A central feature of abortion is the time lag between triggering events and the cessation of development apparent several days later. Because of this lag, studies usually require blind sampling of the ovaries, before abortion can be seen. The samples might include individuals that are not going to abort, causing measurements to have an underlying baseline of normal ovary development. Differences between treatments become small and variable depending on how many normal ovaries are included in the sample.
No early markers, physiological or molecular, are available for identifying aborting ovaries. The closest alternatives seem to be treatments that ensure virtually all the ovaries will abort or develop normally, as in Fig. 1. With this approach, the baseline of normal ovaries is small or non-existent in the abortion treatments, and the fate of the sampled ovaries can be assigned confidently. Constant, repeatable growth environments with high radiation are valuable tools for producing these treatments. Highly controlled glasshouses or field sites with stable and reproducible weather patterns are also useful.
Another feature of abortion is the reduced size of the ovaries. At low w, ovary growth is inhibited and fresh weight fails to accumulate. The fresh weight may even decline because of dehydration. Expression of data on a fresh-weight basis thus gives highly variable results. A similar argument applies to ovary dry weight. Changes in ovary constituents (sugars, ABA, mRNA, etc.) may be greater or less than changes in fresh or dry weights, and the concentrations will rise or fall depending on which factor changes the most. Most important, concentrations changing from altered fresh or dry weights will change similarly for all the other cell constituents, rendering their significance debatable.
There seems to be no totally satisfactory solution to this problem. The best that can be offered is to determine the content of constituents in the whole ovary rather than concentrations. By making the whole-ovary measurements before treatments are imposed and monitoring the change as the treatments progress, in vivo fluxes are revealed. Fluxes into or out of the ovary result from rates of enzymatic and physiological processes, and accumulation of the constituent indicates the difference between influx and efflux, i.e. whether production (phloem delivery, transcription activity, metabolite production) exceeds use (breakdown, metabolism, consumption in biosynthesis). In an ovary that increases in size, these principles still apply but the change in size also may account for some of the change in content. As a result, measuring contents before treatment (organ development is identical) and following the changes during the treatment (organ development differs) will indicate whether production exceeds use regardless of whether size increases or not.
The upward slope in Fig. 2 indicates that delivery exceeds use after the treatment is imposed. The effect of increased size cannot be determined from this approach, but delivery clearly exceeds use in any event. Conversely, if the constituent is consumed faster than it is delivered, there is a downward slope. Zero slope means the constituent is available at the same level as before the treatments were imposed. The advantage of this whole organ, time-based approach is that flux information for the organ is obtained from simple measurements of content regardless of changes in organ size, fresh weight, or dry weight. Because the fluxes indicate how rapidly enzymatic and metabolic processes are acting in vivo in whole metabolic systems, the information is readily interpretable for development.
One of the more intriguing features of early reproductive growth is the similarity of the biochemical events leading to pollen sterility and ovary abortion at low w
(Saini and Westgate, 2000
). Both involve decreases in invertase activity (Zinselmeier et al.
; Sheoran and Saini, 1996
). Both result in starch depletion in the affected organs (Zinselmeier et al.
; Lalonde et al.
). Some of the effects can be induced by ABA applications to the parent (Saini and Westgate, 2000
). It is tempting to seek a unifying cause, and one possibility is an involvement of photosynthate production. High ABA can close stomata and inhibit photosynthesis, in some ways mimicking the effect of low w
on photosynthesis. Does ABA have its effects by inhibiting photosynthesis and decreasing the sugar stream to the ovaries, as seems to be true for low w
It should be noted that these changes take place during early water limitation, which is of importance to agriculture. They often begin before visible symptoms appear and do not involve severe desiccation where other enzyme and gene mechanisms might come into play (Kramer and Boyer, 1995). It also is noteworthy that many enzyme and gene changes are set into motion, but the key will be finding those few changes that have an impact on plant performance, i.e. that are limiting.
It is remarkable that abortion of newly formed zygotes can be so complete while the parent plant can recover so fully. In maize, about half of the shoot dry weight is normally in the kernels at the end of the growing season. A water deficit at pollination can shrink this fraction to zero. The farmer is left with a reasonable vegetative crop but no grain. Important enzymes and genes must determine these two developmental fates, and revealing them is likely to require knowledge of whole plant physiology, biochemical regulation of the physiology, and the particular genes linking them together. With these genes identified, it may be possible to prevent the irreversible fate of the ovaries and pollen on plants subjected to low w.