Introduction

Plant growth and productivity are adversely affected by several types of stresses such as drought, salinity, threshold temperatures and pathogen attack. To cope with these challenges, plants have evolved molecular systems consequently developing adaptive responses through physiological, biochemical and molecular changes (Bohnert et al. 1995). A number of stress-responsive genes have been reported to be involved in the alleviation of stress-induced cellular damage; one of the class is germin-like proteins (GLPs) which is believed to provide resistance to plants against biotic and abiotic stresses.

GLPs are widely distributed in plants and seem to perform a range of metabolic roles from structural to catalytic. The catalytic roles, mostly include oxalate oxidase (OXO) for germins and superoxide dismutase (SOD) activity for GLPs (Yasmin et al. 2015); nevertheless, serine protease inhibitory activity (Segarra et al. 2003), ADP glucose pyrophosphatase (Rodriguez-Lopez et al. 2001), polyphenol oxidase (Cheng et al. 2014) and cysteine peptidase activities (de Freitas et al. 2016) have also been associated with GLPs.

Germin and GLPs are not only expressed at specific developmental stages, but many members are also responsive to biotic and abiotic stresses (Jiang et al. 2007; Dunwell et al. 2008). GLP overexpression in transgenic plants has been associated with improved tolerance to pathogen infection and salt stress (Wang et al. 2013); thus insight into GLP promoters and their corresponding DNA-binding proteins becomes highly relevant. There are several reports regarding in silico analysis of GLP promoters. Sassaki et al. (2014) analyzed EgGLP promoter and found cis-elements implicated in light, auxin, abscisic acid and salicylic acid responsiveness. Himmelbach et al. (2010) characterized the seven W-boxes found to be essential for induction by salicylic acid and providing defense against pathogen infection.

Mahmood et al. (2007) cloned OsRGLP2 promoter, performed heterologous expression in tobacco and observed strong GUS expression in transgenics during wounding/mechanical stress, salt and dehydration. Hussain (2015) performed 5′ deletion analysis of the OsRGLP2 promoter and observed that full length as well as two deleted fragments of 776 and 565 bp were responsive to wound, salinity and temperature stresses, indicating that the OsRGLP2 promoter may serve as a location where multiple signaling pathways might integrate to produce a response to stress stimuli. However, experimental evidence regarding the identification and functional characterization of transcription factors which might interact with cis-acting regulatory elements of the GLPs promoters were required. Therefore, in the present study, putative transcription factors that interact with cis-acting regulatory elements of OsRGLP2 region were identified by in silico analysis followed by their cloning in E. coli, overexpression and purification. Interaction of purified proteins with fragments of OsRGLP2 promoter was investigated by electrophoretic mobility shift assays (EMSAs). Results revealed that OsRGLP2 promoter can interact with OsWRKY71, OsDOF18 and OsMYB1 proteins which play a role in regulating OsRGLP2 gene transcription.

Materials and methods

Computational analysis of OsRGLP2 promoter

The OsRGLP2 promoter region (accession no. DQ414400.1) was screened for the presence of putative stress-responsive elements as listed in PLACE (http://www.dna.affrc.go.jp/PLACE/signalscan.html). The elements detected by PLACE were further validated by other online plant cis-element databases, PlantCARE (http://bioinformatics.psb.ugent.be/webtools/plantcare/html/), ConSite (http://consite.genereg.net/), and JASPAR (http://jaspar.binf.ku.dk/). Target genes were investigated using the gene expression database called Genevestigator (Plant Biology) (https://genevestigator.com/gv/plant.jsp) to identify WRKY71 (Os02g0181300), MYB1 (Os01g0850400) and DOF18 (Os08g0490100) that appeared to show up-regulation during various stresses. Data were acquired for abiotic stresses including cold, heat, drought and salt stresses.

Sequence retrieval and sequence alignment for the DNA-binding domain identification

Nucleotide and protein sequences of selected stress-responsive transcription factors were obtained from NCBI (http://www.ncbi.nlm.nih.gov/). Protein sequences of AtMYB1 and ZmDOF1 were subjected to BLAST (http://blast.ncbi.nlm.nih.gov/Blast.cgi) and search was performed against a non-redundant database, but was limited to include only hits within the O. sativa genome. For identification of DNA-binding domains, protein sequences of various plant proteins were obtained from BLAST result and multiple alignment was performed using TCOFFEE (http://tcoffee.crg.cat/). ESPript (http://espript.ibcp.fr/ESPript/ESPript/) was then used for producing graphical representations of the multiple alignments. Secondary structures of DNA-binding domains were predicted using PSIPRED (http://bioinf.cs.ucl.ac.uk/psipred/).

Preparation of crude nuclear protein extract

To analyze the possible physical interactions of predicted sites with specific transcription factors, nuclear proteins were extracted from Oryza sativa. For this purpose, O. sativa cv KS282 seeds obtained from Rice Program, National Agricultural Research Centre, Islamabad, were germinated on half-strength MS medium at 25 °C for 10 days. Five grams of 10-day-old seedling mass was ground in an ice-cold mortar and pestle in liquid nitrogen and nuclear proteins were extracted according to the protocol of Escobar et al. (2001). The nuclear protein extract was dialyzed in the EMSA binding buffer.

Production and purification of recombinant proteins

To validate the interaction of target transcription factors with the respective putative binding sites, target proteins were produced as purified recombinant proteins. For this purpose, total RNA was isolated using RNeasy Plant Mini Kit (Qiagen) followed by cDNA synthesis using Oligo (dT) 18 Primer and RevertAid Premium Reverse Transcriptase (Thermoscientific). cDNA was then used as template for PCR amplification of DNA-binding domains of OsWRKY71 and OsMYB1, while cDNA clone (J065152E11) acquired from NIAS was used as template for OsDOF18. The DNA sequences coding for DNA-binding domains, including 15–20 amino acids flanking N and C terminus, were amplified using specific primers containing BamHI and XhoI restriction sites (Supplementary Table 1) and cloned into predigested pGEX4T-1 vector. Cloning of correct fragments in recombinant plasmids was confirmed by restriction digestion and commercial sequencing. The recombinant plasmids were transformed into BL21 (DE3) and expression was induced with 1 mM isopropyl β-d-thiogalactopyranoside (IPTG) for 3 h when OD600 reached about 0.6. The cells were harvested by centrifuging (3500g, 20 min, 4 °C) and resuspended in phosphate-buffer saline (PBS) containing 1% triton X-100. The resuspended pellet was sonicated (6 cycles, 30 s) followed by incubation on ice for 30 min. The supernatant was then collected after centrifugation for 20 min at 15,000g at 4 °C and adsorbed onto glutathione-Sepharose beads for 30 min at 4 °C, after which the beads were washed twice in PBS and fusion proteins were eluted using elution buffer. The apparent molecular mass and purity of eluted proteins were analyzed by 15% SDS-PAGE according to standard protocols. Semi-purified proteins were subjected to cation exchange chromatography and the eluents were dialyzed in EMSA binding buffer [50 mM HEPES; pH 7.5, 40 mM KCl, 12.5 mM MgCl2, 2 mM DTT, 1 mM PMSF 0.2 mM EDTA, 20% (v/v) glycerol].

Electrophoretic mobility shift assays (EMSAs)

Oligonucleotides (30 bp) containing WRKY, MYB or DOF binding sites located in the OsRGLP2 promoter were annealed by heating at 95 °C for 10 min and then slowly cooled to room temperature in TEN buffer [Tris (10 mM, pH 8), 1 mM EDTA, 100 mM NaCl]. A DIG Gel Shift kit (Roche) was used to detect DNA–protein interaction. Digoxigenin (DIG) was labeled at the 3′ end of the double-stranded oligonucleotides. The labeled probes were then purified by ethanol precipitation. The pellet was air dried and resuspended in ddH2O to a final concentration of 2.5 pmol/µL. The labeling efficiency was tested by spotting serial dilutions on nylon membrane along with control labeled oligonucleotides provided with the kit.

DIG-labeled (50 fmol), unlabeled (10 pmol) or mutated labeled oligonucleotides (50 fmol) were mixed with GST fusion proteins or a crude nuclear protein extract (NPE) in binding buffer and kept at room temperature for 30 min. The reaction products were fractionated through a pre-run 6% non-denaturing polyacrylamide gel for 90 min in 0.5× TBE buffer at 4 °C. The proteins were then electrophoretically transferred onto a positively charged nylon membrane (Roche) by applying a constant current of 300 mA for 30 min. The DNA was cross-linked to the membrane using a UV stratalinker (Stratagene). The nylon membrane was blocked in 1× blocking reagent for 1 h at room temperature and then incubated with a 1:20,000 dilution of alkaline phosphatase coupled anti-DIG antibody for 30 min. The membrane was washed twice with washing buffer [0.1 M maleic acid (pH 7.5), 0.15 M NaCl, 0.3% (v/v) Tween 20] for 20 min each. After 5 min equilibration in detection buffer (100 mM Tris/HCl, pH 9.5, 100 mM NaCl), the membrane was carefully placed on a plastic sheet and 1:100 dilution of the CSPD substrate was then added on the membrane dropwise near the edges and then the whole membrane was covered by tilting it. The membrane was then incubated at room temperature for 5 min and subsequently placed in a plastic folder and incubated at 37 °C for another 15 min. Finally, the membrane was exposed to X-ray film for 3 h to capture the chemiluminescent signal.

Results

In silico analysis of the OsRGLP2 promoter

In the OsRGLP2 gene promoter region, many key cis-elements that were responsive to abiotic stresses were identified by PLACE, for instance MYB binding site, MYC binding site, DRE elements (dehydration-responsive element), DOF binding site, ABRE, GT-1 element, BIHD1 binding site, ANAERO2CONSENSUS, CBF (C-repeat binding site) and W-Box (Supplementary Table 2). Putative cis-elements were further validated by analyzing the promoter sequence in JASPAR, PlantCare and Consite tools. All the tools shared the presence of MYB, W-box and DOF elements. JASPAR also predicted the presence of bZIP and NAC025 binding sites. Fourteen putative W-boxes, 14 DOF and 4 MYB1 binding sites were found common in the OsRGLP2 promoter region on plus and negative DNA strands (Fig. 1). Therefore, these three sites were chosen for further analysis and it was assumed that WRKY, DOF and MYB proteins may be involved in transcriptional regulation of OsRGLP2 gene under abiotic stresses.

Fig. 1
figure 1

Position of different stress-responsive cis-elements on the 1109 bp OsRGLP2 promoter region

Full size image

OsWRKY71 was predicted by PLACE as an interaction partner of W-box located in the OsRGLP2 promoter. To better understand the secondary structure of the WRKY71 DNA-binding domain, multiple alignments were carried out with WRKY proteins from different plants. Results revealed that it contains a single, highly conserved WRKY domain and zinc finger-like motif at its C terminus (C–X5–C–X23–H–X1–H) that places it in group IIA of the WRKY family (Supplementary Fig. 1). Based on the putative cis-element binding sites, sequences of AtMYB1 and ZmDOF1 (predicted by PLACE, Consite and JASPAR) were retrieved from the NCBI database. Homologous MYB and DOF sequences in rice genome were identified by BLAST against AtMYB1 and ZmDOF1 query sequences. The result showed AtMYB1 to have 42% sequence similarity (all in the DNA-binding domain region) with Os01g0850400 which is similar to partial OsMYB1 protein sequence (accession no. CAE00856.1). Multiple sequence alignment with MYB proteins from different plants showed that it contains two conserved MYB DNA-binding domains at the N-terminus and belongs to R2R3-type subfamily of MYB DNA proteins (Supplementary Fig. 2).

The protein sequence of ZmDOF1 when subjected to BLAST search revealed 54% amino acid sequence similarity with putative rice protein. Protein functional analysis by InterProScan showed identity with an InterPro accession number IPR003851 Znf_DOF, thereby confirming its identity to DOF-like proteins. This protein sequence was found in the Database of Rice Transcription Factors (DRTF) under locus ID Os08g38220 named as OsDOF18/MNB1A. It encodes a unique DNA-binding domain of 55 amino acids encompassing a C2–C2 zinc finger (Supplementary Fig. 3). To investigate the association of OsWRKY71, OsMYB1 and OsDOF18 transcripts with stress responses, an examination of global expression was performed using the Genevestigator (Supplementary Table 3). The transcript level of all genes was significantly up-regulated under different abiotic stresses to which the OsRGLP2 promoter is responsive including salt and drought.

Binding of recombinant GST-fused OsWRKY71, OsMYB1 and OsDOF18 DNA-binding domains with the OsRGLP2 promoter region

The possible physical interaction of in silico identified transcription factors with OsRGLP2 promoter was explored by producing GST-fused recombinant DNA-binding domains in E. coli BL21 and purified as GST/WRKY71 (167–271), GST/DOF18 (1–148) and GST/MYB1 (88–222) (Supplementary Fig. 4). To test the hypothesis on the regulation of the OsRGLP2 promoter by OsWRKY71, OsMYB1 and OsDOF18, electrophoretic mobility shift assays (EMSAs) were performed to assess the potential binding of these proteins to the OsRGLP2 promoter region. WRKY proteins are generally considered to bind to the consensus W-box sequence C/TTGACC/T (Eulgem and Somssich 2007). The OsRGLP2 promoter contains four true W-boxes; two true W-boxes TTGACT at positions −800 and +491, and two true W-boxes CTGACC at positions −129 and −340, although a number of TGAC core sequences are present. As a first step toward characterization of OsWRKY71 binding sites in the OsRGLP2 promoter, three double-stranded 30 bp labeled promoter fragments containing either true W-box or core W-box sequences were assayed for their ability to bind to a purified GST/WRKY71 alone or in the presence of wild-type and mutant DNA fragments. There was only one true W-box and one TGAC cores in the WRKY71-Probe1 (WRKY71-P1) (Fig. 2a). No binding signal was detected in negative controls, while OsWRKY71 bound to the wild-type WRKY71-P1. Binding specificity was demonstrated by the competition experiment with 200-fold excess of unlabeled fragment which outcompeted the binding to the probe, suggesting that the OsRGLP2 promoter fragment specifically interacted with WRKY71. No shift was observed when the TGAC sequences (core W-box) were mutated to AAAA. However, a competitor with both elements mutated (mTGACs) could not abolish the binding of the protein to the probe. Similar results were observed for WRKY71-P2 and WRKY71-P3 probes that had only one true W-box or one W-box core sequence, respectively (Fig. 2b, c). No shift was observed in lanes containing the labeled probe without protein and labeled probe with GST in case of WRKY71-P2. GST–OsWRKY71 bound directly to WRKY71-P2 when incubated with the labeled probe. To confirm binding, W-box sequence (TGAC) was changed to AAAA and EMSA was performed using labeled mutated probe. This mutated W-box containing EMSA probe failed to bind with GST–OsWRKY71. There was no shift in the negative control of the labeled mutant probe. For competition, 200-fold molar excess of unlabeled cold WRKY71-P2 was added to the EMSA reaction before addition of the labeled probe. Intriguingly, cold EMSA probe competes well with the radiolabeled probe. When excess of mutated cold EMSA probe was added to the reaction, it failed to compete with the labeled W-box containing a fragment of the OsRGLP2 promoter (Fig. 2b). With WRKY71-P3, there was no shift in both negative controls, i.e., labeled probe without protein (GST and WRKY71) and labeled probe with GST. GST–OsWRKY71 bound to the core sequence and a shift was observed, but this binding was weak as compared to true W-box sequence. The specificity of shift was further established through competition and mutation assays. The shifted band was totally competed out by 200-fold molar excess of unlabeled cold probe. It was observed that there was no shift when the core sequence (TGAC) was mutated to AAAA. Excess of the unlabeled mutant probe also failed to abolish the WRKY–DNA complex with labeled OsWRKY71-P3. Taken together, the data suggested that OsWRKY71 can recognize the OsRGLP2 promoter in vitro, binding specifically to TGAC containing cis-elements.

Fig. 2
figure 2

In vitro DNA-binding ability of OsWRKY71 with DIG-labeled OsRGLP2 fragments containing core and true W-boxes. EMSA of recombinant OsWRKY71 with fragment 1 containing one core and one true W-box (a), with fragment 2 containing one true W-box (b), and with fragment 3 containing one core W-box on both strands (c). In gels a, b, c lanes are in the same order. Lane 1 OsWRKY71 oligos negative control, lane 2 GST-negative control, lane 3 1 µg GST–OsWRKY71 with OsWRKY71 oligos, lane 4 with 200 molar excess of unlabeled OsWRKY71-2 oligos, lane 5 with mutant labeled oligos, lane 6 200-fold molar excess of unlabeled mutant oligos, lane 7 mutant labeled oligos negative control

Full size image

To investigate the function of the several AAAG motifs in the OsRGLP2 5′-flanking region, DIG-labeled DNA Probes 1 and 2 representing two and one AAAG motifs, respectively, were used in EMSA. When DOF18-P1 was incubated with the GST/OsDOF18 protein, a shifted complex was observed that was competed out when the probe was incubated with 200 molar excess of the corresponding unlabeled oligonucleotides (Fig. 3a). As expected, this binding was not produced when the control GST protein was used in the assay. The same result was obtained when DOF18-P2 containing a single DOF-box was used (Fig. 3b), but the shift was faint in comparison to a probe containing two DOF boxes. The specificity of the interaction was also confirmed using variants of these probes. The binding of the GST/OsDOF18 to the DOF-box motif in the OsRGLP2 gene promoter was abolished when the core AAAG sequence was mutated to AGAC. As shown in Fig. 3b, the mutant version of the DOF probe was not bound by the GST/OsDOF18 protein or was not able to compete the binding of the corresponding wild-type probe at 200 molar excess of mutant DOF18 probe. These results entail greater DOF18 binding to a complex of two AAAG motifs than a single motif (Fig. 3c), implying an additive binding effect due to the presence of a greater number of DOF elements. The presence of multiple DOF18 binding sites in the OsRGLP2 promoter is thus suggestive of the greater DOF18 binding at multiple sites to enhance its effect.

Fig. 3
figure 3

In vitro DNA-binding ability of OsDOF18 with DIG-labeled OsRGLP2 fragment containing two or one AAAG motif. a, b EMSA of recombinant OsDOF18 with OsRGLP2 promoter fragment containing two and one AAAG motifs, respectively. In both a, b, lanes are in the same order. Lane 1 OsDOF18 oligos negative control, lane 2 GST-negative control, lane 3 1 µg GST–OsDOF18 with OsDOF18 oligos, lane 4 with 200 molar excess of unlabeled OsDOF18 oligos, lane 5 with mutant labeled oligos, lane 6 200-fold molar excess of unlabeled mutant oligos, lane 7 mutant labeled oligos negative control. c DOF18-P1 show more specificity for OsDOF18 as compared to DOF18-P2. Lane 1 OsDOF18-P1 oligos negative control, lane 2 GST-negative control, lane 3 1 µg GST–OsDOF18 with OsDOF18-P1, lane 4 with 200 molar excess of unlabeled OsDOF18-P1 oligos, lane 5 with 400 molar excess of unlabeled OsDOF18-P1 oligos, lane 6 OsDOF18-P2 oligos negative control, lane 7 GST-negative control, lane 8 1 µg GST–OsDOF18 with OsDOF18-P2, lane 9 with 200 molar excess of unlabeled OsDOF18-P2 oligos, lane 10 with 400 molar excess of unlabeled OsDOF18-P2 oligos

Full size image

MYB transcription factors are reported to recognize specific binding motifs, A/TAACCA and C/TAACG/TG (Abe et al. 2003). In the OsRGLP2 promoter, MYB-binding motifs (5′-TAACCA-3′ and 5′-AAACCA-3′) were also found at positions −137 and −446. As shown in Fig. 4, the GST/OsMYB1 fusion protein was able to bind to the DIG-labeled MYB1-P1 and MYB1-P2. The unlabeled probes competed with DIG-labeled probe, suggesting that the signal from the probe represents the specific binding ability of the OsMYB1 with the OsRGLP2 promoter fragments. When the MYB-box (WAACCA) was mutated to TCATGA, no shift was observed. In addition, probes with a mutation in the MYB-binding sequence failed to compete with the wild-type form of the probe. These results demonstrated that OsMYB1 was able to directly and specifically bind to the MYB binding site of the OsRGLP2 promoter in vitro.

Fig. 4
figure 4

In vitro DNA-binding ability of OsMYB1 with DIG-labeled OsRGLP2 fragments containing AAACCA or TAACCA MYB binding site, respectively. a, b EMSA of recombinant OsMYB1 with a fragment containing AAACCA and TAACCA motifs, respectively. Both in a, b lanes are in the same order. Lane 1 OsMYB1 oligos negative control, lane 2 GST-negative control, lane 3 1 µg GST–OsMYB1 with OsMYB1 oligos, lane 4 with 200 molar excess of unlabeled OsMYB1 oligos, lane 5 with mutant labeled oligos, lane 6 200 fold molar excess of unlabeled mutant oligos, lane 7 mutant labeled oligos negative control

Full size image

The interactions of rice nuclear protein extract with all probes were also investigated by EMSAs. It was observed that shifts appeared as multiple bands in case of WRKY71-P2 and P3 (Fig. 5). The specificity of binding was also demonstrated by competitive EMSA. When WRKY71-P2 and WRKY71-P3 sequences were scanned with PLACE, Consite, JASPAR and PlantCare, these promoter regions were also found to contain binding sites for proteins other than WRKY71 (data not shown). The appearance of multiple bands may be a consequence of binding of different proteins to the employed probes as a result of binding different proteins (in addition to the three tested in this study) or transcription factor with and without modulatory proteins.

Fig. 5
figure 5

In vitro DNA-binding ability of nuclear protein extract with OsRGLP2 promoter fragments. In each gel, lane 1 7 µg nuclear protein extract with labeled probes, and lane 2 with 200 molar excess of respective unlabeled probe. The arrows represent the specific shift due to binding of protein to cis-acting elements on probe regions. The specific shift faded away when an excess of unlabeled probes were incubated with nuclear protein extract

Full size image

Discussion

Transcriptional regulation of plant genes plays a vital role in the activation of inducible plant defense response. Many studies have been conducted on the expression of GLP genes during development and stress conditions (Li et al. 2010), but still little is known about the mechanism of their regulation. OsRGLP2 promoter showed strong expression of GUS gene in tobacco plants during wounding/mechanical stress, salt and dehydration (Mahmood et al. 2007). The publication of the genome sequences for rice plant and the development of recent bioinformatics tools like Genevestigator (http://genevestigator.com), TRANSFAC (http://www.biobase-international.com/product/transcription-factor-binding-sites), have hastened the studies on the regulation of expression of genes of interest. The cis-elements play a vital part in the transcriptional regulation of gene expression, controlling complicated abiotic stresses to increase the tolerance of plants under fluctuation environments. In our report, different bioinformatics tools were used to identify the putative stress-related cis-acting regulatory elements located in the OsRGLP2 promoter region. Significant variations were observed in the occurrence of stress-related cis- acting elements in the promoter including DOF, MYB, WRKY, ABRE, DREB and GT-1 elements. The OsRGLP2 promoter contains more than ten binding sites for WRKY and DOF proteins and three sites for MYB proteins, which indicates that these genes might be strongly associated with functions under different stresses. However, to validate the computational prediction of transcription factors binding sites within the promoter, it is essential to experimentally investigate the interaction between putative cis-elements and corresponding binding factors. To examine the expression of OsWRKY71, OsMYB1 and OsDOF18 under various abiotic stress conditions in rice, we took advantage of the available microarray from Genevestigator. Analyses of microarray data demonstrated that the expression of OsWRKY71, OsMYB1 and OsDOF18 genes is regulated by various abiotic stress conditions.

W-boxes with typical TGAC core motif forming binding sites for WRKY proteins are reported to be involved in plant development and abiotic and biotic stresses (Eulgem and Somssich 2007; Himmelbach et al. 2010). AAAG motif is a core site for binding of DOF proteins, a type of zinc finger regulatory protein that plays diverse roles in plant gene expression such as in developmental stages (Konishi and Yanagisawa 2007) and different stresses (Kang et al. 2003). MYB transcription factors bind to the WAACCA motif and respond to various environmental stimuli (Yu et al. 2012) and plant metabolism (Goicoechea et al. 2005). The high copy number of W-boxes, WAACCA and AAAG motifs in the upstream regulatory region of OsRGLP2 may describe them as stress-responsive proteins and these cis-regulatory elements may participate in the regulation of its expression to wounding, drought, salt and cold stresses.

DNA-binding domains recognize the regulatory motif sequence and binds with it through short motif, mostly an α-helix or a β-sheet. These short motifs make contact with the major groove of double-stranded DNA. Usually, the DNA–protein contact spread out across 5 bp, with a relative high affinity and sequence specificity (Ciolkowski et al. 2008). It is observed that neighboring nucleotides of cis-elements contribute in determining what type of transcription factors will be recruited (Ciolkowski et al. 2008). Moreover, the small differences that exist between DNA-binding domains could play a role in fine-tuning of their function (Bergholtz et al. 2001). OsWRKY71 was predicted as an interaction partner of W-box in the OsRGLP2 promoter. Zhang et al. (2004) functionally characterized OsWRKY71 and found it to be overexpressed in the aleurone layer/cells of rice and to act as a negative regulator of GA signaling. OsWRKY71 interacts with the GA-inducible Amy32b promoter and represses the GAMYB-mediated GA signaling in aleurone cells (Xie et al. 2006). Liu et al. (2007) observed that OsWRKY71 expression was up-regulated in response to wounding and pathogen infection in rice. OsMYB1 and OsDOF18 sequences retrieved from O. sativa genome, followed by a prediction of their secondary structures, suggested that they might interact with DNA. MYB and DOF proteins are involved in various physiological processes, including abiotic stress response (Yu et al. 2012), but the role of OsMYB1 and OsDOF18 in biotic and abiotic stresses still need to be addressed.

Scientists use diverse methods to study the DNA–protein interaction: for instance, yeast one-hybrid, phage display technique and EMSA. In this study, DIG-labeled EMSA was used to scan the potential transcriptional regulatory factors for OsRGLP2 promoter in vitro. The W-boxes are present in the promoter regions of various plant defense providing genes, including WRKY itself. Both computational and functional studies have found clusters of W-boxes in the promoters of stress-inducible genes: for instance, occurrence of W-boxes in the promoters of PR-1, a marker gene for SAR (Maleck et al. 2000). Multiple W-boxes have a synergistic effect on transcription (Mare et al. 2004). There are four putative W-boxes on OsRGLP2 promoter: two (TTGACT) W-boxes on the positive strand and two (CTGACC) on the negative strand. In addition, many TGAC core sequences were also found. The present study demonstrated that OsWRKY71 interacts with a W-box core sequence (TGAC) as well as extended W-box (C/TTGACC/T) and controls the expression of the OsRGLP2 gene. Adjacent sequences of core W-box may aid in/act synergistically to enhance specificity and affinity (Mare et al. 2004). Co-expression and up-regulation of WRKY and GLPs were observed in rice upon infection with Fusarium fujikuroi (Matic et al. 2016).

In the OsRGLP2 promoter, MYB-binding motifs (5′-TAACCA-3′ and 5′-AAACCA-3′) were found. The cloned DNA-binding domain of OsMYB1 was able to bind both these sequences. Binding was confirmed by competitor EMSA and EMSA with mutant oligonucleotides. Nectarin-I proteins belong to the GLP family and its promoter analysis revealed the presence of consensus MYB DNA binding site (Carter and Thornburg 2003). Binding of NtMYB305 with Nectarin-I promoter was evaluated by EMSA and it was found that R2R3 MYB DNA-binding domain containing NtMYB305 is able to interact with the Nectarin-I promoter (Liu et al. 2009). Similarly, to check the binding of OsDOF18 with AAAG motifs located in the OsRGLP2 promoter, EMSA was carried out with probes containing one or two AAAG motifs. The specificity of interaction was confirmed by adding an excess of unlabeled probes and mutant probes. Shifts were also observed with all the probes when EMSA was carried out with nuclear proteins extracted from rice. The shifted bands were specific as confirmed by competition with an excess of unlabeled probes. We therefore conclude that the OsMYB1 and OsDOF18 transcription factors can directly interact with the OsRGLP2 promoter and transactivate OsRGLP2 expression.

In conclusion, the present study highlights the in silico identification of stress-responsive cis-regulatory elements in the promoter of the OsRGLP2 gene and their corresponding binding proteins. The DNA-binding domains of the identified proteins OsWRKY71, OsMYB1 and OsDOF18 were able to interact with the OsRGLP2 promoter as confirmed by EMSA. OsWRKY71 is characterized for its role in biotic stresses, but no work has been done on its role in response to abiotic stresses. OsMYB1 and OsDOF18 are the proteins yet to be characterized in rice. The availability of more information regarding these proteins in rice may further enhance our understanding about gene expression and its manipulation during biotic as well as abiotic stresses.

Author contribution statement

FD performed most of the laboratory work and participated in all other activities. TS participated in laboratory experiments. TM was involved in planning the study. COS participated in laboratory work and bioinformatics analysis. KS was involved in planning the experiments and provided laboratory facilities in Denmark. SMSN planned and supervised all the work, provided laboratory facilities and also handled the manuscript as corresponding author. All authors read and approved the final manuscript.