Evolution of differential regulatory mechanisms can lead to quite distinct physiological attributes. In the present study, we have identified one such regulatory schema that regulates osa-miR408 and responds differentially in drought-sensitive and -tolerant indica rice varieties. A comparison of the drought stress response in drought-sensitive (Pusa Basmati 1 and IR64) and drought-tolerant (Nagina 22 and Vandana) indica rice varieties revealed that, during drought stress, levels of miR408 transcript decrease significantly in sensitive cultivars, whereas they remain elevated in the tolerant cultivars. The trend is reflected in young seedlings, as well as in flag leaf and spikelets of adult plants (heading stage). Members of the plastocyanin-like protein family targeted by miR408 also show the inverse expression profile and thus accumulate at a lower level in tolerant cultivars during drought. Interestingly, some members of this family are implicated in maintaining the cellular redox state and spikelet fertility in Arabidopsis. An investigation of miR408 loci (including promoter) in all four cultivars did not reveal any significant sequence variation indicating an involvement of the upstream regulatory schema. Indeed, a similar variety-specific stress response was found in the Oryza sativa squamosa promoter-binding-like 9 transcription factor that regulates miR408 expression. We further demonstrate that drought-mediated induction of miR408 in Nagina 22 is regulated by [Ca2+]cyt levels. However, [Ca2+]cyt does not appear to regulate miR408 levels in Pusa Basmati 1, suggesting a variety-specific evolution of regulatory schema in rice.
5′ RNA Ligase Mediated Rapid Amplification of cDNA Ends
Pusa Basmati 1
Tolerance to abiotic stress conditions involves the complicated interplay of numerous molecular components. Decades of research in plant biology have generated a wealth of information regarding the molecular mechanisms of tolerance to abiotic stress conditions ; however, we are still a long way from achieving a complete understanding. One very important component of the stress tolerance regulatory mechanisms comprises the small RNAs, especially the microRNAs (miRNAs). Plants have a diverse small RNA population consisting of miRNAs, trans-acting small interfering RNAs (siRNAs), natural antisense siRNAs (nat-siRNAs), repeat-associated siRNAs (rasiRNAs) and long-siRNAs (lsiRNAs) . Most of the miRNAs are known to be highly conserved across various species  to the extent that the miRNA biogenesis machinery components are even found in archaebacteria and eubacteria . They are known to regulate plant development and physiological processes such as leaf and flower differentiation, floral identity, the auxin response, etc. [5, 6]. The pleiotropic developmental defects observed in Arabidopsis mutants ago1, dcl1, hen1, hyl1 and hst are impaired for miRNA biogenesis and their mode of action emphasizes the global regulatory nature of these small molecules [7-11]. Undoubtedly, they have also been implicated in governing cellular processes involving both biotic and abiotic stress adaptation mechanisms in plants, such as antibacterial resistance [12, 13], the response to fungal infection , cold, drought and salinity [15, 16], mechanical stress , UV-B radiation , heavy metal starvation , sulfate/phosphate starvation [20-23], the nitrogen starvation response , the nitrate response , copper deficiency [26-29] and wounding . Analysis of conserved miRNAs in different plants has shown that miRNAs are regulated by distinct regulatory mechanism to the extent that the same miRNA responds quite differently to various developmental and environmental cues in different plants. Thus, miRNAs are themselves regulated by a very dynamic regulatory mechanism.
In the present study, we exploit this regulatory dynamism to identify the miRNA(s) that may play a nodal role in the regulation of the stress tolerance mechanism of rice. A huge biodiversity exists for rice and studies have identified several rice varieties/cultivars with a natural robustness that enables them to withstand stress conditions. We selected four such indica rice cultivars: Pusa Basmati 1 (PB1), IR64, Nagina 22 (N22) and Vandana. These varieties contrast with respect to their tolerance to drought. Although PB1 and IR64 are sensitive but high-yielding rice varieties, N22 and Vandana are known for their tolerance to drought conditions. We analyzed the selected stress responsive and conserved miRNAs in these varieties with the aim of identifying the miRNAs that may be regulated differentially under similar drought stress conditions. We report at least one rice miRNA, miR408, which tends to behave in a totally opposite manner in the drought tolerant varieties compared to the sensitive varieties under same stress regime. The pattern is reflected in both young and adult tissues. The difference in the expression profile could not be attributed to any sequence-based variation but, instead, to variation in the upstream signal transduction pathway. We further explore the regulatory mechanism controlling miR408 expression and demonstrate that the stress-mediated expression of osa-miR408 is mediated by abscisic acid and cytosolic Ca2+ levels. We also found that miR408 expression is Ca2+ regulated only in N22 but not in PB1. Moreover, we found evidence for transcription factor Oryza sativa squamosa promoter-binding-like 9 (OsSPL9) being involved in such differential regulation.
osa-miR408 is regulated by a variety-specific drought response
The present study aimed to analyze the behaviour of drought responsive miRNAs in drought tolerant and sensitive indica rice cultivars. Four drought responsive miRNAs (osa-miR398, osa-miR319, osa-miR171 and osa-miR408) were selected and their drought stress-mediated modulation of steady-state mature transcript levels was analyzed. To explore the possibility of variety-specific stress regulatory mechanisms, the analysis was carried out simultaneously in two different varieties of indica rice: PB1 and N22. PB1 is an elite basmati rice variety that is relatively sensitive to drought stress , whereas N22 is a deep-rooted drought tolerant variety . Young seedlings (2 weeks old) of both rice varieties were subjected to air dehydration for 1, 3 and 6 h before harvesting. Real-time RT-PCR analysis revealed interesting expression patterns between N22 and PB1. osa-miR398 showed a gradual decrease in transcript levels with an increasing duration of stress (Fig. 1A), whereas osa-miR319 rapidly decreased after 1 h of stress but later showed signs of recovery with an increased duration of stress (Fig. 1B). Nevertheless, the expression pattern of both these miRNAs was similar in the two rice varieties. By contrast, osa-miR171 remained unchanged in stress-treated PB1 seedlings; however, in N22, the transcript levels appears to markedly increase after 3 h of stress but again settled down after 6 h of drought stress (Fig. 1C). A more distinct variety-specific difference in the expression profiles was observed in the case of osa-miR408, which showed a distinct decrease in mature transcript levels in PB1. However, in N22, the levels remained significantly higher compared to the control, even after 6 h of dehydration stress (Fig. 1D). Thus, a clear variety-specific drought response was reflected by the transcript levels of osa-miR408. We further investigated the phenomenon and, because drought stress is most devastating if encountered prior to anthesis , we extended the study to assess the variety-specific stress response in mature field-grown plants subjected to drought stress at the heading stage. Both PB1 and N22 plants were grown in the field and drought conditions were simulated by withholding the water supply until the soil moisture decreased below 15%. Flag leaf and spikelet tissue were harvested from plants at the heading stage. Quantitative RT-PCR analysis revealed that transcript levels of osa-miR408 decreased significantly in both flag leaf and spikelet tissue of drought stress-treated PB1 plants, whereas they showed significant up-regulation in N22 plants under the same conditions (Fig. 2). Thus, miR408 demonstrated a variety-specific drought stress response at both the seedling and heading stages of rice development.
A variety-specific response is not apparent under an abiotic stress other than drought
Because miR408 showed a clear variety-specific drought response, we ventured to further characterize the behaviour of this miRNA in response to different abiotic stresses. We investigated whether the variety-specific response of miR408 is common to other abiotic stress conditions as well or whether it is specific to drought only. Accordingly, young seedlings of both rice varieties were subjected to various abiotic stress regimes. Analysis with respect to the response to salinity, high temperature and cold conditions reveals that mature transcript levels of miR408 increase gradually in response to all stress conditions (Fig. 3). The trend was similar in both rice varieties, indicating that, although the expression of miR408 is regulated under abiotic stress conditions in general, the variety-specific stress response is restricted under drought conditions only.
Drought tolerant rice cultivars resist the drop in miR408 levels under water deficit conditions
Comparative expression profiling of miR408 levels between PB1 and N22 indica rice cultivars clearly indicates that the levels of miR408 tend to remain high under drought stress conditions in the tolerant cultivar N22. Such a regulatory profile may be specific to N22 per se or could also be a general characteristic of drought tolerant rice varieties. To check this, we further profiled the expression of miR408 in IR64 and Vandana cultivars of indica rice. IR64, similar to PB1, is drought sensitive , whereas Vandana is a drought tolerant cultivar . Expression analysis in 3-week-old seedlings subjected to drought stress revealed that, in IR64, miR408 levels fall significantly under water stress in a fashion similar to PB1. However, in Vandana, the miRNA levels remain stabilized and do not decrease under water stress conditions. Thus, although the levels of miR408 increase under water stress in N22, the levels remain close to control conditions in Vandana, clearly indicating that both drought tolerant varieties resist the drop in the transcript levels of miR408 under water stress conditions (Fig. 4).
miR408-mediated regulation of plant specific blue copper-binding plastocyanin-like proteins is widely conserved across distant plant species. In rice, the plastocyanin-like domain-containing gene family comprises 62 members subdivided into uclacyanin-like proteins (UCLs); stellacyanin-like proteins (SCLs) and early-nodulin like proteins (ENODLs) . Out of these 62 members, 13 OsUCLs (OsUCL4, 7, 30 , 6, 15, 16 , 12, 29, 31 (starBase) , 5, 8, 9, 17 (PMRD) ) are reported as targets by different sources, of which OsUCL7 [5, 41], OsUCL9  and OsUCL30 have been validated by 5′ RACE [37, 38]. We investigated whether the observed variety-specific expression of miR408 is also reflected in the expression profile of its target genes. Indeed, a majority of the targets show a low accumulation in N22 compared to PB1 in flag leaf (Fig. 5A,B). A total of ten (OsUCL4, 6, 7, 8, 9, 12, 15, 16, 17 and 29) out of 13 predicted targets show a variety-specific response in flag leaf, whereas only four (OsUCL5, 7, 17 and 29) (Fig. 5C,D) show such a phenomenon in the spikelets, indicating a distinct tissue-specific response. The transcript levels of these genes were significantly decreased in flag leaf of N22. This clearly showed an anti-correlation with the increased levels of miR408 transcripts in similar tissues under similar conditions. By contrast, an increased accumulation of these transcripts in flag leaf of PB1 was observed in plants under similar drought regimes, which again demonstrates an anti-correlation with the reduced levels of miR408 in the corresponding tissue. Interestingly, OsUCL7, 17 and 29 show variety-specific accumulation in both flag leaf and spikelets. Because few target genes do not show variety-specific accumulation in response to miR408 accumulation, we investigated whether they are actually targeted by miR408 in N22. We performed 5′ RNA ligase mediated rapid amplification of cDNA ends (5′ RLM-RACE)  to verify whether OsUCL30 (LOC_Os08g37670) is targeted by miR408 in N22 as well. Indeed, OsUCL30 is targeted by miR408 (Fig. 5F); however, its transcript accumulation is stable under drought conditions and does not correspond to the variation in miR408 expression levels. On further analysis, we found that the transcript levels of OsUCL30 are under tissue-specific regulation under control conditions such that it has a higher expression level in flag leaf than in spikelets (Fig. 5E). This correlates perfectly with the higher abundance of miR408 transcript levels in spikelet compared to flag leaf (Fig. 5E). Other than plantacyanin, miR408 also targets laccase genes LAC3, LAC12 and LAC13 in Arabidopsis [27, 43]. There are 17 annotated laccase gene members in Arabidopsis [27, 44] and 28 members in rice (tigr, version 7.0, http://rice.plantbiology.msu.edu/). Although, the targeting of laccase by miR408 in rice is predicted by Archak and Nagaraju  and the psRNATarget database  (although with low probability), analyses performed with the help of degradome data (starBase database) , as well as prediction software (pmrd database) , do not support the targeting of laccases by miR408.
The MiR408 loci including the promoter is considerably conserved in rice varieties
Because miR408 appears to be regulated very dynamically in different varieties of indica rice, we aimed to determine whether the cultivar-specific accumulation of the mature miR408 transcript is a result of differential regulation of transcription or subsequent processing. Cis-elements in promoter regions have been broadly reported to be involved in the expression of stress-induced genes [47-50] and, thus, any difference in their occurrence may result in the distinct stress response of this miRNA. Accordingly, subsequent analyses were performed aiming to identify any significant sequence-based variation in the precursor, as well as the promoter region, of miR408 in different cultivars of rice (i.e. PB1, IR64 and Vandana) by cloning the specified region. The corresponding N22 reference sequence was obtained from the genome sequence available in the laboratory (S. Mathur and S. Raghuvanshi, unpublished data). Among the many cis-acting elements identified in the promoter region (Table S1), those motifs related to drought, salt, abscisic acid (ABA)-mediated response, copper response, HY5 binding sites, etc., are highlighted (Fig. S1). Some of the common drought, salt and ABA responsive cis-elements, such as MYBCORE, ACGTATERD1, MYCATRD22, MYCCONSENSUSAT, ABRELATERD1, CBFHV, MYB1AT, MYCATERD1, GT1GMSCAM4, AGCBOXNPGLB, PYRIMIDINEBOXHVEPB1, EECCRCAH1, WRKY710S and LTRECOREATCOR15, were identified in the promoter region using the PLACE database . A comparison of the promoter regions identified some indels between N22, Vandana, PB1 and IR64; however, none of the indels appeared to disrupt any known cis-acting element. Moreover, in Arabidopsis, miR408 is known to be regulated by HY5 (ELONGATED HYPOCOTYL 5), a positive regulator of photomorphogenesis . HY5 is a bZIP transcription factor that binds to several light-responsive cis-acting elements, including the Z-box and CT-box. Both these elements are well conserved in the promoter region of miR408 in N22, IR64, PB1 and Vandana. Similarly, a comparison of the precursor sequences in the four varieties revealed identical sequences, except for a substitution at position 146 bp in N22. Thus, the differential expression of miR408 in these rice varieties in response to drought conditions cannot be accounted for by differences in their precursor or promoter sequences.
Transcription factor OsSPL9 may contribute to the varietal regulation of miR408
osa-miR408 is known to be regulated by two transcription factors: HY5  and SPL7  in Arabidopsis. Because the binding sites of both transcription factors are considerably conserved, we considered that the observed variety-specific stress response might be a result of the differential accumulation of the transcription factors under drought stress conditions in both indica rice varieties. We checked the expression pattern of HY5 under similar drought stress conditions in adult field grown rice N22 and PB1 plants. We could not detect any significant variation in the transcript levels of HY5 under drought stress conditions in flag leaf of both rice varieties (Fig. 6A). Because HY5 is involved in the regulation of a light-mediated response and is known to bind miR408 promoter, we explored the light-regulated expression of miR408. Indeed, the expression of miR408 is significantly reduced in etiolated rice seedlings, indicating light-mediated regulation of miR408 expression (Fig. 6C). Thus, HY5 appears to mediate the light-responsive but not the drought regulated expression of miR408 in rice. By contrast, AtSPL7 controls the transcription of copper responsive miRs, including miR408 in Arabidopsis , in response to copper deficiency and sucrose levels  via binding to the GTAC motif. Promoter analysis of miR408 in N22 reveals the presence of 28 GTAC elements (Table S1). It has been reported that the rice orthologue of AtSPL7 is OsSPL9 . We explored whether the variety-specific opposite response of miR408 between N22 and PB1 could be a result of the differential expression of OsSPL9 in the two cultivars in response to drought. Expression analysis clearly revealed that OsSPL9 is up-regulated in N22 flag leaf (heading stage) compared to PB1 under a similar drought stress regime (Fig. 6B). Thus, the variety-specific drought stress response of miR408 may be mediated by the upstream transcription factor OsSPL9.
Drought tolerant and sensitive varieties have evolved distinct Ca2+-mediated regulation of miR408
The expression of miR408 is clearly modulated under most abiotic stress conditions. Moreover, the stress response appears to be variety-specific under drought stress, indicating that the expression of miR408 is under very dynamic regulation. This could partly be a result of the dynamic behaviour of upstream transcription factor OsSPL9. To better understand the regulatory mechanism, we studied the involvement of ABA and calcium in the regulation of miR408 transcription. Ca2+ is reported to be elevated in the cytosol upon perception of stresses such as cold, salinity and drought, as well as hormones such as ABA [55, 56]. Several Ca2+ responsive genes were found to have ABRE core- and classical cis-elements in the upstream promoter regions . Promoter analysis of miR-408 delineated several ABA responsive cis-acting elements (Table S1). Analysis with young rice seedlings treated with 100 μm ABA clearly showed that miR408 is up-regulated by ABA. However, the response to ABA is similar in both N22 and PB1 (Fig. 7A). We subsequently studied the involvement of cytosolic Ca2+ levels on the expression of miR408. An external supply of Ca2+ ions (CaCl2) resulted in the decrease of miR408 transcript levels in young N22 seedlings (Fig. 7C), whereas treatment with EGTA, which is a Ca2+ ion chelator, increases the transcript accumulation of miR408, indicating an active role of Ca2+ in the regulation of miR408 transcription. To further validate the role of Ca2+, several calcium channel blockers were used (Fig. 7B,C). Application of lanthanum chloride (LaCl3), a competitive inhibitor of Ca2+ channels and verapamil, an inhibitor of the L-type voltage-dependent calcium channels located on the plasma membrane of cells, leads to an increase in the transcript accumulation of miR408. The effect of LaCl3 appears to be more effective than verapamil (Fig. 7B). Similarly, application of the inhibitors blocking internal Ca2+ channels [ruthenium red, which blocks flux through mitochondrial membranes, and lithium chloride (LiCl), which blocks the inositol trisphosphate gated endoplasmic reticulum-located channels] also leads to an increase in the expression of miR408. Thus, both internal and external stores of calcium appear to be involved in the regulation of miR408 transcription in N22. Further exploration reveals that Ca2+ involvement also appears to be variety-specific, with PB1 showing no significant response to an external supply of Ca2+ ions or various inhibitors. This difference in the response to Ca2+ might be a result of the additional components of the signalling pathway, such as the intermediate kinases, phosphatases, transcription factors, etc. The difference observed above is further reflected in a subsequent experiment where the involvement of Ca2+ in the drought-mediated modulation of miR408 transcript accumulation is checked. To accomplish this, we treated young rice seedlings with various Ca2+ channel inhibitors and then subjected them to air dehydration. The data obtained reveal that pre-treatment with Ca2+ channel blockers manages to reduce the stress-induced expression of miR408, thereby establishing that Ca2+ is at least partly responsible for the stress-induced expression of miR408 in N22 (Fig. 7D). A similar analysis performed on PB1 revealed that pre-treatment with calcium channel blockers followed by air dehydration is unable to stop the dehydration stress-induced decrease in the miR408 expression (Fig. 7E). Thus, it could be postulated that the cytosolic calcium level might be involved in the control of the dehydration-induced response of miR408 transcript accumulation only in N22 and not in PB1.
The response to abiotic stress conditions is a complicated phenomenon involving the interplay of numerous genetic elements. Although all plants respond to stress conditions, only a few are able to elicit a robust response. To identify the miRNA genes that may play pivotal role in regulation of drought response, we undertook comparative expression profiling in drought tolerant and sensitive indica rice varieties. This enabled us to identify at least one miRNA (i.e. miR408) that is under differential drought stress-mediated regulation in a tolerant variety compared to sensitive varieties. miR408 is also considered to be a powerful modulator of vegetative growth and its constitutive expression in the spl7 mutant background was able to rescue growth defects in Arabidopsis . Previous studies have shown that miR408 is up-regulated in response to drought in Medicago truncatula [59, 60], Arabidopsis  and Hordeum vulgare  but down-regulated in rice cv. IRAT109 . However, an early up-regulation followed by decreased levels in miR408 expression over time was reported in rice japonica var. Zhonghua 11 . In rice, miR408 has a unique expression profile specific to indica rice varieties that are relatively tolerant to drought conditions. At both the young and adult stages, miR408 transcript levels stay very close to control levels or are slightly up-regulated in N22, which is a drought tolerant variety. However, under similar conditions, transcript levels are significantly down-regulated in the sensitive variety PB1. The contrast is more apparent in the spikelet tissue at the heading stage where N22 is able to stabilize miR408 levels but, in PB1, levels decrease significantly. Similarly, in IR64, which is also a sensitive variety, miR408 transcript levels fall appreciably under drought stress. By contrast, in Vandana, which is a drought tolerant variety similar to N22, the transcript levels do not fall but remain at almost control levels. Thus, in both sensitive varieties, the levels of miR408 fall significantly, whereas, in tolerant varieties, the levels are either elevated or close to control levels, indicating that both of the tolerant varieties tend to resist the drop in expression of miR408. Because the expression profiles have been validated in multiple developmental stages (young and adult plants), as well as in multiple tissue (young leaf, flag leaf and heading spikelet), confidence with respect to the unique expression pattern of miR408 is significantly high. Similar stress tolerance biased expression levels of miR408 were also reported in Zea mays to short-term waterlogging . Furthermore, such a varietal difference was found to be specific to drought stress and not apparent in other abiotic stresses, such as salinity, heat, cold, etc. The increased expression levels of miR408 in both PB1 and N22 in response to salt and cold are consistent with the findings of previous studies conducted in Vigna and Arabidopsis . The results of the present study are also in accordance with previous studies where miR408 was shown to be induced in response to 3 h of salt stress (300 mm) in Populus tremula .
Plastocyanin-like copper containing protein family members are well characterized conserved targets of miR408. miR408 associate with both AGO1 and AGO2 in a redundant manner to mediate plantacyanin regulation in Arabidopsis . MiR408-mediated regulation of plantacyanins is widely explored in response to copper and sucrose in Arabidopsis where the dynamics of miR408 contributes towards the maintenance of copper homeostasis [27, 28]. In addition to plantacyanin-like, miR408 is also known to target other copper proteins, such as copper P1B-ATPase in Medicago truncatula . In addition to its function under abiotic conditions, miR408 also plays a role in mechanical stress because it is induced by short-term tension and compression stresses in xylem tissue in poplar . The variety-specific dynamism of miR408 expression pattern is also reflected in the steady-state levels of the transcripts of its target genes. In both flag leaf and spikelet tissue, several target genes show a reduced accumulation in N22, whereas they maintain relatively high levels in PB1, which corresponds to miR408 levels by way of anti-correlation. At2 g02850, an orthologue of one of the target genes, OsUCL7 (LOC_Os03g15340) , which shows dehydration stress-induced variety-specific accumulation in both flag leaf and spikelet (Fig. 5A–D), has been studied in Arabidopsis. It has been reported that over-expression of the orthologous plantacyanin gene (At2 g02850) results in a reduced level of seed-set as a result of a lack of anther dehiscence . Studies are in progress in the laboratory aiming to explore whether stabilizing the levels of plastocyanin-like protein via miR408 comprises one adaptation that allows the drought tolerant varieties to set seeds even under adverse conditions.
Another interesting observation was the extent of correlation of miR408 levels with that of its target genes. In flag leaf, although most targets showed expected anti-correlation, only four targets showed anti-correlation in spikelets under drought/dehydration stress-regulated expression. We explored the phenomenon further and demonstrated that at least one of the target genes not showing anti-correlation under stress-regulated expression is still cleaved by miR408, although the transcript levels do not modulate according to stress-regulated miR408 levels. However, the correlation is true when assessed under tissue-specific regulatory cues. This indicates the complex nature of the relationship between miRNA and its target genes.
Because we could clearly establish that osa-miR408 is regulated quite distinctly in the tolerant and sensitive rice varieties, we further explored the molecular basis of this variability. Variation in the mature miRNA transcript accumulation could be regulated at the transcriptional level and also at the post-transcriptional level through the regulation of its processing. To characterize regulation at the transcriptional level, we compared the putative promoter region of miR408 in all four rice cultivars. Analysis of osa-miR408 genomic loci, including the precursor and upstream 2-kb sequence in N22, PB1, IR64 and Vandana, does not reveal any significant sequence variation among the varieties. This clearly indicates that the variety-specific response of miR408 is a result of some upstream signal transduction components. Interestingly, the expression of miR408 is known to be regulated by two transcription factors (HY5 and SPL7) in Arabidopsis. Although HY5, a b-ZIP transcription factor, is a known regulator of photomorphogenesis, some SPL member genes are known to be involved in juvenile to adult phase transition in Arabidopsis via miR156. Both of these transcription factors appear to regulate different aspects of miR408 expression. HY5 regulates light-regulated expression, whereas OsSPL9 may be involved in the drought-regulated response. We found that drought stress induces the expression of OsSPL9 in N22, whereas the levels fall significantly in PB1. This presents an interesting scenario where both the transcription factor (OsSPL9) and the downstream target gene (miR408) show a variety-specific drought stress response. Moreover, the phenomenon is only restricted to drought stress conditions and is not common to abiotic stress in general because both N22 and PB1 showed up-regulation on heat, cold and salt stress in a similar fashion.
The data obtained in the present study clearly indicate that, from the perspective of miR408, different rice varieties appear to have evolved different gene expression regulatory mechanisms. To clearly understand the regulatory schema controlling miR408 expression, we explored the role of ABA and cytosolic Ca2+ in the regulation of its transcription. Indeed, ABA positively regulates the expression of osa-miR408 in both N22 and PB1, indicating regulation by the ABA-dependent stress signal transduction pathway. The observed up-regulation of osa-miR408 in response to ABA was consistent with the findings reported in rice  and P. tremula . With the help of calcium channel inhibitors, it could be clearly established that both internal and external calcium stores are responsible for the calcium-mediated response of miR408. A stress regime including calcium channel inhibitors followed by water-deficit clearly demonstrated that Ca2+ is at least partially involved in the drought stress-regulated accumulation of osa-mir408 mature transcript. Although cytosolic Ca2+ apparently signals to down-regulate miR408 in N22, it has no effect on its expression in PB1. Thus, both the varieties appear to have evolved different Ca2+-mediated signal transduction mechanisms for regulation of the expression of the same gene (miR408). Further studies are required to identify the exact nature of variation in the Ca2+ signal transduction cascade. Because the response to external ABA and Ca2+ application is inverse, it appears that the response to ABA is Ca2+-independent. ABA regulation of gene expression via Ca2+-dependent and independent pathways has been reported previously [69, 70]. We suggest a possibile difference in the stress response of miR408 via differential signal transduction mediated by Ca2+ and its downstream interacting factors, such as CaM/CML/CAMTAs. The involvement of as yet unidentified proteins regulated by ABA and Ca2+ independently is another possible likelihood. For example, CDPK/CPK proteins that respond to a rise in cytosolic Ca2+ may form alternate complexes with different calcineurin B-like proteins (inducible or non-inducible by ABA) rendering ABA-independent and ABA-dependent pathways . Similarly, there is evidence for ABA-dependent stress responses that are not channelled via Ca2+ changes but, instead, are routed through different proteins, such as kinases [69, 72]. The differential expression or structural/biochemical properties of these signal transduction components in PB1 during the course of evolution may explain why miR408 does not respond to elevated [Ca2+]cyt levels.
In conclusion, by performing comparative expression profiling between the sensitive (PB1 and IR64) and tolerant (N22 and Vandana) indica rice varieties, we were able to demonstrate the evolution of distinct signal transduction schema for the regulation of miR408 expression under drought conditions in different indica rice varieties. However, further studies are required to directly implicate miR408 in the regulation of the drought stress response. We consider this phenomenon to have a wider occurrence in nature and studies in our laboratory have been initiated aiming to identify the drought-specific regulatory schema controlling the regulation of miRNAs in rice.
Materials and methods
Plant growth conditions and stress treatment
Seeds of Oryza sativa L. ssp. indica cultivar PB1, N22, IR64 and Vandana were sterilized as described previously  and soaked in reverse osmosis water overnight. Seeds were grown on muslin cloth tied over a tray containing rice growth medium for 2 weeks in a culture room at 28 ± 2 °C under a 16 : 8 h photocycle. For stress treatments, 1-week-old seedlings were transferred into pots containing an equal proportion of Soilrite (Keltech Energies Ltd., Bangalore, India) and soil supplied with rice growth medium. Three-week-old seedlings were subjected to salinity stress (200 mm NaCl in medium), cold stress (4 °C) and heat stress (42 °C). Shoots were harvested from control and stressed seedlings at 30 min and at 1, 3, 6 and 24 h. One-week-old seedlings were treated for 3 h with ABA (100 μm). For analysis of Ca2+ signalling, treatments were given with CaCl2 (10 mm), EGTA (10 mm), LaCl3 (5 mm), LiCl (5 mm), verapamil (100 μm), ruthenium red (100 μm), LaCl3 (5 mm) + verapamil (50 μm), LiCl (5 mm) + ruthenium red (100 μm) for 3 h. For the etiolation experiments, seeds were sterilized and grown for 7 days in the absence of light in an enclosed chamber and shoots were harvested in the dark. Drought treatment was also given to field grown mature plants of PB1 and N22 by withholding the water supply for 12 days. Soil moisture was monitored using a Hydra Probe Soil Moisture Sensor (Stevens Water Monitoring Systems Inc., Portland, OR, USA). Flag leaf and spikelet at heading stage were collected from control plants and drought-treated plants that showed a leaf rolling phenotype with a soil moisture content reaching below 15%. Chemicals used were obtained from Sigma-Aldrich (Munich, Germany).
Total RNA was extracted from the harvested tissues using TRI Reagent (Sigma-Aldrich) followed by DNaseI treatment (Fermentas, Hanover, MD, USA). Small RNA samples enriched using 4 m LiCl were polyadenylated using a Poly(A) Tailing kit (Ambion, Austin, TX, USA) and 2 μg of each sample was reversed transcribed with miR_oligodT_RTQ  primer for first-strand cDNA synthesis using SuperScript II Reverse Transcriptase (Invitrogen, Calsbad, CA, USA). To analyze the expression of the mature miRNAs, Quantitative RT-PCR was performed using TaqMan Fast Universal PCR Master Mix (Applied Biosystems, Foster City, CA, USA) with RTQ universal reverse primer , miRNA specific forward primer and fluorogenic probe in a StepOnePlus Real-Time PCR System (Applied Biosystems) in accordance with the manufacturer's instructions. The miRNA expression level was normalized using 5S as an endogenous control. The data are reported as the mean of five or six biological replicates with at least three technical repeats for each biological repeat. To analyze the abundance of miRNA target genes, DNaseI-treated total RNA samples were reversed transcribed with oligodT primer using SuperScript II. Real-time PCR was performed using Fast SYBR Green Master Mix (Applied Biosystems) with actin as a normalizing control.
5′ RLM-RACE for target validation of OsUCL30 (LOC_Os08g37670) was performed as described previously . The mRNA population enriched from total RNA using an Oligotex mRNA mini kit (Qiagen, Hilden, Germany) was ligated with 5′-RNA adapter using RNA T4 Ligase (New England Biolabs, Hitchin, UK). Ligated mRNA molecules were reversed transcribed with 3′-adapter (dT) primer using SuperScript II. The resulting cDNA was used for RACE with 5′-adapter primer and gene-specific primer and the amplicon thus obtained was cloned in pUC19 and at least six colonies were sequenced to identify the cut site of miR408 in UCL30.
Promoter and precursor cloning
2-kb upstream regions from the precursor start site of miR408 along with precursor of length 213 bp were PCR amplified using TaKaRa Ex Taq Polymerase (TaKaRa, Tokyo, Japan) cloned from PB1 in pUC19 (Invitrogen) and from IR64 and Vandana in pGEMT vector (Promega, Madison, WI, USA). At least three clones for each were confirmed by sequencing. The 2-kb promoter region of miR408 was also taken from the in-house generated genome sequence of N22. clc main workbench (CLC bio, Aarhus, Denmark) was used for the alignment of promoter sequences. All primer sequences are provided in Table S2.
This work was supported by the Department of Biotechnology, Government of India. We also thank Dr A. K. Singh (Division of Genetics, Indian Agricultural Research Institute, New Delhi, India) for kindly providing us with the seeds of different rice varieties.