Understanding Rice Sterility Induced By Drought--Investigation Of Rice Response To Drought By Physiology And Proteomics

Reproductive-stage drought is a major threat to rainfed rice production. However, identification of the most drought-sensitive process has been hampered by the long time (~2 weeks) that plants take to create water deficits through transpiration in the field. Through the use of a proper soil/root mass ratio, we have developed a greenhouse procedure for rapid induction of water deficit (3-6 days). This procedure allows the impact of drought stress on the four major yield components (panicles m"2, spikelets panicle"1, % filled grain, 1000 grain weight) to be assessed throughout the reproductive stage in a range of cultivars. Water deficit reduced the yield of reference cultivar IR64 maximally when it occurred at heading. Six days of deficit caused an 80% yield loss, principally by blocking grain filling through pollen sterility. Anther development is highly vulnerable to water deficit but African tropical japonica land race Moroberekan is more tolerant than IR64 in terms of maintaining pollen starch accumulation, grain filled percentage and overall yield (especially top part of panicle) under water-deficit. Maintaining starch in pollen grains is very important for improving drought tolerance. In order to have a better understanding of the water-deficit impact on anther (pollen) development, we employed the proteomics to analysis the protein expression profiles and metabolite profiles in rice anthers under water-deficit along with anther development. Over 2000 anther proteins from anther tissue were separated by 2D electrophoresis followed with silver staining for analytical gels or colloidal Coomassie brilliant blue staining for preparative gels. Individual protein abundance was quantified with software (Melanie 3) and 35 protein spots were identified with mass spectrometry (MALDI-TOF MS and /or ESI-TOF MS/MS) and bioinformatics. There were at least 22 anther protein spots detected to be responsive (up- or down-regulated) significantly to the water-deficit in IR64 while only 11 protein spots in Moroberekan. Among those protein spots, only 7 spots were shown to respond in the same changing pattern in two varieties. After rewatering, although there were 14 protein spots in IR64 and 4 spots in Moroberekan, which could be partially recovered , there were still 8 protein spots for IR64 and 12 spots for Moroberekan, which were still affected by the stress-later-effect. Comparing theprotein expression at heading both in the control and water-stressed treatment, in the anther of IR64, 14 protein spots were up-regulated and 29 protein spots were down-regulated while in the anther of Moroberekan, 18 protein spots were up-regulated and 24 protein spots were down-regulated. Although the mitosis division of pollen was affected by water deficit, the stressed pollen could still reach about the same stage as that in the control which was revealed by DAPI staining. Mass spectrometry of identification proteins revealed that the carbohydrate metabolism and glycolysis, protein turn over and degradation system, flavonoid synthesis pathway, nitrogen metabolism and storage protein synthesis and so on were interrupted by the drought stress which differed in two varieties. Our data also supported that there was no serious oxidative stress in anthers when rice experieced water-deficit but there was dramatic ABA accumulation in anthers under drought.