The Plant Cell, Vol.
7, 797-807, July 1995 Q 1995 American Society of Plant Physiologists
Regulation of Photosynthesis in C3 and C4 Plants:
A Molecular Appmach
Robert T. Furbank and Wllliam C. Taylor’
Commonwealth Scientific and Industrial Research Organization, Division of Plant Industry, GPO Box 1600, and Cooperative
Research Centre for Plant Science, Australian National University, GPO Box 475, Canberra 2601, Australia
INTRODUCTION
Most plants use the C3 pathway of photosynthesis,also called
the photosyntheticcarbon reduction cycle (PCR), shown in Fig-
ure 1A. C3 plants have a single chloroplast type that performs
all of the reactions that convert light energy into the chem-
ical energy that is used to fix COp and to synthesize the re-
duced carbon compounds upon which all life depends.
Ribulose-i,5-bisphosphate carboxylaseloxygenase (Rubisco)
catalyzes primary carbon fixation, in which a fivecarbon sugar
phosphate, ribulose-l,5-bisphosphate (RuBP), and COp are
converted to two molecules of the threecarbon compound
3-phosphoglycerate (hence the name C3). Phosphoglycerate
is then phosphorylated and reduced by the products of the
light reactions of photosynthesis(ATP and NADPH) to produce
triose phosphate(TP). TP can be exported from the chloroplast
via the chloroplast envelope phosphate (Pi) transporter to the
cytosol and used in the synthesis of sucrose, which is then
translocated throughout the plant (see Sonnewald et al., 1994),
or it can be retained within the chloroplast for starch synthe-
sis or recycling to RuBP. Rubisco also catalyzes the fixation
of Opin a process known as photorespiration, which competes
directly with fixation of COp. At air levels of Coa, for every
three COpmolecules fixed by Rubisco to form 3-phosphoglyc-
erate, approximately one O2 molecule is fixed, producing
Sphosphoglycerate and 3-phosphoglycolate(Figure 1A). Be-
cause 3-phosphoglycolate cannot be used in the PCR cycle,
it must be recycled to phosphoglyceratevia the photorespira-
tory pathway, expending ATP and NADPH. This competition
between 0 2 and COp and the energy costs associated with
recycling phosphoglycolate largely determine the efficiency
of C3 photosynthesis in air (Hatch, 1988; Woodrow and Berry,
1988).
The C4 pathway is a complexadaptationof the C3 pathway
that overcomes the limitationof photorespiration and is found
in a diverse collection of species, many of which grow in hot
climates with sporadic rainfall. The C4 pathway effectively
suppresses photorespirationby elevating the C02concentra-
tion at the site of Rubisco using a biochemical C02 pump. C4
plants have two chloroplast types, each found in a specialized
cell type. Leaves of C4 plants show extensive vascularization,
To whom correspondence should be addressed.
with a ring of bundle sheath (8) cells surrounding each vein
and an outer ring of mesophyll (M) cells surroundingthe bun-
dle sheath. The development of this socalled Kranz anatomy
and the cell-specific compartmentalization of C4 enzymes are
important features of C4 photosynthesis (Hatch, 1988, and
references therein). C02fixation in these plants is a two-step
process. Atmospheric C02 is initially fixed in the cytosol of M
cells by phosphoenolpyruvate carboxylase (PEPC)to form the
four-carbon dicarboxylic acid oxaloacetate (hence the name
C4), which is converted to malate or aspartate (Figure 16).
These C4 acids then diffuse into the inner ring of B cells,
where they are decarboxylated in the chloroplasts. The COp
produced is then refixed by Rubisco. The mechanism of decar-
boxylation in B chloroplasts varies among the three different
C4 types. We confine our discussion to the most extensively
studied type, the NADP-malic enzyme (ME) type, which is
named for its B cell decarboxylating activity.
The key feature of C4 photosynthesis is the compartmen-
talization of activities into two specialized cell and chloroplast
types. Rubisco and the C3 PCR cycle are found in the inner
ring of B cells. These cells are separated from the mesophyll
cells and from the air in the intercellular spaces by a lamella
that is highly resistant to the diffusion of C02 (Hatch, 1988).
Thus, by virtue of this two-stage Coe fixation pathway, the
mesophyll-locatedC4 cycle acts as a biochemical CO2 pump
to increase the concentration of COp in the bundle sheath an
estimated 10-fold over atmospheric concentrations. The net
result is that the oxygenase activity of Rubisco is effectively
suppressed and the PCR cycle operates more efficiently. C4
plants show higher rates of photosynthesis at high light inten-
sities and high temperatures due to the increased efficiency
of the PCR cycle (Hatch, 1988). In favorable environments, C4
plants outperform C3 plants, making them the most produc-
tive crops and the worst weeds. Maize, sugarcane, sorghum,
and amaranth are examples of C4 crops, and nutgrass
(Cyperus mtundus),crabgrass (Dighria sanquinalis),and barn-
yard grass (Echinochloa crusgalli) are some of the worst C4
weeds.
Although the pathways of C3 and C4 photosynthesis are
well established and the propertiesof individualenzymes are
reasonably well understood, there remain severa1 areas of
798 The Plant Cell
A relative ignorance. One concerns the process of assembly of
Cyl b6f
ADP ATP
CHLOROPLAST
STARCH i for the transition from heterotrophic to photoautotrophic growth
SUCROSE
B encoding photosynthetic proteins. High levels of nuclear-
Figure 1. Simplified Schemes of the C3 and C4 Photosynthetic
Pathways.
(A) C3 pathway (photosyntheticcarbon reduction or PCR cycle), em-
phasizingthe enzymes that have been manipulated using molecular
genetic techniques. Carbon initiallyfixed by Rubisco is phosphorylated
and reducedby the productsof the light reactions(ATP and NADPH).
The reduced three-carbon sugar-phosphate (triose phosphate, TP)
can be either exported from the chloroplast for sucrose synthesis via
the chloroplastenvelope Pitransporter(PiTRANS) or retainedfor starch
the Rubisco enzyme. CA, carbonic anhydrase, catalyzes HC03
Coe; RUBSCO, ribulose-l,5-bisphosphate carboxylase-oxygenase,
.-+
synthesis or recycling to ribulose bisphosphate, the COnacceptor for
catalyzes ribulose bisphosphate + COn -, 2X Sphosphoglycerateor
ribulose bisphosphate + O2-,Sphosphoglycerate + phosphoglyco-
-
late; GAPDH, glyceraldehyde Sphosphatedehydrogenase, catalyzes
1,Sdiphosphoglycerate+ NADPH glyceraldehydeSphosphate +
NADP+; FBPase, fructose l,Sbisphosphatase, catalyzes fructose
bisphosphate fructoseSphosphate + Pi; PRK, phosphoribulokinase,
-t
catalyzes ribuloseSphosphate+ ATP- ribulose bisphosphate + ADI?
(B) C4pathway, simplified to describe only the NADP-malic enzyme
type, which transports malatefrom the mesophyllto the bundlesheath
chloroplasts.Compartmentationof enzymes belweenthe two cell types
is shown. COn (in the form of bicarbonate) is fixed by the enzyme
PEPC to form the C4 acid oxaloacetate (OAA), which is reduced by
NADPH from the light reactions to form malate (MAL),the C, acid that
is transported to the bundle sheath cells. Malate is decarboxylated
thelphotosyntheticapparatus during chloroplast development.
How are the synthesis and assembly of these multimeric pro-
tein complexes regulated? Another concerns the ways the net
activity of each pathway is controlled as the plant encounters
a range of environmental conditions each day and through-
out its life. We show how molecular genetic techniques are
contributing to a new understanding of these issues.
HOW GENE REGULATION CONTROLS C3 AND Ca
ENZYME ACTlVlTlES
The rapid assembly of the photosynthetic apparatus is crucial
in newly germinated seedlings. Rapid assembly is made pos-
sible by high levels of the cytosolic and chloroplastic mRNAs
encoded mRNAs result from a combination of environmental
and developmental controls on gene transcription. The small
gene families coding for the small subunit (SSU) of Rubisco
(the RbcS genes) and the chlorophyll alb apoproteins of the
photosystemII light-hawesting complex (the Cab genes, which
have recently been renamed the Lhcb genes) have served as
paradigmsfor these studies. Light, acting through phytochrome
and blue light photoreceptors, induces high rates of transcrip-
tion of these genes (Tobin and Silverthorne, 1985; Thompson
and White, 1991). cis-Acting DNA sequences responsible for
light-induced transcription have been identified by fusing non-
coding regions of the RbcS and Cab genes to reporter genes
and measuring their activities in transgenic plants (Gilmartin
et al., 1990). The signal transduction pathway from the pho-
toreceptors to the transcription factors that activate transcription
of Cab and RbcS genes is currently being dissected by mo-
lecular genetic approaches, primarily in Arabidopsis (Deng,
1994). Light also acts on the transcription of Cab and RbcS
genes by setting a circadian clock that has a major effect on
the transcription of these genes and probably a lesser effect
on the transcription of other nuclear genes (Giuliano et al.,
1988; Nagy et al., 1988; Taylor, 1989a).
Although developmental regulation has been less well stud-
ied than light regulation, its existence is evident from the
restriction of Cab and RbcS gene expression to chlorenchy-
in the bundle sheath cells, where the C02 released is fixed by the
PCR cycle in much the same way as in C3plants. The threecarbon
compound pyruvate (PYR) diffuses back to the mesophyll, where it
is phosphorylated by ATP to regeneratethe carbon acceptor phos-
-
phoenolpyruvate (PEP). PEPC, phosphoenolpyruvatecarboxylase,
-
catalyzes phosphoenolpyruvate+ HC03- oxaloacetate; MDH, ma-
late dehydrogenase. catalyzes oxaloacetate + NADPH malate +
NADP; NADPME, NADP-malic enzyme, catalyzes malate + NADP
-,pyruvate + NADPH; PWK, pyruvate orthophosphate dikinase, cata-
lyzes pyruvate + ATP + Pi -,phosphoenolpyruvate + AMP + PPi.

mous leaf cells in C3 plants (Muller et al., 1980; Eckes et al.,
1985). A signal from the developing chloroplast is required for
high-leve1transcription of Cab and RbcS genes, even when
light signals are fully activated (Oelmuller, 1989; Taylor, 1989b).
This chloroplast signal may be at least partially responsible
for spatial regulation of nuclear genes coding for chloroplast
proteins.
Three main features distinguish C4 plants from C3 plants:
(1) the differentiationof the two cell and chloroplast types, (2)
the presence of an additional set of genes, and (3) a mecha-
nism regulatingthe cell-specificexpression of these additional
genes. Very little is known about how the differentiation of the
two cell types is regulated. The positionof cells relative to de-
veloping veins seems to determine their fate, with those in close
association with the vein becoming B cells and those at least
one cell removed from the vein becoming M cells (Nelson and
Langdale, 1992).
Genes encoding C4 enzymes are, in most cases, members
of small gene families. Other members of each gene family
code for isozymic forms that perform nonphotosynthetic func-
tions; these non-C4 genes are usually expressed at low levels
in a wide range of cells. C4 enzymes are found at high levels
in either B or M cells. Therefore, genes encoding C4 enzymes
must have evolved new programs of gene expression.
Two experimental approaches have been used to study the
mechanisms regulating the cell-specific expression of C4
genes. One is the use of gene and antibody probes to localize
mRNAs and proteins at various stages in leaf development.
These studies have shown that light induces high-leve1expres-
sion of C4 genes, whereas cell-specific expression of some
C4 genes is controlled by a light-independent developmental
program. However, for other genes, light controls both high-
level expression and cell specificity. For instance, Sheen and
Bogorad (1985) found that transcripts for both RbcS and rbcL
(a chloroplast gene that codes for the Rubisco large subunit,
LSU) are present in both B and M cells of etiolatedmaize leaves.
Light is necessary to induce high levels of both transcripts and
to suppress their accumulation in M cells. In contrast, light does
not affect the spatial pattern of C4 mRNA accumulation in
amaranth leaves (Wang et al., 1993). However, RbcS and rbcL
transcripts and proteins can be detected in both cell types at
early stages of development but then become restricted to B
cells (Wang et al., 1992). Accumulation of transcripts encod-
ing C4 enzymes appears to be cell specific in both plants
(Nelson and Langdale, 1989), although Wang et al. (1992)
detected transcripts coding for PEPC in both cell types of
amaranth. However, as they pointed out, it is possible that the
B cell transcripts encode a nonphotosynthetic isoform.
A complementary approach isto identify cis-acting DNA se-
quences responsible for cell-specific expression and to use
these sequences to unravel the mechanismsthat control high-
level, cell-specificexpression. Until recently, the lack of an ef-
ficient, stable transformation system for any C4 plant has
slowed progress using this approach. Therefore, one strategy
has been to study the expression of C4 gene promoters in
Regulation of Photosynthesis 799
transgenic C3 plants to identify sequences responsible for
high-level, organ-specific expression.
F? Westhoff and colleagues have used this strategy to de-
termine how the C4 members of the gene family encoding
PEPC differ from members coding for nonphotosynthetic iso-
forms. Sequence comparisons indicated that in the C4 dicot
Haveria frinervia, these Ppc genes fall into four subfamilies
(Hermans and Westhoff, 1990). The PpcA subfamily codes for
the C4 isoform, and its members are expressed at high lev-
els in M cells. The C3 species E pringleialso has PpcA genes,
but these are expressed at low levels in leaves, roots, and
stems. The PpcA genes of these two species are more similar
to one another than either is to Ppc genes of other subfami-
lies within the same species. Stockhaus et al. (1995) determined
that the E frinervia PpcA gene has unique cis-acting sequences
that are at least partially responsible for its high-leve1expres-
sion in leaves by comparing the expression in transgenic
tobacco of gene fusions in which 5' regions from E frinervia
and E pringlei PpcA genes were fused to the P-glucuronidase
A(gusA) reporter gene. Sequences between -2118 and -500
of the E frinervia PpcA promoter conferred high-leve1reporter
gene expression in tobacco leaves, primarily in the palisade
parenchyma, whereas promoter sequences from the E prin-
g1ei PpcA conferred lower level gusA expression in roots and
stems and very low expression in leaves. Whether the preferem
tia1 expressionof the E frinervia PpcA gene in tobacco palisade
parenchyma is due to M cell-specific control remains to be
seen.
Similar experiments by M. Matsuoka and colleagues re-
vealed that cis-acting sequences from the 5' regions of the
maize RbcS, Pepc, and Ppdk (whichcodes for M cell pyruvate
orthophosphate dikinase, PPdK) genes confer light-regulated
reporter gene expression in transgenic rice plants (Matsuoka
et al., 1993,1994). In all three cases, leaf M cells preferentially
expressed the reporter gene, despite the fact that RbcS is B
cell specific in maize. Although rice leaves have B cells, these
contain only a few chloroplasts, which may account for the rel-
atively greater expression in M cells. Nove1 cis-acting
sequences at the 5' ends of some C4 genes are therefore an
important component of the mechanismthat controls high-leve1
leaf-specific and light-regulated expression of these genes.
Transient expression studies of C4 gene promoter con-
structs introduced into leaf cells of C4 plants have also
provided clues about cell-specificregulation. Schaffner and
Sheen (1992) showed that B'sequences from a maize C4 Aspc
gene that are not present in a closely related n0n-G gene
confer high-leve1light-regulatedexpression in maize leaf pro-
toplasts. Using microprojectile bombardment of maize leaf
sections, Bansal and Bogorad (1993) have identifiedseparate
sites in the upstream region of a maize Cab gene that control
light responses and cell specificity. Although Cab expression
is not strictly cell specific in most C4 plants, this particular
gene is expressed primarily in M cells. This cell preference
appears to be due to the combination of enhanced expres-
sion in M cells and suppression of B cell activity. Transient
800 The Plant Cell
expression experiments (T. Nelson, personal communication)
have shown that 5’sequences of RbcS genes from F: trinervia
(C4) and F: pringlei (C3) specify different expression patterns
in E frherVia leaves. These results imply that, for at least some
C4 genes, the cis-acting sequences controlling cell specific-
ity are located at the 5’end of the gene and are found only
in the genes from C4 species.
To study further the mechanisms controlling C4 gene ex-
pression and the regulation of C4 enzyme activities (see
following sections), we have developed an Agrobacterium-
mediated transformation system for the C4 dicot Flaveria bi-
dentis (Chitty et al., 1994).This system is reasonably fast and
efficient, giving transformed plants 15 to 20 weeks after ex-
plant cocultivation,and it has already providedsome interesting
insights. In E bidentis, two genes code for chloroplastic forms
of ME, one of which, MeA, encodes the C4 form, which is ex-
pressed at high levels and in a light-regulatedfashion in 6 cells
(J.S. Marshall, J.D. Stubbs, and W.C. Taylor, unpublisheddata).
E bidenfis plants stably transformed with a series of 5‘MeA
sequences fused to gusA show very low GUS activity in leaves
but high levels in meristems and moderately high levels in
stems and anthers (J.S. Marshall, J.A. Chitty, J.D. Stubbs, and
W.C. Taylor, unpublished data). The longest of these constructs
had 2.2 kb of 5‘ noncoding sequence, whereas the shortest
had 0.39kb. However, when 5.8 kb of sequence at the 3’end
of MeA was added to the longest S’construct, high-levelleaf
expressionwas attained. Preliminaryanalysis of primary trans-
genic plants suggests that this expression occurs primarily in
6 cells.
In Flaveria species, PPdK is encoded by a single gene, Wk
(E. Rosche and I? Westhoff, personal communication; C.J.
Chastain, M. Matsuoka, and W.C. Taylor, unpublished data).
Pdk encodes two different isoforms, a prevalent chloroplast
form located in M cells and a presumably nonchloroplastform,
the transcript of which is found in all organs at very low levels.
Primary transformants with 1.5kb of Wk 5‘ sequence fused
to gusA show high levels of GUS activity in leaves (W.C. Taylor,
J.A. Chitty, and E. Rosche, unpublishedobservations). Tran-
scription of the mRNA encoding the nonchloroplast form of
PPdK is driven by a promoter located in the large 6-kb intron
of the Wk gene; similar results have been obtained with the
maize Wk gene, which also encodes both a C4 isoform and
a nonchloroplast form (Glackin and Grula, 1990;Sheen, 1991).
These studies show that environmental cues and develop-
mental programs that use positive and negative regulatory
mechanisms control the accumulation of mRNAs coding for
photosynthetic proteins. This regulation affects the timing of
protein synthesis and its cellular localization, but it does not
always directly control the quantities of proteins. Regulation
of protein quantity also occurs at the stage of assembly of mul-
timeric complexes in the chloroplast. For example, when
synthesis of LSU is inhibited, unassembled SSU polypeptides
are rapidly degraded so that stoichiometric amounts of both
Rubiscosubunits accumulate (Schmidt and Mishkind, 1983).
The assembly process is poorly understood, especially its
quantitative aspects. How the quantity of each enzyme is
determined as the C3 and C4 pathways are assembled during
chloroplast development is unknown. Regulation of enzyme
quantity is also important in the mature leaf cell, because part
of the plant’s response to environmental changes can be to
change absolute amounts of enzymes as well as their activities.
METABOLIC ENGINEERING OF PHOTOSYNTHESIS
The response of photosynthetic rate to environmental
parameters has been well characterized for a wide range of
plants (for review, see Woodrow and Berry, 1988).From the
kinetic characteristics of Rubisco (Figure l), it has been
possible to construct comprehensive models of photosynthe-
sis that accurately predict the response of carbon fixation in
C3 plants to particular ambient C02 concentrations (von
Caemmerer and Farquhar, 1981). In the case of C3 plants in
air and saturating light, it is now clear that the photosynthetic
rate is largely, although not entirely, governed by the amount
and kinetic characteristics of Rubisco. The remainingfractional
control of photosynthetic rate is distributed among the other
enzymes, and the degree of limitation by each step presum-
ably varies with environmental conditions, such as light
regimes, in a manner that is difficult to assess (Figure 2). This
kind of regulationoccursviacovalent modification and allosteric
effects on enzyme activity and in the short term (that is, hours)
does not involve changes in the amount of enzyme.
Figure 1 illustratesthe complexity of regulation in the pho-
tosynthetic system, showing potentially regulatory enzymes
in both the C3 and C4 pathways. In C4 plants, determining
which enzymes control flux is further complicated by the ad-
ditional complexity of cell specialization. Superimposed upon
the regulationof the PCR cycle enzymes is regulation of the
enzymes involved in the C02 concentrating mechanism of the
C4 pathway.
From in vitro studies, we know that the activities of a num-
ber of enzymes in the PCR cycle are capable of responding
to changes in light and could potentially limit photosynthetic
flux. This response occurs either indirectly, via changes in the
stromal pH and Mg2+concentration, both of which increase
on illumination, or directly, by reduction/oxidation of the en-
zyme via the thioredoxin system, a signal transduction pathway
responsive to the redox state of photosystem I (Figure 3;
Buchanan, 1991).In the case of Rubisco, activation is medi-
ated by a specific activating protein (Rubisco activase; Salvucci
et al., 1985),which senses chloroplast energy status. Such
a complex set of light-responsive regulatory mechanisms is
necessary for two reasons. First, a crude on/off switch is re-
quired to prevent “futile cycles” occurring between respiratory
and catabolic processes and the photosynthetic pathway, which
share the same biochemical intermediates. Second, the ma-
jor environmental variable that plants experience is short-term
change in light intensity (light flecks, for example, are com-
mon in many closed canopy crops, forests, and grasslands).
It appears that a multienzyme control mechanism has evolved

O 100% 200%
Wild
< TYPe
antisense overexpression
EXTRACTABLE E N N M E ACTlVlTY
Figure2. The Responseof Photosynthetic Carbon Fixationto Changes
in Enzyme Activity.
Enzyme activity can be reduced to below wild-type levels by antisense
suppressionor elevated to above wild-type levels by overexpression.
The effecton photosyntheticflux dependson whether the enzyme (a)
is essentially “nonlimiting” over a range of enzyme activities, (b) has
control over flux but ‘%o-limits”along with the activity of other enzymes,
or (c) is classically “limiting.”
for conservation of metabolites within the PCR cycle and for
intrinsic stability of the photosynthetic system during such
transients, when the photosynthetic flux can change 10-fold
or more in seconds (see Woodrow and Berry, 1988; Geiger
and Servaites, 1994).
Although we have amassed a great deal of information on
the regulatory properties of enzymes in vitro, evidence for their
individual contributions to controlling photosynthetic flux in vivo
is largely circumstantial. This is because it is often difficult
to extrapolate regulatory properties and kinetic characteris-
tics of enzymes in vitro to the cellular environment of the intact
plant. Estimation of inhibitor/activator concentrations and sub-
strate levelswithin cellular compartments in vivo and the effect
of the high protein concentration present in cells can become
insurmountable problems for such studies (see Ashton, 1982).
Recombinant DNA technology and plant genetic transforma-
tion have provided us with excellent tools to get around some
of these obstacles. Using the techniques of gene suppression
and overexpression,it is possible to alter the amount of a single
Regulation of Photosynthesis 801
enzyme in a transgenic plant, thus generating a series of mu-
tants with a range of enzyme activities from below to above
wild-type levels. Depending on the importance of this enzyme
in determining photosynthetic rate, the phenotypic effects of
changes in the leve1of expression may vary widely (see Knight
and Gray, 1992). Figure 2 shows the expected responseof pho-
tosynthetic rate to varying enzyme levels in three hypothetical
cases: (1) in which the enzyme is present in considerable ex-
cess; (2) in which the enzyme is ‘%o-limiting:’ along with other
enzymes in the pathway; and (3)in which the enzyme is
A Thloredoxlnmedlatedenzyme activation
B Phosphorylatlon/dephosphorylatlonregulatlon of PPdK
c Phosphorylatlonldephosphorylatlonregulatlonof PEP carboxylase
Klnase
E -/T*
E-P
“iflactivated“ ATP “Activated
(Mal sensitive) 4 (Mal insensitive)
Phosphatase
Figure 3. Three Examples of Enzyme Regulation by Covalent
Modification.
(A) Regulation by the thioredoxin system. An enzyme is activatedwhen
a disulfide bridge on the protein is reduced by the regulatory protein
thioredoxin-m (ThR-m). Thioredoxin’s redox state is determined by the
redox state of ferredoxin (Fd), the terminal electron acceptor of pho-
tosystem I, and consequently is responsive to the rate of electron
transport and light intensity.
(E) Phosphorylation regulation of PPdK. PDRP, the PPdK regulatory
protein, catalyzes both the dephosphorylation of inactive enzyme (E-P)
to active enzyme (E) and the reverse reaction. The relative rates of
these reactions are believed to respond to the energy status of the
chloroplast.
(C) Phosphorylation regulation of PEPC. In contrast to PPdK, PEPC
is activatedwhen phosphorylated by a protein kinase and becomes
insensitiveto the metabolic inhibitor malate. The “inactivateddephos-
phorylatedform still has activity but is sensitiveto inhibitionby malate.
802 The Plant Cell
“limiting” for photosynthetic flux. A quantitative assessment
of an enzyme’s role in each class can be made using control
analysis (Kacser and Burns, 1973). This mathematical treat-
ment of biochemicalregulationassigns a “control coefficient”
between O and 1, with O indicating no control over flux by the
enzyme and 1 indicating that the enzyme is limiting. Results
from studies on diverse organisms indicatethat control of flux
through a pathway is frequentl9 shared between a number of
steps.
Metabolic engineering can be used to control not only the
quantity of enzyme present in a transgenic plant but also its
“quality.” Both site-directedmutagenesis and expression of het-
erologous enzymes allow the significance of the regulatory
propertiesof an enzyme to be investigated in vivo. Removing
the amino acids responsible for light regulationof an enzyme,
for example, is a powerful tool in assessing the importance
of this property to the plant. The following sections outline some
of the approaches used to alter enzyme quantity and quality;
a unifying theme throughout is the role of individual enzymes
in controlling photosynthetic rate in vivo.
Antisense Suppression of Key Photosynthetic
Enzymes
C3 Pathway
By far the most widely used approach to the genetic manipu-
lation of photosynthesisto date has been the use of antisense
RNA technology to produce transgenic plants with reduced
levels of key photosynthetic enzymes. “Metabolic engineering”
of photosynthesiswas first reported by Rodermel et al. (1988),
who transformedtobacco with a full-length antisense RNA con-
struct targeted to the SSU of Rubisco, using a constitutive
promoter. These transformants showed a substantial reduc-
tion in RbcS transcript levels as well as in Rubisco protein levels
and enzyme activity (Rodermelet al., 1988; Quick et al., 1991).
However, Rubisco activity could be reduced by up to 40%
before even a marginaleffect on photosynthesis would be ob-
served. In contrast, similar experiments by Hudson et al. (1992)
resulted in a range of phenotypes in which both photosynthe-
sis and growth were adversely affected. Careful analysis of
these plants has revealed that when plants are grown at high
light and at atmospheric COn concentrations, Rubisco activ-
ity exerts a high degree of control over photosynthetic carbon
flux and under these conditions can be considered a limiting
enzyme (Hudson et al., 1992). The apparent inconsistencies
between these observations appear to relate mostly to the light
intensities used for growth of transformants, that is, environ-
ment cabinet illumination in the former case (Quick et al., 1991)
and full sunlight in the latter (Hudson et al., 1992). These
transgenic plants have far wider uses than in the study of pho-
tosynthetic flux. For example, because Rubisco is also the
major protein present in leaves, it is extremely important in
the nitrogen relations and nutritive value of higher plants, mak-
ing these transformants valuable in studying nitrogen use and
allocation (Masle et al., 1993). In addition, these plants have
also been used to study the relationshipbetween photosynths
sis and growth and regulation of stomatal aperture (Evans et
al., 1994).
The list of photosynthetic enzymes whose expression has
been reduced in C3 plants is steadily growing, as indicated
in Table 1. In most cases, soluble stromal enzymes of pho-
tosynthetic carbon metabolism have been targeted: Rubisco,
Rubisco activase, fructose-1,Sbisphosphatase (FBPase),
glyceraldehyde-3-phosphatedehydrogenase (GAPDH), phos-
phoribulokinase (PRK), and carbonic anhydrase. More recently,
membraneproteins,such as translocators(chloroplast Pi trans-
locator) and thylakoid membrane proteins (Reiske Fe-S center,
ATPase, and the 10-kD polypeptide of photosystem I I (PSII]),
have also been manipulated. As discussed earlier, the range
of phenotypes observed in these experiments can vary enor-
mously. Many enzymes, such as FBPase (Kosmann et al.,
1994), GAPDH (Price et al., 1995a), and the Pi translocator
(Riesmeier et ai., 1993), seem to fall into the intermediate or
“co-limiting”category. For example, a 70% decrease in FBPase
activity resulted in only a 20% decrease in maximum photosyn-
thetic rate (Kosmann et al., 1994). For some enzymes, such
as the thylakoid PSll 10-kD polypeptide (whose function is un-
known; Stockhaus et al., 1990) and chloroplastic carbonic
anhydrase (Price et al., 1994), little or no phenotype is evident
when their levels are reducedby antisense suppression, sug-
gesting that these proteins fall into the “nonlimiting” category.
The interpretation of the phenotypic effects of an antisense
construct on photosynthesis and growth of the transformed
plants is not always straightforward.The importance of growth
conditions in influencing phenotype should not be underesti-
mated, as demonstrated by the effect of light intensity during
growth on the interpretation of the Rubisco antisense experi-
ments described previously. It also appears that in many
instances, the relationship between steady state mRNA lev-
els and protein levels in transgenic plants expressing antisense
RNA may not be simple. Attempting to reduce levels of the
23- and 33-kD polypeptides of the PSll oxygen-evolving
complex and the Reiske Fe-S center of the cytochrome bdf
complex, Palomares et ai. (1993) recently reported up to a 90%
reduction in target mRNA levels in transgenic plants without
a discernible effect on protein levels or phenotype. In contrast,
Price et al. (1995b) produced slow-growing transgenic tobacco
with reduced Reiske Fe-S and electron transport rates; how-
ever, a 93 to 94% reduction in message level gave only a 60
to 86% reduction in protein. The phenotype was often unsta-
ble and could be ameliorated by high growth irradiance,
suggesting that low mRNA levels are not limiting under some
conditions or that the endogenous sense transcript can
“swamp” out the antisense mRNA at high irradiance. Similar
nonlinear relationshipsof protein levelto transcript level have
also been observed in plants transformed with an antisense
gene targeted to the chloroplast envelope Pi translocator
(Riesmeier et al., 1993). If this proves to be a widespread
phenomenon, it may be difficult in some cases to produce a
series of mutants with a wide range of enzyme activities.
 Regulation of Photosynthesis 803

Table 1. Antisense Suppression of Key Photosynthetic Enzymes in C3 Plants

Plant
Species Target Enzyme Phenotype Reference
Tobacco Rubisco Reduced maximal rate of Quick et al. (1991)
photosynthesis and growth (if
enzyme activity below 60% of
wild-type levels)
Tobacco Rubisco As above, but more severe over the Hudson et al. (1992)
entire range of Rubisco activities
Tobacco Rubisco activase Photosynthesis and growth impaired Mate et al. (1993)
at air C o e levels; requires high
COpfor growth
Tobacco PRK Severe photoinhibition and stunting G.S. Hudson (personal
even at low light communication)
Potato Pi translocator Maximal photosynthesis reduced 40 Riesmeier et al. (1993)
to 60% by 20 to 30% decrease in
translocation; early growth retard-
ed; stores more starch
Tobacco Reiske Fe-S Photosynthesis inhibited and linearly Price et al. (1995b)
related to Fe-S content at high
light; slow growth
Tobacco ATP synthase Photosynthesis inhibited Price et al. (1995b)
Tobacco GAPDH Photosynthesis unaffected until Price et al. (1995a)
GAPDH less than 20% of wild-
type levels
Potato FBPase Maximal photosynthesis impaired Kosmann et al. (1994)
when FBPase below 40% of wild-
type levels; growth impaired when
FBPase below 15% of wild-type
levels
Tobacco Carbonic anhydrase No phenotype with CA below 2% of Price et al. (1994)
wild-type levels; small effect on
COp conductance
Potato 10-kD polypeptide No phenotype even when less than Stockhaus et al. (1990)
3% of wild-type levels are attained

Because it is desirable for biochemical analysis of flux control
to produce such a series (Figure 2; see Woodrow and Berry,
1988), this could limit the usefulness of the antisense tech-
nique for some applications.
C4 Pathway
Use of the antisense approach in studying C4 photosynthe-
sis has so far been impeded by the lack of efficient and reliable
genetic transformation systems for Cq plants, and attempts to
use this approach to study C4 photosynthesis have only re-
cently been reported. The key regulatory enzymes unique to
the C4 pathway are shown in Figure 1B. PPdK, PEPC, and
NADP-dependentmalate dehydrogenase (NADP-MDH) are all
localized in M cells, with PPdK and NADP-MDH located in the
chloroplast and PEPC in the cytosol. The extractable activity
of PPdK is only just sufficient to account for observed rates
of photosynthesis, and the activity of all three enzymes in vivo
is increased markedly by illumination(Hatch, 1988), suggest-
ing that these enzymes are potentially rate limiting.
Rubisco is also a good candidate for the antisense approach
because, in C4 plants, Rubisco levels are already reduced by
up to 50% on a chlorophyll basis as compared with C3 plants.
Less Rubisco is required in these plants due to the C02-con-
centrating function of the pathway, which increases effective
carbon flux through Rubisco (see Hatch, 1988). This reduc-
tion in the amount of Rubisco required improves the nitrogen
use efficiency of C4 plants considerably over their C3 coun-
terparts. It is not known whether a further reduction in Rubisco
activity via antisense RNA suppression could limit flux through
C4 photosynthesis.
So far, antisense constructs targeted to the B cell enzyme
Rubisco and the M cell enzymes PPdK and NADP-MDH have
been transformed into E bidentis. F: bidentis plants transformed
with a full-lengthantisense construct targeted to RbcS mRNA
show a range of Rubisco activities, from 10 to 1000/0 of wild
type (Furbank et al., 1994). mRNA levels were reduced roughly
in proportion to protein levels, as was seen in similar experi-
ments with C3 plants (Rodermel et al., 1988). Plants with
reduced Rubisco levels show a stunted phenotype, with
804 The Plant Cell
proportionally reduced photosynthetic capacity at high light
and over a range of C02 concentrations. Interestingly, there
appears to be a strong regulatory mechanism linking Rubisco
activity to the rate of the mesophyll C4cycle. Although the lev-
els of mesophyllenzymes were unaffectedin the transformants,
down-regulation of the mesophyll C4 cycle turnover was ob-
served in intact leaves, presumably via down-regulation of
enzyme activity. The biochemical basis for such a mechanism
is currentlyunknown, and further physiological and biochem-
ical analysis of the progeny of these transformants is under
way. These plants will be a valuable tool for understanding
photosynthetic regulationin C4 plants because, unlike the CB
case, the role of Rubisco activity in determining photosynthetic
rate in C4 plants under varying environmental conditions has
not been extensively modeled.
Analysis of F: bidentis transformants containing other pho-
tosynthetic antisense constructs is still in its early stages.
NADP-MDH has been reduced by as much as 60°/o from wild-
type levels (S.J. Trevanion, R.T. Furbank, and A.R. Ashton,
unpublishedobservations); in these plants, there is a commen-
surate reduction in the light-saturated rate of photosynthesis
but only a slight effect on growth.
In the case of PPdK, antisense transformants show up to
a 90% reduction in both enzyme activity and transcript levels
(Furbank et al., 1994).Transformants containing a full-length
antisense construct fall into two groups: in one, enzyme ac-
tivities range from 10to 20% of wild type; in the other, enzyme
activities are normal. No transformants showed an intermedi-
ate level of PPdK, although some plants transformed with an
0.8-kb antisense construct correspondingto the 5’ end of the
Wk cDNA show a reduction in Wk transcript levels but no
reduction in extractable enzyme activity. The phenotype of
plants with 20% or less PPdK activity is severe. Regenerants
are incapable of photo-autotrophic growth in air and can be
maintained only on exogenous sucrose or in an atmosphere
containing lO/o COn. Plants deprived of exogenous sucrose
show rapid and severe photoinhibitionin air, even at very low
light intensity. One explanation for the all-or-none nature of
the phenotype in these experiments could be that the rela-
tionship between transcript level and protein level is nonlinear
or possibly even sigmoidal. Thus, when the transcript level
drops below a critical threshold, a catastrophic falloff in pro-
tein levels may occur.
I
Site-Directed Mutagenesis and “Overexpression”
High-leve1expression of a native or heterologous protein to
increase flux through a pathway, divert flux, or change the
regulatory properties of an enzyme is a powerful tool for un-
derstanding regulation of photosynthesis. So far, these
techniques have been more commonly applied to the path-
ways of sucrose and starch biosynthesisthan to photosynthesis
per se (Sonnewald et al., 1994;Stitt, 1994).A classic example
of this approach is the expression of a bacterial ADP-glucose
pyrophosphorylasein plants to “short circuit” regulation of this
step and increase starch accumulation (see Martin and Smith,
1995,this issue). Site-directed mutagenesisto alter the kinetic
properties of Rubisco has been moderately successful in vitro;
however, the heteropolymericnature of the protein and the fact
that LSU is chloroplast encoded have prevented expression
of the mutated holoenzyme in vivo. It is also generally accepted
that overexpression of genes in plants is more difficult than
antisense suppression due to cosuppression (Napoli et al.,
1990;van der Krol et ai., 1990)and the possible inactivation
of the overexpressed protein by endogenous regulatory sys-
tems (for example, Sonnewald et al., 1994).
In the past two decades, it has become apparent that a ma-
jor form of enzyme regulation involves covalent modification
of specific amino acid residues either by the formation of di-
sulfide bridges or by protein kinase-mediated phosphorylation
(Figure 3). This type of regulation in plants responds directly
to light (via the thioredoxinsystem) or to cellular energy reserves
such as ATI? With sequences of more photosyntheticenzymes
appearing in genetic databases almost daily, this area of en-
zyme regulation is particularly amenable to study by expression
of recombinant protein both in vitro and in vivo. For example,
NADP-MDH is a light-activated enzyme in the C4 pathway
(Figure 1)that undergoes reductive activation in the light via
photosynthetic electron transfer and the thioredoxin system
(Buchanan, 1991).Reductionof a disulfide bridge between two
cysteine residues in NADP-MDH by reduced thioredoxin
activates the enzyme (Figure 3).NADP-MDH protein is homol-
ogous to the NAD-dependent nonphotosynthetic form of the
enzyme, but it also has C- and N-terminal extensions that
have been implicated in conferring the unique regulatory
properties of the photosynthetic enzyme. Using site-directed
mutagenesis, the pairs of cysteine residues responsible for
redox activation at both the N and C termini of the sorghum
enzyme have been identified (see lssakidis et al., 1994).Mu-
tant forms of this enzyme that are not inactivated by oxidation
and thus should not be inactivated in the dark have been pro-
duced. Flaveria and tobacco have been transformed with these
constructs to examine the role of light activationof this enzyme
in vivo (S.J. Trevanion, A.R. Ashton, and I. Issakidis, personal
communication).
Two other key enzymes in C4 photosynthesis are regulated
by covalent modification:PPdK and PEPC (Figures 1 and 3).
PEPC, the primary CO,-fixing enzyme of the C4 pathway, is
“activated in vivo when phosphorylatedby a specific protein
kinase (for review, see Budde and Chollet, 1988). This activa-
tion does not result in an increase in the V, of the enzyme
but manifests itself by a decrease in sensitivityto the inhibitor
malate. In the sorghum protein, phosphorylationof a serine
residue is responsible for this activation. Substitution of this
serine in a recombinant enzyme expressed in Escherichia coli
prevents this activation (Duff et ai., 1993).This region of the
amino acid sequence appears to be highly conserved among
PEPC enzymes from many sources, suggesting that this

regulatory mechanism may be universal. So far, no transfor-
mation experiments have been performed with a recombinant
enzyme in higher plants; therefore, it is not possibleto assess
the importance of this regulatory mechanism in vivo. The sig-
na1 transduction chain that controls the activity of the protein
serine kinase in responseto environmentalconditions remains
somewhat of a mystery, as does the identity of the kinase itself.
PPdK, which is responsible for regenerating PEP, the ac-
ceptor for COp fixation in the mesophyll chloroplast, has long
been recognized as a light-activated enzyme (Figure 1;
reviewed in Hatch, 1988). It is now known that, in contrast to
PEPC, this enzyme is inactivated by phosphorylationof a threo-
nine residue (Ashton and Hatch, 1983; Ashton et al., 1984).
This residue is not the site that is phosphorylatedduring catal-
ysis, which is a nearby histidine residue. The enzyme that
catalyzes regulatory phosphorylation of PPdK, the PPdK
regulatory protein, is unusual in that it uses ADP rather than
ATP as the phosphate donor. In addition, this same enzyme
catalyzes the removal of the phosphate group, reactivating
PPdK by phosphorolysis rather than by hydrolysis (Figure 3;
Ashton et ai., 1984). It is not clear how the activity of the PPdK
regulatory protein itself is controlled, although high pyruvate
levels appear to block inactivation (Burnell et ai., 1986). The
regulatory protein is not abundant in leaves, and purification
to homogeneity has proven elusive (Smith et al., 1994), as has
the isolation of its gene.
The gene coding for PPdK has now been cloned from a num-
ber of species (see previous discussion). Preliminary work has
been done expressing the native enzyme from maize in E. coh
(A.R. Ashton and R.T. Furbank, unpublishedobservations)and
from representativesof the Haveria genus (J.N. Burnell, per-
sonal communication). By mutagenesis of the regulatory
threonine and expression of the recombinant protein in E. Coli
and higher plants, it should be possible to examine the regula-
tory significance of phosphorylation.
One of the greatest challenges in the area of photosynthe-
sis is to use the large body of information available in the
literature on the enzymes of the pathway and their regulation
to understand and perhaps improve photosynthetic perfor-
mance in whole plants. As discussed earlier, the major barrier
to this research has been the difficulty of performing biochem-
ical manipulations in vivo in a precise and interpretable fashion.
Early indications suggest that metabolic engineering, coupled
with more traditional biochemical and physiological ap-
proaches, may provide the means to achieve this aim.
ACKNOWLEDGMENTS
We thank our colleagues who communicated work in progress and
apologize to those whose contributions to the field were omitted due
to space limitations. We are grateful to Tony Ashton for helpful com-
ments on the manuscript.
Regulation of Photosynthesis 805
REFERENCES
Ashton, A.R. (1982).A role for ribulose-l,5-bisphosphatecarboxylase
as a metabolite buffer. FEBS Lett. 145, 1-7.
Ashton, A.R., and Hatch, M.D. (1983).Regulation of C4photosynthe-
sis: Regulation of pyruvate, Pi dikinase by ADP-dependent
phosphorylation and dephosphorylation. Biochem. Biophys. Res.
Commun. 115, 53-60.
Ashton, A.R., Burnell, J.N., and Hatch, M.D. (1984).Regulation of
C, photosynthesis: lnactivation of pyruvate, Pi dikinase by ADP-
dependent phosphorylation and activation by phosphorolysis. Arch.
Biochem. Biophys. 230,492-503.
Bansal, K.C., and Bogorad, L. (1993).Cell type-preferred expres-
sion of maize cab-m: Repression in bundle sheath cells and
enhancement in mesophyll cells. Proc. Natl. Acad. Sci. USA 90,
4057-4061.
Buchanan, 6.6. (1991).Regulation of COn assimilation in oxygenic
photosynthesis: The ferredoxinhhioredoxin system. Arch. Biochem.
Biophys. 288, 1-8.
Budde, R.J.A., and Chollet, R. (1988).Regulation of enzyme activity
in plants by reversible phosphorylation. Physiol. Plant. 72,435-439.
Burnell, J.N., Jenklns, C.L.D., and Hatch, M.D. (1986).Regulation
of C4photosynthesis: A role for pyruvate in regulating pyruvate, Pi
dikinase activity in vivo. Aust. J. Plant Physiol. 13, 203-210.
Chitty, J.A., Furbank, R.T., Marshall, J.S., Chen, Z., andTaylor, W.C.
(1994).Genetic transformation of the C, plant, flavefia bidentis.
Plant J. 6,949-956.
Deng, X.-W. (1994).Fresh view of light signal transduction in plants.
Cell 76,423-426.
Duff, S.M.G., Leplniec, L., Cretln, C., Andreo, C., Condon, S.A.,
Sarath, G., Vldal, J., Gadal, P., and Chollet, R. (1993).An en-
gineered change in the L-malate sensitivity of a site directed mutant
of sorghum phosphoenolpyruvate carboxylase: The effect of sequen-
tia1 mutagenesis and S-carboxymethylation at position 8. Arch.
Biochem. Biophys. 306, 272-276.
Eckes, P., Schell, J., and Willmitzer, L. (1985).Organ-specific ex-
pression of three leafktem specific cDNAs from potato is regulated
by light and correlated with chloroplast development. MOI.Gen. Ge-
net. 199, 216-221.
Evans, J.R., von Caemmerer, S., Setchell, B.A., and Hudson, G.S.
(1994).The relationship between COntransfer conductance and leaf
anatomy in transgenic tobacco with a reduced content of Rubisco.
Aust. J. Plant Physiol. 21, 475-495.
Furbank, R.T., Chltty, J.A., Taylor, W.C., and Jenkins, C.L.D. (1994).
Genetic manipulation of C4 photosynthesis: Antisense of key photc-
synthetic enzymes in Flaveria bidentis. Plant Physiol. 105, 23.
Gelger, D.R., and Servaites, J.C. (1994).Dynamics of self-regulation
of photosynthetic carbon metabolism. Plant Physiol. Biochem. 32,
173-183.
Gilmartin, P.M., Sarokin, L., Memellnk, J., and Chua, N.-H. (1990).
Molecular light switches for plant genes. Plant Cell 2, 369-378.
Giuliano, G., Hoffman, N.E., Ko, K., Scolnik, P.A., andcashmore,
A.R. (1988).A light-entrained circadian clock controls transcription
of severa1 plant genes. EMBO J. 7, 3635-3642.
Glackin, C.A., and Grula, J.W. (1990).Organ-specific transcripts of
different size and abundance derive from the same pyruvate
806 The Plant Cell
orthophosphate dikinase gene in maize. Proc. Natl. Acad. Sci. USA
67,3004-3008.
Hatch, M.D. (1988). C4 photosynthesis: A unique blend of modified
biochemistry, anatomy and ultrastructure. Biochim. Biophys. Acta
895, 81-106.
Hermans, J., and Westhoff, P. (1990). Analysis of expression and evolu-
tionary relationships of phosphoenolpyruvate carboxylase genes
in Flaveria trinervia (C,) and I? pringlei (C3).
MOI.Gen. Genet. 224,
459-468.
Hudson, G.S., Evans, J.R., von Caemmerer, S., Anrldsson, Y.B.C.,
and Andrews, T.J. (1992). Reduction of ribulose-1.5-bisphosphate
carboxylaseloxygenase content by antisense RNA reduces pho-
tosynthesis in transgenic tobacco plants. Plant Physiol. 98,294-302.
Issakidis, E., Saarlnen, M., Decottlgnles, P., Jacquot, J.-P., Cretin,
C, Gadal, P., and Mlglnlac-Maslow, M. (1994). ldentification and
characterizationof the second regulatory disulphide bridge of recom-
binant sorghum leaf NADP-malatedehydrogenase. J. Biol. Chem.
269, 3511-3517.
Kacser, H., and Burns, J.A. (1973). The control of flux. Symp. SOC.
Exp. Biol. 27, 65-104.
Knight, J.S., and Gray, J.C. (1992). Transgenic plants as a tool for
the study of photosynthetic carbon assimilation. In Trends in Pho-
tosynthesis Research, J. Barber, M.G. Guerrero, and H. Medrano,
eds (Andover, MA: Intercept), pp. 267-277.
Kosmann, J., Sonnewald, U., and Wlllmltzer, L. (1994). Reduction
of the chloroplast fructose-1,I-bisphosphatase in transgenic potato
plants impairs photosynthesis and plant growth. Plant J. 6,637-650.
Martln, C., and Smlth, A.M. (1995). Starch biosynthesis. Plant Cell
7, 971-985.
Masle, J., Hudson, G.S., and Badger, M.R. (1993). Effects of am-
bient COp concentration on growth and nitrogen use in tobacco
(Nicotiana tabacum) plants transformed with an antisense gene to
the small subunit of ribulose-l,5-bisphosphatecarboxylaseloxygen-
ase. Plant Physiol. 103, 1075-1088.
Mate, C.J., Hudson, G.S., von Caemmerer, S., Evans, J.R., and
Andrews, T.J. (1993). Reduction of ribulose bisphosphate carboxy-
lase activase levels in tobacco (Nicotiana tabacum) by antisense
RNA reduces ribulose bisphosphate carboxylase carbamylation and
impairs photosynthesis. Plant Physiol. 102, 1119-1128.
Matsuoka, M., Tada, Y., Fujlmura, T., and Kano-Murakami,Y. (1993).
Tissue-specific light-regulated expression directed by the promoter
of a C4 gene, maize pyruvate, orthophosphate dikinase, in a C3
plant, rice. Proc. Natl. Acad. Sci. USA 90, 9586-9590.
Matsuoka, M., Kyozuka, J., Shlmamoto, K., and Kano-Murakami,
Y. (1994). The promoters of two carboxylases in a C4 plant (maize)
direct cell-specific, light-regulated expression in a C3plant (rice).
Plant J. 6, 311-319.
Muller, M., Viro, M., Balke, C., and Kloppstech, K. (1980). Poly-
adenylated mRNA for the light-hawesting chlorophyll alb protein:
Its presence in green and absence in chloroplast-free plant cells.
Planta 148, 444-447.
Nagy, F., Kay, S.A., and Chua, N.-H. (1988). A circadian clock regu-
lates transcription of the wheat Ceb-7 gene. Genes Dev. 2,376-382.
Napoli, C., Lemleux, C., and Jorgensen, R. (1990). lntroduction of
a chimeric chalcone synthase gene into petunia results in revers-
ible co-suppression of homologous genes in trens. Plant Cell 2,
279-289.
Nelson, T., and Langdale, J.A. (1989). Patterns of leaf development
in C4 plants. Plant Cell 1, 3-13.
Nelson, T., and Langdale, J.A. (1992). Developmentalgenetics of C,
photosynthesis.Annu. Rev. Plant Physiol. Plant MOI.Biol. 43,2547.
Oelmuller, R. (1989). Photooxidative destruction of chloroplasts and
its effect on nuclear gene expression and extraplastidic enzyme lev-
els. Photochem. Photobiol. 49, 229-239.
Palomares, R., Herrman, R.G., and Oelmuller, R. (1993). Antisense
RNA for components associated with the oxygen evolving complex
and the Rieske ironlsulfur protein of the tobacco thylakoid mem-
brane suppresses accumulation of mRNA but not protein. Planta
190, 305-312.
Prlce, G.D., von Caemmerer, S., Evans, J.R., Yu, J.-W., Lloyd, J.,
Oja, V., Kell, P., Harrlson, K., Gallagher, A., and Badger, M.R.
(1994). Specific reduction of the chloroplast carbonic anhydrase ac-
tivity by antisense RNA in transgenic tobacco plants has a minor
effect on photosynthetic COp assimilation. Planta 193, 331-340.
Prlce, G.D., Evans, J.R., von Caemmerer, S., RI, J.-W., and Badger,
M.R. (1995a). Specific reduction of chloroplast glyceraldehyde-3-
phosphate dehydrogenase activity by antisense RNA reduces C02
assimilation via a reduction in RuBP regeneration in transgenic
tobacco plants. Planta 195, 369-378.
Prlce, G.D., Yu, J.-W., von Caemmerer, S., Evans, J.R., Chow, W.S.,
Anderson, J.M., Hurry, V., and Badger, M.R. (1995b). Studying
the central roles of the chloroplast cytochrome B6/f and the ATP
synthase complexes in transgenic tobacco: Transformationwith an-
tisense RNAdirected against the Reiske FeS and the ATPd nuclear
encoded polypeptides. Aust. J. Plant Physiol., in press.
Quick, W.P., Schurr, U., Schelbe, R., Schultze, E.-D., Rodermel,
S.R., Bogorad, L., and Stltt, M. (1991). Decreased ribulose-1s-bis-
phosphate carboxylaseoxygenasein transgenic tobacco transformed
with “antisense” rbcS. I. lmpact on photosynthesis in ambient growth
conditions. Planta 183, 542-554.
Riesmeier, J.W., Flugge, U.4, Schulz, B., Heineke, D., Heldt, H.W.,
Wlllmltzer, L., and Frommer, W.B. (1993). Antisense repression
of the chloroplast triose phosphate translocator affects carbon par-
titioning in transgenic potato plants. Proc. Natl. Acad. Sci. USA 90,
6160-6164.
Rodermel, S.R., Abbott, M.S., and Bogorad, L. (1988). Nuclear-
organelle interactions: Nuclear antisense gene inhibits ribulose
bisphosphate carboxylase enzyme levels in transformed tobacco
plants. Cell 55, 673-684.
Salvuccl, M.E., Portis, A.R., Jr., and Ogren, W.L. (1985). Asoluble
chloroplast protein catalyzes ribulose-bisphosphate carboxylasel
oxygenase activation in vivo. Photosynthesis Res. 7, 193-201.
Schaffner, A.R., and Sheen, J. (1992). Maize C4 photosynthesis in-
volves differential regulation of phosphoenolpyruvate carboxylase
genes. Plant J. 2, 221-232.
Schmidt, G.W., and Mlshklnd, M.L. (1983). Rapid degradation of un-
assembled ribulose-l,5-bisphosphate carboxylase small subunits
in chloroplasts. Proc. Natl. Acad. Sci. USA 80, 2632-2636.
Sheen, J. (1991). Molecular mechanisms underlying the differential
expression of maize pyruvate, orthophosphate dikinase genes. Plant
Cell 3, 225-245.
Sheen, JrY, and Bogorad, L. (1985). Differential expression of the
ribulose bisphosphate carboxylase large subunit gene in bundle
sheath and mesophyll cells of developing maize leaves is influenced
by light. Plant Physiol. 79, 1072-1076.

Smlth, C.M., Duff, S.M.G., and Chollet, R. (1994). Partia1 purifica-
tion and characterization of maize-leaf pyruvate, Pi dikinase
regulatory protein: A low abundance mesophyll chloroplast stromal
protein. Arch. Biochem. Biophys. 308, 200-206.
Sonnewald, U., Lerchl, J., Zrenner, I?.,and Frommer, W. (1994).
Manipulation of sink-source relations in transgenic plants. Plant Cell
Environ. 17, 649-658.
Stltt, M. (1994). Manipulation of carbohydrate partitioning. Curr. Opin.
Biotechnol. 5, 137-143.
Stockhaus, J., Hofer, M., Renger, G., Westoff, P., Wydrzynski, T.,
and Willmltzer, L. (1990). Anti-sense RNA efficiently inhibits for-
mation of the 10 kd polypeptide of photosystem II in transgenic potato
plants:Analysis of the role of the 10 kd protein. EM60 J. 9,3013-3021.
Stockhaus,J., Poetsch, W,, Stelnmuller, K., and Westhoff, P.(1995).
Evolution of the C
,. phosphoenolpyruvatecarboxylase promoter of
the C4dicot F/averia frinervia:An expression analysis in the C3plant
tobacco. MOI. Gen. Genet. 245, 286-293.
Taylor, W.C. (1989a). Transcriptional regulation by a circadian rhythm.
Plant Cell 1, 259-264.
Taylor, W.C. (1989b). Regulatory interactions between nuclear and
plastid genomes. Annu. Rev. Plant Physiol. Plant MOI. Biol. 40,
211-233.
Regulation of Photosynthesis 807
Thompson, W.F., and Whlte, M.J. (1991). Physiological and molecu-
lar studies of light-regulated nuclear genes in higher plants. Annu.
Rev. Plant Physiol. Plant MOI. Biol. 42, 423-466.
Tobln, E.M., and Sllverthorne, J. (1985). Light regulation of gene ex-
pression in higher plants. Annu. Rev. Plant Physiol. 36, 569-593.
van der Krol, A.R., Yur, L.A., Beld, M., MOI, J.N.M., and Stultje,
A.R. (1990). Flavonoid genes in petunia: Addition of a limited num-
ber of gene copies may lead to a suppression of gene expression.
Plant Cell 2, 291-299.
von Caemmerer, S.,and Faquhar, G.D. (1981). Some relationships
between the biochemistry of photosynthesis and the gas exchange
of leaves. Planta 153, 376-387.
Wang, J.-L., Klesslg, D.F., and Berry, J.O. (1992). Regulation of C4
gene expression in developing amaranth leaves. Plant Cell 4,
173-184.
Wang, J.-L., Long, J.J., Hotchklss, T., and Berry, J.O. (1993). C4
photosynthetic gene expression in light- and dark-grown amaranth
cotyledons. Plant Physiol. 102, 1085-1093.
Woodrow, I.E., and Berry, J.A. (1988). Enzymatic regulationof pho-
tosynthetic COpfixation in C3plants. Annu. Rev. Plant Physiol. Plant
MOI. Biol. 39, 533-594.