Introduction

Virus induced gene silencing (VIGS) is a powerful loss of function technology for functional characterization of plant genes. It exploits host RNA interference (RNAi) pathway to cleave the introduced recombinant viral transcript containing plant gene fragments, into siRNAs which subsequently degrades the endogenous target gene transcript in a sequence specific manner (Lu 2003; Burch-Smith et al. 2004; Unver and Budak 2009). VIGS transiently down-regulates the expression of target genes, which leads to the emergence of rapid loss of function phenotype in plants within a matter of 2–3 weeks. This system is now considered as powerful tool for plant functional genomics, especially for those species which are recalcitrant to transformation (Bernacki et al. 2010; Zhang et al. 2010). Because the method is rapid, robust and lends itself to high-throughput techniques, VIGS acts as an outstanding alternative platform for large-scale functional screening of target genes (Ye et al. 2009; Yuan et al. 2011). The significance of VIGS for the functional analysis of genes in dicots (Arabidopsis, bean, tobacco, tomato, etc.) as well monocots (barley, rice, wheat, maize etc.) is well characterized (Holzberg et al. 2002; Scofield et al. 2005; Van Der Linde et al. 2011; Lee et al. 2012; Mei et al. 2016; Wang et al. 2016). For example, VIGS vectors have been derived from Tobacco rattle virus (Ratcliff et al. 2001; Liu et al. 2002; Bachan and Dinesh-Kumar 2012), Apple latent spherical virus (Sasaki et al. 2011), Bean pod mottle virus (Zhang et al. 2009, 2010) for dicots (tomato, apple, soybean) and Rice tungro bacilliform virus (Purkayastha et al. 2010), Barley stripe mosaic virus (Holzberg et al. 2002; Lee et al. 2012), Brome mosaic virus (Ding et al. 2006; Van Der Linde et al. 2011), Cucumber mosaic virus (Wang et al. 2016) and Foxtail mosaic virus (Liu et al. 2016; Mei et al. 2016) for monocots (rice, maize and wheat) and employed for functional characterization of genes. Interestingly, among the recently developed VIGS vectors, the Bamboo mosaic virus (BaMV)-based vector has been used for the functional analysis of genes in monocots as well as dicots (Liou et al. 2014).

Environmental factors play a crucial role for plant growth and development and also influence VIGS efficiency in plants. TRV-based gene silencing efficiency in tomato is enhanced at low temperature and low humidity (Fu et al. 2006). Similar observations were reported in cotton plants (Tuttle et al. 2008). Conversely, gene silencing has also been reported to be inhibited at low temperature in Nicotiana benthamiana protoplasts transfected with Cymbidium ringspot virus (Szittya et al. 2003). VIGS-mediated gene silencing efficiency is dependent upon the concentration of agroinocula and temperature (Wang et al. 2013) and insert orientation (Lacomme et al. 2003; Hein et al. 2005; Zhang et al. 2010). Therefore, conditions need to be standardized to achieve optimal and consistent silencing with any VIGS system.

We had earlier developed a gene silencing system for rice using RTBV based VIGS system (pRTBV-MVIGS, Purkayastha et al. 2010). To standardize the optimal conditions for the use of this system, we took inoculation method, temperature and humidity into account using the marker gene, phytoene desaturase (pds). For further extension of pRTBV-MVIGS mediated silencing, an additional marker gene, magnesium chelatase (chl) was also used, which produces a visible phenotype upon silencing. Magnesium chelatase participates in the biosynthetic pathway for chlorophyll synthesis and chloroplast development by catalyzing the insertion of Mg2+ into protoporphyrin IX, the last common intermediate precursor in chlorophyll biosynthesis (Zhang et al. 2006). Its deficiency causes yellowing of leaf (chlorosis) in higher plants. We used the gene encoding one of the three subunits of magnesium chelatase, chlH of rice in sense, antisense and hairpin orientation in pRTBV-MVIGS system, to investigate the effect of insert orientation on the efficiency of VIGS in rice. Finally, we silenced the gene Xa21, well known to confer resistance to the widespread bacterial pathogen Xanthomonas oryzae pv. oryzae, (Xoo) responsible for Bacterial leaf blight (BLB, Song et al. 1997), in rice cv. PB6 (Ellur et al. 2016) to study the potential of this VIGS system for rice functional genomics.

Materials and methods

Construction of recombinant viral vector

The recombinant pRTBV-MVIGS containing as-pds partial fragment (530 bp from rice cloned in antisense orientation, has already been described earlier (Purkayastha et al. 2010). The partial O. sativa chlH gene (Genbank accession no. EU569725.1) was amplified from rice cDNA using gene-specific primers having appropriate restriction enzyme sites and cloned in sense (318 bp), antisense (350 bp) and hairpin orientations (sense, 318 bp + antisense 318 bp) in the vector pHannibal (Table 1; Wesley et al. 2001). The inserts were then subcloned into pRTBV-MVIGS. Rice (O. sativa) Xa21gene fragment (560 bp, Genbank accession no. AB212798.1) was PCR amplified with gene specific primers (Table 1), using cDNA made from rice cv. PB6 carrying intact Xa21. PCR amplified Xa21 fragment was cloned first in T/A vector, followed by subcloning in pRTBV-MVIGS vector respectively in antisense orientation. All constructs (empty pRTBV-MVIGS as well as recombinants) were transformed in the Agrobacterium (strain EHA-105) for agroinoculation in rice plants.

Table 1 Sequence and origin of primers used to amplify chlH subunit fragment of Mg chelatase gene and Xa21 fragment
Full size table

Agroinoculation for silencing pds and chlH

Rice plants cv. PB1 (20 day old) were agroinoculated (Purkayastha et al. 2010) with different compositions of resuspension buffers and kept at variable temperature and humidity (Table 2) to standardize the silencing efficiency of the pds gene. The effect of growing the secondary culture of agrobacterium (sub-culturing primary agrobacterium culture into larger volume of LB broth) in the presence of acetosyringone was also assessed. To study the effect of insert orientation on chlH silencing efficiency, The secondary agrobacterium culture for each construct was supplemented with 1 mM MES and 20 μM acetosyringone and the resulting pellet was reconstituted in resuspension buffer (10 mM MgCl2, 10 mM MES and 500 μM acetosyringone) having OD 0.6–1.0 and incubated for 3 h at room temperature. Rice plants were agroinoculated with 1 ml syringe at the meristematic region and kept at 28 °C and 80% humidity in a growth chamber.

Table 2 Comparative silencing efficiency of pds gene showing leaf-streaking phenotype at variable temperature, humidity and composition of resuspension buffer in rice plants inoculated with pRTBV-MVIGS and pRTBV-MVIGS-aspds, at 20 days post inoculation
Full size table

Silencing phenotype and real-time PCR

The plants inoculated with recombinant pRTBV-MVIGS having pds and chlH partial cDNA fragments were observed for corresponding gene silencing phenotypes in the emerging leaves. The emerging leaves showing silenced phenotype were harvested and further validated for gene silencing through quantitative estimation of the corresponding transcript levels by real-time PCR. The harvested leaves were used for total RNA isolation using RNeasy plant mini kit (Qiagen) and cDNA was synthesized from 2 μg isolated RNA using high capacity cDNA synthesis kit (Applied Biosystems).

The Real-time PCR primers (Table 1) specific to pds, chlH and the endogenous control ubiquitin 5 (UBQ5) were designed using Primer Express® software (Applied Biosystem). cDNA from the plant inoculated with pRTBV-MVIGS and pRTBV-MVIGS-pds/chlH was used for qRT-PCR using Fast SYBR® Green Master Mix (Applied Biosystem) according to manufacturer’s instructions. The relative transcript accumulation of pds and chlH in the emerging leaves from pRTBV-MVIGS-pds/chlH inoculated plants were calculated and compared from pRTBV-MVIGS inoculated plants using comparative CT method (Schmittgen and Livak 2008).

Chlorophyll content and chlorophyll fluorescence

Chlorophyll content was estimated from the emerging leaves at 20 dpi in the pRTBV-MVIGS and pRTBV-MVIGS-chlH inoculated plants using methods of Arnon (1949) and Hiscox and Israelstam (1980) with slight modifications. Chlorophyll was extracted without grinding the leaf tissues by heating the 100 mg of leaf samples in 3 ml DMSO at 65 °C in a water bath for 1 h, the extracted solution volume was made up to 5 ml with DMSO. The extracted chlorophyll was measured at wavelengths of 645 and 663 nm using spectrophotometer for each sample respectively as described (Arnon 1949). Chlorophyll fluorescence was measured from the emerging intact leaves in 30 min dark adapted pRTBV-MVIGS and pRTBV-MVIGS-chlH inoculated plants using portable Pulse modulated chlorophyll fluorescence meter (Walz GmbH, Effeltrich Germany). Chlorophyll fluorescence was recorded in triplicate for each emerged leaf from both pRTBV-MVIGS and pRTBV-MVIGS-chlH inoculated plants. The chlorophyll fluorescence values (F v/F m) obtained, represent the maximum quantum yield of Photosystem II (PS II), which is the ratio of variable (F v) to maximum fluorescence (F m, Puteh et al. 2013).

Agroinoculation, detection of Xa21 silencing and BLB disease susceptibility

Rice cv PB6 (BLB susceptible) and PB6 carrying Xa21 gene plants (BLB resistant) were first inoculated with Xoo (BXO43) using clip inoculation method (Babu et al. 2003) to analyse the BLB disease susceptibility at 7 days post inoculation in susceptible and resistant plants. pRTBV-MVIGS and pRTBV-MVIGS-Xa21 were agroinoculated in 20 days old PB6-Xa21 plants using standard method of rice agroinoculation (Purkayastha et al. 2010). The agroinoculated rice plants were screened for Xa21 silencing at 15 day post agroinoculation through real time PCR as described above using gene-specific primers for Xa21 and ubiquitin (UBQ5, as an internal control, Accession no. AK061988). The mock-silenced (using pRTBV-MVIGS) and Xa21-silenced (using pRTBV-MVIGS-Xa21) plants were challenged with Xoo using clip inoculation method and kept 28 °C temperature and 80% humidity. BLB disease incidences were measured at 7 day post-Xoo inoculation in both non-silenced and Xa21-silenced plants by measuring the mean lesion lengths (MLL) and average percentage diseased leaf area (%DLA) in the Xoo inoculated plants (two leaves/plant). The %DLA was calculated using the formula: %DLA = lesion area/leaf area × 100.

Results

Variable growth conditions affect pds silencing in rice

The efficiency of pRTBV-MVIGS mediated rice pds silencing at variable temperature, humidity and composition of resuspension buffer (MgCl2, MES and acetosyringone) was determined under three different sets (I, II and III) of plant growth conditions (Table 2). At 20 day post inoculation (dpi), varying percentages of plants (6–60%) inoculated with pRTBV-MVIGS-pds displayed white streaks (photobleaching) in the emerging leaves, the plants inoculated with empty pRTBV-MVIGS and un-inoculated plants did not show any streaks in emerging leaves (Fig. 1A, B). Out of three sets (I, II, III) used, set III produced the most efficient gene silencing.

Fig. 1
figure 1

A Phenotype in the emerging leaves of rice plants at 20 dpi. a Leaf of a un-inoculated plant did not show any silencing phenotypic symptoms, b leaf of pRTBV-MVIGS (empty vector) inoculated plant did not show any silencing phenotype and c leaves of pRTBV-MVIGS + Antisense-pds inoculated plant showed white streaking phenotype pointed in yellow arrowhead. B Enlarged picture of individual leaf at 20 dpi: Un-inoculated, pRTBV-MVIGS (empty vector) and pRTBV-MVIGS + Antisense-pds inoculated. C pds transcript accumulation in plants inoculated with pRTBV-MVIGS and pRTBV-MVIGS-Antisense-pds, labels at x-axis shows the individual plants. UI un-inoculated plant, MVIGS pRTBV-MVIGS, empty vector and pds (1–6) pRTBV-MVIGS-Antisense-pds. Relative transcript level of samples is shown at y-axis. The error bars indicate the standard deviations of the mean of three independent biological replicates (color figure online)

Full size image

The accumulation of pds transcripts in plants inoculated with pRTBV-MVIGS and pRTBV-MVIGS-pds (showing white streaking phenotype) was determined by real time PCR from set III. Relative pds transcript levels for individual samples were plotted as a bar diagram (Fig. 1C), which showed 40–80% decrease in transcript accumulation of pds in pRTBV-MVIGS-pds inoculated plants as compared to pRTBV-MVIGS plants. The percentage silencing values shown represents the average of three samples for each plant. The relative pds transcript accumulation obtained thus corroborate pds silencing phenotype in inoculated rice plants.

VIGS mediated chlH silencing develops orientation specific phenotype

Rice plants inoculated with pRTBV MVIGS-Sense/Antisense/Hairpin chlH constructs showed leaf-yellowing phenotypes in the emerging leaves at 20 dpi with variable intensities. The antisense and hairpin chlH constructs were found to exhibit almost equal leaf-yellowing, which was more intense as compared to plants inoculated using sense chlH construct (Fig. 2a–d; Table 3). The plants inoculated with pRTBV-MVIGS (empty vector) did not display any leaf yellowing in the emerging leaves, similar to un-inoculated plants.

Fig. 2
figure 2

Silencing phenotypes in the emerging leaves of rice plants inoculated with a pRTBV-MVIGS (empty vector), b pRTBV-MVIGS + Sense-chlH, c pRTBV-MVIGS + Antisense-chlH and d pRTBV-MVIGS + Hairpin-chlH

Full size image
Table 3 Number of plants inoculated and number of plants showing leaf yellowing phenotype inoculated with pRTBV-MVIGS, pRTBV-MVIGS + Sense(S) chlH, pRTBV-MVIGS + Antisense(AS) chlH and pRTBV-MVIGS + Hairpin(HP) chlH constructs in three (s. no. 1–3) independent experiments
Full size table

Silencing of chlH affects chlorophyll content and maximum quantum yield (F v/F m)

The amount of chlorophyll a, b and total chlorophyll were measured in emerging leaves of plants inoculated with either pRTBV-MVIGS or pRTBV-MVIGS-chlH (sense/antisense/hairpin). In plants inoculated with pRTBV-MVIGS, three different forms of chlorophyll (a, b and total) measurements were found to be in the range of 0.32–0.35, 0.18–0.2 and 0.44–0.48 mg per g of fresh leaf weight respectively in plants. On the other hand, plants inoculated with pRTBV-MVIGS-chlH showed chlorophyll a, b and total chlorophyll in a range of 0.22–0.3, 0.12–0.16 and 0.32–0.4 mg per g of fresh leaf weight respectively across sense, antisense and hairpin constructs, representing significantly low chlorophyll accumulation as compared to pRTBV-MVIGS inoculated plants. Average chlorophyll contents (chl-a/chl-b/total-chl) in the plants agroinoculated with antisense chlH (0.23 ± 0.02/0.13 ± 0.01/0.34 ± 0.02) and hairpin chlH (0.24 ± 0.02/0.13 ± 0.01/0.34 ± 0.02) were comparable, but lower than plants inoculated with sense chlH (0.28 ± 0.01/0.14 ± 0.01/0.37 ± 0.02) milligram per gram of fresh leaf weight respectively. Each value represents the mean ± standard deviation of six replicates with statistical significance between means for each construct (p < 0.01; Student’s t test) (Fig. 3a–c).

Fig. 3
figure 3

Chlorophyll contents and chlorophyll fluorescence in inoculated plants at 20 dpi. Chlorophyll contents (chl a, chl b and total chl) in pRTBV-MVIGS + Sense-chlH (a), pRTBV-MVIGS + Antisense-chlH (b), pRTBV-MVIGS + Hairpin-chlH (c) inoculated plants as compared to pRTBV-MVIGS (empty vector) inoculated plants. In a–c, MVIGS and S/AS/HP-chlH (1–6) at the x-axis represent individual plants inoculated with pRTBV-MVIGS (empty vector) and pRTBV-MVIGS-S/AS/HP-chlH constructs. The experiment had three biological replicates and bar values with error bars indicate the mean value ± standard deviation with statistically significant differences between means for each construct and empty vector (*p < 0.01; Student’s t test). d Comparison of chlorophyll fluorescence represented as F v/F m ratio in pRTBV-MVIGS and pRTBV-MVIGS + Sense/Antisense/Hairpin-chlH inoculated plants. The F v/F m ratio in pRTBV-MVIGS inoculated plants was taken as control. Numbers 1–6 at x-axis indicates individual plants. The experiment was conducted in three biological replicates and line graph with error bars indicate the mean value ± standard deviation with statistically significant differences between means for each construct (*p < 0.05; Student’s t test)

Full size image

In addition to chlorophyll contents, chlorophyll fluorescence from pRTBV-MVIGS and pRTBV-MVIGS-chlH was also determined using pulse modulated chlorophyll fluorescence meter to validate the reduction in chlorophyll biosynthesis. Chlorophyll fluorescence (F v/F m) represents the maximum quantum yield of PS II (first protein complex in the light-dependent reactions of oxygenic photosynthesis) located in the thylakoid membrane of plant leaves. The pRTBV-MVIGS inoculated plants showed the F v/F m values above 0.8 while these values varied between 0.7–0.8 among sense, antisense and hairpin chlH inoculated plants. Like chlorophyll contents, average F v/F m values were comparable in plants inoculated with antisense (0.75 ± 0.01) and hairpin (0.74 ± 0.01) chlH, but lesser than those inoculated with sense chlH (0.79 ± 0.01) showing mean F v/F m ± standard deviation of six replicates with statistical significance between means for each construct (p < 0.05; Student’s t test, Fig. 3d).

Transcript accumulation in chlH silenced plants

In order to determine the dependence of the orientation of chlH on the efficiency of silencing, the chlH transcript accumulation in plants inoculated with pRTBV-MVIGS-sense chlH/antisense chlH/hairpin chlH showing leaf yellowing phenotype at 20 dpi, was validated through real-time PCR. The decrease in relative chlH transcript accumulation was 20–50% in plants inoculated with pRTBV-MVIGS-sense chlH, 40–80% in plants inoculated with pRTBV-MVIGS-antisense chlH and 30–75% in plants inoculated with pRTBV-MVIGS-hairpin chlH respectively as compared to the transcript level in pRTBV-MVIGS inoculated plants (Fig. 4a–c). The chlH transcript accumulation obtained in plants inoculated with sense, antisense and hairpin chlH thus corresponded with their respective leaf yellowing phenotypes.

Fig. 4
figure 4

Bar graph representation of chlH silencing in agroinoculated plants: a relative chlH transcript accumulation of pRTBV-MVIGS + Sense-chlH inoculated plants as compared pRTBV-MVIGS inoculated plants. b Relative chlH transcript accumulation of pRTBV-MVIGS + Antisense-chlH inoculated plants as compared pRTBV-MVIGS inoculated plants. c Relative chlH transcript accumulation of pRTBV-MVIGS + Hairpin chlH inoculated plants as compared pRTBV-MVIGS inoculated plants. In a–c, UI (un-inoculated), MVIGS and S/AS/HP-chlH (1–6) at the x-axis represent individual plants inoculated with pRTBV-MVIGS (empty vector) and pRTBV-MVIGS-S/AS/HP-chlH constructs. The error bars indicate the standard deviation of the mean value and experiment was repeated three times

Full size image

Silencing the BLB resistant gene, Xa21

Bacterial leaf blight disease susceptible and resistant rice cvs. PB6 and PB6 carrying Xa21 (PB6-Xa21), were clip inoculated with Xoo, which resulted in severe necrotic lesions in PB6 plants as compared to mild lesions in the resistant variety PB6-Xa21 at 7 day post clip inoculation (Fig. 5a). Further, Xa21 was silenced in PB6-Xa21 plants by agro-inoculating with pRTBV-MVIGS-Xa21 (Xa21 in antisense orientation), along with PB6-Xa21 plants inoculated with pRTBV-MVIGS alone as control. At 15 dpi, 50–90% decrease in the relative transcript accumulation of Xa21 was observed in pRTBV-MVIGS-Xa21 as compared to pRTBV-MVIGS alone in agro-inoculated plants (Fig. 5b).

Fig. 5
figure 5

BLB disease susceptibility in non-silenced and silenced plants: a PB6 (susceptible) and PB6-Xa21 (resistant) plant leaves inoculated with Xoo. b Non-silenced (pRTBV-MVIGS inoculated) PB6-Xa21 plants showed nominal disease symptoms with very less necrotic leaf lesion in Xoo inoculated leaves. c Xa21-silenced (pRTBV-MVIGS-Xa21 inoculated) PB6-Xa21 plants showed much greater disease symptoms with higher lesion lengths in the Xoo inoculated leaves as compared to non-silenced plants. MVIGS and Xa21 (1–6) at x-axis represent individual plants inoculated with pRTBV-MVIGS and pRTBV-MVIGS-Xa21 respectively. The error bars indicate the standard deviation of the mean value and experiment was repeated three times

Full size image

BLB disease susceptibility was scored in plants inoculated with pRTBV-MVIGS (non-silenced) and those with pRTBV-MVIGS-Xa21 (Xa21 silenced) to validate the effect of Xa21 silencing. Both non-silenced and Xa21-silenced plants were then challenged with Xoo. At 7 dpi, more severe symptoms (Fig. 5c) were observed in plants silenced for Xa21 (lesion lengths of 3.0–11.5 cm) as compared to non-silenced plants (1.0–3.2 cm), the difference being statistically significant (p < 0.001; Student’s t test). The leaf lesion lengths, total leaf length and leaf width were also measured in all inoculated leaves of the plants to determine the mean lesion lengths and percentage diseased leaf area (%DLA) in both non-silenced and Xa21-silenced plants respectively. A total of fifteen plants were tested for analyzing lesion length and %DLA. The average mean lesion length (average cm ± SD) of the total number of non-silenced and Xa-21 silenced plants was measured to be 1.75 ± 0.71 and 5.02 ± 2.42 respectively. Similarly, average %DLA of non-silenced and Xa-21 silenced plants was calculated to be 8.75 ± 4.3 and 21.9 ± 9.7 showing statistical significance in the values (p < 0.05; Student’s t test) respectively. Figure 6a, b shows the increased mean lesion lengths and %DLA upon Xa21 silencing in a graphical manner (Fig. 6a, b).

Fig. 6
figure 6

Evaluation of BLB disease pathogenicity: a Graphical representation of mean lesion lengths (two leaves/plant) with error bars (mean ± standard deviation) for Xoo inoculated leaves in non-silenced (pRTBV-MVIGS inoculated) and Xa21 silenced (pRTBV-MVIGS-Xa21 inoculated) plants. An asterisk represents a statistically significant difference in values between two treatments (*p < 0.001; Student’s t test). b Average percentage diseased leaf area (%DLA, two leaves/plant) of Xoo inoculated leaves in non-silenced (pRTBV-MVIGS inoculated) and Xa21 silenced (pRTBV-MVIGS-Xa21 inoculated) plants. Numbers (1–15) at x-axis in a and b represent individual plants inoculated with pRTBV-MVIGS and pRTBV-MVIGS-Xa21. An asterisk represents statistically significant differences in the values between two treatments (*p < 0.05; Student’s t test). The experiment was repeated thrice

Full size image

Discussion

Virus induced gene silencing is potentially a very powerful tool to determine gene functions in the important food crop rice, especially to meet the emerging challenges of global warming and shrinking land and water resources, factors most likely to impact rice productivity in the not-so-distant future. Here, we standardize the optimal gene silencing conditions for use of the pRTBV-MVIGS system developed earlier by us for rice (Purkayastha and Dasgupta 2009; Purkayastha et al. 2010; Kant et al. 2015). Here, first using the marker gene pds, we tested various conditions of agroinoculation and growth conditions of rice plants, on the efficiency of gene silencing. The inclusion of 20 µM acetosyringone in the growth medium of the agrobacterium during its growth prior to inoculation to the plant (secondary culture) dramatically increased the pds silencing efficiency, as assessed by the number of plants showing the streaking phenotype in the emerging leaves. Acetosyringone is a known inducer of T-DNA transfer from Agrobacterium to plant cells (Godwin et al. 1991; He et al. 2010; Manfroi et al. 2015), but its enhancing effect, when added at the secondary culture may indicate a priming event of the T-DNA transfer process, in addition to the actual transfer process during the agro-inoculation. The enhancement of the proportion of plants showing streaking phenotype (indicative of pds silencing) upon slight increase of the temperature at which plants were grown following agro-inoculation, is contrary to some of the earlier reports of enhancement of VIGS silencing at lower temperatures (Fu et al. 2006; Nethra et al. 2006; Tuttle et al. 2008). However, as shown by (Szittya et al. 2003), the RNA silencing machinery is activated at higher temperatures, which should result in better VIGS-mediated silencing. Hence, our results of enhancement of the proportion of plants showing pds silencing at higher temperatures is supported by the above report, although variations of this general theme cannot be ruled out in plants adapted to varying temperatures. Therefore, we conclude that condition III is the most efficient, among the three conditions used, for gene silencing in rice using pRTBV-MVIGS.

Next, we used the marker gene chlH to study the effect of orientation of the insert in the VIGS vector on silencing efficiency. Our observation demonstrates that antisense and hairpin orientations gave rise to higher silencing as compared to sense orientation is somewhat similar to the reports of Holzberg et al. (2002) and Lacomme et al. (2003), both groups having used the BSMV system, that genes inserted in hp or sense-antisense orientation, show higher silencing than when inserted in antisense or sense orientations. Pacak et al. (2010), using the BMV system, Zhang et al. (2009, 2010), using the BPMV system reported that antisense orientations are superior than sense or hairpin in gene silencing. Unlike BSMV, BMV and BPMV, which are single-stranded positive-sense RNA viruses, RTBV is a double-stranded DNA pararetrovirus, which replicates by reverse transcribing its terminally redundant transcript (Hay et al. 1991). The interaction of the replication intermediates of pararetroviral nucleic acids with the plant DCLs has not been reported, but the accumulation of abundant small RNAs from pararetroviral transcripts (Blevins et al. 2011; Rajeswaran et al. 2014) from regions of the transcripts showing strong secondary structures, point towards the involvement of DCLs recognizing such structures. Using the same logic, intermolecular double-stranded molecules formed between the chlH mRNA and antisense chlH RNA formed from the pRTBV-MVIGS construct might also be recognized and processed into small RNAs by DCLs, which may be responsible for the observed silencing of chlH. The sense chlH-mediated silencing is inefficient, possibly because of inefficient recognition by DCLs or similar proteins. The RNAi machinery may also recognize the hairpin construct efficiently, thereby giving almost the same level of silencing as the antisense construct.

Despite the 40–80% reduction in the chlH transcript levels, the reduction of chlorophyll levels was modest (less than 25%) and so was the reduction in the quantum yield (F v/F m ratio). Fu et al. (2006) reported a much higher (90%) reduction in chlorophyll content following TRV-based silencing of pds gene in tomato. Similarly, Liu and Page (2008) reported an average of 90% fall in the nicotine levels in tobacco leaves following TRV-based silencing of putrescine N-methyl transferase, required in the nicotine biosynthetic pathway. The moderate fall in the chlorophyll content observed here could be due to mild silencing in pRTBV-MVIGS, compared to TRV. It could also be due to differences in the metabolic fluxes in the pathways being compared. Interestingly, the chlH-silenced phenotypes observed here (leaf yellowing following inoculation of the antisense and hairpin chlH constructs, Fig. 2c, d) were similar to the reported rice chl1 and chl9 mutant phenotypes (Zhang et al. 2006), in which the chlD and chlI subunits carried missense mutations.

VIGS mediated functional analysis of genes responsible for bacterial pathogens in monocots are limited; however few recent reports exist in dicots. For example; silencing of NbMADS1 in N. benthamiana using potato virus X (PVX) VIGS vector showed compromised regulation of defense responses to HarpinXoo (bacterial elicitor of the pathogenicity (hrp) gene of Xoo; Zhang et al. 2016). The SUMOE2 conjugating enzyme (SCEI) in Solanum peruvianum appeared to be crucial in the defense against Clavibacter michiganensis subsp. michiganensis (Cmm) colonization using Tomato Mottle Virus (ToMoV) based VIGS (Esparza-Araiza et al. 2015). Silencing of genes encoding ribosomal proteins, RPL12 and RPL19 in N. benthamiana showed disease resistance to non-host bacteria P. syringae pv. tomato using TRV-VIGS (Nagaraj et al. 2015). There are also few reports on VIGS mediated silencing of genes against fungal pathogens in monocots; for example, silencing of three genes in barley associated with resistance against powdery mildew, a fungal disease (Hein et al. 2005) and one against the fungal maize smut disease, caused by Ustilago maydis in maize (Van Der Linde et al. 2011). Hence, we decided to use the pRTBV-MVIGS system to silence the well-characterized Xa21 gene, conferring resistance against the bacterial pathogen Xoo, responsible for causing BLB in rice. This would allow us to assess the use of this gene silencing system in characterizing other disease resistance genes. Silencing of Xa21 resulted in similar levels of reduction in Xa21 transcript accumulation in the variety PB6-Xa21 as observed for pds and chlH, illustrating the similarity of the silencing process in all three cases. In most Xa21-silenced plants, there was an increase in the symptoms of BLB in the leaves upon challenge with Xoo (Figs. 5, 6), clearly indicating the robustness of the pRTBV-MVIGS gene silencing system for the study of disease resistance genes. It compares well with the levels of susceptibility reported after silencing of disease resistance genes using established silencing systems, such as TRV in Arabidopsis (Burch-Smith et al. 2006) and with newly-described VIGS systems, such as ALSV in tobacco (Igarashi et al. 2009), BSMV in barley (Hein et al. 2005) and BMV in maize (Van Der Linde et al. 2011). In this report, we have shown that silencing three rice genes, pds, chlH and Xa21 resulted in the same extent of reduction in the target transcript accumulation, namely 40–80% and hence reinforces the VIGS system as a robust gene silencing system for rice. Although, on one hand, silencing of chlH results in decrease in chlorophyll content and F v/F m ratio, the decrease is not the same. On the other hand, silencing of Xa21 transcript accumulation to the same extent, results in increased susceptibility to BLB to a different level.

In this report, the conditions for efficient VIGS-mediated gene silencing in rice, using the pRTBV-MVIGS system, have been standardized. To achieve an efficient gene silencing, the condition III, described in Table 2 should be followed. The robust pRTBV-MVIGS-based gene silencing system for rice can now be used widely for silencing any gene of choice and can be expected to contribute immensely for rice functional genomics.

Author contribution statement

ID conceived the work, RK constructed the clones, conducted the experiments and compiled the results, RK and ID analyzed the results and RK and ID wrote the manuscript.