You are on page 1of 10

Ovule Abortion in Arabidopsis Triggered by Stress1

Kelian Sun, Kimberly Hunt2, and Bernard A. Hauser*

Department of Botany, University of Florida, Gainesville, Florida 326118526

Environmental stresses frequently decrease plant fertility. In Arabidopsis, the effect of salt stress on reproduction was
examined using plants grown in hydroponic medium. Salt stress inhibited microsporogenesis and stamen filament elongation.
Because plants grown in hydroponic media can be rapidly and transiently stressed, the minimum inductive treatment to cause
ovule abortion could be determined. Nearly 90% of the ovules aborted when roots were incubated for 12 h in a hydroponic
medium supplemented with 200 mM NaCl. The anatomical effects of salt stress on maternal organs were distinct from those in
the gametophyte. A fraction of cells in the chalaza and integuments underwent DNA fragmentation and programmed cell
death. While three-fourths of the gametophytes aborted prior to fertilization, DNA fragmentation was not detected in these
cells. Those gametophytes that survived were fertilized and formed embryos. However, very few of these developing embryos
formed seeds; most senesced during seed development. Thus, during seed formation, there were multiple points where stress
could prematurely terminate plant reproduction. These decreases in fecundity are discussed with respect to the hypothesis of
serial adjustment of maternal investment.

As a result of abiotic stresses, crop yields can be 4- to

20-fold less than yields under optimal growth conditions (Boyer, 1982). These abiotic stresses include
water availability, temperature extremes, nutrient levels, fluence rate, soil salinity, and pollution. Among
these stresses, soil salinity is an ancient problem that
had existed long before the advent of agriculture.
Intensive cultivation in areas that receive irrigation can
and has exacerbated the problem of salinity because
ions in irrigation water become concentrated in soils.
Today, around 20% of the cultivated land and almost
50% of the irrigated land on earth is affected by ion
accumulation (Zhu, 2001). Consequently, salinity is
one of the most severe stresses affecting crop yields.
Breeding for salt tolerance has been unsuccessful
because salt tolerance is a complex trait regulated by
many loci (Quesada et al., 2002). At present, there is no
genetic line in crop plants that is both salt tolerant and
high yielding.
High salinity has two major deleterious effects on
plants. The first is caused by water deficit, which results
from increased solute concentration or osmoticum. The
other is ion toxicity, which inhibits enzymatic function
of key biological processes (Zhang and Blumwald,
2001). Along with these primary effects, secondary
stresses, such as oxidative damage, occur because high
concentrations of ions disrupt cellular homeostasis
(Dat et al., 2000). Significant progress has been made
This work was supported by the U.S. Department of Agriculture
Competitive Grants Program (grant no. 20023510012109). K.S. was
supported by a University of Florida Alumni Fellowship.
Present address: Department of Genetics, University of Georgia,
Athens, GA 30602.
* Corresponding author; e-mail, fax
Article, publication date, and citation information can be found at


in understanding the three classes of plant adaptive

responses: homeostasis, detoxification, and regulation
of growth (Zhu, 2002). An example of adaptive detoxification is the transport of excess sodium ions into
the vacuoles through the operation of vacuolar Na1/
H1 antiporters (Apse et al., 1999). Transgenic tomatoes
that overexpressed a Na1/H1 antiporter sequestered
sodium ions in the vacuole, allowing these transgenic plants to flourish, flower, and produce fruit when
grown in 200 mM sodium chloride (Zhang and
Blumwald, 2001). Enzymes, such as Cu/Zn-superoxide
dismutase, glutathione peroxidase, and catalase, actively scavenge and neutralize reactive oxygen species
(Gueta-Dahan et al., 1997). In stressed plants, increased
production of free-radical scavengers and antioxidants
helps prevent oxidative damage to organelles and
enzymes. Salt stress also dramatically reduces the rate
of photosynthesis (Kawasaki et al., 2001), which, in
turn, inhibits the growth rate. In addition, salt stress can
inhibit cell division and expansion, directly affecting
growth (Zhu, 2001).
In response to these deleterious effects, plants have
evolved multiple mechanisms to overcome environmental stresses. One of these mechanisms, a decrease in
plant fecundity, involves aborting ovules and/or pollen, then shunting resources from reproductive activities into metabolic reactions that increase stress
tolerance. Salt stress also can induce or accelerate senescence of the reproductive organs. For example, in a
study by Asch and Wopereis (2001), salinity reduced
the yield of rice approximately 45% as a result of spikelet sterility and, in surviving seed, a reduction in seed
weight. In field-grown cotton, salt stress was a major
cause of seed abortion, reducing both yield and quality
(Davidonis et al., 2000).
While Arabidopsis has been investigated for salt
tolerance (Bohnert et al., 2001), the effects of this stress
on ovule development have not been extensively

Plant Physiology, August 2004, Vol. 135, pp. 23582367, 2004 American Society of Plant Biologists

Ovule Abortion in Arabidopsis

studied in this model species. The development of

ovules in wild-type Arabidopsis has been well documented (Robinson-Beers et al., 1992; Schneitz et al.,
1995). Arabidopsis ovules contain four distinct regions: the nucellus, inner integument, outer integument, and funiculus. The funiculus attaches the ovule
to the placenta. The inner and outer integuments envelope the gametophyte and assist in the transfer of
nutrients to this structure. The nucellus is the site of
gametophyte development. Similar to most other angiosperms, the gametophyte contains seven cells; the
egg cell, two synergid cells, a diploid central cell, and
three antipodal cells. After fertilization, the endosperm nuclei proliferate and facilitate the transport
of nutrients from maternal organs (chalaza and inner
integument) into the embryo axis. During this phase of
embryogenesis, the embryo grows slowly. However,
once endosperm replication is complete, the embryo
develops rapidly, going through globular, heart, and
torpedo stages. After the torpedo stage, the embryo
undergoes rapid cell expansion and crushes the endosperm, and the nutrients from the endosperm are
subsequently absorbed by the embryo (Mansfield and
Briarty, 1992).
Arabidopsis is an excellent system in which to study
plant cell death (PCD) because it is eminently suited to
analysis using molecular tools now available to researchers. However, the anatomy of PCD in ovules
needs to be described before these tools can be utilized.
Evidence presented here reveals that Arabidopsis
ovules abort when exposed to environmental stress.
While abortion can occur prior to or after fertilization,
the anatomy and physiology of PCD differs according to the developmental stage prior to abortion. Consequently, the anatomical details of ovule abortion and
embryo senescence in stressed Arabidopsis plants are

Salt Stress Reduces Seed Production

Arabidopsis plants were transiently salt stressed to

measure the effects this has on plant reproduction.
Hydroponic growth allowed the nutrient status and
solute concentrations to be precisely controlled and
quickly adjusted. In Arabidopsis plants, the addition
of 200 mM NaCl to the hydroponic solution induced
numerous symptoms, including transient wilting, reduced growth, and decreased fertility. On a macroscopic scale, decreased fecundity was noticeable when
fruit length declined (Fig. 1), 7 to 10 d after plants were
stressed. Since Arabidopsis seed development produces signals that regulate fruit expansion (VivianSmith et al., 2001), fruits containing predominantly
aborted ovules and embryos were narrower and shorter
than healthy ones (Fig. 1).
While seeds nearing the end of their development
were unaffected by stress, developing ovules and
Plant Physiol. Vol. 135, 2004

Figure 1. Stress reduces Arabidopsis fruit size. A, The fruits of a healthy

Arabidopsis plant were plump and uniform in size. B, Roots were
stressed in a hydroponic solution supplemented with 200 mM NaCl for
12 h, then returned to the normal hydroponic solution. Three weeks
after this plant was stressed, the fruit size was compared to healthy
controls. The pistils that were developing when the plant was salt
stressed were thinner and shorter (arrows).

young embryos often aborted if plants were stressed.

To determine the minimum inductive time for seed
abortion, sets of plants were transiently stressed for
increasing lengths of time. The number of seeds per
fruit was counted and plotted as a function of the
duration that plants were stressed (Fig. 2). Four hours
of stress resulted in a small, statistically insignificant
reduction in fertility. After 8 h of stress, however,
plants exhibited significant decreases in fecundity
(P 5 0.0089). In samples that were stressed longer than
12 h, maximal decreases in reproduction were observed. In these plants, only 5% of the ovules formed
seeds (Fig. 2). Thus, for plants grown in standard nutrient conditions, a 12-h exposure to these saline conditions was sufficient to inhibit 90% of the ovules from
developing into seeds. Seeds that developed in stressed
plants were planted and proved viable.
The addition of 200 mM NaCl to the hydroponic
medium increased both the osmoticum and the concentration of Na1 and Cl2 ions. The reduction in plant
fertility might be caused either by increases in ion concentration or by decreases in water potential. It seems
unlikely that Na1 ions would be transported to ovules
and reach toxic levels within 12 h because system-wide
symptoms of ionic stress were not observed over the
course of these experiments. To experimentally determine if increases in osmoticum alone would reduce
fertility, mannitol, a nontoxic sugar that does not ionize
in solution, was added to the hydroponic medium. The
addition of equivalent osmomoles of mannitol caused

Sun et al.

Figure 2. The rate of ovule abortion is proportional to the duration of

salt stress. The number of seeds produced per fruit was plotted
as a function of the duration of plant stress. Plants were stressed
by supplementing hydroponic solution with either 200 mM NaCl or
400 mM mannitol. Three days after plants were stressed, fruits were
dissected and examined using a stereomicroscope. Aborted ovules
appeared small and desiccated, while developing ovules remained
plump (inset). Even in healthy plants (0 h stress), several ovules failed to
develop into seeds. When plants are salt or mannitol stressed, however,
many more ovules aborted. Nearly 90% of the ovules aborted if plants
were stressed for 12 h or longer. All of the stressed plants, including the
ones stressed for 24 h, made a few seeds per fruit. Average fecundity is
plotted 1mannitol or 2NaCl 1 SD.

similar decreases in fertility (Fig. 2), thus ruling out

ionic effects as the cause of reduced fertility.
The addition of salt to the hydroponic medium decreased the water potential of the hydroponic media
and indirectly altered the water potential of the roots. A
pressure chamber was used to determine the rate at
which water potential changed in reproductive organs.
The water potential of Arabidopsis inflorescences was
measured before, during, and after stress (Fig. 3). The
initial water potential of inflorescences was 20.5 MPa.
After plants were salt stressed, the inflorescence
water potential rapidly dropped and equilibrated at
20.9 MPa. After the saline solution was replaced with
fresh hydroponic solution, the average inflorescence
water potential reached baseline levels within 3 h.
These data demonstrated salt stress of roots rapidly and
dramatically alters water relations in inflorescences.
Experiments described below were done to determine
if these alterations in water potential caused defects in
pollen development, pollination, ovule development,
or seed maturation. Defects in any one of these processes could lead to low reproduction rates.
The Effects of Stress on Pollen Development and
Filament Growth

Developing microspores are very sensitive to salt

stress (Namuco and OToole, 1986; Westgate and Boyer,
1986). Healthy Arabidopsis microsporocytes each contained a single nucleus that was densely stained by the

histochemical dyes (Fig. 4A), indicating rapid growth

and development. The cells that were salt stressed
when they were undergoing microsporogenesis did not
mature into viable pollen grains. Instead, these cells
became conspicuously vacuolated (Fig. 4B) and the
majority of them senesced within 2 d, leaving anthers
filled with pollen corpses (Fig. 4C). Interestingly, mature pollen was unaffected by even very long periods of
salt stress, as was demonstrated by their viability in
genetic crosses (data not shown). Thus, the extent to
which pollen was affected by stress depended upon the
developmental stage of the stamen.
The growth of filaments causes the stamens to brush
against the stigma and coat it with pollen. In healthy
plants, stamen filaments and petals undergo rapid
growth (Fig. 4E). In response to salt stress, however,
there was a decrease in the growth rate of stamen
filaments (Fig. 4F). While filaments from stressed
plants grew more slowly than their healthy counterparts, they did manage to pollinate the stigma (Fig.
4G). This slower growth delayed pollination by 12 to
24 h, although this delay did not prevent fertilization
of female gametophytes.
Callose Synthesis in Ovules and Embryos

An early indicator of imminent ovule abortion is

callose synthesis (Vishnyakova, 1991). The initial symptoms of ovule abortion, callose synthesis in the endothelium of ovules, were observed 12 h after stressing.
Twenty-four hours poststress, callose was most abundant in the endothelium and developing gametophyte
(Fig. 5, E and F).

Figure 3. The water potential in inflorescences rapidly equilibrates

after roots are stressed. Using a pressure chamber, inflorescence water
potential was measured before, during, and after salt stress. In healthy
plants grown under hydroponic conditions, the water potential in
flowers averaged 20.5 MPa. These plants were moved for 12 h to
a hydroponic medium supplemented with 200 mM NaCl. The water
potential in inflorescences dropped then equilibrated following the
salt-stress treatment. Once salt stress was discontinued, plants rapidly
returned to the previous, normal water potential. Points are the means
6 SD of five plants.
Plant Physiol. Vol. 135, 2004

Ovule Abortion in Arabidopsis

Figure 4. Salt stress disrupted the normal development of Arabidopsis pollen grains, ovules, and embryos. A, In healthy pollen
grains, cytoplasm was dispersed throughout the cell. Arrowheads identify pollen nuclei. B, Prior to pollen abortion, cells became
highly vacuolated. C, Many of these vacuolated pollen cells lysed and collapsed. D, A normal stage-12 flower, with the tips of the
petals just emerging from behind the sepals. Flowers at this stage were stressed and the ovules within monitored over time for
developmental defects. E, After 24 h, the flower shown in D opened and the stamens and petals had markedly increased in size.
As stamen filaments lengthened, anthers (a) coated the stigma (st) with pollen. F, A stage-12 flower that was salt stressed. Eighteen
hours later pollen dehisced, but filaments were too short to pollinate the stigma. G, A stage-12 flower that was salt stressed.
Twenty-four hours later, the filaments elongated sufficiently to pollinate the stigma. H, Within a normal stage-12 flower, the
gametophyte (g) is fully developed and competent to be fertilized. I, One day after salt stress, starch granules (arrows)
accumulated in the gametophyte. J and K, Subsequently, these starch grains were mobilized from ovules, and the cells in the
gametophyte become vacuolated. L, In the gametophyte, subcellular components were lysed and condensed. M, In some cells,
condensation of debris left little evidence that a gametophyte had occupied this space. N, Healthy suspensor (s) and embryo (e).
O, Following salt stress, an embryo aborted at the globular stage of embryogenesis. P, Before they senesced, the embryo and
suspensor became highly vacuolated.
Plant Physiol. Vol. 135, 2004


Sun et al.

Figure 5. Callose accumulated in the

gametophytes and embryo sacs of aborting ovules. Using Nomarsky optics, the
internal anatomy of ovules is shown in A
to C and G to I, while D to F and J to L
show callose fluorescence. The ovules in
A, D, G, and J were from healthy plants,
while those in B, C, E, F, and H to L were
harvested from salt-stressed plants. A,
Anatomy of a healthy ovule. Starchcontaining plastids were present in the
gametophyte (g). B, One day after plants
were stressed, additional starch grains
were produced in the gametophyte. C,
Two days after plants were stressed, the
central portion of the ovule, where the
gametophyte had been, contained cellular
debris. D, In healthy ovules, the integuments (i) synthesized only small amounts
of callose. E and F, In stressed ovules,
gametophyte cells synthesized copious
amounts of callose. G, In healthy ovules,
the embryo (e) developed within the embryo sac (es). J, Only minute amounts of
callose were present in healthy embryo
sacs. H and K, One day after plants were
salt stressed, the embryos synthesized
callose. I, One day later, embryo development ceased at the globular stage, and
large amounts of callose were present in
the embryo, embryo sac, and endothelial
walls (L).

Similarly, callose is a good indicator for embryo

senescence. Very little of this polysaccharide is found
in healthy embryo sacs (Fig. 5J). In stressed plants,
callose was highly abundant in the cell walls of the
suspensor and embryo (Fig. 5K). Later, callose synthesis spreads into the cell walls between the endothelium
and the embryo sac (Fig. 5L).

Degeneration of Cells within the Female Gametophyte

Depending on the stage of ovule development, stress

altered the development of various female reproductive structures. Because the anatomy of Arabidopsis
ovule and embryo development is well documented
(Mansfield and Briarty, 1991; Mansfield et al., 1991;
Robinson-Beers et al., 1992; Schneitz et al., 1995), the
anatomical differences between healthy and stressed
ovules will be described here. In healthy ovules, the
megaspore mother cell undergoes meiosis, yielding
a large megaspore (Fig. 6B). This cell undergoes further
cell divisions and differentiation to form a mature
gametophyte (Fig. 6C). When plants were stressed
during or just prior to meiosis (stage 2-IV), meiosis

concluded and megaspores were produced, but further

development of these into female gametophytes was
blocked. This developmental block usually occurred
when the gametophyte had two nuclei (Fig. 6E), but
occasionally the four-nuclei stage was reached. Once
gametophyte development ceased, gametophytic nuclei dissipated and cells collapsed (data not shown).
Before fertilization, in healthy plants flower and
ovule development are synchronized (Schneitz et al.,
1995). Consequently, stage 3-IV of ovule development
(the gametophyte is completely differentiated but unfertilized) occurs when petals protrude and become
visible from behind the sepals (Fig. 4D). If ovules were
stressed at stage 3-IV, the gametophyte frequently
underwent PCD (Fig. 4). When compared to healthy
ovules (Fig. 4H), the number of starch crystals within
the central cell of stressed gametophytes appeared to
transiently increase (Fig. 4I). During ovule abortion,
these starch deposits disappeared and gametophyte
cells became more vacuolated (Fig. 4, J and K). Subsequently, the vacuole ruptured and the contents of the
gametophyte condensed (Fig. 4, L and M). In stressed
plants, approximately three-fourths of the gametophytes displayed these anatomical characteristics
Plant Physiol. Vol. 135, 2004

Ovule Abortion in Arabidopsis

Figure 6. Salt stress derailed gametophyte development. Ovules from healthy (AC) and stressed (D and E) plants were marked at
stage 2-IV, and ovule development was monitored as a function of time. In terms of developmental time, the stressed ovules
shown in D and E correspond to similar developmental times for the healthy ovules shown in B and C, respectively. A, In an ovule
at stage 2-IV, the megaspore mother cell (mmc) occupies the central region of the nucellus. B, One day later, the megaspore
mother cell went through meiosis, yielding one large haploid megaspore (m). C, This megaspore underwent three rounds of
mitotic division and the products differentiated into a gametophyte. Within the gametophyte, the egg cell (ec), two synergid cells
(sc), two central cells (cc), and three antipodal cells (ac) were present. D, When ovules were stressed at stage 2-IV, they
completed meiosis, although the outer integument (oi) and inner integument (ii) grew more slowly than healthy ovules. E, The
megaspore from this stressed plant went through only one mitotic division, yielding a gametophyte (g) with just two nuclei.

during PCD. The anatomy of the remaining gametophytes was indistinguishable from wild type.
Senescence of Stressed Zygotes and Embryos

While environmental stress caused most female

gametophytes to abort during development, a fraction
of them remained viable. In plants stressed just prior
to anthesis, pistils were pollinated and fertilization
occurred in one-quarter of the ovules. While approximately 12 seeds per silique were fertilized, only an
average of three of these completed their development
to form viable seeds. The majority of the resulting
zygotes formed four-cell or globular-stage embryos but
did not develop further (Fig. 4O). Stressed embryos
commonly were more vacuolated than healthy counterparts (Fig. 4P). The condensed cytoplasm found in
some endothelial cells revealed that these cells underwent PCD (Fig. 4P). Since these cells normally
transport photosynthate into the gametophyte, their
demise should reduce transport rates.

stressed, DNA fragmentation was first detected in a

few cells of the integument and chalaza (Fig. 7). After
48 h, many more cells in these tissues exhibited DNA
fragmentation. This DNA damage was especially
abundant in endothelial and chalazal cells, which
transfer nutrients from the maternal tissues to the
gametophyte. Cells in the carpel wall did not undergo
DNA fragmentation. Although the TUNEL assay can
yield high backgrounds due to nonspecific labeling
artifacts (Wang et al., 1996), the low background in
carpel walls in these assays indicates this was not a
problem. Fragmentation signals occasionally were detected when abortion of cells within the gametophyte
was nearly complete and subcellular compartments,
including nuclei, began degrading and condensing.
Embryo and endosperm nuclei did not undergo DNA
fragmentation (Fig. 7, G and H).

Ovule Abortion in Arabidopsis

DNA Fragmentation and Changes in the Structure

of the Nucleus

The structure of chromatin often changes in cells

undergoing PCD. In some forms of PCD, fragmentation of the DNA between nucleosomes yields a ladder
of 180-bp fragments and multiples of this size. Cells
were assayed for DNA fragmentation using the TdTmediated dUTP nick-end labeling (TUNEL) assay,
which incorporates fluorescent nucleotides where
DNA is nicked. Twenty-four hours after plants were
Plant Physiol. Vol. 135, 2004

As ovules develop into seeds, the associated cells and

tissues go through a number of intricate developmental
and physiological processes. Because of the substantial
plant resources required for reproduction, plants regulate pollen, ovule, and seed development in response
to changing environmental conditions. Under harsh
environmental conditions, developing pollen, ovules,
and/or embryos abort in many seed-bearing plants.
The research described here details how a model plant
species, Arabidopsis, curtails plant reproduction in

Sun et al.

Figure 7. Maternal cells in stressed ovules (ov) underwent extensive DNA nicking or fragmentation. In B, D, and F to H, nicked
DNA was fluorescein labeled (yellow), while cell walls autofluoresced (red). In A, C, and E, DAPI-stained nuclei were observed
(blue). A and B, In healthy pistils, DNA was intact or unbroken. C and D, In stressed pistils, the DNA in many integument nuclei
and the chalaza nuclei was nicked. By contrast, nuclei in the carpel walls (cw), stigma (st), and style were not modified in this
manner. E and F, While maternal cells from stressed ovules exhibited DNA fragmentation, the nuclei in the gametophyte (arrows)
were undamaged. G and H, Embryo and endosperm nuclei from stressed plants did not exhibit any DNA fragmentation or
nicking. tt, Transmitting tract.

response to stress. While salt was used to stress plants,

similar rates of ovule abortion were observed when
using mannitol (Fig. 2). This result indicates that reduced seed production resulted primarily from water
stress, not ionic stress. While stress delayed filament
growth and pollination (Fig. 4), ovule abortion did not
result simply from a lack of synchronization between
pollination and ovule development. Ovules remained
viable for an additional 2 to 3 d in emasculated plants
(data not shown). Thus, ovule abortion was complete or
nearly complete at a time when gametophytes in
healthy plants remained fertile and receptive.
The duration of water stress affected both ovule and
pollen development in Arabidopsis, resulting in significant decreases in seed production (Fig. 2). Depending on the stage of development, stress had three
different effects on female reproduction: (1) gametophyte differentiation was derailed; (2) mature gametophytes underwent PCD; or (3) fertilized ovules
senesced (Figs. 4 and 6). When unfertilized ovules
were stressed just before anthesis, there were two
distinct developmental responses: (1) gametophytic
cells underwent lysis and condensation (Fig. 4); and

(2) many cells in the chalaza and maternal integuments exhibited DNA fragmentation (Fig. 7). PCD of
these maternal cells frequently lagged behind the
breakdown of gametophytic cells. In a number of crop
plants, including bean, rapeseed, maize, and soybean,
stress induces mature gametophytes to undergo PCD
(Moss and Downey, 1971; Sage and Webster, 1990;
Kokubun et al., 2001; Young et al., 2004). Thus, adverse
environmental conditions can cause plants to suspend
normal development of gametophytes, embryos, and
pollen grains, but the phenotype of abortion depends
upon the developmental stage at which stress was

Callose Synthesis

One of the first observable indications an ovule will

abort is the synthesis of callose around the developing
embryo sac (Vishnyakova, 1991). Callose often has
been proposed to isolate dying cells from living ones.
In muskmelon, callose is a cell wall polysaccharide
that acts as a molecular filter, limiting the diffusion of
Plant Physiol. Vol. 135, 2004

Ovule Abortion in Arabidopsis

small sugar molecules between the seed coat and

embryo (Yim and Bradford, 1998). Similarly, callose
may inhibit sugar transport in aborting Arabidopsis
ovules. To adequately test this hypothesis, however,
more extensive experiments need to be done.
Resource Limitation and Embryo Senescence

In many plant species, water deficits attenuate seed

set or yields (Boyer, 1982). Under most growing conditions, sufficient numbers of ovules are initiated to
maximize reproduction. Since most plants do not
experience optimal growth conditions, inevitably some
of these ovules abort. At times of environmental stress,
the nutritional demand of reproduction frequently exceeds the carbon and nitrogen resources allocated to the
ovule. Consequently, the number of ovules initiated
often exceeds the capacity of the gynoecium or, by
extension, the plant to provide adequate nutrition for
them all (Lee and Bazzaz, 1986). Following water stress,
the rate of photosynthesis dramatically drops in maize
(Westgate and Boyer, 1985a), thus reducing sugar transport into the embryo sac, which limits resources for
seed development (Schussler and Westgate, 1995). In
healthy plants, manual removal of some developing
ovules decreases the overall rate of abortion by providing more photosynthate for surviving ovules. For
example, the total number of cucurbit ovules that
matured into seeds remained steady but the rate of
ovule abortion declined when a portion of the ovules
were surgically removed (El-Keblawy and LovellDoust, 1999). The maize and cucurbit examples are
two of many instances where the demand for photosynthate limits seed production.
Resource limitation, however, is only one variable
affecting reproductive success. This point is elegantly
illustrated in experiments on maize plants that were
artificially supplemented with sucrose during seed
development. Westgate and Boyer (1985b) utilized
maize to study the effects of water stress on plant
reproduction. In these experiments, reduction in watering gradually imposed osmotic stress over a period of
days, which allowed plants to acclimate to this stress
(Westgate and Boyer, 1985b). In our experiments, osmotic conditions were changed rapidly so plants did
not have the opportunity to acclimate to osmotic stress.
Nonetheless, Arabidopsis exhibited many of the same
symptoms observed in other plant systems. In maize,
transport of photosynthate into seedlings was partially
corrected by stem-infusion of sucrose (Zinselmeier
et al., 1999; Fisher and Cash-Clark, 2000). If sucrose
deficiency were the sole obstacle to maize reproduction, then stem infusion of sucrose should completely
restore fertility. Insufficient transport of carbohydrates
into the ovary, as opposed to sugar deficiency, forces
seeds to abort (Schussler and Westgate, 1995). Defects
in seed development and sugar transport derive at least
in part from reduced invertase activity (Miller and
Chourey, 1992; Andersen et al., 2002). These plants
exhibit reduced acid invertase activity and carbohyPlant Physiol. Vol. 135, 2004

drate transport into ovules. In these plants, sucrose

accumulated in the apoplast between the maternal
tissue and the developing embryo. Thus, carbohydrate
transport into the embryo sac, as opposed to sucrose
availability, limited seed formation.
Defects in carbohydrate transport into the embryo
sac are not limited to maize. In other species, photosynthate transport becomes occluded in aborting
ovules concurrent with a massive proliferation of integument tissue (Rappaport et al., 1950; Savenchko,
1960; Lacey et al., 1997). In plants where the maternal
tissue proliferates, the ovule remains an excellent sink
tissue, even though the transfer of assimilates into
developing gametophytes or embryos has ceased. In
beans, as in a number of other plants, hormones were
implicated in the regulation of sink strength and seed
abortion (Tamas et al., 1979, 1986). These results indicate that hormones produced by maternal tissues
regulate the transfer of photosynthate into gametophytes and embryos.
In stressed Arabidopsis plants, very few seeds per
siliques matured (Fig. 2), even though one-quarter of
the ovules were fertilized and produced zygotes.
These zygotes formed globular embryos, but the majority of them senesced at this stage of embryogenesis. Embryo and endosperm nuclei showed no signs
of DNA nicking or degradation (Fig. 7), which indicates that embryos did not undergo PCD; rather, they
stopped developing, became quiescent, and senesced.
Similar to what was observed in maize, this failure in
Arabidopsis embryogenesis may result from a cessation of nutrient transfer into the embryo sac. To determine if this hypothesis is tenable, the movement of
assimilates into Arabidopsis embryo sacs needs to be
measured in both healthy and stressed plants.
Stress Can Halt Female Gamete Development

While male meiosis is highly sensitive to stress (Saini,

1997), this does not appear to be true with female
meiosis. When Arabidopsis ovules were stressed during meiosis, they completed this process and formed
megaspores. However, megagametogenesis arrested
before the mitotic divisions were completed (Fig. 6).
Thus, megagametogenesis was inhibited by environmental stress. These results are consistent with the
hypothesis that one or more stress-induced regulators
curtailed reproductive development.


Plants use a variety of mechanisms to regulate reproduction and maximize plant fitness. Lloyd (1980)
proposed that plants adjust the expenditure of maternal resources at a series of developmental stages,
regulating the number of flowers, gametophytes, and
embryos that advance further in development. Arabidopsis ovule abortion and embryo senescence were
regulated phenomena, which agrees with the conclu2365

Sun et al.

sions drawn by Lloyd in his maternal resources hypothesis. This strategy conserves plant resources since
terminating reproduction in an early developmental
stage expends relatively few resources. The resources
conserved by reducing plant fertility, thereby, can be
shunted into processes that permit plants to acclimate
to stressful environmental conditions. Even when
severely stressed, however, Arabidopsis allocates
sufficient resources to produce a few seeds in each
fruit, which ensures the genetic line is continued. We
hypothesize that by readjusting reproduction and resource allocation during stress, Arabidopsis plants
are able to tolerate harsh environmental conditions
and, if moderate environmental conditions return,
increase reproductive output. This hypothesis will
be tested by precisely monitoring resource allocation
to ovules and embryos during and after stress.


Plant Materials and Growing Conditions
Arabidopsis, ecotype Landsberg erecta, were grown under continuous
illumination at approximately 100 mE m22 s21 under hydroponic conditions
using Rockwool supports as described by Gibeaut et al. (1997). Their research
revealed that plants grown under these conditions grew faster and were
healthier than ones grown in soil. Peterson and Krizek (1992) proposed that
this was due to increased aeration and nutrient availability. In this study, the
hydroponic solution contained 5.0 mM KNO3, 2.5 mM KH2PO4, 2.0 mM MgSO4,
50 mM FeEDTA, 2.0 mM Ca(NO3)2, 70 mM H3BO3, 14 mM MnCl2, 0.5 mM CuSO4,
1.0 mM ZnSO4, 0.2 mM NaMoO4, 10 mM NaCl, and 0.01 mM CoCl2. Changing the
hydroponic environment of the roots can induce salt stress. To stress plants,
the hydroponic solution was supplemented with 200 mM NaCl. Plant tissues
analyzed in this experiment were pistils at early stage 3-VI, when megagametogenesis was complete but fertilization had not yet occurred. Ovule
developmental stages were as according to Schneitz et al. (1995).
Water potential was measured using a pressure chamber (PMS Instruments, Corvallis, OR). At each treatment interval, the balance pressure for five
separate inflorescences was measured.

Light Microscopy
For paraffin sections, flowers were fixed overnight in 10% formalin, 5%
acetic acid, and 50% ethanol. Samples were dehydrated in a graded ethanol
series then embedded in Paraplast1. Samples were sectioned at 5-mm
intervals, then stained with Safranin O (Polysciences, Warrington, PA) and
fast green (Johansen, 1940). For plastic sections, pistils were fixed overnight in
3% (v/v) glutaraldehyde in 50 mM sodium cacodylate buffer, pH 7. Samples
were postfixed with 2% osmium tetroxide, dehydrated, and infiltrated
according to Spurr (1969). Polymerized resin was sectioned at 0.5- to 2-mm
intervals and stained with thionine and acridine orange according to Paul
(1980). To observe callose deposition, ovules were cleared in BB41/2 (Herr,
1971) supplemented with 0.001% aniline blue, and observed by fluorescence
microscopy using 4#, 6-diamino-phenylindole (DAPI) excitation and emission
filters. A TUNEL assay detected DNA fragmentation. Using sectioned pistils,
the DeadEnd Fluorometric TUNEL kit (Promega, Madison, WI) labeled
nicked-DNA with fluorescein. A fluorescent DNA intercalating agent, DAPI
(Sigma, St. Louis), stained nuclei. An Axioscope microscope (Zeiss, Thornwood, NY) equipped with a Kodak MDS290 camera (Rochester, NY) captured
microscopic images. Digital images were adjusted for publication using
Adobe Photoshop 5 (San Jose, CA).

We thank Greg Erdos and Karen Kelley of the ICBR Electron Microscopy
Facility (University of Florida) for technical assistance, Tim Martin (Univer-


sity of Florida) for assistance with pressure chambers, and Margaret Joyner
(University of Florida) and two anonymous reviewers for helpful comments.
Received March 26, 2004; returned for revision June 29, 2004; accepted July 1,

Andersen MN, Asch F, Wu Y, Jensen CR, Naested H, Mogensen VO, Koch
KE (2002) Soluble invertase expression is an early target of drought
stress during the critical, abortion-sensitive phase of young ovary
development in maize. Plant Physiol 130: 591604
Apse MP, Aharon GS, Snedden WA, Blumwald E (1999) Salt tolerance
conferred by overexpression of a vacuolar Na1/H1 antiport in Arabidopsis. Science 285: 12561258
Asch F, Wopereis MCS (2001) Responses of field-grown irrigated rice
cultivars to varying levels of floodwater salinity in a semi-arid environment. Field Crops Res 70: 127137
Bohnert HJ, Ayoubi P, Borchert C, Bressan RA, Burnap RL, Cushman JC,
Cushman MA, Deyholos M, Fischer RL, Galbraith D, et al (2001) A
genomics approach towards stress tolerance. Plant Physiol Biochem 39:
Boyer JS (1982) Plant productivity and the environment. Science 218:
Dat J, Vandenabeele S, Vranova E, Van Montagu M, Inze D, Van
Breusegem F (2000) Dual action of the active oxygen species during
plant stress responses. Cell Mol Life Sci 57: 779795
Davidonis GH, Johnson A, Landivar JA (2000) Cotton mote frequency
under rainfed and irrigated condition. J Cotton Sci 4: 19
El-Keblawy A, Lovell-Doust J (1999) Maternal effects in the progeny
generation in zucchini, Cucurbita pepo (Cucurbitaceae). Int J Plant Sci
160: 331339
Fisher DB, Cash-Clark CE (2000) Gradients in water potential and turgor
pressure along the translocation pathway during grain filling in normally watered and water-stressed wheat plants. Plant Physiol 123:
Gibeaut DM, Hulett J, Cramer GR, Seemann JR (1997) Maximal biomass
of Arabidopsis thaliana using a simple, low-maintenance hydroponic
method and favorable environmental conditions. Plant Physiol 115:
Gueta-Dahan Y, Yaniv Z, Zilinskas BA, Ben-Hayyim G (1997) Salt and
oxidative stress: similar and specific responses and their relation to salt
tolerance in citrus. Planta 203: 460469
Herr JM (1971) A new clearing technique for the study of ovule development in angiosperms. Am J Bot 58: 785790
Johansen DA (1940) Plant Microtechnique. McGraw-Hill, New York
Kawasaki S, Borchert C, Deyholos M, Wang H, Brazille S, Kawai K,
Galbraith D, Bohnert HJ (2001) Gene expression profiles during the
initial phase of salt stress in rice. Plant Cell 13: 889905
Kokubun M, Shimada S, Takahashi M (2001) Flower abortion caused by
preanthesis water deficit is not attributed to impairment of pollen in
soybean. Crop Sci 41: 15171521
Lacey EP, Smith S, Case AL (1997) Parental effects on seed mass: seed coat
but not embryo/endosperm effects. Am J Bot 84: 16171620
Lee TD, Bazzaz FA (1986) Maternal regulation of fecundity: nonrandom
ovule abortion of Cassia fasciculata Michx. Oecologia 68: 459465
Lloyd DG (1980) Sexual strategies in plants I. An hypothesis of serial
adjustment of maternal investment during one reproductive session.
New Phytol 86: 6979
Mansfield SG, Briarty LG (1991) Early embryogenesis in Arabidopsis
thaliana. II. The developing embryo. Can J Bot 69: 461476
Mansfield SG, Briarty LG (1992) Cotyledon cell development in Arabidopsis thaliana during reserve formation. Can J Bot 70: 151164
Mansfield SG, Briarty LG, Erni S (1991) Early embryogenesis in Arabidopsis thaliana. I. The mature embryo sac. Can J Bot 69: 447460
Miller ME, Chourey PS (1992) The maize invertase-deficient miniature1
seed mutation is associated with aberrant pedicel and endosperm
development. Plant Cell 4: 297305
Moss GI, Downey LA (1971) Influence of drought stress on female
gametophyte development in corn (Zea mays) and subsequent grain
yield. Crop Sci 11: 368373

Plant Physiol. Vol. 135, 2004

Ovule Abortion in Arabidopsis

Namuco OS, OToole JC (1986) Reproductive stage water stress and

sterility. Crop Sci 26: 317321
Paul RM (1980) The use of thionin and acridine orange in staining semithin sections of plant material embedded in epoxy resin. Stain Technol
55: 195196
Peterson TA, Krizek DT (1992) A flow-through hydroponic system for the
study of root restriction. J Plant Nutr 15: 893911
Quesada V, Garcia-Martinez S, Piqueras P, Ponce MR, Micol JL (2002)
Genetic architecture of NaCl tolerance in Arabidopsis. Plant Physiol
130: 951963
Rappaport J, Satina S, Blakeslee AF (1950) Ovular tumors associated with
incompatible crosses in Datura. Am J Bot 37: 576586
Robinson-Beers K, Pruitt RE, Gasser CS (1992) Ovule development in wildtype Arabidopsis and two female-sterile mutants. Plant Cell 4: 12371249
Sage TL, Webster BD (1990) Seed abortion in Phaseolus vulgaris L. Bot Gaz
151: 167175
Saini HS (1997) Effects of water stress on male gametophyte development
in plants. Sex Plant Reprod 10: 6773
Savenchko MI (1960) Anomalies in the structure of angiosperm ovules. Bot
Sci 130: 1517
Schneitz K, Hulskamp M, Pruitt RE (1995) Wild-type ovule development
in Arabidopsis thaliana: a light microscope study of cleared whole-mount
tissue. Plant J 7: 731749
Schussler JR, Westgate ME (1995) Assimilate flux determines kernel set at
low water potential in maize. Crop Sci 35: 10741080
Spurr RA (1969) A low viscosity epoxy resin embedding medium for
electron microscopy. J Ultrastruct Res 26: 3143
Tamas IA, Kock JL, Mazur BK, Davies PM (1986) Auxin effects on the
correlative interaction among fruits in Phaseolus vulgaris L. Proc Plant
Growth Reg Soc Am, pp 208215
Tamas IA, Wallace DH, Ludford PM, Ozbun JL (1979) Effect of older fruits

Plant Physiol. Vol. 135, 2004

on abortion and abscisic acid concentration of younger fruits in

Phaseolus vulgaris L. Plant Physiol 64: 620622
Vishnyakova MA (1991) Callose as an indicator of sterile ovules. Phytomorphology 41: 245252
Vivian-Smith A, Luo M, Chaudhury A, Koltunow A (2001) Fruit development is actively restricted in the absence of fertilization in
Arabidopsis. Development 128: 23212331
Wang H, Li J, Bostock RM, Gilchrist DG (1996) Apoptosis: a functional
paradigm for programmed plant cell death induced by a host-selective
phytotoxin an invoked during development. Plant Cell 8: 375391
Westgate ME, Boyer JS (1985a) Carbohydrate reserves and reproductive
development at low leaf water potentials. Crop Sci 25: 762769
Westgate ME, Boyer JS (1985b) Osmotic adjustment and the inhibition of
leaf, root, stem and silk growth at low water potentials in maize. Planta
164: 540549
Westgate ME, Boyer JS (1986) Reproduction at low silk and pollen water
potentials in Maize. Crop Sci 26: 951956
Yim K-O, Bradford KJ (1998) Callose deposition is responsible for apoplastic semipermeability of the endosperm envelope of muskmelon
seeds. Plant Physiol 118: 8390
Young LW, Wilen RW, Bonham-Smith PC (2004) High temperature stress
of Brassica napus during flowering reduces micro- and megagametophyte fertility, induces fruit abortion, and disrupts seed production.
J Exp Bot 55: 485495
Zhang HX, Blumwald E (2001) Transgenic salt-tolerant tomato plants
accumulate salt in foliage but not in fruit. Nat Biotechnol 19: 765768
Zhu J-K (2001) Plant salt tolerance. Trends Plant Sci 6: 6671
Zhu J-K (2002) Salt and drought stress signal transduction in plants. Annu
Rev Plant Biol 53: 247273
Zinselmeier C, Jeong B-R, Boyer JS (1999) Starch and the control of kernel
number in maize at low water potentials. Plant Physiol 121: 2535