V. Physiological and Metabolic Bases for Reproductive Failure under Drought

72



HARGURDEEP S. SAINI AND MARK E. WESTGATE



meiosis in PMCs was not always normal in rather severely stressed barley plants

(Skazkin and Zavadskaya, 1957). More recently, an increase in meiotic abnormalities was also noticed in water-stressed rice plants (Namuco and O’Toole,

1986). These included an increase in univalents, lagging chromosomes, noncongression of bivalents in metaphase, and micronuclei formation. The abnormalities

started to increase at relatively moderate stress (leaf ␺ Ϫ1.1 to Ϫ1.9 MPa) and

peaked at ␺ around Ϫ2.2 MPa. Under severe stress (leaf ␺ Ϫ3.5 MPa), the entire

meiotic process was arrested. Whether any of these abnormalities contributed to

pollen sterility is not known, although the cessation of meiosis under severe stress

would certainly do so, unless it is reversed on rewatering. In contrast, no such abnormalities were noticed in wheat PMCs (Saini, 1982), which always complete

meiotic division under stress (Lalonde et al., 1997a; Saini et al., 1984). Meiosis is

apparently also completed in the PMCs of moderately stressed rice because microspores are produced, although their subsequent development fails (Sheoran

and Saini, 1996). Microspores of wheat, with a few exceptions (Lalonde et al.,

1997a), continue to develop normally for several days before their development is

arrested (Lalonde et al., 1997a; Saini et al., 1984). These observations indicate that

a more subtle lesion rather than a catastrophic failure of meiosis is the probable

cause of male sterility. At least in wheat, the direct desiccation of the sporogenous

tissue is not responsible for this developmental arrest (Saini and Aspinall, 1981;

Westgate et al., 1996).

Developmental anatomy of stress-affected anthers gives some promising clues

about the metabolic events that may be linked to the failure of pollen development.

Cereal pollen grains are rich in starch (Franchi et al., 1996). Starch accumulates

late during pollen development and is then used to support pollen germination and

pollen tube growth (Clément et al., 1994; Franchi et al., 1996; Miki-Hirosige and

Nakamura, 1983; Pacini and Franchi, 1988). Pollen grains rendered sterile by

drought or other stresses fail to accumulate starch (Ito, 1978; Saini and Aspinall,

1981, 1982a; Sheoran and Saini, 1996). Water stress also changes the pattern of

starch distribution in anthers and inhibits intine development in pollen grains

(Lalonde et al., 1997a; Saini et al., 1984). The timing of the inhibition of starch

deposition coincides with the appearance of structural lesions during anther development (Lalonde et al., 1997a; Saini et al., 1984), suggesting that a disturbance

in carbohydrate availability and/or metabolism may be involved in this developmental failure. This view is also supported by the observation that increased sucrose uptake increases the grain set in wheat spikes cultured in nutrient solution

(Waters et al., 1984). The addition of ABA to the nutrient solution decreases sucrose uptake and grain set, and both effects are reversed by supplementing the

medium with additional sucrose (Waters et al., 1984). By extrapolating these observations to intact plants, one could argue that because water stress inhibits the

rate of photosynthesis and export of assimilate from leaves (Boyer and McPherson, 1975; Hanson and Hitz, 1982), the supply of sucrose to anthers could be the



REPRODUCTION DURING DROUGHT



73



limiting factor under drought. However, the levels of sucrose and other sugars in

wheat and rice anthers increase in response to drought stress, suggesting that sugar starvation per se may not be the trigger for pollen sterility (Dorion et al., 1996;

Sheoran and Saini, 1996). This accumulation of sugars could, however, be misleading because sucrose utilization in anthers is inhibited by stress. Activities of

soluble acid invertase (Dorion et al., 1996; Sheoran and Saini, 1996) and cell wallbound invertase (J. S. Minhas and H. S. Saini, unpublished) decline dramatically

in wheat and rice anthers and do not recover even after plants are rehydrated. This

is not a generalized effect on anther enzyme activities, particularly in wheat, because the activities of ADP-glucose pyrophosphorylase and soluble and bound

forms of starch synthase are not affected (Dorion et al., 1996). Preliminary results

show that the expression of genes encoding these invertases is also inhibited (J. S.

Minhas and H. S. Saini, unpublished results). The latter effect is not due to a generalized inhibition of transcription because the expression of the ADP-glucose pyrophosphorylase gene is affected little by stress (Lalonde et al., 1997b). In rice anthers, the activities of starch synthases and ADP-glucose pyrophosphorylase

decline somewhat during meiotic-stage stress or soon thereafter (Sheoran and Saini, 1996), but the most dramatic effect of stress is on invertase activity. Invertase

is the dominant enzyme of sucrose cleavage in the anthers of many species, including the two just mentioned (Bryce and Nelsen, 1979; Dorion et al., 1996; Nakamura et al., 19890, 1992; Sheoran and Saini, 1996). A decline in invertase activity

in anthers would interfere with the proper processing of incoming sucrose. The resulting decline in the levels of hexoses needed for biosynthesis and energy generation could jeopardize crucial metabolic and developmental processes. In wheat,

the decline in invertase activity precedes or coincides with the first anatomical signs

of pollen abortion (Dorion et al., 1996; Lalonde et al., 1997a; Saini et al., 1984).

The observed accumulation of sucrose in stress-affected anthers, despite the expected decline in the sucrose supply on inhibition of photosynthesis (Boyer and

McPherson, 1975; Hanson and Hitz, 1982), could be due to the inhibition of invertase (Dorion et al., 1996). The redistribution of starch, particularly its accumulation in the connective tissue (Lalonde et al., 1997a), is consistent with the inhibition of sucrose utilization by stressed anthers. A similar fate of excess sucrose has

been demonstrated in Lilium anthers (Clément and Audran, 1995). The molecular

and metabolic regulation of invertase inhibition in relation to stress-induced pollen

sterility merits close examination because it is the earliest occurring stress-induced

lesion identified to date, and its timing qualifies it as a potential causal event.



B. CARBOHYDRATE AVAILABILITY AND KERNEL ABORTION

The magnitude of decline in leaf ␺ that causes kernel abortion in maize also completely inhibits photosynthesis and leads to a reduction in carbohydrate reserves in



74



HARGURDEEP S. SAINI AND MARK E. WESTGATE



stem (Westgate and Boyer, 1985a). Therefore, it has been suggested that the abortion may be attributable to a curtailed supply of carbohydrates during pollination

(Westgate and Boyer, 1986b). This view is also supported by the observation that

the CO2 enrichment of wheat plants partially overcomes the yield losses due to water stress (Gifford, 1979). Kernel number is reduced in proportion to the inhibition

of photosynthesis by water deficit or varying light intensity (Schussler and Westgate, 1991a). Although carbohydrates continue to accumulate in vegetative sinks

such as leaf and stalk, despite an inhibition of photosynthesis, their movement to

reproductive sinks is restricted severely (Schussler and Westgate, 1991a,b). These

data indicate that early kernel development is dependent on the supply of assimilate from concurrent photosynthesis, which cannot be replaced by the remobilization of reserves stored in other tissues. Because sugar concentrations in the ovaries

do not change and because the sugar uptake by ovaries isolated from stressed plants

is inhibited, kernel set depends on the rate of movement of the current assimilate to

reproductive organs rather than the concentration of sugars per se (Schussler and

Westgate, 1991b). Consistent with this conclusion, experimental manipulations to

either enhance the accumulation of carbohydrate reserves prior to anthesis or reduce the sink size for a fixed availability of assimilates do not diminish the extent

of water stress-induced kernel loss (Schussler and Westgate, 1994; Zinselmeier et

al., 1995a). Infusing liquid culture medium into the stem in a quantity sufficient to

replace carbohydrates lost by the inhibition of photosynthesis during floweringstage water deficit, however, can prevent the abortion of most (ϳ70%) kernels

(Boyle et al., 1991; Zinselmeier et al., 1995b). Stem infusion does not rehydrate the

plant, nor is kernel abortion prevented by the infusion of water alone. The ingredient in the medium that prevents abortion is sucrose, the level of which has to be elevated above that in well-watered ovaries to prevent abortion under stress (Zinselmeier et al., 1995b). Supplemental sucrose sustains the ovary growth rate, which

is correlated with a high starch content and turgor maintenance by osmotic adjustment (J. S. Boyer and C. Zinselmeier, personal communication). These results further support the conclusion that the maintenance of assimilate supply from current

photosynthesis is essential in preventing kernel abortion in stressed plants.

Delivery of photosynthate to the developing ovaries also depends on their metabolic activity (i.e., sink demand) (Jenner, 1982). Several observations indicate that

the capacity to utilize available assimilate is impaired by water stress. The reduction in the level of starch, which is a product of sucrose metabolism in ovaries, is

only partially restored by sucrose infusion (Zinselmeier et al., 1995b,c). Ovaries

isolated from stressed plants take up sucrose less rapidly than those from wellwatered plants (Schussler and Westgate, 1991b). Kernel abortion is not prevented

completely by feeding culture medium or sucrose (Boyle et al., 1991; Zinselmeier

et al., 1995b) or by increasing assimilate supply through cultural or genetic manipulations (Schussler and Westgate, 1991a,b, 1994; Zinselmeier et al., 1995a).

Direct evidence for a metabolic lesion has been furnished by Zinselmeier et al.



REPRODUCTION DURING DROUGHT



75



(1995c), who showed that acid invertase activity in the ovaries of water-stressed

plants was inhibited strongly in parallel with the cessation of ovary growth, an accumulation of sucrose, and a decrease in the level of reducing sugars. Maize

ovaries induced to abort in vitro by high temperature also have low acid invertase

activity (Hanft and Jones, 1986). These metabolic events are remarkably similar

to those observed in stress-affected anthers (Dorion et al., 1996; Sheoran and Saini, 1996) and may point to a common metabolic basis for the failure of these two

sinks to develop during drought. The abortion of kernels (and anthers) in waterstressed plants could be caused by the metabolic block resulting from a decline in

invertase activity. Further support for this conclusion comes from the maize mutant miniature-1, which lacks soluble and wall-bound acid invertase and fails to

produce normal kernels (Miller and Chourey, 1992). How invertase activity could

be modulated by sucrose is discussed in Section VC2.



C. REGULATION OF GRAIN FILLING AND MATURATION

In cereals, the first 6 to 14 days after anthesis set the potential for subsequent

development (Jones, 1994). Drought and high temperature during this period reduce the storage capacity of cereal grains by decreasing the number of endosperm

cells and/or the number of amyloplasts initiated (Artlip et al., 1995; Brocklehurst

et al., 1978; Jones et al., 1985, 1996; Nicolas et al., 1985; Ouattar et al., 1987a).

Neither high temperature nor water deficit affects the supply of sucrose to the endosperm, suggesting that the availability of assimilates, per se, is not limiting

(Nicolas et al., 1984; Ober and Setter, 1990).

Several lines of evidence indicate that hormonal regulation, particularly by

ABA and cytokinins, may be involved. In maize, the ABA content of pedicel/placento-chalazal tissues and endosperm is low, whereas the ABA content of the embryo is high when sink potential is established (Jones and Brenner, 1987). Water

deficits imposed during the first week after pollination increased kernel ABA levels about eight-fold (Ober and Setter, 1990). The ABA apparently is of maternal

origin and decreases to control levels on plant rehydration. The temporary increase

in ABA has little, if any, effect on invertase, sucrose synthase, or starch synthase

activities, measured in vitro, or on subsequent starch accumulation (Ober and Setter, 1990). However, the increase in ABA content in apical kernels of droughted

plants was correlated with an inhibition of endosperm cell division and a decrease

in capacity for starch synthesis (Ober et al., 1991). Similarly, a rapid rise in endosperm ABA content coincides with the disruption of cell division caused by high

temperature (Jones et al., 1985). Exogenous application of ABA has a similar negative impact on endosperm cell division (Mambelli and Setter, 1998; Myers et al.,

1990) Thus, an increase in endosperm ABA concentration may serve to downregulate the kernel sink potential.



76



HARGURDEEP S. SAINI AND MARK E. WESTGATE



In contrast, evidence from maize, rice, barley, and wheat implicates cytokinins

as a positive effector in establishing sink potential. Levels of zeatin and zeatinriboside increase as much as 500-fold after pollination (Dietrich et al., 1995; Lur

and Setter, 1993; Morris et al., 1993; Schreiber, 1990). While maximum cytokinin

levels have been correlated with kernel size, a causal relationship between the two

has not been demonstrated. The application of cytokinins can enhance kernel set

and sometimes increase the yield of cereals (Dietrich et al., 1995; Hradecka and

Petr, 1992). In each case, however, cytokinins were applied prior to the period of

rapid cell division and promoted earlier stages of flower and zygote development.

To our knowledge, no direct evidence shows that peak cytokinin levels decrease

in drought-stressed kernels nor has a benefit from the exogenous application of cytokinins during the rapid cell division phase been demonstrated. Such information

would clarify the role of cytokinins (or perhaps the cytokinin/ABA ratio) in maintaining the kernel sink potential during drought.

Once cell division in the endosperm is complete and the kernel enters into the

linear growth phase (phase II, Fig. 1), the machinery for reserve accumulation is

established for the remainder of kernel growth. Water deficits have little impact on

the rate of kernel growth, but often shorten the duration of filling. The hormonal

regulation of grain fill duration has been implicated by correlative data linking

ABA content with storage protein deposition, late embryo-genesis abundant

(LEA) proteins, and acquisition of desiccation tolerance (Dure, 1997; Ingram and

Bartels, 1996; Quatrano et al., 1983). However, several studies suggest that ABAresponsive events late in grain filling are initiated by dehydration itself rather than

by a precedent hormonal signal (Bochicchio et al., 1993; Chandler et al., 1993;

Iturriaga et al., 1992). Therefore, we need to consider how drought during kernel

filling affects kernel water relations and the desiccation process.

Egli and co-workers have shown that the final mass of soybean embryos can be

manipulated by restricting seed water volume mechanically (Egli et al., 1987) or

osmotically (Egli, 1990). Similarly, the physical restriction of seeds during the period of rapid water uptake leads to decreases in kernel size in wheat (Millet and

Pinthus, 1984), barley (Grafius, 1978), oat (Grafius, 1978), and rice (Murata and

Matsushima, 1975). Moreover, embryos can grow beyond their normal size in culture if cell expansion is allowed to continue and C and N are available (Egli, 1990).

The mechanism by which water uptake (maximum seed volume) is controlled has

not been established, but it has been suggested that a decrease in assimilate supply to the seed alters the osmotic gradient driving water flow into the seed (Egli

and TeKrony, 1997). This idea, however, is not well supported by evidence. For

example, the maximum water content of the maize endosperm is reached during

rapid seed filling (Westgate, 1994; Westgate and Boyer, 1986c), and therefore, it

is difficult to envision that there would be sufficient assimilate to maintain rapid

dry matter accumulation, but insufficient solute to sustain an osmotic gradient for

water flux. There is also no evidence that osmotic conditions within the seed vary



REPRODUCTION DURING DROUGHT



77



during this transition period. The ␺ and osmotic potential of the maize embryo and

endosperm are nearly constant during rapid grain filling when the maximum osmotic (water) volume is reached (Westgate, 1994; Westgate and Boyer, 1986c).

Similarly, in wheat and barley, there is no discernible change in osmotic conditions

within the kernel as they achieve maximum water content early in grain filling

(Barlow et al., 1980; Morris et al., 1991; Sofield et al,. 1977b). In fact, a change

in the osmotic environment within the kernel would not be expected because water uptake continues rapidly throughout grain filling (Sofield et al., 1977b). The

possibility that the yield threshold for cell wall expansion increases at this time is

also unlikely because no change in embryo or endosperm turgor was observed as

the maximum water content was reached (Barlow et al., 1980; Westgate, 1994;

Westgate and Boyer, 1986c).

Nonetheless, data from a number of unrelated studies on kernel development in

cereals show a general correspondence (r2 ϭ 0.86) between final kernel dry weight

and maximum water content during grain filling (Fig. 2). Among these studies,

however, the individual relationships for maize, rye, and triticale are rather tenuous. Triticale kernels, for example, accumulated much less dry matter than expected from their potential volume. Maize kernels achieved a wide range of



Figure 2 Relationship between maximum water content and final kernel dry weight in maize,

wheat, rye, and triticale. Data were calculated from kernel growth curves for maize (Egli and TeKrony, 1997; Westgate and Boyer, 1986c; M. E. Westgate, unpublished), wheat (Brooks et al., 1982; Egli

and TeKrony, 1997; Sofield et al., 1997a,b), triticale, and rye (Saari et al., 1985). Regression equations

include only wheat or maize data. Open symbols, well watered; closed symbols, water stressed.



78



HARGURDEEP S. SAINI AND MARK E. WESTGATE



weights from a maximum water content of about 150 mg. Within a species, the relationship between final kernel weight and water content was not genotype specific (data not shown). In general, kernels from water-stressed plants followed the

same general pattern as those from well-watered plants.

The pattern of net water content during kernel development varies considerably

among the cereals. The general pattern presented in Fig. 1 is typical of maize (Egli

and TeKrony, 1997; Westgate and Boyer, 1986c) and triticale (Saari et al., 1985)

in that the water content begins to decrease during rapid grain filling. Rapid deposition of storage reserves replaces the volume occupied by water, leading to tissue desiccation. In contrast, the water content in wheat, rye, and barley kernels remains fairly constant throughout grain fill and begins to decrease roughly when

the maximum dry matter is attained (Barlow et al., 1980; Brooks et al., 1982; Egli

and TeKrony, 1997; Saari et al., 1985; Sofield et al., 1977b). Sofield et al. (1977b)

provided convincing evidence that the rapid decrease in water content late in filling of wheat kernels was due to the blockage of water uptake caused by lipid deposition in the apoplast of the pigment strand. Regardless of the mechanism by

which cellular desiccation occurs, lack of water late in development could limit

the synthesis of storage reserves even when assimilates are available (Adams and

Rinne, 1980; Westgate, 1994). If so, the general pattern of kernel dry matter accumulation across species, genotypes, and environments should be fairly similar

when expressed on a kernel moisture basis. Figure 3 shows the pattern of kernel

dry weight accumulation with decreasing kernel moisture for maize and wheat in

several unrelated studies in which the kernel water content was monitored during

development. A large variation in the final dry weight within species reflects the

differences in genetic potential as well as environmental effects on kernel growth.

The key feature of these data is that dry matter accumulation ceased at about the

same moisture content regardless of genotype or environment. Kernel development on water-deficient plants follows the same general pattern (Fig. 4).

Whether soil moisture is abundant or severely limiting, dry matter accumulation

ceases at about 30 and 40% moisture in maize and wheat kernels, respectively

(Figs.3 and 4).

The consistency of this relationship implies that a common mechanism may

control the cessation of kernel growth and that this mechanism is coupled to the

water status of the kernel. It is reasonable to assume that the rapid synthesis of end

products such as starch, protein, and oil during grain filling requires optimum coordination among substrate availability, enzyme activation, translation, and transcription (Bewley and Black, 1994). A decrease in any one of these factors during

desiccation could lead to the cessation of dry matter accumulation. In both control

and water-deficient plants, the progressive loss of water from the endosperm and

the restriction of water uptake by the embryo lead to a rapid decrease in endosperm

and embryo ␺s late in grain filling (Westgate, 1994). Concentrations (per gram of

water) of sucrose and amino acids in the endosperm of maize, wheat, barley, and



REPRODUCTION DURING DROUGHT



79



Figure 3 Pattern of kernel dry weight accumulation with decreasing kernel moisture. Data for

maize are adopted from Egli and TeKrony (1997), Westgate and Boyer (1986c) and M. E. Westgate

(unpublished data). Wheat data are adopted from Egli and TeKroney (1997), Sofield et al. (1997a,b),

Barlow et al. (1980), and Brooks et al. (1982). Closed symbols are used to highlight differences between identical genotypes.



triticale remain fairly constant during linear grain fill and then increase as the water content declines (Brooks et al., 1982; Chevalier and Lingle, 1983; Nicolas et

al., 1984; Saari et al., 1985). Therefore, the initial decrease in ␺s must be due, in

large part, to the passive concentration of solutes. Although it has been suggested

that specialized structures within the apoplast may exist to maintain high solute

concentrations within the embryo and endosperm (Bradford, 1994), the fact that

the decrease in kernel ␺s occurs well in advance of physiological maturity strongly suggests that end product formation was not limited by a lack of substrate when



80



HARGURDEEP S. SAINI AND MARK E. WESTGATE



Figure 4 Pattern of dry weight accumulation with decreasing kernel moisture in maize and wheat

kernels sampled from well-watered (closed symbols) and water-deficient (open symbols) plants. Water was withheld at developmental stages indicated by arrows. Data are adopted from Brooks et al.

(1982). and M. E. Westgate (unpublished data).



dry matter accumulation ceased in either well-watered or water-deficient plants

(Sofield et al., 1977b; Westgate, 1994).

Activities of enzymes involved in nitrogen and carbohydrate metabolism in the

endosperms of maize, wheat, and barley decrease late in grain fill (Chevalier and

Lingle, 1983; Doehlert et al., 1986; Muhitch, 1991; Saari et al., 1985; Singletary

et al., 1990). Adams and Rinne (1980) proposed that dehydration inactivated these

enzymes, thus terminating the accumulation of reserves. In vitro studies show that

desiccation does indeed inhibit enzyme activity dramatically, but only at very low

moisture contents typical of dry seeds (Rupley et al., 1983; Stevens and Stevens,



REPRODUCTION DURING DROUGHT



81



1977). Seeds are considered fully “wetted” at about 28% moisture, at which point

the properties of the water resemble those of water in dilute solutions (Vertucci,

1989). The ␺ of such seeds is about Ϫ20 MPa, which is much lower than the Ϫ1.0

to Ϫ2.0 MPa at which grain dry matter accumulation ceases. The complete hydration of macromolecule surfaces occurs at about 95% RH or about Ϫ7.5 MPa

(Vertucci, 1989), which is well below seed ␺ at end of grain filling. Also, enzymes

extracted from physiologically mature seeds retain fairly high levels of activity

(Chevalier and Lingle, 1983; Doehlert et al., 1986; Muhitch, 1991; Saari et al.,

1985). Although such measurements likely overestimate in vivo activity, as enzymes are assayed under ideal conditions, available data suggest that the osmotic

conditions prevalent in the endosperm and embryo late in grain filling do not directly inhibit enzyme activity.

Rather, the loss of enzyme activity probably reflects a decreased capacity for

protein synthesis as water content declines (Bewley, 1981). Greene (1983) and

Kermode et al. (1989) have shown that the decline in protein synthesis late in seed

development is coupled to the level of translatable mRNAs. Premature desiccation of castor bean (Ricinus communis L.) seeds caused the same quantitative and

qualitative changes in developmental mRNAs that occur during normal maturation drying (Kermode et al., 1989). This result implies that desiccation itself may

be a developmental queue for terminating transcription and/or translation of

mRNAs required for storage product formation. To date, studies on these mRNAs

in cereal kernels have been restricted to the period of rapid storage product formation ( Jones, 1978; Marks et al., 1985; Viotta et al., 1975). Such measurements

need to be extended to the later stages of grain filling to determine the fate of developmental mRNAs during tissue desiccation. Similar mRNA profiles in kernels

of well-watered and water-deficient plants would support the hypothesis that the

duration of dry matter accumulation is controlled at the level of transcription and/

or translation.



D. LONG-DISTANCE SIGNALS AS TRIGGERS

OF REPRODUCTIVE FAILURE

1. Role of Hormones in Drought-Induced Male Sterility

Initial inquiries into the physiological basis for water stress-induced male sterility in wheat (see Section IVB) revealed that the water status of spikelets changed

little, despite a substantial decline in leaf ␺ in response to meiotic-stage water

stress (Morgan, 1980b; Saini and Aspinall, 1981). Westgate et al. (1996) later

demonstrated that individual floral organs, including anthers, were capable of osmoregulating effectively to maintain or even increase their turgor during stress. It

is clear, therefore, that the male sterility in water-stress wheat, and perhaps other



82



HARGURDEEP S. SAINI AND MARK E. WESTGATE



plants, is not caused by desiccation of reproductive structures, but results from indirect consequences of a drop in ␺ or turgor elsewhere in the plant (Morgan, 1980b;

Saini and Aspinall, 1981). How the existence of stress in the vegetative parts of

the plant is communicated to the reproductive tissue is an important question.

The fact that sterility is induced only after leaf turgor falls to zero (Morgan,

1980b) suggests that a turgor-responsive process, such as the accumulation of abscisic acid (Aspinall, 1980; Pierce and Raschke, 1980; Walton, 1980), may be involved. The possibility that ABA produced on turgor loss in the vegetative tissues

(e.g., leaf) is translocated to the inflorescence where it triggers events leading to

male sterility, has been studied in some detail, and evidence favoring this hypothesis is summarized next.

a. ABA Accumulation in Reproductive Tissues

Meiotic-stage spikelets and anthers accumulate ABA during water stress, despite no change in their turgor (Morgan, 1980b; Saini and Aspinall, 1982b; Westgate et al., 1996), indicating that the hormone is transported from leaves or other

vegetative tissues. The long-distance transport of ABA to spikes in wheat, and in

other species, has been demonstrated (Goldbach and Goldbach, 1977; Ober and

Setter, 1990; Wolf et al., 1990).

b. Effects of ABA Application

Application of exogenous ABA to spike or leaf causes pollen sterility and loss

of grain set in wheat (Morgan, 1980b; Saini and Aspinall, 1982b; Waters et al.,

1984; Zeng et al., 1985). Both ABA application and water deficit have their maximal effect during meiosis (Morgan, 1980b; Morgan and King, 1984; Saini and Aspinall, 1981, 1982b; Zeng et al., 1985), and both treatments induce male sterility

without affecting female fertility (Saini and Aspinall, 1982b). Stress- and ABAaffected anthers and pollen grains look morphologically similar at maturity (Morgan, 1980b; Saini and Aspinall, 1981, 1982b).

c. ABA Concentrations and Sterility

When ABA and stress treatments induce similar levels of sterility, the spikelet

ABA content resulting from exogenous ABA application is within the same order

of magnitude as when the hormone accumulates endogenously in response to water stress (Saini and Aspinall, 1982b). Grain set correlates negatively and tightly

with ABA levels in anthers, ovaries, and glumes of water-stressed wheat plants

(Westgate et al., 1996).

d. Other Indirect Evidence

Consistent with the just-mentioned observations, differences in the seed set between well-watered plants of two cultivars of wheat correlate inversely with the

ABA content of the spike (Morgan and King, 1984). Spikelets and leaves of a nu-