The southern root-knot nematode (RKN; M. incognita) culture was maintained on tomato and eggplant (Solanum melongena L.). Tomato and eggplant seeds were sterilized (soaked for 20 min in sterile distilled water, 5 min in 70% ethanol, and 15 min in 5% NaOCl and 0.1% Tween 20, and washed four times in sterile distilled water) and germinated on MS agar medium. After 3 weeks, the plant roots were infected with 500 second-stage juveniles (J2s) of RKN. Six weeks later, egg masses were hand-picked and hatched at 28°C in 10–15 ml of sterile water.

Arabidopsis thaliana (Col-0) seeds were surface sterilized (by immersing for 2 min in 70% alcohol and 7 min in 0.1% Mercuric chloride and 0.1% SDS) and stratified for 72 h at 4°C before germination on Gamborg's B-5 medium. Plates were covered with parafilm M® and maintained at 21°C under a 16 h light/8 h dark photoperiod. Fourteen-day-old Arabidopsis seedlings were transferred from the Petri plates to trays containing a sand/vermicompost/cocopeat mixture (1:1:1 w/w). After 3 weeks, each plant was inoculated with 1,000 freshly hatched J2s of RKN using a 1 ml pipette. The trays were maintained in a growth chamber at 21°C and the M. incognita-infected galls and complete roots of A. thaliana were harvested at two different intervals viz., 10 and 21 days post inoculation (dpi). The uninfected root tissue, dissected from the plants at the same time points, served as control samples. At each time point, samples were harvested from both infected and uninfected plants and frozen in liquid nitrogen and stored at −80°C until further use.

Publicly available microarray data for Arabidopsis was analyzed using packages for R-statistical language and using an online microarray resource, Genevestigator (v 3.0) (Zimmermann et al., 2004). Microarray datasets from studies done by different labs and using platforms were difficult to process through our in-house scripts, primarily due to the batch-effects, and differences in numbers and identifiers of probe-sets corresponding to different platforms. Such datasets were analyzed through Genevestigator tool (Hruz et al., 2008); this study is referred to as “Genevestigator” analysis from here on. Independent to the Genevestigator analysis, the comprehensive microarray datasets from “AtGenExpress” study were processed and analyzed through in-house R-language scripts; this is referred to as “stand-alone” analysis from here on. Please see Supplementary Text Presentation 1 for more information on both the Genevestigator and stand-alone analysis.

Genes that preferentially up-regulate upon nematode infection were identified using Genevestigator; stand-alone analysis to identify such nematode-responsive genes could not be performed because the microarray datasets were not available in public domain at that time this study was performed. Genevestigator (v3) classified the microarray data from nematode infection assays into two stages—“early” and “late” based on days post infection. Probe sets that showed an up-regulation (>1.5-fold-change, log2-scale) in both “early” and “late” stages were downloaded in tab-separated file format and used for downstream analysis. On other hand, genes that show root-specific expression were identified using both the Genevestigator and stand-alone analysis. In stand-alone analysis, probe sets that showed a consistent up-regulation (>2-fold-change, log2) across all three contrasts, including with vegetative stem, inflorescence and leaf samples, at each stage were recorded (see Supplementary Text Presentation 1); this process was repeated for three time points (7, 17, and 21 days). Since, “AtGenExpress” data represented select developmental stages, we collected probe sets that are preferentially expressed (>1.5-fold-change, log2-scale) in root samples (six root zones) compared to other tissues or plant samples from all datasets (except AtGenExpress) available in Genevestigator (v3).

Multiple EM for Motif Elicitation (MEME) was used for the discovery of novel putative motifs. A third-order Markov chain background model was prepared from the upstream sequence of all the Arabidopsis genes. The upstream sequences for 27,144 genes were extracted using RSAT (Thomas-Chollier et al., 2008). Repetitive DNA elements were masked from the input sequences using Repeat Masker program (Smyth, 2004). Two different modes, Any Number of Repetitions (ANR) and Zero or One Occurrence per Sequence (ZOOPS) were used with the following parameters:

-nmotifs 30 -minsites 4 -maxsites 12 -minw 6 -maxw 12 -revcomp -dna -mod anr–bfile

Where, number of -minsites and -maxsites is given based on number of input sequences and the third order background model is provided for–bfile.

-nmotifs 30 -minw 6 -maxw 8 -revcomp -dna -mod zoops–bfile

Where, -minw 6–10 and -maxw 8–12 and the third order background model is provided for –bfile.

For ANR mode, we considered 30 motifs ranked by their significance levels of length between 6 and 12 nt, while for ZOOPS mode we had three runs, each predicting 10 statistically ranked motifs corresponding to motif length of 6–8 nt, 8–10 nt, and 10–12 nt, respectively. The statistically-significant motifs were screened for similarity with known transcription factor binding motifs using POXO (Kankainen et al., 2006) (http://ekhidna.biocenter.helsinki.fi/poxo) and STAMP (Mahony and Benos, 2007).

Promoters of NRRS genes identified along with the previously reported genes as positive control (TobRB7, Atcel1, Hahsp17, and Lemmi9) were subjected to comparative analysis for identification of conserved cis-elements. The upstream region from the start codon was extracted to include only the intergenic region (~1.5 kb) using Regulatory Sequence Analysis Tool (RSAT) (http://www.rsat.eu/) (Thomas-Chollier et al., 2008). The known cis-elements present in each of these promoters were collected using two different public resources, PLACE (Higo et al., 1999) and ATcisDB (Davuluri et al., 2003). The set of reported motifs were further used to elucidate conserved cis-elements using a Python script.

Total RNA was isolated from M. incognita-infested galls and whole roots of Arabidopsis using RNeasy mini kit (Qiagen, Germany) following the manufacturer's protocol. The RNA integrity was checked by formaldehyde gel electrophoresis with ethidium bromide dye. Nanodrop spectrophotometer 8000 (Thermo Scientific, USA) was used to calculate purity ratios and quantify total RNA. cDNA was prepared using Protoscript M-MuLV first strand synthesis kit (NEB, USA) using oligo-d (T)23VN primers. cDNA was normalized and re-quantified before qRT-PCR. RT Primers (Supplementary Table 1) were designed from the cDNA sequence of selected genes using “Primer3” portal (http://frodo.wi.mit.edu/primer3/) (Rozen and Skaletsky, 2000).

The cDNA from infected and control samples was quantified and used as template with three biological replicates along with three technical replicates for each biological replicate. To ensure purity of the master mix and reaction mix setup, a non-template control reaction was included in every plate. A 20 μl reaction volume consisting of SYBR FAST qRT-PCR Master Mix (2x) Universal (KAPA Biosystems) and 10 pmol of each primer was used in all qRT-PCR reactions. The qRT-PCR reactions were performed on a StepOnePlus™ Real-Time PCR system with the following cyclic conditions: initial denaturation temperature of 95°C for 10 min followed by 40 cycles of 95°C for 15 s and 61°C for 45 s. The PCR products were exposed to melting curve analysis; the conditions were incubation at 60–95°C with a temperature increment of 0.3°C s⁻¹ (Applied Biosystems®). The threshold cycle values were normalized by plant UBQ10 (Supplementary Table 1) as endogenous control and fold changes of the target gene were calculated by 2^−ΔΔCt method (Livak and Schmittgen, 2001).

The promoter regions (1.5 kb upstream of the start codon) of NRRS genes, At1g74770 and At2g18140, were amplified from the Arabidopsis genomic DNA using the gene specific oligonucleotides (Supplementary Table 1), with flanking restriction sites, BamHI and SalI. The amplified products were eluted from gel using Pure Link Gel Extraction kit (Invitrogen, USA), quantified and digested with BamHI and SalI. The 1.5 kb BamHI/SalI promoter fragment was cloned upstream of the GUS gene, using T4 DNA ligase (NEB, USA), with linearized BamHI/SalI digested pORE-R2 (Coutu et al., 2007) vector and transformed to DH5-α strain of E. coli (NEB, Massachusetts, USA). The prm::GUS fusion constructs were validated by nucleotide sequencing and introduced into Agrobacterium tumefaciens (Smith and Townsend, 1907) strain GV3101. Arabidopsis plants were transformed using the floral dip method (Clough and Bent, 1998). The primary transformants were selected on medium containing kanamycin (50 μg/ml) and further grown to develop T₃ seeds. For each promoter, five independent Arabidopsis transgenic lines were tested for their response to nematode infection.

Histochemical localization of GUS activity was performed with the substrate 5-bromo-4-chloro-3-indolyl-β-D glucuronide (X-Gluc) (Jefferson, 1989). The infected plants were uprooted and washed with water until roots were free of soil. The soil-free plants were immersed in freshly prepared GUS assay buffer [0.5 mM X-Gluc, 0.1 M NaHPO₄ pH 8.0, 0.5 mM K₃Fe(CN)₆, 0.5 mM K₄Fe(CN)₆, 0.01 M EDTA pH 8.0, 20% methanol, and 0.1% Triton X-100] and vacuum infiltrated for 5–10 min in a desiccator. The tubes were then incubated overnight in dark at 37°C. The tissues were cleared by replacing the buffer with 70% ethanol and then imaged with a stereomicroscope (Nikon®, Japan) with an external fiber optic light source. The plants were monitored for GUS activity at 10 and 21 days after nematode infection.

A 349-bp splicing factor sequence (AW828516) was amplified from M. incognita and cloned in pBC06 RNAi vector in sense and anti-sense directions (Yadav et al., 2006). CaMV35 promoter was removed from above vector and replaced with At1g74770 root-specific promoter using Sbf I and BamHI restriction enzymes. All cloning steps were performed according to the protocols described by Sambrook et al. (1989), and the constructs were confirmed by restriction fragment analysis and sequencing. The binary vectors were transferred to A. tumefaciens strain GV3101 by freeze and thaw method (Weigel and Glazebrook, 2006). A. thaliana (Col-0) plants were transformed with pAt1g74770::splicing factor and an empty pBC-06 vector through floral dip method (Jefferson, 1989). The transformed plants were selected on kanamycin (50 μg/ml). T₂ transgenic plants were raised and confirmed through PCR analysis (data not given). Five independent transgenic lines were developed and two transgenic lines were evaluated against M. incognita infection. Nematode infection assays were performed in T₃ plants and the numbers of galls were calculated.

For gene expression study, RNA isolated from dsRNA expressing transgenic plant and control plant used as template, with setup of three biological replicates along with three technical replicates for each sample. To ensure purity of the master mix and reaction mix setup, a non-template control was included in every plate. A 20 μl reaction volume consisting of SYBR FAST qRT-PCR Master Mix (2x) Universal (KAPA Biosystems) and 10 pmol of each primer was used in all qRT-PCR reactions. The qRT-PCR reactions were performed on a StepOnePlus™ Real-Time PCR system with the following cyclic conditions: initial heating temperature of 95°C for 10 min followed by 40 cycles of 95°C for 15 s and 61°C for 45 s. The PCR products were exposed to melting curve analysis; the conditions were incubation at 60–95°C with a temperature increment of 0.3°C s⁻¹ (Applied Biosystems®). The threshold cycle values were normalized by nematode 18S ribosomal RNA (Supplementary Table 1) as endogenous control and fold changes of the target gene were calculated by 2^−ΔΔCt method (Livak and Schmittgen, 2001).

Candidate NRRS genes were identified in three steps: (a) an analyses of microarray data was performed to find genes that were induced by nematode infection and genes with root-specific expression patterns, (b) these root-specific and nematode-responsive gene sets were collated to find genes with desired NRRS gene expression pattern, and (c) finally, a cis-element based screening of candidate NRRS gene promoters was performed to select final candidates for downstream analyses.

The Genevestigator analysis i.e., meta-analysis of microarray datasets identified 1,374 probe sets that showed statistically significant and consistent up-regulation (>1.5-fold-change, log2 scale) in different (n = 6) root zones (Figure 1). On other hand, the stand-alone analysis, comparing root samples with vegetative shoot, inflorescence and leaf samples identified a total of 672 probe sets from three developmental time points (7, 17, and 21 days). In process, we also identified, 79 probe sets that were consistently enriched in root across all three time points investigated in stand-alone analysis (Figure 2A, Supplementary Table 2) and 336 probe sets that were common between Genevestigator and stand-alone analysis (Figure 2B). The low overlap between probes sets from Genevestigator and stand-alone analysis could be because of the additional data processing and summarization steps that Genevestigator performs to reduce batch effects and to provide robust expression estimates from different microarray platforms. In addition to root-specific probe sets, the Genevestigator analysis for nematode-induced genes (>1.5-fold-change, log2 scale) from “early” and “late” stages of nematode infection yielded 780 probe sets, with 42 probe sets showing a constant up-regulation in both stages (Figure 2C, Supplementary Table 3).

These probe sets were converted to Arabidopsis Gene Identifiers (AGIs) and collated to identify candidate NRRS genes. If multiple probes represented a single gene then the probe with highest median expression across samples was selected to represent the gene, and if a single probe represented multiple genes then all of these were included in final set. This conversion resulted in 700 root-specific AGIs from stand-alone analysis, 1,452 root-specific AGIs from meta-analysis and 850 nematode-responsive AGIs (Table 1). The nematode-responsive (NR) AGIs (n = 700) were then compared with root-specific (RS) AGIs from both the Genevestigator analysis (n = 1,452) and the stand-alone analysis (n = 700) to identify genes with nematode responsive and root-specific expression (n = 51). We refer to these 51 genes as candidate NRRS genes; these included 25 and 34 genes from NR vs. RS comparisons, where RS genes corresponded to stand-alone (Supplementary Table 4) and Genevestigator analysis (Supplementary Table 5) respectively. Both these categories NR vs. RS from stand-alone and NR vs. RS from Genevestigator analysis has eight common genes (Supplementary Table 6). Finally, to select genes that have a low expression under normal developmental stages, which is a desired characteristic for gene promoters that will drive expression of RNAi construct, we filtered genes based on average expression values in root tissues (see Supplementary Text Presentation 1). Six genes passed the low basal intensity (LBI) filter and were considered for next step i.e., a cis-element based screening (Table 2).

The six candidate NRRS genes were screened for the presence of cis-regulatory elements (Table 3) that have been experimentally validated to play a role in root-specificity i.e., AS1, Sorlip1, and FaRB7, and typical for promoters of genes that are upregulated in nematode feeding site (NFS) upon infection i.e., TobRB7 box A/B etc. The purpose of this regulatory element based screening was to further reduce the number of genes selected for downstream molecular cloning and in planta validation. Out of six NRRS genes, only three genes—At1g74770, At2g18140, and At1g48670 (Figure 3) included regulatory elements from both the nematode responsive and root specific (NRRS) categories in their promoter regions. ATCEL2, earlier implicated in nematode-responsiveness as well as root specificity used as positive control (Wieczorek et al., 2008). Gene At2g39230, coding for a LOJ protein (Saha et al., 2007) served as negative control and had no nematode-responsive as well as root-specific elements except for single FaRB7 element. Two genes, At5g58780 and At3g29775 lacked root-specific elements whereas the At1g80320 gene promoter contained only the E-box motif out of all the studied elements (n = 12). Surprisingly, except for the E-Box motif, all other putative motifs described by Sukno et al. (2006) were found to be distributed in all three categories of genes, TobRB7, Atcel1, Hahsp17, and Lemmi9 that show elevated expression upon nematode infection in the NFS (Opperman et al., 1994; Escobar et al., 1999, 2003; Sukno et al., 2006), genes that are downregulated in NFS—AtPAl1, AtTIP, and AtANT1 (Hammes et al., 2005) and UBP22/At5g10790 that does not show any significant difference in expression pattern in response to nematode infection (Favery et al., 2004). Therefore, these putative motifs (W-Box, ElRE, ERE, and P-Box) proposed by were not included in this step to filter NRRS gene based on presence of cis-regulatory elements (Table 3).

To find new candidates that may play role in upregulation of nematode-responsive genes in NFS, we used promoters from two different set of genes. First set consisted of genes that are well-studied for role in establishing compatible plant nematode interaction and show elevated expression upon infection (Lemmi9, TobRB7, and Hahsp17) along with the six NRRS genes predicted in this study (At1g74770, At1g48670, At2g18140, At3g29775, At5g58780, and At1g80320). The second set consisted just the six NRRS gene promoters from this study. We identified four motifs which were common between promoters of well-studied nematode-responsive genes and our NRRS genes (Figure 4A). In addition, we found three new motifs that are present only in NRRS genes (Figure 4B). Later on, MEME analysis has been carried out for the promoter region of all 51 genes to identify NRRS motif (Supplementary Figure Presentation 1; Supplementary Figure 1). To identify the transcription factors that might bind to these sites, we used POXO (Kankainen et al., 2006) and STAMP (Mahony and Benos, 2007) but no statistically significant similarity to known motifs was found. Therefore, further characterization of these novel motifs is required to establish their functional roles in NRRS activity.

Genes that passed the cis-regulatory element based criteria were evaluated for expression patterns in response to nematode infection before proceeding for molecular cloning their promoter regions. First, the expression patterns of six genes, UBP7, OXA1, Actin, RPN7, ATPase (Jammes et al., 2005), and UBQ10 (accession number DQ793132.1) was investigated in response to nematode infection for selection of appropriate internal controls. UBQ10 gene showed a consistent expression in both wild-type and nematode infected plants and was used as an internal controls for quantitative real time (qRT) PCR assay. The expression of six NRRS genes—At1g74770, At2g18140, At1g80320, At1g48670, At5g58780, and At3g29775 was studied at two different time points (10 and 21 dpi) along with At5g26530 gene, earlier implicated in nematode-responsiveness, which served as positive control (Kumar et al., 2016). The relative expression of At2g18140 and At1g74770 was maximum at 10 and 21 dpi, respectively as compared to control (Figure 5).

To experimentally confirm the root-specific and nematode-responsive behavior of the shortlisted genes, we generated promoter::GUS constructs for both the putative NRRS genes. Five independent transgenic lines for each promoter construct and 15 plants from each independent line were evaluated. In transgenic plants harboring pAt1g74770::GUS constructs, the nematode infected plants (T3-P6) showed strong GUS expression in roots at 21 dpi (Figure 6A complete plant right side and Figure 6C) as compared to uninfected (control) plants (Figure 6A complete plant left side and Figure 6B). On examination of pAt1g74770::GUS lines under microscope, large numbers of galls were visible along with blue staining throughout the root system (Figure 6D). In transgenic plants harboring pAt2g18140::GUS constructs, the nematode infected transgenic line (T3-P2) revealed maximum GUS activity only in galls during early stages of infection at 10 dpi (Figures 6F,G) as compared to uninfected (control) plants (Figures 6E,H). The microscopic examination revealed strong GUS activity in the gall (Figure 6I). Thus, the histochemical GUS assay of promoter-reporter gene constructs upon nematode infection confirmed the root-specificity and nematode-responsiveness of both the promoters. GUS activity was not detected in control (uninfected) transgenic plants for either of the promoters used.

NRRS promoter (At1g74770) was used for expressing dsRNA of a nematode gene, splicing factor, to evaluate the efficacy of NRRS promoters in inducing HD-RNAi mediated resistance in Arabidopsis. Of the five independent transgenic lines containing promoter At1g74770::splicing factor, two transgenic lines (At1g74770::SF E1 and At1g74770::SF E2) were tested against nematode. For each transgenic line 15 plants were evaluated. Transgenic lines of the splicing factor gene exhibited 20–32% reduction in number of galls compared to control plants (Figure 7). To study the effect of transgenic lines on gene expression in nematodes, females were isolated from transgenic lines expressing splicing factor dsRNA and control plants at 42 dpi. Quantitative Real-Time PCR (qRT-PCR) analysis revealed slightly reduced expression of splicing factor gene in the females isolated from transgenic plants compared to control, indicating that the splicing factor gene, driven by the NRRS promoter, has not been silenced significantly (Figure 8).

The NRRS gene set identified in this study comprises of several putative nematode responsive and root specific genes such as those from family of expansins, peroxidases, and PINFORMED auxin transporter, that have well-established roles in plant–nematode interactions. Expansins facilitate nematode penetration into the roots (Gal et al., 2006; Fudali et al., 2008); up-regulation of expansins and expansin-like genes are reported in developing galls in Arabidopsis (Jammes et al., 2005), tomato (Gal et al., 2006) and in soybean syncytia induced by soybean cyst nematodes (Ithal et al., 2007). Peroxidases play an essential role in cell wall strengthening and are also reported to be up-regulated in nematode-infected roots (Ithal et al., 2007). The up-regulation of both expansins as well as peroxidases is proposed to be part of mechanism that maintains a balance between cell wall loosening and cell wall strengthening during feeding site development of nematodes (Gheysen and Mitchum, 2009). PIN are efflux facilitators that mediate auxin transport; these express in specific regions of roots and show elevated expression upon nematode infection (Grunewald et al., 2009). The presence of these gene members of families that have established roles in mediating successful plant nematode interactions such as ATEXPA14, ATEXP6, ATEXP4, ATEXPB3, and ATAXP4 in nematode-responsive gene set along with ATCYP86 (a peroxidase), ATRFNR2 in NRRS gene set shows the effectiveness of such strategy to identify tissue-specific and pathogen responsive genes. In addition to, the presence of cis-regulatory elements, other than those used for filtering NRRS genes, such as those that play role wounding, pathogen response, plant defense signaling, disease, and pathogen responses and most importantly, the elements involved in organ specificity (Table 3) further supports that validity of approach used for selection of NRRS genes. Thereby, demonstrating that expression-based surveys for identifying condition-specific gene promoters may have a potential utility in developing plants with host-mediated response to biotic and abiotic response.

In planta prm::GUS constructs of both NRRS genes tested in this study (At1g74770 and At2g18140) demonstrates highly restricted expression in roots in response to nematode infection. The At2g18140, expression is highly restricted to galls that encapsulates NFS with no trace in aerial portion of plant. At2g18140 encodes for a peroxidase protein and plant peroxidases have earlier been implicated in host-plant parasitic nematode interaction (Vercauteren et al., 2001; Jammes et al., 2005; Severino et al., 2012), including Coffea canephora sp. Thereby, confirming important and likely conserved roles of peroxidases in response to PPNs. The At1g74770 gene promoter displays GUS expression throughout the root, strong and constitutively present across all cell layers. Thereby, it seems like an ideal candidate to develop nematode-resistant plants.

The number of galls in Arabidopsis transgenic plants expressing pAt1g74770::splicing factor gene was lower (20–32%) compared to control plants. However, transgenic tobacco plants expressing dsRNA of splicing factor gene through CaMV35S promoter exhibited about 95% reduction in gall formation as well as in number of nematode females (Yadav et al., 2006). Similarly, the splicing factor gene driven by 35S promoter in Arabidopsis transgenic lines exhibited up to 71% reduction in gall number (Kumar et al., 2017). In the last few years, several promoters have been identified in various crops including tobacco (Opperman et al., 1994), LEMMI9 in tomato (Escobar et al., 1999), Hahsp17.7G4 in Helianthus annuus (Escobar et al., 2003), AtCel-1 in Arabidopsis (Sukno et al., 2006), AtWRKY23 (Grunewald et al., 2008), and ZmRCP-1 in maize, banana, and plantains (Onyango et al., 2016). None of the promoters identified have been utilized for driving the production of dsRNA of nematode genes for HD-RNAi silencing. Only TobRB7, a gall specific promoter was used to drive M. javanica gene, MjTis11, in tobacco but no sign of gall reduction was observed in nematode infected plants (Fairbairn et al., 2007). The lack of silencing observed in the TobRB7 promoter lines was attributed to the weakness of the TobRB7 promoter. However, we observed up to 32% reduction in nematode infection by using our NRRS promoter as compared to 71% reduction with use of CaMV35S promoter. The CaMV35S promoter, being constitutive in nature is likely to produce more dsRNA as compared to the conditional NRRS promoter used in the present study. However, the use of 35S promoter has to be exercised with caution since dsRNA is produced in all tissues all the time and can lead to undesirable effects in transgenic plants. There is a need to identify and evaluate large number of tissue-specific promoters and use the best ones for developing nematode resistant plants using HD-RNAi approach.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.