Metabolic responses of rice source and sink organs during recovery from combined drought and heat stress in the field

Abstract Background Drought and heat stress effects on rice have been extensively studied, in particular during the sensitive flowering and grain-filling stages. However, in the field these stresses usually occur together because reduced transpirational cooling under drought conditions results in increased plant tissue temperature. In addition, environmental stresses are usually transient and the ability to efficiently recover from stress may be at least as important for overall stress tolerance as the direct stress response itself. Nevertheless, nothing is known about recovery mechanisms after drought and heat stress in rice under field conditions. Results We have used gas chromatography–mass spectrometry–based metabolomics to elucidate the metabolic responses of flag leaves, flowering spikelets, and developing seeds from 3 rice cultivars differing in their drought and heat tolerance to rewatering after stress in the field. Within 60 hours after rewatering, many stress-responsive metabolites returned to their control levels, although recovery was not complete. In addition, control plants showed developmental differences that were revealed by metabolite profiles during 60 hours of post-stress sampling, in particular in developing seeds. Correlation analysis identified several metabolites as marker candidates for the stability of grain yield or quality under conditions of combined drought and heat stress. Conclusions The rewatering responses of stressed plants seemed to be a combination of the reversal of stress effects and reinitiation of development after stress relief. The identified potential markers can be useful in efforts to breed stress-tolerant rice germplasm to ensure food availability under changing climate conditions.


Results
We have used gas chromatography-mass spectrometry (GC-MS)-based metabolomics to elucidate the metabolic responses of flag leaves, flowering spikelets and developing seeds from three rice cultivars differing in their drought and heat tolerance to rewatering after stress in the field. Within 60 h after rewatering, many stressresponsive metabolites returned to their control levels, although recovery was not complete. In addition, control plants showed developmental differences that were revealed by metabolite profiles during 60 h of post-stress sampling, in particular in developing seeds. Correlation analysis identified several metabolites as marker candidates for the stability of grain yield or quality under conditions of combined drought and heat stress.

Conclusions
The rewatering responses of stressed plants seemed to be a combination of the reversal of stress effects and reinitiation of development after stress relief. The identified potential markers can be useful in efforts to breed stress-tolerant rice germplasm to ensure food availability under changing climate conditions. as availability, access, utilization, stability and safety. However, I think that the results of this study are useful only in ensuring food availability under climate changing conditions, there is no improvement from the safety point of view. Therefore, I would suggest changing "food security" for "food availability". Response: We agree with the reviewer and have changed the phrase as suggested in both places.
Data description -Line 108-111: I would suggest moving this paragraph to the result section Response: We would rather leave this information under Data description. We think that it is important at that point in the paper, because it provides the reader with the background necessary to understand the rational of the experiments. Also, these are not new results, we simply cite an earlier paper.
-Line 112: the authors stated that overall they collected 1241 samples, but I cannot understand how to get this number. How many samples/per years? How many samples/per time points? How many biological replicates? Response: We have included a more detailed description of the sample and replicate numbers in Methods/Sample collection now to clarify this point.
Analysis and discussion: -Line 134: Did the author observed any influence of the harvesting years on the PCA plot? Response: There were differences in metabolite composition between years as conditions in the field are never the same in different years. However, we did not analyse these differences further, but rather treated the samples from all years as replicates to obtain robust metabolic responses. We have therefore not elaborated on the point of yearly variation in our paper.
-The authors focused on the primary metabolism changes. However, lipids are well known to be involved in the plant response to stresses. May the authors comment on that? Response: We agree that lipids, along with secondary metabolites and other compound classes may also be of importance. However, since we have no data on these other compound classes it did not seem appropriate for us to speculate on that.
Methods -Line 461-462: three to five replicates. What influence the number of collected replicates per sample? Did the authors analyse all the collected sample replicates? Response: Yes, we analysed all replicates we collected. In most cases we obtained 5 replicates per year, giving us 15 replicates in total across the three years. In some cases it was 4 and in only a few cases it was 3. This is now stated explicitly under Methods/Sample collection to clarify this point. We do not think that the small fraction of samples with less than 14 replicates in total (only about 7%) had any influence on our analyses or the interpretation of the results. nothing is known about the recovery process from combined drought and heat stress in rice and  flag leaves and developing seeds we may hypothesize that seeds showed a higher rate of 174 metabolic change than flag leaves. In particular, the massive reduction in the content of many 175 amino acids and organic acids could argue for a rapid conversion from metabolically active pools 176 to a reserve storage. This is in agreement with metabolomic studies in maize [43] and rice [44] 177 that also found a strong reduction in the levels of many primary metabolites during seed 178 development.  (Fig. 5C). In contrast, among the metabolites that exhibited reduced levels after rewatering, more 236 than one-third (9 metabolites) were common among all three cultivars at 36 h after rewatering 237 (Fig. 5D), out of which six were amino acids (Fig. 5G). N22 showed a more similar response 6G). Conversely, the six metabolites that showed increased levels 12 h after rewatering were all 289 cultivar specific (Fig. 6A) and the same was true for the two metabolites with increased levels 60 290 h after rewatering (Fig. 6 E). At 36 and 60 h after rewatering, the number of metabolites with 291 significantly changed levels that were common between all three cultivars increased to eight 292 ( Fig. 6D) and nine (Fig. 6F), respectively. In addition to ribitol, which was already common 293 between all cultivars 12 h after rewatering, seven amino acids exhibited lower levels relative to  Developing seeds 303 In the PCA of the metabolite profiles of developing seeds (Fig. 1D) the number of metabolites that were significantly influenced by rewatering compared to the 312 stressed state was also quite low (Fig. 7). We only observed one, four and two metabolites that 313 showed an increase 12 h, 36 h, and 60 h after rewatering (Fig. 7A, C, E) and only raffinose 314 content in Anjali was increased at all time points (Fig. 7G). However, it was also increased in  331 We have previously identified potential marker metabolites for tolerance to combined      Secondly, our data suggest that while many stress-responsive metabolites returned to 426 (almost) control levels within three days after stress relief, this was clearly not true for all such 427 metabolites. While this may in part be due to the additional developmental effects on metabolites 428 as discussed above, these persistent metabolic changes are a sign of metabolic imprinting [48].

429
Metabolic imprints may lead to a modified stress response under a recurrent stress situation.      Competing interests 565 The authors declare that they have no financial or non-financial competing interests.