basis of K+ conduction and
selectivity. Science 280, 69-77.
Durell S R and Guy H R 1999 Structural models of the KtrB, TrkH,
and trkl,2 symporters based on the structure of the KcsA K(+)
channel. Biophys. J. 77, 789-807.
Durell S R, Hao Y, Nakamura T, Bakker E P and Guy H R 1999
Evolutionary relationship between K(+) channels and symport-
ers. Biophys J 77, 775-788.
Durell S R, Bakker E P and Guy H R 2000 Does the KdpA subunit
from the high affinity K( + )-translocating P-type KDP-ATPase
52
have a structure similar to that of K(+) channels? Biophys. J.
78,188-199.
Elumalai R P, Nagpal P and Reed J W 2002 A mutation in the Ar-
abidopsis KT2IKUP2 potassium transporter gene affects shoot
cell expansion. Plant Cell 14, 119-131.
Epstein E and Rains D W 1965 Carrier mediated cation transport in
barley roots: kinetic evidence for a spectrum of active sites. Proc.
Natl. Acad. Sci. USA 53, 1320-1324.
Epstein E, Rains D W and Elzam 0 E 1963 Resolution of dual
mechanisms of potassium absorption by barley roots. Proc. Natl.
Acad. Sci. USA 49, 684-692.
Epstein E, Norlyn J D, Rush D W, Kingsbury R W, Kelley D B,
Cunningham G A and Wrona A F 1980 Saline culture crops: a
genetic approach. Science 210, 399-404.
Fairbairn D J, Liu W, Schachtrnan D P, Gomez-Gallego S, Day
S R and Teasdale R D 2000 Characterisation of two distinct
HKTI-like potassium transporters from Eucalyptus camaldu-
lensis. Plant Mol. BioI. 43, 515-525.
Findlay G P, Tyerman S D, Garrill A and Skerrett M 1994 Pump
and K+ inward rectifiers in the plasmalemma of wheat root
protoplasts. J. Membr. BioI. 139, 103-116.
Fink M, Duprat F, Lesage F, Reyes R, Romey G, Heurteaux C and
Lazdunski M 1996 Cloning, functional expression and brain loc-
alization of a novel unconventional outward rectifier K+ channel.
EMBO J. 15, 6854-6862.
Flowers T J 1999 Salinisation and horticultural production. Sci.
Hort. 78, 1-4.
Fu H H and Luan S 1998 AtKUPl: a dual-affinity K+ transporter
fromArabidopsis. Plant Cell 10, 63-73.
Gaber R F, Styles C A and Fink G R 1988 TRKI encodes a plasma
membrane protein required for high-affinity potassium transport
in Saccharomyces cerevisiae. Mol. Cell BioI. 8,2848-2859.
Gamel K and Torre V 2000 The interaction of Na( +) and K(+) in the
pore of cyclic nucleotide-gated channels. Biophys. J. 79, 2475-
2493.
Gassmann W and Schroeder J I 1994 Inward-rectifying K+ channel
currents in root hairs of wheat: a mechanism for aluminum-
sensitive low-affinity K+ uptake and membrane potential con-
trol. Plant Physiol. 105, 1399-1408.
Gassmann W, Rubio F and Schroeder J 11996 Alkali cation selectiv-
ity of the wheat root high-affinity potassium transporter HKTI.
Plant J. 10,869-852.
Gaymard F, Pilot G, Lacombe B, Bouchez D, Bruneau D, Boucherez
J, Michaux-Ferriere N, Thibaud J B and Sentenac H 1998 Iden-
tification and disruption of a plant shaker-like outward channel
involved in K+ release into the xylem sap. Cell 94, 647-655.
Gierth M, Stelzer R and Lehmann H 1998 An analytical microscop-
ical study on the role of the exodermis in apoplastic Rb+(K+)
transport in barley roots. Plant Soil 207, 209-218.
Glauner K S, Mannuzzu L M, Gandhi C S and Isacoff E Y
1999 Spectroscopic mapping of voltage sensor movement in the
Shaker potassium channel. Nature 402, 813-817.
Golldack D, Kamasani U R, Quigley F, Bennett J and Bohnert H
J 1997 Salt stress-dependent expression of a HKTl-type high
affinity potassium transporter in rice. Plant Physiol. 114, S529.
Greenway H and Munns R A 1980 Mechnaisms of salt tolerance in
non-halophytes. Annu. Rev. Plant Physiol. 31, 149-190.
Ha S B, Smith A P, Howden R, Dietrich W M, Bugg S, O'Connell
M J, Goldsbrough P B and Cobbett C S 1999 Phytochelatin
synthase genes from Arabidopsis and the yeast Schizosacchar-
omyces pombe. Plant Cellll, 1153-1164.
Hampe T and Marschner H 1982 Effect of sodium on morphology,
water relations and net photosynthesis in sugar beet leaves. Z.
Pflanzenphysiol. 108, 151-162.
Haro R, Sainz L, Rubio F and Rodriguez-Navarro A 1999 Clon-
ing of two genes encoding potassium transporters in Neurospora
crassa and expression of the corresponding cDNAs in Saccharo-
myces cerevisiae. Mol. Microbiol. 31,511-520.
Hartje S, Zimmermann S, Klonus D and Mueller-Roeber B 2000
Functional characterisation of LKTl, a K+ uptake channel
from tomato root hairs, and comparison with the closely related
potato inwardly rectifying K+ channel SKT 1 after expression in
Xenopus oocytes. Planta 210, 723-731.
Heginbotham L and MacKinnon R 1993 Conduction properties of
the cloned Shaker K+ channel. Biophys. J. 65, 2089-2096.
Heginbotham L, Abramson T and MacKinnon R 1992 A functional
connection between the pores of distantly related ion channels as
revealed by mutant K+ channels. Science 258, 1152-1155.
Henn D K, Baumann A and Kaupp U B 1995 Probing the trans-
membrane topology of cyclic nucleotide-gated ion channels
with a gene fusion approach. Proc. Natl. Acad. Sci. USA 92,
7425-7429.
Hirsch R E, Lewis B D, Spalding E P and Sussman M R 1998 A
role for the AKT 1 potassium channel in plant nutrition. Science
280,918-921.
Horie T, Yoshida K, Nakayama H, Yamada K, Oiki S and Shinmyo
A 2001 Two types of HKT transporters with different properties
ofNa+ and K+ transport in Oryza sativa. Plant J. 27,129-138.
Kato Y, Sakaguchi M, Mori Y, Saito K, Nakamura T, Bakker E P,
Sato Y, Goshima S and Uozumi N 2001 Evidence in support of
a four transmembrane-pore-transmembrane topology model for
the Arabidopsis thaliana N a+ IK+ translocating AtHKTl pro-
tein, a member of the superfamily of K+ transporters. Proc. Natl.
Acad. Sci. USA 98, 6488-6493.
Kawano M, Abuki R, Igarashi K and Kakinuma Y 2000 Evidence
for Na(+) influx via the NtpJ protein of the KtrII K( +) uptake
system in Enterococcus hirae. J. Bacteriol. 182, 2507-2512.
Ketchum K A and Siayman C W 1996 Isolation of an ion channel
gene from Arabidopsis thaliana using the H5 signature sequence
from voltage-dependent K+ channels. FEBS Lett. 378, 19-26.
Ketchum K A, Joiner W J, Sellers A J, Kaczmarek L K and Gold-
stein S A 1995 A new family of outwardly rectifying potassium
channel proteins with two pore domains in tandem. Nature 376,
690-695.
Kim E J, Kwak J M, Uozumi N and Schroeder J I 1998 AtKUPl:
an Arabidopsis gene encoding high-affinity potassium transport
activity. Plant Cell 10, 51-62.
Kleene S J 1999 Both external and internal calcium reduce the sens-
itivity of the olfactory cyclic-nucleotide-gated channel to cAMP.
J. Neurophysiol. 81,2675-2682.
Kohler C, Merkle T and Neuhaus G 1999 Characterisation of a
novel gene family of putative cyclic nucleotide- and calmodulin-
regulated ion channels in Arabidopsis thaliana. Plant J. 18,
97-104.
Lagarde D, Basset M, Lepetit M, Conejero G, Gaymard F, Astruc
S and Grignon C 1996 Tissue-specific expression of Arabidopsis
AKTl gene is consistent with a role in K+ nutrition. Plant J. 9,
195-203.
LaHaye P A and Epstein E 1969 Salt toleration by plants: enhance-
ment with calcium. Science 166, 395-396.
Laimins L A, Rhoads D B, Altendorf K and Epstein W 1978 Iden-
tification of the structural proteins of an ATP-driven potassium
transport system in Escherichia coli. Proc. Natl. Acad. Sci. USA
75, 3216-3219.
Leng Q, Mercier R W, Yao Wand Berkowitz G A 1999 Cloning and
first functional characterization of a plant cyclic nucleotide-gated
cation channel. Plant Physiol. 121,753-761.

Lesage F, Guillemare E, Fink M, Duprat F, Lazdunski M, Romey
G and Barhanin J 1996 TWIK-l, a ubiquitous human weakly
inward rectifying K+ channel with a novel structure. EMBO J.
15,1004--1011.
Lindhauer M G, Haeder H E and Beringer H 1990 Osmotic poten-
tials and solute concentrations in sugar beet plants cultivated with
varying potassium/sodium ratios. Z. Pflanzenernlihr. Bodenk.
153,25-32.
Liu J P and Zhu J K 1998 A calcium sensor homolog required for
plant salt tolerance. Science 280, 1943-1945.
Liu W, Fairbairn D J, Reid R J and Schachtman D P 2001 Char-
acterization of two HKTI homologues from Eucalyptus ca-
maldulensis that display intrinsic osmosensing capability. Plant
Physiol. 127,283-294.
Lopes C M, Zilberberg N and Goldstein S A 2001 Block of
Kcnk3 by protons: evidence that 2-P-domain potassium channel
subunits function as homodimers. J. BioI. Chern. 276, 24449-52.
Maathuis F J M and Sanders D 1993 Energization of potassium
uptake in Arabidopsis thaliana. Planta 191,302-307.
Maathuis F J and Sanders D 1995 Contrasting roles in ion transport
of two K( +)-channel types in root cells of Arabidopsis thaliana.
Planta 197, 456--464.
Maathuis F and Sanders D 2001 Sodium uptake in Arabidopsis
roots is regulated by cyclic nucleotides. Plant Physiol. 127,
1617-1625.
Marschner H 1995 Mineral Nutrition of Higher Plants. Academic
Press, London.
Marschner H, Kylin A and Kuiper P J C 1981 Genotypic differences
in the response of sugar beet plants to replacement of potassium
by sodium. Physiol. Plant 51, 77-82.
Maser P, Thomine S, Schroeder J I, Ward J M, Hirschi K, Sze H,
Talke I N, Amtrnann A, Maathuis F J M, Sanders D, Harper
J F, Tchieu J, Gribskov M, Persans M W, Salt D E, Kim S A
and Guerinot M 2001 Phylogenetic relationships within cation-
transporter families of Arabidopsis thaliana. Plant Physiol. 126,
1646--1667.
Maser P, Hosoo Y, Goshima S, Horie T, Eckelman B, Yamada
K, Yoshida K, Bakker E P, Shinmyo A, Oiki S, Schroeder J I
and Uozumi N 2002 Glycine residues in potassium channel-like
selectivity filters determine potassium selectivity in four-Ioop-
per-subunit HKT transporters from plants. Proc. Natl. Acad. Sci.
USA, in press.
Mengel K and Kirkby E (1982) Principles of Plant Nutrition.
Worblaufen-Bern, Switzerland, International Potash Inst.
Munns R 1993 Physiological processes limiting plant growth in sa-
line soils: some dogmas and hypotheses. Plant Cell Environ. 16,
15-24.
Munro A W, Ritchie G Y, Lamb A J, Douglas R M and Booth
I R 1991 The cloning and DNA sequence of the gene for the
glutathione-regulated potassium-efflux system Kefc of Escheri-
chia coli. Mol. Microbiol. 5, 607-616.
Murguia J R, Belles J M and Serrano R 1995 A salt-sensitive 3' (2'),
5'-bisphosphate nucleotidase involved in sulfate activation. Sci-
ence 267, 232-234.
Nakamura T, Yamamuro N, Stumpe S, Unemoto T and Bakker
E P 1998 Cloning of the trkAH gene cluster and character-
ization of the Trk K(+)-uptake system of Vibrio alginolyticus.
Microbiology 144,2281-2289.
Page R D 1996 TreeView: an application to display phylogenetic
trees on personal computers. Comput. Appl. Biosci. 12, 357-
358.
Qu W, Zhu XO, Moorhouse AJ, Bieri S, Cunningham AM and Barry
PH 2000 Ion permeation and selectivity of wild-type recombin-
53
ant Na+ CNG (rOCN 1) channels in HEK293 cells. J. Membr.
BioI. 178, 137-150.
Quintero F J and Blatt M R 1997 A new family of K+ transporters
from Arabidopsis that are conserved across phyla. FEBS Lett.
415,206--211.
Rigas S, Debrosses G, Haralampidis K, Vicente-Agullo F, Feldmann
K A, Grabov A, Dolan L and Hatzopoulos P 2001 TRHI encodes
a potassium transporter required for tip growth in Arabidopsis
root hairs. Plant Cell 13, 139-151.
Roberts K and Tester M 1995 Inward and outward K+ -selective
currents in the plasma membrane of protoplasts from maize root
cortex and stele. Plant J. 8, 811-825.
Roberts S K and Tester M 1997 A patch clamp study of Na+
transport in maize roots. J. Exp. Bot. 48, 431-440.
Rubio F, Gassmann W and Schroeder J I 1995 Sodium-driven po-
tassium uptake by the plant potassium transporter HKTl and
mutations conferring salt tolerance. Science 270, 1660--1663.
Rubio F, Schwarz M, Gassmann Wand Schroeder J 11999 Genetic
selection of mutations in the high affinity K+ transporter HKTI
that define functions of a loop site for reduced Na+ permeability
and increased Na+ tolerance. J. BioI. Chern. 274, 6839-6847.
Rubio F, Santa-Maria G E and Rodriguez-Navarro A 2000 Clon-
ing of Arabidopsis and barley cDNAs encoding HAK potassium
transporters in root and shoot cells. Physiol. Plant. 109,34-43.
Rus A, Yokoi S, Sharkhuu A, Reddy M, Lee B H, Matsumoto T
K, Koiwa H, Zhu J K, Bressan R A and Hasegawa P M 2001
AtHKTl is a salt tolerance determinant that controls Na(+) entry
into plant roots. Proc. Nat!. Acad. Sci. USA 98, 14150--14155.
Saier M H Jr. 1999 A functional-phylogenetic system for the classi-
fication of transport proteins. J. Cell Biochem. Suppl. 84--94.
Santa-Maria G E, Rubio F, Dubcovsky J and Rodriguez-Navarro A
1997 The HAKI gene of barley is a member of a large gene
family and encodes a high-affinity potassium transporter. Plant
Cell 9, 2281-2289.
Schachtrnan D P and Schroeder J I 1994 Structure and transport
mechanism of a high-affinity potassium uptake transporter from
higher plants. Nature 370, 655-658.
Schachtrnan D P, Tyerman S D and Terry B R 1991 The K+ lNa+ se-
lectivity of a cation channel in the plasma membrane of root cells
does not differ in salt-tolerant and salt-sensitive wheat species.
Plant Physiol. 97, 598-605.
Schachtrnan D P, Kumar R, Schroeder J I and Marsh E L 1997
Molecular and functional characterization of a novel low-affinity
cation transporter (LCTl) in higher plants. Proc. Natl. Acad. Sci.
USA 94,11079-11084.
Schleyer M and Bakker E P 1993 Nucleotide sequence and 3'-end
deletion studies indicate that the K(+)-uptake protein kup from
Escherichia coli is composed of a hydrophobic core linked to a
large and partially essential hydrophilic C terminus. J. Bacteriol.
175, 6925-6931.
Schlosser A, Meldorf M, Stumpe S, Bakker E P and Epstein W
1995 TrkH and its homolog, TrkG, determine the specificity and
kinetics of cation transport by the Trk system of Escherichia coli.
J. Bacteriol. 177,1908-1910.
Schroeder J I 1988 K+ transport properties of K+ channels in the
plasma membrane of Vidafaba guard cells. J. Gen. Physiol. 92,
667-683.
Schroeder J I, Hedrich R and Fernandez J M 1984 Potassium-
selective single channels in guard cell protoplasts of Vida faba.
Nature 312, 361-362.
Schroeder J I, Ward J M and Gassmann W 1994 Perspectives on
the physiology and structure of inward-rectifying K+ channels
in higher plants: biophysical implications for K+ uptake. Annu.
Rev. Biophys. Biomol. Struct. 23,441-471.
54
Schroeder J I, Allen G J, Hugouvieux V, Kwak J M and Waner D
2001 Guard cell signal transduction. Annu. Rev. Plant PhysioL
Plant MoL BioI. 52, 627-658.
Schuurink R C, Shartzer S F, Fath A and Jones R L 1998 Charac-
terization of a calmodulin-binding transporter from the plasma
membrane of barley aleurone. Proc. Nat!. Acad. Sci. USA 95,
1944-1949.
Sentenac H, Bonneaud N, Minet M, Lacroute F, Salmon J M, Gay-
mard F and Grignon C 1992 Cloning and expression in yeast of
a plant potassium ion transport system. Science 256, 663-665.
Shi H, Ishitani M, Kim C and Zhu J K 2000 The Arabidopsis
thaliana salt tolerance gene SOSI encodes a putative Na+lH+
antiporter. Proc. Nat!. Acad. Sci. USA 97, 6896-6901.
Sokolova 0, Kolmakova-Partensky Land Grigorieff N 2001 Three-
dimensional structure of a voltage-gated potassium channel at 2.5
nm resolution. Structure 9, 215-220.
Soldatenkov V A, Velasco J A, Avila M A, Dritschilo A and Not-
ario V 1995 Isolation and characterization of SpTRK, a gene
from Schizosaccharomyces pombe predicted to encode a K+
transporter protein. Gene 161,97-101.
Spalding E P, Hirsch R E, Lewis D R, Qi Z, Sussman M R and
Lewis B D 1999 Potassium uptake supporting plant growth in
the absence of AKTl channel activity: inhibition by ammonium
and stimulation by sodium. J. Gen. PhysioL 113,909-918.
Su H, Golldack D, Katsuhara M, Zhao C and Bohnert H J 2001
Expression and stress-dependent induction of potassium channel
transcripts in the common ice plant. Plant Physiol. 125, 604-
614.
Sunkar R, Kaplan B, Bouche N, Arazi T, Dolev D, Talke IN,
Maathuis F J, Sanders D, Bouchez D and Fromm H 2000 Ex-
pression of a truncated tobacco NtCBP4 channel in transgenic
plants and disruption of the homologous Arabidopsis CNGeI
gene confer Pb2+ tolerance. Plant J. 24, 533-542.
Sze H, Ward J M and Lai S 1992 Vacuolar H(+)-translocating AT-
Pases from plants: structure, function, and isoforms. 1. Bioenerg.
Biomembr. 24, 371-381.
Tarczynski M C, Jensen R G and Bohnert H J 1993 Stress protection
of transgenic tobacco by production of the osmolyte mannitoL
Science 259, 508-510.
Tempel B L, Papazian D M, Schwarz T L, Jan Y N and Jan L Y 1987
Sequence of a probable potassium channel component encoded
at Shaker locus of Drosophila. Science 237, 770-775.
Thiel G, MacRobbie E A and Blatt M R 1992 Membrane trans-
port in stomatal guard cells: the importance of voltage controL J.
Membr. BioL 126, 1-18.
Tholema N, Bakker E P, Suzuki A and Nakamura T 1999 Change to
alanine of one out of four selectivity filter glycines in KtrB causes
a two orders of magnitude decrease in the affinities for both K+
and Na+ of the Na+ dependent K+ uptake system KtrAB from
Vibrio alginolyticus. FEBS Lett. 450, 217-220.
Thompson J D, Gibson T J, Plewniak F, Jeanmougin F and Hig-
gins D G 1997 The CLUSTAL_X windows interface: flexible
strategies for multiple sequence alignment aided by quality
analysis tools. NucL Acids Res. 25, 4876-4882.
Tusnady G E and Simon I 1998 Principles governing amino acid
composition of integral membrane proteins: application to topo-
logy prediction. 1. MoL BioL 283, 489-506.
Tyerman S D and Skerrett 1M 1999 Root ion channels and salinity.
Sci. Hort. 78, 175-235.
Tyerman S D, Skerrett M, Garrill A, Findlay G P and Leigh R A
1997 Pathways for the permeation of Na+ and Cl- into proto-
plasts derived from the cortex of wheat roots. J. Exp. Bot. 48,
459-480.
Uozumi N, Nakamura T, Schroeder J I and Muto S 1998 De-
termination of transmembrane topology of an inward-rectifying
potassium channel from Arabidopsis thaliana based on func-
tional expression in Escherichia coli. Proc. Nat!. Acad. Sci. USA
95, 9773-9778.
Uozumi N, Kim E J, Rubio F, Yamaguchi T, Muto S, Tsuboi A,
Bakker E P, Nakamura T and Schroeder J I 2000 TheArabidopsis
HKTI gene homolog mediates inward Na+ currents in Xenopus
laevis oocytes and Na+ uptake in Saccharomyces cerevisiae.
Plant PhysioL 122, 1249-1260.
Very A A, Robinson M F, Mansfield T A and Sanders D 1998
Guard cell cation channels are involved in Na+ -induced stomatal
closure in a halophyte. Plant J. 14,509-521.
Vogelzang S A and Prins H B A 1994 Patch clamp analysis of the
dominant plasma membrane K+ channel in root cell protoplasts
of Plantago media L. Its significane for the P and K State. J.
Memb. BioL 141, 113-122.
Wang T B, Gassmann W, Rubio F, Schroeder J I and Glass A D 1998
Rapid up-regulation of HKTl, a high-affinity potassium trans-
porter gene, in roots of barley and wheat following withdrawal
of potassium. Plant PhysioL 118,651-659.
Ward J M 2001 Identification of novel families of membrane pro-
teins from the model plant Arabidopsis thaliana. Bioinformatics
17,560-563.
White P J 1999 The molecular mechanism of sodium influx to root
cells. Trends Plant Sci. 4, 245-246.
White P J and Lemtiri-Chlieh F 1995 Potassium currents across the
plasma membrane of protoplasts derived from rye roots: a patch-
clamp study. J. Exp. Bot. 46, 497-511.
Zagotta W N and Siegelbaum S A 1996 Structure and function of
cyclic nucleotide-gated channels. Annu. Rev. Neurosci. 19,235-
263.
Zhu J K 2000 Genetic analysis of plant salt tolerance using Ara-
bidopsis. Plant PhysioL 124,941-948.
Zhu J K, Liu J P and Xiong L M 1998 Genetic analysis of salt
tolerance in Arabidopsis: evidence for a critical role of potassium
nutrition. Plant Cell 10, 1181-1191.
Nutrient functions
 Plant and Soil 247: 57-67,2002.
© 2002 Kluwer Academic Publishers.
57

A proposed role for copper ions in cell wall loosening

Stephen C. Fry I , Janice G. Miller & Jo C. Dumville

The Edinburgh Cell Wall Group, Institute of Cell and Molecular Biology, The University of Edinburgh, Daniel
Rutherford Building, The King's Buildings, Mayfield Road, Edinburgh EH93JH, u.K. 1 Corresponding author*

Received 19 September 2001. Accepted in revised form 8 March 2002

Key words: ascorbate, cell expansion, cell wall loosening, copper, fruit ripening, hydroxyl radicals, polysaccharides

Abstract
We present the hypothesis that Cu 2+, complexed with cell wall polymers (especially histidine-rich glycoproteins),
can be reduced to Cu+ by apoplastic electron donors such as ascorbate and superoxide, and that the Cu+ can then
undergo a Fenton reaction with apoplastic hydrogen peroxide (H202) to generate hydroxyl radicals (,OH). These
can cause non-enzymic scission of wall polysaccharides and may loosen the cell wall. In tomato fruit cell walls,
'OH radicals cause pectin solubilisation. However, we show that acidic pectins do not sequester Cu 2+ sufficiently
strongly to prevent 'OH-attack on other molecules (xyloglucan or aromatics) present in the same solution. We
describe two approaches by which apoplastic 'OH can be detected in vivo: (a) detection of a chemical 'fingerprint'
characteristic of' OH -attacked cell wall polysaccharides; (b) infiltration into the apoplast of membrane-impermeant
[3H]benzoyl-amides which react with 'OH to yield 3H20. Data support apoplastic 'OH production in vivo, and its
enhancement during pear (Pyrus communis) fruit ripening, but have not so far supported its proposed enhancement
during auxin-stimulated cell expansion in maize (Zea mays) coleoptiles.

Abbreviations: A - dehydro-L-ascorbate; AH2 - L-ascorbate; Ara - L-arabinose; BzG3 - N-benzoyl-

glycylglycylglycine; BzK5Me - N-benzoyl-pentalysine methyl ester; Bz-PA - N-benzoyl-polyallylamine; Cys -
L-cysteine; Gal - o-galactose; GalA - o-galacturonic acid; Glc - o-glucose; His - L-histidine; Hyp - 4-hydroxy-
L-proline; IAA - indole 3-acetic acid; Lys - L-Iysine; Man - D-mannose; MLG - mixed-linkage ,8-(1 ~3),
(l ~4)-o-glucan Mr - relative molecular mass; Pro - L-proline; SOD - superoxide dismutase

Introduction (Tanaka et aI., 1995) and amine oxidase (Rossi et aI.,

Hypothesis
Copper is an essential element for plants, and defi-
ciency causes growth defects such as sinuous stems
and branches (Dell, 1994; Turnbull et aI., 1994). Cop-
per's unique redox properties enable it to serve an
indispensable role in the action of a number of redox
proteins such as plastocyanin, cytochrome oxidase,
laccase (phenol oxidase), ascorbate oxidase, amine
oxidases and some superoxide dismutases (SODs). An
inadequate copper supply measurably reduces the in-
vivo activity of some of these enzymes, e.g. SOD
1995).
In the present paper, we propose the hypothesis
that, in addition to its established roles in redox en-
zymes, copper serves a non-enzymic role in the plant
cell wall. Specifically, we suggest that wall-bound
Cu2+ can cause the production of the hydroxyl rad-
ical (,OH), an exceedingly reactive molecule which
can cleave structural polysaccharides and may thereby
loosen the cell wall, promoting certain physiologic-
ally important processes such as cell expansion, fruit
softening and organ abscission.
Copper in plant cell walls

* FAX No.: +44-131-650-5392. Stems, leaves, fruits and many roots all typically have
E-mail: S.Fry@Ed.Ac.UK Cu 2+ concentrations falling within a relatively narrow
58
range (3-30 mg/kg dry mass), regardless of the cop-
per concentration of the soil in which the plant grew
(Jarvis, 1981; Paul and Southgate, 1978). Much of
the Cu2+, especially in roots, is tightly bound in the
cell wall and can only be displaced by other Cu 2+
ions (for example, 63Cu2+ can displace the radioactive
64Cu 2+), by the chemically similar Pb2+, or by 0.1 M
HCl (Iwasaki et aI., 1990; Walker and Webb, 1981).
Much of the soluble copper in the apoplast is
complexed with organic ligands. In the xylem sap of
various plants, these ligands include histidine (His)
and nicotianamine (Kramer et aI., 1996; Liao et aI.,
2000). The proportion of Cu 2+ that is bound in this
way is pH-dependent: reducing the pH from 5 to 4
results in a ~30-fold increase in the proportion of free
(non-complexed) Cu2+ (Liao et al., 2000).
In plants supplied with high concentrations of
Cu2+, e.g. 200 fLM in the soil-water, the concentration
in the shoots usually remains relatively constant; the
excess Cu 2+ is retained in the cell walls of the root at
up to 1.5 mg of Cu per g of dry root cell walls (Jarvis,
1978).
The insoluble copper in cell walls must be bound
to polymeric, structural components which, however,
are poorly understood. One possible class of binding
sites in the primary cell wall is acidic polysaccharides,
especially pectins [polysaccharides rich in galacturon-
ate (GalA) residues; O'Neill et aI., 1990]. However,
pectins do not bind all the Cu2+ present in typical
biological systems (Nikdel et aI., 1991; Schlemmer,
1989; van Cutsem and Gillet 1982; Wucherpfen-
nig, 1992). Another candidate polymeric ligand is a
neutral polysaccharide, e.g. xyloglucan, which forms
an insoluble Cu 2+complex, often used in xyloglucan
purification (Jones and Stoodley, 1965; Rao, 1959).
There is, however, no evidence that Cu 2+ binds to such
polysaccharides with a high affinity.
Another candidate, more likely to involve high-
affinity binding, is a sub-set of the cell wall's
(glyco )proteins, especially those rich in His and/or
Cys residues. For example, a wall-associated galactose
(Gal)-rich glycoprotein from Nicotiana styles is rich
in hydroxyproline (Hyp), Pro, Lys and His (Sommer-
Knudsen et aI., 1996). Sommer-Knudsen and Bacic
(1997) have developed a simple method for compar-
ing the metal-ion-binding affinities of proteins. Using
this method, their laboratory has shown that the Nico-
tiana style glycoprotein has a high affinity for Cu 2+
(Sommer-Knudsen et aI., 1997; J. Sommer-Knudsen,
M. Verkuijlen, A. Bacic and A.E. Clarke, personal
communication), presumably owing to its high His
content. Several other Hyp-rich glycoproteins, in-
cluding many extensins (Biggs and Fry, 1990; Bown
et aI., 1993; Stuart and Varner, 1980) and some
arabinogalactan-proteins (Baldwin et aI., 2001; Kiel-
iszewski et aI., 1992), also have His-rich domains,
and may therefore have a high affinity for Cu 2+, al-
though their metal-binding properties have not yet
been studied experimentally.
Another class of strongly copper-binding proteins
is the metallothioneins (see below), whose genes are
often highly expressed during fruit ripening (Davies
and Robinson, 2000) but which do not appear to occur
in the cell wall.
Additional wall-associated, copper-binding pro-
teins may await discovery. For example, Drew and
Gatehouse (1994) found that a Pisum mutant that
cannot lignify its pod endocarp fails to express a
gene which probably codes for a blue, type-l copper-
protein. With a calculated Mr of 19 400, this protein
is smaller than a typicallaccase. In addition, a small
copper-protein with an unusual Ser- and Hyp-rich C-
terminal domain has been isolated from cucumber
fruit (Mann et al., 1992). Other wall proteins of pos-
sible relevance to copper-binding are the thionins -
small, Cys-rich polypeptides found in various parts of
the plant cell, including the wall, especially in cereals
and mistletoe (Bohlmann, 1994; Sakurai, 1998).
Copper has also been reported to bind to high-
Mr phenolics (possibly 'tannins') in hyphal walls of
the fungus Neurospora crassa (Suresh and Subraman-
yam, 1998); plant cell walls are also known to contain
diverse phenolics (Wallace and Fry, 1994).
Wall-localised copper may generate hydroxyl radicals
(OH)
Cu 2+ in cell walls is of interest because its reduction-
product, Cu+, can undergo a non-enzymic 'Fenton
reaction' with H202, producing ·OH radicals:
Cu+ + H202 --+ ·OH + OH- + Cu 2+ (1)
The rate-constant for this reaction is 4700 M- 1 s-1,
which is 60 x greater than that for the corresponding
reaction with Fe2+ (Halliwell and Gutteridge, 1990).
The two Fenton reactants (Cu+ and H202) can
both be produced in the apoplast in vivo:
(i) the reduction of Cu 2+ to Cu+ may be effected
non-enzymically by any of a wide range of elec-
tron donors (Davies et aI., 1991; Fry, 1998; Ohta
et aI., 2000; Tabbi et aI., 2001). Three such
electron donors that probably occur in the apo-

plast in vivo are ascorbate (Burkey, 1999; Taka-
hama, 1993; Zheng et aI., 2000), dehydroascorbate
(Burkey, 1999; Takahama 1993) and superoxide
(02'-; Frahry and Schopfer, 2001). In addition,
Chlamydomonas cells, and possibly also some
tissues of higher plants, can directly reduce ex-
tracellular Cu2+ to Cu+ by means of a 'cupric
reductase' , apparently located in the plasma mem-
brane (Weger, 1999).
(ii) the presence of H202 in plant cell waIls has fre-
quently been demonstrated, although its precise
metabolic origin has been debated (Federico and
Angelini, 1986; Frahry and Schopfer, 1998; Lin
and Kao, 2001; Zarra et aI., 1999).
Given the probable presence of these two sub-
strates (Cu+ and H202), it seems likely that 'OH rad-
ical production via the non-enzymic Fenton reaction
(Equation (1)) occurs in plant cell waIls in vivo.
Ascorbate is able non-enzymically to reduce both
Cu2+ to Cu+ and 02 to H202. Therefore, in the
presence of traces of Cu 2+, an aerated solution of
ascorbate is sufficient to generate 'OH radicals via the
reactions:
AH2 + 02 -+ A + H202 (2)
1 1
-AH2 + Cu 2+ -+ -A + H+ + Cu+ (3)
2 2
(where AH2 = ascorbate and A = dehydroascorbate)
followed by Equation (1) (Fry, 1998). Apoplastic
ascorbate is thus of particular interest. The Cu 2+ is
re-cycled (Equations (2), (3), (1)); therefore only
minute traces are required to support the continuous
production of 'OH by ascorbate + 02, with the overall
stoichiometry:
4 AH2 + 3 02 -[Cu 2+]-+ 4 A + 4 'OH + 2 H20 (4)
In this scheme, copper effectively acts as a 'catalyst',
and trace concentrations suffice.
Binding of copper to ligands can promote or inhibit
its participation in the Fenton reaction
Since most of the apoplastic copper is tightly bound to
organic ligands, it might be suggested that this copper
would be 'out of play' as far as the Fenton reaction
is concerned. However, ligand-associated copper may
have an enhanced ability to participate in the Fenton
reaction.
Transition metal ions attached to some artificial
ligands can be reduced and react with H202 to gener-
ate 'OH radicals, whereas those attached to other lig-
ands cannot. The effect may be complex: for example,
59
Cu2+ -phenanthroline, Fe3+ -neocuproine and Fe3+ -
EDTA complexes can readily be reduced and undergo
the Fenton reaction, whereas Fe3+ -phenanthroline,
Cu 2+ - neocuproine and Cu2+ -EDTA complexes can-
not (Aruoma et aI., 1991; Burkitt, 1994).
Cu2+ ions complexed to some natural ligands have
also been studied for their ability to generate 'OH rad-
icals. For example, in mammalian Cu-Zn-SOD an
increase in the Cu:Zn ratio can result in excessive
production of active oxygen species - the enzyme
becomes a 'superoxide synthase' (Estevez et aI., 1999;
Yim et aI., 1993). This provides an example ofligand-
associated copper possibly promoting' OH production.
It has also been suggested that over-expression of Mn-
SOD in maize leaves could result in increased H202
production (Kingston-Smith and Foyer, 2000), which
might also lead to 'OH production.
Metallothioneins are Cys-rich proteins that can
bind Cu and/or Zn. In the Zn2+ form, metaIlothioneins
may act as anti-oxidants, capable of scavenging 'OH
radicals by reaction with Cys residues. However, the
same proteins can also bind Cu2+ ions, which are then
reduced to Cu+ by the Cys residues; the Cu+ formed
can react with H202 to generate 'OH radicals and,
under these circumstances, metallothioneins are 'pro-
oxidants' (Oikawa et aI., 1995; Sakurai et aI., 1994;
Suzuki et aI., 1996).
Cu 2+ that is complexed with His or citrate at
typical apoplastic pH values can still be reduced by
ascorbate and then participate in the Fenton reaction
(Fry, 1998). The same may be true of Cu2+ that is
bound to His-rich cell-waIl glycoproteins. Cu2+ that is
associated with certain other biologically relevant lig-
ands such as poly-His, cysteine or oxalate is partially
prevented from participating - although in the case
of oxalate the effect is only strong when the ligand
exceeds about 10 mM (Fry, 1998). The apoplastic pH
may playa role in governing the ability of Cu 2+ to
generate 'OH radicals since lowering the pH by 1 unit
(as occurs during auxin-induced 'acid growth') greatly
decreases the affinity of Cu2+ for histidine and related
ligands (Liao et aI., 2000).
The affinity of Cu+ for organic ligands has been
less thoroughly studied than has that of Cu2+ owing
to the instability of Cu+ in aqueous solution. Never-
theless, it is noteworthy that in the presence of Cu2+ -
binding polymers, copper-dependent (,OH-mediated)
scission of polysaccharides can continue progress-
ively, ultimately leading to many scission events per
copper atom (see below; Tabbi et aI., 2001). This
suggests that the copper ions can move from one poly-
60
mer molecule to another, and this ability is probably
due to the difference between Cu+ and Cu2+ in their
affinities for diverse ligands (Parker, 1981).
Hydroxyl radicals can cause scission of cell wall
polymers: possible role in wall loosening
The ability of wall-bound copper to generate apo-
plastic ·OH radicals in living plant cells is proposed
to be of biological significance in cell wall 'loosen-
ing'. ·OH radicals readily cause the non-enzymic
scission of wall polysaccharides in vitro, including xy-
loglucan, pectins and mixed-linkage ,8-(1-+3),(1-+4)-
glucan (MLG) (A. Bayne, personal communication;
Fry, 1998; Tabbi et al., 2001). Xyloglucan is a key
polysaccharide in primary cell walls, especially of
dicots: it is thought to act as a molecular 'tether',
inter-connecting adjacent microfibrils and thus limit-
ing cell expansion (Fry, 1989; Hayashi, 1989). MLG
may play a similar role in graminaceous monocots.
The presence of only 1 Cu2+ ion per 150 000 sugar
residues (i.e., 1 Cu2+ ion per 20-30 xyloglucan mo-
lecules of Mr rv 106) is enough to sustain an appre-
ciable rate of ascorbate-induced scission of xyloglucan
chains (Fry, 1998). The chemical mechanism of ·OH-
induced polysaccharide scission is understood, at least
in outline (Fry et aI., 2001; Miller and Fry, 2001;
Schuchmann and von Sonntag, 1978).
Plant cell walls undergo controlled loosening in
vivo, enabling cell expansion and cell separation.
These processes are important in germination, primary
growth, fruit softening and organ abscission, but are
incompletely understood. Wall loosening is widely
assumed to be mediated by wall-localised proteins
- hydrolases, transglycosylases and expansins (Cos-
grove, 1999; Fry, 1995; ). Little attention has been
paid to the possibility that polysaccharides in the walls
of living plant cells also suffer copper-dependent,
non-enzymic scission.
Promotion of polysaccharide scission by Cu+-
generated apoplastic ·OH radicals would be likely to
contribute to wall loosening, and thus cell expansion,
fruit softening etc. in vivo. These considerations may
explain some of the following observations:
(i) Exogenous ascorbate, which reduces Cu2+ to
Cu+, can promote cell expansion, e.g. in
onion roots, especially under conditions favouring
ascorbate oxidation (Hidalgo et al., 1991).
(ii) In soya-bean hypocotyls, dithiothreitol inhibits
both NADH oxidase activity (which forms apo-
plastic superoxide, another agent that can reduce
Cu2+ to Cu+) and auxin-stimulated cell extension
(which is dependent on wall loosening) (Morn! et
al.,1995).
(iii) In maize coleoptiles, exogenous hexacyano-fer-
rate(III) (= 'ferricyanide') rapidly promotes elong-
ation, probably by activating a plasmalemmar
NADH oxidase (Carrasco-Luna et al., 1995).
(iv) In wheat seeds, levels of mRNA for oxalate
oxidase, an H202-generating wall enzyme, rise
sharply at the explosive onset of embryo cell ex-
pansion which occurs 5 to 10 h after imbibition
(Lane et al., 1993).
(v) Wall peroxidase activity, which would compete for
available H202, often correlates negatively with
growth rate (Fry, 1986).
(vi) Pectin solubilisation and degradation occur in
fruits which possess no detectable pectin-cleaving
enzymes, e.g. persimmon (Cutillas-Iturralde et al.,
1993) and kiwifruit (Redgwell et al., 1992).
(vii) Endogenous H202 formation was proposed to
contribute to pear fruit softening (Brennan and
Frenkel, 1977).
(viii) Exogenous Cu2+ promotes the elongation of
Nicotiana pollen tubes (Read et al., 1993).
The 'taming' of the hydroxyl radical by plants?
It might be argued that Cu-dependent wall loosen-
ing, via the non-enzymic scission of polysaccharides
by non-enzymically produced ·OH radicals, would be
a 'hazardous' and poorly controlled mechanism by
which to promote cell expansion or fruit softening -
after all, the ·OH radical is widely seen as 'detrimental
to life' because it can cause mutation, membrane dam-
age and protein denaturation. However, in defence
of the hypothesis presented here, we point out the
following observations.
• Fruits are short-lived (their raison d' etre is often to
be eaten by animals) and therefore mutations and
membrane damage in ripening fruit tissues would
be inconsequential.
• ·OH radicals are so extremely reactive that they
have a very short half-life in a biological milieu.
They are estimated to be capable of diffusing no
further than about 1 nm before reacting with an
organic molecule (Griffiths and Lunec, 1996). This
distance of migration may be compared with the
thickness of a typical primary cell wall (about
100 nm). ·OH radicals are thus effectively site-
specific oxidants: if produced at a particular loca-
tion within the wall matrix or middle lamella, they
 61

would be likely to attack a nearby structural poly-

saccharide chain, and thus promote wall loosening,
with little danger of their reaching the lipids of
the plasma membrane, still less the DNA of the
nucleus.
The site-specificity of production of ·OH radicals, at
wall-bound copper ions, could ultimately be under
the control of the synthesis and secretion of His-rich
glycoproteins and other copper-binding ligands. Fur-
thermore, the initiation and termination of ·OH pro-
duction, and thus of wall loosening, could be regulated
by the activity of proteins, such as
• synthesis, secretion and enzymic oxidation of
ascorbate and the re-absorption of its oxidation-
product, dehydroascorbate (Rautenkranz et al.,
1994; Srnirnoff and Wheeler, 2000);
• production of superoxide e.g. by the action of
NAD(P)H oxidases (Morrt~ et aI., 1988);
• production of H202, e.g. by wall-bound oxidases
(see above);
• removal ofH202, e.g. by the action of wall-bound
peroxidases (Wallace and Fry, 1994).
Therefore, the process of non-enzymic polysaccharide
scission could be just as tightly controlled by pro- Figure 1. Three radioactive probes tested as potential indicat-
teins, and ultimately by gene expression, as any other 0rs of the presence of ·OH radicals in the apoplast. BzG3 =
postulated wall-loosening mechanism such as the ac- N-benzoyl-glycylglycylglycine; BzK5Me = N-benzoyl-pentalysine
tion of xyloglucan endotransglycosylases, cellulases methyl ester; Bz-PA =N-benzoyl-polyallylamine. BzG3 is available
commercially; the others were synthesised in the authors' laboratory
or expansins. (unpublished).

Experiments and results: Recent work to
investigate Cu2+ -dependent polysaccharide
scission
Membrane-impermeant probes for apoplastic ·OB
radicals
It is difficult to demonstrate reliably the presence of
·OH radicals in vivo, and particularly difficult to show
that they are present in the apoplast, where they would
need to be to effect cell wall loosening. Our laboratory
is working on two complementary approaches for the
detection of naturally occurring apoplastic hydroxyl
radicals in healthy plant tissues.
In the first approach, we have developed
three membrane-impermeant, radioactive, aromatic
'probes' to test for the apoplastic production of ·OH
radicals. Aromatic compounds readily react with ·OH
radicals to produce phenols (e.g. phenylalanine pro-
duces a mixture of 0-, m- and p-tyrosine) and have
been used to demonstrate the production of ·OH
radicals in animal tissues, e.g. in arthritic connect-
ive tissue (Ghiselli, 1998; Grootveld and Halliwell,
1986). Unfortunately, in the plant apoplast, measur-
ing the phenolic products is not reliable since they
would be rapidly oxidised by peroxidases. We have
therefore synthesised a series of radioactive, aromatic
probes, which react with ·OH to release a detect-
able product (radioactive water, 3H20), whether or
not the organic products react further. The probes
used are N-eH]benzoyl-glycylglycylglycine (BzG3),
N-[3H]benzoyl-pentalysine methyl ester (BzKsMe)
and N-[3H]benzoyl-polyallylamine (Bz-PA) (Figure
1). The 3H20 released upon reaction with ·OH radicals
can be separated from the remaining unreacted probe
molecules and measured by scintillation-counting. In
the cases of BzKsMe and Bz-PA, the probe is polyca-
tionic and may thus become anchored to acidic poly-
saccharides of the cell wall, facilitating detection of
wall-localised ·OH radicals.
A set of l-cm segments of various plant organs
(maize coleoptiles and mesocotyls, and the vegetat-
62

ive stems, peduncles, petioles or styles of various

other plants) were incubated in 20 mM MES (Na+)
12000
buffer, pH 6.1, containing the probe molecule, e.g.
[3H]Bz-PA, with or without auxin (IAA; 20 I£M).
After incubation (normally for 2 h), the yield of 3H20
Oligosaccharides
was assayed in a sample of the spent incubation me- 10000
of DP>3
dium. Maize coleoptiles, Tropaeolum peduncles, and
Russian vine internodes showed a clear promotion /
of elongation by 20 I£M IAA. The apparent yield S 8000
of 3H20 was in the range 1-4% of the total 3H, ~0-
suggesting the production of apoplastic 'OH radicals. S
However, there was no evidence for enhanced 3H20 .0 Glc
production accompanying auxin-promoted growth. :E..., 6000
u
The IAA-induced elongation of maize coleoptiles 0
oj

was examined over a time-course and shown to be ;.e

easily measurable within 1 h and to continue for
at least 5.5 h. Abrasion of the coleoptile, and re-
moval of the internal leaf had little effect. Again, no
IAA-promotion of apoplastic 'OH production could
be detected in maize coleoptiles. Similar results were
obtained with the other radioactive probes shown in
Figure 1.
We are continuing to work with the three radio-
active probes to explore the natural production of
apoplastic 'OH radicals in various plant materials.
Recognising 'OH-attacked polysaccharides by
'fingerprinting'
The second approach to the detection of 'OH radic-
als in the apoplast in vivo is to look for a molecular
'fingerprint' diagnostic of the action of these rad-
icals on wall polysaccharides. When pure cell wall
polysaccharides (xyloglucan and pectin) were treated
in vitro with ascorbate-Cu 2+-02 mixtures, which
non-enzymically generate 'OH radicals, not only did
partial polysaccharide scission occur but also oxo
groups were introduced which were readily detect-
able by reaction with NaB 3H4: the products included
3H-Iabelled sugar residues, including some, such as
ribose, not known to be present in plant cell wall
polysaccharides (Fry et aI., 2001; Miller and Fry,
2001).
Similar results were obtained when MLG, a major
polysaccharide of primary cell walls of grasses, was
treated with ascorbate-Cu2+ -02-generated'OH radic-
als in vitro. After treatment with NaB3H4 followed
by lichenase digestion, the 'OH-attacked MLG yiel-
ded a series of radioactive products, apparently di- to
penta-saccharides, as determined by paper chromato-
graphy. Subsequent digestion with Driselase (a fungal
oj
0::
4000 Gal
o 10 20 30 40
Distance from origin (em)
Figure 2. 'Borotritiide-fingerprints' of polysaccharide samples
showing evidence for 'OH radical attack. Authentic MLG was
treated with 'OH in vitro, then dialysed, treated with NaB3H4 (to
convert oxo groups to radioactive alcohols), and digested with Dri-
selase (see Fry et aI., 2001). The digestion products were subjected
to paper chromatography in ethyl acetate/pyridinelH20 (8:2: 1 by
volume) for 48 h; the resulting profile of radioactivity is shown
(--). In addition, cell walls were isolated from maize cole-
optile segments which had been incubated with or without 20 fJ..M
IAA for 60 min (a long enough period to give readily detectable
IAA-dependent elongation). The isolated walls were treated with
NaB3H4 and Driselase as above and the products were analysed on
the same paper chromatogram as the MLG digest ( - = no IAA;
- - - - = with IAA).
enzyme mixture containing numerous endo- and exo-
glycosidases) yielded peaks of eH]G1c, eH]Man and
eH]Gal as well as several 3H-disaccharides and lar-
ger Driselase-indigestible 3H-products (Figure 2, thick
line).
The same approach, using NaB3H4, can also be
applied to wall polysaccharides present in plant tis-
sues that have not been deliberately treated with ·OH.
This approach has been tested on auxin-treated maize
coleoptiles and on ripening pear fruit - two con-
trasting systems which, however, are both undergoing
wall-loosening.
 63

~- _ -fI (a) Polysaccharide scissio

O.S
+H20 2 +---$ ................
..... 0.5 Ascorbate ........................ ~
....~
....
"C
M Ascorbate \ \
~ ~
(',;j

....c,;:: 0.3
--If--£-_
- ---8
\
\
"'" "'" \ \
(',;j

£ Q,)
0.2

"",\
0.1

0.0 4-----I-i--.1.--------L---------'-'~
0.00 0.01 0.1
~
[Pectin] (% w/v)
2000

i 1800

1SOD

e 1400

fr
'-'
1200

1000

800

SOD

400

[Pectin] (% w/v)
Figure 3. Effects of added pectin on the Cu2+ -induced, 'OH-mediated oxidation of a polysaccharide and of a model aromatic substrate. Each
reaction mixture (l.l mL final volume) contained xyloglucan (0.77%, w/v), [3H1BzG3 (0.25 JLM; 0.125 JLCi/ml) and CUS04 (2 JLM) in acetate
buffer (Na+, 100 mM, pH 4.7). Other additives, as indicated, were ascorbate (0 or I mM), H202 (0 or I mM) and citrus pectin (see x-axis).
After 18 h at 25 DC in an open tube (i.e., + 02), the specific fluidity of each solution was assayed as described (Fry, 1998) to give an indication
of polysaccharide scission (a), and, in the same solutions, the yield of 3H2 0 was measured as an indication of BzG3 oxidation (b). Each
treatment was set up in duplicate, and both the replicate data-points are shown.

Auxin-treated maize coleoptiles
Cell walls were isolated from maize coleoptiles (pre-
incubated with or without IAA) and then treated with
NaB3H4, which would react with the oxo groups
that are known to be introduced into polysaccharides
during 'OH radical attack.
These cell walls did become 3H-Iabelled. However,
their digestion products did not exactly co-migrate
with authentic MLG-products on paper chromato-
graphy after digestion with either lichenase (data not
shown) or Driselase (Figure 2). The major identified
product was [3H]Gal; however, there was no con-
sistent effect of IAA pre-treatment on the yield of
this. We were thus unable to detect evidence for any
auxin-induced increase in 'OH radical attack on either
MLG or other polysaccharides in the walls of maize
coleoptiles.
64
Ripening pear fruit
When the same type of experiment was conducted
on ripening pear fruit, positive results were obtained:
cell walls from fully ripe pear flesh incorporated more
than twice as much 3H into their polysaccharides as
did crisp, unripe fruit (Fry et aI., 2001). The 3H_
polysaccharides released 3H-oligosaccharides upon
digestion with Driselase. These observations indicate
that during pear fruit ripening the cell wall polysac-
charides do undergo chemical changes indicative of
'OH radical attack. This supports the hypothesis that
apoplastic 'OH radicals are generated in vivo during
an episode of wall loosening.
Solubilisation of pectin by ascorbate ± Cu2+
If wall-bound Cu 2+ is responsible for cell wall loosen-
ing during fruit ripening, we would predict that the
addition of ascorbate to isolated cell walls should
result in a promotion of polysaccharide chain cleav-
age (see Equations (1)-(3), above). Recent evidence
supports this prediction in the case of tomato fruit
cell walls (Dumville, 2001); details will be presented
elsewhere.
The presence of acidic pectins does not prevent the
Cu2+ -ascorbate- 02-dependent scission of other
substrates
It might be proposed that in the presence of acidic
pectins, essentially all the Cu2+ would bind to these
polyanions and that other polysaccharides in the same
environment would be almost free of Cu2+. Since
Cu2+ is thought to be a site-specific generator of
'OH radicals, this type of consideration could help to
determine which polysaccharides are targets for non-
enzymic scission by the action of 'OH radicals in the
walls of living plant cells.
However, our recent experimental observations in
vitro indicate that this is not the case. A series of solu-
tions was set up, each containing 2 fLM Cu 2+ in 100
mM acetate buffer, pH 4.7, and two potential sub-
strates that can be attacked by 'OH: xyloglucan and
eH]BzG3 (Figure 1). Xyloglucan scission by 'OH at-
tack can be assayed viscometrically (Fry, 1998), and
eH]BzG3 oxidation in the same solution can simul-
taneously be assayed as 3H20 production (see above).
The effects of all permutations of the following ad-
ditives were tested: pectin (0, 0.012, 0.12 or 1.2%,
w/v), ascorbate (0 or 1 mM) and H202 (0 or 1 mM)
(Figure 3).
In line with earlier experiments (Fry, 1998), H202
alone did little to promote xyloglucan scission; it
also caused little oxidation of BzG3. Ascorbate, es-
pecially in the presence of H202, strongly promoted
both xyloglucan scission and BzG3 oxidation. Pectin
induced a small but significant promotion of BzG3 ox-
idation, especially in the presence ofH202, suggesting
that pectin can mimic ascorbate in reducing Cu2+ to
Cu+, as required for the Fenton reaction. [Free GalA
(Fry, 1998) and oligogalacturonides (unpublished)
also have this ability, whereas neutral sugars (Glc,
Gal and Ara), galactonate and fructose 6-phosphate
do not (Fry, 1998).] The promotion of BzG3 oxid-
ation by pectin indicates that pectin either does not
strongly bind Cu2+ or that pectin-bound Cu 2+ can
readily be reduced to Cu+ and participate in Fenton
reactions, generating 'OH radicals that are not imme-
diately dissipated by reacting with the pectin molecule
itself.
Support for this conclusion came from the observa-
tion that pectin (0.012 or 0.12%) does not appreciably
interfere with the effect of ascorbate (either with or
without H202) on the scission of xyloglucan or on the
oxidation of BzG3 (Figure 3). Pectin at 1.2% (w/v)
gave a ~50% inhibition of BzG3 oxidation, but such
concentrated pectin probably acted principally to scav-
enge . OH rather than to sequester Cu2+. The effect
of 1.2% pectin on xyloglucan scission could not be
assayed reliably (by viscometry) owing to the high
viscosity of 1.2% pectin itself.
The results show that pectic polysaccharides do not
block (indeed, they somewhat promote) the ability of
Cu2+ to be reduced and then to generate 'OH radicals
via the Fenton reaction. Also, pectins do not interfere
(by scavenging) more than any other carbohydrate (at
the same concentration) in the attack of 'OH on other
target molecules present in the same solution.
Conclusion
In conclusion, we suggest that a detailed study of the
binding sites and roles of copper within the primary
cell walls of higher plants is required. In particular,
the conclusion that wall-bound Cu2+ is likely to act
as a site-specific pro-oxidant, leading to the cleavage
of near-neighbouring polysaccharide chains, suggests
that a knowledge of the exact location of bound Cu2+
within the cell wall could hold valuable clues about
the processes of cell expansion, fruit ripening and leaf
abscission.

Acknowledgements
We are very grateful to Miss Sandra C. Sharples and
Mr Andrew Bayne for preliminary data and thank
the Biotechnology and Biological Sciences Research
Councul (U.K.) for funding this work.
References
Aruoma 0 I, Halliwell B, Gajewski E and Dizdaroglu M 1991
Copper-ion-dependent damage to the bases of DNA in the
presence of hydrogen peroxide. Biochem. J. 273, 601-604.
Baldwin T C, Domingo C, Schindler T, Seetharaman G, Stacey N
and Roberts K 2001 DcAGPl, a secreted arabinogalactan pro-
tein, is related to a family of basic proline-rich proteins. Plant
Mo!. Bio!. 45, 421-435.
Biggs K J and Fry S C 1990 Solubilization of covalently bound
extensin from capsicum cell walls. Plant Physio!. 92, 197-204.
Bohlmann H 1994 The role ofthionins in plant protection. Crit. Rev.
Plant Sci. 13, 1-16.
Bown D P, Bolwell G P and Gatehouse J A 1993 Characterization of
potato (Solanum tuberosum L.) extensins - a novel extensin-like
cDNA from dormant tubers. Gene 134, 229-233.
Brennan T and Frenkel C 1977 Involvement of hydrogen peroxide in
the regulation of senescence in pear. Plant Physio!. 59,411-416.
Burkey K 0 1999 Effects of ozone on apoplast/cytoplasm par-
titioning of ascorbic acid in snap bean. Physio!. Plant. 107,
188-193.
Burkitt M J 1994 Copper-DNA adducts. Meth. Enzymo!. 234, 66-
79.
Buxton G V, Greenstock C L, Helman W P and Ross A B 1988 Crit-
ical review of rate constants for reactions of hydrated electrons,
hydrogen atoms and hydroxyl radicals (-OHl'O-) in aqueous
solution. J. Phys. Chern. Ref. Data 17, 513-886.
Carrasco-Luna J, Calatayud A, GonzaIez-Dar6s F and del Valle-
Tasc6n S 1995 Hexacyanoferrate (III) stimulation of elongation
in coleoptile segments from Zea mays L. Protoplasma 184,
63-71.
Cosgrove D J 1999 Enzymes and other agents that enhance cell wall
extensibility. Annu. Rev. Plant Physio!. Plant Mo!. Bio!. 50, 391-
417.
Cutillas-Iturralde A, Zarra I and Lorences E P 1993 Metabolism of
cell wall polysaccharides from persimmon fruit. Pectin solubil-
ization during fruit ripening occurs in the apparent absence of
polygalacturonase activity. Physio!. Plant. 89, 369-375.
Davies C and Robinson S P 2000 Differential screening indicates a
dramatic change in mRNA profiles during grape berry ripening.
Cloning and characterization of cDNAs encoding putative cell
wall and stress response proteins. Plant Physio!. 122,803-812.
Davies M B, Austin J and Partridge D A 1991 Vitamin C: its Chem-
istry and Biochemistry. Royal Society of Chemistry, Cambridge.
pp 1-154.
Dell B 1994 Copper nutrition of Eucalyptus maculata Hook seed-
lings - requirements for growth, distribution of copper and the
diagnosis of copper deficiency. Plant Soil 167, 181-187.
Drew J E and Gatehouse J A 1994 Isolation and characterization
of a pea pod cDNA-encoding a putative blue copper protein
correlated with lignin deposition. J. Exp. Bot. 45, 1873-1884.
Dumville J C 2001 Apoplastic carbohydrate metabolism during
tomato fruit ripening. PhD Thesis, The University of Edinburgh.
65
Estevez A G, Crow J P, Sampson J B, Reiter C, Zhuang Y X,
Richardson G J, Tarpey M M, Barbeito L and Beckman J S
1999 Induction of nitric oxide-dependent apoptosis in motor
neurons by zinc-deficient superoxide dismutase. Science 286,
2498-2500.
Federico R and Angelini R 1986 Occurrence of diarnine oxidase in
the apoplast of pea epicotyls. Planta 167, 300-302.
Frahry G and Schopfer P 2001 NADH-stimulated, cyanide-resistant
superoxide production in maize coleoptiles analyzed with a
tetrazolium-based assay. Planta 212, 175-183.
Frahry G and Schopfer P 1998 Hydrogen peroxide production by
roots and its stimulation by exogenous NADH. Physio!. Plant.
103, 395-404.
Fry S C 1986 Cross-linking of matrix polymers in the growing cell
walls of angiosperms. Annu. Rev. Plant Physio!. 37, 165-186.
Fry S C 1989 Cellulases, hemicelluloses and auxin-stimulated
growth: a possible relationship. Physio!. Plant. 75, 532-536.
Fry S C 1995 Polysaccharide-modifying enzymes in the plant cell
wal!. Annu. Rev. Plant Physio!. Plant Mo!. Bio!. 46, 497-520.
Fry S C 1998 Oxidative scission of plant cell wall polysaccharides
by ascorbate-induced hydroxyl radicals. Biochem. 1. 332, 507-
515.
Fry S C, Dumville J C and Miller J G 2001 Fingerprinting of poly-
saccharides attacked by hydroxyl radicals in vitro and in the cell
walls of ripening pear fruit. Biochem. J. 357, 729-735.
Ghiselli A 1998 Aromatic hydroxylation: salicylic acid as a probe
for measuring hydroxyl radical production. In Free Radical and
Antioxidant Protocols. Ed. D. Armstrong. pp. 89-100. Humana
Press, Totowa, New Jersey.
Griffiths H R and Lunec J 1996. Investigating the effects of oxy-
gen free radicals on carbohydrates in biological systems. In Free
Radicals: a Practical Approach. Eds. NA Punchard and FJ Kelly.
pp. 185-200. IRL Press, Oxford.
Grootveld M and Halliwell B 1986 Aromatic hydroxylation as a po-
tential measure of hydroxyl radical formation in vivo. Biochem.
1. 237, 499-504.
Halliwell B and Gutteridge J M C 1990 Role of free radicals
and catalytic metal ions in human disease: an overview. Meth.
Enzymo!. 186, 1-85.
Hayashi T 1989 Xyloglucans in the primary cell wal!. Annu. Rev.
Plant Physio!. Plant Mo!. Bio!. 40, 139-168.
Hidalgo A, Garcia-Herdugo G, Gonzalez-Reyes J A, Morre D J
and Navas P 1991 Ascorbate free-radical stimulates onion root-
growth by increasing cell elongation. Bot. Gaz. 152, 282-288.
Iwasaki K, Sakurai K and Takahashi E 1990 Copper-binding by the
root cell walls of Italian ryegrass and red-clover. Soil Sci. Plant
Nutr. 36,431-439.
Jarvis S C 1978 Copper uptake and accumulation by perennial
ryegrass grown in soil and solution culture. J. Sci. Food Agric.
29, 12-18.
Jarvis S C 1981 Copper concentrations in plants and their re-
lationships to soil properties. In Copper in Soils and Plants.
Eds. IF Loneragan, AD Robson and RD Graham. pp. 265-285.
Academic Press, Sydney.
Jones J K N and Stoodley R J 1965 Fractionation using copper
complexes. In Methods in Carbohydrate Chemistry, Vol 5. Ed.
RL Whistler. pp. 36-38. Academic Press, New York.
Kieliszewski M J, Kamyab A, Leykam J F and Lamport D T A 1992
A histidine-rich extensin from Zea mays is an arabinogalactan
protein. Plant Physiol. 99, 538-547.
Kingston-Smith A H and Foyer C H 2000 Overexpression of
Mn-superoxide dismutase in maize leaves leads to increased
monodehydroascorbate reductase, dehydroascorbate reductase
and glutathione reductase activities. 1. Exp. Bot. 51, 1867-1877.
66
Kramer V, Cotter-Howells J D, Charnock J M, Baker A J M and
Smith J A C 1996 Free histidine as a metal chelator in plants that
accumulate nickel. Nature 379, 635-638.
Lane B G, Dunwell J M, Ray J A, Schmitt M R and Cuming A C
1993 Germin, a protein marker of early plant development, is an
oxalate oxidase. J. BioI. Chern. 268, 12239-12242.
Liao M T, Hedley M J, Woolley D J, Brooks R R and Nichols M
A 2000 Copper uptake and translocation in chicory (Cichorium
intybus L. cv Grasslands Puna) and tomato (Lycopersicon escu-
lentum Mill. cv Rondy) plants grown in NFT system. II. The role
of nicotianamine and histidine in xylem sap copper transport.
Plant Soil 223, 243-252.
Lin C C and Kao C H 2001 Abscisic acid induced changes in cell
wall peroxidase activity and hydrogen peroxide level in roots of
rice seedlings. Plant Sci. 160,323-329.
Mann K, Schafer W, Thoenes V, Messerschmidt A, Mehrabian Z
and Nalbandyan R 1992 The amino-acid-sequence of a type-I
copper protein with an unusual serine-rich and hydroxyproline-
rich C-terminal domain isolated from cucumber peelings. FEBS
Lett. 314, 220-223.
Miller J G and Fry S C 200 I Characteristics of xyloglucan after
attack by hydroxyl radicals. Carbohydr. Res. 332, 389-403.
Mom~, D J, Brightman A 0, Hidalgo A and Navas P 1995 Select-
ive inhibition of auxin-stimulated NADH oxidase activity and
elongation growth of soybean hypocotyls by thiol reagents. Plant
Physiol. 107, 1285-1291.
Mom~ D J, Brightman A 0, Wu L-Y, Barr R, Leak B and Crane F
L 1988 Role of plasma membrane redox activities in elongation
growth in plants. Physiol. Plant. 73, 187-193.
Nikdel S, Mackellar D G and Rezaaiyan R 1991 Analysis of
mineral-content and amount of chelated minerals in citrus juice
by inductively coupled plasma emission- spectroscopy. J. Agric.
Food Chern. 39, 1773-1775.
Ohta Y, Shiraishi N, Nishikawa T and Nishikimi M 2000 Copper-
catalyzed autoxidations of GSH and L-ascorbic acid: mutual
inhibition of the respective oxidations by their coexistence.
Biochim. Biophys. Acta 1474, 378-382.
Oikawa S, Kurasaki M, Kojima Y and Kawanishi S 1995 Oxidative
and non-oxidative mechanisms of site-specific DNA cleavage in-
duced by copper-containing metallothioneins. Biochemistry 34,
8763-8770.
O'Neill M, Albersheim P and Darvill A G 1990 The pectic polysac-
charides of primary cell walls. In Methods in Plant Biochemistry
2. Ed. PM Dey. pp. 415-441. Academic Press, London.
Parker A J 1981 The chemistry of copper. In Copper in Soils and
Plants. Eds. JF Loneragan, AD Robson, RD Graham. pp. 1-24.
Academic Press, Sydney.
Paul A A and Southgate D T A 1978 McCance and Widdowson's
The Composition of Foods. Her Majesty's Stationery Office,
London.
Rao P S 1959 Tamarind. In Industrial Gums. Ed. RL Whistler. pp.
461-504. Academic Press, New York.
Rautenkranz A A F, Li L, Machler F, Martinoia E and Oertli J J 1994
Transport of ascorbic and dehydroascorbic acids across proto-
plast and vacuole membranes isolated from barley (Hordeum
vulgare L. cv Gerbel) leaves. Plant Physiol. 106, 187-193.
Read S M, Clarke A E and Bacic A 1993 Stimulation of growth of
cultured Nicotiana tabacum W-38 pollen tubes by poly(ethylene
glycol) and Cu-(II) salts. Protoplasma 177, 1-14.
Redgwell R J, Melton L D and Brasch D J 1992 Cell-wall dissolu-
tion in ripening kiwifruit (Actinidia deliciosa) - solubilization of
the pectic polymers. Plant Physiol. 98, 71-81.
Rossi A, Maccarrone M, Buganza M and Agro A F 1995 Exogen-
ously added copper does not increase the production of copper
amine oxidase in lentil seedlings. Gene 154, 167-170.
Sakurai H, Satoh H, Hatanaka A, Sawada T, Kawano K, Hagino T
and Nakajima K 1994 Vnusual generation of hydroxyl radicals
in hepatic copper-metallothionein of lec (Long-Evans cinnamon)
rats in the presence of hydrogen-peroxide. Biochem. Biophys.
Res.Commun.199,313-318.
Sakurai N 1998 Dynamic function and regulation of apoplast in the
plant body. J. Plant Res. 111, 133-148.
Schlemmer V 1989 Studies of the binding of copper, zinc and
calcium to pectin, alginate, carrageenan and gum guar in HC03-
C02 buffer. Food Chern. 32, 223-234.
Schopfer P, Plachy C and Frahry G 2001 Release of reactive
oxygen intermediates (superoxide radicals, hydrogen peroxide,
hydroxyl radicals) and peroxidase in germinating radish seeds
controlled by light, gibberellin, abscisic acid. Plant Physiol. 125,
1591-1602.
Schuchmann M N and von Sonntag C 1978 The effect of oxygen
on the OH-radical-induced scission of the glycosidic linkage of
cellobiose. Int. J. Radiat. BioI. 34, 397-400.
Smimoff N and Wheeler G L 2000 Ascorbic acid in plants: bio-
synthesis and function. Crit. Rev. Biochem. Mol. BioI. 35,
291-314.
Sommer-Knudsen J and Bacic A 1997 A micro-scale method for
determining relative metal-binding affinities of proteins. Mol.
Biotechnol. 8,215-218.
Sommer-Knudsen J, Clarke A E and Bacic A 1996 A galactose-rich,
cell-wall glycoprotein from styles of Nicotiana alata. Plant J. 9,
71-83.
Sommer-Knudsen J, Clarke A E and Bacic A 1997 Proline-
and hydroxyproline-rich gene products in the sexual tissues of
flowers. Sex. Plant Reprod. 10, 253-260.
Sommer-Knudsen J, Bacic A and Clarke A E 1997 A metal-binding,
cell-wall glycoprotein from Nicotiana alata. J. Exp. Bot. 48
Suppl,102.
Stuart D A and Varner J E 1980 Purification and characterization of
a salt-extractable hydroxyproline-rich glycoprotein from aerated
carrot discs. Plant Physiol. 66, 787-792.
Suresh K and Subramanyam C 1998 Polyphenols are involved in
copper binding to cell walls of Neurospora crassa. J. Inorg.
Biochem.69,209-215.
Suzuki K T, Rui M, Veda J and Ozawa T 1996 Production of hy-
droxyl radicals by copper-containing metallothionein: Roles as
prooxidant. Toxicol. Appl. Pharmacol. 141,231-237.
Tabbi G, Fry S C and Bonomo R P 2001 ESR study of the
non-enzymic scission of xyloglucan by an ascorbate-H202-
copper system: the involvement of the hydroxyl radical and the
degradation of ascorbate. J. Inorg. Biochem. 84, 179-187.
Takahama V 1993 Redox state of ascorbic acid in the apoplast of
stems of Kalanchoe daigremontiana. Physiol. Plant. 89, 791-
798.
Tanaka K, Takio S and Satoh T 1995 Inactivation of the cytoso-
lic Cu/Zn-superoxide dismutase induced by copper deficiency in
suspension-cultured cells of Marchantia paleacea var diptera. J.
Plant Physiol. 146,361-365.
Turnbull C R A, Beadle C L, West P W and Cromer R N 1994 Cop-
per deficiency a probable cause of stem deformity in fertilized
Eucalyptus nitens. Can. J. Forest Res. 24, 1434-1439.
van Cutsem P and Gillet C 1982 Activity coefficients and selectivity
values of Cu++, Zn++ and Ca++ ions adsorbed in the Nitella
fiexilis L. cell wall during triangular ion exchanges. J. Exp. Bot.
33, 847-853.

Walker C D and Webb J 1981 Copper in plants: forms and beha-
viour. In Copper in Soils and Plants. Eds. JF Loneragan, AD
Robson and RD Graham. pp. 189-212. Academic Press, Sydney.
Wallace G and Fry S C 1994 Phenolic components of the plant cell
wall. Int. Rev. Cytol. 151,229-267.
Weger H G 1999 Ferric and cupric reductase activities in the green
alga Chlamydomonas reinhardtii: experiments using iron-limited
chemostats. Planta 207, 377-384.
Wucherpfennig K 1992 Removing iron and copper ions out of bever-
ages by means of selective complexing agents. 1. Singling out
complexing agents. Deut. Lebensm.-Rundschau 88, 313-317.
67
Yim M B, Chock P Band Stadtman E R 1993 Enzyme function of
copper, zinc superoxide dismutase as a free-radical generator. J.
BioI. Chern. 268, 4099-4105.
Zarra I, Sanchez M, Queijeiro E, Pena M J and Revilla G 1999 The
cell wall stiffening mechanism in Pinus pinaster Aiton: regula-
tion by apoplastic levels of ascorbate and hydrogen peroxide. J.
Sci. Food Agric. 79, 416-420.
Zheng Y B, Lyons T, Ollerenshaw J H and Bames J D 2000
Ascorbate in the leaf apoplast is a factor mediating ozone resist-
ance in Plantago major. Plant Physiol. Biochem. 38, 403-411.
The role of the apoplast in plant mineral nutrition
 Plant and Soil 247: 71-80,2002.
© 2002 Kluwer Academic Publishers.
71

The functions of cell wall polysaccharides in composition and architecture

revealed through mutations

Nicholas C. Carpita1,3 & Maureen C. McCann2

1Department of Botany and Plant Pathology, Purdue University, West Lafayette, IN 47907-[[55, U.S.A.
2Department of Cell and Developmental Biology, John Innes Centre, Norwich NR4 7UH, u.K. 3Corresponding
author

Received 19 September 2001. Accepted in revised form 25 February 2002

Key words: Arabidopsis, cellulose, cell wall, FfIR spectroscopy, mutants, pectins

Abstract
The plant cell wall is a highly organized composite of many different polysaccharides, proteins and aromatic sub-
stances. These complex matrices define the shape of each individual cell, and ultimately, they are the determinants
of plant morphology. The fine structures of the major angiosperm cell wall polysaccharides have been character-
ized, but it is not well understood how these polysaccharides are assembled into a metabolically active architecture.
Cell wall biogenesis and remodeling may be partitioned into six major stages of development (precursor synthesis,
polymerization, secretion, assembly, rearrangement and disassembly), and to date, a handful of mutations have
been identified that affect the composition and structure in each of these stages. To greatly augment this collection,
we have initiated a program to use Fourier transform infrared spectroscopy as a high through-put screen to identify
a broad range of cell-wall mutants of Arabidopsis and maize. We anticipate that such mutants will be useful to probe
the impact of the individual components and their metabolism on basic processes of plant growth and development.
The structures of dicot and grass walls, the identification of representative cell wall mutants, and the use of a novel
spectroscopic screen to identify many more cell wall mutants, are briefly reviewed.

Abbreviations: FfIR - Fourier transform infrared; GAX - glucuronoarabinoxylan; HGA - homogalacturonan;

RG - rhamnogalacturonan; XET - xyloglucan endo-transglycosylases; XyG - xyloglucan

Introduction of cellulose and cross-linking xyloglucans (XyGs) and

The primary cell walls of plants are made up of two,
and sometimes three, structurally independent but in-
teracting networks. The fundamental framework of
cellulose microfibrils and cross-linking glycans lies
embedded in a second network of matrix pectic poly-
saccharides (Carpita and Gibeaut, 1993; McCann and
Roberts, 1991). The third independent network con-
sists of the structural proteins or a phenylpropanoid
network. Briefly, Arabidopsis (Zablackis et al., 1995)
and other dicots and the non-commelinoid monocots
possess 'Type I' walls (for review, see Carpita and
Gibeaut, 1993), which contain about equal amounts
* FAX No: +1-765-494-0363.
E-mail: carpita@btny.purdue.edu
various minor amounts of arabinoxylans, glucoman-
nans and galacto-glucomannans. The cellulose-XyG
framework is embedded in a pectin matrix that con-
trols several physiological properties, such as wall
porosity, charge density and microfibril spacing. The
two major pectins are homogalacturonans (HGAs) and
rhamnogalacturonan I (RG I). Some HGAs are cross-
linked into junction zones by Ca2+, whereas other
HGAs and RGs are cross-linked by ester linkages to
other pectins or polymers held more tightly in the wall
matrix and can only be released from the wall by de-
esterifying agents (for review, see Willats et aI., 2001).
An unusual polysaccharide called RG II, a relatively
minor pectic component, has the richest diversity of
sugars and linkage structures known. Neutral poly-
72
mers (arabinans or galactans) are pinned at one end
to the RG I pectic backbone, but extend into, and
are highly mobile in, the wall pores. Some Type I
walls contain several types of structural proteins that
may interact with the pectin network. The various
structural proteins could form intermolecular bridges
with other proteins without necessarily binding to the
polysaccharide components.
The walls of cereals and related monocots are
different
Maize and other commelinoid monocots possess a dif-
ferent kind of primary wall, a 'Type II' wall (Carpita,
1996; Carpita and Gibeaut, 1993). They contain cellu-
lose microfibrils of the same structure as those of the
Type I wall, but glucuronoarabinoxylans (GAXs) are
the principal polymers that interlock the microfibrils.
Unbranched GAXs can hydrogen bond to cellulose
or to each other. The attachment of arabinose and
glucuronic acid side-groups to the xylan backbone
of GAXs prevents the formation of hydrogen bonds,
diminishing the extent of cross-linking between two
unbranched GAX chains or GAX to cellulose. In gen-
eral, grasses are pectin-poor, but with the exception
of the lack of fucose, the pectins they do contain
are similar in structure to those of dicots. When
grass cells begin to elongate, they accumulate mixed-
linked (1 ~3),(1~4),B-D-glucansin addition to GAX.
The mixed-linkage ,B-glucans are unique to the Po-
ales (Smith and Harris, 1999) and are an example of
a developmental-stage-specific polysaccharide associ-
ated with cell expansion (Carpita, 1996).
Grasses, which have very little structural protein
compared with dicots and non-commelinoid mono-
cots, have extensive interconnecting networks of
phenylpropanoids that form primarily when cells stop
expanding (Iiyama et al., 1994). Their richness in
phenolic substances gives species with Type II walls
the distinguishing property of autofluorescence in their
primary walls (Rudall and Caddick, 1994). In the
non-lignified walls, the principal hydroxycinnamate is
ferulic acid, whereas in the lignified walls, both fer-
ulic and p-coumaric acid are found. Ferulic acid is
esterfied to the C-5 of arabinofuranose sidechains of
arabinoxylans and they can dimerize, either via a pho-
tochemical reaction leading to cyclodimers (Ford and
Hartley, 1989) or via radical-mediated dimerization
which gives rise to diferulate dehydrodimers (Ralph
et al., 1995).
Cell-wall mutants will help plant biologists to
deduce gene function
Over 17% of the 25 498 Arabidopsis genes have signal
peptides, and over 400 proteins have been identified
that reside in the wall (Arabidopsis Genome Initiative,
2000). If just one-half of the proteins with signal pep-
tides function in the biosynthesis, assembly and modi-
fication of the walls, then well over 2000 genes are
likely to participate in wall biogenesis and remodeling
during plant development. If all the cytosolic proteins
that function in substrate generation are included, the
number increases significantly. Some integral plasma
membrane-associated proteins that function in wall as-
sembly, such as cellulose synthase, do not contain
signal peptides. Thus, it is likely that some 15% ofthe
Arabidopsis genome is dedicated to cell wall biogen-
esis and modification. Only a few gene families that
encode wall-relevant genes have been characterized,
and it is estimated that over one-half of wall-relevant
genes have yet to be annotated.
The major steps in wall biogenesis and modi-
fication can be divided into six specific stages: (1)
the synthesis of monomer building blocks, such as
nucleotide-sugars and monolignols; (2) the biosyn-
thesis of oligomers and polysaccharides at the plasma
membrane and ER-Golgi apparatus; (3) the targeting
and secretion of Golgi-derived materials; (4) the as-
sembly and architectural patterning of polymers; (5)
dynamic rearrangement during cell growth and dif-
ferentiation; (6) wall disassembly and catabolism of
polymers.
For some of these stages, such as the generation
of known substrates, complete knowledge of the bio-
chemical pathways has led to discovery of many of
the genes encoding the enzymes involved in the cata-
lysis. For other stages, such as wall assembly, we can
only speculate on the nature of the proteins that might
participate. Recently, forward (Reiter et al., 1997;
Turner and Somerville, 1997) and reverse (Reiter,
1998) genetic approaches have provided new insights
on wall-relevant genes. Forward genetic approaches
have historically been hampered by difficult technical
problems associated with characterization of polymer
synthesis in vitro and of higher order architectural
assembly and rearrangement during growth.
The first broad screen for cell-wall mutants was
conducted by Wolf-Dieter Reiter in the laboratory
of Chris Somerville, where they selected 11 loci re-
sponsible for certain cell-wall monosaccharides to be
over- or under-represented when compared to wild-

type (Reiter, 1998; Reiter et al., 1997). Among these
mutants, called 'mur' (from the Latin murus, mean-
ing 'wall'), were three in which fucose was either
completely absent (murl) or under-represented by
about 50% (mur2 and mur3). The murl mutation was
mapped to a gene that encoded a 4,6-dehydratase that
functions in the conversion of GDP-Man to GDP-
Fuc (Bonin et al., 1998). Because the mutation lies
in a gene that is responsible for the synthesis of the
substrate for every polymer containing fucose, all
polymers lack this sugar. For XyG, a majority of
the galactosyl extensions of the xylosyl residues that
would normally contain L-Fuc are left naked, but
about 20% of them contain L-Gal in place of L-Fuc
(Zablackis et al., 1996). The MUR4 gene is thought
to encode a C4 epimerase that interconverts UDP-
Xyl and UDP-Ara (Burget and Reiter, 1999). The
mur4 mutant has approximately one-half of the nor-
mal arabinose content in its leaf cell walls (Reiter
et al., 1997), and an alternative pathway to UDP-
Ara may provide the limited amount of arabinose in
these plants. With several candidate mutations of the
nucleotide-sugar interconversion pathway, the wild-
type phenotype can be restored by feeding plants with
the deficient sugar, which can then form the nucle-
otide sugar by salvage pathways involving C-1 kinases
and NDP-pyrophosphorylases. Such is the case with
murl (Reiter et aI., 1993) and mur4 (Burget and Reiter,
1999).
Mutants with altered cellulose deposition have
been identified directly by loss of birefringence, for
example in trichomes (Potikha and Delmer, 1995).
Other cell-wall mutations have appeared serendipit-
ously in the course of screens for developmental phen-
otypes, such as a temperature-sensitive radial cell-
swelling mutant (rsw1) that was partially cellulose-
deficient (Baskin et al., 1992) or plants with a
collapsed xylem phenotype (Turner and Somerville,
1997). Delmer et al. (Pear et al., 1996) identified
a cDNA produced exclusively during secondary wall
formation in cotton fibers that is most likely a can-
didate for the catalytic subunit of cellulose synthase
(CesA). Arioli et al. (1998) restored normal cellu-
lose synthesis in the Arabidopsis rsw 1 mutant by
genetic complementation with a CesA gene. Simil-
arly, the irregular xylem mutants irx1 and irx3, which
have reduced cellulose content in the secondary walls
of tracheary elements (Turner and Somerville, 1997)
have been traced to defects in CesA genes (Taylor et
al., 1999,2000). A survey of the Arabidopsis database
revealed several dozen related genes that could func-
73
tion in the synthesis, not only of cellulose, but perhaps
also of many other cross-linking glycans that com-
prise the plant cell wall. Synthases of all cross-linking
glycans containing (1-+4)I3-D-glucosyl, mannosyl, or
xylosyl backbones must overcome the steric problem
of flipping one sugar almost 1800 with respect to
each neighboring sugar (Carpita and Vergara, 1998;
Gibeaut and Carpita, 1994). One of the conserved
features of the CesA and cellulose-synthase-like (CsT)
gene families (Richmond and Somerville, 2001) are
four motifs responsible for UDP-sugar binding and
catalysis in 13(1-+4)-linked polymers. Such motifs in
large and diverse families may mean that identifica-
tion of the cross-linking glycan backbone polymerases
is close at hand. Mutants may also reveal unpredict-
able roles for known enzymes and proteins. For ex-
ample, a unique membrane-associated endo-glucanase
(Brummell et al., 1997) appears to be involved in
cellulose synthesis, as judged by a loss-of-function
mutant called korrigan in Arabidopsis (Nicol et aI.,
1998).
The complex polysaccharides of the plants' extra-
cellular matrix are not random mixtures of sugars and
sugar linkages but assemblies of discrete unit struc-
tures. For example, the unit structure of the predomin-
ant XyG in flowering plants consists of a unit structure
of a cellotetraose unit to which xylose units are ad-
ded to the 0-6 position of three contiguous glucosyl
residues (Figure 1). To about one-half of these sub-
units, an a-L-fucosyl-(l-+ 2)I3-D-galactosyl-(1-+ 2)-
side-group is added specifically to the xylose residue
closest to the reducing end of the subunit. In Ar-
abidopsis and other dicots, about one-half of these
fucogalacto-side-chains also contain an unfucosylated
I3-D-galactosyl-(l-+ 2)- extension on the middle xy-
lose unit. Less frequently, an a-L-arabinosyl unit is
added to the 0-2 position of the glucan chain specific-
ally at the xylosyl-bearing unit closest to the reducing
end (Figure 1).
Although the XyG polymer may be assembled
from various combinations of these basic subunits, the
reason for such specificity of linkage structure is inher-
ent in the composition of the synthase complex that
makes the polymer. One may estimate that at least
eight unique enzyme activities are needed to synthes-
ize a fully decorated XyG unit (Figure 2). Because
of the difficulties that investigators have experienced
in purifying such complexes from their membrane
locations in an active form, molecular biological ap-
proaches have now been employed to identify the
genes that encode the components of the XyG syn-
 -.l
.j::..

(Fucogalacto-)Xylogl ucans

.....

[~]
u-L-Fuc~
.l.I\)
I3-D-Gaf,::
.l.I\)
u-D-XyG: u-D-Xyr.:;: u-D-Xy(.: u-D:Xy~
.l. .l. .l. .l.
m m m m
~4)I3-D-Glc-(1 ~4)I3-D-Glc-(1 ~4)tl::.D-Glc-(1 ~4)I3-D-GIC-(1 ~4)I3-D-Glc-(1~4)I3-D-Glgj1 ~4)I3-D-~-(1 ~4)I3-D-Glc-(1 ~4)I3-D-Glc-(1 ~
w W N
i i i
u-D-Xyl-S u-D-Xy~ [u-L-Arad

~--------------------------~I~I----------------------------~
XXXG XXFG
Figure 1. The fundamental building block of (fucogalacto)xyloglucan is a cellotetraose unit to which xylose units are added to the 0-6 position of the three contiguous glucosyl residues.
Oligomers containing these unit structures are generated by hydrolysis of xyloglucan by a Trichoderma cellulase that is restricted to cleavage of the reducing end of unbranched glucosyl
residues (arrows). A single-letter nomenclature, referring to the terminal subtending sugar or G for glucose in the backbone, is shown for two subunit structures (Fry et al., 1993). To about
one-half of these fundamental subunits (XXXG), an a-L-fucosyl-(1---*2)fJ-D-galactosyl-O---*2)- extension is added to the xylose residue closest to the reducing end (XXFG). In Arabidopsis
and other dicots, about one-half of these fucogalacto-extended subunits also contain a fJ-D-galactosyl-O---*2)- extension on the middle xylose unit. Less frequently, an a-L-arabinosyl unit is
added to the 0-2 position of the glucan chain specifically at the xylosyl-bearing unit closest to the reducing end.

Figure 2. An estimated minimum of eight enzymes is required to
construct a xyloglucan. A glucan synthase (1) that synthesizes the
backbone is coordinated with three distinct xylosyl transferases (2,
3 and 4). It is doubtful that a single xylosyl transferase adds all three
residues because of spatial and symmetry considerations. A galacto-
syl transferase (5) extends the first xylosyl residue, and a fucosyl
transferase (6) extends the side-group further. Mutations have been
identified that are defective in the activities of these enzymes. A
second galactosyl transferase (7) adds this residue to the middle
xylosyl residue. This activity is enhanced in mutants that fail to
produce the galactosyl transferase (5). An arabinosyl transferase (8)
adds this residue to the glucan chain, precisely at the 0-2 position
of the glucosyl residue containing the trisaccharide side-group.
thase. The most powerful means of gene identification
is the characterization of mutants that are missing
one or more of the components of synthesis. Mur2
and mur3 are mutations in genes that enclose en-
zymes associated with XyG synthase specifically; the
MUR2 gene encodes the XyG-specific fucosyl trans-
ferase (Vanzin et aI., 2002) and MUR3 encodes the
pre-requisite galactosyl transferase that initially ex-
tends the xylosyl residue (Madson et aI., personal
communication).
The murl mutation causes plants to be slightly
stunted and results in a loss of tensile strength of the
floral stem (Reiter et aI., 1993). The loss in tensile
strength was attributed to the loss of fucose from
XyG, because modeling studies had suggested that
the fucosylated trisaccharide plays a role in physically
straightening the glucan backbone to facilitate binding
to cellulose microfibrils (Levy et aI., 1991). Hence,
75
loss of fucose meant loss of this function, and a com-
promised XyG-cellulose binding would be expected
to result in loss of cell wall strength. However, mur2
and mur3 plants, despite complete loss of the trisac-
charide function, are phenotypically indistinguishable
from wild-type in every respect, and tensile strength
of the floral stalk is comparable to wild-type (Madson
et aI., personal communication). These findings indic-
ate that the loss of wall strength in the murl plants
may be a consequence of a loss of fucose from pectic
polysaccharides.
The synthesis of XyGs is relatively simple when
compared to the synthesis of some pectic polysacchar-
ides. Synthesis of HGA requires a minimum of two en-
zymes, as the monomeric backbone GalA residues are
subsequently methyl esterified (Doong and Mohnen,
1998) and synthesis of RG I requires three: two
to assemble the a-L-Rha-(l---+ 2)-a-D-GalA-(1---+4)-
backbone unit, with additional acetyl esters on some
of the GalA residues. Several neutral sugar chains
are added to the rhamnosyl units at certain stages of
cell development, notably the arabinans, galactans,
and arabinogalactans (Carpita and Gibeaut, 1993).
The mur5 and muro mutations also result in reduced
amounts of cell-wall arabinose (Reiter, 1998), but they
cannot be rescued by feeding of arabinose, and there-
fore, the genes affected encode enzymes involved in
downstream events in wall biogenesis.
One of the most complex polysaccharides ever dis-
covered in nature is RG II, with at least 21 enzymes
required for its synthesis (Figure 3). The function of
each monosaccharide in the structure is far from un-
derstood, but one of the essential functions of the RG
II polymer is to cross-link to form boron di-diester
dimers, which limits pore size (Fleischer et aI., 1998)
and increases tensile strength. Walls swell when plants
are grown on boron-deficient media (Ishii et aI., 2001).
Identification of all the genes involves in the synthesis
and assembling of RG II will be a daunting task.
Secretion mutants related to vesicle targeting have
been well characterized in yeast, but comparable de-
fects in plants have not (Staehelin and Moore, 1995).
Judging from the yeast paradigm, such mutations
would be predicted to result in a general loss in wall
mass with concomitant accumulation of cell-wall con-
taining vesicles in the cytoplasm. A single secretion
mutant, called knolle, with a defective syntaxin gene,
does not complete formation of the cell plate and cells
are not correctly partitioned (Lukowitz et aI., 1996).
Wall assembly mechanisms are not understood at
all, and we cannot predict the impact of defects to
 -.J
0\

~G)

@)
~

HO
~~:
Figure 3. At least 21 enzymes are required to assemble a rhamnogalacturonan II molecule. This is one of the most complex biopolymers in Nature and contains several rare sugars, including
3-deoxy-D-manno-2-octulosonic acid (KDO), 3-deoxY-D-lyxo-2-heptulosonic acid (DHA), aceric acid (3-C'-carboxy-5-deoxY-L-xylose), apiose, 2-0-methyl xylose, and 2-0-methyl fucose.
The molecule dirnerizes to form boron di-diesters with the first apiose residues (2). This dimerization is associated with increases in tensile strength and decreased porosity.

these processes. Important mutants to identify are the
enzymes and proteins that function in wall rearrange-
ment during growth, such as expansins and xyloglucan
endotransglycosylases (XETs). Exogenous expansin
can alone trigger leaf primordium formation in the
epidermal cells of an apical dome (Fleming et al.,
1997) and so mutants might be expected to have
severe developmental phenotypes. Since many hy-
drolases function as transglycosylases, enzymes such
as XET may have a strict transglycosylase activity
(Nishitani, 1995). Such activities may serve roles in
wall assembly rather than in wall dynamics and dis-
assembly (Thompson et aI., 1997). One member of
the XET family is touch-inducible (Xu et aI., 1995).
Dwarf and miniature mutants represent a visible class
of mutants, whose basis may lie in defects in wall
dynamics caused by lack of XETs (Akamatsu et al.,
1999; Verica and Medford, 1997).
Hydrolysis of polymers during wall disassembly
results in production of several monosaccharides that
are salvaged back into the nucleotide-sugar pool (Re-
iter and Vanzin, 2001). Defects in these pathways
would result in an inability to recycle sugars and may
lead to reduced content of those sugars in the wall.
Many of these sugars are also phytotoxic. In fact,
an arabinose C-1 kinase mutant was discovered be-
cause the plants are unable to grow in the presence
of the high concentrations of arabinose (Dolezal and
Cobbett, 1991).
Antisense inhibition experiments have been of
equivocal value in assessing the role in growth and
development of cell-wall related proteins because of
the spectrum of functions that individual members
of a gene family may possess. Two additional reas-
ons for the difficulty in interpreting the phenotypes
of antisense plants stem from either excess enzyme
levels such that pathways can proceed normally with
as little as 1% of the normal level, or genetic redund-
ancy. Meaningful studies may need to await discovery
of activity-controlling proteins revealed as develop-
mental mutants. Many other developmental mutants
exist whose defects are likely to be traced to cell
wall components (Benfey et al., 1993; Lolle et al.,
1997; Reynolds et aI., 1998; Smith et al., 1996), al-
though their functions in wall architecture remain to
be elucidated.
77
A new high through-put method to identify
cell-wall mutants
Given the enormous range of cellular, developmental,
and molecular events that the six stages of cell-wall
biogenesis and remodeling comprise, and the paucity
of cell-wall mutants in hand, we need to augment the
mutant collection substantially. We have developed a
powerful method to do this. Fourier transform infrared
(FTIR) microspectroscopy is an extremely rapid, non-
invasive vibrational spectroscopic method that can
quantitatively detect a range of functional groups (Mc-
Cann et aI., 1992, 1997; Sene et al., 1994). We have
optimized FTIR as a high through-put screen for cell
wall mutants (Chen et aI., 1998). The advantages of
FTIR spectroscopy for a mutant screen are many-fold.
The time taken to acquire a spectrum of a sample is
roughly 30 s. It is non-destructive and does not require
derivatization of the sample, so further assays can be
applied to the same sample once potential mutants are
identified. Several functional groups absorb infrared
radiation at characteristic frequencies, making assign-
ments of some specific wall components possible. In
particular, frequencies corresponding to carboxylic
and phenolic esters and the carboxylate stretches of
uronic acids are clearly resolved in the spectrum (Fig-
ure 4A). As chemically different specimens can give
rise to visually similar spectra, or may have very
subtle differences, we have used chemometric meth-
ods to analyze these multivariate data. By using data
compression methods such as Principal Component
Analysis, we can transform a data set comprising a
large number of inter-correlated variates into a new
set of a small number of variables (Figure 4B). The
derived variables are known as 'loadings', and are un-
correlated and ordered such that the first few account
for the majority of the variance that was spread across
all of the original variates (Kemsley, 1998). Use of this
algorithm enables one to analyze very large popula-
tions of spectra (Figure 4B). Additional chemometric
methods can be applied, including Linear Discrimin-
ant Analysis, which tests if all members of a given
population vary significantly from all members of a
wild-type population, and other algorithms that test if
a specimen is a member of the 'wild-type' class at the
95 or 99% probability level (Kemsley, 1998).
We are also using reverse-genetic approaches to
identify cell wall biogenesis-related genes in Ara-
bidopsis and maize. Cloning of these genes is simpli-
fied because of the use of insertional mutants, either
transposon-tagged maize or T-DNA transformed Ara-
78

B o• ChIorite-NaOH
Control
•

• o

,
PCl
1734 U38 1734

1800 1600 1400 1200 1000 800

Wavenumber, em-1 Pe2Loading
Figure 4. How Discriminant Analysis of Fourier transformed infrared spectra is used to differentiate cell walls of different composition. (A)
Spectra are shown of ten samples each of maize walls from control and those sequentially treated with acidic sodium chlorite to oxidize phenolic
substances and 0.1 m NaOH to remove mostly highly substituted GAXs. A difference spectrum shows that ester (1238 and 1734 cm- 1) and
phenolic residues account for the major differences revealed in a difference spectrum of the averaged spectra of these two groups. (B) Principal
Components 1 and 2 account for over 80% of the total variation among spectra of both samples (30 each) and completely resolve them. The
'loading' of Principal Component 2, which is the primary discriminator, illustrates the differences are primarily due to ester linkages and
phenolic residues lost in the treated walls.

bidopsis. A major practical goal is to generate plants
with genetically defined variation in composition and
architecture to permit assessment of modifications on
wall properties and plant development. Several of
these mutants may be useful to determine the role
of the apoplasm in regulating inorganic nutrient up-
take. As the mutations we identify by spectroscopy
are characterized, the plant biology community will
be informed of them through a web site, and a sys-
tem will be devised to provide seeds and clones to the
community through established centers.
Acknowledgements
Characterization of the mur2 and mur3 mutants is
supported by a grant from the U.S. Department of
Agriculture-National Research Initiative Competitive
Grants Program (to N.C.C.) and the joint program to
develop the use of FTIR spectroscopy to identify cell-
wall mutants is supported by grants from the U.S.
National Science Foundation Genome Research Pro-
gram (to N.C.C.) and the U.K. Biotechnological and
Biological Science Research Council (to M.C.M.).
M.C.M. is grateful for a University Research Fellow-
ship fom the Royal Society. Journal paper No. 16,
xxx of the Purdue University Agricultural Experiment
Station.
References
Akarnatsu T, Hanzawa Y, Ohtake Y, Takahashi T, Nishitani K and
Komeda Y 1999 Expression of endoxyloglucan transferase genes
in acaulis mutants of Arabidopsis. Plant Physiol. 121,715-721.
Arabidopsis Genome Initiative 2000 Analysis of the genome se-
quence of the flowering plant Arabidopsis thaliana. Nature 408,
796-815.
Arioli T, Peng L C, Betzner A S, Burn J, Wittke W et al. 1998 Mo-
lecular analysis of cellulose biosynthesis in Arabidopsis. Science
279,717-720.
Baskin T I, Betzner A S, Hoggart R, Cork A and Williamson, R E
1992 Root morphology mutants in Arabidopsis thaliana. Aust. J.
Plant Physiol. 19,427-437.
Benfey P N, Linstead P J, Roberts R, Schiefelbein J W, Hauser M
T and Aeschbacher R A 1993 Root development in Arabidop-
sis - four mutants with dramatically altered root morphogenesis.
Development 119,57-70.
Bonin C P, Potter I, Vanzin G F and Reiter W-D 1997 The MURI
gene of Arabidopsis thaliana encodes an isoform of GDP-D-
mannose-4,6-dehydratase, catalyzing the first step in the de novo

synthesis of GDP-L-fucose. Proc. Natl. Acad. Sci. U.S.A. 94,
2085-2090.
Brummell D A, Catahi C, Lashbrook, C C and Bennett A B 1997
A membrane-anchored E-type endo-l,4-fi-glucanase is localized
on Golgi and plasma membranes of higher plants. Proc. Natl.
Acad. Sci. U.S.A. 94, 4794-4799.
Burget E G and Reiter W-D 1999 The mur4 mutant of Arabidopsis is
partially defective in the de novo synthesis of uridine diphospho
L-arabinose. Plant Physiol. 121, 383-389.
Carpita N C 1996 Structure and biogenesis of the cell walls of
grasses. Annu. Rev. Plant Physiol. Plant Mol. BioI. 47,445-476.
Carpita N C and Gibeaut D M 1993 Structural models of the primary
cell walls in flowering plants: consistency of molecular structure
with the physical properties of the walls during growth. Plant J.
3,1-30.
Carpita N and Vergara C 1998 A recipe for cellulose. Science 279,
672-673.
Chen L-M, Carpita N C, Reiter W-D, Wilson R H, Jeffries C and
McCann M C 1998 A rapid method to screen for cell wall
mutants using discriminant analysis of Fourier transform infrared
°
spectra. Plant J. 16, 385-392.
Dolezal and Cobbett C S 1991 Arabinose kinase-deficient mutant
of Arabidopsis thaliana. Plant Physiol. 96, 1255-1260.
Doong R L and Mohnen D 1998 Solubilization and characteriz-
ation of a galacturonosyltransferase that synthesizes the pectic
polysaccharide homogalacturonan. Plant J. 13, 363-374.
Ford C W and Hartley R D 1989 GCIMS characterization of cyclodi-
mers from p-coumaric and ferulic acids by photodimerization - a
possible factor influencing cell wall biodegradability. J. Sci. Food
Agric. 46, 310-310.
Fleischer A, Titel C and Ehwald R 1998 The boron requirement
and cell wall properties of growing and stationary suspension-
cultured Chenopodium album L. cells. Plant Physiol. 117, 1401-
1410.
Fleming A J, McQueen Mason S, Mandel T, Kuhlemeier C 1997
Induction of leaf primordia by the cell wall protein expansion.
Science 276, 1415-1418.
Fry S C, York W S, Albersheim P, Darvill A, Hayashi T et
al. 1993 An unambiguous nomenclature for xyloglucan-derived
oligosaccharides. Physiol. Plant. 89, 1-3.
Gibeaut D M and Carpita N C 1994 Biosynthesis of plant cell wall
polysaccharides. FASEB J. 8, 904--915.
Iiyama K, Lam T B T and Stone B A 1994 Covalent cross-links in
the cell wall. Plant Physiol. 104,315-320.
Ishii T, Matsunaga T and Hayashi N 2001 Formation of rham-
nogalacturonan II-borate dimer in pectin determines cell wall
thickness of pumpkin tissue. Plant Physiol. 126, 1698-1705.
Kemsley E K 1998 Discriminant Analysis of Spectroscopic Data.
John Wiley and Sons, Chichester, U.K. 179 pp.
Levy S, York W S, Stuikeprill R, Meyer Band Staehelin LA 1991
Simulations of the static and dynamic molecular conformations
of xyloglucan - the role of the fucosylated side-chain in surface-
specific side-chain folding. Plant J. 1, 195-215.
Lolle S J, Berlyn G P, Engstrom E M, Krolikowski K M, Reiter W-D
and Pruitt R E 1997 Developmental regulation of cell interactions
in the Arabidopsis fiddleheadl mutant: A role for the epidermal
cell wall and cuticle. Devel. BioI. 189,311-321.
Lukowitz W, Mayer U and Jurgens G 1996 Cytokinesis in the
Arabidopsis embryo involves the syntaxin-related KNOLLE gene
product. Cell 84, 61-71.
McCann M C and Roberts K 1991 Architecture of the primary cell
wall. In The Cytoskeletal Basis of Plant Growth and Form. Ed.
C W Lloyd. pp. 109-129. Academic Press, London.
79
McCann M C, Hammouri M, Wilson R, Belton P and Roberts K
1992 Fourier transform infrared microspectroscopy is a new way
to look at plant cell walls. Plant Physiol. 100, 1940--1947.
McCann M C, Chen L, Roberts K, Kemsley E K, Sene C, Carpita
N C, Stacey N J and Wilson R H 1997 Infrared microspectro-
scopy: sampling heterogeneity in plant cell wall composition and
architecture. Physiol. Plant. 100, 729-738.
Nicol F, His I, Jauneau A, Vernhettes S, Canut H and H6fte H 1998
A plasma membrane-bound putative endo-l,4-fi-D-glucanase
is required for normal wall assembly and cell elongation in
Arabidopsis. EMBO J. 17,5563-5576.
Nishitani K 1995 Endo-xyloglucan transferase, a new class of
transferase involved in cell wall construction. J. Plant Res. 108,
137-148.
Pear J R, Kawagoe Y, Schreckengost W E, Delmer D P, Stalker D M
1996 Higher plants contain homologs of the bacterial celA genes
encoding the catalytic subunit of cellulose synthase. Proc. Natl.
Acad. Sci. U.S.A. 93, 12637-12642.
Potikha T and Delmer D P 1995 A mutant of Arabidopsis thali-
ana displaying altered patterns of cellulose deposition. Plant J.
7,453-460.
Ralph J, Grabber J G and Hatfield R D 1995 Lignin-ferulate cross-
links in grasses - active incorporation of ferulate polysaccharide
esters into ryegrass lignins. Carbohydr. Res. 275,167-178.
Reiter W-D 1998 Arabidopsis thaliana as a model system to study
synthesis, structure and function of the plant cell wall. Plant
Physiol. Biochem. 36, 167-176.
Reiter W-D, Chapple C C S and Somerville C R 1993 Altered
growth and cell walls in a fucose-deficient mutant of Arabidop-
sis. Science 261, 1032-1035.
Reiter W-D, Chapple C and Somerville C R 1997 Mutants of
Arabidopsis thaliana with altered cell wall polysaccharide com-
position. Plant J. 12,335-345.
Reiter W-D and Vanzin G F 2001 Molecular genetics of nucleotide
sugar interconversion pathways in plants. Plant Mol. BioI. 47,
95-113.
Reynolds J 0, Eisses J F and Sylvester A W 1998 Balancing division
and expansion during maize leaf morphogenesis: analysis of the
mutant, warty1. Development 125,259-268.
Richmond T A and Somerville C R 2001 Integrative approaches to
determining Csi function. Plant Mol. BioI. 47, 131-143.
Rudall P J and Caddick L R 1994 Investigation of the presence of
phenolic compounds in monocotyledonous cell walls using UV
fluorescence microscopy. Ann. Bot. 74,483-491.
Sene C F B, McCann M C, Wilson R H and Grinter R 1994
Fourier-transform Raman and Fourier-transform infrared spec-
troscopy - an investigation of five higher plant cell walls and
their components. Plant Physiol. 106, 1623-163.
Smith B G and Harris P J 1999 The polysaccharide composition of
Poales cell walls: Poaceae cell walls are not unique. Biochem.
System Ecol. 27, 33-53.
Smith L G, Hake S and Sylvester AW 1996 The tangledl mutation
alters cell division orientations throughout maize leaf develop-
ment without altering leaf shape. Development 122, 481-489.
Staehelin L A and Moore I 1995 The plant Golgi apparatus - struc-
ture, functional organization and trafficking mechanisms. Annu.
Rev. Plant Physiol. Plant Mol. BioI. 46, 261-288.
Taylor N G, Scheible W, Cutler S, Somerville C R and Turner S R
1999 The IRREGULAR XYLEM3 locus of Arabidopsis encodes
a cellulose synthase required for secondary cell wall synthesis.
Plant Cell 11, 769-779.
Taylor N G, Laurie S and Turner S R 2000 Multiple cellulose syn-
thase catalytic subunits are required for cellulose synthesis in
Arabidopsis. Plant Cell 12,2529-2539.
80
Thompson J E, Smith R C and Fry S C 1997 Xyloglucan undergoes
interpolymeric transglycosylation during binding to the plant cell
wall in vivo: Evidence from C-13/H -3 dual labelling and isopyc-
nic centrifugation in caesium trifiuoroacetate. Biochem J. 327,
699-708.
Turner S R and Somerville C R 1997 Collapsed xylem phenotype of
Arabidopsis identifies mutants deficient in cellulose deposition
in the secondary cell wall. Plant Cell 9, 689-701.
Vanzin G F, Madson M, Carpita N C, Raikhel N V, Keegstra K,
Reiter W D 2002 The mur2 mutant of Arabidopsis thaliana lacks
fucosylated xyloglucan because of a lesion in fucosyltransferase
AtFUTI. Proc. Natl. Acad. Sci. USA 99,3340-3345.
Verica J A and Medford J 11997 Modified MERI5 expression alters
cell expansion in transgenic Arabidopsis plants. Plant Sci. 125,
201-210.
Willats W G T, McCartney L, Mackie Wand Knox J P 2001 Pec-
tin: cell biology and prospects for functional analysis. Plant Mol.
BioI. 47, 9-27.
Xu W, Purugganan M M, Polisensky D H, Antosiewicz D M,
Fry S C and Braam J 1995 Arabidopsis TCH4, regulated by
hormones and the environment, encodes a xyloglucan endo-
transglycosylase. Plant Cell 7, 1555-1567.
Zablackis E, Huang J, Muller B, Darvill A G and Albersheim P 1995
Structure of plant cell walls. 34. Characterization of the cell-wall
polysaccharides of Arabidopsis thaliana leaves. Plant Physiol.
107, 1129-1138.
Zablackis E, York W S, Pauly M, Hantus S, Reiter W-D, Chapple C
C S, Albersheim P and Darvill A 1996 Substitution of L-fucose
by L-galactose in cell walls of Arabidopsis murl. Science 272,
1808-1810.
Mineral nutrition: plant quality and plant health
 Plant and Soil 247: 83-90,2002.
© 2002 Kluwer Academic Publishers.
83

The impact of mineral nutrients in food crops on global human health

Ross. M. Welch
USDA, ARS, U.S. Plant, Soil and Nutrition Laboratory, Cornell University, Tower Road, Ithaca, NY 14853, USA
(E-mail: rmw1@cornell.edu)

Key words: food quality, food systems, human nutrition, global nutrition, malnutrition, micronutrients, mineral
nutrition, nutrient deficiencies, plant nutrition

Abstract
Nutrient sufficiency is the basis of good health, productive lives and longevity for everyone. Nutrient availability to
people is primarily determined by the output of foods produced from agricultural systems. If agricultural systems
fail to provide enough food diversity and quantity to satisfy all the nutrients essential to human life, people will
suffer, societies will deteriorate and national development efforts will stagnate. Importantly, plant foods provide
most of the nutrients that feed the developing world. Unfortunately, as a result of population pressures, many global
food systems are not currently providing enough micronutrients to assure adequate micronutrient intakes for all
people. This has resulted in an increasing prevalence of micronutrient deficiencies (e.g., iron deficiency, vitamin
A deficiency, and iodine deficiency disorders) that now afflicts over three billion people globally mostly among
resource-poor women, infants and children in developing countries. The consequences of micronutrient malnutri-
tion are profound and alarming for human existence. Agricultural approaches to finding sustainable solutions to
this problem are urgently needed. This review presents some ways in which plant nutritionists can contribute to
preventing micronutrient malnutrition in sustainable ways.

Introduction • Vitamin A deficiency can lead to poor night vis-

Humans are dependent on consuming enough diverse
foods to provide all the required nutrients to sus-
tain life. If food systems do not provide sufficient
quantities and enough diversity of foods to meet these
needs continuously, malnutrition will ensue among
certain population groups, especially the poor, and
their health and welfare will deteriorate (Welch et
aI., 1997). Starvation and regional famines are the
commonly recognized results of severe calorie/protein
malnutrition, and preventing these outcomes was the
paramount goal driving the agricultural 'green revolu-
tion' during the latter half of the 20th century. How-
ever, other more subtle consequences of malnutrition
include: impaired immune function, increased mor-
tality and morbidity rates, lower worker productivity,
diminished intellectual performance, less educational
attainment, a lower livelihood, higher birth rates and a
lower standard of living for all those affected (Welch
and Graham, 1999).
Some consequences of micronutrient deficiences
ion, eye lesions and, in severe cases, perman-
ent blindness; increased illness and death from
infections.
• Iodine deficiency can cause goiter, mental retarda-
tion, brain damage and reproductive failure.
• Iron deficiency can cause nutritional anemia, prob-
lem pregnancies, stunted growth, lower resist-
ance to infections, long-term impairment in mental
function, decreased productivity and food-energy
conversion and impaired neural motor develop-
ment.
• Zinc deficiency can cause growth retardation,
delayed skeletal and sexual maturity, dermatitis,
diarrhoea, alopecia and defects in immune func-
tion with resulting increase in susceptibility to
infection.
Because agriculture is the primary source of all
nutrients required for human life, those national agri-
cultural systems that do not provide sufficient nutrient
output to meet these nutritional needs will ultimately
fail, as well the food systems dependent on them. Pos-
84

High Risk Areas

Iron deficient
• Warrin A & Iron def"lCient
• Iodine, Warrin A & Iron deficient

Figure 1. Global distribution of Fe, vitamin A and I deficiencies (map modified from Sanghvi, 1996).

,. -____________________________
~ 2~
sibly, changes in cropping systems over the past 60 '"'"";' I!II!I Cereal Productloo ;
years that resulted during the 'green revolution' are ...
...~ Iill1I Pulse ProductiOn .;,
200 CJ Population 200 ~
contributing to the dysfunction of food systems that
can not meet all the nutritional needs of billions of I. 1~ 1~~
E
e
people in various developing world regions (Graham c c
and Welch, 2000; Graham et aI., 2001).
""
U
::J
i:;

In retrospect, the 'green revolution' resulted in

~
Q.
100 - 100 I
many positive outcomes related to human nutrition. ..••
.5

50 50 ..•..
.5

For example, between 1960 and 1990 globally: ~

u ~
u
.s .5
• rice production more than doubled :F- ;F.

• per capita food availability increased by 37%
• per capita calories available per day increased by
35%
• real food prices declined by 50%
However, there were some unforseen con-
sequences of this agricultural revolution that have had
profound affects on human health, felicity and world
development.
Today, there are over 3.7 billion iron-deficient in-
dividuals and about 1 billion people that are or are
at risk of developing iodine deficiency disorders (see
Figure 1 for the global extent of problem). Addition-
ally, there are over 200 million people that are vitamin
A deficient (World Health Organization, 1999). Other
micronutrient deficiencies (e.g., Zn, Se, vitamin C,
vitamin D, and folic acid deficiencies) may be as wide
spread as iron, iodine and vitamin A deficiencies, but
there are no reliable data to confirm this although
circumstantial evidence suggests that this may be so
(Combs et aI., 1996; World Health Organization,
1999).
'".. n
0 0
.!I
~ i ~
r
1;]
~ i~
.!:
Figure 2. Trends in cereal and pulse production in some develop-
ing countries and in populations between 1965 to 1999 (data from
FAOSTAT 1999).
What changes in agricultural systems might be
contributing to the rise in micronutrient malnutrition
globally? Figure 2 shows the trends in cereal and
pulse production and in human populations for certain
developing regions and nations from 1965 to 1999.
Clearly, cereal production has exceeded population
growth in these countries and regions. However, pulse
production has not kept pace with population growth
during this period. These crop production trends show
that while the agricultural systems in these countries
and regions were able to produce adequate quantities
of cereal grains to supply enough energy and pro-
tein to prevent widespread famines, they have not
been able to sustain the production of pulses. This
 85

"'.£
c
u
1996
Table 2. Effect of milling and polishing on micronutrient con-
centrations in rice grain (data from Salunkhe and Deshpande,
1991)
~
E
oo Micronutrient] Whole Polished %
o brown rice Reduction
oo
!:. rice
c
o Iron (mglkg) 30 10 67
~

e
:>
Copper (mglkg) 3.3 2.9 12
Q. Manganese (mg/kg) 17.6 10.9 62
~ 100 Zinc (mg/kg) 18 13 30
o
(5 Thiamin (mglkg) 3.4 0.7 80
Riboflavin (mglkg) 0.5 0.3 40
Niacin (mglkg) 47 16 66
Vitamin B6 (mg/kg) 6.2 0.4 94
Folic Acid (J.Lglkg) 200 160 20
Figure 3. Global production of food crops in 1996 (data from Mann,
1997). Pantothenic Acid (mglkg) 20 10 50
Biotin (J.Lglkg) 120 50 58
Vitamin E (IU/kg) 20 10 50
Table 1. Concentrations of Fe and Zn in
whole cereal grains and pulse seeds
] Dry weight basis.

Plant food Fe Zn
(J.Lg g I , dry weight)
Many food systems which sustain life in a num-
Brown rice] 4 15 ber of countries today cannot produce enough of all
Whole maize] 21 22 the nutrients needed to satisfy human requirements.
Whole wheat2 37 31 For example, Figure 4 shows the nutrient output of
Mung bean 3 87 41 two food systems (a banana-based and a grain-based
Blackgram3 139 36 system) in Uganda that feed numerous individuals in
Cowpea3 67 45 that country (McIntyre et aI., 2001). Figure 4 clearly
Soybean4 97 43 show that the banana-based system does not provide
Phaseolus bean5 65 37 enough vitamin A, Zn and Ca, and the grain-based
] Data from Wolnik et aI., 1985.
system does not provide enough Zn and Ca to meet
2Data from Wolnik et aI., 1983. the Recommended Dietary Allowances (RDAs) for
3Unpublished data provided by R. M. Welch. these nutrients established by the National Research
4Data from Holland et a!., 1991. Council, NAS, USA. If one uses the Recommended
5Data from Peck et aI., 1980.
Nutrient Intake (RNI) value of 59 mg day-l for Fe
(i.e., for adult females between the ages of 19 and 50
preliminarily recommended by the joint FAOIWHO
has affected the supply of available micronutrients Expert Consultation (FAOIWHO, 2000)) instead of the
produced in these nations because pulses are much USA's RDA for Fe of 15 mg day-l then neither food
richer sources of micronutrients (Table 1) compared system could supply enough Fe to meet the needs of
to cereals which are usually milled and processed be- people dependent on these systems for nourishment.
fore consumption, further reducing their micronutrient These types of studies clearly indicate that many food
content (Table 2) (Salunkhe and Deshpande, 1991). systems are failing to provide enough nutrients to sus-
Furthermore, micronutrient-rich fruit and vegetable tain healthy, active and productive lives for all citizens
production has not kept pace with population growth in many developing nations (Welch et aI., 1997).
in many nations (Welch and Graham, 1999). Finally,
the decrease in food crop diversity brought about dur-
ing the 'green revolution's' drive to improve cereal Improving plants as sources of micronutrients
production and expand cereal crop acreage most likely
is also a contributing factor (see Figure 3) (Graham et What can plant nutritionists do to enhance the nu-
al.,2001). tritional quality of plant foods? Cultural practices
86
Table 3. Effects of increasing Zn and Se supplies to wheat plants grown in nutrient solutions on the concentration of Zn and
Se in mature wheat grain and bioavailable amounts of Zn and Se in the grain when fed to Zn-depleted rats in a single meal
(data from House and Welch, 1989)

Zn Se Znin Sein Bioavailable Bioavailable

supplied supplied grain grain Zn Se
(JLM) (JLM) (JLg g-l, dry wt.) (JLg g-l, dry wt.) (JLg absorbed from meal) (JLg absorbed from meal)

1.0 0.3 8.8 0.6 5.9 0.41

1.0 1.5 9.3 3.8 5.5 2.03
5.0 0.3 33.0 0.6 18.7 0.42
5.0 1.5 36.1 4.3 15.4 1.91

200~--------------.-~~~~~----~
Banana-based food system Grain-based food system cussion briefly presents some examples of how some
~ 180 cultural and agronomic practices could be used to en-
160 hance the micronutrient output garnered from farming
C3
a: 140
.s., systems .
120
j 100 --- Cultural practices
"5 80
Co
"5
0 60
Both macronutrient fertilizers containing N, P, K, and
1: ., S, and certain micronutrient fertilizers containing, for
.""5 40

I.
example Zn, Ni, and Se, can have significant ef-
~.
z 20
fects on the accumulation of micronutrients in edible
0
« ., «c: ., plant products (Allaway, 1986; Grunes and Allaway,
.,c: ~" ·ec:
c: c:
.,c:
>- >- c:
2' N u.. (J 2' ]j N u.. (J

Q.
w
e ·e
W .l!l Q.
5 ~
Figure 4. Output of nutrients (i.e., energy, protein, vitamin A,
Zn, Fe and Ca) for two Ugandan food systems (banana-based and
grain-based) relative to the recommended daily allowance (RDA)
expressed as a percent of required nutrient intake (i.e., percent of
RDA). Values were calculated on a consumption unit basis (data
from McIntyre et al., 2001).
used by agriculturalists can affect the output of avail-
able nutrients from agricultural systems (Salunkhe an.d
Deshpande, 1991; Welch, 1995). However, current
agricultural practices are almost always directed at
maximizing production while minimizing costs. Re-
cently, in some nations, preserving the environment
is becoming a more important objective of agriculture
(i.e., 'sustainable' agricultural goals). Maximizing nu-
trient output of farming systems has never been a
purport of either agriCUlture or of public policy. Yet,
scientific knowledge is available that could greatly im-
prove the micronutrients output of farming systems,
and the available micronutrient content of the food
crops produced. The debilitating effects ofmicronutri-
ent malnutrition on people and societies and its current
magnitude in developing nations certainly testifies to
the need to consider doing so now. The following dis-
1985). Other micronutrient fertilizers have very little
if any effect on the amount of the micronutrient ac-
cumulated in edible seeds and grains when they are
applied to soils or when used as foliar sprays (Welch,
1986). This is especially true for those micronutri-
ent elements with limited phloem sap mobility such
as Fe, B, V and Cr. Some examples of the effects
of fertilizer practices on the micronutrient concentra-
tions in edible plant parts are given below. For more
detailed information concerning the effects of fertiliz-
ation practices on micronutrient accumulation in plant
foods refer to Allaway (1975), Grunes and Allaway
(1985), Karmas and Harris (1988), Nagy and War-
dowski (1988), Salunkhe and Desai (1988) and Welch
(1997).
Excessive N fertilizers can adversely affect the ac-
cumulation of vitamin C in various vegetable crops
such as lettuce (Lactuca sativa L.), beets (Beta vul-
garis cicla L.), kale (Brassica oleracea acephala DC.),
endive (Cychorium endivia L.), and brussels-sprouts
(Brassica oleracea gemmifera DC.) by as much as
26% (cited in Salunkhe and Desai, 1988). However,
increasing the amount of K fertilizer supplied to these
crops significantly increased their vitamin C con-
tent by about 8-20% depending on the species. The
concentration of ,B-carotene in carrot (Ducus carota

subsp. carota sativus (Hoffm.) Arcang.) roots in-
creased at first harvest in response to increases in the N
supplied. ,8-carotene increased from 113 mg 100 g-l
root in those plants supplied 0.3 g N per pot to 126 mg
100 g-l root dry weight (about 12% increase) for
plants treated with 2.4 g N per pot (Habben, 1972 cited
in Salunkhe and Desai, 1988). By the third harvest the
increase in ,8-carotene level resulting from increasing
N supply was only about 7%, but the late harvest res-
ulted in an increase in the level of ,8-carotene even
in the lowest N treatment from 113 to 136 mg 100
g-l demonstrating a large effect of harvest date on
,8-carotene content of carrot.
Macronutrient treatments can influence the con-
centration of ,8-carotene and other micronutrients in
carrots (Welch, 1997). Vereecke, in 1979 (cited in
Salunkhe and Desai, 1988) reported results of studies
concerning the effects of combined N, P, K and Mg
fertilizers on ,8-carotene, Fe, Mn, Zn and Cu in carrot
root. Treatments containing N, P and Mg increased
the accumulation of ,8-carotene in carrot roots by 42%
compared to roots from unfertilized control plants.
Adding K to the combined N and P fertilizer treat-
ments increased the ,8-carotene by 27% over-fertilized
plants not receiving K. Removal of Mg from the fertil-
izer mix lowered the increase in ,8-carotene from 42 to
30%. Apparently, Mg was required for maximum ,8-
carotene production in carrot roots furnished adequate
N, P, and K. Vereecke also reported the effects of these
treatments on Fe, Mn, Zn and Cu in carrot leaves at
two harvest dates. Large effects were found on the Mn
and Fe content of the leaves, but the other micronutri-
ents determined were not greatly effected by fertilizer
treatments. Combined treatments with N, P, K and Mg
increased leaf Fe and Mn concentrations at a late har-
vest by 20% (from 194 to 234 tLg Fe g-l dry wt.) and
43% (56 to 136 tLg Mn g-l dry wt.), respectively.
The vitamin C concentration in fruits is also af-
fected by macronutrient fertilizers. As with vegetable
crops, excessive N fertilization has been reported to
reduce vitamin C concentration in the fruits of several
species including oranges, lemons, mandarins, canta-
loupe, and apple. Also, higher rates of K fertilization
are associated with greater concentration of vitamin C
in fruits (Nagy and Wardowski, 1988). Apparently, the
effects of Zn, Mg, Mn, and Cu fertilization on increas-
ing vitamin C concentration in citrus fruits is limited
to soils that are deficient in these elements. Supplying
more of these elements than is required for optimum
yield does not increase further vitamin C level in the
fruit.
87
High·Zn Supply
Seed-Zn Concentration ( Ilg gO') Bioavailable Zn ( Ilg)
Figure 5. Effects of increasing Zn supply to pea (Pisum sativum
L.) plants, grown in 65Zn radiolabeled nutrient solutions, on Zn
concentration (JLg g-l) and bioavailable Zn (JLg Zn g-dieC 1) in
mature pea seeds fed in single meals to Zn-depleted rats (data from
Welch et al., 1974).
For certain essential micronutrient elements (e.g.,
Zn, Ni, and Se), increasing their supply to food crops
can result in significant increases in their concentra-
tions in edible plant products. For example, increasing
the supply of Zn to pea plants (Pisum sativum L.) at
levels in excess of that required for maximum yield has
been shown to increase the concentration of bioavail-
able Zn in pea seeds (Peck et al., 1980; Welch et
al., 1974). Furthermore, increasing the supply of Zn
and Se to wheat, Triticum aestivum L. (see Figure 5)
(House and Welch, 1989) is also effective. This has
also been shown for navy beans (Phaseolus vulgaris
L.) as well as other crops (Moraghan, 1994; Peck et
al.,1980).
For iron, providing more to plants than required
to sustain growth does little to further increase the
Fe in edible seeds and grains (for example see Welch
and Van Campen, 1975). The accumulation of mi-
cronutrient elements in seeds and grains is controlled
by a number of processes including root-cell uptake,
root-shoot transfer, and the ability of leaf tissues to
load these nutrients into the vascular phloem ele-
ments which are ultimately responsible for delivering
these nutrients to developing seeds and grains via
the phloem sap (Welch, 1986). Phloem loading and
unloading of these nutrients is tightly controlled by
poorly understood hemostatic mechanisms in the plant
and further research should be carried out to under-
stand these processes if we are to significantly increase
certain micronutrient elements, such as Fe, in staple
seeds and grains (Welch, 1995).
88
Soil amendments are frequently used by farmers to
adjust soil pH and to enhance plant growth properties
of soils. Using lime (CaC03) raises soil pH, permit-
ting acid-intolerant legume species to grow in soils
that would otherwise be to acidic for their growth. It
is also used to supply Ca to plants. However, adding
lime may depress the uptake of Zn, Cu, Fe, and Co,
and increase the uptake of Se and Mo by plants. A
high soil-pH favors the oxidation of reduced forms of
Se such as Se- 2 and Se03 -2 to the more soluble and
plant-available Se04-2 anion. Gypsum (CaS04) and
elemental S are used to decrease the pH of alkaline
soils as well as to provide S for plant uptake and to
ameliorate high-Na alkali soils. Using gypsum on al-
kaline soils could increase plant-available Fe, Mn, Zn,
Cu, and Co by decreasing alkaline soil pH.
The use of farm-yard manures and other forms
of organic matter can also change plant-available mi-
cronutrients by changing both the physical and biolo-
gical characteristics of the soil. In many circumstances
these changes improve soil physical structure and wa-
ter holding capacity, resulting in more extensive root
development and enhanced soil microflora and fauna
activity, all of which can affect available micronutri-
ent levels in soil to plants (Stevenson, 1991, 1994).
However, very few controlled experiments have been
done to determine which types of organic matter prac-
tices significantly enhance or depress the levels of
micronutrients in edible portions of major food crops.
More research should be carried out to understand
the impact of various types of organic matter on crop
nutritional quality.
Genetic manipulation
Breeding for micronutrient-dense varieties of staple
foods is also a powerful tool to use in the fight
against micronutrient malnutrition. Recent findings
from an international project among several Con-
sultative Group on International Agricultural Research
(CGIAR) centers (including the International Food
Policy Research Institute (IFPRI), the International
Rice Research Institute (IRRI) , the Centro Interna-
tional de Agricultura Tropical (CIAT) and the Centro
International de Mejoramiento de Maiz y Trigo (CIM-
MYT)), the University of Adelaide, Waite Campus,
and the USDA, ARS U.S. Plant, Soil and Nutrition
Laboratory show that it is possible to breed for en-
hanced levels of iron, zinc and provitamin A caroten-
oids in edible portions of rice, wheat, maize, beans
and cassava (see Graham et aI., 2001, for an extensive
review of these findings). The findings show:
• there is enough genetic variation in concentrations
of these micronutrients among lines in the major
germplasm banks to justify selection
• micronutrient-density traits are stable across envir-
onments
• it is possible to combine micronutrient-rich traits
with high yield
• genetic control is fairly simple making breeding
for the traits economical
• several limiting micronutrients can be improved
simultaneously
• both seedling vigor and nutritional quality are im-
proved through genetically modifying seeds with
micronutrient enrichment traits
Combining both human nutrition with improved
agricultural productivity from such breeding efforts
results in extremely high costlbenefit ratios for invest-
ing in this type of micronutrient intervention (i.e., bet-
ter than 1:50). Furthermore, the adoption and spread
of micronutrient-enriched seeds by farmers can be
driven by profit incentives because micronutrient en-
riched seeds increase crop productivity when planted
to micronutrient-poor soils (Graham et aI., 2001). The
benefits can be disseminated widely and they are sus-
tainable once developed unlike current micronutrient
interventions that rely on supplements or food fortific-
ants (Graham et aI., 2001; Graham and Welch, 2000;
Welch et aI., 1997).
Importantly, not only will it be necessary to in-
crease the content of micronutrients in plant foods,
but also the absorption and utilization (i.e., bioavail-
ability) of micronutrients in meals containing plant
foods must be improved (Graham and Welch, 1996).
Plant foods can contain various substances, both an-
tinutrient and promoter substances, that interact with
micronutrients to either lower or enhance, respect-
ively, their bioavailability to humans (Graham et aI.,
2001). These substances can also be manipulated by
agricultural practices (Graham et aI., 2001) and by
genetic manipulation (Lucca et aI., 2001; Watson,
1995).
Modern recombinant DNA technology can be used
to enhance the nutritional quality of food crops such
as increasing the amount and bioavailability of mi-
cronutrients in plants (DellaPenna, 2001; Forssard et
aI., 2000; Goto et aI., 1999,2000; Kishore and Shew-
maker, 1999; Lucca et aI., 2000, 2001). 'Golden rice'
is such an example where recombinant DNA techno-
logy was used to improve the provitamin A content

of rice-grain endosperm (Ye et aI., 2000). Another
example was reported by Goto et al. (1999). They en-
riched rice-grain endosperm Fe by transforming rice
plants using a phytoferritin (a major protein storage
form of Fe in plants) gene from soybean and a rice en-
dosperm promoter gene. They were able to more then
double the rice-grain concentration of Fe (i.e., from
about 14 f.Lg Fe g-1 in the non-transformed rice to
about 37 f.Lg Fe g-1 in the transformed rice) using this
procedure. Lucca et ai. (2001) also reported enriching
the Fe concentration in rice-grain endosperm using the
phytoferritin gene from pea plants.
A recent human study supports the contention that
the Fe in soybean-phytoferritin is highly bioavailable
to humans. The bioavailable Fe in intrinsically ra-
diolabeled soybeans was determined in Fe-depleted
women. The absorption of the labeled Fe averaged
27% of the dose when the soybeans were fed either
as a soybean soup or as a soybean muffin in single
meals (personal communication from Elizabeth C.
Theil, Children's Hospital Oakland Research Institute,
Oakland, CA). This finding suggests that, possibly,
other staple food crops should be genetically modified
to accumulate more phytoferritin in their edible por-
tions. Doing so would give another powerful tool to
nutritionists to use in addressing iron deficiency in the
developing world.
Synergies exist between various micronutrients
and their bioavailability to humans from plant food
sources. For example, the bioavailability of Fe from
cereal grains to humans is enhanced by ,8-carotene
(Graham et aI., 2001). Apparently, this interaction
between Fe and ,8-carotene alleviates the negative
effects of the antinutrient, phytate, on reducing Fe ab-
sorption from the human gut. Because cereal grains
contain substantial amounts of phytate, modifying
cereal grains to contain greater levels of both Fe and,8-
carotene would result in greatly improved nutritional
value of cereal grains for humans. Modem transgenic
technologies provide new tools to do such transforma-
tions. However, ingesting ,8-carotene-rich plant foods
(such as carrots and sweet potatoes) with meals con-
taining cereal grains could produce similar synergies
and making more of these foods available in food
systems is also an important tool to use in reducing
micronutrient malnutrition.
89
Conclusions
Agriculture must change in ways that will closely link
food production to human health and nutritional re-
quirements. Holistic food system models hold promise
in providing sustainable intervention.s to these com-
plex nutrition and health problems (Combs et aI.,
1996). Plant nutritionists can and should play an im-
portant role in such efforts in the future. Sustainable
solutions to micronutrient malnutrition can only be
found in forming a nexus between agricultural pro-
duction and human health. Because the magnitude of
the problem is so great we must use every tool at our
disposal to eliminate this scourge from the world.
References
Allaway WH 1975 The effects of soils and fertilizers on human and
animal nutrition. USDA-ARS Agriculture Information Bulletin
No. 378. Washington, DC, U.S. Government Printing Office.
Allaway WH 1986 Soil-plant-animal and human intrrelationships in
trace element nutrition. In Trace Elements in Human and Animal
Nutrition. Ed. W. Mertz. pp. 465-488. Academic Press, Or-
lando, San Diego, New York, Austin, London, Montreal, Sydney,
Tokyo, Toronto.
Combs GF Jr., Welch RM, Duxbury JM, Uphoff NT and Nesheim
MC 1996 Food-Based Approaches to Preventing Micronutrient
Malnutrition: an International Research Agenda, Cornell In-
ternational Institute for Food, Agriculture, and Development,
Cornell University, Ithaca, NY pp. 1-68.
DellaPenna D 2001 Nutritional genomics: manipulating plant mi-
cronutrients to improve human health. Science 285, 375-379.
FAOIWHO 2000 Preliminary report on recommended nutrient in-
takes. Joint FAOIWHO Expert Consultation on Human Vitamin
and Mineral Requirements, FAO, Bangkok, Thailand, 21-30
September 1998, Food and Agricultural Organization of the
United Nations and the World Health Organization, Rome. pp.
1-16.
Frossard E, Bucher M, Machler F, Mozafar A and Hurrell R 2000
Review. Potential for increasing the content and bioavailability
of Fe, Zn and Ca in plants for human nutrition. J. Sci. Fd. Agric.
80,861-879.
Goto F, Yoshihara T, Shigemoto N, Toki S and Takaiwa F 1999
Iron fortification of rice seed by the soybean ferritin gene. Nat.
Biotechno!. 17, 282-286.
Goto F, Yoshihara T and Saiki H 2000 Iron accumulation and
enhanced growth in trasgenic lettuce plants expressing hte irojn-
binding protein ferritin. Theor. App!. Genet. 100, 658-664.
Graham RD and Welch RM 1996 Breeding for staple-food crops
with high micronutrient density. Agricultural Strategies for Mi-
cronutrients Working Paper 3, International Food Policy Re-
search Institute, Washington, DC. pp. 1-72.
Graham RD and Welch RM 2000 Plant food micronutrient compos-
ition and human nutrition. Commun. Soil Sci. Plant Anal. 31,
1627-1640.
Graham RD, Welch RM and Bouis HE 2001 Addressing micronu-
trient malnutrition through enhancing the nutritional quality of
staple foods: principles, perspectives and knowledge gaps. Adv.
Agron. 70, 77-142.
90
Grunes DL and Allaway W H 1985 Nutritional quality of plants in
relation to fertilizer use. In Fertilizer Technology and Use. Ed.
O. P. Engelstad. pp. 589--619. Soil Science Society of America,
Madison; WI.
Holland B, Welch AA, Unwin lD, Buss DH, Paul AA and Southg-
ate DAT 1991 McCance and Widdowson's The Composition of
Foods, The Royal Society of Chemistry, Letchworth, UK 462 pp.
House WA and Welch RM 1989 Bioavailability of and interactions
between zinc and selenium in rats fed wheat grain intrinsically
labeled with 65Zn and 75Se. J. Nutr. 119,916-921.
Karmas E and Harris R S 1988 Nutritional Evaluation of Food Pro-
cessing, Avi Book, Van Nostrand Reinhold Co., New York. 786
pp.
Kishore G and Shewmaker C 1999 Biotechnology: enhancing hu-
man nutrition in developing and developed worlds. Proc. Natl.
Acad. Sci. U.S.A. 96, 5968-5972.
Lucca P, Wunn J, Hurrell RF and Potrykus I 2000 Development of
iron-rich rice and improvement of its absorption in humans by
genetic engineering. J. Plant Nutr. 23,1983-1988.
Lucca P, Hurrell R and Potrykus 12001 Genetic engineering ap-
proaches to improve the bioavailability and the level of iron in
rice grains. Theor. Appl. Genet. 102,392-397.
Mann C 1997 Reseeding the green revolution. Science 277, 1038-
1043.
McIntyre BD, Bouldin DR, Urey GH and Kizito F 2001 Modeling
cropping strategies to improve human nutrition in Uganda. Agric.
Syst. 67, 105-120.
Moraghan JT 1994 Accumulation of zinc, phosphorus, and mag-
nesium by navy bean seed. J. Plant Nutr. 17, 1111-1125.
Nagy S and Wardowski WF 1988 Effects of agriCUltural practices,
handling, processing, and storage on fruits, In Nutritional Eval-
uation of Food Processing. Eds. E Karmas and RS Harris. pp.
73-100. Avi Book, Van Nostrand Reinhold Co., New York.
Peck N, Grunes DL, Welch RM and MacDonald GE 1980 Nutri-
tional quality of vegetable crops as affected by phosphorus and
zinc fertilizers. Agron. J. 72, 528-534.
Salunkhe DK and Desai BB 1988 Effects of agriCUltural practices,
handling, processing, and storage on vegetables. In Nutritional
Evaluation of Food Processing. Eds. E. Karmas and RS Harris.
pp. 23-71. Avi Book, Van Nostrand Reinhold Co., New York.
Salunkhe DK and Deshpande SS 1991 Foods of Plant Origin: Pro-
duction, Technology, and Human Nutrition, AVI Book, Van
Nostran Reinhold, New York 501 pp.
Sanghvi TG 1996 Economic rationale for investing in micronutri-
ent programs. A policy brief based on new analyses. Office of
Nutrition, Bureau for Research and Development, United States
Agency for International Development, Washington, DC.
Stevenson FJ 1991 Organic matter-micronutrient reactions in soil.
In Micronutrients in Agriculture. Eds. JJ Mortvedt et al. pp. 145-
186. Soil Sci. Soc. Am., Madison, WI.
Stevenson FJ 1994 Humus Chemistry: Genesis, Composition, Re-
actions. Wiley, New York. 496 pp.
Watson JM 1995 Improving the nutritional and functional qualities
of plant food through genetic engineering. Proc. Nutr. Soc. Aust.
19,73-79.
Welch RM 1986 Effects of nutrient deficiencies on seed production
and quality. Adv. Plant Nutr. 2,205-247.
Welch RM 1995 Micronutrient nutrition of plants. Crit. Rev. Plant
Sci. 14,49-82. Welch, R M 1997 Agronomic problems related
to provitamin A carotenoid-rich plants. Eur. J. Clin. Nutr. 51,
S34-S38.
Welch RM and Graham RD 1999 A new paradigm for world
agriCUlture: meeting human needs - Productive, sustainable,
nutritious. Field Crops Res. 60, 1-10.
Welch RM and Van Campen DR 1975 Iron availability to rats from
soybeans. J. Nutr. 105,253-256.
Welch RM, House WA and Allaway WH 1974 Availability of zinc
from pea seeds to rats. J Nutr. 104, 733-740.
Welch RM, Combs GF Jr and Duxbury JM 1997 Toward a Greener
revolution. Issues Sci. Tech. 14, 50-58.
Wolnik KA, Frick FL, Capar SG, Braude GL, Meyer MW, Satzger
RD and Kuennen RW 1983 Elements in major raw agriCUltural
crops in the United States. 2. Other elements in lettuce, peanuts,
potatoes, soybeans, sweet corn, and wheat. J. Agric. Food Chern.
31, 1244-1249.
Wolnik KA, Frick FL, Capar SG, Braude GL, Meyer MW, Satzger
RD, Bonnie E and Gaston CM 1985 Elements in major raw agri-
cultural crops in the United States. 3. Cadmium, lead, and eleven
other elements in carrots, field corn, onions, rice, spinach, and
tomatoes. J. Agric. Food Chern. 33, 807-811.
World Health Organization 1999 Malnutrition Worldwide. http:
www.who.intlnutlmalnutrition_worldwide.htrn, 1-13. World
Health Organization, Geneva.
Ye X, Al-Babili S, Kliili A, Zhang J, Lucca P, Beyer P and Potrykus
I 2000 Engineering the provitamin A (b-carotene) biosynthetic
pathway into (carotenoid-free) rice endosperm. Science 287,
303-305.
Salinity and plant - soil - water relations
 Plant and Soil 247: 93-105,2002.
93
© 2002 Kluwer Academic Publishers.

Avenues for increasing salt tolerance of crops, and the role of

physiologically based selection traits

Rana Munns 1,5 , Shazia Husainl, Anna Rita RiveUi 2 , Richard A. James l , A.G. (Tony) Condon l ,
Megan P. Lindsay 1, Evans S. Lagudah 1, Daniel P. Schachtman3 & Ray A. Hare4
lCSIRO Plant Industry, GPO Box 1600, Canberra ACT 2601, Australia. 2Department of Plant Production, Uni-
versity of Basilicata, Potenza, Italy. 3 Danforth Plant Science Center, Campus Box 1098, St. Louis, MO 63105,
USA. 4 NSW Agriculture, RMB 944, Tamworth NSW 2340, Australia. 5 Corresponding author*

Key words: durum wheat, molecular markers, salinity, sodium

Abstract
Increased salt tolerance is needed for crops grown in areas at risk of salinisation. This requires new genetic sources
of salt tolerance, and more efficient techniques for identifying salt-tolerant germplasm, so that new genes for
tolerance can be introduced into crop cultivars. Screening a large number of genotypes for salt tolerance is not
easy. Salt tolerance is achieved through the control of salt movement into and through the plant, and salt-specific
effects on growth are seen only after long periods of time. Early effects on growth and metabolism are likely due
to osmotic effects of the salt, that is to the salt in the soil solution. To avoid the necessity of growing plants for
long periods of time to measure biomass or yield, practical selection techniques can be based on physiological
traits. We illustrate this with current work on durum wheat, on selection for the trait of sodium exclusion. We have
explored a wide range of genetic diversity, identified a new source of sodium exclusion, confirmed that the trait has
a high heritability, checked for possible penalties associated with the trait, and are currently developing molecular
markers. This illustrates the potential for marker-assisted selection based on sound physiological principles in
producing salt-tolerant crop cultivars.

The problem
About 7% of the world's total land area is affected
by salt, as is a similar percentage of its arable land
(Ghassemi et aI., 1995; Szabo1cs, 1994). The area is
still increasing as a result of irrigation or land clearing
(FAOSTAT).
The major salinity problem in Australia is 'dryland
salinity' , i.e., it results from land clearing. It is due to
rising water tables resulting from clearing the original
native vegetation, which consisted largely of perennial
species. Water tables rise because annual crops allow
more rainwater to escape their root systems than the
original perennial vegetation. In the Australian wheat
belt, an average of about 30 mm of water is added to
the groundwater each year (Dunin et aI., 2001). This
raises the water table by about half a metre each year,
* FAX No: 61 262465399.
E-mail: rana.munns@csiro.au
carrying salts that have accumulated in the soil. When
the water table eventually reaches the surface the water
evaporates, leaving any salt behind. In Australia, this
process may ultimately affect a very large proportion
of the area that has been cleared for farming. Predic-
tions by a recent national land and water resources
audit (ANRA, 2001) indicate that by the year 2050, as
many as 17 million ha will be salinised, or be at risk of
salinisation. This area represents a third of Australia's
agricultural land area.
Irrigation systems are particularly prone to salin-
isation, with about half the existing irrigation systems
of the world now under the influence of salinisation or
waterlogging, due to either low quality irrigation wa-
ter, or to excessive leaching and subsequent rising wa-
ter tables (Szabo1cs, 1994). Irrigation schemes cover
only 15% of the cultivated land of the world, but as
irrigated land has at least twice the productivity of
rainfed land, it may produce one third of the world's
food.
94
The solution
Salinisation can be restricted by changed farm man-
agement practices. Irrigated agriculture can be sus-
tained by better irrigation practices such as adoption
of partial root zone drying methodology, and drip
or micro-jet irrigation to optimise use of water. The
spread of dryland salinity can be contained by redu-
cing the amount of water passing beyond the roots.
This can be done by re-introducing deep-rooted per-
ennial plants, that continue to grow and use water
during the seasons that do not support annual crop
plants. This may restore the balance between rainfall
and water use, thus preventing rising water tables and
the movement of salt to the soil surface.
Farming systems can change to incorporate peren-
nials in rotation with annual crops (phase farming),
in mixed plantings (alley farming, intercropping), or
in site-specific plantings (precision farming). Preci-
sion farming can identify areas giving consistently
poor yield, and these can be excluded from cropping,
as such areas are especially prone to 'leakage' (i.e.,
allowing rainwater to escape below the roots). In pre-
cision farming, areas of high production can also be
identified, and these sites can be planted with cultivars
of high vigour that use water effectively during the
growing season, and consume most of the available
soil water. Phase farming, in which several years of
pasture are rotated with several years of crop, can
make use of deep-rooted pasture plants to dry the
deep subsoil, thereby creating a buffer zone to hold
any water that escapes the crops. Trials in Australia
have shown that the deep-rooted perennial lucerne
(Medicago alfalfa) can lower the water table suffi-
ciently to allow subsequent cropping (Ridley et aI.,
2001). Such practices will rely on plants that have a
high degree of salt tolerance. Tolerance will be re-
quired for the 'de-watering' species, but also for the
annual crops to follow, as salt will be left in the soil
when the water table is lowered. Salt tolerance in crops
will also allow the more effective use of poor quality
irrigation water.
Diversity in salt tolerance between species
Salt tolerance is usually assessed as the percent bio-
mass production in saline versus control conditions
over a prolonged period of time. Figure 1 shows the ef-
fect of salinity on lupin (one of the most salt-sensitive
crop species), barley (one of the most tolerant), as well
as two halophytes that are useful forage in salt-affected
soils. The data shown in the figure are for plant dry
weight after a period of about a month in a range of
salinities. The data are from experiments in which the
salinity increased after plants were established in non-
saline conditions, not for experiments when salt was
added at germination. The figure indicates that in a
field where the salinity rises to 100 mM NaCl (about
10 dS m -1), lupins, and in fact most legume species,
will die before maturity, while crops such as wheat and
barley will produce a reduced yield. Even barley dies
at salt concentrations higher than 250 mM NaCI (about
25 dS m- 1, or 50% seawater). Only halophytes will
cope with soils where the watertable has brought salt
to the surface, as the water in the topsoil will contain
salts at concentrations higher than seawater.
The effects of salinity on barley and lupin probably
span the extremes of salt tolerance of crops. Wheat
(Triticum aestivum) is usually considered less tolerant
than barley, but there is such difference between geno-
types that it is difficult to make a categorical statement.
Bean (Phaseolus vulgaris) is one of the most salt-
sensitive species, but for this species, like so many,
the supply of additional Ca2+ is crucial for the salt
tolerance (Lahaye and Epstein, 1971), and again it is
difficult to generalise. Rice (Oryza sativa) is regarded
as one of the more salt-sensitive crops, which is cer-
tainly true when grain yield is considered (Khatun et
aI., 1995; Maas and Hoffman, 1977). However, veget-
ative growth of some rice cultivars can be surprisingly
tolerant of salinity, at least when adequate Ca2+ is
supplied (Muhammed et aI., 1987).
Another criterion of salt tolerance of crops is their
yield in saline versus non-saline conditions. A survey
of salt tolerance of crops, vegetables and fruit trees
has been published by Maas and Hoffman (1977), and
updated by Francois and Maas (1994). They show for
each species a threshold salinity below which there is
no reduction in yield, and then a regression for the
reduction in yield with increasing salinity. The data in
some cases are for a single cultivar of the species, or a
limited number of cultivars at a single site, so they are
not necessarily representative of the species. However,
the data are useful in that they show the wide range
of tolerance across species, and also show that yield
has a different pattern of response than does vegetative
biomass. Yield always shows a threshold in response
to a range of salinities (Maas and Hoffman, 1977),
but with young plants a threshold in growth is rarely
seen. With plants exposed to salinity at an early stage
of seedling development there are linear reductions in
 95

125

100

.--..
(5
.....
"-
c 75
0
u

---..ccf2......
$: 50
0
"-
(!J

25

Kaller grass Saltbush

Barley
O~-'-----------'-----------.-----------r----------,
o 200 400 600 800
NaCI concentration in soil (mM)
Figure 1. Growth of four diverse plant species to a range of salinity levels. The saltbush species is Atriplex amnicola, a halophyte of Western
Australia (data from Aslam et ai., 1986). Kallar grass, Diplachne (syn.Leptochloa) Jusca, is widespread in many continents including Australia
(data from B. Myers and D. West, pers. commun.). Barley (Hordeum vulgare) is one of the most salt-tolerant crops (data from Greenway, 1962;
Rawson et ai., 1988) and white lupin (Lupinus albus) is one of the most sensitive (data fromleschke et ai., 1986).

both leaf area expansion and total plant biomass with
increasing salinity, as shown in Figure 1.
There is probably a great diversity in salt toler-
ance within species that has not been fully explored.
One reason for this is the difficulty of measuring the
tolerance of salinity as distinct from the tolerance of
water or osmotic stress, and the difficulty of screen-
ing large numbers of individuals for small, repeatable
and quantifiable differences in biomass production, let
alone yield.
How to screen for small differences in salt
tolerance within species
Differences in salt tolerance between closely related
species are difficult to quantify, as the growth reduc-
tion depends so much on the period of time over which
the plants have grown in saline conditions. Salinity
lowers the water potential of the roots, and this quickly
causes reductions in growth rate, along with a suite
of metabolic changes identical to those caused by wa-
ter stress (reviewed by Munns, 2002). Later, there
may be salt-specific effects that impact on growth or
senescence.
The first few days or weeks in salinity may reveal
no differences in growth response between species that
have quite different reputations for salt tolerance. For
example, durum wheat, Triticum turgidum ssp. durum
is much more salt-sensitive than bread wheat, Triticum
aestivum (Francois et aI., 1986; Rawson et al., 1988),
yet over short periods of time in salinity we found
no differences between durum and bread wheat cul-
tivars (Munns et aI., 1995). In a comparison between
20 cultivars of wheat, barley and triticale we found
no significant differences between the leaf elongation
rate in the first 10 days of salinisation of any cultivar,
including that of the one that ultimately turned out to
be the most sensitive (a durum wheat) and the one (a
barley) that turned out to be the most tolerant (Rawson
et al., 1988). Similar results have been obtained re-
cently with other wheat lines that have a reputation of
differing in salt tolerance. Four weeks of growth at 150
mM NaCI was insufficient time for difference in salt
tolerance between genotypes to show up (Rivelli et aI.,
2002), including bread and durum wheat cultivars that
were known to differ in salt tolerances in the field.
96
5 A two-phase growth response has been shown
---
4
CI
J::.
CI
3
'iii
3:
~
-
'C 2 that a growth difference was clearly seen (Fortmeier
iij
0
I-
Time after NaCI added (d)
Figure 2. Two accessions of the diploid wheat progenitor Ae.
tauschii grown in supported hydroponics in control solution (closed
symbols) and in 150 mM N aCI with supplemental Ca2+ (open sym-
bols). Circles denote the tolerant accession, triangles the sensitive
one. The arrow marks the time at which symptoms of salt injury
could be seen on the sensitive accession; at that time the proportion
of dead leaves was 10% for the sensitive and 1% for the tolerant
accession (Munns et al., 1995). A similar result is given in Fortmeier
and Schubert (1995).
These data are consistent with the concept of a two-
phase growth response to salinity (Munns, 1993). The
first phase of growth reduction is quickly apparent, and
is due to the salt outside the roots. It can be called
a water stress or osmotic phase, for which there is
surprisingly little genotypic difference. Then there is a
second phase of growth reduction, which takes time to
develop, and is associated with advanced senescence
of older leaves. This presumably results from internal
injury due to salts accumulating in these transpiring
leaves to excessive levels. If excessive amounts of salt
enter a plant, salt will eventually rise to toxic levels
in the older transpiring leaves, causing premature sen-
escence and reducing the photosynthetic capacity of
the plant to a level that cannot sustain further growth
(Munns, 1993). The cause of the injury is probably
due to the salt load exceeding the ability of the cells
to compartmentalise salts in the vacuole. Salts then
would rapidly build up in the cytoplasm and inhibit
enzyme activity. Alternatively, they might build up in
the cell walls and dehydrate the cell. Evidence for ions
accumulating to high concentrations in the apoplast of
leaves has been found in rice (Flowers et aI., 1991),
but not maize (Mtihling and Uiuchli, 2002).
clearly for maize and wheat cultivars. Two maize
cultivars with 2-fold differences in rates of Na+ accu-
mulation in leaves had the same growth reduction for
15 days in 80 mMNaCl (Crameret aI., 1994). Another
two maize cultivars, again with 2-fold differences in
Na+ accumulation, had the same growth reduction for
4 weeks in 100 mMNaCI, and it was not until 8 weeks
and Schubert, 1995). Similar results were found in
wheat (Munns et aI., 1995). Two closely-related wheat
genotypes that differed in rates of Na+ accumulation
had the same growth reduction for 4 weeks in 150
mM NaCI, and it was not until after 4 weeks that a
growth difference between the genotypes was clearly
seen (Figure 2). However, within 2 weeks dead leaves
became visible on the more sensitive genotype (Figure
2), and the rate of leaf death was clearly greater on the
sensitive than the tolerant genotype. Once the num-
ber of dead leaves increased above about 20% of the
total, plant growth slowed down and many individuals
started to die (Munns et aI., 1995).
With rice, also, a clear distinction has been made
between the initial effects of salinity, from which re-
covery is possible, and the long-term effects that result
from the accumulation of salt within expanded leaves
(Yeo et aI., 1991).
These observations illustrate the principle that the
initial growth reduction is due to the osmotic effect of
the salt outside the roots, and that what distinguishes
a salt-sensitive plant from a more tolerant one is the
inability to prevent salt from reaching toxic levels in
the transpiring leaves, which takes some time.
The length of time required before growth differ-
ences between genotypes can be seen depends on the
salinity and the degree of salt tolerance of the spe-
cies. The second phase will start earlier in plants that
are poor excluders of Na+, such as lupins or beans,
and when salinities are higher. It will also start earlier
when root temperatures are higher. For plants such as
rice that are grown at high temperatures, 10-15 days in
salinity is sufficient to generate differences in biomass
between genotypes that correlate well with differences
in yield (Aslam et aI., 1993).
The labour and space demands of these long ex-
periments makes this impractical for screening large
numbers of genotypes, or selecting salt-tolerant pro-
geny. This means that our knowledge of physiological
mechanisms should be used to identify traits that can
be employed for rapid and cost-effective selection
techniques.

5) excretion by salt
4) loading of phloem
3) unloading of xylem
2) loading of xylem
1) uptake from soil
Figure 3. Control points at which salt transport is regulated. These
are: (1) selectivity of uptake from the soil solution; (2) loading of
the xylem; (3) removal of salt from the xylem in the upper part of
the plant; (4) loading of the phloem; and (5) excretion through salt
glands or bladders. For a salt tolerant plant growing for some time
in a soil solution of 100 mM N aCl, the root concentrations of N a+
and Cl- are typically about 50 mM, the xylem concentration about
5 mM, and the concentration in the oldest leaf as high as 500 mM.
See text for explanation.
Mechanisms of salt tolerance
Mechanisms of salt tolerance take place at three levels
of organisation: whole plant, cellular and molecular.
Control at the whole plant level
Physiological mechanisms conferring exclusion that
operate at the cellular and whole plant level have been
described in previous reviews (Greenway and Munns,
1980; Uiuchli, 1984; Munns et aI., 1983; Pitman,
1984; Storey and Walker, 1999), and with particu-
lar reference to selectivity for K+ over Na+ (Jeschke,
1984; Jeschke and Hartung, 2000). Salt tolerance de-
pends on the ability of the plant to control the transport
of salt at five sites (Figure 3), as summarised below:
1. Selectivity of uptake by root cells. It is still unclear
which cell types control the selectivity of ions from
the soil solution. The initial uptake ofNa+ and Cl-
could occur at the epidermis, at the exodermis, or
if soil solution flows apoplastically across the root
cortex, it would occur at the endodermis.
2. Loading of the xylem. There is evidence for a pref-
erentialloading of K+ rather than Na+ by the cells
of the stele.
3. Removal of salt from the xylem in the upper part
of the roots, the stem, petiole or leaf sheaths. In
97
many species, Na+ is retained in the upper part of
the root system and in the lower part of the shoot,
indicating an exchange of K+ for Na+ by the cells
in the stele of the roots or in the vascular bundles
in stems and petioles.
4. Loading of the phloem. There is little retransloca-
tion of Na+ or Cl- in the phloem, particularly in
the more tolerant species. This ensures that salt is
not exported to growing tissues of the shoot.
5. Excretion through salt glands or bladders. Only
halophytes have these specialised cell types.
All halophytes have well-developed mechanisms to
control the uptake, transport and excretion of salt.
Glycophytes rely on the first three mechanisms, and
exhibit these mechanisms to various degrees. Genetic
variation within a given species, or between closely
related species, has in most cases been identified as
due to different degrees of control of salt uptake by
roots, or in loading of the xylem. Genetic variation in
Na+ loading of the xylem appears to explain differ-
ences in Na+ accumulation and hence salt tolerance
between Triticum species (Gorham et aI., 1990). Ex-
clusion is particularly important for perennials, the
leaves of which may live for a year or more; there is
greater need to regulate the incoming salt load over
a much longer period of time than for annual species
whose leaves may live for only 1 month.
There are contributory features that function to
maintain low rates of salt accumulation in leaves. High
shoot root ratios and high intrinsic growth rates (Pit-
man, 1984), and absence of an apoplastic pathway in
roots (Garcia et al., 1997) all will serve to reduce the
rate at which salt enters the transpiration stream and
accumulates in the shoot.
Control at the organelle level: ion compartmentation
There is no evidence of adaptations in enzymes to the
presence of salt (reviewed by Munns et aI., 1983),
so mechanisms for salt tolerance at the cellular level
involve keeping the salt out of the cytoplasm, and se-
questering it in the vacuole of the cell. That this occurs
in most species is indicated by the high concentrations
found in leaves that are still functioning normally, con-
centrations well over 200 rnM, yet we know that these
same concentrations will completely repress enzyme
activity in vitro (Munns et aI., 1983). Generally, Na+
starts to inhibit most enzymes at a concentration above
100 rnM. The concentration at which Cl- becomes
toxic is even less well defined, but is probably in the
same range as that for Na+. If Na+ and Cl- are se-
98
K carrier
K-selective
channels
non-selective channels
NaIR antiporter Na
H
Figure 4. Mechanisms of Na+ transport in higher plants. Regulation of Na+ uptake across the plasmalemma would come from restricted
uptake by selective cation transporters and channels. coupled with efflux by the antiporter. The antiporter on the tonoplast sequesters Na+ in
the vacuole. Adapted from Amtmann and Sanders (1999).
questered in the vacuole of the cell, K+ and organic
solutes should accumulate in the cytoplasm and or-
ganelles to balance the osmotic pressure of the ions
in the vacuole. The organic solutes that accumulate
most commonly under salinity are proline and glycine
betaine, although other molecules can accumulate to
lesser degrees (summarised in Hasegawa et aI., 2000,
their Figure 1).
Control at the molecular level: ion transporters
The ion channels and transporters that regulate the
net movement of salt across cell membranes have
been recently reviewed (Amtmann and Sanders, 1999;
Blumwald, 2000; Schachtman and Liu, 1999; Tyer-
man and Skerritt, 1999). The mechanisms that control
Na+ transport are summarised in Figure 4. There is
no specific Na+ transporter, Na+ entry being gained
by competition with other cations, in particular K+.
Na+ could enter the cell through high affinity K+
carriers or through low affinity channels called non-
selective cation channels that are strongly influenced
by Ca2+. These cation channels could allow entry of
large amounts of Na+ from a highly saline soil if not
adequately regulated (Amtmann and Sanders, 1999).
Na+ can be effluxed from the cytoplasm through
Na+/H+ antiporters, driven by the pH gradient across
PLASMALEMMA
VACUOLE
the plasmalemma (Blumwald, 2000). These transport
processes all work together to control the rate of net
uptake of Na+ by a cell (Figure 4). Intracellular com-
partmentation is by a vacuolar Na+/H+ antiporter,
driven by a pH gradient across the tonoplast (Blum-
wald et aI., 2000). The transporters that maintain low
Na+ concentrations in organelles such as chloroplasts
and mitochondria are not known. In some species, CI-
transport is associated with salt tolerance. Mechan-
isms that control Cl- movement across membranes
have been comprehensively reviewed by White and
Broadley (2001).
StrategiesJor increasing salt tolerance: and the
importance oj traits
There are two main avenues for improving salt tol-
erance of a given crop or cultivated species. These
are (1) searching amongst natural diversity within the
species, or closely related and inter-fertile species,
and (2) genetic engineering. With both avenues, back-
crossing into cultivars or advanced breeding lines will
be required. This requires precise screening of pro-
geny, using either a quantitative trait or a molecular
marker for that trait. Screening for a trait associated
with a specific mechanism is preferable to screening
for salt tolerance itself, as measuring the effect of salt

on biomass or yield of a large number of lines is not
feasible. As mentioned above, plants need to be grown
for lengthy periods of time, and controls need to be
included, as the source of the salt tolerance could come
from a parent or transformant that is taller or shorter or
has a different growth rate than the cultivar into which
the germplasm or gene is being introduced. In the
field, the major drawback is the heterogenous nature
of salinity within paddocks and between sites.
Traits for salt tolerance that have been used to
screen germplasm collections have included rates of
Na+ or Cl- accumulation in leaves, degree of leaf
injury, seedling root length, and germination percent-
age. The most successful relate to rates of Na+ or
Cl- accumulation in leaves, measured as the increase
in salt in a given leaf over a fixed period of time.
Sodium accumulation in leaves has been shown to re-
late to salt tolerance in genotypes of rice (Yeo and
Flowers, 1986) and diploid wheat, Aegilops tauschii
(Schachtman et aI., 1991). Salt tolerance at germin-
ation is easy to measure, but little or no relation
between salt tolerance at germination and that of the
seedling or adult plant has been found in any species
examined, including barley (e.g., Mano and Takeda
1997), bread wheat (Kingsbury and Epstein, 1984),
and durum wheat (Almansouri et aI., 2001). In the
approach described below, we use the rate of Na+
accumulation in a given leaf as a non-destructive and
accurate quantitative trait.
Molecular markers for these traits can provide
an efficient selection technique in breeding pro-
grams. Some success has been reported for combining
physiological traits in rice (lRRI, 1997). Molecular
markers would be particularly useful for pyramiding
different traits for salt tolerance (Flowers et aI., 2000;
Yeo and Flowers, 1986), and additionally for incorpor-
ating characters associated with other accompanying
stresses, such as drought or waterlogging. We are
therefore attempting to identify molecular markers for
salt tolerance in wheat.
Physiology and genetics of sodium exclusion
Our current work is focussed on improving the salt
tolerance of durum wheat (Triticum turgidum ssp.
durum). Cultivated durum wheat is more salt-sensitive
than bread wheat (Triticum aestivum), a feature that
restricts its expansion into areas with sodic or saline
soils. The trait that we have targeted is low rates of
Na+ uptake. Screening for this by a non-destructive
99
method (Na+ concentration in a given leaf 10 days
after emergence, in plants at 150 mM NaCl) has al-
lowed us to identify novel sources of Na+ exclusion
in ancient landraces of durum wheat (Munns et aI.,
2000).
In wheat, salt tolerance is associated with low rates
of transport of Na+ to shoots with high selectivity for
K+ over Na+; there is little genotypic variation in
rates of Cl- transport (Gorham, 1990). Bread wheat
cultivars (hexaploid, AABBDD genomes) have a low
rate of Na+ accumulation and enhanced K+ IN a+ dis-
crimination, a character located on the long arm of
chromosome 4D (Gorham et aI., 1987). This char-
acter is controlled by a single locus (Knal) and has
been linked to molecular markers on the distal third
of chromosome 4DL (Dubcovsky et aI., 1996). The
gene or genes associated with this locus have not been
identified. Durum wheat cultivars (tetraploid, AABB
genomes) have high rates of Na+ accumulation and
poor K+lNa+ discrimination (Gorham et aI., 1987),
and are less salt-tolerant than bread wheat. One ap-
proach to improve the salt tolerance of durum wheat
has been to create novel germplasm with low accumu-
lation of Na+ and enhanced K+ lNa+ discrimination
by homologous recombination with chromosome 4D
(Dvorak et aI., 1994). This, however, brings in un-
wanted genetic material on the translocated chromo-
some segment, which cannot be eliminated. Another
approach is to search for natural genetic diversity on
the A or B genomes, and this is the approach we have
taken.
(a) Genetic variation in N a + exclusion
In order to introduce salt tolerance into current durum
wheat from sources other than the D genome, we
searched for genetic variation in salt tolerance across
a wide range of durum-related tetraploids representing
five Triticum turgidum sub-species (durum, carth-
licum, turgidum, turanicum, polonicum). Selections
were screened for low Na+ uptake and its associ-
ated enhanced K+ IN a+ discrimination. Wide genetic
variation was found (Munns et aI., 2000). Low Na+
accumulation (and high K+ lNa+ discrimination) of
similar magnitude to that of bread wheat was found in
a landrace from the sub-species durum (Munns et aI.,
2000). Figure 5 illustrates the range of genetic vari-
ation in Na+ uptake that exists in the Triticum genus.
The low Na+ values for the bread wheat 'Janz' are
typical of bread wheat cultivars, and the higher Na+
values for the durum wheats 'Tamaroi' and 'Wollaroi'
100
1-400
(c) Does Na+ exclusion carry a penalty for water
1200 relations?
~~

rs. 1000 Four wheat genotypes with contrasting rates of Na+

~ accumulation were selected to see if Na+ exclusion
~ 800
c: resulted in poorer water relations during exposure to
,g
I" 600 saline conditions. One genotype was the low-Na+
.,<.>
C'
landrace used in the durum improvement program de-
c:
8 -400
scribed above. Another was the bread wheat 'Janz'
0 ..

2
200
o while the fourth was a durum landrace with extremely
HIgjINa' T....rol Low Na+ Janz
durum (durum)
"-bI~11)1
(durum) durum (br.ad)
Iond.... landrace
Figure 5. Sodium accumulation over 10 days in leaves of wheat
genotypes. Plants were grown in supported hydroponics in 150
mM NaCI with half-strength Hoagland's solution and supplemental
Ca2+. Shown are the durum landraces with the highest and low-
est Na+ accumulation of the tetraploids screened by Munns et aI.
(2000), two current durum cultvars (Tamaroi and Wollaroi) and a
represtative bread wheat (Janz).
are typical of current durum wheat cultivars. Most
landraces examined had Na+ values as high or higher
than the cultivars, but a few had very low values. The
landrace with the lowest rate of Na+ accumulation is
shown in Figure 5, and this trait is being incorpor-
ated into the durum cultivars through a backcrossing
program.
(b) Inheritance of Na+ exclusion
The low-Na+ durum landrace (Figure 5) was crossed
with the durum cultivar Tamaroi. The phenotypes of
the F 1, F2 individuals and F2:3 families from this cross
were determined by measuring the Na+ accumulation
in leaf 3 at 10 days after emergence, in plants grown
at 150 rnM NaCl. The Fl progeny were intermediate
between the parents in Na+ accumulation, but about
10% of the F2 progeny had Na+ levels as low as the
low-Na+ parent. The distribution of Na+ accumula-
tion in individuals in the F2 population and the F2:3
families indicated that more than one gene was in-
volved, most likely two or three genes of major effect
(data not shown). Inheritance of this trait was assessed
by regressing a selection of the fifteen highest and low-
est F2:3 family means with the corresponding single
F2 plant values. This showed that the trait was highly
heritable with a high narrow sense heritability of 0.79
(±0.07) and therefore viable for use in a breeding
program.
with similarly low rate of Na+ uptake. A third was the
durum wheat 'Tamaroi' with high rates of Na+ uptake
high rates ofNa+ uptake. The Na+ levels of these four
genotypes are shown in Figure 5. Plants were grown
in supported hydroponics, with and without 150 rnM
NaCl, and sampled for water relations, biomass, and
ion accumulation over time. The results, as described
by Rivelli et aI. (2002), showed that there was little
difference between genotypes in the effect of salinity
on water relations, as indicated by their water po-
tential, estimated turgor, and relative water content.
Osmotic adjustment occurred in all genotypes, with
one of the low-Na+ genotypes having the greatest os-
motic adjustment, and one of the high-Na+ genotypes
having the lowest osmotic adjustment. In the low-Na+
genotypes, osmotic adjustment was enabled in part by
the higher K+, as Na+ exclusion was always asso-
ciated with maintenance of higher K+ levels. Other
solutes, such as Cl- and organic solutes, also played
a part (Rivelli et aI., 2002). Over the 4-week period
of this experiment, there was no significant differ-
ence between genotypes in the effect of salinity on
growth. These data indicate that selecting lines with
low Na+ accumulation for the purpose of improving
salt tolerance is unlikely to introduce adverse effects
on plant-water relations or growth.
The reason why Na+ exclusion might not impose a
limitation to osmotic adjustment is that it is generally
associated with K+ accumulation. We asked the ques-
tion: would Na+ exclusion carry a penalty for water
relations in soils with low K+?
(d) Does low K+ supply affect peiformance of
low-Na+ lines?
To answer this question, four genotypes with contrast-
ing rates of Na+ accumulation were selected. These
were basically the same as the four genotypes used in
the water relations experiment described above, dif-
fering only that the durum cultivar was 'Wollaroi' (see
Figure 5). Plants were grown for 2 weeks at two K+
levels, one representing an enriched soil, 3.3 rnM,

the same as half-strength Hoagland's nutrient solution,
the other representing a low-K+ soil, 0.5 mM. There
were three salinity levels: 1, 100 and 150 mM NaCl,
with supplemental Ca2+ to overcome the salt effect on
Ca2+ activity of the external solution.
As expected, the low K+ supply reduced K+ up-
take and increased Na+ uptake in all genotypes, but
to different degrees depending on the ion, the salin-
ity level and the genotype (data not shown). At 150
mM NaCl, the increased uptake of Na+ entirely com-
pensated for the decreased K+ uptake, so that the sum
of (K+ + Na+) was not significantly affected in any
genotype by the low K+ supply (Table 1). At 100 mM
NaCl, the decreased K+ uptake was not balanced by
the increased uptake of Na+ in two of the four gen-
otypes, so there was a small but significant decrease
in the sum of (K+ +Na+) in two genotypes. Only one
of these was a low-Na+ genotype. At 1 mM NaCl the
sum of (K+ + Na+) was significantly reduced in all
genotypes, as the decrease in K+ uptake was greater
than the increase in Na+ uptake (Table 1). The growth
rate was not affected by the low K+ treatment at any of
the three salinity levels over this experimental period
of2 weeks.
The experiment therefore showed that there were
no effects of K+ supply on the accumulation of Na+
or K+ in either the shoot (Table 1) or the root (data not
shown) that would restrict the osmotic adjustment of
the low-Na+ genotypes in saline soil. Cl- accumula-
tion was not affected by the low K+ supply (data not
shown). Curiously, Ca2+ uptake was enhanced by low
K+ supply, by about one third, at all salinities and in
all genotypes. Thus the higher external K+ competed
with both Na+ and Ca2+ for uptake.
In summary, we identified a new source of Na+
exclusion that can be incorporated into modem durum
cultivars with no growth penalty. Isogenic lines with
high versus low Na+ accumulation are being de-
veloped, and will provide germplasm that can be
evaluated in the field to test the concept that low Na+
accumulation increases biomass production and yield
of durum wheat in saline soil.
Molecular markers for physiological traits
The development of molecular markers for physiolo-
gical traits has made significant headway in recent
years with the advancement of new technologies. Con-
sequently, the use of molecular markers in breeding
programs is increasing rapidly as they have been
101
shown to greatly improve the efficiency of the breed-
ing programs. Although the application of molecular
markers is relatively straightforward, the development
of robust markers that are reliable across a wide range
of backgrounds can be quite difficult, and is entirely
dependent on an accurate phenotype screen. Under-
standing the physiology of sodium uptake is critical
to the development of a reliable and accurate phen-
otype test, and thereby to the identification of QTLs
and molecular markers.
QTLs (Quantitative Trait !eoci) for salt tolerance
have been described in several cereal species, includ-
ing rice (Flowers et aI., 2000; Koyama et aI., 2001)
barley (Ellis et aI., 1997; Mano and Takeda, 1997),
and bread wheat (Semikhodskii et aI., 1997). How-
ever, these studies have not yet yielded robust markers
that can be used across a range of germplasm, signi-
ficant associations between the trait and the marker
being confined to the populations in which they were
derived. The success of these studies could be lim-
ited by the small amount of genetic diversity present
within modem cultivars, and the use of parental lines
with small differences in the traits. Our approach has
been to seek a wider genetic diversity than exists in
modem populations. This is possible with tetraploid
and hexaploid wheats, as the progenitors of modem
durum and bread wheats may have been derived from
a limited germplasm base, and there may be genetic
diversity present in original populations of the diploid
ancestors that is not yet exploited.
Aegilops tauschii - diploid wheat
To develop markers for salt tolerance in bread wheat
(AABBDD), we worked with Ae. tauschii, the diploid
progenitor of bread wheat (DD). This species (syn. Ae.
squarrosa, Triticum tauschii) was chosen because the
D genome was shown to be responsible for the en-
hanced sodium exclusion of bread wheat as compared
to durum wheat (Shah et aI., 1987).
We searched a large collection of Ae. tauschii
accessions for sodium exclusion (Schachtman et aI.,
1991). Accessions containing high and low sodium in
the most recent fully expanded leaf (leaf 5) were selec-
ted and crossed to construct F2 populations that could
be mapped in search of markers linked to sodium
exclusion. Three populations were created by cross-
ing the selected accessions: CPIl10835 xCPIl10791,
CPIl10664 x AUS 18905 and CPIl10664 x CPIl10791.
Progeny from the Fl were grown in salt and Na+
concentrations in leaf 5 were measured. Populations
102
Table 1. Effect of low K+ supply (0.5 mM versus 3.3 mM K+ in the soil solution)
on the sum of (K+ +Na+) concentration in the whole shoot of wheat genotypes with
low and high Na+ uptake rates (details in text). Plants were grown for 2 weeks at
three salinities: 1,100 and 150 mMNaCl with supplemental Ca2+(2, 8 and 10 mM
Ca2+, respectively). Asterisks denote significant differences at the P = 0.05 level.
Genetic difference in leaf N a+ concentrations in the four genotypes are shown in
Figure 5 for 150 mMNaCl

K+ +Na+ (mmolg- 1 DW)

1 mMNaCl 100mMNaCl l50mMNaCl
External K+ (mM) 3.3 0.5 3.3 0.5 3.3 0.5

Low-Na+ durum landrace 1.21 1.00* 1.32 1.28 1.52 1.47

Low-Na+ cv 'Janz' 1.26 0.96* 1.42 1.21* 1.52 1.48
High-Na+ cv 'Wollaroi' 1.24 1.01* 1.70 1.54* 1.85 1.70
High-Na+ durum landrace 1.22 0.97* 1.80 1.67 1.77 1.69

(74-96 individuals) were skewed to the low sodium
parent, and there appeared to be transgressive se-
gregation which produced individuals with lower Na+
concentrations in leaf 5 than the parental accessions
(Schachtman, 1991). These results indicated that mul-
tiple genes were associated with the sodium exclusion
trait, although genes of major effect that confer sodium
exclusion were evident. The population created from
CPI110664xAUS18905 was used for mapping, and
43 out of 150 available RFLP probes were found to
be polymorphic. A linkage map was then constructed
with the 38 linked markers with five markers remain-
ing unlinked. Regions of significant QTLs for Na+
concentrations in leaf 5 were calculated along por-
tions of the skeletal linkage groups for the Ae. tauschii
genome using MAPMAKER QTL software. A single
QTL (LOD score=2.1) was found on chromosome 4
between the markers that encode sequences for the
loci controlling the early germination protein, germin,
and the 7S globulin gene, 7SglobA. These markers are
linked to the short arm of chromosome 4. At the time
the analysis was completed markers on the long arm
of the chromosome 4 were not available. With addi-
tional polymorphic markers the statistical significance
of the QTL analysis could have been increased. In
addition our analysis revealed that the sodium exclu-
sion trait showed a low narrow sense heritability of
0.11, indicating that either the phenotyping was not
sufficiently precise, or that there was a strong envir-
onmental influence (a large genotype by environment
interaction). Phenotyping under a number of different
environments would have clarified this. These results
provided guidance for later work.
This mapping activity provided information on the
dominance of the sodium exclusion trait and sug-
gested that genetic improvement in sodium exclusion
could be achieved by breeding. To test this, we de-
termined whether the differences in salt tolerance and
Na+ accumulation in the different Ae. tauschii ac-
cessions would be expressed in synthetic hexaploid
wheat, by crossing three of them with a common
tetraploid wheat (Schachtman et al., 1992). The salt
tolerance of the synthetic hexaploids was greater than
the tetraploid parents primarily due to the maintenance
of kernel weight. The synthetic hexaploids varied in
salt tolerance according to the salt tolerance of the Ae.
tauschii used in the cross, demonstrating that genes
for salt tolerance from the diploid are expressed at the
hexaploid level.
Durum wheat
The construction of a genetic map in durum wheat
(AABB) and subsequent development of molecular
markers for the trait of sodium exclusion is of par-
ticular interest to our group. As mentioned earlier,
durum wheat lacks the D genome and the associ-
ated trait of sodium exclusion. We are attempting to
find a source of this trait on the A or B genome, in
old durum wheat landraces or related tetraploid spe-
cies, and to introduce this trait into Australian durum
cultivars. Improving the salt tolerance of Australian
durum cultivars will enable the continued growth of
these high-yielding and high-value crops in the Aus-
tralian wheat belt, which is faced with rising water
tables and the risk of dryland salinity (ANRA, 2001).

A population segregating for the low Na+ uptake
trait was developed from a cross between the low-
Na+ landrace and the cultivar 'Tamaroi' (see Figure
5), for which phenotypes of the F2 individuals and
F2:3 families were determined by measuring the Na+
accumulation in leaves 10 days after emergence, as
described earlier. This population had shown a high
level of heritability. DNA extracted from these indi-
viduals provided the material for genotypic analysis
and the construction of the genetic map. Construc-
tion of a genetic linkage map based on AFLPs and
microsatellites was initiated, to identify the chromo-
somal regions of major effect on Na+ accumulation.
Initially, 144 AFLP primer combinations were used to
identify polymorphisms between the parental lines and
bulked segregants of the 15 highest Na+ uptake lines
and 15 lowest Na+ uptake lines. Twenty-three primer
combinations were polymorphic, and with an aver-
age of five sites of polymorphism with each primer
combination, approximately 100 polymorphic bands
were available for scoring across the F2 population.
In addition, a screen of the parental lines using 112
micro satellites evenly distributed through the durum
wheat genome identified a group of micro satellite
markers that were polymorphic. Using a high strin-
gency mapping approach (P = 0.001) several linkage
groups were identified. Those with known map loca-
tions were identified and the marker densities in the
identified regions were increased using RFLPs. In-
terval mapping using MapManager QTX version 13b
revealed a QTL located on chromosome 2AL. This
QTL showed significant association with the trait hav-
ing a LOD score of 7.5. The analysis has shown that
this locus accounts for half of the phenotypic variation
of this trait. Other linkage groups did not have signi-
ficant association with the trait. Results indicate that
the allelic contribution to the QTL located on chromo-
some 2A was predominantly from the low-Na uptake
parent. Robust molecular markers for this locus are
being developed.
Concluding remarks
Transformation techniques available for most crop
species make it possible to manipulate the expression
of genes involved in the control of transport of Na+
across membranes. There are various candidate genes
from higher plants, as indicated in Figure 4, and also
some yeast-specific ones (Schachtman and Liu, 1999),
for control of transport of Na+ across membranes.
103
Some of these have been overexpresssed in model sys-
tems, with subsequent increase in salt tolerance. The
most dramatic responses have been with the vacuolar
antiporter AtNHXl in Arabidopsis (Apse et aI., 1999)
and tomato (Zhang and Blumwald, 2001).
However, the lack of community acceptance of
genetic engineering of crop species means that other
approaches need to be taken at present. Molecular
markers offer a way around this current impasse; user-
friendly markers can be developed from germplasm
with contrasting phenotypes, i.e., quantifiable trait
differences, using QTL or bulk segregant analysis,
or from genes of known function. They can then
be used to follow the inheritance of the trait during
backcrossing into cultivars.
We conclude that there is considerable natural ge-
netic variation in transport processes controlling the
uptake and accumulation of Na+ and Cl- that is yet
to be utilised for increasing the salt tolerance of crop
species. With an understanding of the function of these
transporters at the whole plant level, this genetic vari-
ation can be exploited for developing molecular mark-
ers to track the introduction of salt-tolerant germplasm
into cultivars by conventional breeding methods, and
ultimately for identifying genes that can be used for
transformation when salt tolerance in closely related
germplasm cannot be found.
In summary, modem molecular techniques offer
new approaches to improving salt tolerance of crops.
Possibly a combination of all approaches, old and new,
will be the most productive. Identifying physiological
traits and key genes, and understanding mechanisms
at the cellular and whole plant level, is central to all
approaches.
Acknowledgement
This work is supported by the Grains Research and
Development Corporation.
References
Almansouri M, Kinet J -M and Lutts S 200 I Effect of salt and os-
motic stresses on germination in durum wheat (Triticum durum
Desf.). Plant Soil 231, 243-254.
ANRA 2001 Australian Natural Resources Atlas.
http://audit.ea.gov.au.
Apse M P, Aharon G S, Snedden W A and Blumwald E 1999 Salt
tolerance conferred by overexpression of a vacuolar N a+IH+
antiporter in Arabidopsis. Science 285, 1256-1258.
104
Aslam M, Qureshi R H and Ahmed N 1993 A rapid screening tech-
nique for salt tolerance in rice (Oryza sativa L.). Plant Soil 150,
99-107.
Amtmann A and Sanders D 1999 Mechanisms of Na+ uptake by
plant cells. Adv. Bot. Res. 29, 76-112.
Blumwald E 2000 Sodium transport and salt tolerance in plant cells.
Curro Opin. Cell BioI. 12,431-434.
Blumwald E, Aharon G S and Apse M P 2000 Sodium transport in
plants. Biochim. Biophys. Acta 1465, 140-151.
Cramer G R, Alberico G J and Schmidt C 1994 Salt tolerance is not
associated with the sodium accumulation of two maize hybrids.
Aust. J. Plant Physiol. 21, 675-692.
Dubcovsky J, Santa Maria G, Epstein E, Luo M C and Dvorak
J 1996 Mapping of the K+/Na+ discrimination locus KnaI in
wheat. Theor. Appl. Genet. 2, 448-454.
Dunin F X, Smith C J, Zegelin S J, Leuning R, Denmead 0 T and
Poss R 2001 Water balance changes in a crop sequence with
lucerne. Aust. J. Agric. Res. 52, 247-261.
Dvorak J, Noaman M M, Goyal S and Gorham J 1994 Enhancement
of the salt tolerance of Triticum turgidum L. by the KnaI locus
transferred from the Triticum aestivum L. chromosome 4D by
hornoeologous recombination. Theor. Appl. Genet. 87, 872-877.
Ellis R P, Forster B P, Waugh R, Bonar N, Handley L L, Robinson
D, Gordon D C and Powell W 1997 Mapping physiological traits
in barley. New Phytol. 137,149-157
FAOSTAT Statistics database. http://appsJao.org/cgi-bin/nph-db.pl.
Flowers T J, Hajibagheri M A and Yeo A R 1991 Ion accumulation
in the cell walls of rice plants growing under saline conditions:
evidence for the Oertli hypothesis. Plant Cell Environ. 14, 319-
325.
Flowers T J, Koyama M L, Flowers S A, Sudhakar C, Singh K P
and Yeo A R 2000 QTL: their place in engineering tolerance of
rice to salinity. J. Exp. Bot. 51, 99-106.
Fortmeier R and Schubert S 1995 Salt tolerance of maize (Zea mays
L.): the role of sodium exclusion. Plant Cell Environ. 18, 1041-
1047.
Francois L E and Maas E V 1994 Crop response and management
on salt-affected soils. In Handbook of Plant and Crop Stress. Ed.
M Pessarakli. pp. 149-181. Marcel Dekker, New York, USA.
Francois L E, Maas E V, Donovan T J and Youngs V L 1986 Ef-
fect of salinity on grain yield and quality, vegetative growth,
and germination of semi-dwarf and durum wheat. Agron. J. 78,
1053-1058.
Garcia A, Rizzo C A, Ud-Din J, Bartos S L, Senadhira D, Flowers
T J and Yeo A R 1997 Sodium and potassium transport to the
xylem are inherited independently in rice, and the mechanism
of sodium:potassium selectivity differs between rice and wheat.
Plant Cell Environ. 20,1167-1174.
Ghassemi F, Jakeman A J and Nix H A 1995 Salinisation ofland and
water resources: Human causes, extent, management and case
studies. UNSW Press, Sydney, Australia, and CAB International,
Wallingford, UK.
Gorham J 1990 Salt tolerance in the Triticeae: KINa discrimination
in synthetic hexaploid wheats. J. Exp. Bot. 41, 623-627.
Gorham J, Hardy C, Wyn Jones R G, Joppa L R and Law C N 1987
Chromosomal location of a KINa discrimination character in the
D genome of wheat. Theor. Appl. Genet. 74, 584-588.
Gorham J, Wyn Jones R G and Bristol A 1990 Partial characteriz-
ation of the trait for enhanced K+ -Na+ discrimination in the D
genome of wheat. Planta 180,590-597.
Greenway H 1962 Plant response to saline substrates I. Growth
and ion uptake of several varieties of Hordeum during and after
sodium chloride treatment. Aust. J. BioI. Sci. 15, 16--38.
Greenway H and Munns R 1980 Mechanisms of salt tolerance in
nonhalophytes. Annu. Rev. Plant Physiol. 31, 149-190.
Hasegawa PM, Bressan R A, Zhu J-K and Bohnert H J 2000 Plant
cellular and molecular responses to high salinity. Annu. Rev.
Plant Physiol. Plant Mol. BioI. 51, 463-499.
IRRI 1997 Rice varieties boost yield and improve saline soils. In
Partners Making a Difference. Eds. C Dedolph, G Hettel. IRRI,
Manila. p. 37.
Jeschke W D 1984 K+ -Na+ exchange at cellular membranes, in-
tracellular compartmentation of cations, and salt tolerance. In
Salinity Tolerance in Plants: Strategies for Crop Improvement.
Ed. R C Staples. pp. 37-66. Wiley, New York.
Jeschke W D and Hartung W 2000 Root-shoot interactions in
mineral nutrition. Plant Soil 226, 57-69.
Khatun S, Rizzo C A and Flowers T J 1995 Genotypic variation in
the effect of salinity on fertility in rice. Plant Soil 173, 239-250.
Kingsbury R W and Epstein E 1984 Selection for salt-resistant
spring wheat. Crop Sci. 24, 310-314.
Koyama M L, Levesley A, Koebner R M D, Flowers T J and Yeo A
R 2001 Quantitative trait loci for component physiological traits
determining salt tolerance in rice. Plant Physiol. 125, 406-422.
Lauchli A 1984 Salt exclusion: an adaptation of legumes for crops
and pastures under saline conditions. In Salinity Tolerance in
Plants: Strategies for Crop Improvement. Ed. R C Staples. pp.
171-187. Wiley, New York.
Layahe P A and Epstein E 1971 Calcium and salt toleration by bean
plants. Physiol. Plant. 25, 213-218.
Maas E V and Hoffman G J 1977 Crop salt tolerance - current
assessment. ASCE J. Irrig. Drain. Div. 103, 115-134.
Mano Y and Takeda K 1997 Mapping quantitative trait loci for
salt tolerance at germination and the seedling stage in barley
(Hordeum vulgare L.). Euphytica 94,263-272.
Muhammed S, Akbar M and Neue H U 1987 Effect of NaiCa and
NaJK ratios in saline culture solution on the growth and mineral
nutrition of rice (Oryza sativa L.). Plant Soil 104, 57-62.
Miihling K H and Lauchli A 2002 Effect of salt stress on growth and
cation compartmentation in leaves of two plant species differing
in salt tolerance. J Plant Physiol. 159, 137-146.
Munns R 1993 Physiological processes limiting plant growth in sa-
line soil: some dogmas and hypotheses. Plant Cell Environ. 16,
15-24.
Munns R 2002 Comparative physiology of salt and water stress.
Plant Cell Environ. 25, 239-250.
Munns R, Greenway H. and Kirst G 0 1983 Halotolerant euk-
aryotes. In Physiological Plant Ecology. III. Responses to the
Chemical and Biological Environment. Eds. 0 L Lange, P S
Nobel, C B Osmond and H Zeigler. pp. 59-135. Encycl. Plant
Physiol., New Series, Vol. 12C. Springer, Berlin.
Munns R, Schachtman D P and Condon A G 1995 The significance
of a two-phase growth response to salinity in wheat and barley.
Aust. J. Plant Physiol. 22, 561-569.
Munns R, Hare R A, James R A and Rebetzke G J 2000 Genetic
variation for improving the salt tolerance of durum wheat. Aust.
J. Agric. Res. 51, 69-74.
Pitman M G 1984 Transport across the root and shoot/root in-
teractions. In Salinity Tolerance in Plants: Strategies for Crop
Improvement. Ed. R C Staples. pp. 93-123. Wiley, New York.
Rawson H M, Richards R A and Munns R 1988 An examination of
selection criteria for salt-tolerance in wheat, barley and triticale
genotypes. Aust. J. Agric. Res. 39,759-772.
Ridley A M, Christy B, Dunin F X, Haines P J, Wilson K F and
Ellington A 2001 Lucerne in crop rotations on the Riverine Plains
1: The soil water balance. Aust. J. Agric. Res. 52, 279-293.

Rivelli A R, James R A, Munns R and Condon A G 2002 Effect of
salinity on water relations and growth of wheat genotypes with
contrasting sodium uptake. Funct. Plant Bioi 29 (Issue No 9). (in
press).
Schachtman D P 1991 Physiology and genetics of sodium accu-
mulation and salt tolerance in Triticum species. PhD Thesis,
Australian National University.
Schachtman D and Liu W 1999 Molecular pieces to the puzzle of
the interaction between potassium and sodium uptake in plants.
Trends Plant Sci. 4, 281-287.
Schachtman D P, Munns R and Whitecross M I 1991 Variation of
sodium exclusion and salt tolerance in Triticum tauschii. Crop
Sci. 31, 992-997.
Schachtman D P, Lagudah E S and Munns R 1992 The expression of
salt tolerance from Triticum tauschii in hexaploid wheat. Theor.
App!. Genet. 84,714-719.
Sernikhodskii A G, Quarrie S A and Snape J W 1997 Mapping
quantitative trait loci for salinity responses in wheat. In Drought
and Plant Production. pp. 83-92. Proceedings of Lipenski Vir
Meeting, Serbia 1996.
Shah S H, Gorham J, Forster B P and Wyn Jones R G 1987 Salt
105
tolerance in the Triticeae: the contribution of the D genome to
cation selectivity in hexaploid wheat. J. Exp. Bot. 38, 254-269.
Storey R and Walker R R 1999 Citrus and salinity. Sci. Hort. 78,
39-81.
Szabolcs I 1994 Soils and salinisation. In Handbook of Plant and
Crop Stress. Ed. M Pessarakali. pp. 3-11. Marcel Dekker, New
York.
Tyerman S D and Skerrett M 1999 Root ion channels and salinity.
Sci. Hort. 78, 175-235.
White P J and Broadley M R 2001 Chloride in soils and its uptake
and movement within the plant: a review. Ann. Bot. 88, 967-988.
Yeo A R and Flowers T J 1986 Salinity resistance in rice (Oryza
sativa L.) and a pyramiding approach to breeding varieties for
saline soils. Aust. J. Plant Physio!. 13, 161-173.
Yeo A R, Lee K S, Izard P, Boursier P J and Flowers T J 1991 Short-
and long-term effects of salinity on leaf growth in rice (Oryza
sativa L.). J. Exp. Bot. 42, 881-889.
Zhang H X and Blumwald E 2001 Transgenic salt-tolerant tomato
plants accumulate salt in foliage but not in fruit. Nat. Biotech.
19,756-768.
Mineral element toxicity and resistance
 Plant and Soil 247: 109-119,2002.
109
© 2002 Kluwer Academic Publishers.

Mechanisms of metal resistance in plants: aluminum and heavy metals

Leon V. Kochian1,3, Nicole S. Pence l , Deborah L.D. Letham l , Miguel A. Pineros l , Jurandir V.
Magalhaes 1, Owen A. Hoekenga1 & David F. Garvin2
1 U.S. Plant, Soil and Nutrition Laboratory, USDA -ARS, Cornell University, Ithaca, NY 14853, USA. 2 Plant Science
Research Unit, USDA -ARS, 411 Borlaug Hall, University ofMinnesota, St. Paul, MN 551 08, USA. 3 Corresponding
author*

Key words: aluminum resistance, anion channel, gene expression, heavy metal transport, metal hyperaccumulation,
organic acid exudation

Abstract
Plants have evolved sophisticated mechanisms to deal with toxic levels of metals in the soil. In this paper, an
overview of recent progress with regards to understanding fundamental molecular and physiological mechanisms
underlying plant resistance to both aluminum (AI) and heavy metals is presented. The discussion of plant Al
resistance will focus on recent advances in our understanding of a mechanism based on Al exclusion from the root
apex, which is facilitated by AI-activated exudation of organic acid anions. The consideration of heavy metal
resistance will focus on research into a metal hyperaccumulating plant species, the ZnlCd hyperaccumulator,
Thlaspi caerulescens, as an example for plant heavy metal research. Based on the specific cases considered in
this paper, it appears that quite different strategies are used for Al and heavy metal resistance. For AI, our current
understanding of a resistance mechanism based on excluding soil-borne Al from the root apex is presented. For
heavy metals, a totally different strategy based on extreme tolerance and metal hyperaccumulation is described for a
hyperaccumulatorplant species that has evolved on naturally metalliferous soils. The reason these two strategies are
the focus of this paper is that, currently, they are the best understood mechanisms of metal resistance in terrestrial
plants. However, it is likely that other mechanisms of Al and/or heavy metal resistance are also operating in certain
plant species, and there may be common features shared for dealing with Al and heavy resistance. Future research
may uncover a number of novel metal resistance mechanisms in plants. Certainly the complex genetics of AI
resistance in some crop plant species, such as rice and maize, suggests that a number of presently unidentified
mechanisms are part of an overall strategy of metal resistance in crop plants.

Introduction when cultivated on the significant areas of acid soils

The molecular and physiological basis for crop plant
interactions with the environment has attracted consid-
erable interest in recent years. Within the general area
of plant responses to the environment, one research
area that has been the focus for many laboratories
deals with the underlying mechanisms and strategies
plants employ to tolerate potentially toxic levels of
metals in the soil. A significant portion of this type
of research has dealt with mechanisms that plants em-
ploy to deal with toxic levels of soil aluminum (AI)
* FAX No: +1-607-255-2459.
E-mail: Ivkl @cornell.edu
throughout the world. The focus of the research on
plant Al toxicity has been to ultimately identify the
genes conferring AI resistance, in order to assist in
the development of crops that will be better suited for
growth on acid soils. Another important area of metal
resistance research involves molecular and physiolo-
gical mechanisms of plant heavy metal resistance. One
reason for increased interest in plant heavy metal in-
teractions has been the recent attention on the use of
plants to remediate heavy metal contaminated soils.
Interest in this concept, termed phytoremediation, has
in turn been driven partly by the growing awareness by
the scientific community of the existence of a number
of metal hyperaccumulating plant species. In a num-
110
ber of laboratories, metal hyperaccumulator plants are
currently being studied to gain a better understand-
ing of molecular mechanisms that confer heavy metal
hyperaccumulation and extreme metal resistance in
plants.
In this paper, plant mechanisms of Al and heavy
metal resistance will be compared and contrasted. The
discussion of Al resistance mechanisms will deal with
recent work by a number of laboratories on mechan-
isms of Al exclusion from the root apex in crop species
that depend on AI-activated exudation of organic acid
anions from the root tip. For the section on plant heavy
metal resistance mechanisms, we will focus on metal
hyperaccumulators as a dramatic example of tolerance
to high levels of toxic metals in the soil, as well as
the ability to accumulate those metals to extremely
high levels in the shoot. For our examination of heavy
metal hyperaccumulation, our laboratory's molecular
and physiological investigations of the Zn/Cd hyper-
accumulator, Thlaspi caerulescens, will be used as an
example.
Plant Al resistance
As aluminum (AI) is the most abundant metal and
third most abundant element in the earth's crust, plants
have evolved in a soil environment where the roots
are exposed to potentially high levels of aluminum.
Fortunately, phytotoxic forms of Al are relatively in-
soluble at alkaline, neutral or mildly acidic soil pH
values. However, at soil pH values at or below 5,
the rhizotoxic Al species, AI3+, is solubilized into
the soil solution, inhibiting root growth and function
and thus reducing crop yields. Acid soils limit ag-
ricultural productivity in many regions of the world.
Approximately 30% of the world's total land area con-
sist of acid soils, and it has been estimated that over
50% of the world's potentially arable lands are acidic
(von Uexkiill and Mutert, 1995). Acid soils limit the
growth of crops in many developing countries where
food production is critical. Furthermore, in developed
countries such as the United States, high-input farming
practices such as the extensive use of ammonia fertil-
izers are causing further acidification of agricultural
soils. The addition of lime to acid soils can help to
correct soil acidity, and this is a common agronomic
practice in developed countries. However, the cost of
purchasing and transporting lime to the farm precludes
liming as an effective strategy for low-income farm-
ers, and liming is not effective in correcting subsoil
acidity. Therefore, considerable effort has been put
into developing crop genotypes expressing increased
Al resistance. For nearly 100 years, breeding pro-
grams have been effective in producing AI-resistant
crop varieties, in particular for cereals such as wheat
and maize (see, for example, Beckmann, 1976). Re-
cently, there also has been a focus on molecular and
physiological mechanisms of AI resistance. One of the
primary goals of this type of research is the identific-
ation of Al resistance genes that can then be used via
biotechnology to increase Al resistance in crop species
for which significant natural variation in this trait does
not exist.
Physiology ofAl resistance
Two classes of mechanisms have been proposed to
account for Al resistance: mechanisms that allow the
plant to tolerate Al accumulation in the symplasm
(AI tolerance), and those which exclude Al from the
root apex (AI exclusion) (Delhaize and Ryan, 1995;
Kochian, 1995; Kochian and Jones, 1997). There
have been many different mechanisms proposed for
AI tolerance and exclusion in the literature, with little
evidence supporting most of these hypotheses. How-
ever, recent experimental evidence has been presented
supporting the role of organic acid anion exudation
from the root apex as a major mechanism of AI exclu-
sion. The root tip is the site where Al resistance genes
must act to control resistance mechanisms, as this is
the site of AI toxicity (Ryan et aI., 1993; Sivaguru
and Horst, 1998). In a number of plant species (e.g.,
wheat, maize, buckwheat, rye, taro, snapbean) it has
been shown that Al resistance by exclusion appears to
be mediated by AI-activated release of organic acid an-
ions such as malate, oxalate, or citrate, which chelate
AI3+ in the rhizosphere and prevent its entry into the
root apex (Ma et aI., 2001).
When one considers all of the evidence in sup-
port of AI-induced organic acid release from the root
as a bona fide Al resistance mechanism in plants, a
very strong case in support of this hypothesis can be
made. These findings include: (1) a strong correlation
between AI resistance and AI-induced organic acid
anion release in a range of different plant species (Del-
haize et al., 1993a,b; Ma and Miyasaka, 1998; Ma et
aI., 1997a,b; Miyasaka et aI., 1991; Pellet et aI., 1995;
Ryan et al., 1995a; Zheng et al., 1998); (2) an excellent
correlation between the degree of Al resistance and
magnitude of AI-induced root malate release in differ-
ent wheat genotypes varying in Al resistance (ranging

from AI-sensitive to very AI-resistant) (Papemik et al.,
2000; Ryan et aI., 1995b); (3) the addition of organic
acids (malate, citrate or oxalate) to root bathing solu-
tions ameliorates Al toxicity in AI-sensitive varieties
(see, for example, Delhaize et aI., 1993b; Pellet et aI.,
1997; Zheng et aI., 1998); (4) complexes of Al with
di- and tricarboxylic organic acids are not transported
across membranes or absorbed by roots (Akeson and
Munns, 1989; Shi and Haug, 1990); (5) AI-induced
malate release genetically cosegregates with Al res-
istance and Al exclusion from the wheat root apex
(Delhaize et al., 1993a,b); (6) the rapid release of or-
ganic acid anions is consistent with the time frame for
the onset of Al resistance (see, for example, Ryan et
aI., 1995a); (7) over-expression of the bacterial cit-
rate synthase gene in tobacco and papaya resulted in
increased citrate levels in roots, increased citrate exud-
ation, and a significant increase in Al resistance (de la
Fuente et aI., 1997); and (8) an AI-gated anion channel
has been identified in protoplasts isolated from the root
apex of AI-resistant wheat and maize, which is a good
candidate to be the transport system facilitating AI-
induced organic acid anion release (Kollmeier et aI.,
2001; Pineros and Kochian, 2000; Ryan et aI., 1997;
Zhang et aI., 2000).
The role of anion channels in the release of organic
acid anions by roots
The numerous physiological studies investigating Al
resistance mechanisms based on AI-activated organic
acid anion exudation have provided strong evidence
that the transport of organic acid anions across the
root-cell plasma membrane and not organic acid syn-
thesis is the key, rate limiting step in this process.
In studies on AI-resistant and -sensitive genotypes of
both wheat and rye it has been clearly shown that Al
exposure only stimulates organic acid anion release
from the root apex and has no effect on root apical
organic acid concentrations. This stimulation occurs
primarily in the resistant genotypes (Delhaize et aI.,
1993b; Li et aI., 2000; Ryan et al., 1995, 2001).
Furthermore, the AI-activated organic acid anion ex-
udation continues for many hours without significant
changes in any key enzymes involved in organic acid
synthesis (PEP carboxylase, malate dehydrogenase,
isocitrate dehydrogenase, citrate synthase). Based on
studies such as these, it is widely accepted that Al
is activating a plasma membrane organic acid anion
transporter, and this transporter plays a central role
111
AI A'IiYlrioo orR""I Organic Add Ex"dlUon
Apoplasm
(pH 4.0-4.5)
Figure 1. Possible mechanisms of AI-activated organic acid anion
release involved in Al resistance via Al exclusion from the root
apex. In this model, Al activation of an anion channel is proposed to
be central to this resistance mechanism. This can be accomplished
by: (1) the binding of extracellular A13+ directly to the channel,
which effects opening; (2) AI3+ binds to an unknown receptor in the
plasma membrane, and mediates channel opening through a mem-
brane-localized signal pathway; and (3) A13+ enters the cell and
triggers opening through a signal transduction cascade that involves
both cytoplasmic and plasma membrane components.
in Al resistance (Delhaize and Ryan, 1995; Kochian,
1995; Kochian and Jones, 1997; Ryan et aI., 2001).
Organic acids in the cytoplasm (pH 7) will be
largely deprotonated and exist as anions, such that
the AI-activated system involved in organic acid anion
release and plant Al resistance likely involves an an-
ion transporter. The thermodynamics of both inorganic
and organic anion transport across the root-cell plasma
membrane are such that there is a large outwardly dir-
ected electrochemical gradient for anions. Therefore,
activation (i.e., opening) of plasma membrane anion
channels will result in a large anion efflux, and it is
likely that anion channels (permeable to organic acid
anions) will constitute the transport mechanism via
which AI-induced organic acid anion exudation oc-
curs. As mentioned above, Ryan and co-workers have
used the patch clamp technique on protoplasts isolated
from the root apex of the AI-resistant wheat isoline ex-
pressing Altl to identify a novel AI3+ -activated anion
channel that could be the release pathway for malate
exudation involved in Al resistance (Ryan et aI., 1997).
More recently, they showed that this channel me-
diates malate transport in wheat roots and is inhibited
by compounds that block both members of the CLC
and ABC gene families of ion transporters (Zhang
et aI., 2001). They also showed that although this
AI-activated anion current occurs in protoplasts from
the root apex of both AI-resistant and -sensitive gen-
otypes, the current occurred much more frequently,
112
exhibited a greater anion flux, and remained active
longer in protoplasts isolated from the AI-resistant
isoline (Zhang et aI., 2001). These findings indic-
ate that the same Al resistance machinery appears to
operate in both AI-resistant and -sensitive genotypes,
but the system is more active or abundant in the AI-
resistant plants. As shown in Figure 1, there are three
possible scenarios for Al regulating this anion chan-
nel that mediates organic acid anion efflux: (1) AI3+
binds directly to the channel and effects opening; (2)
AI3+ binds to an unknown receptor in the plasma
membrane, and mediates channel opening through
a membrane-localized signal pathway; and (3) AI3+
enters the cell and triggers opening through a signal
transduction cascade that involves both cytoplasmic
and plasma membrane components.
Recent work from our lab and from the labs of
Horst and Hedrich have identified an AI-gated anion
channel in the plasma membrane of protoplasts isol-
ated from AI-resistant maize, where Al is excluded
from the root tip via AI-induced citrate release (Koll-
meier et aI., 2001; Pineros and Kochian, 2001). In
the work by Kollmeier and colleagues, it was shown
that this AI-activated channel mediates the transport
of citrate, and is more active in root tip cells from
AI-resistant versus AI-sensitive maize. In the work by
Pineros and Kochian the potentially important result
that AI-activation of the anion channel was observed in
excised patches of the plasma membrane was presen-
ted. This finding indicates that the Al responsive
machinery is localized to the plasma membrane, and
either directly involves gating of the anion channel
by Al or acts via a closely associated membrane re-
ceptor. These findings suggest that research should be
directed to both plasma membrane organic acid an-
ion transporters as well as regulatory proteins which
may be closely associated with these transporters, as
candidates for Al resistance genes.
Genetics and molecular biology ofAl resistance
Significant variation for Al resistance is well known in
many plant species and has led to a number of studies
of inheritance for Al resistance (Foy, 1988). Wheat has
been a widely studied crop species with regards to the
genetics of Al resistance, and in wheat Al resistance
is often monogenic (Camargo, 1984; Delhaize et aI.,
1993a; Kerridge and Kronstad, 1968; Riede and An-
derson, 1996; Somers and Gustafson, 1995). However,
there is also evidence to suggest that in certain very
AI-resistant wheat cultivars, this trait is controlled by
two genes (Berzonsky, 1992; Camargo, 1981). In our
laboratories, a recent genetic analysis of Al resistance
in barley (Tang et aI., 2000) and sorghum (Magalhaes,
Garvin, and Kochian, unpublished results) indicates
that as in wheat, Al resistance appears to be a genetic-
ally simple trait. However, in certain other grain crop
species, such as maize and rice, Al resistance appears
to be a more complex character. The findings from
several laboratories have indicated that in maize, Al
resistance is conferred by multiple genes (Magnavaca
et aI., 1987; Sawazaki and Furlani, 1986); recent work
on Al resistance in rice indicates a similar level of
genetic complexity (Nguyen et al., 2001; Wu et al.,
2000).
AI-inducible resistance genes
Many researchers have assumed that expression of
Al resistance genes is induced by Al exposure, and
this assumption has been the driving force for sev-
eral studies in wheat aimed at cloning Al resistance
genes. However, the considerable body of physiolo-
gical information concerning Al resistance in wheat,
which involves AI-activation of root apical malate ex-
udation, suggests that Al resistance is not inducible.
In wheat, all of the biochemical machinery for root
Al exclusion via malate release appears to be in place
before exposure to AI, and Al exposure appears to trig-
ger this response at the level of protein activity rather
than at the level of gene expression. Thus, attempts to
clone AI-inducible Al resistance genes have identified
stress response genes that are induced equally well
in AI-resistant and -sensitive genotypes, usually well
after the phenotypic expression of resistance is seen
(see, for example, Richards et aI., 1994; Snowden and
Gardner, 1993).
In contrast to wheat there appears to be a different
pattern of AI-activated organic acid anion release in
certain other plant species, suggesting that Al resist-
ance can also be an inducible process. In AI-resistant
genotypes of rye and Cassia tora, there is a lag of up
to 12 h between Al exposure and organic acid anion
release, and the rate of organic acid anion release con-
tinues to increase upon continued exposure to Al (Li
et aI., 2000; Ma, 2000; Ma et aI., 1997b). Based on
these observations, the authors have speculated that Al
induction of resistance genes occurs. Our own work on
Al resistance in maize and sorghum strongly suggests
that Al resistance is inducible in both of these crop
species. In our laboratory, it was shown that Al res-
istance in maize involves Al activation of organic acid

Aluminum Resistance in Sorghum
Period of AI exposure (days)
Figure 2. Influence of Al exposure (39 fLM A13+ activity in full nutrient solution) on root growth in AI-resistant and -sensitive parental lines
of sorghum. Root growth is presented as per cent of control root growth measured in the absence of Al.
anion release (citrate) and Al exclusion from the root
apex (Pellet et aI., 1995). Al exposure rapidly triggers
a low level of citrate exudation localized to the root
apex of AI-resistant but not AI-sensitive maize geno-
types. As shown in Pellet et al. (1995), upon continued
Al exposure, the rate of citrate exudation increases
continuously over the first 48 h of Al exposure. After
48 h the rate of citrate release levels off to a constant
high rate. These physiologically based experiments
suggest there is a constitutive level of Al resistance in
maize; superimposed upon this is a resistance mechan-
ism that is activated by Al exposure and could involve
AI-induced expression of resistance genes.
In our recent work with sorghum, it was found that
Al resistance is due to Al exclusion from the root apex
mediated by AI-activated citrate release. As shown in
Figure 2, continued exposure to Al in hydroponic me-
dia over a 6-day period induces an increasing level of
Al resistance in the resistant parent, as measured by
root growth in Al compared to control (-AI) plants.
These findings indicate that as in maize, Al resistance
appears to be AI-inducible in sorghum.
A primary goal for the ongoing Al resistance re-
search in a number of laboratories is the isolation and
characterization of Al resistance genes. This will be
important both in gaining a complete understanding of
this potentially complex trait, and also for use in im-
proving the Al resistance of a number of crop species.
Based on what is already known about the physiology
and genetics of Al resistance, several different ap-
proaches are being taken by researchers attempting to
isolate Al resistance genes. First, the recent identific-
ation of a major Al resistance mechanism involving
113
--+-- AI Resistant Line
--- AI Sensitive Line
AI-activation of organic acid anion efflux has provided
several targets for candidate genes, including root
plasma membrane organic acid anion transporters as
well as other proteins that may interact with these
transporters. As physiological and molecular genetic
research on Al resistance in rice and sorghum moves
forward, it may be possible to use map-based cloning
approaches to identify Al resistance genes in these two
model plant species. Finally, the possibility that resist-
ance genes may be AI-inducible in some crop plant
species suggests that genomics-based approaches may
provide an additional avenue for the cloning of Al
resistance genes.
Heavy metal hyperaccumulation and resistance
Terrestrial plants have evolved sophisticated strategies
for the acquisition of relatively unavailable micronu-
trients such as Zn, Mn, Cu, Fe and Ni from the soil.
As these essential micronutrients are also highly re-
active and potentially toxic to plants, micronutrient
uptake, transport and accumulation is by necessity
highly coordinated and regulated. Because many mi-
cronutrients are also heavy metals, contamination of
the soil environment with heavy metals is in reality
the accumulation of high levels of either essential mi-
cronutrients (Zn, Mn, Cu, Ni), or metals that can
act as micronutrient analogues (such as Cd, Pb, or
Hg). Recently, progress has been made in elucidating
the molecular and physiological mechanisms of plant
micronutrientlheavy metal accumulation and homeo-
stasis. There are a number oflabs throughout the world
114
conducting research on plant heavy metal accumu-
lation and resistance using metal hyperaccumulating
plant species as model plants. The goal of this type
of research is to provide the basic understanding and
molecular tools that ultimately can be used to de-
velop improved plant species for the remediation of
metal-contaminated soils.
Contamination of soils with heavy metals is a ser-
ious worldwide problem both for human health and
agriculture (Gairola et aI., 1992; Mazess and Barden,
1991; Ryan et aI., 1982). Cleanup of hazardous wastes
by the currently used engineering-based technologies
has been estimated to cost $400 billion dollars in the
U.S. alone (Salt et aI., 1995). Recently, there has been
considerable interest in the use of terrestrial plants
as an alternative, 'green technology' for the phytore-
mediation of surface soils contaminated with toxic
heavy metals (Chaney, 1983; Cunningham and Ow,
1996; Raskin et aI., 1997; Salt et aI., 1995, 1998).
A major factor behind the recent interest in phytore-
mediation of metal polluted soils has been the growing
awareness by the scientific community of the existence
of a number of metal hyperaccumulating plant species.
Over 200 terrestrial species have been reported that are
endemic to metalliferous soils and can tolerate and ac-
cumulate high levels of heavy metals such as Zn, Cd,
Cu, and Ni in their shoots. These plants were coined
hyperaccumulators by Brooks et aI. (1977). The ex-
istence of these interesting metal hyperaccumulator
species suggests that the genetic potential exists for
phytoremediation to be successful. Most of these hy-
peraccumulator species, however, are small and slow
growing, and because they produce limited shoot bio-
mass their potential for large-scale decontamination of
polluted soils is limited (Ebbs et aI., 1997). Trans-
ferring the genes expressing the hyperaccumulating
phenotype to higher shoot biomass-producing plants
has been suggested as an avenue for enhancing the
potential of phytoremediation as a viable commercial
technology (Brown et aI., 1995a). Progress towards
this goal, however, has been hindered by a lack of
understanding of the basic molecular, biochemical and
physiological mechanisms involved in heavy metal
hyperaccumulation.
One of the best known metal hyperaccumulat-
ors is Thlaspi caerulescens, which is a member of
the Brassicaceae family and a Cd/Zn hyperaccumu-
lator. Certain ecotypes of Thlaspi caerulescens have
been shown to accumulate and tolerate up to 3000
p,g g-l DW Cd in the shoots (typical shoot levels
are between 0.1 and 10 p,g g-l DW) and 40000 p,g
g-l DW Zn (normal foliar Zn levels for hydroponic-
ally grown plants are around 100-200 fLg g-l DW,
while 30 fLg g-l DW is considered adequate) (Brown
et aI., 1995a,b). Additionally, certain ecotypes of T.
caerulescens have been reported to accumulate high
levels of other metals, including Ni and Co (Baker,
1981; Baker and Brooks, 1989; Brown et aI., 1995b).
The unique physiology of heavy metal transport and
resistance in Thlaspi caerulescens makes it a very in-
teresting experimental system for basic research aimed
at elucidating plant mechanisms of heavy metal hyper-
accumulation.
Physiology oJZn2+ transport in hyperaccumulator
and non-accumulator species oJThlaspi
We initially conducted physiological studies that fo-
cused on the use of radiotracer flux techniques
( 65 Zn2+) to characterize Zn transport and compart-
mentation in Thlaspi caerulescens and a related non-
accumulator, Thlaspi arvense. These studies indicated
that a number of ZrrlCd transport sites in the root
and shoot contribute to the hyperaccumulation trait
in T. caerulescens. As detailed in Lasat et aI. (1996,
1998) these include: (1) a stimulated metal influx
across the root-cell plasma membrane; (2) reduced
metal sequestration in the root vacuole; (3) increased
xylem-localized metal loading into xylem and sub-
sequent translocation to the shoot; and (4) stimulated
metal influx across the leaf cell plasma membrane and
enhanced storage in the leaf vacuole. The stimulated
ZrrlCd influx across the root-cell plasma membrane
was studied in the greatest detail and it was shown
that root Zn absorption is mediated by a saturable,
high affinity Zn2+ transporter while the concentration-
dependent kinetics for Cd influx were linear and non-
saturating. The saturating Zn influx system had a
similar affinity for Zn2+ in the two Thlaspi species
(Km for root Zn2+ uptake was 6 and 8 fLM, in T.
caerulescens and T. arvense, respectively). However,
there was a 5-fold larger Vmax for root Zn uptake in T.
caerulescens compared with T. arvense (Lasat et aI.,
1996). While root Cd influx followed linear, first-order
concentration dependent kinetics in the two Thlaspi
species, it was 2-3-fold higher in T. caerulescens.
These findings suggest that particularly the increased
saturable Zn uptake in T. caerulescens could be due
to a higher density of Zn transporters in the root-cell
plasma membrane.

Molecular basis of zinc hyperaccumulation in Thlaspi
caerulescens
A molecular characterization of plant heavy metal hy-
peraccumulation was initiated by cloning a Zn trans-
porter cDNA from T. caerulescens via functional com-
plementation in yeast. The Saccharomyces cerevisiae
mutant, ZHY3, defective in the high and low affinity
Zn transporters, ZRTI and ZRT2, has a much higher
Zn requirement for growth than the parental wild
type yeast (Zhao and Eide, 1997). The ZHY3 strain
was transformed with a T. caerulescens cDNA lib-
rary constructed in the yeast expression vector, pFL61.
Screening of 350000 yeast transformants for growth
on low Zn media resulted in the identification of ZNTl
(for Zn !ransporter) (see Pence et aI., 2000).
Expression of ZNTl in ZHY3 restored growth on
low Zn media to that of the parental wild type yeast.
The predicted open reading frame for ZNTl is 379
amino acids in length and demonstrates significant se-
quence identity with the Arabidopsis genes ZlP4 and
lRT1, which encode putative Zn and Fe transporters,
respectively (Eide et aI., 1996; Grotz et aI., 1998)
and are members of the ZIP family of micronutrient
transport proteins (Eng et aI., 1998; Guerinot, 2000).
ZNTl shares the structural features exhibited by other
members of this family, including eight putative trans-
membrane domains and a highly hydrophilic cytoplas-
mic region predicted to reside between transmembrane
domains three and four. This putative cytoplasmic do-
main contains a series of histidine repeats, which may
define a metal binding region for the transporter. The
similarities in predicted amino acid sequence and pro-
tein structure between ZNTI and other members of
the ZIP family suggest that ZNTl is an integral mem-
brane protein that mediates Zn2+ transport across a
cell membrane.
To test the hypothesis that ZNTI is a Zn transport
protein, ZNTl was expressed in yeast (ZHY3) for ra-
diotracer (65 Zn2+ and l09Cd2 +) flux experiments that
were used to determine the concentration dependent
kinetics of Zn2+ and Cd2+ influx mediated by ZNTl
in ZHY3. We found that ZNTl mediated saturable Zn
uptake that conformed to Michaelis-Menten kinetics
with a Krn of 7.5 p,M, as well as low affinity Cd2+
influx in yeast that follows first-order (linear) transport
kinetics (Pence et aI., 2000). The kinetic properties
for Zn2+ and Cd2 + uptake mediated by ZNTl in yeast
are very similar to what we have previously seen for
Zn2+ and Cd2+ uptake in T. caerulescens roots (Las at
et aI., 1996). These results are consistent with the hy-
115
Tc Ta
A [Zn) (11M) in growth solution
o 1 50 0 1 10
ZNTI
ZNTl-arvense
B Root Zn transport kinetic parameters
4 6 5 2 2 2 K...
244 271 76 80 43 43 Vmax
Figure 3. Influence of plant Zn status on ZNTl expression and
root Zn2 + uptake. Seedlings of T. caerulescens (lanes labeled Tc)
and T. arvense (lanes labeled Ta) were grown in nutrient solution
containing 0, 1 and 10 or 50 fLM Zn. Root RNA was isolated
for Northern analysis and hybridized to gene-specific probes for
ZNTl in T. caerulescens and for the ZNTl homolog in T. arvense.
Also, 65Zn flux studies were conducted to determine the concentra-
tion-dependent kinetics of root Zn uptake in the Thlaspi seedlings.
The Km and Vmax for high affinity 65Zn uptake are presented below
the Northern.
pothesis that ZNTl encodes a root plasma membrane
Zn2+ /Cd2+ transporter.
A 5-fold increase in the Vrnax for root Zn2+ in-
flux in T. caerulescens compared with T. arvense in an
earlier study led us to speculate that there are a greater
number of Zn transporters per unit area of root-cell
plasma membrane in the hyperaccumulator (Las at et
aI., 1996). In order to test this further, seedlings of
both Thlaspi species were grown under Zn deficient
(0 p,M), 'normal' Zn conditions (1 p,M) and high Zn
conditions (50 p,M Zn for T. caerulescens and 10 p,M
Zn for T. arvense). Then, the concentration-dependent
kinetics of root Zn2+ influx and ZNTl Northern ana-
lysis using RNA isolated from roots and shoots of both
Thlaspi species grown under all three Zn treatments
was conducted (Figure 3 and Pence et aI, 2000). When
gene specific probes for ZNTl or its homolog from
T. arvense were used, we found that ZNTl transcript
abundance was dramatically higher in roots and shoots
of T. caerulescens grown under 'normal' Zn (l p,M)
and Zn-deficient conditions (0 p,M) as compared with
T. arvense. This is consistent with the hypothesis that
Zn hyperaccumulation in T. caerulescens is due, in
part, to increased expression of Zn transporters in the
root and shoot.
The root Zn transport data and ZNTl transcript
abundance as a function of seedling Zn status provides
116
Table 1. ZIP family micronutrient transporters cloned in Thlaspi caerulescens
Thlaspi caerulescens Arabidopsis homolog
ZNT1 ZIP4(U95973)
ZNT2 ZIP2(AF033536)
ZNT3 ZIP3(AF033537)
ZNT4 IRT3(AAF27669)
ZNT5 ZIP5(AAB7l447)
ZIP 1, 7 and 9 have been cloned in Arabidopsis and are being used to isolate T. caerulescens homologs.
insights into the regulation of Zn transporters in hy-
peraccumulator and nonaccumulator plants. A close
correlation between ZNTl expression (Figure 3A) and
the Vmax for root Zn2+ influx (Figure 3B) was found
in both Thlaspi species. In T. arvense, growth on 'nor-
mal' (1 fLM) or high Zn (10 fLM) had no effect on the
low level of root ZNTl expression or the small root
Zn2+ influx that was observed (Vmax of 43 nmol gm h).
Only when T. arvense plants were made Zn-deficient,
was the moderate increase in ZNTI expression and
root Zn2+ influx seen (increase in Vmax to 80 nmol
gm- l h- l ). Quantification of root transcript abund-
ance from the data in the Northern blot indicated that
Zn deficiency caused a 2-fold increase in T. arvense
mRNA abundance, which correlates with the 2-fold
enhancement of Vmax .
The response of root Zn uptake to changes in plant
Zn status in T. caerulescens were found to be qual-
itatively similar to the responses in T. anJense when
seedlings were grown in a wide range of Zn concen-
trations in the nutrient solution (0-500 fLM Zn). That
is, T. caerulescens seedlings grown in 0 and 1 fLM Zn
had very high level of ZNTl expression as well as a
considerably larger Vmax for root Zn2+ influx (Vmax
values of 244 and 271) in comparison with T. arvense.
However, when T. caerulescens seedlings were grown
on levels of Zn ranging from 50 to 500 fLM (which
are comparable to levels of available Zn2+ for Zn-
contaminated soils), a significant down-regulation in
ZNTI expression and reduction in root Zn2+ uptake
were observed (Vmax reduced to 76 nmol gm- l h- 1
and a 6-fold reduction in root mRNA abundance). Al-
though growth on high Zn reduced ZNTl expression
and Zn uptake in T. caerulescens, they were still 4-
and 2-fold higher, respectively, than in Zn-sufficient
T. arvense. Thus, it appears that an alteration in the
regulation of Zn transport by Zn status, and not a con-
stitutive increase in Zn transporter gene expression,
plays a role in Zn hyperaccumulation.
% identity between Length in T. caerulescens
T. caerulescens and A. thaliana
88 Full length
88 Mising 28 bp at 5' end
87 Missing 200 bp at 5' end
86 Full length
86 Full length
If a Zn responsive regulatory scheme similar to
that in yeast exists in higher plants, how might it
be altered to cause the enhanced Zn transporter gene
expression and Zn hyperaccumulation observed in T.
caerulescens? One possibility involves a mutation in a
putative Zn responsive transcriptional activator, which
would alter Zn-dependent down-regulation of ZNTl
expression. Such a mutation in a Zn-responsive tran-
scriptional activator, ZAP1, has been isolated in yeast
(Zhao and Eide, 1997; Zhao et aI., 1998). The semi-
dominant mutant allele, ZAP1-1 uP, results from a
substitution of a serine for a cysteine residue in the
N terminal region, and causes high level of expression
of the yeast Zn transporters under Zn replete condi-
tions. Thus, specific alterations in a Zn-responsive
transcriptional activator or Zn- responsive elements
in transporter gene promoters possibly play an im-
portant role in heavy metal hyperaccumulation in T.
caerulescens.
Involvement of multiple micronutrient transporter
genes in metal hyperaccumulation
The results presented above suggest that ZNTl could
play an important role in the root and leaf metal trans-
port processes resulting in hyperaccumulation, based
on the correlations between the magnitude of root
Zn2+ influx and ZNTl expression. We have found
using in situ RT-PCR techniques that ZNTI expres-
sion is localized to the root cortex, further supporting
a role for this transporter in root metal absorption
from the soil (Letham et aI., 2000). Information from
the Arabidopsis genome sequencing efforts as well as
advances from other laboratories (Eng et aI., 1998;
Grotz et aI., 1999) have shown that there are a large
number of members of the ZIPIIRT family of mi-
cronutrient transporters in Arabidopsis including 14
possible ZIPs and three IRTs. Indications of a sim-
ilarly large transport gene family in T. caerulescens

came from Southern analysis of ZNTI under low strin-
gency, which yielded a complex pattern indicative of
ZNTl hybridization to other members of the ZIP fam-
ily in Thlaspi (data not shown). As it appears that
ZNTI plays an important role in heavy metal hyperac-
cumulation in T. caerulescens, the existence of other
closely related genes certainly raises the possibility
that other members of this family also are important
to metal hyperaccumulation. To get a start on this area
of research we have cloned four other T. caerulescens
homologs of the ZIP gene family and are in the pro-
cess of isolating three more; these are summarized in
Table 1.
From the findings presented in Figure 3, we
wondered whether in addition to ZNTl, other T. caer-
ulescens genes involved in metal hyperaccumulation
may be expressed to much higher levels in T. cae rules-
cens compared with the nonaccumulator, T. arvense.
Northern analysis has been conducted with some of
the other ZIP genes as well as other heavy metal-
related genes. It was found that at least two other
members of the ZIP family of metal transporters in
T. caerulescens as well as homo logs of Arabidop-
sis metallothionein and Nramp genes were expressed
to much higher levels in T. caerulescens compared
with T. arvense. It should be noted that when the
blot was probed with three different AtNramp genes,
only a homolog to AtNramp2 exhibited increased tran-
script abundance, suggesting that overexpression is
specific to particular genes in individual heavy metal
transporter gene families.
The ZIP micronutrient transporter gene family in
higher plants could be regulated by a Zn-dependent
transcriptional activator as has been shown for yeast,
where the transcription factor ZAP 1 interacts with
Zn-responsive elements in the promoters of the ZIP
homologs, ZRTI-3 (Zhao and Eide, 1997; Zhao et aI.,
1998). The results described above indicating that at
least one Nramp gene also is up regulated in T. caer-
ulescens suggests that in plants, a Zn-dependent factor
could also regulate expression of genes from other,
'non-ZIP' transporter gene families. Further support
for this possibility comes from our recent observation
that expression of at least one ZAT homolog (a putat-
ive vacuolar Zn transporter from the Cation Diffusion
Facilitator [CDF] family of micronutrient transport-
ers) is up regulated in T. caerulescens. Also, work
from Dr. Eide's laboratory has shown that in yeast,
ZAP 1 can regulate the expression of a number of
different genes in a Zn-dependent manner, including
the ZRCI gene which encodes a vacuolar Znlheavy
117
metal transporter from the CDF family of transport-
ers (Lyons et aI., 2000). Taken together, all of these
pieces of evidence provide circumstantial support for
a model of heavy metal hyperaccumulation in T. caer-
ulescens that involves alterations in the 'normal' reg-
ulation of micronutrient homeostasis that occurs in
non-accumulator plant species.
Conclusions
In this paper, we have described specific mechanisms
that plants employ to deal with toxic levels of alu-
minum and heavy metals in the soil. Based on an ana-
lysis of the current literature, quite different strategies
appear to be used for Al and heavy metal resistance.
For AI, most of the evidence points to an Al resistance
mechanism based on exclusion of Al from the root
apex. This involves AI-activation of a transporter loc-
alized to the root-cell plasma membrane that mediates
the release of organic acid anions into the rhizosphere.
These organic anions complex and detoxify AI3+ in
the soil. For heavy metals, a totally different strategy
based on extreme tolerance and metal hyperaccumu-
lation was described for a hyperaccumulator plant
species that has evolved on naturally metalliferous
soils. The reason these two strategies were the focus
of this paper was that currently, they are the best un-
derstood mechanisms of metal resistance in terrestrial
plants. Ho\vever, it is likely that other mechanisms of
Al and/or heavy metal resistance are also operating
in certain plant species, and there may be common
features shared for dealing with Al and heavy metals.
For example, a second Al resistance mechanism has
recently been described in hydrangea and buckwheat
that involves internal detoxification of accumulated Al
by organic acids (citrate and oxalate) (Ma et aI., 1997;
1998). This mechanism allows these plant species to
accumulate Al in their leaves to quite high levels in
hydrangea (3000 ppm) and moderately high levels in
buckwheat (450 ppm). In comparison, plant species
such as wheat, which employ the Al exclusion mech-
anism described in this paper, accumulate less than 50
ppm Al in their leaves (Ma, 2000). It is likely that
future research on plant metal resistance will uncover
novel mechanisms of metal tolerance that currently
have not been described.
118
References
Akeson M and Munns D N 1989 Lipid bilayer permeation by neutral
aluminum citrate and by three alpha hydroxy carboxylic acids.
Biochim. Biophys. Acta 984, 200-206.
Baker A J M 1981 Accumulators and excluders-strategies in the
response of plants to heavy metals. J. Plant Nutr. 3, 643-654.
Baker A J M and Brooks R R 1989 Terrestrial higher plants which
hyperaccumulate metallic elements. Biorecovery 1, 81-97.
Beckmann I 1976 Cultivation and breeding of wheat in the south of
Brazil. Proceedings of Workshop on Plant Adaptation to Mineral
Stress in Problem Soil. pp 409-416. Beltsville, MD.
Berzonsky W A 1992 The genomic inheritance of aluminum resist-
ance in 'Atlas 66' wheat. Genome 35, 689-693.
Brooks R R, Lee J, Reeves R and Jaffre T 1977 Detection of nick-
e1iferous rocks by analysis of herbarium specimens of indicator
plants. J. Geochem. Explor. 7,49-58.
Brown S L, Chaney R L, Angle J S and Baker A J M 1995a Zinc and
cadmium uptake by hyperaccumulator Thlaspi caerulescens and
metal tolerant Silene vulgaris grown on sludge-amended soils.
Environ. Sci. Technol. 29, 1581-1585.
Brown S L, Chaney R L, Angle J S and Baker A J M 1995b Zinc
and cadmium uptake by hyperaccumulator Thlaspi caerulescens
grown in nutrient solution. Soil Sci. Soc. Am. J. 59, 125-133.
Camargo CEO 1981 Me1horamento do trigo. I. Hereditariedade do
tolerancia a toxicidade do aluminio. Bragantia 40, 33-45.
Camargo CEO 1984 Melhoramento do trigo. VI. Hereditar-
iedade de tolerancia a tres concentracoes de aluminio em solucao
nutritiva. Bragantia 43, 279-291.
Chaney R L 1983 Plant uptake of inorganic waste constituents. In
Land Treatment of Hazardous Wastes. Eds. J F Parr, P B Marsh
and J M KIa. pp 50-76. Noyes Data Corp, Park Ridge, NJ.
Cunningham S D and Ow D W 1996 Promise and prospects of
phytoremediation. Plant Physiol. 110,715-719.
De 1a Fuente J M, Ramirez-Rodriguez V, Cabrera-Ponce J Land
Herrera-Estrella L 1997 Science 276, 1566-1568.
Delhaize E and Ryan P R 1995 Aluminum toxicity and resistance in
plants. Plant Physiol. 107,315-321.
Deihaize E, Craig S, Beaton CD, Bennet R J, Jagadish V C and
Randall P J 1993a Aluminum resistance in wheat (Triticum aes-
tivum L.) I. Uptake and distribution of aluminum in root apices.
Plant Physiol. 103, 685-693.
Delhaize E, Ryan P R and Randall P J 1993b Aluminum resist-
ance in wheat (Triticum aestivum L.): II. Aluminum-stimulated
excretion of malic acid from root apices. Plant Physiol. 103,
695-702.
Ebbs S D, Lasat M M, Brady D J, Cornish J, Gordon R and Ko-
chian L V 1997 Phytoextraction of cadmium and zinc from a
contaminated site. J. Environ. Qual. 26, 1424-1430.
Eide D, Broderius M, Fett J and Guerinot M L 1996 A novel iron-
regulated metal transporter from plants identified by functional
expression in yeast. Proc. Natl. Acad. Sci. USA 93, 5624-28.
Eng B H, Guerinot M L, Eide D and Saier M H 1998 Sequence
analyses and phylogenetic characterization of the ZIP family of
metal ion transport proteins. J. Membr. BioI. 166, 1-7.
Foy C D 1988 Plant adaptation to acid, aluminum-toxic soils.
Commun. Soil Sci. Plant Anal. 19,959-987.
Gairo1a C G, Wagner G J and Diana J N 1992 Tobacco, Cd and
health. J Smoking ReI. Dis. 3, 3-6.
Grotz N, Fox T, Connolly E, Park W, Guerinot M and Eide D 1998
Identification of a family of zinc transporter genes from Ara-
bidopsis that respond to zinc deficiency. Proc. Natl. Acad. Sci.
USA 95,7220-7224.
Guerinot M L 2000 The ZIP family of metal transporters. Biochim.
Biophys. Acta 1465, 190-198.
Kerridge P C and Kronstad W E 1968 Evidence of genetic resistance
to aluminum toxicity in wheat (Triticum aestivum ViiI., Host).
Crop Sci. 60, 710-711.
Kochian L V 1995 Cellular mechanisms of aluminum toxicity and
resistance in plants. Annu. Rev. Plant Physiol. Plant Mol. BioI.
46, 237-260.
Kochian L V and Jones D L 1997 Aluminum toxicity and resistance
in plants. In Research Issues in Aluminum Toxicity. Eds. R Yokel
and M S Golub. pp 69-90. Taylor and Francis, Washington, DC.
Kollmeier M, Dietrich M, Bauer C S, Horst W J and Hedrich R
2001 Aluminum activates a citrate-permeable anion channel in
the aluminum-sensitive zone of the maize root apex. A compar-
ison between an aluminum-sensitive and an aluminum-resistant
cu1tivar. Plant Physiol. 126, 397-410.
Lasat M M, Baker A J M and Kochian L V 1996 Physiological char-
acterization of root Zn2+ absorption and translocation to shoots
in Zn hyperaccumulator and nonaccumulator species of Thlaspi.
Plant Physiol. 112, 1715-1722.
Lasat M M, Baker A J M and Kochian L V 1998 Altered zinc
compartmentation in the root symplasm and stimulated Zn 2+
absorption into the leaf as mechanisms involved in zinc hy-
peraccumulation in Thlaspi caerulescens. Plant Physiol. 118,
875-883.
Letham D L D, Pence N S, Pineros M A, Papoian A and Kochian
L V 2000 Molecular characterization of ZnlCd uptake in the hy-
peraccumulator Thlaspi caerulescens including in situ RT-PCR
localization and characterization of the heavy metal transporter
ZNT1. Plant Physiol. Abstracts, p. 156.
Li X F, Ma J F and Matsumoto H 2000 Pattern of aluminum-induced
secretion of organic acids differs between rye and wheat. Plant
Physiol. 123, 1537-1544.
Lyons T J, Gasch A, Gaither L A, Botstein D, Brown P 0 and
Eide D 2000 Genome-wide characterization of the Zaplp zinc-
responsive regulon in yeast. Proc. Natl. Acad. Sci. USA 97,
7957-7962.
Ma J F, Zheng S J, Hiradate S and Matsumoto H 1997a Detoxifying
aluminum witli buckwheat. Nature 390, 569-570.
Ma J F, Zheng S J and Matsumoto H 1997b Specific secretion of cit-
ric acid induced by Al stress in Cassia tara L. Plant Cell Physiol.
38,1019-1025.
Ma J F 2000 Role of organic acids in detoxification of aluminum in
higher plants. Plant Cell Physiol. 41, 383-390.
Ma Z and Miyasaka S C 1998 Oxalate exudation by taro in response
to aluminum. Plant Physiol. 118, 861.
Ma J F, Ryan P R and Delhaize E 2001 Aluminum tolerance in
plants and the complexing role of organic acids. Trends Plant
Sci. 6, 273-278.
Magnavaca R, Gardner C 0 and Clark R B 1987 Inheritance of alu-
minum resistance in maize. In Genetic Aspects of Plant Mineral
Nutrition. Eds. H W Gabe1man and B C Loughman. pp. 201-212.
Martinus Nijhoff, Dordrecht, The Netherlands.
Mazess R and Barden H 1991 Bone density in premenopausal wo-
men; effect of age, dietary intake, physical activity, smoking and
birth control pills. Am. J. C1in. Nutr. 53, 132-142.
Miyasaka S C, Buta J G, Howell R K and Foy C D 1991 Mechanism
of aluminum resistance in snapbeans. Root exudation of citric
acid. Plant Physiol. 96, 737-743.
Nguyen V T, Burow M D, Nguyen H T, Le B T and Paterson
A H 2001 Molecular mapping of genes conferring aluminum
resistance in rice (Oryza sativa L.). Theor. Appl. Genet. 102,
1002-1010.

Papemik L A, Bethea A S, Singleton T E, Magalhaes J V, Garvin
D F and Kochian L V 2001 Physiological basis of reduced AI
resistance in ditelosomic lines of Chinese Spring wheat. Planta
212,829-834.
Pellet D M, Grunes D L and Kochian L V 1995 Organic acid exud-
ation as a mechanism of AI-resistance in Zea mays. Planta 197,
788-795.
Pellet D M, Papemik L A and Kochian L V 1996 Multiple aluminum
resistance mechanisms in wheat: the role of root apical phosphate
and malate exudation. Plant Physiol. 112,591-597.
Pence N S, Larsen P B, Ebbs S D, Lasat M M, Letham D L D,
Garvin D F, Eide D and Kochian L V 2000 The molecular basis
for heavy metal hyperaccumulation in Thlaspi caerulescens.
Proc. Natl. Acad. Sci. USA 97, 4956-4960.
Pineros M A and Kochian L V 2001 A patch clamp study on
the physiology of aluminum toxicity and resistance in Zea
mays: identification and characterization of A1 3+ -induced anion
channels. Plant Physiol. 124, 1-14.
Raskin I, Smith R D and Salt D E 1997 Phytoremediation of metals:
using plants to remove pollutants from the environment. Curro
Opin. Biotech. 8,221-226.
Richards K D, Schott E J, Sharma Y K, Davis K R and Gardner R
C 1998 Aluminum induces oxidative stress genes in Arabidopsis
thaliana. Plant Physiol. 116,409-418.
Riede C R and Anderson J A 1996 Linkage of RFLP markers to an
aluminum resistance gene in wheat. Crop Sci. 36, 905-909.
Ryan J A, Pahren H R and Lucas J B 1982 Controlling cadmium
in the human food chain: a review and rationale based on health
effects. Environ. Res. 28, 251-302.
Ryan P R, DiTomaso J M and Kochian L V 1993 Aluminum toxicity
in roots: an investigation of spatial sensitivity and the role of the
root cap. J. Exp. Bot. 44, 437-446.
Ryan P R, Delhaize E and Randall P J 1995a Characterization of AI-
stimulated efflux of malate from the apices of AI-resistant wheat
roots. Planta 196, 103-110.
Ryan P R, Delhaize E and Randall P J 1995b Malate efflux from
root apices: evidence for a general mechanism of AI-resistance
in wheat. Aust. J. Plant Physiol. 22, 531-536.
Ryan P R, Skerrett M, Findlay G P ruld Dclhaize E 1997 Aluminum
activates an anion channel in the apical cells of wheat roots. Proc.
Natl. Acad. Sci. USA 94,6547-6552.
Ryan P R, Delhaize E and Jones D L 2001 Function and mechanism
of organic anion exudation from plant roots. Annu. Rev. Plant
Physiol. Plant Mol. BioI. 52,527-560.
Salt D E, Blaylock M, Nanda Kumar P B A, Dushenkov V, Ens-
119
ley B D, Chet I and Raskin I 1995 Phytoremediation: a novel
strategy for the removal of toxic metals from the environment
using plants. Biotechnology 13,468-474.
Salt D E, Smith R D and Raskin I 1998 Phytoremediation. Annu.
Rev. Plant Physiol. Plant Mol. BioI. 49, 643-668.
Sawazaki E and Furlani E R 1986 Genetica da tolerancia ao alu-
mino em linhagens de milho cateto. pp. 382-392. XVI Congresso
Nacional de Milho e Sorgo. Belo Horizonte, Brazil.
Shi Band Haug A 1990 Aluminum uptake by neuroblastoma cells.
J. Neurochem. 55,551-558.
Sivaguru M and Horst W J 1998 The distal part of the transition
zone is the most aluminum sensitive apical root zone of maize.
Plant Physiol. 116, 115-163.
Snowden K C and Gardner R C 1993 Five genes induced by alu-
minum in wheat (Triticum aestivum L.) roots. Plant Physiol. 103,
855-861.
Somers D J and Gustafson J P 1995 The expression of aluminum
stress induced polypeptides in a population segregating for alu-
minum resistance in wheat (Triticum aestivum L.). Genome 38,
1213-1220.
Tang Y, Sorrells M E, Kochian L V and Garvin D F 2000 Identifica-
tion of RFLP markers linked to barley aluminum resistance gene
Alp. Crop Sci. 40, 778-782.
von Uexkiill H R and Mutert E 1995 Global extent, development
and economic impact of acid soils. In Plant-Soil Interactions at
Low pH: Principles and Management. Eds. Date R A, Grundon
N J, Raymet G E and Probert M E. pp. 5-19. Kluwer Academic
Publishers, Dordrecht, The Netherlands.
Wu P, Liao C Y, Hu B, Yi K K, Jin W Z, Ni J J and He C 2000 QTLs
and epistasis for aluminum resistance in rice (Oryza sativa L.) at
different seedling stages. Theor. Appl. Genet. 100, 1295-1303.
Zhang W-H, Ryan P R and Tyerman S D 2001 Malate-permeable
channels and cation channels activated by aluminum in the apical
cells of wheat roots. Plant Physiol. 125, 1459-1472.
Zhao H and Eide D J 1997 Zap I p, a metalloregulatory protein
involved in zinc-responsive transcriptional regulation in Sacchar-
omyces cerevisiae. Mol. Cell BioI. 17,5044-5052.
Zhao H. Butler E, Rodgers J, Spizzo T, Duesterhoeft Sand Eide
D 1998 Regulation of zinc homeostasis in yeast by binding of
the ZAPI transcriptional activator to zinc-responsive promoter
elements. J. BioI. Chern. 273, 28713-28720.
Zheng S J, Ma J F and Matsumoto H 1998 High aluminum resist-
ance in buckwheat. I. AI-induced specific secretion of oxalic acid
from root tips. Plant Physiol. 117,745-751.
Nutrient acquisition: mechanisms and modelling
 Plant and Soil 247: 123-130, 2002.
123
© 2002 Kluwer Academic Publishers.

Use of modelling to understand nutrient acquisition by plants

G. J. D. Kirk 1,*
lInternational Rice Research Institute, MCPO Box 3127, 1271 Makati City, Philippines. Present address:
Department of Plant Sciences, University of Cambridge, Downing Street, Cambridge CB2 3EA, UK

Received 11 October 2001. Accepted in revised form 20 August 2002

Key words: modelling, nutrient acquisition, rhizosphere, rice

Abstract
The definition, testing and uses of explanatory models in plant nutrition are discussed and contrasted with descript-
ive, predictive models. Two examples are given. First a model of phosphorus acquisition by rice plants growing in
soils that undergo flooding and drainage, in which changes in the soil induced by roots are important. The changes
are, in flooded anaerobic soil, oxidation of mobile reductants and consequent production of H+ ions in the soil as
well as direct release of H+ ions from the roots to balance excess intake of cations over anions; and in non-flooded
aerobic soil, release of organic anions from the roots. The model's predictions are tested against experimental data.
Second a model of N acquisition by rice plants in flooded anaerobic soil, in which the uptake properties of the roots
are important. The contribution to N uptake of N03' formed in the rhizosphere of rice in flooded soil is compared
with that of NHt from the bulk soil using a model and absorption properties of rice roots measured in nutrient
culture. The models' uses and limitations are discussed.

Introduction ample, equations for the transport of a nutrient

Models can be used to understand complex systems, to
study experimentally inaccessible systems, as a frame-
work for discussion, especially across disciplines, and
to identify gaps in knowledge or understanding. In this
paper, I discuss the characteristics of explanatory mod-
els aimed at such uses - as distinct from descriptive,
predictive models with a practical aim - and illustrate
their use in understanding complex problems in plant
nutrition.
Explanatory models are distinguished from pre-
dictive models by their different structure and content
and by the requirements for their testing (Nye, 1992).
All models contain:
1. Facts. For example, in models of nutrient trans-
fer through soil into roots, the initial concentration
of the nutrient in the soil and the soil moisture
content.
2. Mechanistic relations, derived from first principles
and the laws of physics and chemistry. For ex-
* FAX No: +44-122-333-3953.
E-mail: gjdk2@cam.ac.uk
through soil by mass flow and diffusion.
3. Empirical relations. For example, Michaelis-
Menten equations for the relation between the
influx of a nutrient into a root and its concentration
in solution at the root surface, or for the dynam-
ics of microbial populations in the rhizosphere.
These are necessary because it is not possible
to describe biological systems in purely mechan-
istic terms. Preferably, any empirical relations are
derived from lower order, mechanistic models.
In explanatory models mechanistic relations dominate,
and it is important that any empirical inputs are derived
as far as possible independent of the model's output.
Otherwise the whole exercise becomes an elaborate
form of curve fitting. As such the model may be a use-
ful summary of empirical knowledge, but it cannot be
used to judge the truth of any supposed mechanisms.
Nye (1992) discusses the different requirements for
testing predictive and explanatory models and defines
separate terms to distinguish them: predictive mod-
els are merely verified by comparing their output with
available empirical data, whereas explanatory models
124

35
~~ 1.4
o Flooded ~

"0 - - Flooded
v Moist '~
CD 30
0 1.2 ..... Moist
E o "0
C
~
c 1.0 ~ 25 Flooded then moist
'iii OJ o Flooded
I!! g 20 "V Moist
.S 0.8
!!l
c o Flooded then moist
"0
.!! III o
III
:;
0.6
Mt =4.93x 10. 10 X ..Jt Q. 15
E ~
0.4 ~
~
u
CD
.>< 10
U III
III
0.2 15.
a.. ~
5 0
a..
0.0
0 B --'T"
0 2 4 6 8 10
0 10 20 30 40
Time (days)
Time following planting (days)
Figure 1. Time courses ofP accumulation in a resin sink in contact
with columns of soil maintained under flooded, moist or flooded Figure 2. Measured uptake ofP by rice plants grown under the three
then moist water regimes. Data are means±SE of four replicates. different water regimes in Figure 1 (points) and the calculated up-
Lines are fits to the equation for diffusion from a semi-infinite take without solubilization (lines) using the diffusion coefficients of
source to an infinite planar sink: M t = 2[P]oo.J Dpt In where Mt is P in the soils derived from the data in Figure I. From Huguenin-Elie
the amount of P accumulated after time t, [P]oo the concentration of et al. (2002).
diffusible P in the soil bulk and Dp the P diffusion coefficient. From
Huguenin-Elie et al. (2002).
diffusion to roots is faster. Subsequent drainage has
the reverse effects, and the fluctuating water regime
are corroborated by rigorously testing them at their disrupts mycorrhizal populations (Ellis, 1998; Hag et
extremes. They can then be used as research tools or aI., 1987; Miller, 2000).
as sub-models in higher order models. Kirk and Saleque (1995) and Saleque and Kirk
I here give two examples of explanatory models. (1995) developed and corroborated an explanatory
Both examples are derived from the early models of model of P solubilization by rice in flooded anaer-
nutrient uptake developed in the 1960s and 1970s, obic soil, and Kirk (1999) and Kirk et al. (1999)
which treated roots as passive sinks to which nutri- did the same for solubilization by rice in moist aer-
ents were delivered by mass flow and diffusion. These obic soil. This work is summarized in Kirk (2002).
models worked reasonably well for the more soluble In flooded soil, solubilization occurs as a result
nutrient ions in well-fertilized soils, but for other cases of acidification caused by: (a) oxidation of Fe(II)
they failed and it became evident that the root played and other mobile reductants by root-released 02
a more active and complex role. In the first example, (4Fe2+ +02+ lOH20=4Fe(OHh+8H+); and (b) export
root-induced changes in the soil modifying nutrient of H+ from the roots to balance excess intake of
availability are important. In the second, the uptake cations over anions under NHt
nutrition. The model
properties of the roots themselves are important. A describes the movement of H+ away from the root and
more extensive review of models of the rhizosphere oxidation zone by acid-base transfer, the interaction
has recently been made by Darrah and Roose (2001). between H+ and soil P, and the simultaneous removal
of P by diffusion towards the root as well as away from
it. In moist aerobic soil, solubilization occurs as a res-
Phosphorus acquisition by rice under fluctuating ult of excretion of organic anions from the roots. The
water regimes model describes the movement of the organic anion
away from a root, its simultaneous consumption by
Much of the world's rice is grown under rainfed low- microbes and reaction with the soil in solubilizing P,
land conditions in which fields are kept flooded as far and the simultaneous movement of P away.
as possible, but they may become drained at some Huguenin-Elie et al. (2002) have used these mod-
point during the growing season. Under these condi- els to investigate P uptake by rice under fluctuating
tions, deficiency of P is often a problem. Flooding the water regimes. They grew rice in moist, flooded and
soil tends to increase the availability of P to roots both flooded then moist soils, and compared the measured
because the P becomes more soluble and because its P uptake with that calculated using the models. The
 125

35 7.0
(a)
30
'"""' p 6.5
Q) 25 pH at inter-root boundary uptake
"0
.!:
>.
0 20 . 6.0

.s
0>

15 a
I
Co
Q)
.><: 5.5
co pH at root surface
a.
::J 10
D..
5.0
5 0

0 4.5
0 10 20 30 40 50

Time following planting (days)

3.0 0 7.0

~
..a,
"":l
2.5
2.0
1.5
~
'"""' -4

(5

.s -12
E
-8
r
./
-_- -- --- ----
[Pd
6.5

pH .............. 6.0
...............
I
0..
/ ' ...................... 5.5
!!:. 1.0
E:' /
';:j -16
0.5 .'
/ (b) 42 days
5.0
-20 --L 4.5
0.0
0.0 0.1 0.2 0.3 0.4 0.5

Distance from root surface (mm)

Figure 3. (a) Measured P uptake by rice plants in the flooded soil in Figure 2 (points) and calculated uptake with solubilization by acidification
(solid lines). plus the calculated pHs at the root surface and at the mid-point between adjacent roots (dotted lines). (b) Calculated changes in
concentration ofP in the whole soil (Ll[P]), concentrations ofP in the soil solution ([PL]) and pH near a root after 42 days. From Huguenin-Elie
et aL (2002).

original models were developed in planar geometry
so as to simplify testing them against concentration
profiles of P and solubilizing agents measured in the
soil close to roots. Huguenin-Elie et al. (2002) conver-
ted the models into cylindrical geometry and allowed
for the increasing density of the root system over time
and the consequent overlapping of the zones of root
influence. The results are summarized as follows.
Calculated uptake without solubilization
To determine the diffusion coefficients of P in the soils
under the three water regimes and the concentrations
of diffusible P, Huguenin-Elie et al. (2002) measured
diffusion of P to a resin sink and fitted the appro-
priate diffusion equations (Figure 1). They then used
these values in conjunction with measurements of root
length density over time to calculate how much P the
plants would take up if there was no solubilization
and the roots acted as simple sinks without otherwise
influencing the soil. Figure 2 shows the observed res-
ults for the three water regimes. In the flooded soil,
uptake was three-times that in the moist soil. In all
three water regimes, the calculated uptake was an or-
der of magnitude smaller than the measured. In each
case, therefore, the plants relied upon solubilization
for most of their P.
Calculated uptake with solubilization
Figures 3 and 4 show the calculated uptake in the
flooded and moist soils allowing for solubilization. In
both cases, the model input parameters were measured
or estimated independent of the measured P uptake, so
the comparison between the observed and predicted
results provides a rigorous test of the models.
In the flooded soil, the uptake was well predicted
by the model and is therefore consistent with solu-
bilization by acidification resulting from oxidation of
Fe(II) and imbalance between the intake of cations and
126

35 ,..----,-----,-----,----,-------, 2.0
(a)
30
~

',- 1.5
Q)
"C
25 [e] at root surface
~
c:
'~
~ 20 [C) at inter-root (5
boundary
.s
Ol
1.0 E
Q) 15 .s
""15.
tv
P uptake 2:
::::J 10
a. 0.5
5 ....
y
~

0 L - _ - = " " - ' - _ - = = - ' - -_ _--'-_ _---'_ _----' 0.0

0 10 20 30 40 50
Time following planting (days)

0.6 0.0 .---.--,--.--,--.--,--,--,--,----, 2.0

............ ...........................................
0.5 ~~ -0.2 [e) 1.5
~

------
~
~
::l.
0.4
(5
E
-0.4 f- ------- 1.0
~
(5
[POd
;:; 0.3
9:. 0.2
.s -0.6 1-
/'" 8[P)
E
.s
Ii' /'"
0.5 2:
0.1 <i-O.aV (b) 42 days
0.0 _1.01/ 0.0
0.0 0.1 0.2 0.3 0.4 0.5
Distance from root surface (mm)
Figure 4. (a) Measured P uptake by rice plants in the moist soil in Figure 2 (points) and calculated uptake with solubilization by organic anion
excretion (solid lines), plus the calculated concentrations of citrate at the root surface and at the mid-point between adjacent roots (dotted lines).
(b) Calculated change in concentrations of P in the whole soil (~[PD, concentrations of P in the soil solution ([Prj) and concentrations of
citrate in the whole soil ([CD near a root after 42 days. From Huguenin-Elie et al. (2002).

anions. In the moist soil, the uptake was well pre-
dicted after 31 and 42 days, though in the earlier stages
it is under-predicted. Huguenin-Elie et al. offer two
explanations for this. Firstly, there may have been a
declining release of organic anions from the roots over
time - the calculations assuming a constant release -
as would be expected if the deficiency of P in the plant
became acute in the later stages. Secondly, excretion
of organic anions is generally found to be localized
along the root length, not uniformly distributed as as-
sumed in the model, and the effect of this may have
been to increase the proportion of the solubilized P
captured by the roots in the early stages before the
root system was very dense. Further experiments are
required to test these hypotheses, but they illustrate the
power of explanatory models to test understanding and
indicate where further research is needed.
In the flooded then moist soil, uptake declined
sharply as the soil was drained because P became im-
mobilized in the soil. However, the final uptake was
similar to that in the continuously moist soil, indicat-
ing that some of the immobilized P was re-solubilized
by roots, possibly by excretion of organic anions.
Again, the model indicates where further research is
needed.
In developing the models, the 'solubilizing effect'
of the solubilizing agent was defined in such a way that
the interaction between P and the solubilizing agent
could be measured unambiguously. The solubilizing
effect is defined as the amount of P than must be re-
moved from the soil for a given uniform addition of
solubilizing agent to leave the concentration of P in
the soil solution unchanged. This quantity is equal to
the ratio of the buffer powers of the soil for P and the
solubilizing agent multiplied by an interaction coeffi-
cient. It follows from this definition that the interaction
coefficient must be measured at constant concentration
of P in the soil solution by providing an adequate sink
 127

10 2.0
mmol plant' day' (a) 0
(b)
0 mean
o +N
0.4
p-
..
....... 8 +N 1.5 fo·
root surface

~
' 0
c:: 0.3 ~ 0

a.
I'll 0
0
g 0
-
0

6 0.2 c::
"0 0 0 8 ··0·
0
E :;:l
g 0.1 0
0
:J
"0
1.0
0
0
rn
~
!
4 0.0 .=
+~
ft- -N 0

:x: -I-
~
:J

•
z 0.5
.....
~.

2 2;.
.....
• •
.....
.,. ::e:: ..a
0.0 .L

25 30 35 40 45 50 55 20 25 30 35 40 45 50 55

I I I I I I

5000 (c) -N
5000 (d) +N
<f'
E
6
0
9 <f'
E
"C 8
I
"C
E
E- 0 0
O
E
E-
~ ~ 0

I
1000 ~ 1000 I:- -
e
~
·iii
c::
·iii
c:: 6
..
Q) Q)
"C 500 Q "C 500
.c::
.c::
• C, •
§
Cl
c::
~
t') ~
c::

15 0 15
0
0::: 100
• 0
0::: 100
•
I:- ...,

P I I I I I

20 25 30 35 40 45 50 55 20 25 30 35 40 45 50 55

Time following germination (days)

Figure 5. (a) Measured time-course of N uptake by rice plants in large pots of flooded soil with (0) and without (e) added ammoniacal-N;
the lines are fitted logistic curves; the inset shows the instantaneous rates of uptake calculated from the slopes of the logistic curves. (b) The
corresponding changes in concentrations of NHt
in the soil solution: solid horizontal lines are mean NHt concentrations in the soil solution;
broken lines are calculated NHt
concentrations in solution at the root surface. (c) and (d) The corresponding observed root length densities
(points) and calculated minimum root length densities required to explain the observed rates of N uptake (lines). The full lines were calculated
using measured mean soil solution NHt ±SE; the broken lines with soil solution NHt
derived from exchangeable ±SE. From Kirk and NHt
Solivas (1997).

for the P solubilized. In published measurements of
solubilizing effects this condition is often not met and
consequently the solubilizing effect is underestimated.
Nitrogen acquisition by rice in flooded soil
The second example concerns the efficiency with
which nitrogen fertilizer is used by rice in flooded soil.
Nitrogen broadcast into the floodwater in ricefields
can be absorbed rapidly by the plants at rates largely
independent of the properties of the root system, but
it is otherwise rapidly lost by volatilization into the
atmosphere. By contrast the uptake of N from the soil,
which is important in the overall N economy of the
ricefield, is likely to be limited by the properties of the
root system. Because the average recovery of N fertil-
izer in irrigated rice crops across Asia is less than 30%
(IRRI, 1998), devising ways of improving the capture
of N by the plants is of huge practical concern.
Calculated importance of root properties versus
transport through the soil
Kirk and Solivas (1997) have developed an explanat-
ory model of N acquisition by rice in flooded soil with
128

which to quantify the importance of root properties 9

versus rates of delivery of N through the soil. In the
model the principal form of N absorbed is taken to be 8
N0 3
NHt. The root demand for N per unit root length is
described with equations relating the influx of NHt to 7
,-...
its concentration in the soil solution at the root surface, :c
given the N status of the plant. The rate of transport 6
~
of NHt through the soil is described by equations u..
"';-
for mass flow and diffusion. Kirk and Solivas used Cl 5
(5
the model to calculate the minimum root length re- E
quired to explain measured rates of N uptake from the 2> 4
><
measured concentrations of NHt in the soil solution :::l
;;::
c:: 3
over the course of plant growth and measured max- z
imum rates of uptake per unit root length. The plants
were grown in large pots of flooded soil with ammo- 2
niacal N fertilizer mixed uniformly into the soil so
that the concentration in the floodwater was minimal.
The parameters for the kinetics of NHt absorption by
the roots were derived from independent experiments
o 100 200 300 400 500
in nutrient culture using the tracer 13N (Kronzucker
et aI., 1998a,b; Wang et aI., 1993). The plants were IN] in external solution (~M)

grown under a range of N levels prior to determining
the kinetics of NHt influx; the parameters used in the
model were for plants under moderate N deficiency
and indicate maximal rates of uptake.
Figure 5a shows the measured N uptake by the
plants and Figure 5b the corresponding measured con-
centrations of NHt in the soil solution and calculated
concentrations at the root surface. Figures 5c and d
show the corresponding observed root length densit-
ies and the calculated minimum root length densities
required to explain the uptake. With root uptake para-
meters assigned values such that influx was maximal
within realistic ranges, the minimum root length dens-
ities required to explain uptake were similar to meas-
ured root length densities in both the N-fertilized and
unfertilized soil. This suggests that most if not all of
the root length was active in uptake and that uptake
per unit root length was near maximal. A sensitiv-
ity analysis showed that the necessary minimum root
length is very sensitive to root uptake properties within
appropriate ranges. It also indicated that transport of
NHt to the roots - which is mainly by diffusion - will
generally not limit uptake in well-puddled soils, but
it may greatly limit uptake in puddled soils that have
been drained and re-flooded and in unpuddled flooded
soils.
That almost the whole of the root system was
evidently active in absorbing NHt is surprising: in
general N uptake by dryland crops can be accounted
for with only a fraction of the root length. An import-
Figure 6. Influx of N into roots of intact rice plants grown on 100
{LM N as either NO;- or NHt: (a) concentration-dependence of
NO;- and NHt influx in 4-week-old plants; (b) induction of NO;-
uptake in 3 or 4-week-old plants deprived of N for 24 h before
re-supply at 100 {LM for the indicated periods. Bars indicate SE
(n :::: 16). From Kronzucker et al. (2000).
ant difference is that in dryland soils the main form of
plant-available N is NO;-, and, because NO;- is not
adsorbed on the soil solid, for an equal quantity of N
its concentration in the soil solution is at least an or-
der of magnitude larger than that of NHt, and greatly
exceeds Km for root transport systems. Therefore up-
take rates per unit root length are larger and a smaller
root length is required. Nonetheless, the calculations
indicate an unlikely lack of margin for error in the root
length. A possible explanation is that N species other
than NHt are also being absorbed - such as NO;- or
amino acids - as shown in the following section.
Calculations of the fonn of N taken up
Kronzucker et al. (1999, 2000) found that lowland
rice grown hydroponically is exceptionally efficient in
absorbing NO;- , raising the possibility that rice grow-
ing in flooded soil may absorb significant amounts
of NO;- formed by nitrification in the rhizosphere.
This is important because: (a) this NO;- is otherwise
lost through denitrification in the soil bulk; and (b)
plant growth and yield are generally improved when

80
~
,.:;
c:
0
~ 60
(5
'"
'0
'"
.5
0....
z 40
X 970
Z c:
ON
c:
0 20
""~
CD
<>
c:
0
()
2 3
Distance from root surface (mm)
Figure 7. Calculated concentration profiles of 02, NO), NHt and
Fe2+ in a flooded soil near a rice root after 10 days of root-soil
contact. The parameter values used in the calculations are realistic
for a healthy root growing in an unexceptional low land rice soil
From Kirk and Kronzucker (2000).
plants absorb their N as a mixture of NO.3 and NHt
compared with either on its own.
The evidence for the unusually efficient absorp-
tion of NO.3 is as follows. First, steady-state influx
of NO.3 and NHt measured with 13N tracer followed
Michaelis-Menten kinetics over the relevant concen-
tration range (Figure 6), and \l.'Ilax for NO.3 was some
40% larger than that for NHt and Km 50% smaller.
Secondly, measurements of the change in NO.3 influx
over time following its re-supply to plants deprived of
NO.3 showed exceptionally rapid induction of NO.3
uptake. Peak rates of influx occurred within 2 h of
re-supply. For comparison, in white spruce, which is
not well adapted to using NO.3, full induction takes
several days, and in barley, which is considered one of
the most efficient NO.3 users among higher plants, full
induction takes up to 24 h. Thirdly, sub-cellular pool
sizes and fluxes, estimated from the kinetics of 13N
efflux out of labelled roots, indicated highly efficient
NO.3 use: while similar proportions of incoming NHt
and NO.3 were channelled into assimilation and to the
cell vacuole, the proportion of NO.3 translocated to the
shoot was larger and that lost by efflux out of roots was
smaller.
Based on the above observations Kirk and
Kronzucker (2000) developed a model with which to
assess how far rice growing in flooded soil can absorb
129
1000 NO.3 formed in the rhizosphere. The following pro-
cesses are allowed for: (a) transport of 02 away from
a root and its consumption in microbial processes -
990 ~
,.:; in addition to nitrification - and oxidation of mobile
c:
0 reductants such as Fe2+; (b) transport of NHt towards
:g the root and its consumption in nitrification and uptake
(5
980 '" at the root surface; and (c) transport of NO.3 formed
'0
'"
.5 from NHt towards the root and its consumption in
"CD denitrification and uptake by the root. Uptake of NHt
u..
0
and NO.3 by the root are described by the Michaelis-
~ Menten relations obtained by Kronzucker et al. (1999,
E
CD 2000).
960 0
<>
c:
Figure 7 shows the calculated concentration pro-
()
files of 02, Fe2+, NHt and NO.3 near a root for
realistic flooded soil conditions. The corresponding
950
uptake of NO.3 into the root was about half that of
NHt, i.e. a third of the total N uptake. The amount
of N denitrified in 10 days in the calculations corres-
ponds to about 10% of the NHt initially in the volume
of soil influenced by the root. This is of the order of
maximum rates of denitrification reported in the liter-
ature for rice in flooded soil, indicating that the model
parameter values are indeed realistic.
The calculations are sensitive to the rate of delivery
of 02 through the root. Reported rates of 02 release
from rice roots vary by more than two orders of mag-
nitude depending on the experimental system used and
the strength of the external sink for 02. Models of root
aeration show that rates of release at the upper end of
this range can be sustained by rice roots with typical
lengths, respiration rates and other characteristics. The
rate of 02 release in the calculations in Figure 7 is in
the middle of the range.
The calculations are also sensitive to the rate of 02
consumption in the soil. The maximum rate of nitrific-
ation is some function of the maximum overall rate
of microbial 02 consumption, both being functions
of the supply of oxidizable organic substrates. The
maximum rate of denitrification will also depend on
the supply of organic substrates and will be related
to the maximum rate of microbial 02 consumption.
Thus the rates of nitrification and denitrification are
very sensitive to reducing conditions in the soil. Based
on results with a range of reduced soils, the maximum
rate of microbial 02 consumption was set at 100 nmol
dm- 3 s-1 in the model calculations and the maximum
rates of nitrification and denitrification set at half and
a quarter of this.
The model indicates that quite large amounts of
NO.3 may be absorbed by rice in flooded soils, per-
130
haps as much as a third of the total N absorbed if soil
conditions and water management prevent very thor-
ough soil reduction. Thus NO.3 uptake by rice may
be more important than thought hitherto, and this may
have important consequences for the management of
rice crops.
References
Darrah P R and Roose T 2001 Modeling the rhizosphere. In The
Rhizosphere: Biochemistry and Organic Substances at the Soil-
Plant Interface. Eds R Pinton, Z Varanini and P Nannipieri. pp
327-372. Marcel Dekker, New York.
Ellis J R 1998 Post flood syndrome and vesicular-arbuscular mycor-
rhizal fungi. J. Prod. Agric. 11, 200-204.
Huguenin-Elie 0, Kirk G J D and Frossard E 2002 Phosphorus up-
take by rice from soil that is moist, flooded or flooded then moist.
Eur. J. Soil Sci., in press.
Hag L L, Rosales AM, Elazegui F A and Mew T W 1987 Changes in
the population of infective endomycorrhizal fungi in a rice-based
cropping system. Plant Soil 103, 67-73.
IRRI (International Rice Research Institute) 1998 Reversing Trends
of Declining Productivity in Intensive Irrigated Rice Systems.
Progress Report 1997. International Rice Research Institute
Manila.
Kirk G J D 1999 A model of phosphate solubilization by organic
anion excretion from plant roots. Eur. J. Soil Sci. 50, 369-378.
Kirk G J D 2002 Modelling root-induced solubilization of nutrients.
Plant Soil, in press.
Kirk G J D and Kronzucker H J 2000 Nitrogen uptake by rice roots.
In Carbon and Nitrogen Dynamics in Flooded Soils. Eds G J D
Kirk and D C Olk. pp. 147-162. IRRI, Manila.
Kirk G J D and Saleque M A 1995 Solubilization of phosphate by
rice plants growing in reduced soil: prediction of the amount
solubilized and the resultant increase in uptake. Eur. J. Soil Sci.
46,247-255.
Kirk G J D, Santos E E and Santos M B 1999 Phosphate solubiliza-
tion by organic anion excretion from rice growing in aerobic soil:
rates of excretion and decomposition, effects on rhizosphere pH,
and effects on phosphate solubility and uptake. New PhytoL 142,
185-200.
Kirk G J D and Solivas J L 1997 On the extent to which root
properties and transport through the soil limit nitrogen uptake
by lowland rice. Eur. J. Soil Sci. 48, 613-621.
Kronzucker H J, Kirk G J D, Siddiqi M Y and Glass ADM 1998a
Effects of hypoxia on 13NHt fluxes in rice roots: kinetics and
compartmental analysis. Plant PhysioL 116,581-587.
Kronzucker H J, Schjoerring J K, Erner Y, Kirk G J D, Siddiqi M
Y and Glass ADM 1998b Dynamic interactions between root
NHt influx and long-distance N translocation in rice: insights
into negative feedback processes. Plant Cell PhysioL 39, 1287-
1293.
Kronzucker H J, Siddiqi M Y, Glass ADM and Kirk G J D 1999
Nitrate-ammonium synergism in rice: a subcellular flux analysis.
Plant PhysioL 119, 1041-1045.
Kronzucker H J, Glass A.D M, Siddiqi M Y and Kirk G J D 2000
Comparative kinetic analysis of ammonium and nitrate acquisi-
tion by tropicallowland rice: implications for rice cultivation and
yield potential. New PhytoL 145,471-476.
Miller S P 2000 Arbuscular mycorrhizal colonization of semi-
aquatic grasses along a wide hydrologic gradient. New PhytoL
' 145, 145-155.
Nye P H 1992 Towards the quantitative control of crop production
and qUality. I. The role of computer models in soil and plant
research. J. Plant Nutr. 15, 1121-1150.
Saleque M A and Kirk G J D 1995 Root-induced solubilization of
phosphate in the rhizosphere of lowland rice. New PhytoL 129,
325-336.
Wang Y M, Siddiqi M Y, Ruth T J and Glass ADM 1993 Am-
monium uptake by rice roots. II. Kinetics of 13NHt influx across
the plasmalemma. Plant PhysioL 103, 1259-1267.
Soil organisms/plant interactions
 ..... Plant and Soil 247: 133-139,2002.
• 133
'l1li © 2002 Kluwer Academic Publishers.

Malate plays a central role in plant nutrition

J. Schulze 1,2, M. Tesfaye3 , R. H. M. G. Litjens 4 , B. Bucciarelli2 , G. Trepp 2,5, S. Miller2 , D.

Samac 3,6, D. Allan7 & C. P. Vance 2,6,8
1Institut fuer Bodenkunde und Pflanzenernaehrung der Martin-Luther-Universitaet, Halle-Wittenberg, Halle, Ger-
many. 2Department of Agronomy and Plant Genetics, University of Minnesota, 1991 Upper Buford Circle, St.
Paul, MN 55108, USA. 3 Department of Plant Pathology, University of Minnesota, 1991 Upper Buford Circle, St.
Paul, MN 55108, USA. 4Fontys University, Eindhoven, Eindhoven, The Netherlands. 5ETH Zurich, Switzerland.
6USDA-ARS, Plant Science Research Unit, 1991 Upper Buford Circle, 411 Borlaug Hall, St. Paul, MN 55108,
USA. 7 Department of Soil, Water, and Climate, University of Minnesota, 1991 Upper Buford Circle, St. Paul, MN
55108, USA. 8 Corresponding author*

Received 21 December 2002. Accepted in revised form 24 May 2002

Key words: aluminum tolerance, malate dehydrogenase, nitrogen fixation, organic acids, phosphorus stress

Abstract
Malate occupies a central role in plant metabolism. Its importance in plant mineral nutrition is reflected by the role
it plays in symbiotic nitrogen fixation, phosphorus acquisition, and aluminum tolerance. In nitrogen-fixing root
nodules, malate is the primary substrate for bacteroid respiration, thus fueling nitrogenase. Malate also provides
the carbon skeletons for assimilation of fixed nitrogen into amino acids. During phosphorus deficiency, malate
is frequently secreted from roots to release unavailable forms of phosphorus. Malate is also involved with plant
adaptation to aluminum toxicity. To define the genetic and biochemical regulation of malate formation in plant
nutrition we have isolated and characterized genes involved in malate metabolism from nitrogen-fixing root nodules
of alfalfa and those involved in organic acid excretion from phosphorus-deficient proteoid roots of white lupin.
Moreover, we have overexpressed malate dehydrogenase in alfalfa in attempts to improve nutrient acquisition.
This report is an overview of our efforts to understand and modify malate metabolism, particularly in the legumes
alfalfa and white lupin.

Introduction aluminum (AI) tolerance (Gietl, 1992; Kochian, 1995;

Malate is a key product of plant metabolism and
thought by many (see Lance and Rustin, 1984; Mar-
tinoia and Rentsch, 1994) to be the ultimate product
of glycolysis, rather than pyruvate. Functional roles
for malate in plants are quite diverse including, but
not limited to: respiration and energy generation, pho-
tosynthesis (both C3 and C4), fatty acid oxidation,
lignin biosynthesis, pulvinal and stomatal function, ni-
trogen (N2) fixation and amino acid biosynthesis, ion
balance, uptake of phosphorus (P) and iron (Fe), and
* FAX No: +1-651-649-5058
E-mail: vance004@tc.urnn.edu
Martinoia and Rentsch, 1994).
In N2-fixing nodules, malate is the predominant
source of energy for bacteroid respiration (Driscoll
and Finan, 1993) and provides a significant portion
of the carbon skeletons for assimilation of fixed N2
(Rosendahl et aI., 1990). Malate may also be in-
volved in regulation of the nodule oxygen diffusion
barrier through an osmoelectrical mechanism (Den-
ison, 1998; Galvez et aI., 2000; Vance and Heichel,
1991). The critical role that malate plays in root nod-
ules is evidenced by the fact that ineffective nodules,
whether induced by changes in either the bacterial
or plant genotype, have strikingly reduced levels of
malate as compared to effective nodules. Moreover,
mutations in rhizobia that block organic acid use res-
134
SUCROSE
PHOTOSYNTHATE
HEXOSES
,lJ.
PEP
jJ!
ll[D
OXALOACETATE
+
MALATE
Figure 1. Diagrammatic scheme for the degradation of photosynthate (sucrose) to malate. Due to its diverse functions in plants, malate is
targeted to many subcellular locations. Because malate functions in a wide range of processes in distinct cellular locations, at least five forms
of malate dehydrogenase (MDH) can be found in plants including: cytosolic, glyoxysomal, peroxisomal, plastidic, and mitochondrial. In root
nodules, MDH may also be associated with symbiosomes. SS=sucrose synthase; PEPC=phosphoeno[pyruvate carboxylase.
ult in ineffective nodules while those that block amino
acid and carbohydrate use generally have no effect on
N2 fixation (Driscoll and Finan, 1993; Ronson et aI.,
1981).
Phosphorus is the limiting element in many soils
(Bieleski, 1973; Vance, 2001). In such soils plants
display adaptive features that aid in acquisition of P
(Raghothama, 1999; Schactman et aI., 1998). White
lupin forms a unique root developmental adaptation
as a substitute for a mycorrhizal association, the de-
velopment of proteoid or cluster roots (Skene, 2001 ;
Watt and Evans, 1999). Proteoid roots not only in-
crease the root surface area by more than 100-fold to
aid in exploration for P but they also release phenom-
enal amounts of organic acids, primarily malate and
citrate (Dinkelaker et aI., 1995; Skene, 2001). Release
of organic acids helps to increase the availability of
mineral P by solubilizing Ca-, Fe- and AI-phosphates
(Ryan et aI., 2001). Exudation of malate and citrate
from proteoid roots may represent as much as 25 % of
total photosynthate (Gardner et aI., 1983; Johnson et
aI., 1996a).
Synthesis of malate is catalyzed by the enzyme
malate dehydrogenase (MDH: EC 1.1.1.82) through
the reversible reduction of oxaloacetate to malate. Be-
cause this enzyme is important in several metabolic
pathways in higher plants, it occurs in multiple forms
PEPC I
that differ in subcellular localization (Figure 1) and
cofactor specificity (Gietl, 1992). Although the reac-
tion favors malate production, whether oxaloacetate or
malate forms depends upon physiological conditions
and enzyme location.
Chloroplasts contain an NADP-dependent MDH
(pIMDH) that plays a critical role in balancing re-
ducing equivalents between the cytosol and stroma.
Plants also contain at least 4 NAD-dependent MDHs
which are found in: (i) the cytosol (cMDH) and perox-
isomes (pMDH) involved in malate-aspartate shuttles;
(ii) the mitochondria (mMDH) involved in the TCA
cycle; (iii) the glyoxysomes (gMDH) functioning in
,B-oxidation (Gietl, 1992; Miller et aI. , 1998); and
(iv) root nodules (neMDH) functioning in N2-fixation
and N assimilation (Miller et aI. , 1998). The enzyme
has been purified from several plant sources and anti-
bodies produced against plMDH, gMDH, and pMDH.
The p and gMDHs are serologically indistinguishable,
while the plMDH is antigenically unique.
Because malate plays such a pivotal role in
nitrogen-fixing root nodules and in acquisition of P by
proteoid roots, we thought it important to understand
the biochemical and genetic factors affecting malate
synthesis. This report provides an overview of our
laboratory's efforts to understand and modify malate
synthesis.

Chloroplast
Glyoxysomal
Mitochondrial
Cytosolic
Nodule-Enhanced
Figure 2. Steady state transcript abundance of the various fonns of malate dehydrogenase (MDH) in alfalfa organs: R=root; N=nodule;
C=cotyledon; S=stem; L=leaf. Transcript size in kb is noted to the right of each blot. Equivalent amounts of polyA + RNA, I J.l.g, were loaded
in each lane as measured by 32p poly(U) hybridization (adapted from Miller et al. 1998).
Alfalfa root nodules
Using strategies involving complementation of an E.
coli MDH- mutant, RT-PCR, and screening with
heterologous probes, we cloned five distinct MDH
cDNAs (Miller et aI., 1998). These cDNAs encoded
chloroplast (pI), mitochondrial (m), glyoxysomal (g),
cytosolic (c), and nodule-enhanced (ne) MDHs. The
cDNAs representing c- and neMDH were novel, this
being the first report of their isolation.
RNA blot analysis was used to assess the relat-
ive abundance of each MDH mRNA transcript (Figure
2). There appeared to be fairly uniform abundance of
mMDH, gMDH, and cMDH in roots, nodules, coty-
ledons, stems, and leaves. By comparison, plMDH
was found only in green tissues, particularly leaves.
Transcripts for neMDH were found predominantly in
root nodules with very minor amounts in other tissues.
The discovery of a nodule form of MDH is consist-
ent with biochemical studies of pea nodules by Appels
and Haaker (1988) and lupin nodules by Ratajczak et
al. (1989), suggesting a root nodule form of MDH
135
Kb
I 1.7
I 1.&
I 1.&
... ..1 1.&
I 1.7
R N C S L
catalyzing high in vitro rates of malate production.
Moreover, Fedorova et al. (1999) isolated from pea
nodules a neMDH cDNA having striking similarity to
the alfalfa homolog.
Because the alfalfa neMDH and cMDH are novel
forms, we did extensive biochemical characterization
of these proteins. Having both forms of MDH ex-
pressed in an E. coli MDH- mutant allowed us to isol-
ate the recombinant enzymes, perform kinetics, and
produce antibodies. Kinetic parameters for neMDH
were strikingly different than those for cMDH. The
Km analysis for oxaloacetate showed that neMDH
had a 7-fold greater affinity for this substrate as com-
pared to cMDH. The neMDH enzyme turnover rates
for malate oxidation and oxaloacetate reduction were
4-fold and 3D-fold greater, respectively, than those for
cMDH. Even more surprising, the neMDH specificity
constants for oxaloacetate and NADH were 70- to 100-
fold greater than those for malate and NAD. These
data suggest that neMDH catalysis dramatically favors
malate production and neMDH may be a useful target
to develop plants with modified malate metabolism.
136
ENZYME cMDH neMDH MIX
ACTIVITY
Figure 3. Immunologic specificity of antibodies prepared to cytoso-
lic (c) and nodule-enhanced (ne) malate dehydrogenase (MDHJ.
Nodule protein was subjected to native polyacrylamide gel elec-
trophoresis and tben stained for MDH activity (Enzyme Activity)
or transferred to nitrocellulose and immunostained witb antibodies
to cytosolic MDH (cMDH). nodule-enhanced MDH (neMDH). or
a mixture of botb antibodies (Mix). Note tbe specificity of c- and
neMDH antibodies for specific isozymes of MDH activity.
Antibodies prepared to neMDH and cMDH were
highly specific (Figure 3) showing little to no cross
reactivity. Having high-quality specific antibodies al-
lowed us to assess steady state quantities of ne- and
cMDH protein. Similar to RNA blot analysis, neMDH
protein was greater in nodules than any other tissue ex-
amined. Substantial amounts of neMDH protein were
found in meristematic tissue, with much less in roots,
cotyledons, stems, and leaves. Likewise, cMDH pro-
tein was uniformly expressed in all tissues examined
as were cMDH transcripts. Accompanying immuno-
precipitation experiments showed that neMDH com-
prises the largest proportion of total nodule MDH,
about 50%.
Most recently we isolated and sequenced the al-
falfa genes encoding c- and neMDH, respectively
(Figure 4). Cytosolic MDH was comprised of seven
exons interrupted by six introns. This gene struc-
ture was conserved in the Arabidopsis genome, how-
ever, introns in Arabidopsis were much smaller than
those in alfalfa. Preliminary sequencing of Medicago
truncatula cMDH showed that intron-exon structure
and sequence similarity was very conserved between
alfalfa and M. truncatula. By comparison, alfalfa
neMDH contained only one intron, and it was loc-
ated in the 5' -untranslated region of the gene. Again,
intron structure was conserved between alfalfa, M.
truncatula and Arabidopsis.
White lupin proteoid roots
White lupin is adapted for growth on neutral to acidic
soils and can flourish in nutrient-poor environments
that would limit most plants (Dinkelaker et al., 1995;
Gilbert et al., 1998). Although it does not form a
mycorrhizal symbiosis, white lupin can effectively ac-
quire soil P through the development of proteoid roots
and their subsequent exudation of malic and citric
acids (Johnson et al., 1994, 1996a, b; Neumann et
al., 2000). Proteoid roots are densely clustered ter-
tiary roots having prolific development of root hairs
(Dinkelaker et al., 1995; Skene, 2001). Their form-
ation is suppressed when soil P levels are sufficient,
but even when P is adequate, as much as 5-10% of
the root mass may be proteoid (Johnson et al., 1984;
Skene, 2001). Proteoid root formation can increase
the root surface area by more than 100-fold. The mal-
ate and citrate exuded from proteoid roots allows for
the displacement of bound P from Al3+ -, Fe3+ - and
Ca2+ -phosphates, thus increasing available P. A fur-
ther adaptation to low P is that proteoid roots have
enhanced P uptake and increased expression of phos-
phate transporter genes (Liu et al., 2001; Neumann et
al., 1999).
We (Johnson et al., 1994, 1996a, b) as well as oth-
ers (Keerthisinghe et al., 1998; Neumann et al., 1999)
have shown that proteoid roots of white lupin excrete
phenomenal amounts of citrate and malate during P-
deficiency. The amount of organic acids released in
root exudates of white lupin in response to P stress can
represent 11-23% of total photosynthate (Dinkelaker
et al., 1995; Gardner et al., 1983). We showed, us-
ing in vivo 14C02-labeling, that during a 70-h period
227 t-tmol citrate and 156 t-tmol malate were recovered
from -P root exudates as compared to 11 t-tmol cit-
rate and 9 t-tmol malate from +P roots (Johnson et
al., 1996a). In efforts to understand the mechanisms
underlying P-stress induced organic acid exudation
from proteoid roots the activity of enzymes related to
malate and citrate synthesis was evaluated. We found
that increased organic acid synthesis and exudation
is mediated by increased specific activities (Table 1)
of cMDH, phosphoenolpyruvate carboxylase (PEPC),
and citrate synthase (CS) (Johnson et al., 1994, 1996a,
b). Increased MDH and PEPC activity in proteoid
roots is accompanied by enhanced PEPC and MDH
mRNA and protein (Johnson et al., 1996b; C. Uhde-
Stone, C. Vance, and D. Allan data not shown). More
recently, enhanced citrate synthesis was also shown to
be related to reduced degradation as aconitase activ-
 137

cMDH
Il I2 I3 I4 IS I6

neMDH

500 bp
Figure 4. Exon-intron structure of alfalfa cytosolic, cMDH, and nodule-enhanced, neMDH, genes. The genes are distinctly different in structure,
6 introns are found in the coding region of cMDH while none are found in the coding region of neMDH.

Table 1.Specific activity of malic dehydrogenase (MDH), phosphoenolpyruvate carboxylase (PEPC), and citrate synthase (CS) in proteoid
and normal roots of white lupin. Data adapted from Johnson et al. (1994, 1996b) and Gilbert et al. (l99S)

Root Type MDH PEPC CS

fLmol NADH-min-1·mg protein- 1 fLmol acetyl CoA-min-1·mg protein- 1
Normal +P 7.9 a O.OSa O.lla
Normal-P S.2a O.27 b o.osa
Proteoid +P O.13 a O.lOa
Proteoid -P 0.41 e 0.14b

Each value is the mean of at least nine determinations. Means in each column labeled with the same letter are not different as determined by
LSD (P:sO.OS).

Table 2. In vivo C02 fixation and recovery of fixed C in root exudates of phosphorus-stressed white lupin. ND=Not Detectable. Adapted from
Johnson et al. (1996a)

C02 labeling Treatment rn~

'---'--' L fivo:at~ ...... n
.1.1.1'...... " .......1 .. .1
rra. ............ "::. ...... ,
_.lvvVvv.lJ

site Total exudate Citrate Malate

nmol.min-1.g FW- 1 fLmol C
Root +P 2.6±0.S 0.1±0.1 ND ND
Root -P 11.l±1.7 2.4±0.3 1.0±0.6 0.S±0.3
Shoot +P 72.3±24.7 0.4±0.3 ND ND
Shoot -P 102.0±36.1 S.S±1.9 3.S±2.0 1.3±1.3

Mean±SD for five independent determinations. Differences are significant at P:sO.OS.

ity in proteoid roots of P-stressed plants is reduced PEPC would require catalysis to malate and citrate via
(Neumann et aI., 1999). MDH (Figure 1) and CS, respectively. Thus, MDH
In vivo-labeling experiments in which shoots and plays a role in providing organic acids for exudation
roots were independently labeled with 14C02 and root from P-stressed proteoid roots as an adaptive strategy
exudates subsequently collected (Johnson et aI., 1994, to acquire unavailable P under infertile conditions.
1996a, b) showed that 25-35% of the label found in
exuded citrate and malate was derived through root
C02 fixation via root PEPC (Table 2). For signific- Transgenic enhancement of malate
ant amounts of radioactivity to be found in organic
acids after root labeling, the oxaloacetate derived from Malate has a pivotal role in acquisition of N through
symbiotic N2 fixation and P through root exudation.
138
Table 3. Steady state mRNA abundance in trans-
genic plants overexpressing neMDH. Relative abundance
measured as incorporation of 32 P from labeled probe into
neMDHRNA
Transgenic Tissue
line
Leaf Root Root tip Nodule
cpm in neMDH RNA
Control 1400 1497 2580 7085
MDH 16-27 2360 4169 4537 8649
MDH 20-10 2370 6838 9928 12600
Mean of two independent determinations.
Moreover, increased synthesis of citrate and mal-
ate is intimately related with plant tolerance to Al
stress (Delhaize and Ryan, 1995; Kochian, 1995).
Because malate is a critical component of plant nutri-
ent acquisition and adaptation to environmental stress,
we hypothesized that improving malate synthesis via
transgenic technology might be an effective strategy
for improving plant nutrition. Although controver-
sial (Delhaize et aI., 2001), precedence for transgenic
approaches to increasing organic acid synthesis and
improving plant nutrition has been demonstrated by
overexpression of citrate synthase (de la Fuente et aI.,
1997; Koyama et aI., 1999; Lopez-Bucio et aI., 2000).
Overexpression of citrate synthase resulted in plants
having improved P nutrition and resistance to AI.
To overexpress MDH we developed a chimeric
gene construct with neMDH being under the control
of the constitutive 35S promoter. Alfalfa was trans-
formed with the 35S::neMDH construct and trans-
genic plants recovered. Of nine independent trans-
formants showing increased neMDH activity in root
tips, the two showing the greatest enhancement of
MDH activity, MDH 20-10 and 16-27, were selected
for further study (Tesfaye et aI., 2001).
Both lines showed that increased MDH activity
was accompanied by greater MDH protein (Tesfaye et
aI., 2001) and mRNA (Table 3) as compared to the
controls. Moreover, during a 24-h period, MDH 16-
27 plants exuded significantly more citrate, malate,
oxalate, succinate, and acetate into the rhizosphere
than did the controls. By comparison, MDH 20-10
plants also tended to exude more organic acids than
the control but less than MDH 16-27 (Tesfaye et aI.,
2001).
To determine whether overexpression of neMDH
affected P acquisition and Al tolerance, plants were
grown in highly weathered acid soil and hydroponic
solution containing various concentrations of AI. The
Table 4. Shoot and root P content of transgenic
plants overexpressing neMDH. Data adapted from
Tesfaye et aI., 2001
Transgenic line Shoot P RootP
fLg plant- 1
Control 96a 34a
MDH 16--27 136b 85 b
MDH 20-10 14l b 83 b
Each value is the mean of at least three determ-
inations. Means in each column labeled with the
same letter are not different as determined by LSD
(P::o:O.05).
shoot and root P accumulation in transgenic plants
was greater than the control (Table 4). Furthermore,
root elongation of transgenic plants were signific-
antly more tolerant to Al than was that of the control
(Tesfaye et aI., 2001).
To assess whether overexpression of neMDH has
any effect on N2 fixation, nodulated plants of MDH
16-27 and a control transgenic line PARC-lOO were
exposed to l5N2 gas for 24 hand N2 fixation de-
termined. Preliminary results from plants tested at
8 weeks after inoculation showed significantly im-
proved N2 fixation as evaluated by nodule efficiency
(/tmol N·gF nodule- l .h- 1). The nodule efficiency of
PARC-lOO was 12.36 while that of MDH 16-27 was
16.33, demonstrating enhanced N2 fixation potential
in nodules ofMDH 16-27 plants.
Overview
Malate plays a central role in plant nutrition. Physiolo-
gical, biochemical, and molecular studies with alfalfa
and white lupin provide evidence for the importance
of malate in N2 fixation, acquisition of P from in-
fertile soils, and tolerance to Al stress. At least five
separate forms of MDH are found in alfalfa and each
is encoded by distinct genes. The unusual kinetic
parameters of alfalfa neMDH make it an excellent can-
didate for enhancing malate synthesis via transgenic
technology. Exudation of the organic acids malate
and citrate from P-stress induced proteoid roots of
white lupin increases the availability of sparingly sol-
uble metal phosphates thereby providing P under low
P conditions. Synthesis of malate and citrate in P-
stress proteoid roots requires the concerted action of
the enzymes PEPC, MDH, and CS. Overexpression of
neMDH in alfalfa results in plants with higher P ac-
cumulation, improved tolerance to AI, and increased

root nodule efficiency. It will be important to estab-
lish whether further enhancement of the expression of
neMDH in alfalfa results in improved P uptake and
greater tolerance to AI. In addition, it will be crit-
ical to determine whether increased nodule efficiency
translates into improved N nutrition.
References
Appels M A and Haaker H 1988 Identification of cytoplasmic
nodule-associated forms of malate dehydrogenase involved in
the symbiosis between Rhizobium leguminosarum and Pisum
sativum. Eur. J. Biochem. 171,515-522.
Bieleski R L 1973 Phosphate pools, phosphate transport, and
phosphate availability. Annu. Rev. Plant Physiol. 24, 225-252.
de la Fuente J M, Ramirez-Rodrigues V, Cabreta-Ponce J L and
Herrera-Estrella L 1997 Aluminum tolerance in transgenic plants
by alteration of citrate synthesis. Science 276, 1566-1568.
Delhaize E and Ryan P R 1995 Aluminum toxicity and tolerance in
plants. Plant Physiol. 107,315-321.
Delhaize E, Hebb D M and Ryan P P 2001 Expression of a
Pseudomonas aeruginosa citrate synthase gene in tobacco is not
associated with either enhanced citrate accumulation or efflux.
Plant Physiol. 125,2059-2067.
Denison R F 1998 Decreased oxygen permeability: a universal
stress response in legume root nodules. Bot. Acta 111, 191-192.
Dinkelaker B, Hengeler C and Marschner H 1995 Distribution and
function of proteoid roots. Bot. Acta 108, 183-200.
Driscoll B T and Finan T M 1993 NAD+ -dependent malic enzyme
of Rhizobium meliloti is required for symbiotic nitrogen fixation.
Mol. Microbiol. 7, 8655-8873.
Fedorova M, Tikhonovich I A and Vance C P 1999 Expression of C-
assimilating enzymes in pea Pisum sativum root nodules. Plant,
Cell Environ. 22, 1249-1262.
Galvez S, Hirsch A M, Wycoff K L, Hunt S, Layzell D, Kondorosi
A and Crespi M 2000 Oxygen regulation of a nodule-located
carbonic anhydrase in alfalfa. Plant Physiol. 124, 1059-1068.
Gardner W K, Barber D A and Parberry D G 1983 The acquisition
of phosphorus by Lupinus albus L. III. The probable mechan-
ism by which phosphorus movement in the soil/root interface is
enhanced. Plant Soil 70, 107-124.
Gietl C 1992 Malate dehydrogenase isoenzymes: cellular locations
and role in the flow of metabolites between cytoplasm and cell
organelles. Biochem. Biophys. Acta 1100, 217-234.
Gilbert G A, Vance C P and Allan D L 1998 Regulation of white
lupin metabolism by phosphorus availability. In Phosphorus in
Plant Biology: Regulatory Roles in Molecular, Cellular, Organis-
mic and Ecosystem Processes. Eds. J P Lynch and J Deikman. pp.
157-167. American Society Plant Physiology, Rockville, MD.
Johnson J F, Allan D L and Vance C P 1994 Phosphorus stress-
induced proteoid roots show altered metabolism in Lupinus
albus. Plant Physiol. 104, 657-665.
Johnson J F, Allan, D L, Vance C P and Weiblen g 1996a Root
carbon dioxide fixation by phosphorus-deficient Lupinus albus.
Contribution to organic acid exudation by proteoid roots. Plant
Physiol. 112, 19-30.
Johnson J M F, Vance C P and Allan D L 1996b Phosphorus de-
ficiency in Lupinus albus. Altered lateral root development and
enhanced expression of phosphoenolpyruvate carboxylase. Plant
Physiol. 112,31-41.
139
Keerthisinghe G, Hocking P J, Ryan P R and Delhaize E 1998 Effect
of phosphorus supply on the formation and function of proteoid
roots of white lupin. Plant Cell Environ. 21,467-478.
Kochian L V 1995 Cellular mechanisms of aluminum toxicity and
resistance in plants. Annu. Rev. Plant Physiol. Plant Mol. BioI.
46, 237-260.
Koyama H, Takita E, Kawamura A, Hara T and Shibata D 1999
Over-expression of a mitochondrial citrate synthase gene im-
proves the growth of carrot cells in AI-phosphate medium. Plant
Cell Physiol. 40, 482-488.
Lance C and Rustin P 1984 The central role of malate in plant
metabolism. Physiol. Veg. 22, 625-641.
Liu J, Uhde-Stone C, Li A, Vance C P and Allan D L 2001 A phos-
phate transporter with enhanced expression in proteoid roots.
Plant Soil 237, 257-266.
Lopez-Bucio J, de La Vega 0 M, Guevara-Garcia A and Herrera-
Estrella L 2000 Enhanced phosphorus uptake in transgenic plants
that overproduce citrate. Nature Biotech. 18,450-453.
Martinoia E and Rentsch D 1994 Malate compartmentation - re-
sponses to a complex metabolism. Annu. Rev. Plant Physiol.
Plant Mol. BioI. 45, 447-467.
Miller S S, Driscoll B T, Gregerson R G, Gantt J S and Vance C
P 1998 Alfalfa malate dehydrogenase (MDH): molecular clon-
ing and characterization of five different forms reveals a unique
nodule-enhanced MDH. Plant J 15,173-184.
Neumann G, Massonneau A, Martinoia E and Romheld V 1999
Physiological adaptations to phosphorus deficiency during pro-
teoid root development in white lupin. Planta 208, 373-382.
Neumann G, Massonneau A, Langlade N, Dinkelaker B, Hengeler
C, Romheld V and Martinoia E 2000 Physiological aspects of
cluster root function and development in phosphorus-deficient
white lupin (Lupinus albus L.). Ann. Bot. 85,909-919.
Raghothama K G 1999 Phosphate acquisition. Annu. Rev. Plant
Physiol. Plant Mol. BioI. 50, 665-693.
Ratajczak L, Ratajczak W and Koroniak D 1989 Detection of nod-
ule specific forms of malate dehydrogenase from root nodules of
Lupinus luteus. Biochem. Physiol. Pflanzen 184, 243-248.
Ronson C W, Lyttleton P and Robertson J G 1981 C4-dicarboxylate
transport mutants of Rhizobium trifoli form ineffective nodules
on Trifolium repens. Proc. Natl. Acad. Sci. USA 78, 4284-4288.
Rosendahl L, Vance C P and Pedersen W B 1990 Products of dark
C02 fixation in pea root nodules support bacteroid metabolism.
Plant Physiol. 93, 12-19.
Ryan P R, Delhaize E and Jones D L 2001 Function and mechan-
ism of organic anion exudation from plants. Annu. Rev. Plant
Physiol. Plant Mol. BioI. 52, 527-560.
Schactman D P, Reid R J and Ayling S M 1998 Phosphorus uptake
by plants: from soil to cell. Plant Physiol. 116,447-453.
Skene K R 2001 Cluster roots: model experimental tools for key
biological problems. J Exp. Bot. 52,479-485.
Tesfaye M, Temple S J, Allan D L, Vance C P and Samac D A
2001 Over-expression of malate dehydrogenase in transgenic al-
falfa enhances organic acid synthesis and confers tolerance to
aluminum. Plant Physiol. 127, 1836-1844.
Vance C P 2001 Symbiotic nitrogen fixation and phosphorus acquis-
ition. Plant nutrition in a world of declining renewable resources.
Plant Physiol. 127,390--397.
Vance C P and Heichel G H 1991 Carbon in N2 fixation: limitation
or exquisite adaptation? Annu. Rev. Plant Physiol. Plant Mol.
BioI. 42, 373-392.
Watt M and Evans J R 1999 Proteoid roots: physiology and
development. Plant Physiol. 121,317-323.
Fertilizer use with regard to optimum yield and environment
 ... Plant and Soil 247: 143-149, 2002.
143
" © 2002 Kluwer Academic Publishers.

Precision agriculture: a challenge for crop nutrition management

P. C. Robert*
Precision agriculture Center, University of Minnesota. 1991 Upper Buford Circle, St. Paul, MN 55108, USA

Received 11 October 2001. Accepted in revised fonn 15 August 2002

Key words: corn, crop nutrition, precision agriculture, site-specific management, soybean, wheat

Abstract
Precision agriculture was initiated in the mid 1980s, using newly available technologies, to improve the application
of fertilizers by varying rates and blends as needed within fields. Presently, the concept has been adapted to a variety
of practices, crops, and countries. Its adoption varies significantly by cropping system, regions, and countries but
it is progressively introduced or evaluated around the world. Several types of challenges limit a broader adoption:
socio-economical, agronomical, and technological. Socio-economical barriers are principally costs and lack of
skills. Agronomical challenges are lack of basic information, inadequate sampling and scouting procedures, ab-
sence of site-specific fertilizer recommendations, misuse of information, and lack of qualified agronomic services.
There are multiple technological barriers that relate to machinery, sensor, GPS, software, and remote sensing.
However, these barriers will be progressively lifted and precision agriculture will be a significant component of the
agricultural system of the future. It offers a variety of potential benefits in profitability, productivity, sustainability,
crop quality, food safety, environmental protection, on-farm quality of life, and rural economic development.

Introduction

Herb Dechant, farmer in Oberlin, Ohio, said "Even

though I would be considered an 'old timer' due to
my years in farming, I am not going kicking and
screaming into the 21 st century, fearing the changes
that computerization will bring us. In fact, I look for-
ward to the future of agriculture. It is my opinion that,
no what how many acres you farm or what type of
farm you operate, in order to keep up with current Figure 1. Soil spatial variability in Minnesota.
and future demands, you must make it a priority to
acquaint yourself with the techniques involved in pre-
cision agriculture and begin incorporating them into
your method of farming. It is the future for farmers outcome was to perceive potential benefits of crop
(Dechant, 1999)." management by zones within fields rather than whole
Better field investigation methods including soil fields for increased profitability and environmental
survey, soil sampling, aerial photography, and crop protection. At the same time, new technologies - mi-
scouting, from the mid 1970s into the early 1980s, crocomputers, geographic information systems (GIS),
resulted in a better awareness of soil and crop condi- and global positioning systems (GPS) - became avail-
tion variability within fields (Figure 1). An important able and made possible the acquisition, processing,
and utilization of spatial field data as well as the de-
* FAX No: +1-612-624-4223. velopment of a new kind of farm machinery with
E-mail: probert@soils.umn.edu computerized controllers and sensors (Figure 2).
144
Figure 2. Modem fertilizer applicator computerized system (From AGCO/AgChem).
This was the beginning of a new agricultural
management concept called initially 'farming by soil
types' then 'site-specific management (SSM),. Today,
it is commonly named 'Precision Farming or Preci-
sion Agriculture'. It quickly generated strong interest,
mainly because of associated new technologies, the
concept made good sense, and it offered new routes
for agri-industries and agri-businesses. The concept of
SSM was initiated for application of fertilizers to com
(Zea mays L.), soybeans (Glycine max (L.) Merr.), and
wheat (Triticum aestivum L.) crops in the Midwest,
Plain, and Northwest regions of the U.S. Presently,
it has been adapted to a variety of practices, crops,
and countries. From variable rate applications of fer-
tilizers, it now encompasses all management practices
on a spatial and temporal basis: sampling, tillage,
planting (rate and variety), fertilizing (rate, blend, and
type including manure), crop protection product ap-
plications (rate and mix), harvesting, and irrigation
(Figure 3). The SSM concept has been considered for
most common cropping systems and some specialty
crops, worldwide. From com, soybean, and small
grains, it now applies to a variety of plants such as rice
(Oryza L.), cotton (Gossypium L.), peanut (Arachis
hypogaea L.); sugar beet (Beta vulgaris L.) and sugar
cane (Saccharum L.); potato (Solanum L.) and ve-
getables; orchards and vineyards; and more recently
tropical crops such as palm oil (Elaeis guineensis L.)
and banana (Musa paradisiaca L.). Furthermore, it is
entering new domains such as precision turf manage-
ment, precision pasture and range management, and
precision tree management. From the Midwest, it has
entered most U.S. states and Canadian provinces; most
countries of western Europe and it is considered in
some eastern European countries; South and Cent-
ral America, particularly Argentina, Brazil, Chile,
Costa Rica, and Mexico; Australia and New Zealand;
and Asia, for example, Japan, Malaysia, and Indone-
sia. Some research and development projects are now
starting in China.
Precision Agriculture (PA) is not just the addition
of new technologies but it is rather an information re-
volution, made possible by new technologies that res-
ult in a higher level, a more precise farm management
system. PA is the application of a holistic management
strategy that uses information technology to bring data
from multiple sources to bear on decisions associated
with agricultural production, marketing, finance, and
personnel (Olson, 1998). In addition to increased prof-
itability, as a result of better management practices and
the development of farm information systems, PA can
bring additional benefits such as:
• increased crop quality,
• improved sustainability,
• lower management risk (Lowenberg-Deboer and
Aghib,200l),
• food safety associated to product traceability,
• environmental protection,
• rural development through new skills transferable
to other activities.
If the emphasis is made on the technology or in-
formation technology aspect of PA, it must be noted
that the basic concept of PA can be applied at differ-
ent technology levels (Blackmore et aI., 1994). Spatial
variability, natural or man-made, exists in most any
fields, whatever the size, and crop production can
benefit from improved basic information and SSM.
However, presently, the adoption in the U.S. of
PA for fertilizer management is quite variable, from
a few percent to as much as 40 percent (of tillable
land) in some regions such as the sugar beet grow-
ing area of the Red River Valley in Minnesota and

Figure 3. The concept of precision agriculture can be applied to most agricultural practices based on information technology (from the Potash
and Phosphate Institute).
North Dakota (Robert, 1999). The adoption of the
concept, including for crop nutrition, has been slower
than initially thought because of several kinds of signi-
ficant challenges: socio-economical, agronomical, and
technological.
Challenge: socio-economical barriers
Costs
The implementation of PA practices on the farm gen-
erates additional costs considered as excessive by
many farmers, particularly at a time of low commod-
ity prices. This has been repeatedly demonstrated by
an annual survey on the adoption of PA in the U.S.
sponsored by the Purdue University Center for Ag-
ricultural Business and the Farm Chemical magazine
(Akridge and Whipker, 1999). The 1999 survey indic-
ates that principal barriers of the adoption of precision
practices by farmers (Table 1) are cost (61%), slow
145
adopter (34%), cropping program not adapted (24%),
and lack of skills (19%).
Most common expenditures include new services;
acquisition of new technologies such as sensors for
yield monitoring, GPS receiver and correction sig-
nal subscription, GIS software for data storage, pro-
cessing, analysis, and making prescriptions; new or
modified farm equipment; and a high-quality com-
puter. A major barrier for site-specific fertilization is
the cost of making a satisfactory field soil nutrient
status map. Another service cost is data processing
for the development of site-specific application maps.
Most farmers do not have skills or time to do it. In
some U.S. regions and other countries, an additional
important barrier is the lack of agronomic services.
Variable rate fertilization costs are higher. When ap-
plied by local dealers, the cost for variable rate ap-
plication of fertilizers is generally between $12 and 25
per hectare depending on the number of services and
products sold to the farms . When applied by farmers,
new or upgraded equipment must be purchased.
146
Table 1. Barriers to the adoption of precision agriculture practices
by producers (from Akridge and Whipker, 1999)
1999
% of respondents
Cost 61
Slow adopters 34
Cropping program not adapted 24
Lack of management expertise 19
Cost-benefit analyses
Innovative producers that have used PA for several
years, and paid the additional costs, have generally
benefited from increased productivity and profitabil-
ity. Unfortunately, most farmers do not consider the
additional PA costs as an investment and don't perform
appropriate cost-benefit analyses.
Skills
PA adoption surveys (Akridge and Whipker, 1999;
NRC, 1997) have consistently found that age, attitude,
and education of producers are significant barriers.
They show that a majority of them are over 55 year old,
only have a partial or complete high school education,
and have limited interest in changing substantially
cropping practices and starting to use computerized
systems. Presently, until the new computer educated
generation arrives on the farm, only the more progress-
ive, innovative farmers have adopted PA. PA requires
new skills and only a limited number of farmers have
them or willing to get them. This may explain in
part why the adoption of PA in the U.S. seems to
have reached a plateau but is still higher than Europe
(Pedersen et aI., 2001).
Challenge: Agronomical barriers
Early adopters that applied some careful and effi-
cient PA practices such as yield mapping, selected
site-specific practices, including crop nutrition, and
precise farm information systems, have in most cases
benefited financially from improved soil and input
management, crop yield and quality, and marketing.
They have progressively modified some of their prac-
tices using site-specific information. However they
have found that present PA agronomics are still in
infancy and have major barriers.
Basic soil information
Present country soil surveys commonly have scales of
1:20000 or smaller. In the U.S., the standard county
U.S.DA-NRCS soil survey (1:12000-1:24000 scale)
available for most agricultural land was designed for
general planning use. It should not be used for PA
management, particularly for crop nutrition, because
of dissimilar soil inclusions, with significant different
soil characteristics, within mapping units and man-
made variability. There is a need to develop new
mapping methods based on new technologies (GIS,
GPS, remote sensing, terrain modeling) to create more
detailed soil maps (e.g., 1:5000 scale) (Johnson and
Robert, 1998).
Basic terrain information
PA research results show the value of the landscape or
Digital Elevation Model (DEM) for site-specific man-
agement, particularly the development of management
zones. Strong correlations between landscape posi-
tions and soil properties, fertilizer requirements, weed
distributions, and yields have been found (Robert et
al., 2000). For example, Nugteren et al. (1999) found
that protein levels exhibit less variation on the side
slope landscape positions (Figure 4). But topographic
maps at an appropriate scale are not readily available.
They must be surveyed, field by field, using topo-
graphic survey equipment, high-quality differential
GPS, or costly remote sensing techniques.
Soil sampling
Presently, the most common practice is grid sampling.
A recent U.S. survey on the adoption of PA (Akridge
and Whipker, 1999) indicates that 36% of deal-
ers/consultants used a sampling grid size greater than
1 hectare to lower costs. But research indicates that
resulting nutrient status maps provide inaccurate spa-
tial nutrient needs if the sampling size is greater than
0.4 to 1 hectare (Mallarino and Wittry, 2000). Another
common practice in the Midwest for selecting soil
sampling sites, 31% of dealers (Akridge and Whip-
ker, 1999), is the use of the standard soil survey at
the 1:20,000 scale. This practice reduces the num-
ber of soil samples. But the soil survey was never
intended for site-specific use and will generally res-
ult in a poor characterization of soil nutrient status
because of soil type variability within map units and
past field management. Presently, a preferred method
is directed or smart sampling based on management

Protein cokriged wI Elevation
Figure 4. Relationship between corn protein content and landscape position.
zones (Franzen et aI., 2000). However, this method
requires basic information such as soil characteristics,
landscape parameters, yield maps, crop management
historical and present information, and aerial imagery.
Crop monitoring
Monitoring of field crops during the growing season
for soil water status, nutrient needs, particularly ni-
trogen, pests (weeds, insects, etc.) problems, crop
maturation, varies greatly in intensity and quality de-
pending on the crop, farm, and region. Most scouting
executed by producers and consultants are generally
quick visual observations executed from the road or
on a small fraction of the field. Sampling techniques
to map distribution and intensity of a crop need or
problem are rarely adequate.
Site-specific recommendations
The utilization of conventional fertilization rate re-
commendations from University Extension Services
or from agrobusiness laboratories do not optimize site-
specific nutrient management as demonstrated by a
number of experiments for various crops, soil land-
scape conditions, and regions. But nutrient recom-
mendations corresponding to within field site-specific
characteristics are rarely available (Pierce and Nowak,
1999). They require on farm experiments executed by
farmers while most of them do not have the tech-
nical know-how. This essential need will demand
new programs from universities and agricultural ex-
147
r= 0.5232 Protein
7.8%
7.5%
7.4%
7.0%
6.2%
tension services to train field agronomists and farmers
in implementing meaningful trials.
Inefficient management and use of agronomic
information
PA is an information rich based management. This is
creating major problems to farmers used to make de-
cisions from limited records and rules. Some farmers
have understood the value of information, have accu-
mulated several years of all kinds of data but do not
know how to utilized them while others used immedi-
ately limited data without careful analysis and corrob-
orating data. For example, yield map information has
been often misused for fertilizer recommendations.
Low yields can be the result of many factors other than
soil nutrient levels Deductions are often to quickly
made without careful study of site-specific natural re-
sources conditions, sensor calibration, software bias,
field management history, machinery performance,
season special events, weather status, and previous
yield maps.
Agroconsultants
PA requires new skills most often missing on farms .
Presently, most farmers depend on dealer services,
agroconsultants, and cooperative personnel to use site-
specific fertilizer management and detailed spatial in-
formation systems. If these services are often available
in the Midwest, they are generally rare or absent in
other regions and countries.
148
Challenge: technological barriers
The concept of PA was linked to technology develop-
ment from the beginning. The past 10 years, progress
has been made in many PA technological areas. How-
ever, there is still a very significant need to improve
present PA technologies and develop new ones for
machinery, sensor, GPS, GIS, software, and remote
sensing.
Farm equipment
The early PA fertilizer spreaders were developed for
the Midwest conditions using bulk blending for grain
crops. PA machinery is not always available for other
fertilization practices, crops, and local natural and
socio-economical conditions. Equipment standards
are still in development for PA computerized systems,
machinery, and software. Studies of the accuracy of
fertilizer application equipment based on spinner, air,
and injection seem to indicate a need for improvement.
Trained workers capable of using efficiently the higher
technology equipment and systems may be difficult to
find in rural areas. We found as-applied map, recorded
by the on-board VRT computer during applications,
of inputs showing portions of fields where no fertilizer
was applied and overlapping strips doubling the input.
Another very common problem is the lack of proper
calibration of the equipment.
Sensors
Sensors are an essential part of PA. Grain yield sensors
have demonstrated well the value of site-specific in-
formation and the potential for site-specific manage-
ment. However, similar barriers exist for other sensors:
they do not exist, do not fit the local needs, lack accur-
acy, and are poorly used or calibrated. For example,
reliable yield sensors, essential for crop nutrition, do
not exist yet for a variety of crops and are rarely ad-
equately calibrated. In my view, a very critical barrier
is the lack of a real time soil nutrient sensor to ef-
ficiently create soil nutrient maps (Shibusawa et aI.,
2000).
CPS
GPS receivers are readily available for agricultural use
but the correction signal, required for PA, may not
be easily available without a costly annual subscrip-
tion. Advanced PA farmers are waiting for a higher
precision signal, in the centimeter range, for better
placement of inputs, faster applications, and night
operations.
Software
Software for data acquisition, processing, and analysis
in a GIS environment has made significant progress
in ease of use, openness, spatial analysis and display
(3D) but much more is still needed. In particular, farm-
ers and consultants are progressively overwhelmed by
the amount of information. Expert systems or decision
tool systems will have to be developed to optim-
ize the use of multi-layers and multi-years spatial
information.
Remote sensing
Research has shown for several decades that remote
sensing techniques have a variety of applications to
agriculture, including fertilizer management, using
ground, plane, or satellite based imageries. However
its use on the farm is still rare and faces mUltiple
barriers: lack of understanding of applications and
benefits, costs, availability of imagery (cloud cover),
quality and resolution of imagery, delay in getting
the imagery, and lack of skills in rural areas for the
interpretation of imagery.
Conclusion
Recent U.S. surveys seem to indicate a continuing
increase in the adoption of PA by dealers and farm-
ers, a significant variability by regions and crops,
and a recent slow down. In addition to worldwide
low prices for most crops, lowering farm profitabil-
ity and motivations to change management practices,
several challenges limit the use of PA practices. These
barriers are of socio-economical, agronomic ai, and
technological nature. They must be corrected before
PA adoption expands from the progressive farmers to
the majority of farmers. This is not specific to PA
but common to the adoption of new technologies. A
broad adoption of the tractor took more than 30 years.
When these barriers are lifted, site-specific practices,
including crop nutrition, will be an essential aspect of
the agricultural system of the 21st century because it
offers a variety of potential benefits in profitability,
productivity, sustainability, crop quality, food safety,
environmental protection, on-farm quality of life, and
rural economic development.

References
Akridge J T and Whipker L D 1999 1999 Precision agricultural
services and enhanced seed dealership survey results. Center
for Agricultural Business, Purdue University. West Lafayette. In
Staff Paper No. 99-6.
Blackmore B S, Wheeler P N, Morris J, Morris R M and Jones R
J A 1994 In Proceedings of the 2nd International Conference on
Precision Agriculture. Eds. PRobert et al. 27-30 March 1994.
Minneapolis, MN. ASA, CSSA, SSSA, Madison, WI.
Dechant H 1999 21st century fanning. In InfoAg99. 9-11 August
1999. West Lafayette, IN.
Franzen D W, Halvorson A D and Hofman V L 2000 Management
zones for soil N and P levels in the Northern Great plains. In
Proceedings of the Fifth International Conference on Precision
Agriculture. Eds. PRobert et al. 16-19 July 2000. Minneapolis,
MN. ASA, CSSA, SSSA, Madison, WI.
Johnson R K and Robert P C 1998. A new methodology for devel-
oping applied soil surveys for precision agriculture. pp. 265. In
Agronomy Abstracts. Baltimore, MD.
Lowenberg-DeBoer J and Aghib A 1999 Average returns and risk
characteristics of site specific P and K management: Eastern
cornbelt on-farm trial results. J. Prod. Agricul. 12 (2), 276-282.
Mallarino A P and Wittry D J 2000 Identifying cost-effective soil
sampling schemes for variable-rate fertilization and liming. In
Proceedings of the Fifth International Conference on Precision
Agriculture. Eds. pRobert et al. 16-19 July 2000. Minneapolis,
MN. ASA, CSSA, SSSA, Madison, WI.
149
National Research Council (NRC). 1997. Precision Agriculture in
the 21st Century. National Academy Press. Washington, DC.
Nugteren A A, Robert P C and Fairchild D 1999 Corn grain quality
as affected by soil properties, management, and landscape. In
Agronomy abstracts. ASA, CSSA, SSSA, Madison, WI.
Olson K 1998 Precision agriculture: current economic and envir-
onmental issues. Sixth Joint Conference on Food, Agriculture,
and the Environment. 31 Aug - 2 Sept 1998. University of
Minnesota, St. Paul, MN.
Pedersen S M, Ferguson R B and Lark R M 2001 A multinational
survey of precision fanning early adopters. Farm Management
11 (3), 147-155.
Pierce F J and Nowak P 1999 Aspects of precision agriculture. In
Advances in Agriculture. Ed. D L Sparks. pp. 1-85. Academic
Press.
Robert P C 1999 Precision Agriculture: Status and Research Needs.
In Proceedings of the 2nd European Conference on Precision
Agriculture. Ed. J Stafford. 12-15 July 1999. Odense, Danemark.
SCI, London, UK.
Robert P C, Rust R H and Larson W E (Eds.) 2000 Proceedings of
the Fifth International Conference on Precision Agriculture. 16-
19 July 2000. Minneapolis, MN. ASA, CSSA, SSSA, Madison,
WI.
Shibusawa W S, Made Anom I, Sato H, Sasao A and Blackmore
S 2000. In Proceedings of the Fifth International Conference
on Precision Agriculture. Eds. PRobert et al. 16-19 July 2000.
Minneapolis, MN. ASA, CSSA, SSSA, Madison, WI.
Plant nutrition and sustainable development
 Plant and Soil 247: 153-175,2002.
© 2002 Kluwer Academic Publishers.
153

Plant nutrient management for enhanced productivity in intensive grain

production systems of the United States and Asia

A. Dobermann 1 & K.G. Cassman

IDepartment of Agronomy and Horticulture, University of Nebraska P.O. Box 830915 Lincoln, NE 68583-0915,
USA. Corresponding author*

Key words: fertilizer, germplasm, maize, nutrient use efficiency, rice, site-specific management

Abstract
Are present nutrient management recommendations for the world's major cereal cropping systems adequate to
sustain the productivity gains required to meet food demand while also assuring acceptable standards of environ-
mental quality? To address this question, the current nutrient management approaches and their scientific basis in
large-scale, mechanized maize (Zea mays L.)-based cropping systems of the USA and more labor-intensive, small-
scale irrigated rice (Oryza sativa L.) production systems in Asia were evaluated. The principal challenges in both
systems are similar: (1) there is no compelling evidence for significant increases in the genetic yield potential in
both systems during the past 30 years, (2) farm yields are presently about 40-65% of the attainable yield potential,
and (3) nutrient management mostly relies on approaches that do not account for the dynamic nature of crop
response to the environment. Because average farm yield levels of 70-80% of the attainable yield potential are
necessary to meet expected food demand in the next 30 years, research must seek to develop nutrient management
approaches that optimize profit, preserve soil quality, and protect natural resources in systems that consistently
produce at these high yield levels. Achieving these goals will require novel strategies for more precise plant
nutrient management tailored to the technologies, dynamics and spatial scales relevant to each system. Significant
advances in soil chemistry, crop physiology, plant nutrition, molecular biology, and information technology must be
combined in this effort. Future field-oriented plant nutrition research must be of a more strategic, interdisciplinary,
and quantitative nature. Systems approaches at micro- to meso-scales are required for gaining a more quantitative
understanding of crop response to nutrients based on interactions among the essential crop nutrient requirements
and on response to dynamic environmental conditions.

Abbreviations: PFPN - partial factor productivity of applied N (kg grain per kg N applied); REN - apparent recovery
efficiency of applied N (kg increase in N uptake per kg N applied); SSNM - site-specific nutrient management

Introduction itability, and nutrient efficiency that can be exploited

Increases in future food production will largely come
from today's most intensively used agricultural land,
but these systems must also meet stricter environ-
mental standards. Hence, an ecological intensification
(Cassman, 1999) of cereal production systems is jus-
tified by concerns about food security, the availability
of adequate land and water resources (Young, 1999),
and protection of natural resources. It appears feasible
because of the existence of large gaps in yield, prof-
* FAX No.: +1-402-472-7904; E-mail: adobennann2@unl.edu
with greater precision of soil and crop management
(Cassman, 1999). The absolute increase in demand for
cereals during 1997-2020 will be as large (about 650
million metric tons) as the increase in demand during
the preceding 23 years even though annual relative
growth rates in world cereal demand are expected to
decline to about 1% per year (Rosegrant et aI., 2001).
There is considerable uncertainty in such estimates,
but a yield increase of 20-30% over a period of 20
years represents a significant challenge. For tropical
areas, average rice yields would have to increase to
154
about 80% of the climate-adjusted yield potential of
presently available germplasm (Dobermann, 2000).
The first objective of this paper is to compare nu-
trient management practices in two of the world's
most important cereal production systems: rainfed and
irrigated maize-based cropping systems in the north-
central USA and irrigated rice systems in south and
southeast Asia. Secondly, we review plant nutrition
issues that are relevant for raising yields to levels that
may provide an optimal combination of minimal en-
vironmental impact, greatest profit, and sustainable
food production for both systems. Specific questions
addressed include: (i) What is the potential for signi-
ficant improvements in nutrient use efficiency through
germplasm improvement and how will other trends
in crop genetic improvement affect nutrient manage-
ment? (ii) Is current scientific knowledge of crop
response to nutrients and environmental conditions
sufficiently robust to make significant improvements
in nutrient use efficiency? (iii) How can farmers better
account for spatial and temporal variation in indigen-
ous nutrient supply and crop nutrient demand? (iv)
Can productivity be raised further without significant
negative environmental impacts?
Nutrient management in intensive maize systems
in the USA
Rainfed and irrigated systems in which maize (Zea
mays L.) is grown either in rotation with soybean
(Glycine max L.) or as a continuous monocrop are
the predominant cropping systems in North America.
About 30 million ha of maize are harvested annually
for grain in the USA, of which eleven states in the
Corn Belt produce more than 210 million t or 35%
of global maize supply (Table 1). Environmental con-
ditions in this region are favorable, soils are deep and
fertile, input use is relatively high, and farms are large.
Individual fields are typically >50 ha and maize is
produced in highly mechanized systems with an av-
erage labor input of only 6 h ha- I per crop. During
the past 35 years, average maize yields have increased
linearly at a rate of 109 kg ha- I per year (Figure 1),
mainly due to the adoption of improved crop manage-
ment technologies and genetic improvement of maize
hybrids that complements these management practices
(Duvick and Cassman, 1999). Average maize yields
now approach 9 Mg ha- I , but progressive farmers
routinely harvest 11-13 Mg ha -I. Despite this steady
yield gain, however, present maize yields are only
about 40-50% of the estimated climate-adjusted ge-
netic yield potential of current maize hybrids and there
is little compelling evidence that yield potential has in-
creased significantly in the past 30 years (Duvick and
Cassman, 1999).
Fertilizer recommendations are based on soil test-
ing in about half of the maize area. The use of anhyd-
rous ammonia and fluid N fertilizers as the primary N
sources is another unique feature of maize systems in
the USA, which account for 80% of global anhydrous
ammonia consumption and 67% of nitrogen solutions
(IFA, 2000). Fertilizer rates used on maize are typic-
ally within ranges of 94-185 kg N ha -I, 10-34 kg P
ha- I , and 0-86 kg K ha- I (Table 2) but large differ-
ences exist among states and among farms within each
state (Padgitt et aI., 2000). For example, average K use
in the top 11 maize states ranges from 10 kg K ha -I in
Nebraska to 120 kg K ha- I in Indiana, average N use
from 100 kg N ha- I in Wisconsin to 180 kg N ha- I
in Illinois (Figure 2). Regional relationships between
grain yield and fertilizer rate tend to be inconsistent
because fertilizer management decisions are largely
made before planting and are not adjusted during the
growing season in response to climatic conditions (N)
or are based on long-term management programs (P,
K). For example, in the U.S. Corn Belt positive cor-
relations were observed between grain yield and rates
of N, P, and K applied to maize in 2000, whereas
no such relationship was observed in 1999 although
the average maize yield was similar in both years
(Figure 2).
Commercial fertilizer use rose sharply in the 1960s
and 1970s in response to the adoption of responsive
maize hybrids and favorable economic forces (Uri,
1998). However, maize yield increases since 1980
were achieved with stagnating fertilizer-N use and
declining rates of P and K, leading to significant
increases in the partial factor productivity (PFP, kg
grain per kg nutrient applied) of these macronutrients
(Figure 1). Average grain output per unit N applied
increased from 41 kg kg-I in 1980 to 58 kg kg-I in
2000. Three factors have probably contributed to the
improvement in N fertilizer efficiency: (i) increased
yields and more vigorous crop growth associated with
increased stress tolerance of modern hybrids (Duvick
and Cassman, 1999), (ii) improved management of
production factors other than N such as conservation
tillage, seed quality and higher plant densities, and
(iii) improved N management. Improvements in N
management include some reductions in fall-applied
N fertilizer with a shift to applications in spring or at
 155

";'
9
co (a) Grain yield
.c: 8

-
0)
::iE 7
32 6
Q)
'>,
I: 5
=109 kg ha- 1 y(l
...co
(!)
4
b
~ =0.77

160 70
140 ";'

60 .x::0)
120
-a.z
0)
.x::
100 50
80 40 LL
on a.
uv
";'
30
CO
..c:
~,....
40 (c) Phosphorus 350
-
co 0)
.c: .x::
0)1:
35 ";'
.x::
-co
,_ 300 .x::0)
30

- -
Q) ... 0)
250 .x::
...Q)~.c:'i 25
- 0)

a.
,!::! 20 a.
200 LL
co 15 a.
1::Q) >0
150
LL E 10
Q)
0::
80 180
70 160':-
0)
60 140 .x::0)
50
40 \ b =3.65 kg kg- 1 y(l
120 .x::
~
100 a.
-
30 ~ =0.71 LL
r~../'-./A...-/------ 80 a.
20 --------'---/~'v/-----"v
10 60
1965 1970 1975 1980 1985 1990 1995 2000
Figure 1. Trends in grain yield, fertilizer use (filled circles), partial factor productivity of fertilizer nutrients (open circles, PFP =kg grain yield
per kg nutrient applied), and nutrient removal with grain (dashed line, kg element ha -\ ) in maize grown in the USA. Trend lines were fitted to
the period from 1980 to 2000. Yield data: Mean annual maize yields, National Agricultural Statistics Service, USDA http://www.usda.gov/nass;
Fertilizer data: Mean N, P, and K amounts applied to maize, USDA Annual Cropping Practices Surveys of more than 2000 farms representing
80-90% of the maize area, http://www.ers.usda.gov. Nutrient removal with grain was calculated by assuming average concentration of 1.4% N,
0.27% P, and 0.35% K in grain.

planting, greater use of split N fertilizer applications that give N 'credits' for manure, legume rotations,
rather than a single large N application, and develop- and residual soil nitrate (Shapiro et aI., 2001). In ad-
ment and extension of N fertilizer recommendations dition, nitrification or urease inhibitors are used on
156
Table 1. Comparison of intensive rice and maize systems. Values shown refer to the main grain crop in the
cropping system, i.e., maize for the USA and rice for Asia

Irrigated and rainfed maize Intensive irrigated rice

north-central USA 1 south and southeast Asia

Predominant cropping systems 2 Single crop Double and triple crop

M-S, M-M, M-M-S, M-S-S R-R, R-W, R-R-R, R-R-M
Estimated land area (million ha)3 45 41
Annually harvested area (million ha)4 25 66
Share of global production (%) 35 58
Average grain yield (Mg ha- 1 per crop)5 8.7 5.3
Annual grain production (Mg ha- I year- I )6 7-11 8-12
Grain yield (% of yield potential) 7 40-50 60-65
Cropping technologies large fields small fields
mechanized labor-intensive
Average labor use (h ha -I per crop) 8 6 115-2150

1 Maize: top 11 maize-producing states in the Corn Belt that account for 85% of the U.S. maize and soybean
production (Iowa, Illinois, Nebraska, Minnesota, Indiana, Ohio, South Dakota, Kansas, Wisconsin, Missouri,
Michigan), 1999-2000 (NASS, 2001).
2 R - rice; W - wheat; M - maize; S - soybean. Rice: two to three crops grown per year. Maize: continuous
maize or 2 to 3-year crop rotations with only one crop grown per year.
3 Estimated area occupied by the predominant cropping systems. Rice: 24 million ha double- and triple-crop
continuous rice systems (Huke and Huke, 1997) and 17 million ha irrigated rice-wheat systems. Irrigated rice-
wheat area was estimated to be 7.5 M ha in China, 7 M ha in India, 1.6 M ha in Pakistan, 0.75 M ha in
Bangladesh, and 0.5 M ha in Nepal. These estimates were revised from previously published numbers (Ladha
et aI., 2000b; Timsina and Connor, 2001; Woodhead et al., 1994) by (a) taking into account recent declines in
the rice and wheat areas in China by about 2-3 M ha as well as small increases in other countries (FAO, 2001)
and (b) assuming that about 65% of the 10.8 M ha R-W area in India is fully irrigated (Woodhead et aI., 1994).
Maize: Estimated from harvested areas in different crop rotations: 4 M ha continuous maize + 2 x 13.5 M ha
maize-soybean rotation + 2 x 7 M ha maize in other rotations = 45 M ha land area.
4 Rice: Total world harvested area of irrigated rice is about 76 M ha. Of this, about 10 M ha is irrigated rice
grown in temperate climate (single crop, about 9 M ha) and irrigated rice grown in cropping systems other than
those included here (FAO, 2001; IRRI, 1997). Maize: Annually harvested corn area. Of this, about 4.5 M ha
is continuous maize, 13.5 M ha is maize grown in annually alternating maize-soybean rotation, and 7 M ha is
maize grown in other crop rotations (Padgitt et al., 2000).
5 Average yield in 2000 based on regional production statistics (FAO, 2001; NASS, 2001). Irrigated rice yield
was calculated by assuming a 56% share of the total rice area.
6 Typical range of annual grain production. Rice: two to three crops per year, 25-75% quartile range of 205
farms (Dobermann, 2000). Maize: one crop per year, yield range achieved by most farmers.
7 Rice: assuming an average simulated climatic yield potential of about 8.1-8.5 Mg ha- 1 (Matthews et al.,
1995). Maize: assuming a yield potential of about 18-22 Mg ha- I achieved in field plots with near-optimal
~rowth (Duvick and Cassman, 1999).
Includes paid and unpaid labor. Rice: range of average labor use in seven key irrigated rice domains of south
and southeast Asia (Moya et aI., 2002). Maize: average of maize farms in the north-central region of the USA
surveyed in 1996 (Foreman, 2001).

about 14% of the maize area (Table 2). Despite the
progress made in increasing N use efficiency, recent
on-farm data indicate that on average only 37% of the
applied fertilizer-N is taken up by maize (Cassman
et aI., 2002). Management control points for N are
different for irrigated and rainfed maize, but on-farm
data are not available to evaluate differences in N use
efficiency in more detail. Recovery efficiencies of ap-
plied N (REN, kg increase in plant N accumulation per
kg N applied) also are highly variable because almost
80% of the N is applied before crop emergence, which
makes it vulnerable to losses during the crop estab-
lishment phase before the crop can establish an active
root system. Only 14% of the maize area receives split
applications of N after planting (Padgitt et aI., 2000).
During the past 35 years maize farmers have made
considerable investments in soil conservation meas-
ures and in building soil fertility through P and K
applications that exceeded crop removal (Figure 1).
Only grain is removed and all crop residues are re-

- 10 (a)
';"111
~ 0 0 ·
~9 o~.
• ••~.
"i. o. 0
.~ 8 • .0
"e
C) 7 .0
100 120 140 160 180
Average N rate (kg ha·1)
2000: b=0.016, ~=0.34
1999: b=0.002, ~=0.01
Figure 2. Correlation between maize yield and the average rates of fertilizer nutrients applied in the top 11 maize states of the USA in 1999 and
2000. Yield data: Mean annual maize yields, National Agricultural Statistics Service, USDA http://www.usda.gov/nass; Fertilizer data: Mean
N, P, and K amounts applied to maize, USDA Annual Cropping Practices Surveys, http://www.ers.usda.gov.
cycled. At current average yield levels, maize grain
removes 105 kg N, 20 kg P, and 26 kg K ha -1 per crop.
Fertilizer rates used by maize farmers in the USA com
belt since 1965 exceeded this net nutrient removal, but
the difference is declining in recent years (Figure 1).
For example, the average P surplus decreased from 13
kg P ha- 1 per crop in 1980-1984 to just 4 kg P ha- 1
per crop in 1996-2000. Since the late 1970s, USA
maize farmers have been taking advantage of residual
soil P and K supplies built up by previous nutrient
applications (Uri, 1998), but large differences exist
within the region. Across the Com Belt, about 50% of
all soil samples analyzed each year test in the medium
or higher soil test P categories (Bray-l P equivalent
larger than 25 mg kg-I). This average proportion has
remained virtually unchanged since 1975, but ranges
from 13% in South Dakota to 83% in Michigan (PPI,
2001). More than 50% of all soils test above 160 mg
K kg- 1 (1 N NH4 -acetate equivalent, range from 23%
in Michigan to 89% in Wisconsin). Average soil test
K levels have declined since 1980 in the eastern part
of the Com Belt (Illinois, Indiana, and Ohio), whereas
they increased in Iowa and Minnesota or remained un-
changed in states with large native soil K reserves such
as Nebraska or Kansas (PPI, 2001).
Nutrients supplied with farmyard manure affect the
general trends shown in Figure 1. At present, 17%
of the maize area and 6% of soybean area receive an
application of livestock manure, but maize-based sys-
tems account for 89% of the total manured area in the
USA (Padgitt et al., 2000). The top 11 maize states
shown in Table 1 produce about 425 000 tons of recov-
erable manure-N (manure nutrients available for land
application after deduction of losses due to storage and
transportation) and 270000 tons recoverable manure-
157
(b) (c)
• 2000
o 1999
15 20 25 30 35 40 20 40 60 80 100 120
Average P rate (kg ha·1) Average K rate (kg ha·1)
2000: b=0.086, ~=0.73 2000: b=0.017, ~=0.57
1999: b=0.006, ~=O.OO 1999: b=0.002, ~=0.01
P, which is equivalent to roughly 40% of the available
manure nutrients for land application in the USA (Kel-
logg et al., 2000). Assuming a similar share during
the past 20 years and an increase in the total manure
amount of roughly 20% from 1982 to 1997 (Kellogg
et aI., 2000), the average annual manure nutrient in-
put on land used for maize-based cropping systems
increased from 6 kg P and 10 kg N ha- 1 in 1982 to
just 7 kg P and 11 kg N ha -1 in 1997. Because manure
is mostly applied on arable land in close proximity to
livestock operations, the actual rate of nutrients ap-
plied in manure is typically quite high on manured
land. In states where manure production is high re-
lative to the available land area and crop nutrient
removal, the percentage of soil samples testing in high
and very categories of soil test P has been increasing in
recent years (PPI, 2001). In Michigan and Wisconsin,
for example, 40-50% of all soil samples analyzed test
higher than 50 mg P kg- 1 (Bray-l P equivalent, PPI,
2001). Comprehensive nutrient management planning
has become one of the key environmental and eco-
nomic challenges for the Com Belt region because
of the continuing trend towards large-scale industrial
livestock production systems (Gollehon et al., 2001).
Nutrient management in intensive rice systems in
Asia
Irrigated double- and triple-crop rice systems occupy
about 41 million ha in south and southeast Asia and
contribute about 58% of global rice supplies (Table 1).
There are at least 50 million irrigated rice farms in
Asia because farm size is typically small, ranging from
0.3 ha in densely populated areas such as the Red
158
Table 2. Nutrient management practices and nitrogen use efficiency in intensive rice and maize systems.
Values shown refer to the main grain crop grown, i.e., maize for the USA and rice for Asia

Irrigated and rainfed maize Intensive irrigated rice

north-central USA I south and southeast Asia

Fertilizer recommendations State-specific, often soil Country-specific, blanket

test-based for large areas
Sample grain yield (Mg ha -I crop -I ) I 9.4 (7.5-10.4) 5.2 (4.0-5.9)
Fertilizer-N use (kg N ha- I crop-I)I 146 (94-185) 111(86-138)
Fertilizer-P use (kg P ha- I crop-I)I 22 (10-34) 18 (11-25)
Fertilizer-K use (kg K ha- I crop-I)I 49 (0-86) 17 (0-46)
Predominant N application technology NH3, UAN solution Urea
knife, dribble, broadcast Broadcast
Soil testing (% of area)2 46 Negligible
Plant tissue testing (% of area)2 2 Negligible
Nitrogen inhibitor used (% of area)2 14 Negligible
Number of N applications per crop3 1.8 2.6
N applied before crop emergence (%)4 77 33
Partial productivity of N (kg kg -I) 1,5 60 (48-80) 45 (32-59)
Recovery efficiency of N (%)6 37 (30) 31 (18)

I Rice: medians and 25-75% quartile ranges of 207 rice-rice and rice-wheat farms in India, China,
Thailand, Vietnam, Indonesia and the Philippines surveyed from 1995 to 1997 (Dobermann, 2000).
Maize: medians and 25-75% quartile ranges of a sample of 4712 farms in Iowa, lllinois, Nebraska,
Minnesota, Indiana, Ohio, South Dakota, Wisconsin, Missouri, and Michigan surveyed in 1994 (USDA,
1994 Cropping Practices Survey, unofficial data files, http://www.ers.usda.gov). Note that 1994 was a
high-yielding year in most states.
2 Rice: no exact numbers are available, but this practice is not common. Maize: average of
1859 farms surveyed in 1999 (USDA, 1999 Cropping Practices Survey, unofficial data files,
http://www.ers.usda.gov).
3 Rice: average of farms at seven sites in China, India, Indonesia, Philippines, Thailand, and Vietnam
and range of means among sites, 1995-97 (Moya et aI., 2002). Maize: average of 1922 farms surveyed
in 2000 (USDA, 2001).
4 Nitrogen applied before or at planting (% of total N amount). Rice: average of farms at seven sites in
China, India, Indonesia, Philippines, Thailand, and Vietnam and range of means among sites, 1995-97
(Moya et al., 2002). Maize: average of 1859 farms surveyed in 1999 (USDA, 1999 Cropping Practices
Survey, unofficial data files, http://www.ers.usda.gov).
5 Partial factor productivity of fertilizer N = kg grain yield per kg N applied.
6 Recovery efficiency is the proportion of applied N fertilizer that is taken up by the crop and is
determined by the difference in the total amount of N measured in aboveground biomass at maturity
in replicated plots that receive N fertilizer and a control plot without applied N. Rice: mean and standard
deviation of four consecutive rice crops at 179 sites in key irrigated rice domains of Asia, 1997-2000
(Dobermann et al., 2002). Maize: mean and standard deviation of 38 on-farm experiments conducted
in Illinois, Michigan, Minnesota, Missouri, Nebraska, and Wisconsin during 1995-1999 (North Central
Regional Research Project NC-218; Cassman et al., 2002).

River Delta of North Vietnam to more than 4 ha in
areas of southern and northern India and central Thai-
land (Moya et aI., 2002). Individual fields are even
smaller (0.2-0.5 ha) so that only small machinery can
be used. Labor input is high, ranging from 115 h ha- I
per crop in areas where rice is direct-seeded to more
than 2000 h ha -I per crop in transplanted rice fields
(Table 1). Favorable climate and access to irrigation
water allow farmers to grow two to three crops each
year. Although soils that support these systems vary
widely in quality, relatively high levels of fertilizer and
pesticide use are typical in most intensive rice pro-
duction systems. Rice-rice and rice-wheat (Triticum
aestivum L.) are the two dominant cropping systems.
Rice yields in Asia increased at an average rate
of 2.5% yeac l from 1967 to 1984, but yield growth
rates dropped to 1.2% from 1984 to 1996 (Dawe and
Dobermann, 1999). In some large rice production do-
mains where farmers were early adopters of modern
irrigated rice production technologies, yields appear

to have stagnated since the mid-1980s (Cassman and
Dobermann, 2001) although the current average ir-
rigated rice yield of 5.3 Mg ha- 1 per crop is only
60-65% of the climate-adjusted yield potential across
Asia (Table 1). Rice accounts for 15-85% of the total
fertilizer consumption in major rice-producing coun-
tries in Asia, but accurate numbers and their changes
over time are difficult to obtain. Published estimates
of fertilizer use on rice are either derived from es-
timated shares of rice in total fertilizer consumption
(Hossain and Singh, 2000) or expert opinions about
rates applied to different crops (IFA, 1999). It is gener-
ally concluded that the impressive gains in rice yields
during the 1960s and 1970s were associated with in-
creased use of fertilizers, particularly urea-N, whereas
growth in fertilizer consumption has slowed in recent
years (Hossain and Singh, 2000).
Regular surveys of rice farms in Central Luzon,
Philippines conducted by the International Rice Re-
search Institute since 1966 represent one of the few
sources of on-farm data on fertilizer use trends on rice
in Asia (Figure 3). After an initial steep rise due to the
adoption of modem varieties and fertilizers, rice yields
in Central Luzon during the past 20 years have fluctu-
ated around 3.2 Mg ha- 1 in the wet season and 4 Mg
ha- 1 in the dry season. Nevertheless, fertilizer use
continued to increase during the same period, leading
to a steady decline in PFP of applied nutrients since
the mid-1970s (Figure 3). Similar yield trends are ob-
served in other irrigated rice domains (Cassman and
Dobermann, 2001) so that it is likely that a stagnation
or even decline in PFP of fertilizer has become a com-
mon feature in rice systems of developing countries
in Asia. There is no indication that this has led to a
build-up of mineral N levels in the soil or an increase
in the indigenous N supply, suggesting that the extra
N remains in organic soil N forms that are less plant
available (Cassman et aI., 1998; Olk et aI., 1996).
A network for strategic on-farm research in key ir-
rigated rice domains of Asia was established in 1994
(Dobermann et aI., 2002). Data from this study show
that most irrigated rice farmers apply 86-138 kg N,
11-25 kg P, and 0-46 kg K ha- 1 crop-l (Table 2),
mostly broadcast by hand and with little use of spe-
cial products such as slow-release fertilizers or N
inhibitors. Typically, about one third of the N is ap-
plied before crop emergence and split application is
common. Nevertheless, fertilizer-N efficiency has not
increased substantially during the past two decades.
Average REN in the late 1990s was only 0.31 kg kg- 1
(Table 2), which compares to an average REN of 0.30
159
kg kg- 1 (0.26 kg kg- 1 in wet season and 0.33 kg kg- 1
in dry season rice) measured in 236 experiments con-
ducted with irrigated rice in Indonesia during the early
1970s (van Keulen, 1977). Only 20% of all farmers
achieve REN >0.5 kg kg- 1 which is comparable to
N efficiency typically measured in well-managed ex-
periments (Dobermann et aI., 2002). Large variability
in indigenous nutrient supplies among rice fields was
found to be a general feature of intensive rice systems
in Asia (Cassman et aI., 1996; Olk et aI., 1999; Wang
et aI., 2001). Fertilizer rates, particularly N, are typic-
ally not adjust to this spatial and temporal variability,
resulting in a lack of congruence between nutrient sup-
ply and crop demand, sub-optimal yield and low N use
efficiency (Cassman et aI., 1998).
The Green Revolution in Asia significantly altered
nutrient cycling in lowland rice systems so that con-
cern was raised about negative nutrient input-output
balances or other threats to sustainability such as
a general decline in soil quality (Greenland, 1997).
There is a paucity of data to address such issues be-
cause detailed on-farm studies of nutrient inputs and
outputs rarely exist and the few studies conducted
at research sites are unlikely to adequately repres-
ent the wide range of production environments where
rice is grown. An attempt to estimate the average
NPK input-output balance in an irrigated rice sys-
tem under present conditions is presented in Table 3,
but several assumptions must be validated in future
research. Most uncertain is the N input from biolo-
gical N fixation and nutrient inputs and losses from
sources such as manure, rain, and irrigation. Informa-
tion about net losses of nutrients from crop residues is
also scarce. Compared to maize systems in the USA,
larger amounts of rice crop residues are removed for
use as fuel or fodder, or burned to facilitate more rapid
planting of the subsequent rice crop.
Despite the low REN and large gaseous losses of
fertilizer-N in many farms, N is sequestered in intens-
ive rice systems with long periods of flooding. Field
experiments with more detailed measurements support
this conclusion and often suggest even larger positive
N balances than the one shown in Table 3 (Ladha et
aI., 2000a; Witt et aI., 2000). However, the accumu-
lating N is likely stored in organic matter pools that
are not easily plant available (Cassman et aI., 1998).
At comparable total soil N levels, the average indi-
genous supply of plant available N during a growing
season is almost 3-fold larger in a typical USA maize
field than in a lowland rice field of Asia (Cassman et
aI., 2002). Although N mineralization can be briefly
160

";" -ca
.c
5

-
C)
:::E
"C 3
'i
'>.
r::: 2 ___ Sample of farms, wet season
I! -0- Central Luzon region, wet season
C) -0- Central Luzon region, dry season

100 q (b) Nitrogen

80
\
\
b = 2.62 kg ha-1 y(1
r2 = 0.95 250
200
-
";"

~
C)

-z
\ C)

60 \ ___ Fertilizer rate 150 ~

\
\ --0- PFP
40 100 0..
LL
20 50 0..

-
";"
ca
.c 25 400
-
-...
";"
C)
~
20 350 ~C)
.!ca 15 300 ~
-
C)

...
Q) 10
"-
"-
250 0..
a..
,!:::! "-0------0... 200 LL
........
:e
Q)
5 ........
........
'0
150 0..
LL

14
12 /
/
/ '" (d) Potassium

~ '"
900
800 -
";"
C)

-
~
10 / 700
'" "
C)
/
8 / 600 ~

b =0.34 kg ha-1 y(1

6 c5 \ r2 =0.75 500 ~
0..
4 \ ...-0 __ 400 LL
\ 0..
2 b---- ---- ---0 300

1965 1970 1975 1980 1985 1990 1995 2000

Figure 3. Trends in grain yield, fertilizer use, and partial factor productivity of fertilizer nutrients (PFP =kg grain yield per kg nutrient applied)
in irrigated rice areas of Central Luzon, Philippines. Rice is grown in continuous annual double crop systems (dry season and wet season).
Yield data (paddy): Mean of regular loop surveys of 58-146 farms and regional yields obtained from official statistics. Fertilizer data: Mean of
loop surveys of 58-146 farms conducted by IRRI (data provided by D. Dawe and M. Hossain, Social Sciences Division, IRRI).

accelerated by measures such as soil drying or drain- et al., 2000). At the average farm level, phosphorus
age, there is little indication that the indigenous N applications generally appear to be in balance with rice
supply can be easily increased over time (Dobermann yield increases and P removal, although P deficiency
 161
Table 3. Estimated average input-output balance of N, P, and K in intensive rice systems of South
and Southeast Asia with an average yield of 5.2 Mg ha -1 (Dobermann and Witt, 2000)

Inputs and outputs 1 N (kg ha- 1 crop-I) P (kg ha- 1 crop-I) K (kg ha- 1 crop-I)

Inputs:
Fertilizer 117 18 17
Farmyard manure 5 2 5
BNF2 50 0 0
Outputs:
Gaseous losses 3 87 0 0
Net removal with grain 58 12 13
Net removal with straw4 20 2 35

Input-output balance +7 +6 -26

1 Estimates are based on medians of fertilizer input, apparent recovery efficiency of applied nutrients,
crop uptake, and crop residue amount measured for two consecutive rice crops in 207 farms in China,
India, Indonesia, the Philippines, Thailand, and Vietnam (1995-1997, Reversing Trends of Declining
Productivity in Intensive Irrigated Rice Systems, On-farm monitoring database, June 2000, IRRI.).
At most sites, surface water with low nutrient content is used for irrigation so that the assumption
was made that nutrient inputs from irrigation and rainwater are roughly equal to leaching losses
(Dobermann et al., 1998). Average nutrient concentrations in irrigation water samples collected at all
sites from 1999 to 2000 (N=125) were 2 mg N L -1,0.5 mg P L -1, and 3 mg K L -1. At an average
water use of 500 mm per crop (1000 mm irrigation in a dry season and no supplemental irrigation in
a wet season), this amounts to inputs of 10 kg N, 2.5 kg P, and 15 kg K ha- 1 crop-I, but leaching
losses were not measured.
2 An average input of N from biological N fixation (BNF) of 50 kg ha -1 was assumed (Koyama and
App, 1979), but actual BNF inputs vary from about 28 to 51 kg ha- 1 per crop (Cassman et al., 1998).
3 Gaseous N losses were estimated from the fertilizer and manure N input and the measured recovery
efficiency of applied fertilizer N, assuming that continuous fertilizer use does not result in a significant
increase in residual mineral N in the soil.
4 Net removal with straw includes nutrients lost due to removal or burning of crop residues and
was estimated from plant nutrient accumulation in straw, amount of residue remaining, and the
predominant crop residue management practice at each site.

still occurs in some areas. In contrast, potassium de-
ficiency is likely to become an emerging constraint
in many rice areas because there is a large negative
K balance (Table 3), which is consistent with results
from long-term experiments (Dobermann et aI., 1998).
Net removal of K from rice fields averages about 26
kg ha- 1 per crop or 52 kg ha- 1 on an annual basis in
double-crop systems. At a similar level of annual grain
production, the K input-output balance in maize sys-
tems of the USA appears to be about +30 kg ha -1 per
year because K application rates are larger and fewer
residues are removed.
Multifaceted future nutrient management needs
The comparison of USA maize and Asian rice sys-
tems suggests that the latter have lower nutrient use
efficiency and average yields that are closer to the
climate-adjusted yield potential than those of maize
in North America. Available on-farm data indicate
that the average REN is only 30% in rice and 37% in
maize, whereas recovery efficiencies of 50-80% can
be achieved in field experiments with good manage-
ment in both environments (Cassman et aI., 2002). As
expected, nutrient efficiency (expressed as PFP) ini-
tially dropped in both systems during the early years
of adoption of modem varieties/hybrids because fertil-
izer use increased rapidly from previously low levels.
However, PFP has increased since the early 1980s in
USA maize systems while the limited data available
suggest that it has not increased in the major irrigated
rice production domains of south and southeast Asia.
Differences in economic systems and government
roles in farm programs as well as improved technolo-
gies supported by a strong research and extension sys-
tem appear to be major reasons for the steady increase
in nutrient efficiency in USA maize systems. Im-
proved technologies include adoption of conservation
tillage, hybrids with multiple tolerances to stresses,
high seed quality, better weed control, soil testing
162
and locally calibrated fertilizer recommendations, and
timely planting and new fertilizer application tech-
niques made possible by rapid improvements in mech-
anized equipment. In contrast, rapid initial advances
in rice yields were mainly achieved through adoption
of 'seed-and-fertilizer packages'. While USA farmers
have conserved or even increased soil nutrient stocks,
rice farmers in Asia appear to be depleting soil K re-
serves. While USA farmers use high-quality hybrid
seed, rice farmers in Asia mostly rely on poorer-
quality local seed sources because of less private
sector involvement in seed production. The impact
of technological progress on nutrient use efficiency
is further illustrated by the trends observed in Japan.
There, in contrast to most other Asian countries, PFP
of macronutrients in irrigated rice has increased since
the early 1980s, when fertilizer use began to decline
because higher grain quality became important and en-
vironmental concerns stimulated the adoption of more
sophisticated management technologies such as deep
placement, splitting of N applications, models ac-
counting for N mineralization, and the increased use
of slow release fertilizers (Suzuki, 1997).
Despite these differences, similar challenges for
plant nutrition exist in these two major cereals pro-
duction systems because exploitable gaps in yield and
nutrient use efficiency between current farm averages
and attainable levels are similar. The need for more
precise and diverse methods of nutrient management
will be driven by a number of factors. First, germplasm
improvement will widen the range of nutrient man-
agement solutions required for specific needs. Second,
as yields of current varieties and hybrids approach
the yield ceiling, the margin of error between nutri-
ent excess and deficiency decreases markedly. This
is because of the non-linear nature of the relation-
ships between nutrient uptake and grain yield uptake
requirements per unit yield increase as yields ex-
ceed about 70% of the yield potential (Witt et aI.,
1999). Therefore, a greater quantitative knowledge
about crop response to nutrients and balanced plant
nutrition is required to manage crops at high yield
levels. Third, dynamic, site-specific nutrient manage-
ment of small units such as single fields or areas within
them will be required to overcome the current mis-
match of fertilizer rates and crop nutrient demand at
the farm level. Fourth, environmental concerns such
as pollution of drinking water by nitrate, eutroph-
ication of streams, lakes, and coastal marine envir-
onments, as well as net contributions of agricultural
systems to global warming will force the development
of improved management practices further enhancing
nutrient use efficiency.
Potential role of germ plasm improvement
In both USA maize and Asian rice systems, germ-
plasm improvement during the past 30 years has res-
ulted in greater yield stability as a result of substantial
increases in tolerance to abiotic and biotic stresses
(Duvick and Cassman, 1999; Peng et aI., 1999; Tollen-
aar and Wu, 1999). Less certain is whether there has
been a significant increase in yield potential. It is also
uncertain whether major scientific breakthroughs in
complex traits such as yield potential or traits that dir-
ectly confer improvements in nutrient uptake efficien-
cies or physiological requirements can be achieved.
Breeding genotypes that produce more grain per unit
nutrient uptake in the plant appears questionable be-
cause it is difficult to further increase the harvest
index and because relationships between crop growth
rates and internal nutrient requirements appear to be
tightly conserved (Burns et aI., 1997). In rice, ef-
forts are in progress to develop new plant types with
a 25% larger yield potential (Peng et aI., 1999), C4-
photosynthesis characteristics (Sheehy et aI., 2000),
or increased contributions from biological N fixation
(Ladha and Reddy, 2000). Because such traits are un-
der complex genetic control, it is not likely that these
efforts will have measurable impact in the near future.
There is little evidence of comparable efforts on maize
in either public or private sector research. Instead,
the private sector seed industry continues to focus on
improving yield stability and stress tolerance through a
multi-location selection process coupled with molecu-
lar approaches to incorporate specific traits for pest
and herbicide resistance, and end-use quality.
Over the short-term, conventional and molecular
breeding and biotechnology will probably contribute
most by facilitating the development of crop geno-
types with improved growth fitness and specific grain
qualities. Growth fitness traits that affect nutrient man-
agement include (i) rapid early vegetative growth to
reduce the period of inefficient resource use, (ii) root
architecture for increased soil exploration and nu-
trient acquisition, and (iii) tolerance to abiotic and
biotic stresses. Breeders can make contributions to im-
proving nutrient efficiency by developing genotypes
in which growth and spatial distribution of roots are
more congruent with the release dynamics and the
spatial distribution of nutrients in the soil. Genotypes

with more rapid leaf area development that increases
biomass accumulation during the crop establishment
phase are likely to improve N use efficiency by in-
creasing N uptake during the period of greatest soil
N supply immediately after planting.
Although genotypic variation in nutrient uptake
kinetics has been reported for rice (Teo et aI., 1995)
and maize (Baligar and Barber, 1979), field measure-
ments and simulation models suggested that nutrient
uptake capacities of root systems in current varieties
and hybrids are unlikely to be a significant constraint
to increasing nutrient use efficiency in intensive ag-
ricultural systems with adequate water supply, fertile
soils and high fertilizer use (Burns, 1980; Kirk and
Solivas , 1997; Peng and Cassman, 1998). However,
major changes in cropping practices have occurred in
both maize and rice systems, which re-emphasize the
need for research on genotypic variation and genetic
control of root architecture. In the USA, conservation
tillage (no-till, ridge-till, or mulch-till) has increased
to 47% of the total area planted to maize and soybeans
(Padgitt et aI., 2000). Long-term no-tillage may lead
to surface accumulation of crop residue, P and K due
to broadcast fertilizer application, uneven nutrient ex-
traction by crop roots, and annual return of residues
to the surface (Mackay et aI., 1987; Vyn and Jan-
ovicek, 2001). Soil acidification may occur in zones
with annual N injection (Bouman et al., 1995). In Asia,
direct-seeding of rice has replaced transplanting as the
dominant form of crop establishment in many areas. A
move from transplanted rice to direct-seeding may be
associated with lower N uptake from indigenous soil
resources (Peng et aI., 1996) or lower yields due to
imperfect control of factors affecting nutrient use effi-
ciency (Dobermann et aI., 2002). Direct-seeding is of-
ten associated with shallow soil tillage and high plant
density. This results in a shallower root system and a
smaller soil volume from which nutrients are extracted
by the plants, but it also changes the dynamics of nutri-
ent uptake due to greater plant competition. Therefore,
larger potential may exist for genetic manipulation of
root architecture than attempting to exploit small dif-
ferences in root nutrient influx parameters or internal
nutrient utilization among cultivars or hybrids.
Genotypic differences in crop response to nutrient
supply have often been described (Parks, 1985), but
their physiological basis is poorly understood. Mo-
lecular tools may help obtain a better understanding
of the genetic controls for tolerance to stresses, and
thereby facilitate selection of germplasm better adap-
ted to different soils or crop management practices.
163
On marginal lands, farmers may be able to grow crops
where drought and/or poor soil conditions have limited
crop production in the past (Wood et aI., 2000). On
favorable land, traits such as increased nutrient recov-
ery, lodging resistance or host-plant resistance to pests
or certain pesticides help minimize yield losses and/or
reduce production costs. Thus, future approaches for
fertilizer management must also take into account ef-
fects on stress tolerance traits so that the targeted yield
goals can be more consistently realized.
In 2001, 63% of the U.S. soybean area was
planted with herbicide-tolerant transgenic varieties,
16% of the maize area was planted with Bt-maize
and 7% with herbicide-tolerant maize (USDA, 2001,
http://www.usda.gov/nass). To date, adoption of trans-
genic rice varieties has been minimal, but is likely
to increase in the future (Conway and Toenniessen,
1999). The extent to which nutrient requirements
and management of transgenic crops might differ
from conventional varieties or hybrids has not re-
ceived much attention (PPI, 1999). As long as genetic
modifications do not alter plant traits or biochemical
pathways that increase the genetic yield potential or
harvest index, crop nutrient requirements are likely
to be similar to those of non-transformed crops. This
is probably true for the current generation of trans-
genic crops adopted by farmers. Field experiments
conducted in Nebraska found only slight differences
in yields of herbicide resistant soybeans as compared
to their non-transformed sister lines (Elmore et aI.,
2001). Research on rape showed no significant differ-
ences in yield and oil content response to N among two
transgenic hybrids and two non-transformed cultivars
(Schuster and Rathke, 2001).
Future generations of transgenic crops, however,
may have more pronounced differences in yield po-
tential due to manipulations of more complex traits
such as biochemical pathways involved in photosyn-
thesis (Zeigler, 2001) or increased nutrient acquisi-
tion by more vigorous root systems. Improvements
in grain quality through genetic engineering may also
alter nutrient management requirements. Examples
include low-phytate maize for human nutrition (Men-
doza et aI., 2001) or to reduce P content in manure
(Waldroup et al., 2000), high-oil maize (Lambert and
Hallauer, 1994), plants grown for functional foods or
nutraceuticals (Dillard and German, 2000), vitamin A-
enriched rice (Potrykus, 2001), or high phytase rice
with an increased iron content (Lucca et al., 2001).
Many obstacles must be overcome until these break-
throughs in biotechnology will have impact at the
164
farm level, particularly in developing countries (Zei-
gler, 2001). Applied plant nutrition research should
not lag behind such developments because the large-
scale investments made into biotechnology will lead
to rapid advances in the foreseeable future. Research
is also necessary to study the ecological consequences
of long-term use of transgenic crops on soil processes
and nutrient cycling. For example, recent chemical
analysis suggested that the lignin content of Bt-corn
hybrids was 33-97% higher than that of their re-
spective non-Bt isolines (Saxena and Stotzky, 2001).
Reasons for this are not understood, but such differ-
ences are likely to affect pest resistance, non-target
organisms, and the decomposition of crop residues.
Quantitative understanding of yield response to
nutrients
Progress in fundamental soil and plant research has
had insufficient impact on theoretical and practical
concepts for nutrient management in intensive agri-
culture. Researchers in the USA and Asia have mostly
developed fertilizer recommendations based on empir-
ical yield-input relationships, whereas less attention
has been paid to more quantitative plant nutrition con-
cepts. In the USA, nutrient management issues are
generally discussed within the framework of mobile
(N) versus immobile (P, K) nutrients or nutrients that
are environmentally sensitive (N, P) and those without
known environmental risk (Havlin et aI., 1999). Farm-
ers rely on soil testing to determine field-specific
application needs (Table 2), but the fertilizer recom-
mendations for maize vary widely among states in
the corn belt (Table 4). Algorithms for estimating N
rates often include a yield goal and credits for crop
rotation and other sources of N input, but the use of
soil testing varies and in some states such as Iowa
the recommendations do not explicitly account for
most of these components (Table 4). Fertilizer-N al-
gorithms that are based on soil tests tend to overpredict
N rates in years with poor response to fertilizer due
to unfavorable climate or inaccurate soil N03 testing
(Bundy et aI., 1999). Recommendations for man-
aging 'immobile' nutrients such as P and K follow
concepts 'sufficiency-deficiency correction', 'buildup
and maintenance', or 'replenishment of crop removal'
(Hergert et aI., 1997). Critical soil test levels for P
and K (derived from relative yield response curves)
vary somewhat among states and soil types in the corn
belt, but they have changed little since Bray's original
research in Illinois conducted in the late 1930s and
early 1940s (Bray, 1944, 1945, 1954). In most cases,
economics of fertilizer use are not included in the fer-
tilizer recommendation algorithms. In Asia, fertilizer
recommendations for rice are mostly based upon em-
pirical yield response functions that are extended on
a district or regional scale. Soil testing and plant tis-
sue analysis are rarely used (Table 2), mainly because
infrastructure and commercial soil testing services are
lacking (ESCAPIFAOIUNIDO, 1994). However, even
if soil testing were available, many existing soil tests
are of limited use for irrigated rice because they of-
ten fail to accurately predict the indigenous nutrient
supply under field conditions. Routine soil tests often
fail to extract soil nutrient fractions that are important
for nutrient availability under flooded soil conditions.
Moreover, in an irrigated rice field, the indigenous
nutrient supply during a growing season is also much
affected by the dynamics of flooding and drying cycles
as well as nutrient inputs from sources other than
the soil, such as irrigation and biological N2 fixa-
tion in the soil-floodwater system (Cassman et aI.,
1996; Doberrnann et aI., 1996; Yadvinder-Singh et aI.,
2000).
Recommendations such as those shown in Table 4
were the result of multi -site calibration and correlation
research, but their principles were developed during a
time when yields of maize or rice were halftoday's av-
erage yields. Errors associated with sampling and soil
testing (sampling density and depth, laboratory variab-
ility), widely varying interpretation of soil test values
and the relative insensitivity of current recommend-
ations to different soil types and crop management
practices have raised concern that the 'correlation and
calibration' approach cannot keep pace with changes
in intensified cropping systems (Hergert et aI., 1997).
The correlation/calibration yield-response approach
would require frequent empirical verification and up-
dating of recommendations in response to changes
in cropping, but the requirement for multi-year and
multi-location evaluation is both costly and slow.
In the future, fertilizer recommendation algorithms
must be considerably more robust and accurate than
current approaches. They must accommodate different
crops, cropping systems, crop management technolo-
gies, soil conditions, and climate-driven yield poten-
tial. Single levels in a recommendation should then be
based on standard conditions (e.g., maize planted in
May on a no-till deep silt loam soil in Nebraska; rice
direct-sown on a clay soil in the dry season in South
Vietnam) that take into account the major factors gov-
Table 4. Examples of university fertilizer recommendations for maize in the USA

State Nitrogen! Phosphorus2 Potassium3 Source

Iowa All pre-emergence N:
(IA) Maize after maize: 150-200lb/acre
Maize after soybean: 100-1501b/acre
Pre-emergence + in-season application:
Pre-emergence:
Maize after maize: 50-125Ib/acre
Maize after soybean: 0-75lb/acre
Sidedress: N = (CL - N03) x 8
Sufficiency concept. Tables based on
five topsoil soil test categories (very low,
low, optimum, high, very high) and two
subsoil P levels (low, high). No P if STL
> 15 ppm (Olsen) or >20 ppm (Bray,
Mehlich).
Sufficiency concept. Tables based on (Blackmer and
five topsoil soil test categories (very low, Voss, 1997)
low, optimum, high, very high), two
subsoil P levels (low, high), and two soil
texture categories (fine, sandy). No Kif
STL > 130 ppm.

Nebraska N = -35 + (1.2 x YG) - (8 x N03) - (0.14) Sufficiency concept. Tabular values for Sufficiency concept. Tabular values for (Shapiro et al.,
(NE) x YG x SOM) - N credits two modes of application (bruadcast, two modes of application (broadcast, 2001
N credits: maize 0 band) based on five topsoil soil test band) based on five topsoil soil test
soybeans 45 categories (very low, low, optimum, categories (vey low, low, optimum,
N from irrigation water high, very high). No. P if STL > 10 ppm high, very high). No Kif STL > 125 ppm.
manure (Olsen) or > 15 ppm (Bray).

Illinois N = 1.2 x YG - N credits Buildup and maintenance concept for Buildup and maintenance concept for (Hoeft and Peck,
(IL) N credits: maize 0 three geographical zones of different two geographical zones of difference CEC. 1999)
soybeans 40 subsoil-P supply. Target STL 40-50 Target STl260-300 lb/acre. No Kif
N from other chemicals) lb/acre. No P if STL >60-70 lb/acre. STL >360-400 lb/acre.
N from irrigation water CR: 0.431b P20Slbu yield. CR: 0.281b K20Slbu yield.

Minnesota Standard: Yield goal-based: Yield goal-based: (Rehm et al.,

(MN) Tables based on previous crop (6), SOM Bray-I: P = (0.700 - 0.035 STL) x YG K =(1.166 - 0.0073 STL) x YG 2000)
(2 levels), YG (categories), N03 credit. Olsen: P = (0.700 - 0.044 STL) x YG Full amount if broadcast, reduced if row-
WestemMN: Full amount if broadcast, reduced if row- applied. No K if STL > 175 ppm.
N = 1.2 x YG - N03 - other N credits applied. No P if STL > 20 ppm (Olsen) or
N credits: maize 0 >25 ppm (Bray).
soybeans 40

! N =N rate (lb N/acre); YG =yield goal or yield potential (bu/acre); N credit =credits given for N supply from previous crop or irrigation; N03 =soil test nitrate-N
level (ppm); CL = critical soil test nitrate-N level; SOM = soil organic matter content (%). Soil tests used: late spring nitrate in 0-30 cm depth (ppm, IA), soil organic
matter in topsoil (%, NE), fall or spring soil nitrate-N (ppm) in 0-120 cm (NE) or in 0-60 cm (MN).
2 P =Prate (lb P20S/acre); YG =yield goal or yield potential (bu/acre); STL =soil test level; CL =critical soil test level. ML =maintenance soil test plateau. CR =
crop nutrient removal per unit harvested yield. Soil tests used: Bray-l (IA, IL, NE, MN), Olsen (IA, NE, MN), Mehlich-3 (IA).
3 K =K rate (lb K20/acre); YG =yield goal or yield potential (bu/acre); STL =soil test level; CL =critical soil test level. ML =maintenance soil test plateau; CR =
crop nutrient removal per unit harvested yield. Soil tests used: 1 N Ammonium acetate (IA, IL, NE, MN, IA), Mehlich-3 (IA) ..-
0\
Ul
166
erning crop response to the nutrient of interest. Such
refinements can be made at different levels of com-
plexity such that a general recommendation can be
broken down into more meaningful and detailed spe-
cific recommendations. However, a key challenge is
to improve the prediction of soil nutrient supply, fer-
tilizer efficiency, plant nutrient accumulation, and its
effect on yield in absolute terms (Sinclair and Park,
1993; Witt et aI., 1999). Future improvements in es-
timating optimal fertilizer rates (F) will depend on
how researchers will be able to solve the general
equation:
Ya = !(Ym , UI, U2, ... Ux)
FI = (UI - h)/RI
where Ym is climatic and genetic yield potential, Ya
is attainable nutrient-limited yield, Fx is amount of
applied fertilizer, Ux is amount of nutrient in the plant,
Ix is supply of nutrient from indigenous sources, Rx is
fraction of nutrient recovered in the plant, and 1 to x
denote each of the essential plant nutrients.
Although several process-oriented crop simulation
models have been developed for maize and rice, their
use for solving this equation in practical nutrient man-
agement appears limited. Input data required by such
models are often not available and most of them can-
not account for nutrients other than N. Attempts have
been made to model the complete soil-plant P (Green-
wood et al., 2001) and K cycle (Greenwood and
Karpinets, 1997) in a more applied but still process-
oriented context. Another alternative is a robust, step-
wise empirical model that encompasses a wide range
of conditions as opposed to a narrowly defined local
calibration or response curve. The QUEFTS model
(Janssen et aI., 1990; Smaling and Janssen, 1993) is
such an empirical solution because it allows estim-
ating the fertilizer requirement as a function of (i)
climatic yield potential, (ii) the relationship between
grain yield and plant accumulation of N, P, and K,
(iii) the potential indigenous N, P, and K supplies,
and (iv) recovery efficiencies of fertilizer N, P, and
K. In this approach (i) can be estimated using a val-
idated crop simulation model, (iii) must be measured
using a soil test or a crop-based estimate, and (iv)
is usually adjusted to local soil types and cropping
conditions. Estimates for (ii) can be obtained from a
generic relationship between grain yield and nutrient
accumulation obtained from a large database from a
wide range of production environments to account for
nutrient interactions and differences in yield potential
(Witt et aI., 1999).
Practitioners estimate crop nutrient uptake per unit
biomass or yield (goal) using a single number (rule of
thumb). Using such numbers, long-term research on
maize in Nebraska concluded that crop removal-based
fertilizer recommendations led to an uneconomically
high use of P and K fertilizer with no significant yield
gains over a sufficiency approach based on critical soil
test levels (Olson et aI., 1982). However, questions
must be raised whether correct estimates of crop nutri-
ent requirements per unit yield (or crop removal coef-
ficient) are currently used because they (i) are typically
derived from field experiments conducted at only few
sites, which are most often located at research stations
with high background levels of indigenous soil nutri-
ent supply, (ii) assume linearity between crop yield
and nutrient accumulation, and (iii) do not account for
nutrient interactions and climatic yield potential as a
driving force for optimal nutrient requirements (Witt
et aI., 1999). There is generally a close relationship
between dry matter and nutrient accumulation across
a wide range of sites and varieties, described by the
same non-linear function for both maize and rice (Fig-
ure 4a). However, the relationship becomes scattered
when grain yield is plotted against plant N accumu-
lation. There are different plateaus for rice and maize
that are related to differences in the genetic and cli-
matic yield potential, but there is also large variation
within each species, which is caused by a multitude of
yield-limiting factors as well as excess of certain nu-
trients. In both crops, non-linear average relationships
suggest decreasing internal efficiency of nutrients as
yields approach the yield potential (Figure 4b). The
same principles hold true for other nutrients such as P
and K (data not shown). Using a single 'crop removal
coefficient' may therefore lead to erroneous nutrient
use and low efficiency.
In earlier work, C.T. de Wit and later H. van Keu-
len (van Keulen, 1977, 1986; van Keulen and Van
Heemst, 1982) studied the relationship between yield
and plant nutrient accumulation for several crops, in-
cluding maize and rice. They showed a linear range
followed by a parabolic plateau and concluded that
an upper boundary exists at which a nutrient's con-
centration in grain (and straw) becomes diluted to the
maximum possible extent when that nutrient is the sole
factor limiting yield. In the QUEFTS model, Janssen
et aI. (1990) expanded this work by using two linear
boundaries that described the range from maximum
 167

;:- 30000 (a)

'co
.c 25000 Boundary of maximum dilution 0
en
-= 20000
.lII:

Q)

~ 15000
~ 10000
"C

So 5000
... o
•
o
Rice (avg. 1.0% N in OM)
Maize (avg. 1.2% N in OM)

- 16000
":"co 14000
.c
(b)
o

:E 12000
c
-
~ 10000
"C
(j)
8000
o
o
Maize
y =-3710 + 995 XO.5
R2 = 0.76
.~ 6000
Rice
4000 y =-1573 + 643 XO. 5
R2 = 0.55
2000
o
o 50 100 150 200 250 300 350
Plant N accumulation (kg N ha-1)
Figure 4. Relationships between total plant dry matter or grain yield and plant nitrogen accumulation in maize and rice. Irrigated rice: on-farm
and research station experiments conducted in Asia during 1995-2000 (n = 1658). Maize: on-farm and research station experiments conducted
in the north-central USA during 1995-2000 (n=470). Dashed lines indicate the apparent boundary of maximum dilution of nitrogen in the plant,
whereas the solid lines show the average internal efficiency for each environment as obtained from non-linear regression.

accumulation to maximum dilution of N, P, and K in from 2 to 11 kg P, and from 16 to 50 kg K per 1000

maize. Those 'envelopes' were then mathematically
combined into linear parabolic plateau curves of op-
timal (balanced) nutrition of all three macronutrients.
Witt et aI. (1999) demonstrated how this method can
be used to develop families of yield-optimal NPK ac-
cumulation curves for rice grown in environments with
different climatic yield potential across Asia. Their
model predicted a linear increase in grain yield of
rice if nutrients are taken up in balanced amounts of
14.7 kg N, 2.6 kg P, and 14.5 kg K per 1000 kg
of grain yield, until yields reached about 70-80% of
the climate-adjusted yield potential. This compares to
published estimates that range from 15 to 24 kg N,
kg yield (Witt et aI., 1999). This comparison indicates
that literature data appear to overestimate nutrient re-
quirements if they are not based on experiments that
represent the whole range of farming conditions.
Related research on defining generic nutrient dilu-
tion curves during crop growth describe the decline in
critical plant N concentration with increasing crop bio-
mass accumulation (Greenwood et aI., 1990; Sheehy
et aI., 1998). Attempts are also being made to combine
those concepts with the yield-nutrient accumulation
relationships used in QUEFTS (Witt et aI., 2001).
All these approaches have a common foundation in
that they seek to develop a quantitative understanding
168
of crop nutrient requirements in a robust format that
is suitable for practical nutrient management across
a wide range of environments. However, except for
recent research in irrigated rice (Dobermann et aI.,
2002), such an approach has not been evaluated by the
soil testing and fertilizer management community in
the USA and Asia.
Quantitative approaches are particularly suitable
for favorable production environments, because in
most years the yield response to nutrients is not
severely confounded by other abiotic or biotic stresses.
Their generic framework simultaneously accounts for
interactions amongst macronutrient requirements and
also allows the estimation of economic return from
fertilizer application (Janssen, 1998). To improve the
accuracy of this approach, however, future research
must clarify a number of unanswered questions. First,
available experimental data suggest that the upper
boundary of maximum nutrient dilution in the plant at
harvest is also non-linear (Figure 4b), not linear as as-
sumed in models such as QUEFTS. Second, there is no
clear theoretical justification for using a boundary line
describing maximum accumulation of a nutrient in the
plant because these are mostly situations of disturbed
growth due to factors other than nutrients. Third, in
modeling the nutrient interactions, the same weight is
given to N, P, and K, whereas it is possible that nu-
trients such as K can be diluted relatively more in the
plant than N before a significant reduction in growth
occurs (Bums et aI., 1997). However, Greenwood and
Stone (2001) have recently shown K dilution curves
for a range of vegetable crops, suggesting that, as
for N, critical and maximum K concentrations during
growth appear to be linearly related to relative growth
rate. Critical P or K dilution curves analogous to those
determined for N have not been published for cereal
crops. Fourth, the nutrient requirements of a crop must
be examined in relation to yield potential, with par-
ticular emphasis on requirements at yield levels that
are 80% or more of the yield potential ceiling. A re-
lated issue is whether the linear range of the optimal
relationship between grain yield and plant nutrient ac-
cumulation will simply extend further with additional
increases in yield potential.
Generic approaches for site-specific nutrient
management
Precision farming or site-specific management in the
USA has focused on managing spatial variability of
nutrients within large fields by variable application of
N, P, K, or lime using local fertilizer algorithms in
combination with soil samples collected from grids
or 'soil management zones' within a field (Pierce
and Nowak, 1999). With the exception of liming,
many of the case studies conducted so far have failed
to demonstrate significant agronomic, economic, or
environmental benefits over uniform applications (Fer-
guson et aI., 2002; Lowenberg-DeBoer and Swin-
ton, 1997; Pierce and Nowak, 1999; Wibawa et aI.,
1993). Reasons for this failure appear to result from:
(i) insufficient characterization of spatial variation in
indigenous nutrient supply (including sampling and
laboratory error) and yield goals, (ii) use of empir-
ical, single-nutrient fertilizer algorithms that are not
suited for site-specific management (Ferguson et aI.,
2002), and (iii) insufficient post-emergence adjust-
ment of N rates and timing of application to account
for differences between the actual yield potential and
the average climatic conditions that were assumed for
making the fertilizer recommendation. Understanding
is lacking of spatial cause-effect relationships that can
be quantified, generalized and extrapolated. Reliance
on grid soil sampling and inherently imprecise meas-
urements such as soil test P and K for developing
variable rate fertilizer application maps has largely
proven unsuccessful. There is also a lack of mul-
tivariate response functions that can estimate the yield
response to inputs, site characteristics, and varying
plant density (Bullock et aI., 1998). Currently, there
are no standards for soil sampling designs, sampling
intensity, or methods of interpolation used in creating
nutrient management maps and rarely is there inform-
ation provided about the quality of such maps (Pierce
and Nowak, 1999). Quantitative propagation of errors
is not well understood. Errors must be partitioned into
those caused by unresolved spatial variation (due to
sampling, soil testing, and interpolation), uncertainty
about crop response models, and the application er-
ror associated with equipment performance. Recent
studies suggest that it is unlikely that the classical
soil sampling and soil testing approach can become a
basis for precise nutrient management because the soil
chemical analytical cost is large and because of the
sampling, analytical, and interpolation error inherently
associated with it (Viscarra Rossel and McBratney,
1998). Taking into account all uncertainties involved
may lead to the conclusion that the optimum is reached
with an uniform application of inputs (Viscarra Rossel
et aI., 2001), but no such studies have been conducted

to compare different nutrients in environments with
different yield potential and risk.
Site-specific nutrient management (SSNM) should
be more broadly defined as the dynamic, location-
specific management of nutrients in a particular crop-
ping season to optimize the congruence of supply
and demand of nutrients according to their differ-
ences in cycling through soil-plant systems (Dober-
mann et aI., 2002). This definition accounts for (i)
regional and seasonal differences in yield potential
and crop nutrient demand, (ii) between- as well as
within-field spatial variability in indigenous nutrient
supply, (iii) within-season dynamics of soil N sup-
ply and crop N demand, and (iv) location-specific
cropping systems and crop management practices. A
generic SSNM concept must then consider the de-
terminants and governing forces of (i) pre-emergence
and long-term management of macro- and micronu-
trients and (ii) post-emergence (in-season) adjustment
of N to account for the seasonal variability in growth
and yield potential (Figure 5). Such a concept also
takes into account the principal differences in man-
aging nitrogen, less mobile nutrients such as P and
K, and micronutrients (Dobermann and White, 1999).
From 1997 to 2000, a field-specific variant of the
SSNM strategy shown in Figure 5 was developed
and evaluated in on-farm experiments at 179 sites in
eight irrigated rice domains of Asia (Dobermann et
aI., 2002; Wang et aI., 2001). Because significant
field-to-field variability existed and within-field vari-
ability operated over short distances that were difficult
to cope with (Dobermann et aI., 1995, 1997), man-
aging the variability among fields was identified as
highest priority. Fertilizer application rates for N, P,
and K were estimated for individual fields by account-
ing for the indigenous nutrient supply, yield goal,
and nutrient demand as a function of the interactions
between uptake requirements for N, P, and K (Witt
et aI., 1999). Crop-based estimates of the indigenous
nutrient supply in nutrient omission plots were used
because soil testing methods did not sufficiently pre-
dict this parameter. Different N management schemes
were developed for each domain to account for re-
gional variation in the primary factors driving N use
efficiency (Figure 5). Average grain yield increased by
0.5 Mg ha- I (11 %) and N fertilizer rate decreased by
5 kg N ha -I with field-specific management compared
to the baseline farmers' fertilizer practice. Farmers'
practices typically relied on a large N fertilizer ap-
plication early in the season, when the capacity for
crop uptake was small, and one or two additional N
169
topdressings. In contrast, field-specific management
utilized two to four topdressings that were applied
to achieve greater synchrony with crop demand, and
individual doses of pre-plant or topdressed N were
smaller than those applied by farmers' practices. As
a result, mean REN increased from 30% with farm-
ers' practices to 40% with field-specific management.
On average, profit increased by US$46 ha- 1 per crop
through the use of field-specific management. These
results highlight the potential for SSNM in small-scale
farming systems in developing countries, provided the
technologies chosen match the systems' biophysical
and socioeconomic characteristics.
Key avenues for developing a similar improved
SSNM strategy for mechanized maize farming in-
clude (i) thematic mapping of soil properties using a
combination of spatially dense auxiliary information
(on-the-go soil sensors, remote sensing, yield maps)
and destructive soil sampling, and (ii) decision aids
for in-season N management in a large field. Repro-
ducible procedures for delineation of functional soil
zones for site-specific management (van Alphen and
Stoorvogel, 2000) must be developed so that such
'zones' can be managed with greater precision and
by using more precise nutrient models and recom-
mendations. Many in-season N management concepts
based on post-emergence soil or plant indicators were
proposed for maize in the past, but most of them
have not found practical acceptance due to their high
cost or difficulties in implementing them in routine
farming (Schroeder et aI., 2000). Methods can be
broadly divided into three approaches: (i) corrective N
management, (ii) predictive N management, and (iii)
predictive-corrective N management. In each of these,
post-emergence N applications can be homogeneously
applied to the whole field or, in the most advanced
sense, varied over very short distances if a sensor
is attached to a fertilizer spreader with variable-rate
capabilities.
Corrective methods employ diagnostic tools such
as a chlorophyll meter (Varvel et aI., 1997), remote
sensing (Blackmer et aI., 1996), or on-the-go sensors
(Lammel et aI., 2001) to determine the need for an
N topdressing. Although technology development is
proceeding rapidly, the ability to interpret remotely
sensed information about canopy N status to estimate
the amount of fertilizer-N needed has proven to be a
difficult challenge. At present, this approach relies on
empirical comparison with an over- or under-fertilized
reference strip to assess whether an additional yield
response to N is likely to occur. However, if the
170
(II) Post-emergence nitrogen management
Determinant
Yield potential (actual)
N input requirement
Fraction of N applied
(I) Pre-emergence nutrient management
Determinant
Yield potential (mean, variability)
Attainable yield (goal)
NPK input requirement
Fraction and form of NPK applied
- - - - .. Data acquisition
4 Interpretation and management
Figure 5. Detenninants and their controlling factors in a general, dynamic, site-specific nutrient management scheme for non-legume field
crops (P - planting; H - harvest).
diagnostic tools used would allow establishing quant-
itative relationships between reflectance and biomass
(Bouman et aI., 1992) and between reflectance and
nitrogen status (Blackmer et al., 1996), future im-
provements in data interpretation can be made by
applying concepts such as critical N dilution curves
for a certain yield target (Greenwood et al., 1990;
Witt et al., 2001). Moreover, corrective approaches re-
quire careful N management at all key growth stages to
avoid N deficiency occurring at critical growth stages.
If N deficiency occurs during early vegetative growth
of maize, correcting it with late-season N applications
is unlikely to fully compensate for the yield loss as-
sociated with yield components formed during early
growth (Binder et aI., 2000).
In contrast, predictive N management can be ac-
complished in real-time by using soil/crop models in
combination with actual weather data (van Alphen and
Stoorvogel, 2000). In predictive-corrective N man-
agement, this approach may be combined with other
methods of crop diagnosis. For example, remote sens-
ing data can be used as a forcing function in a crop
growth model to improve the prediction by taking into
account actual growth data (Bouman, 1995). With in-
Governing forces
Climate (actual)
Crop growth: biomass and N concentration
Crop architecture, water management, pest
incidence, prices
Governing forces
Genotype, climate (historical)
Soil potential (quality)
Indigenous nutrient supply
Tillage, crop establishment method, water
management, soil mineralogy, prices
creasing availability of weather and remotely sensed
data over the internet, these dynamic N management
approaches are likely to become more attractive and
cost-effective to maize farmers in the USA. The chal-
lenge will be to develop simplified crop models to
estimate yield and the additional N needed to achieve
it in response to actual crop conditions.
Managing the local and global environment
Increasing the amount of fertilizer used without an
associated increase in nutrient efficiency may have
adverse consequences for the environment and hu-
man health that are not reflected in the costs and
returns of agricultural production (Uri, 1998). The
main fertilizer compounds or transformation products
contaminating water and air resources are N20, NH3,
N03, soluble phosphates, and traces of heavy metals.
Preliminary estimates for the UK (Pretty et aI., 2000)
and Germany (van der Ploeg et aI., 2002) suggest that
the external costs of agriculture may be as large as the
total worth of all farm goods produced, with abundant
N use contributing to about one third of these costs.

Consequently, regulation of fertilizer use through laws
or taxes has become part of intensive agriculture in
developed countries, including maize farming in the
USA (Uri, 1998). Unfortunately, little field research
has been conducted to address the whole spectrum of
agronomic and environmental consequences of crop
intensification. Significant differences exist between
maize systems in the USA and rice systems in Asia. In
both, available estimates of external costs are sketchy
at best because of difficulties in obtaining accurate
numbers of the nutrient cycling processes involved.
Moreover, fertilizer use affects crop growth and this
enhancement of biomass production may be associ-
ated with positive as well as negative effects on the
environment.
Nitrate contamination of water supplies is not a
major issue when rice is grown in anaerobic flooded
systems such as double-cropped rice (Bouman et al.,
2001). However, intensive maize-based cropping sys-
tems in the com belt may contribute significantly to
nitrate contamination of groundwater (Spalding and
Exner, 1993) or of whole regional watersheds such
as the Mississippi River Basin feeding into the Gulf
of Mexico (CAST, 1999). The Mississippi River ex-
ports about 1.8 million tons of N each year and it is
estimated that agriculture's annual share is about 2-3
kg N ha- 1 agricultural land - an equivalent to a total
fertilizer value of $410 million year- 1 (CAST, 1999).
Several states have therefore implemented regulation
governing fertilizer use by farmers in these areas. In
Nebraska, best management practices required by law
depend on the nitrate concentration in groundwater. In
highly contaminated areas (Phase III areas with > 20
ppm nitrate), irrigation water must be tested for nitrate,
irrigation applications must be metered, soils must be
analyzed for nitrate to 1 m depth annually on every
field, fall and winter applications of fertilizers are pro-
hibited, and spring applications must be split or must
use an approved N inhibitor. Subsequent studies have
shown that widespread adoption of these management
practices has led to decreasing nitrate concentrations
in the groundwater, but also to benefits for farmers be-
cause it enabled them to reduce fertilizer use without
affecting crop yield (Bosch et aI., 1995; Fuglie and
Bosch, 1995).
Besides such local and regional environmental
effects, maize systems have a significantly under-
utilized carbon (C) sequestration potential (Collins et
al., 1999), which is related to the amount of bio-
mass (crop residues) produced and thereby dependent
on crop rotation, fertilizer use and nutrient efficiency
171
(Halvorson et aI., 1999). However, potentially positive
effects of sequestering C in such agricultural sys-
tems may be offset by emissions of other greenhouse
gases such as N20 or high energy use (Robertson
et aI., 2000) if yields and nutrient use efficiency are
below attainable levels. Similarly, although irrigated
rice monoculture systems in Asia are known to se-
quester carbon (Bronson et aI., 1997a) and emit little
N20 (Bronson et al., 1997b), recent estimates suggest
that they contribute about 2-5% to the global meth-
ane (C~) budget (Matthews et al., 2000). Methane
emission can be managed through a variety of means,
including organic and inorganic amendments as well
as crop management practices that also affect nutrient
dynamics (Wassmann et aI., 2000).
In general, the effects of improved nutrient man-
agement on environmental quality and other external-
ities are likely to be positive if they combine yield in-
creases with increases in fertilizer use efficiency. Yield
level, amount and decomposition of crop residues,
soil organic matter, and soil N and P dynamics are
important determinants of greenhouse gas emissions
that can be manipulated through plant nutrition. Every
increase in grain production that comes from higher
yields per unit area contributes to sequestration of C
and reduces the pressure to expand cultivated area to
natural ecosystems or marginal land. Increased REN
potentially results in less N runoff and leaching and
reduced gaseous N losses into the environment. In-
creased PFPN reduces the amount of fertilizer needed
to produce a unit of grain, which will reduce C02
emissions resulting from the use of fossil energy to
produce fertilizer-No
In summary, assessing the local, regional, and
global consequences of nutrient application in agri-
culture on the environment and human health must
become an integral component of future agronomic
research. Intensive agricultural systems can probably
be designed in which an optimal balance of productiv-
ity, soil C sequestration, nitrate leaching, and emission
of greenhouse gases is achieved through increased
yield, more efficient use of fertilizers, conservation
tillage, and irrigation. Educational programs in com-
bination with non-regulatory incentives (Uri, 1998)
that motivate farmers to increase nutrient use effi-
ciency are preferable over regulatory levies because
the latter expose farmers in one country to a com-
petitive disadvantage (van der Ploeg et aI., 2002).
Multilateral agreements such as the Kyoto Protocol
may further help reducing the external costs of ag-
 172
riculture in the future, but implementation of such
international agreements is laden with obstacles.
Conclusions
The principal challenges to improving yields, input
use efficiency, profitability, and environmental impact
are similar in large-scale maize and small-scale rice
systems. At present, average yield levels of maize and
rice are only 40-65% of the attainable yield potential
and the average recovery of fertilizer N is less than
40%. The commonly used approaches to soil fertil-
ity and fertilizer research and management may be
insufficient for achieving greater input efficiency be-
cause they are too general and too empirical. Factors
that will drive the need for more precise, dynamic,
and diverse nutrient management approaches are (i)
increasing yield levels that approach current yield po-
tential ceilings in the best farms, (ii) future germplasm
improvement in yield potential, grain quality, stress
tolerance, and adaptation to more intensive manage-
ment practices, (iii) spatial and temporal variability
of soil nutrient supply and crop nutrient demand,
and (iv) threats of regulation due to local and global
environmental concerns.
Robust strategies for site-specific nutrient man-
agement must therefore be based on a quantitative
understanding of seminal relationships between yield
and nutrient uptake and the congruence between nu-
trient supply and crop demand. While the underpin-
ning scientific principles guiding such an approach are
generic, implementation will require consideration of
appropriate technologies designed for different spatial
scales. More resources should be devoted to strategic
field- and on-farm research that follows systems ap-
proaches. Priorities include (i) farm level data on nu-
trient use and nutrient use efficiency, including trends
over time and causes of variability, (ii) use of mod-
els and geospatial techniques to obtain a quantitative
understanding of crop response to spatial and tem-
poral environmental variation, (iii) better approaches
for real-time N management, (iv) greater knowledge
of nutrient management requirements for transgenic
crops with specific end-use traits, and (v) interdiscip-
linary field research that addresses the entire spectrum
of agricultural, ecological, and environmental func-
tions of intensive cropping systems, which are the
foundation of the human food supply.
Acknowledgements
The Swiss Agency for Development and Coopera-
tion (SDC), the International Fertilizer Industry As-
sociation (IFA) , the Potash and Phosphate Institute
(PPIIPPIC), and the International Potash Institute (IPI)
provided funding for the rice research cited through-
out the paper. Many researchers and support staff
participating in the Project on 'Reversing Trends of
Declining Productivity in Intensive Irrigated Rice Sys-
tems' have made contributions to this work. We thank
Dan Walters (University of Nebraska), Christian Witt
and David Dawe (IRRI) for providing data cited in this
paper and many stimulating discussions.
References
Baligar V C and Barber S A 1979 Genotypic differences of corn for
ion uptake. Agron. J. 66. 870-873.
Binder D L, Sander D H and Walters D T 2000 Irrigated maize
response to time of nitrogen application as affected by level of
nitrogen deficiency. Agron. J. 92, 1228-1236.
Blackmer A M and Voss R D 1997 Nitrogen Fertilizer Recommend-
ations for Corn in Iowa. Pm-1714. Iowa State University, Ames,
IA.
Blackmer T M, Schepers J S, Varvel G E and Walter-Shea E A 1996
Nitrogen deficiency detection using reflected shortwave radiation
from irrigated corn canopies. Agron. J. 88, 1-5.
Bosch D, Cook Z and Fuglie K 1995 Voluntary versus mandatory
policies to protect water quality: adoption of nitrogen testing in
Nebraska. Rev. Agric. Econ. 17,13-24.
Bouman BAM 1995 Crop modelling and remote sensing for yield
prediction. Neth. J. Agric. Sci. 43,143-161.
Bouman BAM, van Kasteren H W J and Uenk D 1992 Standard
relations to estimate ground cover and LAI of agricultural crops
from reflectance measurements. Eur. J. Agron. 1,249-262.
Bouman BAM, Castaneda A R and Bhuiyan S I 2001 Nitrate and
pesticide contamination of groundwater under rice-based crop-
ping systems; Evidence from the Philippines. Agric. Ecosyst.
Environ. (in review)
Bouman 0 T, Curtin D, Campbell C A, Biederbeck V 0 and
Ukrainetz H 1995 Soil acidification from long-term use of an-
hydrous ammonia and urea. Soil Sci. Soc. Am. J. 59, 1488-1494.
Bray R H 1944 Soil-plant relationships: I. The quantitative relation
of exchangeable potassium to crop yields and to crop response to
potash additions. Soil Sci. 58, 305-324.
Bray R H 1945 Soil-plant relationships: II. Balanced fertilizer use
through soil tests for potassium and phosphorus. Soil Sci. 59,
463-473.
Bray R H 1954 A nutrient mobility concept of soil-plant relation-
ships. Soil Sci. 78, 9-22.
Bronson K F, Cassman K G, Wassman R, Olk D C, van Noordwijk
M and Garrity D P 1997a Soil carbon dynamics in different crop-
ping systems in principal ecoregions of Asia. In Management of
carbon sequestration in soil. Eds. R Lal, J Kimble, R F Follett
and B A Stewart. pp. 35-57. CRC Press, Boca Raton, FL.
Bronson K F, Neue H U, Singh U and Abao E B Jr. 1997b Auto-
mated chamber measurements of methane and nitrous oxide

flux in a flooded rice soil: I. Residue, nitrogen, and water
management. Soil Sci. Soc. Am. J. 61, 981-987.
Bullock D G, Bullock D S, Nafziger E D, Doerge T A, Paszkiewicz
S R, Carter P R and Peterson T A 1998 Does variable rate seeding
of corn pay? Agron. 1. 90, 830-836.
Bundy L G, Walters D T and Olness A E 1999 Evaluation of soil
nitrate tests for predicting corn nitrogen response in the North
Central Region. NC Reg. Res. Publ. No. 342. Wisconsin Agric.
Expt. Station, Madison, WI. 31 pp.
Burns I G 1980 Influence of the spatial distribution of nitrate on the
uptake of N by plants: A review and a model for rooting depth.
J. Soil Sci. 31,155-173.
Burns I G, Walker R Land Moorby J 1997 How do nutrients drive
growth? Plant Soil 196, 321-325.
Cassman K G 1999 Ecological intensification of cereal production
systems: Yield potential, soil quality, and precision agriculture.
Proc. Natl. Acad. Sci. USA 96,5952-5959.
Cassman K G and Dobermann A 2001 Evolving rice production sys-
tems to meet global demand. In Rice Research and Production in
the 21st Century: Proceedings of a Symposium Honoring Robert
F. Chandler, Jr. June 15-16 2000. Cornell University, Ithaca,
New York. pp. 79-100. International Rice Research Institute,
Manila.
Cassman K G, Dobermann A, Sta.Cruz P C, Gines H C, Samson
M I, Descalsota J P, Alcantara J M, Dizon M A and Olk D C
1996 Soil organic matter and the indigenous nitrogen supply of
intensive irrigateq rice systems in the tropics. Plant Soil 182,
267-278.
Cassman K G, Peng S, Olk D C, Ladha J K, Reichardt W, Dober-
mann A and Singh U 1998 Opportunities for increased nitrogen
use efficiency from improved resource management in irrigated
rice systems. Field Crops Res. 56, 7-38.
Cassman K G, Dobermann A and Walters D T 2002 Agroecosys-
terns, nitrogen-use efficiency, and nitrogen management. Ambio
31,132-140.
CAST 1999 Gulf of Mexico hypoxia: Land and sea interactions.
Task Force Report No. 134. Council for Agricultural Science and
Technology, Ames, IA. 44 pp.
Collins H P, Blevins R L, Bundy L G, Christenson D R, Dick
W A, Huggins D R and Paul E A 1999 Soil carbon dynamics
in corn-based agroecosystems: results from carbon-13 natural
abundance. Soil Sci. Soc. Am. J. 63, 584-591.
Conway G and Toenniessen G 1999 Feeding the world in the
twenty-first century. Nature 402 (6761, Suppl.), C55-C58.
Dawe D and Dobermann A 1999 Defining productivity and yield.
IRRI discussion papers series no. 33. International Rice Research
Institute, Makati City, Philippines.
Dillard C J and German J B 2000 Phytochernicals: nutraceuticals
and human health. J. Sci. Food Agric. 80, 1744-1756.
Dobermann A 2000 Future intensification of irrigated rice systems.
In Redesigning Rice Photosynthesis to Increase Yield. Eds. J
E Sheehy, P L Mitchell and B Hardy. pp. 229-247. Interna-
tional Rice Research Institute, Makati City, Philippines. Elsevier
Science, Amsterdam.
Dobermann A and White P F 1999 Strategies for nutrient manage-
ment in irrigated and rainfed lowland rice systems. Nutr. Cycl.
Agroecosyst. 53, 1-18.
Dobermann A and Witt C 2000 The potential impact of crop intensi-
fication on carbon and nitrogen cycling in intensive rice systems.
In Carbon and Nitrogen Dynamics in Flooded Soils. Eds. G J
D Kirk and D C Olk. pp. 1-25. International Rice Research
Institute, Makati City, Philippines.
173
Dobermann A, Pampolino M F and Neue H U 1995 Spatial and
temporal variation of transplanted rice at the field scale. Agron.
J. 87, 712-720.
Dobermann A, Cassman K G, Sta.Cruz P C, Adviento M A A and
Pampolino M F 1996 Fertilizer inputs, nutrient balance, and soil
nutrient-supplying power in intensive, irrigated rice systems. II.
Effective soil K-supplying capacity. Nutr. Cycl. Agroecosyst. 46,
11-21.
Dobermann A, Goovaerts P and Neue H U 1997 Scale depend-
ent correlations among soil properties in two tropical lowland
ricefields. Soil Sci. Soc. Am. J. 61, 1483-1496.
Dobermann A, Cassman K G, Mamaril C P and Sheehy J E 1998
Management of phosphorus, potassium and sulfur in intensive,
irrigated lowland rice. Field Crops Res. 56, 113-138.
Dobermann A, Dawe D, Roetter R P and Cassman K G 2000 Re-
versal of rice yield decline in a long-term continuous cropping
experiment. Agron. J. 92, 633-643.
Dobermann A, Witt C, Dawe D, Gines G C, Nagarajan R, Satawath-
ananont S, Son T T, Tan P S, Wang G H, Chien N V, Thoa
V T K, Phung C V, Stalin P, Muthukrishnan P, Ravi V, Babu
M, Chatuporn S, Kongchum M, Sun Q, Fu R, Simbahan G C
and Adviento M A A 2002 Site-specific nutrient management
for intensive rice cropping systems in Asia. Field Crops Res. 74,
37-66.
Duvick D N and Cassman K G 1999 Post-green revolution trends in
yield potential of temperate maize in the North-Central United
States. Crop Sci. 39, 1622-1630.
Elmore R W, Roeth F W, Nelson L A, Shapiro C A, Klein R N,
Knezevic S Z and Martin A 2001 Glyphosate-resistant soybean
cultivar yields compared with sister lines. Agron. J. 93, 408-412.
ESCAPIFAOfUNIDO 1994 Proceedings of the FADINAP regional
workshop on cooperation in soil testing for Asia and the Pacific,
Bangkok, 16-18 August 1993. United Nations, New York. 137
p.
FAO 2001 FAOSTAT. http://apps.fao.org. FAO, Rome.
Ferguson R B, Hergert G W, Schepers J S, Gotway C A, Cahoon J
E and Peterson T A 2002 Site-specific nitrogen management of
irrigated maize: yield and soil residual nitrate effects. Soil Sci.
Soc. Am. J. 66, 544-553.
Foreman L F 2001 Characteristics and production costs of U.S. corn
farms. Statistical bulletin no. 974 (www.ers.usda.gov). USDA,
Economic Research Service, Washington, DC.
Fuglie K 0 and Bosch D J 1995 Economic and environmental
implications of soil nitrogen testing - a switching-regression
analysis. Am. J. Agric. Econ. 77, 891-900.
Gollehon N, Caswell M, Ribaudo M, Kellogg R, Lander C and
Letson D 2001 Confined animal production and manure nutri-
ents. Agricultural information bulletin no. 771. USDA, ERS,
Washington, DC.
Greenland D J 1997 The Sustainability of Rice Farming. CAB In-
ternational, International Rice Research Institute, Oxon, Manila.
273 pp.
Greenwood D J and Karpinets TV 1997 Dynamic model for the
effects of K-fertilizer on crop growth, K-uptake and soil-K in
arable cropping. I. Description of the model. Soil Use Manage.
13,178-183.
Greenwood D J and Stone D A 2001 Prediction and measurement of
the decline in critical-K, the maximum-K and total cation plant
concentrations during the growth of field vegetable crops. Ann.
Bot. 82, 871-881.
Greenwood D J, Lemaire G, Gosse G, Cruz P, Draycott A and Neet-
son J J 1990 Decline in percentage N of C3 and C4 crops with
increasing plant mass. Ann. Bot. 66, 425-436.
174
Greenwood D J, Karpinets T V and Stone D A 2001 Dynamic model
for the effects of soil and fertililzer-P on crop growth, P-uptake
and soil-P in arable cropping: model description. Ann. Bot. 88,
279-291.
Halvorson A D, Reule C A and Follett R F 1999 Nitrogen fertiliz-
ation effects on soil carbon and nitrogen in a dryland cropping
system. Soil Sci. Soc. Am. J. 63, 912-917.
Havlin J L, Beaton J D, Tisdale S L and Nelson W L 1999 Soil
fertility and fertilizers. An Introduction to Nutrient Management.
Prentice Hall, Upper Saddle River, NJ. 499 pp.
Hergert G W, Pan W L, Huggins D R, Grove J H and Peck T R 1997
Adequacy of current fertilizer recommendations for site-specific
management. In The Site-Specific Management for Agricultural
Systems. pp. 283-300. ASA, CSSA, SSSA, Madison, WI.
Hoeft R G and Peck T R 1999 Soil testing and fertility. In Illinois
Agronomy Handbook 1999-2000. pp. 78-116. University of
Illinois, Urbana-Champaign, lL.
Hossain M and Singh V P 2000 Fertilizer use in Asian agriculture:
implications for sustaining food security and the environment.
Nutr. Cycl. Agroecosyst. 155-169.
Huke R E and Huke E H 1997 Rice area by type of culture:
South, Southeast, and East Asia. A revised and updated database.
International Rice Research Institute, Manila, Philippines.
IFA 1999 Fertilizer use by crop. IFA, IFDC, FAa, Rome.
IFA 2000 IFADATA statistics from 1973 to 1999. CD-ROM. Paris.
IRRI 1997 Rice Almanac. International Rice Research Institute
(IRRI), Los Banos, Philippines. 181 pp.
Janssen B H 1998 Efficient use of nutrients: an art of balancing.
Field Crops Res. 56, 197-201.
Janssen B H, Guiking F C T, Van der Eijk D, Smaling EM A, Wolf
J and van Reuler H 1990 A system for Quantitative Evaluation
of the Fertility of Tropical Soils (QUEFTS). Geoderma 46, 299-
318.
Kellogg R L, Lander C H, Moffitt D and Gollehon N 2000 Manure
nutrients relative to the capacity of cropland and pasture land
to assimilate nutrients: Spatial and temporal trends for the US
USDA, NRCS, Washington, DC. 93 pp.
Kirk G J D and Solivas J L 1997 On the extent to which root
properties and transport through the soil limit nitrogen uptake
by lowland rice. Eur. J. Soil Sci. 48,613-621.
Koyama T and App A A 1979 Nitrogen balance in flooded rice soils.
In Nitrogen and Rice. pp. 95-104. International Rice Research
Institute (IRRI), Los Banos, Philippines.
Ladha J K, Dawe D, Ventura T S, Singh U, Ventura Wand Watanabe
I 2000a Long-term effects of urea and green manure on rice
yields and nitrogen balance. Soil Sci. Soc. Am. J. 64,1993-2001.
Ladha J K, Fischer K S, Hossain M, Hobbs P R and Hardy B 2000b
Improving the productivity and sustainability of rice-wheat sys-
tems of the Indo-Gangetic plains: A synthesis of NARS-IRRI
partnership research. IRRI discussion paper no. 40. International
Rice Research Institute, Makati City, Philippines.
Ladha J K and Reddy P M 2000 The quest for nitrogen fixa-
tion in rice. International Rice Research Institute, Makati City,
Philippines. 353 pp.
Lambert R J and Hallauer A R 1994 High-oil com hybrids. In
Specialty Corns. pp. 123-145. CRC Press, Boca Raton, FL.
Lammel J, Wollring J and Reusch S 2001 Tractor based remote
sensing for variable rate nitrogen fertilizer application. In Plant
Nutrition - Food security and sustainability of agro-ecosystems.
Eds. W J Horst et al. pp. 694-695. Kluwer, Dordrecht.
Lowenberg-DeBoer J and Swinton S M 1997 Economics of site-
specific management in agronomic crops. In The State of Site-
Specific Management for Agriculture. Eds. F J Pierce and E J
Sadler. pp. 369-396. ASA, CSSA, SSSA, Madison, WI.
Lucca P, Hurrell R and Potrykus I 2001 Genetic engineering ap-
proaches to improve the bioavailability and the level of iron in
rice grains. Theor. Appl. Genet. 102,392-397.
Mackay A D, Kladivko E J, Barber S A and Griffith D R 1987
Phosphorus and potassium uptake by com in conservation tillage
systems. Soil Sci. Soc. Am. J. 51, 970-974.
Matthews R B, Kropff M J, Bachelet D and van Laar H H 1995
Modeling the Impact of Climate Change on Rice Production in
Asia. CAB International, IRRI, Wallingford.
Matthews R B, Wassmann R, Knox J W and Buendia L V 2000
Using a crop/soil simulation model and GIS techniques to as-
sess methane emissions from rice fields in Asia. IV. Up scaling to
national levels. Nutr. Cycl. Agroecosyst. 58, 201-217.
Mendoza C, Viteri F E, Lonnerdal B, Raboy V, Young K A and
Brown K H 2001 Absorption of iron from unmodified maize
and genetically altered, low-phytate maize fortified with ferrous
sulfate or sodium iron EDTA. Am. J. Clin. Nutr. 73, 80-85.
Moya P F, Dawe D, Pabale D, Tiongco M, Chien N V, Devarajan
S, Djatiharti A, Lai N X, Niyomvit L, Ping X H, Redondo G
and Wardana P 2002 The economics of intensively irrigated rice
in Asia. In Increasing Productivity of Intensive Rice Systems
through Site-Specific Nutrient Management. Eds. A Dobermann,
C Witt and D Dawe. Science Publishers, Inc., International Rice
Research Institute, New Delhi, Los Banos (in press).
NASS 2001 National Agricultural Statistics Service
(http://www.usda.gov/nass). USDA, Washington, DC.
Olk D C, Cassman KG, Randall E W, Kinchesh P, Sanger L J and
Anderson J M 1996 Changes in chemical properties of soil or-
ganic matter with intensified rice cropping in tropical lowland
soils. Eur. J. Soil Sci. 47, 293-303.
Olk D C, Cassman K G, Simbahan G C, Sta.Cruz P C, Ab-
dulrachman S, Nagarajan R, Tan P S and Satawathananont
S 1999 Interpreting fertilizer-use efficiency in relation to soil
nutrient-supplying capacity, factor productivity, and agronomic
efficiency. Nutr. Cycl. Agroecosyst. 53, 35-41.
Olson R A, Frank K D, Grabouski P Hand Rehm G W 1982
Economic and agronomic impacts of varied philosophies of soil
testing. Agron. J. 74, 492-499.
Padgitt M, Newton D, Penn Rand Sandretto C 2000 Production
practices for major crops in U.S. agriculture, 1990-97. Statistical
bulletin no. 969. USDA, ERS, Washington, DC.
Parks W L 1985 Interaction of potassium with crop varieties or
hybrids. In Potassium in Agriculture. Ed. R D Munson. pp.
536-558. ASA, CSSA, SSSA, Madison, WI.
Peng Sand Cassman K G 1998 Upper thresholds of nitrogen up-
take rates and associated N fertilizer efficiencies in irrigated rice.
Agron. J. 90, 178-185.
Peng S, Garcia F V, Gines H C, Laza R C, Samson M I, Sanico A L,
Visperas R M and Cassman K G 1996 Nitrogen use efficiency of
irrigated tropical rice established by broadcast wet-seeding and
transplanting. Fertil. Res. 45, 123-134.
Peng S, Cassman K G, Virmani S S, Sheehy J E and Khush G S
1999 Yield potential trends of tropical rice since the release of
IR8 and the challenge of increasing rice yield potential. Crop
Sci. 39, 1552-1559.
Pierce F J and Nowak P 1999 Aspects of precision agriculture. Adv.
Agron. 67, 1-85.
Potrykus I 2001 Golden rice and beyond. Plant Physiol. 125, 1157-
1161.
PPI 1999 Impact of biotechnology on the fertilizer industry.
PPIIPPICIFAR Technical Bulletin 1999-1. PPI, PPIC, FAR,
Norcross, GA. 34 pp.
PPI2001 Soil test levels in North America. PPIPPICIFAR Technical
Bulletin 2001-1. Potash & Phosphate Institute, Norcross, GA.

Pretty J, Brett C, Gee D, Hine R E, Mason C F, Morison J I L, Raven
H, Rayment M D and van der Bijl G 2000 An assessment of the
total external costs of UK agriculture. Agric. Syst. 65, 113-l36.
Rehm G W, Schmitt M, Randall G W, Lamb J A and Eliason
R 2000 Fertilizing corn in Minnesota. FO-3790. University of
Minnesota, Minneapolis, MN.
Robertson G P, Paul E A and Harwood R R 2000 Greenhouse gases
in intensive agriculture: contributions of individual gases to the
radiative forcing of the atmosphere. Science 289,1922-1925.
Rosegrant M W, Paisner M S, Meijer S and Witcover J 2001
Global food projections to 2020: Emerging trends and alternative
futures. IFPRI, Washington, DC.
Saxena D and Stotzky G 2001 Bt corn has a higher lignin content
than non-Bt corn. Am. J. Bot. 88, 1704-1706.
Schroeder J J, Neeteson J J, Oenema 0 and Struik P C 2000 Does
the crop or the soil indicate how to save nitrogen in maize pro-
duction? Reviewing the state of the art. Field Crops Res. 66,
151-164.
Schuster C and Rathke G W 2001 Nitrogen fertilisation of trans-
genic winter oilseed rape. In Plant Nutrition - Food security and
sustainability of agro-ecosystems. XIV International Plant Nu-
trition Colloquium. Eds. W J Horst et al. pp. 336-337. Kluwer,
Dordrecht.
Shapiro C A, Ferguson R B, Hergert G W, Dobermann A and
Wortmann C S 2001 Fertilizer suggestions for corn. NebGuide
G74-l74-A. Univ. of Nebraska Coop. Ext. Service, Lincoln, NE.
Sheehy J E, Dionora M J A, Mitchell P L, Peng S, Cassman K
G, Lemaire G and Williams R L 1998 Critical nitrogen con-
centrations: implications for high-yielding rice (Oryza sativa L.)
cultivars in the tropics. Field Crops Res. 59, 31-41.
Sheehy J E, Mitchell P L and Hardy B (eds) 2000 Redesigning Rice
Photosynthesis to Increase Yield. International Rice Research
Institute, Makati City, Philippines. Elsevier Science, Amsterdam.
Sinclair T R and Park W I 1993 Inadequacy of the Liebig limiting-
factor paradigm for explaining varying crop yields. Agron. J. 85,
742-746.
Smaling EM A and Janssen B H 1993 Calibration of QUEFTS, a
model predicting nutrient uptake and yields from chemical soil
fertility indices. Geoderma 59,21-44.
Spalding R F and Exner M E 1993 Occurrence of nitrate in
groundwater - a review. J. Environ. Qual. 22, 392-402.
Suzuki A 1997 Fertilization of Rice in Japan. Japan FAO Associ-
ation, Tokyo, Japan. 111 pp.
Teo Y H, Beyrouty C A, Norman R J and Gbur E E 1995 Nutrient
uptake relationship to root characteristics of rice. Plant Soil 171,
297-302.
Timsina J and Connor D J 200 1 Productivity and management of
rice-wheat cropping systems: issues and challenges. Field Crops
Res. 69, 93-l32.
Tollenaar M and Wu J 1999 Yield improvement in temperate maize
is attributable to greater stress tolerance. Crop Sci. 39, 1597-
1604.
Uri N D 1998 Environmental considerations in the fertilizer use
decision. Environ. Geol. 34, 103-110.
USDA 2001 Agricultural Chemical Usage. 2000 Field crops sum-
mary. USDA-National Agricultural Statistics Service, Washing-
ton, DC.
van Alphen B J and Stoorvogel J J 2000 A functional approach to
soil characterization in support of precision agriculture. Soil Sci.
Soc. Am. J. 64, 1706-17l3.
van der Ploeg R R, Schweigert P and Bachmann J 2002 Use and
misuse of nitrogen in agriculture: the German story. In Proceed-
ings of the Second International Nitrogen Conference, 14-18
October, Washington DC. (in press)
175
van Keulen H 1977 Nitrogen requirements of rice with special ref-
erence to Java. Ministry of Agriculture, Agricultural Technical
Assistance Programme, Bogor, Indonesia. 67 pp.
van Keulen H 1986 Crop yield and nutrient requirements. In Mod-
elling of Agricultural Production: Weather, soils and crops. Eds.
H van Keulen and J Wolf. pp. 155-181. Pudoc, Wageningen.
van Keulen H and Van Heemst H D J 1982 Crop response to
the supply of macronutrients. Agric. Res. Rept. 916. Pudoc,
Wageningen. 46 pp.
Varvel G E, Schepers J S and Francis D D 1997 Ability for in-season
correction of nitrogen deficiency in corn using chlorophyll
meters. Soil Sci. Soc. Am. J. 61, 1233-1239.
Viscarra Rossel R A and McBratney A B 1998 Soil chemical analyt-
ical accuracy and costs: implications from precision agriculture.
Aust. J. Exp. Agric. 38, 765-775.
Viscarra Rossel R A, Goovaerts P and McBratney A B 2001 Assess-
ment of the production and economic risks of site-specific liming
using geostatistical uncertainty modelling. Environmetrics 12,
699-711.
Vyn T J and Janovicek K J 2001 Potassium placement and till-
age system effects on corn response following long-term no till.
Agron. J. 93, 487-495.
Waldroup P W, Kersey J H, Saleh E A, Fritts C A, Yan F, Stilborn
H L, Crum R C and Raboy V 2000 Nonphytate phosphorus re-
quirement and phosphorus excretion of broiler chicks fed diets
composed of normal or high available phosphate corn with and
without microbial phytase. Poult. Sci. 79, 1451-1459.
Wang G H, Dobermann A, Witt C, Sun Q Z and Fu R X 2001
Performance of site-specific nutrient management for irrigated
rice in southeast China. Agron. J. 93, 869-878.
Wassmann R, Lantin R S, Neue H U, Buendia L V, Corton T M
and Lu Y 2000 Characterization of methane emissions from rice
fields in Asia. III. Mitigation options and future research needs.
Nutr. Cycl. Agroecosyst. 58, 23-36.
Wibawa W D, Dludlu D L, Swenson L J, Hopkins D G and Dahnke
W C 1993 Variable fertilizer application based on yield goal, soil
fertility, and soil map unit. J. Prod. Agric. 6, 255-261.
Witt C, Arah J R M and Siband P 2001 Towards a simple, generic
modelling approach for developing site-specific nitrogen man-
agement strategies in irrigated rice. In Plant Nutrition - Food
security and sustainability of agro-ecosystems. Eds. W J Horst et
al. pp. 730-731. Kluwer, Dordrecht.
Witt C, Cassman K G, Olk D C, Biker U, Liboon S P, Samson M I
and Ottow J C G 2000 Crop rotation and residue management ef-
fects on carbon sequestration, nitrogen cycling and productivity
of irrigated rice systems. Plant Soil 225, 263-278.
Witt C, Dobermann A, Abdulrachman S, Gines H C, Wang G H,
Nagarajan R, Satawathananont S, Son T T, Tan P S, Tiem L V,
Simbahan G C and Olk D C 1999 Internal nutrient efficiencies
of irrigated lowland rice in tropical and subtropical Asia. Field
Crops Res. 63, 1l3-l38.
Wood S, Sebastian K and Scherr S J 2000 Pilot analysis of global
ecosystems: agroecosystems. IFPRI, WRI, Washington, DC.
Woodhead T, Huke R, Huke E and Balababa L 1994 Rice-wheat
atlas of India. IRRI, CIMMYT, Los Banos, Philippines. 147 pp.
Yadvinder-Singh, Dobermann A, Bijay-Singh, Bronson K F and
Khind C S 2000 Optimal phosphorus management strategies for
wheat-rice cropping on a loamy sand. Soil Sci. Soc. Am. J. 64,
l4l3-1422.
Young A 1999 Is there really spare land? A critique of estimates of
available cultivable land in developing countries. Environ. Dev.
Sustainability 1,3-18.
Zeigler R S 2001 Agricultural biotechnology. Reducing poverty in
developing countries. Plant Dis. 85,568-579.
Nutrient dynamics in natural and agricultural ecosystems - Processes and modelling
 Horst et al. (Eds.), Progress in Plant Nutrition: 179-188,2002.
© 2002 Kluwer Academic Publishers.
179

Nitrogen saturation in temperate forest ecosystems: current theory,

remaining questions and recent advances

John D. Aber*
Complex Systems Research Center, University of New Hampshire, Durham, NH 03824, USA

Key words: abiotic immobilization, deposition, foliar N, isotopes, land use, nitrate

Abstract
Nitrogen saturation is a continuum, not a state or threshold. The trajectories of change for several different pro-
cesses along this continuum from N-limited to non-N-limited have been well-defined and can be summarized in
a coherent theory. Forests in many parts of the world and under different climatic regimes follow much the same
pathway. Key changes along this continuum include the inducement of nitrification and nitrate leaching and the
initiation of declines in forests growth. Previous land use history appears to be an important factor determining
the initial degree of N limitation in forests, while vegetation type (evergreen versus deciduous) determines the rate
at which N saturation proceeds. Because of the predictable nature of the syndrome of N saturation, either single
stand characteristics (e.g. foliar N concentration) or indices which integrate several characteristics appear to be
useful in mapping current conditions. Detection of temporal trends in N saturation through these indicators can be
made difficult by the effects of interannual variability in climate or pulse disturbance events. Remaining mysteries
associated with N saturation focus on the mechanisms by which very large amounts of nitrogen continue to be
immobilized into soils even after large cumulative doses have been received. There is a growing body of evidence
suggesting that chemical or abiotic incorporation of mineral N into dissolved organic carbon or solid phase soil
organic matter is an important and very poorly understood pathway in the N cycle of forests. The possible role of
mycorrhizae in the assimilation and stabilization of large amounts of added N remains unclear.

Introduction elevated rates of N deposition. While N deposition

Nitrogen (N) emissions to the atmosphere due to hu-
man activity remain elevated in industrialized regions
of the world and are accelerating in many develop-
ing regions (Galloway, 1995). While the deposition
of sulfur (S) has been reduced over much of the U.S.
and Europe by aggressive environmental protection
policies, reducing N deposition requires regulation of
non-point sources (e.g. automobiles) and reduction
targets in the U.S. and elsewhere are more modest. The
importance of nitrogenous fertilizers in meeting the
growing global demand for food ensures increasing
deposition of ammonium across the developing world,
and especially in Asia (Galloway, 1998).
Much of the developed and most rapidly develop-
ing regions on the Earth are embedded within forest
biomes, such that remaining forests are subjected to
* FAX No: +603-862-0188. E-mail: John.aber@unh.edu
in remote areas may range from 0.1-0.2 g N'm- 2
ye 1 and may be dominated by organic rather than
inorganic forms (Weathers et aI., 2000), forests near
heavily industrialized areas may experience inputs of
more than 2 g N'm- 2 ye 1 . Inputs of ammonium from
intensive agriculture may increase this total to 4-5 g
N'm- 2yr -1 regionally and to over 10 g N'm- 2 yr- 1
locally (Dise et aI., 1998).
The potential negative effects of excess N depos-
ition on forests, and the term 'nitrogen saturation'
were first introduced into the literature only 16 years
ago (Nihlgard, 1985). Since that time, the European
community has supported a well-coordinated program
of research on forest effects (e.g. the NITREX pro-
gram, Wright and van Breeman, 1995; Wright and
Rasmussen, 1998) and has progressed to the definition
and establishment of critical loads (e.g. Nilsson and
Grennfelt, 1988; Henricksen et aI., 1992; Warfvinge
and Sverdrup, 1992; Wright and Rasmussen, 1998).
180
In the U.S., despite the lack of a policy mandate
or organized governmental research program, signi-
ficant research projects have been carried forward
(see especially review by Fenn et aI., 1997). Inter-
national cooperation has been excellent in this field,
and an integrated theory of the interactions between N
deposition and forests has emerged.
The purpose of this paper is to briefly review our
current understanding of nitrogen saturation in forest
ecosystems, identify remaining unanswered questions,
and summarize recent work.
Current Theory
Background
The term nitrogen saturation has been defined in many
ways (Aber et aI., 1989; Agren and Bossata, 1988;
Stoddard, 1994), but all definitions include increases
in N availability over time which alleviate N limit-
ations on biological function resulting in increased
nitrification and nitrate mobility in soils. It is im-
portant to emphasize that N saturation is a process
and not a point. Much of the divergence in definition
results from attempts to say that "forests are N satur-
ated when ... " with different arbitrary points offered
as the defining characteristic, including the initiation
of nitrate leaching, the saturation of N demand by
trees, or equality of N inputs and outputs. N satura-
tion should be presented and discussed as a continuum
from highly N-limited to non-N limited, with many
processes changing continuously, but non-linearly, as
N availability increases.
One initial set of hypotheses regarding N saturation
(Aber et aI., 1989) emphasized nitrification and ni-
trate mobility as key processes controlling ecosystem
response to N deposition, and presented the follow-
ing view of interactions between N availability and
forest processes (see paper for supporting references).
Under background or low N deposition conditions,
plant growth in temperate forests is most frequently
limited by nitrogen availability, while the metabolism
of free living soil microbes is most generally lim-
ited by carbon (or energy). These essential constraints
on the biogeochemistry of forested ecosystems are
self-perpetuating in that plants incorporate available
N released by microbial decomposition into carbon
compounds that are energetically and structurally diffi-
cult to decompose. The partial decomposition of fresh
plant tissues shed as litter results in both a further
Succession, N Deposition
~
Harvest, Fire, Agriculture
III
200
<Il
·i;;J
. 100
;.
.: Methane
!
Consumption
0
0 1
Stage
Figure 1. A summary view of the effects on several important
ecosystem characteristics and processes of chronic N additions to
N-limited forests (after Aber et al., 1998). NPP = Net Primary
Production.
concentration of N in the resulting organic matter (hu-
mus) and a further degradation in the decomposability
of this residue. The rate-limiting step in the nitrogen
cycle of temperate forest ecosystems is thus the humus
decay process, and the largest pool of N in these sys-
tems is in soil organic matter. In addition, published
studies suggested that, in acid forest soils, nitrification
is controlled mainly by competition between plants
and nitrifying microbes such that nitrification tends
to occur only where the supply of ammonium is high
relative to plant demand.
Summary of experimental results
A large number of site-to-watershed-Ievel experiments
and regional-scale surveys have been carried out in
the last decade-plus to increase our understanding of
the process of N saturation. These include chronic N
additions (e.g. Magill et aI., 2000; Norton and Fernan-
dez, 1999; Emmet et al., 1998; Moldan et aI., 1998;
Gundersen, 1998; McNulty et aI., 1996), nitrogen re-
movals or 'clean-rain experiments' (e.g. Beier et aI.,
1998; Boxman et al., 1998) and measurements of N
status and cycling along ambient N deposition gradi-
ents (e.g. McNulty et al., 1991; Dise et al., 1998).
Results from these experiments and observations can
be combined to derive a summary view of the effects
of N additions on ecosystem processes (Figure 1).

Soil C:N, N mineralization and nitrification
A central, early hypothesis in the N saturation liter-
ature was that chronic additions of nitrogen should
lead to increased N content in systems (narrower eN
ratios), followed by higher rates of net N mineral-
ization, nitrification and nitrate leaching. However,
long-term results have qualified this characteristic of
N saturation.
At the Harvard Forest study site (Aber et aI., 1998)
initial increases in N mineralization with chronic N
additions were followed by actual decreases in net N
mineralization by the sixth year of application. Several
older studies have also reported decreases in N miner-
alization under N enriched conditions (e.g. Baath et
aI., 1981 and Soderstrom et aI., 1983, see also review
by Fog et aI., 1988). In the NITREX studies in Europe,
Tietema (1998), using 15N pool dilution, found in-
creases in both gross mineralization and gross immob-
ilization rates across a nitrogen deposition gradient,
but net N mineralization peaked at intermediate N de-
position. Gundersen et aI. (1998) report increases in
field measured net N mineralization rates with N ad-
dition only at N-limited sites, while those with high
initial rates of N cycling showed decreases in net N
mineralization with further N additions.
Two extant hypotheses could explain this decline
in net N mineralization rates. The first is that the ad-
dition of N to soil organic matter occurs in a way that
randomizes chemical bond structures in soil organic
matter, reducing efficiencies of extracellular catabolic
enzymes and decreasing decay rates (e.g. Berg, 1986).
Thus N applications result in higher N concentration
in soil organic matter which reduces decomposabil-
ity and net N mineralization despite a narrower C:N
ratio. The second hypothesis states that the produc-
tion of humus-degrading enzymes by soil microbes
(especially fungi) is suppressed in the presence of el-
evated concentrations of mineral N, thereby reducing
N mineralization (Keyser et aI., 1978; Tien and Myer,
1990; Fog, 1988). The relative importance of these
two mechanisms under field conditions cannot be eval-
uated with available data, although the inhibition of
enzyme production may be especially relevant where
experimental manipulations result in very high rates of
N addition.
One of the most consistent results across studies is
the relationship between soil eN ratio and net nitrific-
ation. Data from a spruce-fir forest in Vermont, USA,
a gradient of spruce-fir forests across New England,
and from the intensive NITREX study sites all show
a strong linear negative relationship between forest
181
floor C:N and net nitrification (McNulty et aI., 1996;
Tietema and Beier, 1995). The only experimental
stand that showed neither a decrease in net N miner-
alization nor an increase in net nitrification was the
Harvard Forest mixed hardwood site (see discussion
below). Moreover, all of these studies yielded a sim-
ilar critical eN value below which nitrification began.
Unfortunately, the size and variability of soil carbon
and nitrogen pools has made detection of changes in
C:N ratios difficult in most experiments, although at
least one regional survey has linked rates of ambient
N deposition with forest floor eN (McNulty et aI.,
1991).
Nitrate losses and nitrogen retention
Increases in nitrate leaching with N additions, or
decreases in 'clean-rain' experiments has also been
consistent across studies. However, decreases in the N
removal experiments were dramatic, immediate, and
proportional to the removals, while increases in losses
were only a small fraction of the experimental addi-
tions (Bredemeier et aI., 1998; Magill et aI., 2000;
Quist et aI., 1999). In broad surveys of nitrate leaching
in response to N deposition in Europe and the U.S.,
Dise and Wright (1995) and Peterjohn et aI. (1996)
showed that nitrate leaching losses averaged much less
than half inorganic N deposition. Similar results have
been compiled for North American sites by Peterjohn
et aI. (1996, see also Christ et aI., 1995). Effect-
ive retention of added N, even when addition rates
are many times ambient, is one of the most difficult
responses to understand, and the processes involved
remain uncertain. This will be discussed in more detail
below.
Foliar N
Increases in foliar N with N additions (and decreases
with removals) are also consistent across experiments
(Bredemeier et aI., 1998; Boxman et aI., 1998; Ma-
gill et aI., 2000; McNulty et aI., 1996). Most studies
also show changes in N:Ca and N:Mg ratios as a result
mainly of changes in foliar N, but shifts in Ca:Al ratios
suggest important changes in the biogeochemistry of
those elements in response to increased nitrification
and nitrate leaching (Gilliam et aI., 1996; Dise and
Wright, 1995; Gundersen et aI., 1998; Magill et aI.,
1997; McNulty et aI., 1996).
Increased foliar N concentration can affect eco-
system carbon balance through increased rates of net
photosynthesis (Reich et aI., 1995; Field and Mooney,
1986). This increase can increase aboveground bio-
182
mass production (the fertilizer effect), and may also
provide additional carbon (and energy) to fuel below-
ground processes. However, above a certain N content,
diversion of N from photosynthetic enzymes (e.g. Ru-
bisco) to secondary metabolites, waste products and
stress response compounds (e.g. arginine, polyamines)
decouples foliar N from carbon gain and may repres-
ent a sensitive indicator of N status and ecosystem
stress (Boxman et aI., 1998; Minocha et al., 2000).
Net primary production
The apparent non-linear response of forest NPP to N
additions is one of the most difficult to present and
most important to policy formation. By far the most
common response to N additions in temperate zone
forests is an increase in tree growth. However, early
work by Schulze (1989) and colleagues demonstrated
that N:Mg imbalances in the foliager could reduce
carbon acquisition and initiate forest decline. Exper-
imental additions in a red pine stand at the Harvard
Forest and a spruce-fir forest in Vermont both demon-
strate significant declines in production of needle-
leaved evergreen trees. These findings suggest that the
significant forest decline in high elevation spruce/fir
forests in the last three decades in Western New Eng-
land and Eastern New York (Siccama et al., 1982;
Hornbeck and Smith, 1985) may have been due in part
to increased N deposition. Boxman et al. (1998) report
increased tree growth in two NITREX sites with high
N cycling following the reduction of N inputs through
under-canopy roof construction.
Fine roots
As a very difficult parameter to measure, there are
fewer data on responses of fine roots to N deposition.
At the Harvard Forest, no changes in fine root bio-
mass were measured through year 4 of N additions,
but root N concentration nearly doubled (Magill et al.,
1997). In the NITREX studies, fine root biomass in-
creased significantly following N exclusion in the site
with the highest N availability (Gundersen et al., 1998;
Boxman et al., 1998). Hendricks et al. (2000) meas-
ured significant changes in fine root substrate quality
across an N fertility gradient, suggesting important
feedbacks between N enrichment and future rates of
N mineralization.
The possible role of land use in preconditioning
ecosystem response to N additions
While the trends in Figure 1 appear consistently across
studies and locations, it is clear that not all stands used
in N addition experiments began at the same location
along the N saturation continuum nor moved at the
same rate to the right as N additions occurred. Two
summary statements can be made here: 1) In general,
broad-leaved deciduous forests are more resistant to
N saturation than needle-leaved evergreen stands and
move more slowly to the right in Figure 1 for a given N
addition, and 2) Initial N status of forests is the result
of previous land use or disturbance history.
In a comparison of responses to N additions in
different forests, a red pine stand and a spruce-fir
stand were shifted completely into a stage three re-
sponse while two deciduous stands receiving similar
additions changed very little (Aber et aI., 1998). In
addition, strong changes in forest response across an
ambient N deposition gradient in the northeastern U.S.
were found for spruce-fir forests, but not for decidu-
ous stands (Lovett and Reuth, 1999, see additional
discussion below).
The statement on land use and disturbance is more
difficult to support. A general comparison between
two stands in the northeastern U.S. showed that the
stand with a more intensive history of extractive land
use and higher N deposition rates was far more resist-
ant to increased N additions than a less disturbed stand
receiving lower N inputs (Magill et al., 1996, 2000).
This suggests that land use, rather than cumulative
deposition, was the more important determinant of N
status. Additionally, data summarisesd from European
sites show that forest floor C:N ratio, reflective of both
past land use and current N status, is significantly
correlated with N leaching losses (Tietema and Beier,
1995; Gundersen et al., 1998). In high elevation sites
in the southeastern U.S., both deposition rate (which
increases with elevation) and previous land use have
been shown to affect nitrate concentrations in streams
(Silsbee and Larson, 1982, 1983; Nodvin et al., 1995).
Related to this is another implication from Figure 1
that undisturbed, mature, or steady-state forests should
be near the middle of stage 2, where nitrogen losses
are sufficient to balance N inputs in deposition. Stands
to the left of this in Figure 1 represent disturbed con-'
ditions in which N has been removed from the site
resulting in a slower N cycle.
Summary indicators of N saturation
Figure 1 suggests an integrated, simultaneous change
in several ecosystem characteristics and processes
with increasing N availability. If this is the case then
it should be possible to derive indicators of N status
based on some or all of these variables which can be

used to ordinate ecosystems along the N saturation
gradient. Gundersen et aI., 1998 have developed such
an indicator based on a series of metrics available for
the intensive NITREX study sties. The indicator ap-
pears to be responsive to the experiments conducted,
increasing with N additions and decreasing with N
removals.
As foliar N is one of the most consistent response
variables in N additions and removal experiments,
this metric alone may prove valuable in mapping
for characterizing N status across landscapes. Recent
work demonstrating that foliar N concentration can be
measured by high-spectral-resolution remote sensing
techniques (Martin and Aber, 1997; Ollinger et aI., in
press; Smith et aI., 2002) suggests that methods for
producing maps of N status at high resolution may be
forthcoming.
Remaining questions and recent advances
Overview
This broad understanding of the process of N satur-
ation allows us to focus on the remaining mysteries.
There are a number of critical questions which may
all be related to the same or similar processes. Two
process-related questions are: 1) How quickly will
stands move toward saturation, and why do evergreens
and deciduous stands move at different rates? and 2)
Why is N retention efficiency so high in many stands,
even those with large accumulations of N inputs for
very high experimental rates of N addition? Another
suite of questions relates to changes in nitrate loss over
space and time: 1) What is the spatial pattern of ni-
trate loss and what underlies this pattern? and 2) What
processes cause large interannual changes in nitrate
loss?
Soil/plant/mycorrhizal interactions
Far more than half of added N is to be found in non-
extractable forms in soils within a year of addition,
based on either 15N tracer data, or simpler mass bal-
ance calculations (Nadelhoffer et aI., 1999a, b; Magill
et aI., 2000). In an earlier paper (Aber et aI., 1998)
we argued that only two processes are consistent with
measurements in terms of efficient retention of N: 1)
abiotic immobilization and 2) mycorrhizal assimila-
tion without increased biomass production. These two
could account for large measured rates of N incor-
poration into soils without increased C02 flux. Both
183
abiotic incorporation and the role of mycorrhizae are
being actively investigated at this time and results are
beginning to appear.
Abiotic incorporation of DIN into soil organic matter
Significant rates of direct chemical incorporation of
NH4 + into SOM have been measured by mass bal-
ance using sterilized soil samples (Nommik, 1970;
Axelsson and Berg, 1988; Schimel and Firestone,
1989; Sen and Chalk, 1995). Measured rates of direct
chemical incorporation of N~ + range from 0.0036
(Nommik, 1970) to 112 mg NH4-N/kg C/day (Sen and
Chalk, 1995). It is difficult to interpret these results
in terms of field N retention rates because the meth-
ods used (incubation times, NH4 + concentrations,
extraction procedures, and the method of soil steriliz-
ation) vary between studies. In particular, all methods
of soil sterilization (gaseous fumigation, solutions of
heavy metal salts, gamma irradiation, autoclaving)
may influence the incorporation of NH4 + by altera-
tion of the substrate. However, Foster et aI. (1985a, b)
have shown that significant immobilization can occur
through reactions between ammonia and organic mat-
ter in forest soils treated with urea, which creates both
high pH and high ammonia concentrations, conditions
which foster this reaction.
Abiotic incorporation of dissolved inorganic nitro-
gen (DIN) into soils has also been estimated by the
rapid (15-30 minute) disappearance of 15N tracer in
pool dilution experiments (Davidson et aI., 1991). This
rapid incorporation was not strongly influenced by
autoclaving of soil cores. Both the speed of immob-
ilization and the insensitivity to autoclaving suggest
that abiotic processes were involved.
In recent studies, Johnson et aI. (2000) report re-
lative rates of biotic and abiotic immobilization for
organic and mineral soil horizons from a wide variety
of forest types with large differences in N status. They
concluded that biotic N immobilization was highly
variable between sites and was inversely related to soil
N concentration. Abiotic immobilization was more
consistent between sites and was unrelated to N con-
centration or pH. Because of these differences, the
fraction of total immobilization due to abiotic pro-
cesses varied from 6-90% and was directly related to
soil N concentration. It is significant that none of the
existing models of forest N dynamics contain an expli-
cit algorithm for abiotic N immobilization, a process
that apparently accounts for nearly half, on average, of
total N incorporation into soil organic matter (Johnson
et aI., 2000).
184
Berntson and Aber (2000) carried out 15N pool di-
lution studies with soils from the chronic N plots at the
Harvard Forest. They showed that nearly all measured
nitrate immobilization in the pine control and high N
plots, and the hardwood high N plot, occurred within
the first few minutes. Only in the hardwood control
stand did significant immobilization occur between 1
and 24 hours. As 'fast' immobilization is commonly
associated with abiotic processes, this suggests a large
role for these processes in N retention at the Harvard
Forest site.
Dail and Davidson (2001) expanded on this study
by comparing the fates of nitrate and nitrite in con-
trol, gamma irradiated and autoclaved soils. 30, 40 and
60% of added nitrate was recovered as soluble organic
N (dissolved organic N or DON), respectively, in soils
receiving these treatments, suggesting rapid and par-
tially abiotic conversion of nitrate to organic forms.
Added nitrite was rapidly converted to insoluble or-
ganic N. The authors conclude that abiotic processes
can be important in N retention, and that no accepted
soil process can account for their results.
The rapid conversion of nitrate to DON could be
a major process for conveying added N into soil or-
ganic matter, even though DON is never a large pool
(Figure 2). Dissolved organic matter (DaM) concen-
trations in soil solution are maintained by a dynamic
equilibrium between large sorption/desorption fluxes.
Chemical incorporation of mineral N into DaM, fol-
lowed by rapid resorption of the N-enriched DaM
(lower DOC:DON ratio, Currie et al., 1996) could
move relatively large amounts of N rapidly into soil
organic matter. To test this idea of a 'DaM con-
veyor' for inorganic N into soil organic matter, we
need to know more about the kinetics and reactions
of mineral N incorporation into DaM and about gross
sorption/desorption rates for DaM.
Additional indirect evidence for abiotic pro-
cesses of N fixation comes from an integrated
tracer/modeling study. Currie and Nadelhoffer (1999)
present the TRACE model and apply the model to 15N
redistribution data from the Harvard Forest. Among
the results presented is the requirement to decouple N
immobilization in soils from the carbon cycle through
microbes, making gross N immobilization not depend-
ent on a carbon substrate for the fueling of biotic
immobilization.
The incorporation of mineral N into soil organic
matter through chemical fixation reactions has been
verified by 15N NMR spectrometry. Laboratory stud-
ies have shown that NHt (Thorn and Mikita, 1992)
NADDITIONS
Soil
Organic
Matter
PLANT UPTAKE
MICROBIAL
IMMOBILIZATION
LEACHING LOSS
Figure 2. The 'DOM conveyor' for incorporation of dissolved in-
organic nitrogen (DIN) into soil organic matter relies on rapid
reactions between mineral N and dissolved organic matter (DOM)
and rapid sorption/desorption rates for DOM.
and other N compounds, such as glycine (Benzing-
Purdie et aI., 1983) undergo reactions with soil organic
matter that result mainly in heterocyclic forms of N
(e.g. indoles and pyrroles) and some amides (Thorn
and Mikita, 1992). Biotic N incorporation in soils
transforms N primarily into amides (Almendros et al.,
1991; Knicker et al., 1993; Clinton et al., 1995).
However, 15N NMR spectra from soil organic mat-
ter and compost residues in which biotic processes
were not excluded suggest that direct chemical fixation
explains 15% or less of the total N retention. Nat-
ural abundance spectra of SaM (Knicker et al., 1993),
spectra from labeled organic matter amendments (Al-
mendros et aI., 1991) and spectra from labeled mineral
N amendments (Clinton et aI., 1995) show that about
85% of N retained occurs in the amide form. This
discrepancy between estimates of the importance of
abiotic N immobilization by NMR methods and soil
15N methods remains to be resolved. Perhaps the
assumption that only heterocyclic N results from abi-
otic immobilization leads to underestimation of this
process by NMR techniques.
Mycorrhizal assimilation
There can be no doubt that mycorrhizae represent the
most important component of forest ecosystems about
which very little is known. Mycorrhizae are ubiquit-
ous in forest soils and can comprise a large fraction
of the total fungal biomass which, in tum, can be
a large fraction of the total microbial biomass (Kill-
ham, 1994). This fraction may be very large under
N-limiting conditions and decline with N saturation

(Tietema, 1998). Mycorrhizae are longer-lived than
free living microbes, act as an extension of root sys-
tems, and are generally thought of as a mechanism
by which plants increase the efficiency of nutrient
and water uptake (Yanai et al., 1995). Still, there is
some uncertainty about the degree to which the plant
or the fungus controls this interaction (Rygiewicz and
Andersen, 1994).
While direct studies of C and N processing by
mycorrhizae in the field have not been undertaken, in-
novative isotopic approaches are beginning to provide
some insights into changes in the role of mycorrhizae
over N availability gradients. For example, Hobbie et
ai. (1999,2000) demonstrate significant changes in the
fractionation of natural abundance 15N between soils
and foliage. By examining 15N in SOM, extractable
N, roots, and foliage, they determined that the de-
gree of fractionation through this system varied with
N status, with fractionation being largest in the sites
with the lowest levels of N availability. A modeling
analysis suggested that increased fractional retention
of N by mycorrhizae was the process most consistent
with measured results. Reduced retention by mycor-
rhizae at high N availability may then contribute to
nitrate leaching.
Spatial and temporal patterns in nitrate losses
Nitrate concentrations in surface waters are a key in-
dicator of N status or the degree of N saturation (Aber
et al., 1998; Stoddard, 1994). If N saturation is a key
process controlling the biogeochemistry of ecosys-
tems, then clear patterns should exist in the landscape
which link the duration and intensity of N deposition
with nitrate loss rates. While examples of changes
in nitrate concentrations over time in areas of heavy
deposition, or across spatial gradients of N depos-
ition have been reported (e.g. Peterjohn et al., 1996;
Murdoch and Stoddard, 1993; McNulty et aI., 1991),
counterexamples are also in the literature (e.g. Aber
and Driscoll, 1997; Murdoch et al., 1998).
It seems clear that previous land use or disturb-
ance history, as well as current species composition
(possibly related to history as well) can override the
effects of N deposition in driving nitrate losses from
forests. For example, Lovett et al. (2000) surveyed 39
streams in the Catskill Mountain region of the north-
eastern US and determined, based on mean nitrate
concentrations and the seasonality of those concen-
trations, that all stages of N saturation were to be
found. Differences could not be related to N depos-
185
ition, topography, or hydrology, and were ascribed
rather to differences in land use history. In contrast, a
clear gradient in nitrate concentrations occurred with
elevation, due to increasing inputs, within a single wa-
tershed with a presumably homogeneous disturbance
history (Lawrence et al., 2000).
An integrated data/modeling analysis of nitrate
losses in and around the Hubbard Brook Research
Forest in New Hampshire (Aber and Driscoll, 1997)
reinforced the prolonged influence of major disturb-
ance events on nitrate losses. Watersheds devegetated
for three years in the 1960s and subjected to a major
fire event in the 1820s both continue to show low ni-
trate losses. On a regional scale, the effects of a major
harvest/burn episode across the White Mountains of
New Hampshire is still visible in soil and stream water
measurements (Goodale et aI., 2000a, b).
Even more subtle 'disturbance' events may al-
ter nitrate losses and cloud the interpretation of time
series data. Murdoch et al. (1998) present evidence for
climatic controls on interannual variability in nitrate
losses, with higher loss rates in wanner years. Mitchell
et al. (1996) attribute a peak in nitrate losses in 1990
to the widespread occurrence of soil frost in the winter
of 1989-90. Aber and Driscoll (1997) found that a re-
latively simple model of C and N dynamics driven by
an actual climate data time series could capture much
of the pattern of interannual changes in nitrate loss at
Hubbard Brook.
Biotic disturbance may playa role as well. Eshle-
man et al. (1998) demonstrate the interaction between
a severe forest defoliation event and dramatic in-
creases in nitrate leaching, with accompanying acid-
ification of stream water.
In terms of species effects, Lovett and Reuth
(1999) showed significant increases in net N mineral-
ization and nitrification along a N deposition gradient
in the northeastern US for stands dominated by sugar
maple (Acer saccharum), but not for stands domin-
ated by beech (Fagus grandifolia). Finzi et ai. (1998)
demonstrate the tendency for species composition to
reinforce existing gradients in the environment. Res-
ults from the Harvard Forest (above) and elsewhere
clearly show that forests dominated by broad-leaved
deciduous versus needle-leaved evergreen species re-
spond very differently to chronic N additions.
Finally, it is becoming clear that the process of
N saturation is not restricted to areas of high pop-
ulation density and industrial activity. Williams et
al. (1996) and Williams and Tonnesson (2000) have
demonstrated clear spatial and temporal trends in N
186
deposition, surface water nitrate and ANC in the Front
Range of the Rocky Mountains in Colorado, and
across the intermountain west in the US. Although
the highest rates of N deposition reported are only 3.2
kg N·ha-l. yr- 1, these rates have increased three-fold
since 1984 in the most heavily impacted areas near
Denver, and show also a 3 to 4-fold range spatially.
A regional survey of lake chemistry in 1985 was re-
peated in 1995 and showed a significant shift toward
higher nitrate concentrations. Similar results can be
found around other urban areas (e.g. Fenn and Poth,
1999, see also review by Fenn et aI., 1997).
Conclusion
Our understanding of the process of N saturation has
increased dramatically since Nihlgard's first paper on
the subject in 1985. There is strong evidence for a
consistent pattern of response across forest ecosys-
tems under a wide range of climatic conditions. The
remaining unknowns involve the timing with which N
saturation occurs and the processes, mainly in the soil
and at the soil/plant interface, that control the parti-
tioning of added N between plants and soil organic
matter. Clear trends of increasing nitrate concentration
in streams over time in areas of high N deposition
can be masked by large or subtle disturbance events
including climatic variation and defoliation.
Acknowledgements
Funding for the research summarized here was
provided by the National Science Foundation - Long-
Term Ecological Research Program (LTER), and the
U.S. Environmental Protection Agency and the USDA
Forest Service.
References
Aber J D and Driscoll C T 1997 Effects of land use, climate vari-
ation and N deposition on N cycling and C storage in northern
hardwood forests. Global Biogeochemical cycles. 11, 639-648.
Aber, J D, McDowell W H, Nadelhoffer K J, Magill A, Berntson
G, Kamakea M, McNulty S G, Currie W, Rustad L and Fernan-
dez I 1998 Nitrogen saturation in temperate forest ecosystems:
hypotheses revisited. BioScience 48, 921-934.
Aber J D, Nadelhoffer K J, Steudler P and Melillo J M 1989 Ni-
trogen saturation in northern forest ecosystems. BioScience 39,
378-386.
Agren G and Bosatta E 1988 Nitrogen saturation of terrestrial
ecosystems. Environ. Poll. 54, 185-197.
Almendros G, Frund R, Gonzales-Vila F J, Haider K M, Knicker
H and Ludemann H D 1991 Analysis of 13C and 15N CPMAS
NMR-spectra of soil organic matter and composts. FEBS Letters
282, 119-121.
Axelsson G and Berg B 1988 Fixation of ammonia ( 15 N) to
Pinus silvestris needle litter in different stages of decomposition.
Scandinavian Journal of Forest Research 3, 273-279.
Baath E, Lundgren B and Soderstrom B 1981 Effects of nitrogen
fertilization on the activity and biomass of fungi and bacteria in
a podzolic soil. Zentralblatt fuer Bakteriologie, Mikrobiologie
und Hygiene 2, 90-98.
Beier C, Blanck K, Bredemeier M, LamersdorfN, Rasmussen Land
Xu Y-J 1998 Field-scale 'clean' rain treatments to two Norway
spruce stands within the EXMAN project - effects on soil solu-
tion chemistry, foliar nutrition and tree growth. For. Ecol. and
Manag. 101,111-124.
Benzing-Purdie L, Ripmeester J A and Preston C M 1983 Elucid-
ation of the nitrogen forms in melanoidins and humic acid by
nitrogen-15 cross polaization-magic angle spinning nuclear mag-
netic resonance spectroscopy. J. Ag. Food Chem. 31, 913-915.
Berg, B 1986 Nutrient release from litter and humus in coniferous
soils - a mini review Scand. J. For. Res. 1,359-369.
Boxman A W, Blanck K, Brandrud T-E, Emmet B A, Gundersen
P, Hogervorst R F, Kjonaas 0 J, Persson H and Timmermann
V 1998 Vegetation and soil biota response to experimentally-
changed nitrogen inputs in coniferous forest ecosystems of the
NITREX project. For. Eco. Manag. 101,65-80.
Bredemeier M, Blanck K, Xu X-J, Tietema A, Boxman A W, Emmet
B, Moldan F, Gundersen P, Schleppi P and Wright R F 1998
Input-output budgets at the NITREX sites. For. Ecol. and Manag.
101,57--64.
Christ M, Zhang Y, Likens G E and Driscoll C T 1995 Nitro-
gen retention capacity of a northern hardwood forest soil under
ammonium sulfate additions. Ecol. App. 5, 802-812.
Clinton P W, Newman R H and Allen RB 1995 Immobilization of
15N in forest litter studied by 15N CPMAS NMR spectroscopy.
Europ. J. of Soil Sci. 46, 551-556.
Currie W S , Aber J D, McDowell W H, Boone R D and Magill A H
1996 Vertical transport of dissolved organic C and N under long-
term N amendments in pine and hardwood forests. Biogeochem.
35,471-505.
Currie W S and N adelhoffer K J 1999 Dynamic redistribution of
isotopically labeled cohorts of nitrogen inputs in two temperate
forests. Ecosystems 2, 4-18.
Dail D B, Davidson E A and Chorover J 2001 Rapid abiotic trans-
formation of nitrate in an acid forest soil. Biogeochem. 54,
131-146.
Davidson E A, Hart S C and Firestone M K 1991 Measuring gross
nitrogen mineralization, immobilization, and nitrification by 15N
isotope dilution in intact soil cores. J. of Soil Sci. 42, 335-349.
Dise N B and Wright R F 1995 Nitrogen leaching in European
forests in relation to nitrogen deposition. For. Ecol. Manag. 71,
153-162.
Dise N B, E Matzner P and Gundersen 1998 Synthesis of nitrogen
pools and fluxes from European forest ecosystems. Wat. Air Soil
Poll. 105, 143-154.
Emmett B A, Reynolds B, Silgram M, Sparks T H and Woods
C 1998 The consequences of chronic nitrogen additions on N
cycling and soil water chemistry in a Sitka spruce stand, North
Wales. For. Ecol. Manag. 101, 165-176.
Eshleman K N, Morgan R P, Webb J R, Deviney F A and Gal-
loway J N 1998 Temporal patterns of nitrogen leakage from
mid-Appalachian forested watersheds, role of insect defoliation.
Wat. Resour. Res. 34, 2005-2016.

Fenn M E and Poth M A 1999 Temporal and spatial trends in stream-
water nitrate concentrations in the San Bernadino Mountains,
Southern California. J. Env. Qual. 28, 822-836.
Fenn, ME, Poth M A, Aber J D, Baron J S, Bormann B T, Johnson D
W, Lemly A D, McNulty S G, Ryan D F and Stottlemyer R 1997
Nitrogen excess in North American ecosystems: predisposing
factors, ecosystem responses and management strategies. EcoL
App. 8, 706-733.
Field C J and Mooney H A, 1986 The photosynthesis-nitrogen rela-
tionship in wild plants. In Givinish, T J (ed.), On the Economy of
Plant Form and Function Cambridge University Press New York
pp.25-55.
Finzi A C, Canham C D and Van Breemen N 1998 Canopy tree-soil
interactions within temperate forests: species effects on pH and
cations. EcoL App. 8,447-454.
Fog K 1988 The effect of added nitrogen on the rate of decomposi-
tion of organic matter. Bio. Rev. 63, 433-462.
Foster N W, Beauchamp E G and Corke C T 1985a Immobilization
of nitrogen-15-labeled urea in a jack pine forest floor. Soil Sci.
Soc. Am. J. 49, 448-452.
Foster N W, Beauchamp E G and Corke C T 1985b Reactions of
15N-Iabeled urea with jack pine forest-floor material. Soil Bio.
and Biochem. 17,699-703.
Galloway J N 1995 Acid deposition: perspectives in time and space.
Wat. Air Soil Poll. 85, 15-24.
Galloway J N 1998 The global nitrogen cycle: changes and con-
sequences. Environ. Poll. 102, 15-24.
Gilliam F S, Adams M B and Yurish B M 1996 Ecosystem nutri-
ent responses to chronic nitrogen inputs at Fernow Experimental
Forest, West VIrginia. Can. J. For. Res. 26, 196-205.
Goodale C L and Aber J D 2001 The long-term effects of land-use
history on nitrogen cycling in northern hardwood forests. EcoL
App. 11,253-267.
Goodale C L, Aber J D and McDowell W H 2000 Long-term effects
of disturbance on organic and inorganic nitrogen export in the
White Mountains, New Hampshire. Ecosystems 3, 433-450.
Gundersen P, Emmet B A, Kjonaas 0 J, Koopmans C J and Tietema
A 1998 Impact of nitrogen deposition on nitrogen cycling in
forests: a synthesis of NITREX data. For. EcoL Manag. 101,
37-56.
Hendricks J J, Aber J D, Nadelhoffer K J and Hallett R D 2000 Ni-
trogen controls on fine root substrate quality in temperate forest
ecosystems. Ecosystems 3, 57---{)9.
Henriksen A, Kamari J, Posch M and Wilander A 1992 Critical
loads to surface waters in the Nordic countries. Ambio 21,
356-363.
Hobbie E A, Macko S A and Shugart HH 1999a Insights into nitro-
gen and carbon dynamics of ectomycorrhizal and saprotrophic
fungi from isotopic evidence. Oecologia 118, 353-360.
Hobbie E A, Macko S A and Williams M 2000 Correlations
between foliar 15 N and nitrogen concentrations may indicate
plant-mycorrhizal interactions. Oecologia 122, 273-283.
Hornbeck J W and Smith R B 1985 Documentation of red spruce
growth decline. Can. J. For. Res. 15, 1199-1201.
Johnson D W 1992 Nitrogen retention in forest soils. J. Env. QuaL
21,1-12.
Johnson D W, Cheng W and Burke I C 2000 Biotic and abiotic
nitrogen retention in a variety of forest soils. Soil Sci. Soc. Am.
J. 64,1503-1514.
Keyser P, Kirk T K and Zeikus J G 1978 Ligninolytic enzyme
systems of Phanerochaete chrysosporium: synthesized in the ab-
sence of lignin in response to nitrogen starvation. J. BacterioL
135,790--797.
187
Knicker H, Frund R and Ludemann D 1993 The chemical nature of
nitrogen in native soil organic matter. Naturwissenschaften 80,
219-221.
Lawrence G B, Lovett G M and Baevsky Y H 2000 Atmospheric
deposition and watershed nitrogen export along an elevational
gradient in the Catskill Mountains, New York. Biogeochemistry
50,21-43.
Lovett G M and Rueth H 1999 Soil nitrogen transformations in
beech and maple stands along a nitrogen deposition gradient.
EcoL App. 9, 1330--1344.
Lovett G M, Weathers K C and Sobczak W V 2000 Nitrogen
saturation and retention in forested watersheds of the Catskill
Mountains, New York. EcoL App. 10, 73-84.
Magill A H, Downs M R, Nadelhoffer K J, Hallett R A and Aber
J D 1996 Forest ecosystem response to four years of chronic
nitrate and sulfate additions at Lead Mountain, Maine. For. EcoL
Manag. 84,29-37.
Magill A H, Aber J D, Hendricks J J, Bowden R D, Melillo J M and
Steudler P 1997 Biogeochemical response of forest ecosystems
to simulated chronic nitrogen deposition. EcoL App. 7, 402-415.
Magill, A , Aber J, Berntson G, McDowell W, Nadelhoffer K,
Melillo J and Steudler P 2000 Long-term nitrogen additions
and nitrogen saturation in two temperate forests. Ecosystems 3,
238-253.
Martin M E and Aber J D 1997 Estimation of forest canopy lignin
and nitrogen concentration and ecosystem processes by high
spectral resolution remote sensing. EcoL App. 7, 431-443.
McNulty S G, Aber J D and Boone R D 1991 Spatial changes in
forest floor and foliar chemistry of spruce-fir forests across New
England. Biogeochemistry 14, 13-29.
McNulty S G, Aber J D and Newman S D 1996 Nitrogen saturation
in a high elevation spruce-fir stand. For. EcoL Manag. 84, 109-
121.
Minocha R, Long S, Magill A, Aber J and McDowell W H 2000
Foliar free polyamine and inorganic ion content in relation to
soil and soil solution chemistry in two fertilized forest stands at
the HArvard Forest, Massachusetts. Plant Soil 222, 119-137.
Mitchell M J, Driscoll C T, Kahl J S, Likens G E, Murdoch P S and
Pardo L H 1996 Climatic control of nitrate loss from forested
watersheds in the northeast United States. Env. Sci. Tech. 30,
2609-2612.
Moldan F and Wright R F 1998 Changes in runoff chemistry after
five years of N addition to a forested catchment at Gardsjon,
Sweden. For. EcoL Manag. 101, 187-198.
Murdoch P S and Stoddard J L 1993 Chemical characteristics and
temporal trends in eight streams of the Catskill Mountains, New
York. Wat. Air Soil PolL 67, 367-396.
Murdoch P S, Bums D A and Lawrence G B 1998 Relation of cli-
mate change to the acidification of surface waters by nitrogen
deposition. Env. Sci. Tech. 32, 1642-1647.
Nadelhoffer K J, Emmett B A, Gundersen P, Kjonaas 0 J, Koop-
mans C J, Schleppi P, Tietema A and Wright R F 1999 Nitrogen
deposition makes a minor contribution to carbon sequestration in
temperate forests. Nature 398, 145-148.
Nadelhoffer, K J , M R Downs and B Fry 1999 Sinks for 15N-
euriched additions to an oak forest and a red pine plantation.
EcoL App. 9, 72-86.
Nihlgard B 1985 The ammonium hypothesis- an additional explan-
ation to the forest decline in Europe. Ambio 14, 2-8.
Nilsson J, Grennfelt P 1988 Critical loads for sulphur and nitrogen
- Report from a workshop held at Skokloster, Sweden, 19-24.
March, 1988 Nordic Council of Ministers Copenhagen 418 p.
Nodvin S C, Van Miegroet H, Lindberg S E, Nicholas N S and
Johnson D W 1995 Acidic deposition, ecosystem processes and
188
nitrogen saturation in a high elevation southern Appalachian
watershed. Wat. Air Soil PolL 85, 1647-1652.
Nommik H 1970 Non-exchangeable binding of ammonium and
amino nitrogen by Norway spruce raw humus. Plant And Soil
33,581-595.
Norton S A and Fernandez I J 1999 The Bear Brook Watershed in
Maine. Dordrecht, Kluwer Academic Publishers.
Ollinger S V, Smith M L, Martin M E, Hallett R A, Goodale C L
and Aber J D 2002 Regional variation in foliar chemistry and soil
nitrogen status among forests of diverse history and composition.
Ecology 83, 339-355.
Peterjohn W T, Adams M B and Gilliam F S 1996 Symptoms
of nitrogen saturation in two central Appalachian hardwood
ecosystems. Biogeochemistry 35, 507-522.
Quist M.E, Nasholm T, Lindeberg J, Johannisson C, Hogbom L and
Hogberg P 1999 Responses of a nitrogen-saturated forest to a
sharp decrease in nitrogen input. J. Env. QuaL 28, 1970-1977.
Reich P B, Kloeppel B, Ellsworth D S and Walters M B 1995 Dif-
ferent photosynthesis-nitrogen relations in deciduous hardwood
and evergreen coniferous tree species. Oecologia 104, 24-30.
Rygiewicz P T and Andersen C P 1994 Mycorrhizae alter quality
and quantity of carbon allocated below ground. Nature 369, 56-
60.
Schimel J P and Firestone M K 1989 Inorganic N incorporation by
coniferous forest floor materiaL Soil Bio. Biochem. 21,41-46.
Schulze E D 1989 Air pollution and forest decline in a spruce (picea
abies) forest. Science 244, 776-783.
Sen S and Chalk P M 1995 Biological interactions between
soil nitrogen and alkaline-hydrolysing nitrogen fertilizers. BioL
Fert.Soils 20,41-48.
Siccama T G, Bliss M and Vogelmann H W 1982 Decline of red
spruce in the Green Mountains of Vermont. BulL Torrey Bot.
Club 109, 162-168.
Silsbee D G and Larson G L 1982 Water quality of streams in the
Great Smoky Mountain National Park. Hydrobiologia 89, 97-
115.
Silsbee D G and Larson G L 1983 Comparison of streams in logged
and unlogged areas of Great Smoky Mountain National Park.
Hydrobiologia 102,99-111.
Smith M L, Ollinger S V, Martin M E, Aber J D, Hallett R A
and Goodale C L 2002. Direct estimation of aboveground forest
productivity through hyperspectral remote sensing of canopy
nitrogen. EcoL App. (in press).
Soderstrom B, Baath E and Lundgren B 1983 Decrease in soil mi-
crobial activity and biomass owing to nitrogen amendments. Can.
J. MicrobioL 29, 1500-1506.
Stoddard J L 1994 Long-term changes in watershed retention of
nitrogen: its causes and aquatic consequences. IN: Baker, L A
ed Environmental Chemistry of Lakes and Reservoirs American
Chemical Society pp 223-284.
Thron K A and Mikita M A 1992 Ammonium fixation by humic
substances: a nitrogen-15 and carbon-13 study. Science of the
Total Environment 113, 67-87.
Tien M and Myer S B 1990 Selection and characterization of
mutants of Phanerochaete chrysosporium exhibiting ligninolytic
activity under nutrient rich conditions. App. Env. MicrobioL 56,
2540-2544.
Tietema A 1998 Microbial carbon and nitrogen dynamics in coni-
ferous forest floor material collected along a European nitrogen
deposition gradient. For. EcoL Manag. 101,29-36.
Tietema A and Beier C 1995 A correlative evaluation of nitrogen
cycling in the forest ecosystems of the EC projects NITREX and
EXMAN. For. EcoL Manag. 71, 143-152.
Warfvinge P and Sverdrup H 1992 Calculating critical loads of acid
deposition with PROFILE - a steady-state soil chemistry modeL
Wat. Air Soil PolL 63, 119-143.
Weathers K C, Lovett G M, Likens G E and Caraco N F M 2000
Cloudwater inputs of nitrogen to forest ecosystems in southern
Chile: forms, fluxes and sources. Ecosystems 3, 590-595.
Williams M W, Baron J S, Caine N, Sommerfeld R and Sanford Jr R
1996 Nitrogen sturation in the rocky mountains. Env. Sci. Tech.
30, 640--646.
Williams M W and Tonnesson K A 2000 Critical loads for inorganic
nitrogen deposition in the Colorado front range, USA. EcoL App.
10, 1648-1665.
Wright R F and Rasmussen L 1998 Introduction to the NITREX and
EXMAN projects. For. EcoL Manag. 101, 1-8.
Wright R F and van Breeman N 1995 The NITREX project: an
introduction. For. EcoL Manag. 71, 1-6.
Yanai R D, Fahey T J and Miller S L 1995 Efficiency of nutrient ac-
quisition by fine roots and rnycorrhizae. In Resource Physiology
of Conifers Academic Press, New York, pp. 75-103.