A major role of model organisms is to facilitate investigation and gain insights into biological mechanisms that can be applied to other organisms. A good example is how research on Arabidopsis impacts on research in the biomedical field [58
], despite the ~1 billion years since divergence. More commonly, research using model species is usually used in comparisons within phylogenetic boundaries, i.e. plant, animal or fungi. Arabidopsis and rice are the forefront plant dicot and monocot models, respectively. A variety of comparisons between these species have been carried out from the genome [59
], proteome [60
] and metabolome levels [61
] revealing insights into conserved and distinct features and responses. At a transcriptome level, a variety of comparisons have also been carried out with a variety of gene families (see introduction for examples). In this study, a comparison of organ and abiotic stress responses was carried out at the global transcriptome level. Given that abiotic stress responses were compared between rice and Arabidopsis, it is important to point out that each of these species has evolved different mechanisms or tolerances optimising their survival in their native climates. This is particularly relevant in terms of optimal climatic conditions, where Arabidopsis is grown in a temperate climate (22°C) [62
], whilst optimal growth conditions for rice is in a tropical climate (28-30°C) [63
]. Therefore, any results derived from comparisons to extreme temperature stress would reveal not only the response to the temperature but also incorporate the natural responses to the respective temperature. The comparisons in this study were done to define similarities and differences at a global level for the first time. While similarities point to conserved processes and provide targets to dissect basic core processes, differences provide opportunities to identify different approaches that have arisen in evolution. Note that differences also represent the divergent responses to the conditions applied, not just a reflection of the basal transcriptome, and thus have the potential to reveal the diversity of responses to altering conditions. While analysing the differences between closely related varieties of a species, such as in Arabidopsis, in which whole genome approaches can now be applied to tens or even hundreds of varieties, has the potential to reveal the diversity in responses within essentially a very similar set of genes, analysis between different species will reveal alternative responses.
When carrying out comparisons between species, the typical assumption is that similar genes (defined as orthologues) are likely to have similar responses. When analysing both transcript response and orthology, it is important to consider the stringency in the definition of orthologues. Previous studies, may have over-estimated the proportion of rice genes that have Arabidopsis orthologues and vice versa, however, these estimates were most likely based on the sequence of genes that had been determined and over-estimated. The data presented here is conservative, in that the definition of orthology is based on the Inparanoid data - which could be considered a gold standard. However, even using functional based definitions of gene function (e.g. Pageman FUNCATs) the responses are still observed to be divergent for certain functional groups as a whole. Thus, irrespective of the strict definition of orthology, or even same the functional categorisation, the responses between Arabidopsis and rice appear to be divergent for specific groups of genes, particularly in response to abiotic stress.
This study revealed that the transcriptional network and underlying cis
regulatory factors of rice and Arabidopsis differs significantly in many aspects. Only by multi-dimensional analysis of gene expression across a multitude of microarray studies in both species, was it possible to get an idea of the transcriptomic flexibility i.e. the proportion of genes showing tissue specific or differential expression under various abiotic stress conditions. Combining expression with orthology revealed the danger in assuming orthology also reflects similarity in transcriptomic or even proteomic response. Firstly, the specific transcriptomes of rice seeds, flower, leaf and root was seen to be distinct to that of Arabidopsis, notably complying with findings from other transcriptomic studies [32
]. This lack of orthology between these "organ specific" genes indicates that the observed morphological differences between rice and Arabidopsis organs are also correlated with distinct gene expression and although correlation is not cause, it does present a possible avenue by which distinct morphology can be obtained. While the overlap in organ specific expression between Arabidopsis and rice is low at the level of gene orthology, for roots the overlap in function for many categories is much higher. Thus, despite differences in roots between monocots and dicots [65
], it appears that at a functional level, the organ specific transcript profiles are more conserved compared to flower, seed and leaf.
In contrast, it was observed that a significantly larger proportion of genes with orthologues were present in each of the abiotic stress responsive sets compared to the percentage of orthologues in the whole genome level. This suggests that although distinct genes may explain morphological differences between species, it may not be the reason for differential tolerances to abiotic stresses. Interestingly, the proportions of genes showing comparable, opposite and unchanging responses to abiotic stresses differed between rice and Arabidopsis. These represent particularly important findings, given that most studies tend to focus on similarities between plants, especially when considering orthologous genes. In rice, there were more than 4 times as many DEGs following drought and salt stress compared to the number of DEGs following these stresses in Arabidopsis, and this cannot only be accounted for by differences in the gene coding content of the respective genomes, as the number of differentially expressed genes in response to heat and cold for both species was approximately the same. In addition, this is also not due to species specific rice gene expression as there is in fact a larger percentage of known Arabidopsis orthologues for the rice genes changing under drought and salt stress (> 50%) compared to the genome percentage (~31% Arabidopsis orthologues). Noticeably, despite differences in the number of DEGs, it was evidenced that the number of DEGs showing opposite responses was much lower than the number of DEGs showing complementary changes, indicating that the response to drought or salt stress is more conserved. Closer examination of the orthologous genes showing common expression responses revealed the conserved down-regulation of translation functions and up-regulation of membrane transport functions in both Arabidopsis and rice. This suggests that under these conditions, energy demanding process such as translation are down-regulated, whilst the change in water content under drought or salt stress affects membrane fluidity in a conserved manner.
Interestingly, the number of DEGs under heat and cold stress were found to be comparable across rice and Arabidopsis. However, it was surprising to find that equal or more orthologous genes responded in the opposite manner as in those responding in a similar manner, providing novel evidence for a divergence in the regulation of these genes in response to temperature extremities. Specifically, an overall up-regulation of rice genes was observed compared to the respective orthologous genes in Arabidopsis following cold or heat stress. This suggests that the regulation of gene expression in response to these temperature extremities has diverged, with genes involved in secondary metabolism, amino acid degradation and redox metabolism being up-regulated under these conditions in rice, whilst the Arabidopsis orthologues remained unchanging or oppositely regulated. Overall, it is important to note that the number of DEGs observed in response to abiotic stress between Arabidopsis and rice cannot only be attributed to the differences in the size of gene families or duplication of genes. A specific example of this principle was seen for the non-symbiotic haemoglobin encoding genes, where the transcripts for all gene family members responded in opposite manners.
Given the general down-regulatory trend observed for the transcripts encoding redox pathway components in Arabidopsis following heat treatment, it was of interest to consider the regulatory mechanisms that may be controlling this down-regulation. Typically, transcriptional control is examined as the main mechanism regulating transcript abundance, however, given the observed down-regulation, a role for mRNA degradation was also considered. A previous study examined the global mRNA degradation rates for Arabidopsis, following transcriptional inhibition [55
]. Brief examination of the mRNA half-lives of the genes encoding redox pathway components (as in the Arabidopsis genes in Figure ) revealed remarkably high mRNA half-lives (> 10 h) for several of these genes, including 2 genes encoding catalases (AT4G35090, AT1G20630), a gene encoding DHAR (AT5G16710), 2 genes encoding dismutases (AT4G25100, AT3G10920), a gene encoding GR (AT3G54660) and the gene encoding a non-symbiotic hemoglobin (AT2G16060). mRNA half-lives of >10 h indicates a high level of stability of these mRNAs, thus it was surprising to note the significant down-regulation of these within only 3 h of heat treatment in Arabidopsis. Similarly, a large down-regulation of transcripts encoding translation functions is seen to occur in both Arabidopsis and rice following drought or salt treatment, again within only 3 h of stress treatment. The observed rapid down-regulation of these transcripts in Arabidopsis was somewhat surprising, given the expected high level of stability of these transcripts [55
] and may be explained by a mechanism of active degradation. A previous study in yeast revealed that in response to changes in oxygen availability, there is active degradation of specific transcripts that occurs faster than the steady state decay rates [66
]. Considering this, it is possible that part of the response mechanism to different abiotic stresses is the active degradation of specific transcripts, e.g. the possible active degradation of transcripts encoding redox pathway components following heat treatment in Arabidopsis.
In terms of transcriptional regulatory processes, it also appears that there were more differences than similarities between rice and Arabidopsis. Global comparison of all over-represented, putative cis
-acting sites within the 1 kb upstream regions of co-expressed orthologous genes revealed some unexpected findings. Specifically, it was observed that even when similar expression patterns were observed under abiotic stress for orthologous genes in both species, the regulatory mechanisms that drive these responses appeared to differ, with some exceptions. Overall, these findings indicate that even when two plants have similar genes (orthologues) in their genome and even if these genes show similarities in their transcriptomic responses, it is possible that the factors regulating this gene expression can differ. To consider these factors in both species, the expression of TFs was considered. Notably, the CAREs and TF families of AP2 and NAC were found to be enriched in both species in response to several abiotic stresses, complying with the previously established conserved regulatory role for these families in both Arabidopsis and rice [53
]. Thus, the outcome of the in silico
analysis in this study was strengthened by the support of previous experimental observation. Notably, although some NAC TFs responded in the same manner between Arabidopsis and rice under certain abiotic stresses, it has also been suggested that NAC TFs may have additional roles in rice [53
]. Other groups of TFs also enriched in the response of both species to specific stresses included the Auxin TFs in response to drought and salt, and bZIP, HSF and GNAT in response to heat. With the exception of the cold responsive set, overall, more families of TFs appeared to be affected in rice, compared to Arabidopsis. This suggests that in rice, the regulatory network in response to abiotic stress is more diverse and that TFs have addition roles in the stress response.
In addition to transcriptional regulation, the changes observed in transcript abundance observed are also the result of post-transcriptional regulation. Interestingly, a brief search of the genes encoding the oppositely regulated transcripts in Arabidopsis revealed some examples of genes with known to be regulated at the post-transcriptional level. For example, a recent study of the regulation of SEN1 (At4g35770) revealed that the transcript level for this gene as well as two other genes; a gene encoding a leucine zipper TF (At2g22430) and an gene with unknown function (At3g26740) are regulated by a controlled mRNA degradation mechanism [67
]. Targets of this controlled mRNA degradation pathway, including the aforementioned genes, are characterised by the presence of the downstream (DST) motif in their 3'UTRs and this motif is associated with rapid degradation of these transcripts [67
]. Notably, in Arabidopsis, all 3 of the aforementioned genes are down-regulated in response to one or more abiotic stresses whilst the rice orthologues of these genes are up-regulated or unchanged in abundance for that stress. In addition, a brief search for the DST motif in the rice orthologues of these genes revealed that the DST motif was not present in these 3'UTRs, providing an example that indicates potential divergence in post-transcriptional regulation between Arabidopsis and rice.
The differences in the transcriptome response of Arabidopsis and rice provides opportunities to identify species specific responses and thus broaden the possibilities of transferring traits from model to crop species. In this respect, the orthologous genes that significantly changed in opposite directions warrant further investigations. Notably, the transcripts of many genes in these various sub-sets encode proteins involved in stress defence and are classified as such. The control and regulation of genes involved in various processes classified as redox, biotic stress and secondary metabolism, which are all intimately associated with stress responses, will likely reveal species specific regulation and response that may offer new insights for translational biology. In particular, the fact that a non-symbiotic hemoglobin gene is regulated in the opposite manner in rice (up) compared to Arabidopsis (down) in all 4 stresses merits investigations into nitric oxide signalling differences that may exist between both plants in response to abiotic stress [68
]. In addition, the finding that transcripts encoding components involved in redox functions also displayed opposite trends under heat stress, suggests that differences in reactive oxygen species or reactive nitrogen species related functions may exist between both species. Although there are reports where the expression of heterologous genes in crop species confers resistance to a stress [23
], this is often accompanied by growth penalties [66
]. One possible reason for the observed growth penalties is the different interaction and/or regulatory environments of genes (or more correctly the proteins they encode) in different species. Thus, although reactive oxygen or nitrogen species are likely to play signalling roles in both Arabidopsis and rice under a variety of conditions, especially stress, the response to these signals may differ significantly.