Introduction

Geminiviruses are plant viruses with single-stranded (ss) circular DNA genome (~ 2.7 kb) encapsidated by twinned isometric geminate particles (18 × 30 nm) and cause huge losses in food and fiber crops across the globe (Fauquet et al. 2003; Varma and Malathi 2003; Krupovic et al. 2009; Hanley-Bowdoin et al. 2013; García-Arenal and Zerbini 2019). They replicate in nuclei of plant cells through double-stranded (ds) DNA intermediates that can act as minichromosomes (Pilartz and Jeske 1992; Ceniceros-Ojeda et al. 2016). Based on genome organization, host range, and vector specificity, these viruses are classified into 9 genera, viz., Becurtovirus, Begomovirus, Capulavirus, Curtovirus, Eragrovirus, Mastrevirus, Grablovirus, Topocuvirus, and Turncurtovirus (Adams et al. 2016; Zerbini et al. 2017).

Leaf curl disease of tomato (ToLCD) is one of the major constraints of tomato production in the Indian sub-continent (Chakraborty et al. 2003; Chakraborty 2008; Ranjan et al. 2014). Tomato leaf curl New Delhi virus (ToLCNDV) and Tomato leaf curl Gujarat virus (ToLCGuV) are two predominant tomato-infecting begomovirus species causing severe ToLCD in India (Padidam et al. 1995; Muniyappa et al. 2000; Chakraborty et al. 2003, 2008; Ranjan et al. 2013). ToLCNDV is a typical bipartite begomovirus containing DNA-A and DNA-B packaged within two separate virions (Padidam et al. 1995; Chakraborty and Kumar 2020). In contrast, ToLCGuV is a monopartite begomovirus which, in association with DNA-B, increases symptom severity (Chakraborty et al. 2003). DNA-A contains six open reading frames (ORFs) each coding different proteins, viz., AC1 (replication associated protein), AC2 (transcriptional activator protein), AC3 (replication enhancer), AC4 (pathogenicity determinant), AV1 (coat protein), and AV2 (pre-coat protein) (Padidam et al. 1995). DNA-B contains two ORFs encoding BV1 (nuclear shuttle protein) and BC1 (movement protein), required for cell-to-cell movement of the virus and long-distance movement through the phloem, respectively (Sanderfoot and Lazarowitz 1996).

Mixed infection of different begomoviruses occurs frequently among field-grown plants. Synergism between two viruses has been reported to occur in plants during simultaneous infection by two distinct viruses, as a result of which infectivity of either one or both the viruses is enhanced (Zhang et al. 2001; Rentería-Canett et al. 2011; Singh et al. 2016; Xia et al. 2016; Zhou et al. 2017). Usually, mixed infection by different species of tomato-infecting begomoviruses leads to more prevalent symptoms as compared to singly infected plants and sometimes may cause the outbreak of epidemics (Torre et al. 2018). One of the best characterized plant viral synergisms was reported between Potato virus X and Potato virus Y (PVY) in tobacco (Rochow and Ross 1955; Goodman and Ross 1974; Vance 1991; Hameed et al. 2014).

Supervirulent pseudorecombination and asymmetric synergism have been demonstrated between two severe tomato-infecting begomoviruses, ToLCNDV and ToLCGuV. Mixed infection of these two viruses leads to development of extremely severe symptoms as well as higher level of DNA-A and DNA-B of ToLCNDV and DNA-B of ToLCGuV (Chakraborty et al. 2008). However, DNA-A level of ToLCGuV remained unaltered, indicative of asymmetric synergism between these two begomoviruses (Chakraborty et al. 2008). Synergistic interaction between ToLCNDV and Tomato leaf curl Palampur virus (ToLCPaV) has been demonstrated previously (Kanakala et al. 2013). Contrary to the above, synergism between pepper-infecting begomoviruses, viz., Pepper huasteco yellow vein virus (PHYVV) and Pepper golden mosaic virus (PepGMV), has led to increased viral DNA accumulation of both the viruses (Rentería-Canett et al. 2011). Synergistic interactions have also been reported between Squash leaf curl virus (SLCV) and Watermelon chlorotic stunt virus (WmCSV), two cucurbit infecting viruses (Sufrin-Ringwald and Lapidot 2011). Despite common occurrence of synergistic interaction among plant viruses, specific role of viral encoded proteins in mediating synergism has not been extensively studied.

A few studies have identified post transcriptional gene silencing (PTGS) suppressors in viral synergism (Mascia and Gallitelli 2016; Singh et al. 2016). For example, HC-Pro encoded by Tobacco etch virus (TEV) (Anandalakshmi et al. 1998; Vance et al. 1995) and 2b protein of Tomato aspermy virus (TAV) (Lewsey et al. 2010) play a crucial role in PTGS suppression and viral synergism (Pallas and García 2011). Proteins encoded by AC2 of East African cassava mosaic Camaroon virus (EACMCV) and AC4 of African cassava mosaic virus (ACMV-[CM]) together contributed to elevated viral DNA titer (Vanitharani et al. 2004). Synergistic effect of PTGS suppressor between C2 of Beet curly top virus (BCTV), a Curtovirus, has been found to be responsible for promoting synergistic interaction with a begomovirus, Tomato yellow leaf curl Sardinia virus (TYLCSV) in Nicotiana benthamiana (Caracuel et al. 2012). However, the role of PTGS suppressors in mediating asymmetric synergism between ToLCNDV and ToLCGuV is not known yet.

In the current study, we identified proteins encoded by ToLCGuV AC2 and AV2 in promoting asymmetric synergism with ToLCNDV in N. benthamiana and tomato. Two different approaches, in planta experiments and leaf disc assays, were undertaken to identify the synergistic effect of viral encoded proteins of ToLCGuV on ToLCNDV pathogenesis. Our results suggest that ToLCGuV AC2 and AV2 encoded proteins are the key viral factors responsible for facilitating asymmetric synergism with ToLCNDV leading to severe leaf curl disease. The present study contributes to both ecological and epidemiological perspectives of plant-virus interactions, because of the frequently occurring mixed infections of ToLCNDV and ToLCGuV in tomato fields.

Materials and methods

Viral clones

The partial tandem repeat constructs of ToLCGuV (AY190290 and AY190291) and ToLCNDV (U15015 and U15017) used in this study have been reported previously (Padidam et al. 1995; Chakraborty et al. 2003). To facilitate site-directed mutagenesis, ToLCNDV DNA-A was re-cloned into the pUC18 (TaKaRa, New Delhi, India) vector at the SacI site, while the available monomeric clone of ToLCGuV was used as a template. The generation of infectious constructs of 5 different mutants of both ToLCNDV and ToLCGuV DNA-A used in this study has been previously described (Basu et al. 2018). Nonsense codons were introduced to replace serine (S5) in ORF AC2 of both ToLCNDV and ToLCGuV, phenyl alanine (F8) in ORF AC4 of ToLCNDV and cysteine (C9) of ToLCGuV, and tryptophan (W2) in ORF AV2 of both ToLCNDV and ToLCGuV (referred to as ∆AC2, ∆AC4, and ∆AV2, respectively). Double mutations were introduced both in serine (S5) and phenyl alanine (F8) in ORF AC2 and AC4 of ToLCNDV and serine (S5) and cysteine (C9) of ToLCGuV (referred to as ∆AC2∆AC4). The double mutants of both serine (S5) and tryptophan (W2) were mutated to introduce nonsense codons in both ToLCNDV and ToLCGuV, respectively (referred to as ∆AC2∆AV2). All the full-length (~ 2.7 kb) infectious clones of wild type (both DNA-A and DNA-B) and five different mutants (DNA-A) of both ToLCNDV and ToLCGuV in pCAMBIA2301 (Yu et al. 2018) already available in the laboratory were used.

Plant inoculation

Test plants (N. benthamiana, Solanum lycopersicum cv. Punjab Chhuhara) were grown in an insect-proof green house with 16:8-h (light/dark) photoperiod and relative humidity of 60%. The confirmed clones (wild type and mutants of both ToLCNDV and ToLCGuV) were mobilized into Agrobacterium tumefaciens strain EHA105 by freeze-thaw method and selected on LB agar containing kanamycin (50 μg/ml) and rifampicin (30 μg/ml). Infectious constructs were agro-inoculated onto test plants following standard protocol (Kumari et al. 2010; Singh et al. 2012). Inoculated plants were maintained until 30 days post-inoculation (dpi) for assessing symptom severity. All the mutations were confirmed by sequencing of PCR-amplified viral target genes from progeny viral DNA isolated from plants inoculated with each viral mutant. All the experiments were repeated three times and symptom severity of inoculated plants was recorded following Chakraborty et al. (2008). Severity of symptoms was scored following 0–4 severity scale where 0 indicates no visible symptom (0) and 4 (++++) indicated highly severe symptoms.

Southern hybridization

Total DNA was extracted from the upper most systemically infected leaves (~ 1 g) from virus- and mock-inoculated plants following Dellaporta et al. (1983). Total DNA (7 μg) was separated on 0.8% agarose gel and transferred to Hybond-N+ membranes (Amersham Corp., Amersham, UK) (Sambrook et al. 2001). Viral DNA was detected by hybridizing blots using a specific radiolabeled probe of either ToLCNDV or ToLCGuV DNA-A. AC1 ORFs of both these viruses were PCR-amplified separately and labeled with [α-32p]-dCTP using random oligonucleotide-primed synthesis method (Feinberg and Vogelstein 1983). Viral bands were detected and quantified using a phosphor image analysis system (Typhoon, Amersham, UK) following standard procedures (Vanitharani et al. 2004; Chellappan et al. 2005; Chakraborty et al. 2008). Band intensities were quantified by image analysis software (Quantity one, Bio-Rad, CA, USA) following Singh et al. (2016). Positions of four replicative forms of begomoviral DNA (OC, open circular; Lin, linear; SC, super-coiled; SS, single-stranded) are indicated on the right side of the blots. Signal intensities of all four forms of bands were quantified using the Quantity One analysis software (Bio-Rad, CA, USA) to calculate the total intensity per lane. Data are presented as percentage values relative to those of singly infected wild-type begomoviral DNA-A along with its cognate DNA-B molecule, used as control and represented as 100%. Probes were specifically tested to preclude that there is no cross-hybridization between DNA-A of these two begomoviruses. Southern blotting experiments were repeated three times, and each time, three replicates were pulled together for each treatment (Singh et al. 2016).

Isolation and detection of siRNAs

Total RNA was isolated from the uppermost systemically infected leaves (~ 1 g). Enrichment of low molecular weight RNAs was performed using 5% polyethylene glycol and 0.5 M NaCl (Hamilton and Baulcombe 1999), separated in 15% Tris-borate-EDTA-urea acrylamide gel and transferred to Hybond-N+ membrane (Amersham Biosciences Corp., Amarsham, UK) using semi-dry electro blotter (GE Amersham, UK). For detection of viral specific siRNAs, [α-32p]-dCTP-labeled DNA probes of overlapping regions between AC1/AC2 and AC2/AC3 of both ToLCNDV (nt 1148-1701) and ToLCGuV (nt 1209-1550) were used. For detection of GFP-specific siRNAs, full-length GFP ORF was PCR-amplified and used as a probe. Hybridization was carried out at 40 °C overnight, and siRNAs were detected and quantified using phosphor imager analysis system following standard procedures (Vanitharani et al. 2004; Chellappan et al. 2005).

Transient replication of viral DNA in tobacco leaf discs

To assess the direct effect of ToLCGuV-encoded proteins in asymmetric synergism, full-length ORFs (AC1, AC2, AC4, and AV2) were cloned into a plant expression vector (pBinAR) (Vinoth Kumar et al. 2012). The ability of cloned viral ORFs under constitutive expression to influence replication of ToLCNDV was tested in N. benthamiana leaf discs following Agrobacterium-mediated inoculation (Horsch and Klee 1986). N. benthamiana leaf discs incubated in pre-conditioning medium (MS medium supplemented with 1 mg/L BAP and 0.1 mg/L IAA for 48 h at 25̊C) were dipped into A. tumefaciens suspensions carrying appropriate viral clones. Leaf discs were dried and placed on co-cultivation medium (MS medium supplemented with 1 mg/L BAP, 0.1 mg/L IAA, and 100 μM acetosyringone) for 48 h and then transferred onto selection medium containing 450 mg/L cefotaxime and 50 mg/L kanamycin. Total genomic DNA was isolated from leaf discs after 8 dpi, and replication of viral DNA was detected by Southern blotting using the probes as described above.

Agro infiltration and GFP imaging

In order to identify viral encoded proteins involved in PTGS suppression, full-length individual ORFs of both mutants and wild type (AC2, AC4, and AV2) of ToLCNDV and ToLCGuV DNA-A were amplified using an appropriate pair of primers (Supplementary Table S1) and cloned into the pTZ57R/T vector (Thermo Fisher Scientific, Waltham, MA). ORFs were finally cloned into a plant expression vector (pBinAR) following digestion with suitable restriction enzymes. The plasmid pBinP1 expressing Hc-Pro protein of Tobacco etch virus (TEV) was used as positive control for the suppressor assays (Cañizares et al. 2006). Transgenic GFP expressing N. benthamiana (16c line) plants were infiltrated with A. tumefaciens containing desired constructs (Hamilton et al. 2002).

To investigate suppressor activity of the viral encoded proteins, Agrobacterium culture containing 35S-GFP cloned in pBinAR was co-infiltrated with either wild type or mutant(s) separately (Voinnet et al. 2000; Baulcombe 2004). siRNAs were isolated from agroinfiltrated leaves at 6 dpi and were probed with GFP-specific probe. Plants were examined daily for GFP fluorescence under long wavelength (365 nm) UV light (UVP, USA) and photographed using a Canon Power shot SX120 IS digital camera (Canon, Tokyo, Japan).

Results

Effect of ToLCGuV mutants on ToLCNDV pathogenesis

To elucidate the role of viral encoded proteins involved in asymmetric synergism between ToLCNDV and ToLCGuV, we used a reverse genetic approach in which different mutants of ToLCGuV DNA-A were inoculated to study their effect on accumulation of ToLCNDV DNA as well as ToLCNDV-specific siRNAs in both N. benthamiana and tomato. Wild type and five different mutants of ToLCGuV (three single mutants, viz., ∆AC2, ∆AC4, and ∆AV2, and two double mutants, viz., ∆AC2∆AC4 and ∆AC2∆AV2) were co-inoculated with ToLCNDV (DNA-A + DNA-B) on both N. benthamiana and tomato and symptom expression patterns of the inoculated plants were monitored over time (Fig. 1a). We studied disease development on N. benthamiana as well as on tomato along with analysis of accumulation of viral DNA and viral specific siRNAs (Figs. 1 and 2; Supplementary Fig. S1; Table 1). N. benthamiana plants inoculated with all these five mutants as well as wild-type ToLCGuV DNA-A were found to be symptomatic when co-inoculated with ToLCNDV (Fig. 1a) and exhibited typical symptoms like downward leaf curling, severe stunting, yellowing, chlorosis, crumpling, and small leaf. While the wild-type virus showed its first symptom at 5 dpi (days past inoculation), all remaining mutants showed the first symptom on 6 dpi except for the AC2 mutant on 8 dpi (Fig. 1 and Table 1).

Fig. 1
figure 1

Effect of ToLCGuV-encoded proteins on ToLCNDV pathogenesis on N. benthamiana. a Symptom expression on test plants infected with combinations as indicated against each representative plant at 21 dpi. b Accumulation of ToLCNDV DNA in test plants co-inoculated with wild type and different mutants of ToLCGuV along with ToLCNDV at 21 dpi. Total genomic DNA stained with ethidium bromide (EtBr) at the bottom of gel serves as loading control. Four replicative forms of viral DNA are indicated: OC, open circular; Lin, linear SC, super-coiled; SS, single-stranded. c Accumulation of ToLCNDV-specific siRNAs at 21 dpi. NA—ToLCNDV DNA-A, NB—ToLCNDV DNA-B, and VA—ToLCGuV DNA-A. Ethidium bromide–stained enriched small RNAs are shown as the loading controls

Full size image
Fig. 2.
figure 2

Effect of ToLCGuV-encoded proteins on ToLCNDV pathogenesis on S. lycopersicum. a Symptom expression on test plants infected with combinations as indicated against each representative plant at 21 dpi. b Accumulation of ToLCNDV DNA in plants co-inoculated with wild type and different mutants of ToLCGuV along with ToLCNDV at 21 dpi. Total genomic DNA stained with ethidium bromide (EtBr) at the bottom of the gel serves as loading control. Four replicative forms of viral DNA are indicated: OC, open circular; Lin, linear SC, super-coiled; SS, single-stranded. c Accumulation of ToLCNDV-specific siRNAs at 21 dpi. NA—ToLCNDV DNA-A, NB—ToLCNDV DNA-B, and VA—ToLCGuV DNA-A. Ethidium bromide–stained enriched small RNAs are shown as the loading controls

Full size image
Table 1 Infectivity of ToLCGV (mutants and wild type) co-inoculated with ToLCNDV on N. benthamiana and S. lycopersicum
Full size table

Wild-type ToLCGuV DNA-A-inoculated plants displayed maximum severity (++++) while co-inoculated with ToLCNDV, but we failed to distinguish a difference in symptom severity among mutants; however, the number of days to attain severe symptoms of infection varied among wild type and 5 distinct mutants (Fig. 1 and Table 1). N. benthamiana and tomato plants attained severe symptoms of viral infection such as severe leaf curling, stunting, yellowing, leaf crumpling, and chlorosis much faster (13 dpi and 14 dpi, respectively) when wild-type ToLCGuV DNA-A was co-inoculated with ToLCNDV DNA-A and its cognate DNA-B molecule. Mutation in AC2 of ToLCGuV delayed initial symptom appearance (8 dpi) on N. benthamiana, as well as appearance of severe symptoms on both N. benthamiana and tomato (20 dpi and 21 dpi, respectively), when co-inoculated with ToLCNDV. All the other mutants of ToLCGuV exhibited severe symptoms of infection when co-inoculated with ToLCNDV, but the number of days to achieve disease severity were intermediate between wildtype and ∆AC2 on both N. benthamiana and tomato (Table 1). A total of 15 plants were inoculated for each treatment and all the plants were found symptomatic following co-infection (Table 1).

ToLCNDV DNA level was measured from the two uppermost symptomatic leaves of N. benthamiana plants at 10 and 21 dpi, respectively (Supplementary Fig. S1C and Fig. 1b). In all the cases, wild-type ToLCNDV DNA accumulation alone was considered 100% in presence of its cognate DNA-B molecule. At 10 dpi, higher accumulation of ToLCNDV DNA-A was detected in N. benthamiana plants infected with ToLCGuV wild type (128%), ∆AC2 (123%), and ∆AV2 (118%) co-inoculated with ToLCNDV DNA-A and DNA-B, while the double mutants (∆AC2∆AC4 and ∆AC2∆AV2) failed to show enhanced accumulation (Supplementary Fig. S1C). At 21 dpi, ToLCGuV wild type and ∆AV2 mutant-infected plants contained higher titer of ToLCNDV, 148% and 133%, respectively) compared to plants infected with ToLCNDV alone (Fig. 1b). Other mutants of ToLCGuV did not yield any significant synergistic interaction with ToLCNDV (Fig. 1b).

To correlate the plausible role of the PTGS pathway with symptom severity and viral DNA accumulation, viral siRNAs were isolated from systemically infected two uppermost leaves of N. benthamiana at 21 dpi. Higher accumulation of ToLCNDV-specific siRNAs was observed in plants infected with wild-type ToLCNDV DNA-A along with its cognate DNA-B component, which was considered 100% (Fig. 1c). However, accumulation of ToLCNDV-specific siRNAs was found to decrease in presence of either wild type or mutants of ToLCGuV DNA-A (Fig. 1d). Accumulation of ToLCNDV siRNAs decreased to 72% in case of ∆AC2, 45% with ∆AC4, and 66% in case of ∆AV2 co-infected plants (Fig. 1c). Plants infected with ToLCNDV DNA along with double mutants of ToLCGuV showed decreased siRNA accumulation (33% with ∆AC2∆AC4 and 62% with ∆AC2∆AV2) (Fig. 1c).

To validate the synergistic effect of ToLCGuV-encoded proteins on ToLCNDV in the natural host tomato, we followed the same combinations used for N. benthamiana inoculation. Tomato plants infected with wild type and the mutants of ToLCGuV and co-inoculated with ToLCNDV were found to exhibit distinguishable symptoms including severe downward leaf curling, yellowing of leaf lamina, and stunting (Fig. 2a). As expected, more severe symptoms were observed in presence of wild-type ToLCGuV DNA-A co-infected with ToLCNDV (6 dpi) and the initial symptom appeared 2 days earlier than tomato plants inoculated with ToLCNDV alone (8 dpi) (Fig. 2a and Table 1). Interestingly, the first symptom appeared on the 8th dpi on tomato plants infected with all the mutants of ToLCGuV except for ∆AC2∆AC4, which showed first symptom on 11th dpi when co-inoculated with ToLCNDV (Table 1). The pattern of symptom severity in tomato exhibited a trend similar to N. benthamiana. Like N. benthamiana, natural host tomato exhibited severe infections much earlier when the wild-type ToLCGuV DNA-A was co-inoculated with ToLCNDV as compared to mutants (approximately 15 dpi), while delays in expressing severe symptoms were recorded on plants co-inoculated with ToLCGuV∆AC2 and ToLCNDV (21 dpi). All the other ToLCGuV mutants co-inoculated with ToLCNDV showed severe symptoms of co-infection between 17 and 18 days on tomato (Table 1).

To further confirm the synergistic effect of ToLCGuV-encoded proteins on ToLCNDV titer in tomato, accumulation of ToLCNDV DNA-A was studied from the upper most symptomatic leaves of plants infected with either wild type or ToLCGuV mutants, as mentioned earlier for N. benthamiana (Fig. 2b). Higher accumulation of ToLCNDV DNA-A was detected in tomato co-infected with wild-type ToLCGuV DNA-A (120%) and ∆AV2 (115%) at 21 dpi while ∆AC2, ∆AC4, and double mutants failed to elevate ToLCNDV accumulation (Fig. 2b).

To investigate synergistic interactions of ToLCGuV-encoded proteins with PTGS pathway, ToLCNDV-specific siRNAs were isolated at 21 dpi from systemically infected upper most symptomatic leaves of tomato. Similar to N. benthamiana, the accumulation of ToLCNDV-specific siRNAs were found to decrease in tomato co-infected with wild type and mutants of ToLCGuV, while higher accumulation was observed from plants infected with wild-type ToLCNDV inoculated with its cognate DNA-B component. Decreased siRNA accumulation was observed in tomato plants co-infected with ToLCGuV wild type (43%), ∆AC2 (62%), ∆AC4 (60%), and ∆AV2 (82%) (Fig. 2c). Tomato plants inoculated with the double mutants (∆AC2∆AC4 and ∆AC2∆AV2) of ToLCGuV also showed low levels of ToLCNDV siRNAs following co-infection with either of these mutants (66% and 78%, respectively) (Fig. 2c).

Effect of ToLCNDV mutants on ToLCGuV pathogenesis

Conversely, to test the role of ToLCNDV-encoded proteins during synergistic interaction with ToLCGuV, we inoculated N. benthamiana plants with all the five mutants of ToLCNDV along with ToLCGuV (DNA-A + DNA-B) (Fig. 3). We studied symptom severity on the co-infected N. benthamiana plants followed by accumulation of ToLCGuV-specific DNA and siRNAs.

Fig. 3
figure 3

Effect of ToLCNDV-encoded proteins on ToLCGuV pathogenesis on N. benthamiana. a Symptom expression on test plants infected with combinations as indicated against each representative plant at 21 dpi. b Accumulation of ToLCGuV DNA in plants co-inoculated with wild type and different mutants of ToLCNDV along with ToLCGuV at 21 dpi. c Accumulation of ToLCGuV-specific siRNAs at 21 dpi. VA—ToLCGuV DNA-A, VB—ToLCGuV DNA-B, and NA—ToLCNDV DNA-A

Full size image

Symptoms appeared on N. benthamiana plants infected with any of the 5 mutants of ToLCNDV in presence of wild-type ToLCGuV (Fig. 3a). All the ToLCNDV mutants except ∆AC2 exhibited distinguishable symptoms which include severe leaf curling, stunting, yellowing, chlorosis, leaf puckering, rolling of leaf lamina, and vein thickening (Fig. 3a, Table 2). Symptom severity was found to be maximum when wild-type ToLCGuV DNA-A was co-inoculated with ToLCNDV on N. benthamiana and the initial symptom appeared on 5th dpi. The infected plants exhibited severe symptoms including leaf curling, yellowing and stunting, and leaf chlorosis. At later stages of infection, the plants developed extreme chlorosis of leaves and died (data not shown). Symptoms on N. benthamiana plants infected with ToLCNDV∆AC2 mutant along with ToLCGuV started to appear from 9 dpi onwards which later on resulted in delayed and attenuated symptoms and the plants exhibited mild-to-moderate downward leaf curling, stunting, and yellowing of leaf lamina (Fig. 3a, Table 2, and Supplementary Fig. S1B). ToLCNDV∆AC4 infected plants showed initial symptoms of leaf curling at 7 dpi, followed by severe symptoms like that of wild type when co-inoculated with ToLCGuV (Fig. 3a, Table 2, and Supplementary Fig. S1B). ∆AC2∆AC4, ∆AV2, and ∆AC2∆AV2 mutants of ToLCNDV failed to indicate notable difference in terms of symptom severity pattern when co-inoculated with ToLCGuV and the initial symptom appeared on test plants around 5–6 dpi (Fig. 3a, Table 2, and Supplementary Fig. S1B). Days to achieve maximum severity greatly varied among wild type and different mutants of ToLCNDV DNA-A when co-inoculated with ToLCNGuV. N. benthamiana plants attained maximum severity much faster when wild-type ToLCNDV DNA-A was co-inoculated with ToLCGuV as compared to mutants (13 dpi), while maximum delay was recorded in the case of plants co-inoculated ToLCNDV∆AC2 and ToLCGuV (19 dpi). All other ToLCNDV mutants co-inoculated with ToLCGuV reached maximum severity between 16 and 19 dpi for N. benthamiana (Table 2). A total of 15 plants were inoculated for each combination, and all the plants developed symptoms following co-infection (Table 2).

Table 2 Infectivity of ToLCNDV (mutants and wild type) co-inoculated with ToLCGV on N. benthamiana
Full size table

Southern blot analysis revealed that ToLCNDV-encoded proteins failed to enhance ToLCGuV DNA-A accumulation. At 10 dpi, decreased accumulation of ToLCGuV DNA-A was observed in the presence of either wild type or any of the ToLCNDV mutants (Supplementary Fig. S1B, S1D). At 21 dpi, ToLCGuV DNA-A accumulation was reduced in the presence of wild type (60%), ∆AC2 (57%), and ∆AC4 (61%) of ToLCNDV while equivalent levels were detected in the presence of ∆AC2∆AC4 (99%) and ∆AV2 (95%). ToLCGuV DNA-A titer was increased 155% when ∆AC2∆AV2 was used for inoculation (Fig. 3b).

Accumulation of ToLCGuV-specific siRNAs was compromised severely in the presence of wild type, ∆AC2, and ∆AC4 of ToLCNDV while a low level was detected in the presence of ∆AV2 and ∆AC2∆AC4 (23% and 90%, respectively). ToLCGuV-specific siRNA accumulation was increased in N. benthamiana in the presence of double mutants of ToLCNDV∆AC2∆AV2 (126%), compared to ToLCGuV-alone-infected plants (Fig. 3c).

Synergistic effect of viral encoded proteins on ToLCNDV and ToLCGuV replication in N. benthamiana leaf disc

A leaf disc assay was carried out to further validate the ability of ToLCGuV-encoded proteins to enhance ToLCNDV replication (Fig. 4 and Supplementary Fig. S2). AC1, AC2, AC4, and AV2 ORFs of ToLCGuV were cloned into a plant expression vector (pBinAR) under 35S promoter and co-infected with ToLCNDV DNA-A and its cognate DNA-B molecule in N. benthamiana leaf disc. Accumulation of ToLCNDV DNA was studied from these infected leaf discs at 6th and 8th dpi (Fig. 4; Supplementary Fig. S2). ToLCNDV DNA-A level was found to be higher in the presence of ToLCGuV AV2 and AC2 (~ 1.8 and 1.5 times, respectively) whereas maximum accumulation was observed when co-infected with ToLCGuV DNA-A (3.4 times in comparison to wild-type ToLCNDV DNA-A and DNA-B) (Fig. 4) at 8 dpi. Similarly, enhanced ToLCNDV DNA-A accumulation was observed (2 times and 1.8 times, respectively) in presence of ToLCGuV AV2 and AC2, at 6 dpi (Supplementary Fig. S2). On the contrary, AC1- and AC4-encoded proteins were unable to exert any synergistic effect on ToLCNDV replication (Fig. 4 and Supplementary Fig. S2). A similar strategy was taken to check the effects of ToLCNDV-encoded proteins on ToLCGuV replication in N. benthamiana leaf disc. None of the 4 ORFs of ToLCNDV DNA-A (AC1, AC2, AC4, and AV2) was found to enhance ToLCGuV DNA-replication in the transient leaf disc assay (Supplementary Fig. S3).

Fig. 4
figure 4

Effect of ToLCGuV-encoded proteins on ToLCNDV replication in N. benthamiana leaf discs at 8 dpi. ORFs were cloned under CaMV 35S promoter

Full size image

AC2 and AV2 of ToLCNDV and AV2 of ToLCGuV are major suppressors of PTGS

Reversal of GFP silencing assays were carried out to determine the ability of viral encoded proteins to suppress gene silencing. Following infiltration, the fluorescence continued to increase and was detected at a very high level at 6 dpi in the case of 16c GFP transgenic N. benthamiana lines infiltrated with AC2 and AV2 of ToLCNDV, and ToLCGuV-AV2 (Fig. 5a, c). Mild suppression of GFP silencing was observed in the case of AC4 of ToLCNDV, and AC2 and AC4 of ToLCGuV (Fig. 5a, c). N. benthamiana 16c plants infiltrated with only 35S-GFP were considered negative control whereas P1/HC-Pro was used as a positive control (Fig. 5a, c). As expected, the mutation in viral ORFs failed to reverse silencing of GFP which indicated that the mutations abolished suppressor activity of these mutant viral encoded proteins (Fig. 5a, c).

Fig. 5
figure 5

RNA silencing suppression activity of tomato-infecting begomoviruses. a and c Leaves of N. benthamiana line 16c plants were co-infiltrated with 35S-GFP and 35S-AC2/AC4 /AV2 of either ToLCNDV (a) or ToLCGuV (c). Mutant ORFs (mAC2, mAC4, and mAV2) of ToLCNDV and ToLCGuV co-infiltrated GFP were also used to check the effect of mutation on these ORFs. 35S-GFP alone and Hc-Pro/pBinP1from TEV co-infiltrated GFP were used as negative and positive controls, respectively. Leaves were photographed at 6 dpi under UV light. ToLCNDV is represented as NA and ToLCGuV as VA. b and d Relative accumulation of GFP-specific siRNAs from leaves of 16c plants at 6 dpi. Total RNA extracted from leaves co-infiltrated with 35S-GFP plus p35S-ToLCNDV (b) and ToLCGuV (d) ORFs. GFP-specific siRNAs were detected from the infiltrated leaves using [α-32P] dCTP-labeled GFP-specific probe. Ethidium bromide–stained enriched small RNAs at the bottom serves as the loading control. mGFP indicates modified GFP

Full size image

GFP-specific siRNAs readily detected from the 16c plants infiltrated with GFP alone were considered 100% (Fig. 5b, d). siRNA accumulation was found to be negligible (9%) in ToLCNDV-AC2-infiltrated plants whereas ToLCNDV-AV2 infiltration resulted in 21% accumulation (Fig. 5b). GFP-specific siRNAs were abundant in ToLCNDV-AC4 (81%) which was not found to be a potent suppressor (Fig. 5b). In the case of ToLCGuV, AV2-infiltrated samples accumulated 35% of GFP-specific siRNAs whereas AC2 and AC4 accumulated siRNAs 81% and 79%, respectively (Fig. 5d). As expected, accumulation of GFP-specific siRNAs was abundant in leaf patches co-infiltrated with viral mutants (Fig. 5b, d).

Discussion

Mixed infection is a very common phenomenon under field conditions and may lead to either synergistic or antagonistic interactions among different viruses by altering both viral and host fitness (Rentería-Canett et al. 2011; Syller 2012; Singh et al. 2016; Syller and Grupa 2016; Xia et al. 2016; Zhou et al. 2017; Moreno and López-Moya 2019). Sometimes, the order of infection of viruses during mixed infection determines the occurrence of synergism or antagonism. For example, mixed infection of Papaya ring spot virus (PRSV) and Papaya mosaic virus (PapMV) cause synergism in papaya, but occurrence of PapMV infection before PRSV results in antagonistic interactions between these two viruses (Chávez-Calvillo et al. 2016). In case of synergism, the appearance of disease symptoms has been found to be more severe as compared to plants infected with single virus. Apart from yield reduction, mixed infected plants also provide possibility of forming new variants or species through recombination and virus evolution leading to severe pathogenesis (Roossinck 1997).

Synergistic association is very common among diverse plant viruses, and involvement of different PTGS suppressors has been implicated during mixed infection (Vanitharani et al. 2004; Caracuel et al. 2012; Syller 2012). Synergistic interactions among geminiviruses have also been reported during mixed infection (Chakraborty et al. 2008; Sufrin-Ringwald and Lapidot 2011; Cuellar et al. 2015; Singh et al. 2016). Tomato-infecting begomoviruses are serious pathogens of solanaceous crops like chili, tomato, and tobacco. ToLCNDV and ToLCGuV are two predominant begomovirus species and their mixed infection resulted in extremely severe symptom where DNA-A and DNA-B of ToLCNDV and DNA-B of ToLCGuV were found to accumulate at much higher level while DNA-A component of ToLCGuV remained unaltered (Chakraborty et al. 2008). Therefore, asymmetric synergism did persist between these two distinct tomato-begomovirus species occurring in India. Although the role of viral encoded proteins in disease development was extensively investigated recently by Basu et al. (2018), their involvement in PTGS suppression and asymmetric synergistic interaction between these two tomato-infecting begomoviruses has not been tested so far.

The present study was conceived to identify the role of PTGS suppressors (AC2, AC4, or AV2) encoded by both ToLCNDV and ToLCGuV in facilitating asymmetric synergism between these begomovirus species. Two different approaches were taken to examine the effect of PTGS suppressor(s) in asymmetric synergism. Firstly, using reverse genetic approach the effect of single and double mutants of both ToLCNDV and ToLCGuV on asymmetric synergism on two different solanaceous hosts, N. benthamiana and tomato were studied in terms of symptom severity, viral titer, and viral siRNA accumulation. All the mutants were found to be infectious when inoculated with wild type of either of the viruses in N. benthamiana and tomato (Figs. 1, 2, and 3). The symptom expression pattern did not provide much information to identify the effect of viral mutants in asymmetric synergism except that the plants infected with both viruses developed more severe symptoms than singly infected plants. Therefore, to identify the plausible role of PTGS suppressor(s) in asymmetric synergism, we further checked the viral DNA accumulation in these co-infected plants. During the early stage of infection, the accumulation of ToLCNDV DNA-A was found to be higher in N. benthamiana co-infected with either ∆AC2 or ∆AV2 of ToLCGuV while ToLCNDV∆AC2 or ∆AV2 failed to elevate ToLCGuV replication (Supplementary Fig. S1C and S1D). But at 21 dpi, only ToLCNDV DNA-A showed elevated accumulation in presence of ToLCGuV∆AV2 while ∆AC2 could not synergistically elevate ToLCNDV titer in both N. benthamiana and tomato. In the early stage of infection, both functional AV2 and AC2 of ToLCGuV can synergistically enhance ToLCNDV accumulation (Supplementary Fig. 1C). However, at a later stage of infection, an enhanced level of ToLCNDV was detected only for ToLCGuVΔAV2 because of the presence of functional AC2 of ToLCGuV. Similarly, ToLCGuVΔAC2 failed to synergistically elevate ToLCNDV replication because of the lack of the major viral suppressor. On the contrary, none of the viral mutants (∆AC2, ∆AC4, or ∆AV2) of ToLCNDV could elevate ToLCGuV accumulation. It is possible that AC2- and AV2-encoded proteins of ToLCNDV interfere with ToLCGuV DNA-A accumulation and thereby severely hamper generation of ToLCGuV-specific siRNAs. A recent study of Basu et al. (2018) established the role of ToLCNDV AC2 and AV2, and ToLCGuV AV2 in disease development, and we have reconfirmed their crucial role for suppression of PTGS by transiently expressing them in transgenic N. benthamiana 16c plants through reversion of GFP silencing assays (Fig. 5). In fact, AC2 and AV2 proteins of ToLCNDV exerted inhibitory effect on ToLCGuV replication which justifies asymmetric synergism between these two tomato-infecting begomoviruses, while none of the ToLCGuV-encoded proteins negatively influenced ToLCNDV DNA-A accumulation. This can presumably happen through suppression of host antiviral silencing and can be better explained by insignificant effects of ToLCGuV-AV2 in pathogenesis unlike AV2 of ToLCNDV (Basu et al. 2018). Interestingly, the double mutant ∆AC2∆AC4 of ToLCGuV showed similar reduction of ToLCNDV-specific siRNAs but failed to influence ToLCNDV accumulation. It is important to mention here that ToLCGuV∆AC2∆AC4 encodes a functional AV2 protein. As we discussed above, ToLCGuV-AV2 contributes insignificantly to pathogenesis unlike ToLCNDV-AV2 (Basu et al. 2018) but serves as a strong PTGS suppressor (Fig. 5a and c). Therefore, ToLCGuV-AV2 functions together with AC2 and AV2 of ToLCNDV to severely reduce siRNA accumulation.

The detection of a higher amount of ToLCNDV-specific siRNAs in both N. benthamiana and tomato inoculated with ToLCNDV alone could be due to the absence of cumulative effects of the viral suppressors encoded by ToLCGuV, which was evident in plants co-infected with either wild type or mutants of ToLCGuV. The synergistic effect of viral suppressors of ToLCGuV resulted in reduced accumulation of ToLCNDV siRNAs which was correlated with levels of ToLCNDV DNA-A accumulation. Cumulative effects of PTGS suppressors encoded by begomoviruses on viral siRNA accumulation have been investigated (Singh et al. 2016; Kumar et al. 2017; Basu et al. 2018). Lower accumulation levels of ToLCNDV-specific siRNAs in N. benthamiana co-infected with ToLCGuV∆AC4 and ∆AC2∆AC4 mutants could be attributed to the presence of functional AV2 of ToLCGuV, but levels of ToLCNDV DNA-A were not decreased much because of the presence of multiple functional suppressors encoded by both of these begomoviruses. Accumulation of ToLCNDV-specific siRNAs was further elevated in the presence of ToLCGuV mutants like ∆AC2, ∆AV2, and ∆AC2∆AV2 because the major suppressors are rendered non-functional. Notably, the level of ToLCNDV DNA-A was found to be higher in the presence of ToLCGuV∆AV2. This might be due to the presence of functional ToLCGuV AC2 and AC4 which synergistically elevates ToLCNDV DNA accumulation. Similarly, in the natural host tomato, enhanced ToLCNDV accumulation was observed in the presence of wild-type ToLCGuV while the accumulation of ToLCNDV-specific siRNAs was found to be low and was further enhanced when the major suppressors of ToLCGuV (such as AC2 and AV2) were made non-functional through mutation (Fig. 2b, c). ToLCNDV DNA-A level was not found to decrease much when co-inoculated with ToLCGuV mutants because the other suppressors contributed to ToLCNDV pathogenesis. Recently, Basu et al. (2018) described the inconsistency of correlation and leaf-specific responses between accumulation of viral DNA and viral specific siRNAs for both ToLCNDV and ToLCGuV, which indicated the involvement of RNA-dependent RNA polymerase 1 (RDR1), an important component of host antiviral silencing (Donaire et al. 2008; Wang et al. 2010). Besides RDR1, several other host factors related to transcriptional gene silencing (TGS) and antiviral PTGS pathways (RDRs, AGOs, and DCLs etc.) have also been investigated because of the involvement of these suppressors in subverting TGS and PTGS-mediated antiviral defense (Garcia-Ruiz et al. 2010; Wang et al. 2010; Ying et al. 2010; Basu et al. 2018; Zarreen and Chakraborty 2020).

Plant defense against invading viruses are mostly mediated through an innate antiviral system constituted by both TGS and PTGS, and the viruses have evolved suppressors to counteract these pathways (Pumplin and Voinnet 2013). In tomato-infecting begomoviruses, TrAP protein is encoded by AC2 and suppresses TGS (Sharma and Ikegami 2010; Amin et al. 2011; Luna et al. 2012; Castillo-González et al. 2015). Similarly, AC4/C4 protein promotes pathogenesis and suppression of antiviral silencing through multiple mechanisms (Fondong 2019). The V2/AV2-encoded pre-coat protein also serves as a pathogenicity determinant and silencing suppressor of both TGS and PTGS (Sharma and Ikegami 2010; Wang et al. 2014; Basu et al. 2018). AV2 of Tomato yellow leaf curl Sardinia virus is known to interact to SlSGS3 for substrate dsRNA recognition (Glick et al. 2008; Fukunaga and Doudna 2009) and does not allow RDR6 to generate duplex RNA (Glick et al. 2008). ToLCNDV AV2 mutant infection resulted in mild symptom induction and reduction of viral DNA level (Padidam et al. 1996). Hence, silencing suppressor activity of AV2 is attributed to block the spread of RNA silencing signals to the systemic leaves (Zrachya et al. 2007; Glick et al. 2008). Therefore, mutation in AV2 resulted in successful PTGS through generation of viral specific siRNAs via RDR6-SGS3-mediated antiviral pathway which finally led to enhanced accumulation of viral siRNAs. It was also found that neither wild type nor ToLCNDV mutants could enhance ToLCGuV replication in N. benthamiana at 10 dpi. Reduced accumulation of ToLCGuV DNA was observed in the presence of either wild type or ∆AC2 or ∆AC4 of ToLCNDV DNA-A at 21 dpi. Similar to ToLCGuV DNA accumulation, ToLCGuV-specific siRNAs were found to be compromised severely in the presence of wild type, ∆AC2, and ∆AC4, and reduced significantly in the absence of ∆AV2.

N. benthamiana leaf disc assay further confirmed the ability of AV2 and AC2 of ToLCGuV in positively influencing ToLCNDV accumulation. ToLCNDV accumulation was reported to be increased in association with ToLCGuV AC2 and AV2 cloned under 35S CaMV promoter in pBinAR (Fig. 4 and Supplementary Fig. S2) than ToLCNDV alone. Furthermore, ToLCNDV AC2 has been identified as a potent suppressor as compared to ToLCGuV AC2 as evidenced by both reversal of GFP silencing and less accumulation of GFP-specific siRNA. It is worth noting that mutation in AC2 diminishes pathogenicity of ToLCNDV in tobacco and tomato indicating host-specific role of viral encoded proteins (Basu et al. 2018). AV2 of ToLCNDV has been found to be a more potent suppressor than ToLCGuV AV2. Despite sharing a high degree of amino acid sequence identity (74%) and having a consensus motif involved in PTGS suppression, the divergent regions of AV2 protein might contribute for individual protein functions. Unlike AV2, ToLCNDV AC2 protein share less degree of identity (58%) with ToLCGuV AC2 which likely contributes to a differential role in pathogenesis in diverse hosts and altered efficacy of ToLCGuV AC2 as a PTGS suppressor.

Besides suppressing TGS and PTGS through virus-encoded suppressors, synergistic interactions among plant viruses lead to repression of various genes involved in host-defense pathways (Herranz et al. 2013; Singh et al. 2016). Synergistic interactions of multiple genomic components of geminiviruses reduce the expression of genes associated with basal resistance (ascorbate peroxidase, thionin, polyphenol oxidase etc.) and genes related to specific host-defense (NBS-LRR) which ultimately result in resistance breakdown in resistant chili varieties (Singh et al. 2016). Expressions of various defense genes have also been reported to downregulate synergistic interactions between Prunus necrotic ringspot virus (PNRSV) and Prunus latent mosaic viroid (PLMVd) (Herranz et al. 2013). Synergistic interactions between Potato virus X and Potato virus Y compromised chloroplast function, upregulation of both carbohydrate and protein degradation, and induction of various defense genes associated with biotic and oxidative stresses to cope with the situation (García-Marcos et al. 2009).

Taken together, in planta experiments, using viral mutants and leaf disc assays clearly suggest that AV2 and AC2 encoded proteins of ToLCGuV synergistically enhance ToLCNDV replication. We have identified TrAP protein (encoded by AC2) and pre-coat protein (encoded by AV2) of ToLCGuV, as the key regulatory proteins responsible for asymmetric synergism between two most prevalent tomato-infecting begomoviruses from India. During asymmetric synergism, ToLCGuV-encoded proteins might promote a suitable cellular environment for efficient replication of ToLCNDV either by blocking the host surveillance system or by blocking any other antiviral silencing pathway.