The Arabidopsis Columbia (Col-0) ecotype wild-type (WT), srp1-1 (SAIL_187_C10), srp1-2 (SAIL_312_E06), aba3-1 (CS157), and 35S::SRP1 transgenic lines were used in this work. The mutants were obtained from Arabidopsis Biological Resource Center¹ (ABRC). 35S::SRP1 transgenic lines were generated by Agrobacterium-mediated transformation in this study. The seeds were sown on 1/2 Murashige and Skoog (MS) medium with 3% (w/v) sucrose and 0.8% agar in a light homoeothermic incubator (24 °C) with 16 h light/8 h dark. After 3 days, the germinated seeds were transferred into pot (R = 3 cm, H = 8 cm) filled with nutritional soil and vermiculite (1:1) and grown under 50 μmol m^-2 s^-1 light at 22°C and 16 h light/8 h dark in a greenhouse. The seeds harvested at the same time were used for the experiment.

One hundred seeds each line were sown in the plate (10 cm × 10 cm × 1 cm) with 1/2 MS medium containing different concentrations of ABA or NaCl in triplicates. The plates with seeds were stored at 4°C in dark for four days, and subsequently incubated in the growth chamber (22°C) with 16 h light/8 h dark. Germination and green cotyledon rates were scored at the indicated times after the incubation. Statistical analysis was performed with three biological replicates. Seeds were counted as germinated when the radicle tip had fully penetrated the seed coat (radicle protrusion) as previously described (Liu et al., 2013).

The coding sequence of SRP1 was cloned into binary vector pCAMBIA1300s which was further transferred into Agrobacterium tumefaciens strain EHA105. The Agrobacterium-mediated transformation in Arabidopsis was conducted as previously described (Clough and Bent, 1998). Transgenic plants were screened in hygromycin (25 mg/L)-containing medium and the positive seedlings were randomly selected to be verified by the primers for the hygromycin resistance gene (Supplementary Table 1). Mutants srp1-1 (SAIL_187_C10) and srp1-2 (SAIL_312_E06) for the SRP1 gene were identified using the primers LP, RP and pCSA110LB for amplification of T-DNA insertion. The mutants or the transgenic line were further verified by semi-quantitative RT-PCR or qRT-PCR analysis. The primers were listed in Supplementary Table 1.

Total RNA were prepared from WT, srp1 and transgenic lines as previously described (Gutierrez et al., 2009). Approximately 5 μg of RNA were treated with DNase I using a DNA free Kit (Thermo Fisher Scientific, Waltham, MA, USA), and cDNA was synthesized by reverse transcribing total RNA using SuperScript III reverse transcriptase (Invitrogen, Foster City, CA, USA) with oligo (dT)₁₂ primer according to the manufacturer’s instructions. Transcript levels of ABA signaling-related genes RD29A (AT5G52310), RD29B (AT5G52300), ABI1 (AT4G26080), ABI2 (AT5G57050), ABI3 (AT3G24650), ABI5 (AT2G36270), EM1 (AT3G51810), and EM6 (AT2G40170) were assessed by real-time quantitative reverse transcription PCR (qRT-PCR). Arabidopsis Actin7 was used as the internal control for the qRT-PCR, respectively (Liu et al., 2007; Yelagandula et al., 2014). The primers were listed in Supplementary Table 1.

The transient expression vector was constructed by cloning SRP1 into pA7-GFP to fuse with GFP. The pA7-GFP vector was used as a control. The vectors were expressed transiently in Arabidopsis leaf protoplasts using the PEG (polyethylene glycol) method (Abel and Theologis, 1994). DAPI (4′,6-diamidino-2-phenylindole) was used to dye the nucleus. The visualization of SRP1-GFP or GFP was observed using the Leica SP2 confocal laser scanning microscope.

The SRP1 coding sequence was cloned into the expression vector pGEX-2T to generate GST-labeled SRP1 protein. The fusion protein was prepared according to the GST Gene Fusion System Handbook (Amersham Biosciences, Piscataway, NJ, USA). The 3′UTR sequences of ABI2/3 dsDNA were cloned with the primers containing the T7 promoter sequence (Rohde et al., 2000; Ohta et al., 2003). The 3′UTRs of ABI2/3 ssRNA was labeled using T7 RNA polymerase and biotin-16 UTP according to the instruction manual (Thermo Fisher Scientific, Waltham, MA, USA). The binding reactions were performed by incubating biotin-labeled RNA with recombinant protein at 23°C for 30 min in 10 μl of binding buffer containing 20 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES) (pH 7.5), 50 mM NaCl, 5 mM 2-mercaptoethanol, 1 mM DL-Dithiothreitol (DTT), 5 μM ZnCl₂ and 4 mM MgCl₂ as described (Qu et al., 2014). Electrophoresis was performed at 4°C in 8% native TRIS borate-acrylamide (29:1) gels, which were transferred to N+ nylon membrane (Roche, Basel, Swiss). The membrane was exposed using the Chemiluminescent Biotin-labeled Nucleic Acid Detection Kit (Thermo Fisher Scientific, Waltham, MA, USA), and scanned by ChemiDoc XRS+ imager (Bio-Rad, Irvine, CA, USA). For competition assays, biotin-labeled RNA and recombinant protein were incubated prior to the addition of unlabeled competitor RNA. The sequences of ABI2 and ABI3 3′UTRs were listed in Supplementary Table 2.

In protoplast system analyzing the effect of SRP1 on the stability of mRNA, the control reporter vector 3× DRE-LUC was constructed by inserting three copies of dehydration response element (DRE, TACCGACAT) and a minimal TATA region of 35S promoter of Cauliflower Mosaic Virus (CaMV) into the upstream of the firefly luciferase (LUC) in pGL3-basic vector. The reporter vectors 3× DRE-LUC-ABI2 3′UTR and 3× DRE-LUC-ABI3 3′UTR were constructed by insertion of ABI2 and ABI3 3′UTR into the downstream of luciferase coding sequence. Effector vectors 35S-DREB1A and 35S-SRP1 were constructed by insertion of the DREB1A (AT4G25480) or SRP1 coding region into PA7 vector (Narusaka et al., 2003). The pPTRL (Renilla reniformis Luciferase driven by 35S promoter) was used as the internal control. The effectors, reporters and internal control were co-transfected into Arabidopsis protoplasts. After culturing for 16 h, the luciferase assays were performed with the Dual-Luciferase Reporter Assay System (Promega, Madison, WI, USA) and quantitated by the GloMax 20/20 luminometer (Promega, Madison, WI, USA). Statistical analysis was performed with three biological replicates.

To analyze the transactivation activity of SRP1, the assay in Arabidopsis protoplast system was performed as previously described (Fang et al., 2015). The reporter vector was constructed by inserting four copies of GAL4 binding element and a minimal TATA region of 35S promoter of CaMV into the upstream of the firefly LUC in pPCV814 vector. The control effector vector 35S-Gal4DB was constructed by inserting the GAL4DB into pRTL2-GUS vector. The SRP1 was cloned into control effector vector to generate 35S-Gal4DB-SRP1 as the effector. The pPTRL was used as the internal control. The effectors, reporters and internal control were co-transfected into Arabidopsis protoplasts. After culturing for 16 h, the luciferase assays were performed with the Dual-Luciferase Reporter Assay System (Promega, Madison, WI, USA) and quantitated by the GloMax 20/20 luminometer (Promega, Madison, WI, USA). Statistical analysis was performed with three biological replicates.

The transactivation activity assays in the yeast were performed using the Matchmaker GAL4 Two-Hybrid System 3 (Clontech, Mountain View, CA, USA). BD-SRP1 was constructed by cloning the full-length CDS of SRP1 into pGBKT7 to fuse with the GAL4 binding domain. The construct and the empty vectors were transformed into the yeast strain AH109. The yeast cells were plated onto synthetic dextrose (SD)/-Trp medium to confirm the transformation. At the same time, the yeast cells were plated onto SD/-His/-Trp/-Leu containing 3 mM 3-amino-1,2,4-triazole medium to screen for transactivation activity.

Previously we reported a stress repressive zinc finger protein SRZ1 in rice (Huang et al., 2008). In present study, we studied the function of SRP1, the homolog of SRZ1 in Arabidopsis. The SRP1 (AT2G17975) gene possesses a translated region of 807-bp that encodes a protein containing 268 amino acids, with a calculated molecular weight of 30.24 kDa. On the basis of conserved domain analysis, four C2C2-type ZnF motifs were found in SRP1 and homologous to mammalian RanBP2-type ZnF domains (Nguyen et al., 2011; O’Connell et al., 2012; Vandevenne et al., 2014) (Figure 1A). The homologous proteins of SRP1 were found widely distributing among plant species, and they contained three or four C2C2-type ZnF motifs (Figures 1B,C). In Arabidopsis, four orthologs of SRP1 were identified, but SRP1 was unique among those five homologous proteins for the extra ZnF motif in N-terminal.

The expression of SRP1 was analyzed in the WT seedlings treated with ABA, salt stress and cold stress. SRP1 expression was down-regulated under the treatment of ABA within 24 h and gradually decreased along with the treatment time (Figures 2A,B). The expression of SRP1 was also repressed by salt or cold stresses (Figures 2C,D). Compared with WT, the expression of SRP1 in aba3-1 was up-regulated (Figure 2E). The aba3-1 is ABA deficient mutant, and the level of endogenous ABA in aba3-1 is reduced as the conversion of ABA-aldehyde to ABA is blocked (Léon-Kloosterziel et al., 1996). These results indicate the expression of SRP1 is repressed by abiotic stress and negatively regulated by ABA. It is of interest to investigate the involvement of SRP1 in ABA signaling as it is significantly down-regulated by ABA (Figure 2B).

The SRP1 T-DNA insertion mutants srp1-1 and srp1-2 (Figure 3A) were verified via genomic PCR (Figure 3B) and semi-quantitative RT-PCR (Figure 3C). The 35S::SRP1 plants (T₄ generation) were used in subsequent analysis. By qRT-PCR verification, it was found that SRP1 was highly expressed in the 35S::SRP1 plant (Figure 3D).

The seeds of the WT, srp1-1. srp1-2 and 35S::SRP1 plants were germinated on 1/2 MS medium supplemented with or without ABA to analyze the sensitivity to ABA. The germination rates among WT, srp1-1. srp1-2, and 35S::SRP1 plants were similar on the medium under control condition (Supplementary Figure 1A). As shown in Figure 4A, the germination rates among WT, srp1-1 and srp1-2 were similar on the 1/2 MS medium supplemented with 1 μM ABA. However, the germination rates of srp1-1 and srp1-2 seeds were significantly higher than WT on the third day when 1.5 or 2 μM exogenous ABA was applied. The germination rates of mutant seeds with 2.5 μM ABA treatment were significantly higher than WT from the third to the sixth day. The germination rate of WT was reduced by 4.97, 29.36, 44.64, or 52.06%, respectively, on the third day when 1, 1.5, 2, or 2.5 μM exogenous ABA was applied. However, the germination rate of srp1-1 was reduced by 0.05, 10.74, 24.38, or 29.33%, while srp1-2 was reduced by 2.05, 10.29, 24.28, or 32.10%. The difference between the mutants and WT was increased with the ABA concentration (Figure 4B). In contrast, the 35S::SRP1 line exhibited significantly lower germination rate than WT under ABA treatment (Figure 4A).

The green cotyledon rates of the WT, srp1-1. srp1-2, and 35S::SRP1 plants were also analyzed to find out the sensitivity of Arabidopsis to ABA during the post-germinative growth stage. The green cotyledon rates among WT, srp1-1. srp1-2, and 35S::SRP1 plants were similar under control condition (Supplementary Figure 1B). Seven days after the incubation on the 1/2 MS medium supplemented with 1 μM ABA, the cotyledon greening efficiencies of srp1-1 and srp1-2 were significantly higher than WT, while 35S::SRP1 line was significantly lower (Figure 5A). The green cotyledon rates of srp1-1 and srp1-2 were significantly higher than WT from the seventh day to the eleventh day when 1 μM exogenous ABA was applied, and the difference between mutants and WT was extremely significant from the ninth day to the eleventh day (Figure 5B). The green cotyledon rates of mutants and WT treated with 1.5, 2, or 2.5 μM exogenous ABA were also analyzed but no significant difference was observed (Figure 5C). Moreover, the green cotyledon rates of mutants and WT were decreased along with the ABA concentration. In contrast, the 35S::SRP1 plants treated with 1 μM exogenous ABA exhibited significantly lower green cotyledon rate than WT from the seventh day after sowing (Figure 5B). These results indicated that the srp1 mutants were more insensitive to exogenous ABA than the WT during the post-germinative growth, while the 35S::SRP1 plant was more sensitive.

In the presence of 100 mM NaCl, the germination rates of srp1-1 and srp1-2 seeds were significantly higher than WT in the first two days (Figure 6A), and the cotyledon greening efficiencies of srp1-1 and srp1-2 were significantly higher than WT on the third day (Figures 6B,C). In contrast, the 35S::SRP1 line exhibited significantly lower germination and green cotyledon rate than WT under NaCl treatment (Figure 6).

The contents of endogenous ABA in srp1 and WT were analyzed and we did not observe the marked differences among the mutant, 35S::SRP1 plant and WT (Supplementary Figure 2). The transcript abundances of ABA signaling and germination-related gene RD29A, RD29B, ABA INSENSITIVE 1 (ABI1), ABA INSENSITIVE 2 (ABI2), ABA INSENSITIVE 3 (ABI3), ABA INSENSITIVE 5 (ABI5), LATE EMBRYOGENESIS ABUNDANT 1 (EM1). and LATE EMBRYOGENESIS ABUNDANT 1 (EM6) were analyzed in the srp1 mutants. As shown in Figures 7A,B, the transcript levels of RD29A and RD29B were repressed by SRP1 knock-out under the control condition, but the transcriptions of ABI1 and ABI2 were enhanced. The transcript levels of RD29A. RD29B. ABI3. ABI5. EM1, and EM6 were significantly less induced in srp1 mutants than in WT under ABA treatment (Figures 7C,D). Interestingly, the transcript level of ABI2 was significantly increased in the srp1 mutants under control and ABA conditions, while that of ABI1 was slightly increased under control condition. In addition, the expressions of those genes in 35S::SRP1 plant were opposite. These results indicate the involvement of SRP1 in ABA signaling.

It has been demonstrated that AU rich element (ARE) could be bound by RBP (Lai and Blackshear, 2001), and the core sequence of ARE is AUUUA. In this study, the distributions of AUUUA sequence in ABA signaling-related genes were analyzed. Interestingly, four AUUUA sequences were found in the 3′UTR of ABI2 while zero or fewer AUUUA sequences were found in other genes (Figure 8A). As shown in Figure 8B, the biotin labeled ABI2 3′UTR mRNA lagged in the wells containing SRP1 protein. Moreover, the signals of lagging probes were reduced along with the addition of ABI2 3′UTR mRNA without labeling. By contrast, the biotin labeled ABI3 3′UTR mRNA did not lag in the wells containing SRP1 protein (Figure 8B). These results indicate the activity of SRP1 in binding to ABI2 3′UTR rather than ABI3 3′UTR. In addition, SRP1 was found with the transcriptional activation activity in yeast GAL4 system (Figure 8C). To further confirm the result, we examined the transcriptional activation activity of SRP1 in Arabidopsis protoplast system. As shown in Figure 8D, SRP1 could increase the luciferase activity by four-fold, indicating the transcriptional activation role of SRP1. The SRP1-green fluorescent protein (GFP) fusion protein was transiently expressed in Arabidopsis protoplasts. The SRP1 was revealed to be localized in nucleus of Arabidopsis protoplast cells (Figure 8E). The result suggested the function of SRP1 was performed in the nucleus.

The effect of SRP1 on ABI2 3′UTR-containing mRNA stability was examined in Arabidopsis protoplast cells to confirm whether the binding of SRP1 to ABI2 3′UTR mediated mRNA silencing (Figure 9). In this experiment, two reporter vectors, LUC-ABI2 3′UTR and LUC-ABI3 3′UTR were constructed by inserting the ABI2 or ABI3 3′UTR into the downstream of luciferase coding sequence in the control reporter LUC (Figure 9A). All of the reporter vectors were trans-activated by dehydration-responsive element-binding protein 1A (DREB1A), but the luciferase activity was reduced if an additional ABI2 3′UTR was fused with luciferase gene (Figure 9B). Importantly, the luciferase activity of LUC-ABI2 3′UTR was reduced by the addition of SRP1 while the luciferase activities of LUC and LUC-ABI3 3′UTR were enhanced (Figure 9B). Considering the transcriptional activity of SRP1 observed in yeast and protoplast systems, all three of these vectors were supposed to be enhanced. Hence, the decreased luciferase activity of LUC-ABI2 3′UTR confirmed the role of SRP1 in reducing the stability of mRNA with 3′UTR of ABI2.