3.10 Comparative expression of selected DE transcripts in WT andStpfld252 leaves
The expression of several DE genes was evaluated by qRT-PCR to validate the results obtained with the RNA microarray experiments. Genes were selected on the basis of their stress responses. Out of 10 DE genes evaluated, 4 showed a strict correlation between the microarray and qRT-PCR data, including the FC, whereas other 5 displayed similar expression patterns in both procedures, but quantitative differences in FC were observed between them (Figure 7). In all 5 cases, the observed differences did not modify cluster assignment.
Chloroplast-targeted Fld prevented the drought-dependent repression of photosynthetic genes to various extents (Figures 5, 6). The PSII core complex protein psbY and the photosynthetic Fd isoform 1 were down-regulated in WT leaves under water limitation, as members of cluster 7 (Figure 5). While expression patterns of the two genes determined by microarray analysis and qRT-PCR were similar, the latter procedure gave significantly lower FC values for drought-dependent repression in the wild type (Figure 7). Instead, the two methods yielded a good quantitative correlation for the PSI component psaK belonging to cluster 12 (Figure 6b). Cluster 7 also contained genes encoding two stress-responsive enzymes of central metabolism regarded as drought molecular markers: a granule-bound starch synthase and nitrate reductase 3 (Azedo-Silva, Osório, Fonseca & Correia 2004). Microarray results for both genes were quantitatively validated by qRT-PCR (Figure 7).
Genes primed by Fld expression included the drought-associated transcription factor ATHB7 (cluster 15, Figure 6c), which has been reported to promote stomatal closure among other abiotic stress responses (Ré, Capella, Bonaventure & Chan 2014). The ATHB7 gene was induced by Fld in the absence of stress and by drought in both lines, displaying one of the highest FC values (Figure 7). The opposite behavior (e.g. , Fld- and drought-dependent repression) was exhibited by a gene encoding an aminocyclopropane-1-carboxylate (ACC) oxidase involved in ethylene biosynthesis, which was a member of cluster 17 (Figure 6c). In both cases, microarray analysis and qRT-PCR displayed quantitative differences that did not modify cluster assignment. Another gene displaying a large FC in the transcriptional profiling wasGAST1 , encoding a protein involved in plant development and used as a molecular marker to monitor changes in endogenous active gibberellin levels (Shi & Olszewski 1998). Stress-dependent repression in both WT and Stpfld leaves matches the patterns observed in cluster 2 (Supplementary Figure S5), although lower FC values were obtained for the drought condition in the qRT-PCR experiment compared to the microarray (Figure 7).
Further, two genes encoding components of the proteasomal system (cluster 3, Supplementary Figure S6) were evaluated: subunit PSα2β, which contributes to the assembly of the 20S proteasome core complex, and RPN9b, a regulatory subunit of the 26S proteasome that is involved in the ATP-dependent degradation of ubiquitinated proteins (Geng, Wenzel & Tansey 2012). Whilst the RPN9b-encoding transcripts showed a good quantitative correlation between the microarray and qRT-PCR assays, those corresponding to PSα2β would instead be allocated to cluster 34 instead of cluster 3 on the basis of the qRT-PCR assay (Figure 7). Noteworthy, the change in cluster assignment was caused by a single significant difference in the FC obtained for stressed WT plants (Figure 7).