Expression of miR394 under salt and drought stresses and ABA treatment
To validate expression of miR394 in response to salt and drought stresses, two week-old Arabidopsis seedlings were exposed to 300 mM salt (NaCl), drought (desiccation) and 100 μM ABA for 6 and 12 h. RNA gel-blot analysis showed that miR394 was slightly induced by NaCl and drought (Figure 1A). Using RT-PCR, we further analyzed whether expression of miR394a and miR394b was also affected by salt and drought exposure. It is shown that both miR394a and miR394b were induced under saline and drought stresses (Figure 1B; Additional file 1: Figure S1). To further investigate expression of miR394a/b, its promoter sequences were retrieved, fused to the β-glucuronidase (GUS) reporter gene, and introduced into transgenic plants. The histochemical analysis showed that heavy GUS staining was detected in pMIR394a/b::GUS seedlings under salt and drought stresses, whereas the relatively light GUS staining was observed in the control seedlings (Figure 1C). The phytohormone ABA mediates plant response to abiotic stresses [1]. Examination of miR394 response to ABA treatment revealed that miR394 was mildly induced (Figure 1A). Similarly, the abundance of both pre-miR394a and pre-miR394b transcripts as well as pMIR394a/b::GUS staining were also slightly enhanced by ABA (Figure 1B,C; Additional file 1: Figure S1). Feedback loops in which miRNA-regulated genes may regulate the transcription of their miRNA have been described before [23]. Our analysis showed that miR394 transcripts were not changed in the 35S:mLCR and lcr mutant plants (Figure 1D); also, quantitative real-time PCR showed no significant deference between 35S:mLCR or lcr mutant plants and wild-type (Additional file 1: Figure S2), indicating that miR394 is not feedback-regulated by LCR.
Expression of LCR under salt and drought stresses and ABA treatment
Although miR394 is induced by salt and drought, whether LCR is regulated by ABA and the abiotic stresses is unknown. To investigate the expression pattern of LCR, we generated constructs encoding a GUS reporter protein and LCR fusion protein under the control of LCR promoter sequence (1.45 kb) and transformed into Arabidopsis. Histochemical staining showed that pLCR::GUS was active at all development stages (Figure 2A). To test the regulation of LCR promoter by abiotic stresses, two week-old seedlings were exposed to NaCl, drought and ABA. Treatments with salt and drought induced higher GUS activity in pLCR::GUS plants compared with the control (Figure 2B). To confirm the expression pattern of GUS staining under the abiotic stresses, we comparatively analyzed the transcripts of GUS in pLCR::GUS plants and LCR in wild-type using qRT-PCR. As shown in Figure 2C,D, the transcript levels of GUS and LCR were up-regulated 6 and 12 h after NaCl and drought treatments. But the GUS level was higher than the LCR level. This suggests that LCR mRNA was partially silenced by miR394 in plants under the stress condition. To investigate whether LCR at a translational level was affected by salt and drought, a western blot study was carried out. Using a monoclonal antibody specifically interacting with LCR protein (52 kDa), we showed that the LCR proteins could be also induced by salt and drought treatments (Figure 2E). Unexpectedly, less GUS staining was detected in seedlings with ABA treatment (Figure 2B). Also, both LCR transcript and protein abundance with ABA was found to be lower than the control (Figure 2C,D,E).
Identification of 35S::MIR394, 35S::m5LCR and lcrmutant lines
To identify the role of miR394 in the stress responses, we examined the transgenic lines over-expressing MIR394a/b (35S::MIR394a/b) and the cleavage-resistant version (35S::m5LCR) [16]. The Western blotting on LCR proteins was analyzed. The abundance of LCR proteins was lower in 35S::MIR394a plants but was higher in 35S::m5LCR plants compared with the wild-type (Figure 3A), indicating that miR394 affects the abundance of LCR proteins.
The genomic sequence of LCR contains two exons (1404 bp), which are interrupted by an intron towards 5′ end; the CDS contains an open reading frame coding for a protein with 467 amino acid residues (Figure 3B). Two independent T-DNA insertion lines, lcr-1 (SALK_016763c) and lcr-2 (SALK_136833c) were used. The mutant lcr-1 has a T-DNA insertion at 5′ UTR, whereas lcr-2 has an insertion in the intron towards the first exon. The two mutants were verified by diagnostic PCR using LCR gene-specific and T-DNA border primers (Figure 3C,D). The lcr-1 and lcr-2 alleles were confirmed by RT-PCR and Western blot, by which the expression of both LCR mRNA and proteins were severely suppressed (Figure 3E,F).
MIR394a over-expressing and lcrmutant plants display salt sensitivity
Plant seeds were placed on the solid MS medium supplemented with 0–150 mM NaCl. The seed germination and post-germination growth responded differently to 50–100 mM NaCl after 4 d treatment (Figure 4A). Following the treatment with 100 mM NaCl, the germination percentages of 35S::MIR394a-1, 35S::MIR394a-2, lcr-1 and lcr-2 seeds were only 10-20%, whereas those of wild-type were 78%. Simultaneously, the cotyledon greening was blocked by 50 or 100 mM NaCl in 35S::MIR394a or lcr plants. Under the same condition, the germinating rate of 35S::m5LCR seeds was up to 94%, and the percentage of cotyledon greening in 35S::m5LCR seedlings was higher than that of wild-type (Figure 4B). Root growth of 35S::MIR394a and lcr seedlings was strongly inhibited by 50–100 mM NaCl relative to the wild-type, but compared with 35S::MIR394a, the root growth of lcr seedlings was more severely inhibited (Figure 4C,D,F,G). By contrast, 35S::m5LCR plants had higher tolerance of roots to NaCl stress. The root elongation of 35S::m5LCR plants with 150 mM NaCl was increased by 30-40% compared with wild type (Figure 4E,H). These results indicate that both lcr mutant and 35S::MIR394a plants are hypersensitive, whereas 35S::m5LCR plants are tolerant to the salt stress.
MIR394a over-expression and lcrmutant plants display drought tolerance
Under the water loss condition, most of wild-type, 35S::MIR394a and mutant plants were withered (Figure 5A,B). After a three-day re-watering, 47-56% of 35S::MIR394a and 68-71% of the lcr mutant plants resumed their growth, whereas only 26% of the wild-type plants survived (Figure 5D). By contrast, the 35S::m5LCR plants were severely wilted after dehydration stress (Figure 5C). After re-watering for 3 d, there was only 18-23% survival (Figure 5E), indicating that the 35S::m5LCR plants are supersensitive to drought stress.
We further examined plant response to drought stress using detached rosette leaves of two week-old seedlings, which were placed on a open-lid petri dish under dim light at room temperature. The reduced fresh weight of leaves was measured over time (0–150 min). Leaves from all leaves showed progressive loss of weight (Figure 5F). To the end of experiment, the fresh weight for 35S::m5LCR plants was only 58-60% of their starting value, and the fresh weight for 35S::MIR394a and lcr mutant plants was 73-78%. The wild-type had an intermediate weight of 67%. From these results it is shown that while 35S::m5LCR plants were hypersensitive to water loss, 35S::MIR394a and lcr mutant plants displayed higher tolerance to drought stress.
MIR394a and LCR are involved in ABA-dependent seed germination and root growth
For 35S::MIR394a and lcr plants, seed germination and cotyledon development were severely inhibited by 0.25-2 μM ABA (Figure 6A-H), but seed germination of 35S::m5LCR plants was unaffected by ABA (even at 2 μM) (Figure 6I-K). Nevertheless, 81-85% of 35S::m5LCR seedlings were able to develop true green cotyledons in the presence of 0.5 μM ABA, with only 18% of wild-type was presented (Figure 6L). A concentration-dependent change was observed, in which the ABA-hypersensitive response of 35S::MIR394a and lcr plants occurred at 0.25 μM of ABA, and when ABA concentrations were up to 1 μM, the cotyledon growth was completely blocked (Figure 6D,H). By contrast, 35-40% of 35S::m5LCR seedling cotyledons were still green and even expanded at 1 μM of ABA, where the growth of wild-type was arrested (Figure 6L).
Consistent with seed germination and cotyledon growth, root elongation of 35S::MIR394a and lcr seedlings was hypersensitive to ABA (Figure 7A,B). The sensitive concentration occurred at 0.5 μM, where the root length of 35S::MIR394a and lcr was only 52.9-64.8% and 46.1-49.8% of wild-type, respectively (Figure 7D,E). Under the same condition, however, the elongation of 35S::m5LCR root showed more resistance to ABA treatment (Figure 7C). At 0.5 and 1 μM of ABA, the root length of 35S::m5LCR plants was 1.44-1.54 and 1.81-1.95 folds over the wild-type, respectively (Figure 7F). We generated transgenic lines over-expressing a modified version of IPS1 (INDUCED BY PHOSPHATE STARVATION 1), which encodes a non-coding RNA [24]. This RNA comprises a short motif highly complementary to miR394 with a small loop at the expected cleavage position, which can sequester miR394, partially release its natural target, and increase LCR abundance [16]. We used target mimicry miR394 (MIM394) transgenic lines to test ABA response in plants and found that phenotype of root growth is very similar to that of 35S::m5LCR plants (Additional file 1: Figure S3). Thus, MIR394 over-expressing plants were hypersensitive to ABA, whereas 35S::m5LCR plants had a converse phenotype with ABA.
MIR394a and LCR are involved in ABA-dependent stomatal closure
Leaves of 4 week-old plants were submerged in stomatal opening solution and treated with ABA at 0, 1, and 10 μM for 2 h. In the absence of ABA, all guard cells on leaves of plants were fully opened (Figure 8A). When 1 and 10 μM ABA was added to the solution, the stomata on the leaves of wild-type and 35S::MIR394a plants were closed. Quantitative analysis using stomatal aperture (the ratio of width to length) showed that there was much stronger stomatal closure in 35S::MIR394a plants than in wild-type with 1 μM ABA (Figure 8B). Conversely, the 35S::m5LCR leaves were insensitive to ABA-induced stomatal closure and most of guard cells were open in the presence of 1 or 10 μM ABA. The stomatal apertures were 1.4- (1 μM ABA) and 5.6- (10 μM ABA) fold over those of wild-type, respectively.
Reactive oxygen species (ROS) is one of the essential signal molecules involved in abscisic acid (ABA)-induced stomatal closure [25]. Leaves of four week-old plants were exposed to 0 or 50 μM ABA for 2 h and treated with nitroblue tetrazolium (NBT) or 3,3’-diaminob enzidine (DAB). Compared with wild-type, leaves of 35S::MIR394a plants exposed to 50 μM ABA were stained intensively, whereas those of 35S::m5LCR plants were stained light with O2–. and H2O2 (Figure 8C,D). These results indicated that ABA-induced ROS abundance could be altered in 35S::MIR394a and 35S::m5LCR plants.
Over-expression of miR394a cannot rescue the ABA insensitivity of abi4-1 and abi5-1
To gain insights into involvement of miR394 in ABA signaling pathway, abi4-1 and abi5-1 mutant plants were individually crossed to the 35S::MIR394a plants. Both ABA-INSENSITIVE (ABI) 4 and ABI5 encode the major transcription factors which act as positive regulators of ABA response [4, 7]. Because 35S::MIR394a and abi4-1 or abi5-1 mutant plants display the opposite phenotypes, analysis of the crossed plants can elucidate the relationship between the genes. preMIR394a were expressed in the abi4-1/35S::MIR394a and abi5-1/35S::MIR394a plants (Figure 9A). Examination of the two cross plants revealed that the phenotypes of abi4-1/35S::MIR394a and abi5-1/35S::MIR394a plants were nearly identical to the phenotype of abi4-1 and abi5-1 plants, respectively [6, 26, 27]. In the presence of 0.5 or 1 μM ABA, both abi4-1/35S::MIR394a and abi5-1/35S::MIR394a plants showed higher seed germination rates, enhanced cotyledon development, and increased root elongation (Figure 9B-E). These results indicate that over-expression of MIR394a in abi4-1 or abi5-1 background resulted in loss of ABA-sensitivity in 35S::MIR394a plants. The data suggest that miR394a does not function in parallel to ABI4 and ABI5, but both ABI4 and ABI5 may work downstream of miR394a.
Expression of ABA and stress responsive genes in 35S::miR394a and 35S::m5LCRplants
To investigate whether ABA and stress-responsive genes could be regulated by miR394 or LCR over-expression, two categories of genes were tested. The first group contained five genes (ABI3, ABI4, ABI5, ABF3, and ABF4) coding for ABA-responsive basic leucine zipper (bZIP) transcription factors that bind to the ABA response element (ABRE) of their targets and function during the plant development and stress responses [4]. After treatment with 100 μM ABA for 6 h, all genes were induced in 35S::MIR394a plants (Figure 10A). However, expression of ABI3, ABI4, and ABI5 were down-regulated in 35S::m5LCR plants.
The second group genes RD29A, KIN1, and RD22 belong to the DRE/CRT (drought responsive/C-repeat) elements-containing class of stress- and ABA-responsive genes [28]. Another gene P5CS encodes enzyme D1-pyrroline-5-carboxylate synthetase involved in proline biosynthesis and ABA or abiotic stress responses [29, 30]. With a similar pattern as ABI4 or ABI5, all these genes RD29A, KIN1, RD22, and P5CS were induced in 35S::MIR394a plants. By contrast, expression of RD29A, KIN1, RD22, and P5CS was down-regulated in 35S::m5LCR plants although the degree of the gene depression varied differently (Figure 10B). These results indicated that miR394 over-expression can modify transcription of ABA- and stress-responsive genes.