The Sw5a gene confers resistance to ToLCNDV and triggers an HR response after direct AC4 effector recognition

Significance

Tomato leaf curl New Delhi virus (ToLCNDV) infection causes severe losses in tomato yield worldwide. Lack of information on resistance (R) genes against ToLCNDV has considerably retarded the pace of crop improvement against this rapidly spreading pathogen. Here, we report an effective defense strategy deployed by a resistant tomato cultivar against ToLCNDV. It employs Sw5a (R gene) that recognizes AC4 protein (viral effector) of ToLCNDV to restrict virus spread. At the transcriptional level, the sly-miR159-SlMyb33 module has been identified as governing gene expression of Sw5a. Thus, we provide a mechanistic insight into sly-miR159-SlMyb33–controlled Sw5a-mediated defense response in tomato against ToLCNDV. These findings could be translated into development of resistance in susceptible cultivars of tomato through modern breeding or molecular approaches.

Abstract

Several attempts have been made to identify antiviral genes against Tomato leaf curl New Delhi virus (ToLCNDV) and related viruses. This has led to the recognition of Ty genes (Ty1-Ty6), which have been successful in developing virus-resistant crops to some extent. Owing to the regular appearance of resistance-breaking strains of these viruses, it is important to identify genes related to resistance. In the present study, we identified a ToLCNDV resistance (R) gene, SlSw5a, in a ToLCNDV-resistant tomato cultivar, H-88-78-1, which lacks the known Ty genes. The expression of SlSw5a is controlled by the transcription factor SlMyb33, which in turn is regulated by microRNA159 (sly-miR159). Virus-induced gene silencing of either SlSw5a or SlMyb33 severely increases the disease symptoms and viral titer in leaves of resistant cultivar. Moreover, in SlMyb33-silenced plants, the relative messenger RNA level of SlSw5a was reduced, suggesting SlSw5a is downstream of the sly-miR159-SlMyb33 module. We also demonstrate that SlSw5a interacts physically with ToLCNDV-AC4 (viral suppressor of RNA silencing) to trigger a hypersensitive response (HR) and generate reactive oxygen species at infection sites to limit the spread of the virus. The “RTSK” motif in the AC4 C terminus is important for the interaction, and its mutation completely abolishes the interaction with Sw5a and HR elicitation. Overall, our research reports an R gene against ToLCNDV and establishes a connection between the upstream miR159-Myb33 module and its downstream target Sw5a to activate HR in the tomato, resulting in geminivirus resistance.

Tomato leaf curl New Delhi virus (ToLCNDV) belongs to the family Geminiviridae, which are characterized by their twin-icosahedral structure with single-stranded DNA as their genome (1–3). ToLCNDV is a bipartite Begomovirus containing two genome components (DNA-A and DNA-B). The DNA-A encodes proteins involved in replication and transcription, whereas DNA-B encodes for proteins required for the movement of virus, on the virion- or complementary-sense orientation (1, 2). The typical symptoms of ToLCNDV infection include leaf curling, stunted growth, and mostly death of the entire plant, leading to severe losses (up to 100%) in tomato production in several parts of the world (1, 3). Reports highlight the involvement of Resistance (R) genes in conferring resistance to viral pathogens through their interaction with the viral effectors. These interactions lead to the activation of hypersensitive responses (HR) to limit the virus's spread, thus showing resistance/tolerance (4, 5). However, the precise mechanism and factors regulating resistance against ToLCNDV remain unknown and need to be explored to better understand plant–virus interaction.

In plants, microRNA (miRNA)-mediated regulation of gene expression has been shown to modulate plant growth, development, and response to various environmental stresses (6, 7). miRNAs are noncoding RNAs of ∼20 to 24 nucleotides (nt) in size and regulate the expression of their target genes through translational inhibition or cleavage of corresponding messenger RNAs [mRNAs (6–9)]. Studies have shown the involvement of miRNAs in response to plant–virus interaction. For example, miRNAs of different families and their targets were differentially expressed upon virus infection of diverse genera like Cowpea severe mosaic Virus, Cotton leaf curl Burewala virus, Potato virus X, and Mungbean Yellow Mosaic India Virus (10–16). Based on in silico prediction and transcript analysis, few studies identified miRNA and their targets during ToLCNDV infection. Enhanced expression of miR159/319 and miR172 was detected in ToLCNDV-infected tomato cv. Pusa Ruby, whereas transcript levels of their predicted targets were reduced (17, 18). Although these studies highlighted the involvement of miRNAs during ToLCNDV infection, a mechanistic understanding remains lacking.

Plants encode several disease R genes that recognize specific effector proteins and establish effector-triggered immunity to initiate hypersensitive resistance response (19). miRNAs have been reported to target R genes (mostly NBS-LRR) and subsequently reduce their cellular abundance (15, 16, 20–24). For example, miR482/2118 were found to cleave different NBS-LRR mRNAs during immune response against bacterial, fungal, and viral infection (15, 20). Similarly, SlNgene (R gene against Tobacco mosaic virus) and Mla transcripts (R genes against powdery mildew fungus) were found to be regulated by miR6019/6020 and miR9863, respectively (16, 24). Overall, these findings suggest that miRNA-mediated gene silencing might act as a key regulator of R gene–mediated defense responses. However, the exact mechanism by which miRNA modulates R gene expression and their effect on the host's virus defense strategies needs further investigation.

In our previous study, we have shown that a naturally resistant tomato cultivar (cv., H-88-78-1) exhibited HR and accumulated a relatively higher level of small RNAs (sRNAs) compared with naturally susceptible cultivar (Punjab Chhuhara) following ToLCNDV infection (25, 26). Previous attempts to identify R genes against Tomato leaf curl virus (ToLCV) and related species have resulted in the identification of Ty genes. Although these genes proved to be important for tolerance/resistance against ToLCV, they are not present in cv. H-88-78-1. Thus, we looked for an alternate approach to identify R genes that might be involved in ToLCNDV resistance. Since several miRNAs are known to target R genes (15, 16, 19–24), we performed a miRNAome in these resistant and susceptible cultivars during ToLCNDV infection. Sly-miR159 was found to be differentially regulated among these cultivars, and it targets SlMyb33. Furthermore, SlMyb33 regulates the Sw5 (R gene)–mediated HR pathway. Sw5 consists of five paralogs, among which Sw5b confers resistance against orthotospoviruses; Tomato spotted wilt virus (TSWV) in tobacco and tomato (27, 28). In our study, SlMyb33 modulates the expression of Sw5a^S paralog, which is responsible for susceptibility against TSWV. This paralog (henceforth SlSw5a) interacts with ToLCNDV AC4 protein to activate HR. The present study evidently reveals a pathway in plant's response to ToLCNDV infection, where the expression of sly-miR159 is associated with SlMyb33-mediated transcriptional activation of SlSw5a, providing resistance to ToLCNDV by triggering the hypersensitive response.

Results

Resistant and Susceptible Cultivars of Tomato Have Uniquely Enriched Endogenous and Novel miRNAs during ToLCNDV Infection.

Previous studies have specified that resistance to ToLCV and related species is linked with the presence of Ty genes (1, 3, 29, 30). Tomato cultivars differing in their resistance level to ToLCNDV (25) (cv. H-88-78-1, resistant; cv. Punjab Chhuhara, susceptible, Fig. 1A) were screened for the presence of Ty genes. It was observed that both cultivars did not carry any functional Ty gene in the homozygous state (SI Appendix, Fig. S1). Apart from R genes–mediated regulation of viruses, miRNAs have been reported as crucial modulators of host–virus interactions, importantly by regulating the expression of R genes resulting in antiviral immunity (10–18). Therefore, to explore if the potential mechanisms providing H-88-78-1 a resistant attribute are associated with miRNAs, sRNA libraries were constructed and sequenced under mock-treated and ToLCNDV-infected conditions in both cultivars.

Fig. 1.

Identification of ToLCNDV-responsive miRNAs. (A) Phenotypes of mock-inoculated and ToLCNDV-infected resistant (H‐88‐78‐1) and susceptible (Punjab Chhuhara) tomato cultivars at 21 dpi. (B) Percentage length distribution of miRNAs identified in mock (M) and infected (I) sRNA libraries from both cultivars. (C) heat map depicting the relative expression pattern of 13 miRNAs in resistant and susceptible cultivars of tomato under mock (M) and ToLCNDV-infected (I) conditions. For normalization, U6 small nuclear RNA was used as endogenous control. (D and E) Expression analysis of SlMyb33 and sly-miR159 by Northern blot and qRT-PCR analysis in the resistant and susceptible cultivar after ToLCNDV infection (n = 3), which reveals contrasting expression of SlMyb33 in comparison with sly-miR159. Each experiment was conducted in three biological replicates. Alpha-Tubulin was used as an endogenous control. *, **, and ***, significant differences between values (i.e., P < 0.05, P < 0.01, and P < 0.001, respectively).

Overall, 60 million high-quality reads were obtained, which were further mapped onto the tomato genome (SI Appendix, Table S1). A total of 182 miRNAs were identified, including 137 known and 45 novel miRNAs (Dataset S1). The size distribution of miRNAs in all the libraries ranged from 20 to 24 nt, with a predominant distribution of 21-nt miRNAs (61.28%) (Fig. 1B). Among these miRNAs, 13 potential miRNAs (9 known and 4 novel) that showed at least 1.5-fold change in the expression pattern between the control and ToLCNDV-infected condition in either of the cultivars were selected. Their expression was further validated through Northern blot analysis. Distinct expression profiles were observed for each miRNA in mock-treated and virus-infected samples, which was in consistent with the deep sequencing data (Fig. 1C and SI Appendix, Fig. S2). Among known miRNAs, sly-miR159 exhibited 3.0-fold up-regulation in susceptible cv. and 4.0-fold down-regulation in the resistant cultivar. However, sly-miR9472-3p was down-regulated by 2.0-fold in the susceptible cultivar and up-regulated by 3.0-fold in the resistant cultivar. On the other hand, miR6024 was up-regulated by 2.4-fold in susceptible cultivar and down-regulated by 1.6-fold in resistant cultivar (Fig. 1C and SI Appendix, Fig. S2). In contrast to known miRNAs, the fold change of novel miRNAs, namely sly-NovmiR2, sly-NovmiR4, sly-NovmiR15, and sly-NovmiR16 were down-regulated in both susceptible and resistant cultivars (Fig. 1C and SI Appendix, Fig. S2). Based on the expression profiling data, sly-miR159 was the highest differentially expressed miRNA in both cultivars during ToLCNDV infection. Since numerous miRNAs are known to negatively regulate their targets, we assumed that contrasting expression of miR159 in both cultivars could be one of the factors regulating infection against ToLCNDV (Fig. 1 D and E). Hence, we selected miR159 for downstream analysis to decipher its role in ToLCNDV infection.

Sly-miR159 Targets Myb33 Transcription Factor in the Tomato.

miRNAs function by negatively regulating the expression of their target (8); therefore, we predicted the targets of miR159 through psRNATarget web server (https://www.zhaolab.org/psRNATarget/). A total of eight putative targets were predicted through in silico analysis [with stringent Minimum Free Energy (MFE) cutoff threshold of ≤2.5; Dataset S2)]. This includes Myb transcription factor family (Solyc01g009070, Solyc01g090530, Solyc06g073640, and Solyc10g019260), Cotton fiber expressed protein 1 (Solyc02g079010), Serine/threonine kinase (Solyc06g008320), Glycosyltransferase-like protein (Solyc07g052640), and Unknown Protein (Solyc12g014120) (Dataset S2). However, RNA Ligase Mediated-Rapid Amplification of cDNA Ends (RLM-RACE) showed the amplification of only Solyc01g009070 (a Myb transcription factor, SlMyb33), overruling the virtual prediction (SI Appendix, Fig. S3A). Sly-miR159 was found to cleave between the 10th and 11th base of its complementary target site in SlMyb33 (SI Appendix, Fig. S3B). Transcript accumulation of SlMyb33 was significantly up-regulated (2.5-fold; P < 0.001) and down-regulated (twofold; P < 0.001) in resistant and susceptible cv., respectively, upon ToLCNDV infection as compared with their respective mock-treated samples (Fig. 1 D and E). To validate that sly-miR159 targets SlMyb33, short tandem target mimic (STTM)-mediated knockdown approach was used to inhibit the expression of sly-miR159 (construct named STTM159). The transient expression of STTM159 in both cultivars showed a dramatic increase in the expression levels of SlMyb33 in STTM159 plants compared with the control plant (SI Appendix, Fig. S3C), suggesting sly-miR159–mediated cleavage of SlMyb33 transcripts.

Silencing of SlMyb33 Exhibited ToLCNDV Susceptibility Attribute in Resistant Cultivar.

To investigate whether SlMyb33 modulates ToLCNDV infection, virus-induced gene silencing (VIGS) was performed in naturally resistant cv. H-88-78-1 (hereafter, H_TRV:SlMyb33) and susceptible cv. Punjab Chhuhara (hereafter, PC_TRV:SlMyb33) followed by virus infection (25). Tomato Phytoene desaturase (SlPDS)-silenced lines (TRV:SlPDS) were used as positive control to assess the efficacy of vector. The characteristic photobleaching effect appeared in the topmost leaves after 10 d postsilencing (dps) and enhanced up to 21 d thereafter in SlPDS-silenced plants (SI Appendix, Fig. S4 A and B). To validate the silencing, transcript abundance of SlMyb33 and SlPDS were examined through qRT-PCR. Compared with mock-inoculated plants (TRV:00), ∼65 to 70% reduction in the transcript abundance of SlMyb33 and SlPDS was observed after VIGS of respective genes in both cultivars (SI Appendix, Fig. S4 C–F).

Furthermore, ToLCNDV-infectivity assay was performed, and phenotypic analysis of SlMyb33-silenced H-88-78-1 plants showed severe ToLCNDV infection symptoms compared with control and mock-inoculated plants (Fig. 2A and SI Appendix, Fig. S4). Disease symptoms of virus infection in SlMyb33-silenced plants include the early appearance of curly leaf and stunted growth (at 8 dpi) compared with the appearance of visible symptoms at 14 dpi in control plants (Fig. 2A and SI Appendix, Table S2). Additionally, ∼80% of SlMyb33-silenced plants displayed a typical leaf-curling phenotype at 21 dpi. Furthermore, SlMyb33 silencing led to >fourfold increase in viral titer in resistant cv. H-88-78-1 compared with mock-inoculated plants at 21 dpi (Fig. 2 B and C). Simultaneously, SlMyb33-silenced Punjab Chhuhara lines (used as negative controls) also showed severe leaf curling (SI Appendix, Fig. S4G and Table S3) and ∼twofold increase in viral titer when compared with control plants at 21 dpi (SI Appendix, Fig. S4 H and I).

Fig. 2.

Functional assessment of SlMyb33 upon ToLCNDV infection by VIGS. (A) Myb33 silencing phenotype of cv. H-88-78-1 under control and ToLCNDV-inoculated conditions. Control, control cv. H-88-78-1; H_TRV:00, cv. H-88-78-1 as a mock inoculated with TRV:00 VIGS vectors; H_TRV:SlMyb33, SlMyb33-silenced cv. H-88-78-1. (B) Southern blot analysis, to study the accumulation of viral titer in ToLCNDV-noninfected (1–3) and infected (4–6) control, mock, and SlMyb33-silenced plants. Genomic DNA was hybridized with ToLCNDV coat protein gene–specific probe (n = 3). Viral replicative forms such as open circular (OC), linear (Lin), supercoiled (SC), and single strand (SS) have been indicated. (C) Relative accumulation of virus genome in the experimental samples. ***, significant difference between the values (i.e., P <0.001).

To rule out the possible off-target silencing due to the higher percentage of sequence homology shared by the other member of Myb gene family, the expression pattern of six potential off-target Myb genes was analyzed in SlMyb33-silenced plants of cv. H-88-78-1. These six genes (Solyc05g007160, Solyc07g054980, Solyc09g010820, Solyc11g072060, Solyc08g082890, and Solyc06g073640) shared ∼21- to 23-nt continuous similarity within the coding region of SlMyb33 (SI Appendix, Fig. S5A). It was observed that none of the other six genes showed significant changes in the expression after SlMyb33 silencing, suggesting no off-target silencing in the samples (SI Appendix, Fig. S5B).

SlMyb33 Possesses DNA Binding Ability.

The Myb transcription factor has been shown to possess DNA binding properties onto the cis-element “TAACAAA” (31, 32). To validate the binding specificity of SlMyb33 to cis-element, “TAACAAA” yeast one-hybrid (Y1H) assay was performed. Yeast cells cotransformed with pGAD:SlMyb33 and pAbAi:Prm (containing the cis-element, “TAACAAA”) grew normally in the presence of a minimal inhibitory concentration of Aureobasidin A (AbA, 100 ng/mL) (SI Appendix, Fig. S6). However, the cells transformed with pAbAi:Prm, and constructs without SlMyb33 failed to grow on it. The cells cotransformed with pGAD:SlMyb33 and ΔpAbAi:Prm (containing the mutated cis-element, “TgAgAAA”) also failed to grow in a minimal inhibitory concentration of AbA (SI Appendix, Fig. S6). These ascertained the specificity of Myb33 in binding “TAACAAA” element in the promoter region.

Sly-miR159-SlMyb33 Promotes the Expression of the Resistance Gene, SlSw5a.

To identify potential targets of SlMyb33, we performed chromatin immunoprecipitation (ChIP), coupled with Illumina sequencing (ChIP-Seq). We transiently overexpressed Myc-tagged SlMyb33 in leaves of cv. H-88-78-1 and used anti–Myc-antibody to immunoprecipitate SlMyb33-associated DNA. A total of ∼170 million reads were obtained from two biological replicates of each input control (IP_rep1 and IP_rep2) and immunoprecipitated samples (Myb33_rep1 and Myb33_rep2) (SI Appendix, Table S4 and Fig. S7). About 68% of SlMyb33-binding elements (MBE) were observed to be located in the promoter regions, within the 1,000-base pairs (bp) upstream region of the transcription start site (TSS) (Fig. 3A).

Fig. 3.

Analysis and validation of SlMyb33-mediated regulation of SlSw5a. (A) Distribution of SlMyb33 binding peaks pattern per 1,000 bp in −5,000- to +5,000-bp region flanking the TSS in the immunoprecipitated samples (Myb33_rep1 and Myb33_rep2). (B) Heat map of qRT-PCR analysis of genes consisting of MBE in their promoter region in order to identify the SlMyb33-mediated regulation of these genes in response to ToLCNDV infection. (C) qRT-PCR analysis of SlSw5a in response to ToLCNDV infection in experimental plants. Figure depicts up-regulation of SlSw5a in resistant cv. and down-regulation in susceptible cv. upon virus infection. Each experiment was conducted in three biological replicates. Alpha-Tubulin was used as an internal control. (D) EMSA to examine the interaction between SlMyb33 and SlSw5a promoter. Binding of SlMyb33 protein onto 102-bp αP³²-dCTP–labeled fragment of SlSw5a promoter is depicted by retarded DNA–protein complex through electrophoretic mobility shift (lane 4). Lane 1 shows free DNA probe only. As negative control, GST was incubated with Sw5a promoter probe (lane 2). Lane 3 shows substantially diminished ability of SlMyb33 to bind onto the labeled mutated counterpart of SlSw5a promoter, SlSw5a (Mut). Specificity of retarded DNA–protein bands was confirmed by competition assay by using with surplus of unlabeled probe (100, 50, and 25 molar excess, depicted through an open triangle, lanes 5 to 7). Signs + or − represent the presence or absence of components. (E) In vivo binding of SlMyb33 onto the SlSw5a promoter. ChIP assay was done by transiently overexpressing SlMyb33 in tomato leaves. Figure depicts the amplification of SlSw5a promoter from the DNA immunoprecipitated from the sample by using c-myc-tag and immunoglobulin G (IgG-negative control) antibody. Moreover, SlMyb33 binding activity onto SlNgene and SlTm2 gene promoters were also observed by ChIP assay. Total input chromatin (input) was used for the linear amplification. Actin7 gene was used as a negative control. Bars represent the SDs (± SD). *, **, and ***, significant differences between values (i.e., P < 0.05, P < 0.01, and P < 0.001, respectively).

Among the 121 potential targets having SlMyb33 binding sites (SI Appendix, Table S5), 21 genes showing ≥twofold enrichment (P < 0.05) in immunoprecipitated samples as compared with input samples were selected, and their expression pattern was examined in SlMyb33-silenced lines (in resistant cv. H-88-78-1) (Table 1 and SI Appendix, Fig. S8). A significant reduction in transcript abundance of all the 21 genes was observed in SlMyb33-silenced plant compared with control and mock-inoculated (TRV:00) plants (Fig. 3B). To correlate these genes in the event of ToLCNDV infection, we evaluated the expression profile of these genes in both susceptible and resistant cultivar of tomato (control versus infected). A gene encoding for a leucine-rich repeat–containing protein, SlSw5a (Solyc09g098130), showed ∼sixfold up-regulation in resistant cultivar and ∼sevenfold down-regulation in susceptible cultivar upon ToLCNDV infection in comparison with respective controls (Fig. 3 B and C). This gene (located on chromosome 9) showed ∼10-fold enrichment in both immunoprecipitated samples compared with input samples (SI Appendix, Fig. S7). To identify 5′ Untranslated Region (UTR) and the precise location of SlMyb33 binding element in SlSw5a promoter, 5′ RACE was performed. Sequence analysis revealed the presence of 244-bp long 5′ UTR in SlSw5a (SI Appendix, Fig. S9A). The SlMyb33 binding domain was located at 269 bp upstream of TSS of SlSw5a (SI Appendix, Fig. S9B).

Table 1.

List of putative genes regulated by SlMyb33 identified by ChIP-Seq

Gene ID	Name of gene	Distance to TSS	Fold enrichment Myb/IP		Myb33_Rep1		Myb33_Rep2
			Myb33_Rep1	Myb33_Rep2	-log10 P value	-log10 q value	-log10 P value	-log10 q value
Solyc01g017070	Ycf2	−455	3.6084	3.0825	82.2457	76.9054	67.5062	62.1588
Solyc01g017160	NAD(P)H-quinone oxidoreductase subunit 2	−301	2.4923	2.1617	50.9977	46.4167	40.0566	35.3909
Solyc01g017430	Cytochrome b6-f complex subunit 4	−477	3.0663	2.6628	71.6516	66.4597	59.7064	54.5529
Solyc01g020200	Ycf1	−200	2.4138	2.0626	44.6544	40.3442	34.9582	30.4639
Solyc01g065640	NAD(P)H-quinone oxidoreductase chain 4	−40	2.0013	1.9747	17.3969	14.1346	17.8257	14.1664
Solyc02g011820	Ycf1	−59	3.5457	2.8280	77.0824	71.8046	59.9934	54.8340
Solyc02g011990	Photosystem II protein D1	−546	2.6267	2.0641	59.5177	54.6219	36.7581	32.2004
Solyc02g012020	ATP synthase subunit b	−108	2.0851	1.9461	19.1500	15.8258	19.0049	15.2930
Solyc02g063080	NAD(P)H-quinone oxidoreductase subunit K	−56	2.4229	1.9447	56.084	51.309	32.1724	27.7844
Solyc04g039770	Ycf2	−425	3.1745	2.7011	66.3492	61.2477	54.5155	49.5008
Solyc04g045550	Ycf1	−67	2.9168	2.3681	69.3019	64.1526	50.1003	45.1912
Solyc07g049220	Serine/threonine protein phosphatase 2A regulatory subunit B	−1580	2.3434	1.9313	48.5808	44.0994	32.2911	27.8996
Solyc08g045620	Ycf2	−135	2.8462	2.4663	54.3765	49.6644	47.7261	42.8711
Solyc08g045870	Ribosomal protein S3	−1236	2.0993	1.9880	21.2639	17.8718	21.0456	17.1334
Solyc08g066020	Long chain base biosynthesis protein 1	−273	2.2752	2.060	35.8045	32.0184	31.3216	26.9645
Solyc09g055840	NAD(P)H-quinone oxidoreductase subunit 2	−48	2.4234	1.9476	53.4407	48.7626	33.0837	28.6618
Solyc09g098130	Sw5 CC-NBS-LRR, resistance protein	−269	11.0768	10.0768	114.4539	108.765	104.4534	94.4539
Solyc10g047410	Photosystem II CP43 reaction center protein	−445	2.4328	1.9675	50.3042	45.7518	33.2327	28.8058
Solyc10g047460	NAD(P)H-quinone oxidoreductase subunit H	−557	2.2010	2.5809	41.4637	37.3069	50.9719	46.0437
Solyc11g044610	NAD(P)H-quinone oxidoreductase chain 4	−267	2.4496	1.8966	54.8154	50.0881	30.9126	26.5697
Solyc12g035890	DNA-directed RNA polymerase subunit alpha	−337	2.3772	2.0060	25.8121	22.2870	19.2054	15.4819

SlMyb33 Binds onto the Promoter Element of SlSw5a but Not onto Other R Genes.

To confirm whether SlMyb33 binds onto the promoter of SlSw5a gene, electrophoretic mobility shift assay (EMSA) was performed. A 102-bp radioactively labeled DNA fragment of SlSw5a promoter consisting of MBE (wild type, TAACAAA) was incubated with recombinant SlMyb33-GST fusion protein (SI Appendix, Fig. S9 C and D), resulting in the appearance of high-affinity protein–DNA complex in the gel (Fig. 3D; lane 4). To further validate the transcriptional regulation of SlSw5a, we mutated MBE by site-directed mutagenesis (SDM), and a mutated probe (SlSw5a Mut, TgAgAAA) was generated. No complex was observed in the presence of mutated SlSw5a promoter region (Fig. 3D; lane 3). Simultaneously, in the competition assay, unlabeled wild-type MBE element was able to compete with labeled wild-type probe in binding with SlMyb33 (Fig. 3D; lanes 5 to 7). Furthermore, ChIP-PCR assay was performed to identify the binding affinities of SlMyb33 onto promoter of other probable known R genes (SlTm2 and SlNgene) of tomato. The leaves of resistant cultivar transiently overexpressing Myb33-cMyc were used. Amplification of only the targeted region of SlSw5a from input control and immunoprecipitated samples confirmed the binding specificity of SlMyb33 specifically to SlSw5a but not to SlTm2 and SlNgene (Fig. 3E). The results suggested that SlMyb33 has the unique capability to regulate the expression of SlSw5a by binding to MBE present in its promoter.

SlSw5a Silencing Confers ToLCNDV Susceptibility to Resistant Cultivar.

To understand whether SlSw5a is involved in resistance against ToLCNDV infection, SlSw5a silencing was performed using the TRV-based VIGS approach in resistant (H_TRV:SlSw5a) and susceptible cv. (PC_TRV:SlSw5a). Silencing of SlSw5a was confirmed by decreased transcript abundance of SlSw5a (>70%) in silenced plants as compared with mock-inoculated (TRV:00) plants (SI Appendix, Fig. S10 A and B). The ToLCNDV-infectivity assay was performed in SlSw5a-silenced plants, and disease symptoms were recorded. At 21dpi, typical viral disease symptoms of stunted growth and severe leaf curling were observed in SlSw5a-silenced plants of both cultivars (Fig. 4A and SI Appendix, Fig. S10C and Tables S6 and S7). To quantify the disease, viral titer was assayed, which showed higher viral accumulation in H_TRV:SlSw5a (>threefold) and PC_TRV:SlSw5a (∼1.5-fold) plants in comparison with respective mock-treated plants (Fig. 4 B and C and SI Appendix, Fig. S10 D and E), suggesting that Sw5a may regulate ToLCNDV infection.

Fig. 4.

Role of SlSw5a in providing resistance against ToLCNDV via HR activation. (A) Phenotypes of cv. H-88-78-1 after SlSw5a silencing under control and ToLCNDV-inoculated conditions. H_TRV:00, cv. H-88-78-1 as a mock inoculated with TRV:00 VIGS vectors; H_TRV:SlSw5a, SlSw5a-silenced cv. H-88-78-1; plants were grown together in the same growth chamber. (B) Southern blot analysis to study the virus accumulation in ToLCNDV-noninfected (1–3) and infected (4–6) control, mock, and SlSw5a-silenced plants (n = 3). Viral replicative forms OC, Lin, SC, and SS have been indicated. Electrophoretic gel stained with ethidium bromide is shown as a loading control. (C) Relative accumulation of virus genome in the experimental samples. Data represents mean ± SD of three independent experiments; ***P < 0.001. (D) Level of viral DNA accumulation in ToLCNDV-noninfected (1–5) and infected (6–11) control, SlSw5a-silenced H-88-78-1 (TRV:SlSw5a) plants, agro-inoculated samples of overexpressed SlMyb33 (OE-SlMyb33, TRV:SlSw5a+OE-SlMyb33, OE-SlMyb33 (PC)) depicting SlSw5a as downstream target of SlMyb33. DNA was probed with ToLCNDV coat protein gene (n = 3). Viral replicative forms OC, Lin, SC, and SS have been indicated. (E) Relative accumulation of viral DNA in the experimental samples. Data represents mean ± SD of three independent experiments; **P < 0.01. (F) Leaves of mock-treated, SlMyb33-silenced H-88-78-1 (TRV:SlMyb33), SlSw5a-silenced H-88-78-1 (TRV:SlSw5a) plants, and SlSw5a-silenced H-88-78-1 plants overexpressing SlMyb33 (TRV:SlSw5a+OE-SlMyb33) were agro infected with ToLCNDV. Dead cell staining was performed by Trypan blue treatment, and oxidative burst–related damage was detected by DAB staining. Leaves of the mock plants showed the visual necrotic response (HR lesions) on infiltration whereas leaves of SlMyb33- and SlSw5a-silenced plants, and SlSw5a-silenced plants overexpressing SlMyb33 showed relatively less production of HR lesions.

Since SlMyb33 binds to the promoter region of SlSw5a and might regulate its expression, overexpression of SlMyb33 should induce the transcript abundance of Sw5a and regulate virus accumulation. To verify this, we transiently overexpressed SlMyb33 in the susceptible cultivar, along with the virus. A significant reduction of virus titer (∼70%) in these transiently overexpressed SlMyb33 leaves was observed, suggesting that SlMyb33 induces the expression of Sw5a, which provides resistance against ToLCNDV (Fig. 4 D and E; lanes 5, 10, and 11). On the contrary, when SlSw5a-silenced plants (in the background of cv. H-88-78-1) were transiently overexpressed with SlMyb33, no significant difference in the viral titer was observed, compared with SlSw5a-silenced plants only (Fig. 4 D and E; lanes 1 to 4 and 6 to 9). These results further demonstrate that SlMyb33 restricts the ToLCNDV infection by specifically regulating the expression of SlSw5a.

SlSw5a Regulates HR and Antioxidant Enzyme Activities during ToLCNDV Infection.

The SlSw5a identified in our study shared 94.5% amino acid (aa) sequence identity with SlSw5a^S (SI Appendix, Fig. S11A), a potential marker linked with TSWV susceptibility in Solanum lycopersicum (27, 28). We observed that SlSw5a from both resistant and susceptible cultivars had a unique single aa mutation (Q/R) in the NB-ARC domain compared with SlSw5a^S (SI Appendix, Fig. S11B). Thus, to evaluate that Sw5a could induce HR upon infection with ToLCNDV, we performed infection assays in leaves of SlMyb33-, SlSw5a-, and SlSw5a-silenced leaves overexpressing SlMyb33 (in the background of cv. H-88-78-1) and compared with mock treated (TRV:00) (Fig. 4F). Characteristic HR lesions were observed in the mock-treated sample leaves, whereas SlMyb33-, SlSw5a-, and SlSw5a-silenced leaves overexpressing SlMyb33 failed to generate HR reaction (Fig. 4F). Accumulation of H₂O₂ in leaves was detected by 3,3′-diaminobenzidine (DAB) staining, and the clear visualization of reddish-brown spots in mock-infected plants indicated higher H₂O₂ accumulation. Stained patches in SlMyb33- and SlSw5a-silenced plants showed a reduced accumulation of H₂O₂ upon virus infection. This was further confirmed by the appearance of dead cell patches when stained with trypan blue (Fig. 4F). To further validate that Q/R reversion is the prerequisite for resistance against ToLCNDV, Money Maker (MM) tomato plants (contains SlSw5a^S without Q/R reversion; SI Appendix, Fig. S12A) were infected with ToLCNDV. The MM plants showed severe symptoms of ToLCNDV infection when compared with control plants at 21 dpi (SI Appendix, Fig. S12 B–D). Simultaneously, reduced stain patches were observed in DAB-stained MM leaves upon ToLCNDV infiltration compared with H-88-78-1, highlighting the importance of Q/R reversion in Sw5a for initiating HR against ToLCNDV (SI Appendix, Fig. S12E and Fig. 4F).

Furthermore, to validate that HR was generated as a result of higher levels of H₂O₂ production, the levels of H₂O₂-scavenging enzymes, catalase (CAT) and ascorbate peroxidase (APX), were observed in all the experimental samples infected with ToLCNDV. The CAT and APX activities were significantly reduced (P < 0.01 and 0.05, respectively) upon virus infection in nonsilenced control and mock-treated cv. H-88-78-1 leaves (SI Appendix, Fig. S13 A and B). On the contrary, an increase in enzymatic activity of CAT and APX was observed in SlMyb33- and SlSw5a-silenced plants upon virus infection, resulting in reduced accumulation of H₂O₂ and leading to the development of disease symptoms (SI Appendix, Fig. S13 A and B). Simultaneously, higher levels of malondialdehyde (MDA) content were observed in all the virus-infected experimental plants. However, the content was relatively lower in both SlMyb33- and SlSw5a-silenced plants (∼21%) as compared with virus-infected control and mock-treated plants (∼52%) (SI Appendix, Fig. S13C). These results highlight a strong correlation of lipid peroxidation with the development of necrotic lesions to inhibit the viral spread in the resistant cultivar. Thus, it suggested a decrease in reactive oxygen species (ROS) level upon silencing of SlMyb33 and SlSw5a could confer ToLCNDV susceptibility to the naturally resistant cv. H-88-78-1.

ToLCNDV AC4 Is the Avirulence Factor of SlSw5a-Mediated HR Response.

The majority of R genes identify a specific pathogen by their effector, termed as avirulence factors (Avr) (19, 33, 34). The Avr of ToLCNDV, which could trigger HR via R genes activation remains unknown. To identify the probable ToLCNDV component involved in triggering the molecular response against ToLCNDV, the interaction of SlSw5a with all virus proteins (Fig. 5A) was analyzed via yeast two-hybrid assay (Y2H). The screening suggested that SlSw5a specifically interacts with AC4 protein (viral suppressor of RNA silencing) of ToLCNDV (Fig. 5B). In GST-pulldown assay too, SlSw5a was found to strongly interact with AC4, and no binding was observed between GST and SlSw5a (Fig. 5C). The physical interaction between SlSw5a and AC4 was also verified in vivo via bimolecular fluorescence complementation (BiFC) and coimmunoprecipitation (Co-IP) assays in tobacco leaves. Upon coinfiltration of Sw5a and AC4, a strong Yellow Fluorescent Protein (YFP) signal was observed in the membrane compared with controls (SI Appendix, Fig. S14). Consistently, in Co-IP assay, AC4 was strongly pulled down by Sw5a, thus validating that AC4 may act as the Avr for SlSw5a (Fig. 5D).

Fig. 5.

Sw5a interacts with AC4. (A) Organization of ToLCDNV genome. Orientations of corresponding ORFs are depicted with arrows. DNA-A–encoded proteins [i.e., AC1 (replication initiation protein); AC2 (transcription activator); AC3 (replication enhancer); AC4 (pathogenesis-related protein); AC5; AV1 (coat protein); AV2 (precoat protein); and AV3] are depicted in arrows according to their orientation in the genome. DNA-B–specific ORFs encoding BV1/NSP (nuclear shuttle protein) and BC1/MP (movement protein) are shown. (B) Yeast two hybrid assays performed between SlSw5a and 10 virus-encoded genes to identify the interacting partner of SlSw5a. (C) AC4 interacts with Sw5a in GST pull down assay. Bacterially expressed GST, AC4-GST, and Sw5a-His fusion proteins were used to perform pull down. GST/AC4-GST (∼1 mg)-bound GSH-Sepharose beads were incubated with purified Sw5a-His protein (∼1 mg) at 4 °C for 2 h and washed with 1× Phosphate Buffered Saline (PBS). Sw5a-His protein was detected by Western blotting with anti-His antibody. AC4-GST was detected with anti-GST antibody and Sw5a-His with anti-His antibody to represent the input control. (D) Sw5a interacts with AC4 by Co-IP assay in tomato. Protein extracted from tomato leaves coinfiltrated with Sw5a-Myc, and AC4-HA or Sw5a-Myc and pSPYCE-EV were immunoprecipitated (IP) by anti–HA antibody–conjugated protein-G magnetic beads, and immunoblot (IB) was developed with anti-Myc antibody. Sw5a-Myc was detected by anti-Myc antibody and AC4-HA by anti-HA antibody to represent the input control. (E) Depicted are Y2H assays. The C terminal of AC4 [AC4 (C)] was found to have very strong interaction with SlSw5a, where aa residues from R46 to K49 were necessary to interact with Sw5a. (F) BiFC analysis confirmed that “RTSK” motif is crucial for interaction of AC4 with SlSw5a in Nicotiana leaf. Fluorescent and bright field images of Nicotiana leaves infiltrated with equimolar mix of different plasmids. The leaves infiltrated with pSPYNE: SlSw5a and pSPYCE: AC4 plasmid DNA shows fluorescence in the cell membrane. No fluorescent signal is detected in leaves co infiltrated with pSPYNE: SlSw5a and pSPYCE: ΔAC4 (AC4 mutant) plasmid and negative controls infiltrated with mixes of vector pSPYNE (empty vector) and pSPYCE: AC4; pSPYNE: Sw5a and pSPYCE (empty vector) plasmids. PM_mCherry; marker for the plasma membrane.

The AC4 open reading frame (ORF) completely overlaps with ORF of AC1 (Replication initiation protein, Rep) (SI Appendix, Fig. S15A). Therefore, to further investigate that AC4 is the inducer of HR associated with SlSw5a and not AC1, we coinfiltrated AC4 and AC1 with SlSw5a in the leaf of susceptible cultivar. The DAB staining showed H₂O₂ production in the leaves expressing AC4 and Sw5a, which was not observed in leaves expressing AC1 and Sw5a (SI Appendix, Fig. S15B). Furthermore, the level of H₂O₂-scavenging capacities of CAT and APX was significantly decreased in leaves coinfiltrated with SlSw5a and AC4 (∼1.5- and ∼2.5-fold, respectively) as compared with SlSw5a-infiltrated (∼1- and ∼1.2-fold, respectively) and SlSw5a and AC1– (∼1- and ∼1.2-fold, respectively) coinfiltrated leaves (SI Appendix, Fig. S15 C and D). The MDA content (∼22%) in Sw5a and AC4–coinfiltrated leaves was also higher than other experimental leaves (∼13%; SI Appendix, Fig. S15E). These results suggested that AC4–SlSw5a interaction induces HR reaction and cell death lesion upon ToLCNDV infection.

To identify the interface residues of AC4 protein responsible for interaction with SlSw5a, a homology model was generated using Phyre2 server (35). Sequence analysis of AC4 using IUPred2 showed the presence of an intrinsically disordered region (IDR) with a small helix of 4 aa (46-RTSK-49) along with a loop at C terminal of AC4 protein (SI Appendix, Fig. S16 A and B). IDRs offer numerous advantages in the context of protein–protein interactions (36), suggesting the possibility of “RTSK” to be working as Molecular Recognition Feature (MoRF) to nucleate interaction with SlSw5a. Furthermore, this motif of four aas, “RTSK,” is present only in AC4 and not in AC1 (SI Appendix, Fig. S16C). To validate the engagement of these four aas in interaction, we cloned N terminal [AC4 (N); aa 1 to 26] and C terminal [AC4 (C); aa 27 to 58] of AC4 and studied its interaction with SlSw5a in yeast. It was observed that SlSw5a have higher interaction specificity toward C-terminal variant of AC4 (Fig. 5E). To validate “RTSK” motif-based interaction of AC4 with SlSw5a, SDM was performed, and these residues were replaced with alanine (AC4 mutant, ΔAC4; SI Appendix, Fig. S16D). The interaction between SlSw5a and AC4 was completely abolished due to mutation in the RTSK motif (Fig. 5 E and F). The binding affinity of SlSw5a toward the RTSK motif of AC4 was additionally confirmed in vivo by BiFC assay. A strong YFP signal was observed in the membrane of Nicotiana benthamiana leaf cotransfected with Sw5a-YFP^N (pSPYNE:SlSw5a) and AC4-YFP^C (pSPYCE:AC4) but not in leaf cotransfected with Sw5a-YFP^N (pSPYNE:SlSw5a) and ΔAC4-YFP^C (pSPYCE:ΔAC4) (Fig. 5F). These results suggest that “RTSK” motif in AC4 is essential for this interaction, although mutation in “RTSK” motif did not influence the localization of AC4 (SI Appendix, Fig. S16E). Next, we performed SDM to pinpoint the key residue involved in interaction within the “RTSK” motif. BiFC assay showed S48 of ToLCNDV AC4 (highly conserved aa in ToLCNDV strains, SI Appendix, Fig. S17) is vital for interaction, as this single-aa mutation led to the abolition of interaction (SI Appendix, Fig. S18A). Notably, expression of Sw5a along with AC4 and its mutants, R46A, T47A, and K49A, produced HR as evident from the visible severe chlorotic spots. In contrast, infiltration of ΔAC4 and S48A showed compromised HR, with no chlorotic spots in resistant cv. H-88-78-1 (SI Appendix, Fig. S18B). Taken together, these results indicate that S48 of ToLCNDV is essential for the interaction between AC4 and Sw5a and HR initiation to limit the virus spread.

Majority of R genes consist of Leucine-rich repeat (LRR) domain, which plays a major role in ligand binding and determines the specificity (19). The LRR domain of SlSw5a shows 98% aa sequence identity with SlSw5a^S and 92% similarity to Sw5b (SI Appendix, Fig. S19A). Previous research has shown that LRR domain of Sw5b plays a role in conferring resistance against TSWV by activating HR when it interacts with TSWV-encoded movement protein NSm (27, 28). Therefore, to check if NSm could activate HR in cv. H-88-78-1 and cv. Punjab Chhuhara, Green Fluorescent Protein (GFP)-tagged NSm (SI Appendix, Fig. S19B) and AC4 were transiently overexpressed in leaves of both cultivars. No HR was observed in resistant and susceptible cultivar leaves upon NSm infiltration (SI Appendix, Fig. S19C). In addition, DAB staining revealed H₂O₂ accumulation at the site of AC4 infiltration in cv. H-88-78-1, whereas no H₂O₂ accumulation was observed at the site of NSm infiltration (SI Appendix, Fig. S19C). This result indicates that Sw5a copies in cv. H-88-78-1 can be specifically triggered by ToLCNDV.

Discussion

Previously, Ty genes have been linked with resistance against ToLCV (3, 29, 30); however, the molecular mechanism in resistant cultivars lacking these resistance genes remains elusive. The present study aimed to decipher the molecular mechanism underlying ToLCNDV resistance in the naturally resistant tomato cv. H-88-78-1 (25, 26), which produces HR on ToLCNDV infection but lacks known Ty genes. We identified an R gene, Sw5a, regulated by the miR159-Myb33 module, to provide resistance against ToLCNDV infection. Moreover, Sw5a physically interacts with AC4 protein of ToLCNDV to trigger HR and limit the virus spread.

Reduced Expression of sly-miR159 Subsequently Promotes Expression of Myb33 in Resistant Cultivar.

The defense arbitrated by R genes is the most sustainable antiviral mechanism (3, 19, 33, 34). Six ToLCV resistance genes (Ty-1, -2, -3, -4, -5 and -6) have been identified to be involved in antiviral defense mechanisms (3, 29, 30). However, tomato cultivars used in this study, H-88-78-1 (resistant) and Punjab Chhuhara (susceptible positive control, no Ty gene is present, 29), lack the above-known R genes (SI Appendix, Fig. S1). We further focused on finding the potential mechanisms providing H-88-78-1 a resistant attribute. Besides, R genes–mediated regulation of viruses, miRNAs have been reported as crucial modulators of host–virus interactions, resulting in antiviral immunity (10–18). Furthermore, miRNAs regulating the expression of R genes have been reported in different crops (15, 16, 20–24). Therefore, profiling of miRNAs and their related target genes was selected as an alternative approach to defining the resistance mechanism involved in cv. H-88-78-1 against ToLCNDV infection.

Notably, we identified three known miRNAs, sly-miR159, sly-miR6024, and sly-miR9472-3p, which were differentially expressed in the resistant and susceptible cultivars (Fig. 1C and SI Appendix, Fig. S2). Sly-miR6024 (up-regulated in susceptible cultivar) has been reported to target defense-related R gene, namely, Rx1 (provides resistance to potato virus X) (16). Therefore, enhanced expression of these miRNAs in susceptible cultivar might inhibit the expression of their target R genes, leading to susceptibility against ToLCNDV. On the other hand, the sly-miR9472-3p (up-regulated in resistant cultivar) has been previously described to target adenylate isoamyltransferase and to regulate the level of cytokinin (37), a negative regulator of defense mechanism against geminivirus infection (38). As a result, increased expression of sly-miR9472-3p may control cytokinin accumulation and provide resistance to ToLCNDV infection.

Previous studies have implicated the role of miR159-Myb regulatory module during plant development (39, 40) and plant–pathogen interaction, including oomycetes (41), fungus (42), and nematode (43). In this study, we found that sly-miR159 was the highest differentially expressed miRNA during ToLCNDV infection. It was down-regulated in resistant and up-regulated in susceptible tomato cultivar during ToLCNDV infection (Fig. 1 D and E). In previous studies, differential expression of sly159 was also observed during ToLCNDV infection (17, 18), supporting our findings. However, the extent to which miR159 regulates plant–virus interaction, especially against ToLCNDV infection, has remained unclear, which prompted us to investigate the targets and downstream components of sly-miR159 with respect to ToLCNDV in cv. H-88-78-1.

Transcription Factor Myb33 Specifically Transcribes the Sw5a Resistance Gene to Provide Geminivirus Resistance.

A nuclear-localized protein in tomato (44) and the Myb transcription factor in Arabidopsis (45, 46) have previously been identified as miR159 targets. It acts as a molecular switch to inhibit the expression of Myb33 and Myb65 in vegetative tissues, limiting plant development by preventing cell proliferation (45, 46). In this study, we demonstrated that SlMyb33 is a target of sly-miR159 (SI Appendix, Fig. S3 A and B). The expression pattern of SlMyb33 was contradictory to sly-miR159 (Fig. 1 D and E), and STTM of miR159 increased the SlMyb33 expression level, indicating that their expression levels are negatively correlated (SI Appendix, Fig. S3C). Furthermore, SlMyb33-silenced plants showed severe ToLCNDV infection symptoms such as stunted growth, leaf curling, and increased viral titer (Fig. 2 A–C). In line with these observations, silencing of SlMyb33 in susceptible plants further enhanced susceptibility toward ToLCNDV (SI Appendix, Fig. S4 G–I). Thus, supporting the notion that being a transcription factor, Myb33 represents a crucial target of miR159 and could be involved in regulating virus infection by transactivating some key downstream proteins.

In the present study, we identified Sw5a (encoding a LRR-containing protein) as a target of Myb33. The ChIP-Seq data showed ∼10-fold enrichment of Sw5a in both immunoprecipitated samples compared with input samples (SI Appendix, Fig. S7). Expression of SlSw5a was down-regulated in SlMyb33-silenced plants (Fig. 3 B and C). Furthermore, transcript abundance of SlSw5a was enhanced in the resistant cultivar and reduced in susceptible cultivar when challenged with the virus, which was in coordination with the expression pattern of SlMyb33 (Fig. 3 B and C). Remarkably, the genomic analysis and molecular assays revealed the binding of SlMyb33 onto the promoter region of SlSw5a (Fig. 3 D and E), suggesting the association of SlMyb33 with SlSw5a promoter mediates the transcriptional regulation of SlSw5a in response to virus infection. Furthermore, SlSw5a silencing resulted in susceptibility to ToLCNDV infection (Fig. 4 A–C), and transient overexpression of SlMyb33 in these silenced plants had no significant impact on disease symptoms and virus accumulation (Fig. 4 D and E). This indicates that SlSw5a is the exclusive target of SlMyb33 in the miR159-SlMyb33 module–based defense pathway against ToLCNDV infection in tomato.

Interaction between SlSw5a and ToLCNDV-AC4 Is Vital to Initiate HR during ToLCNDV Infection.

Sw5 gene cluster has five paralogs (Sw5a-Sw5e), among which only Sw5b has been shown to confer resistance against a broad range of tospoviruses (27, 28). The paralog Sw5a has been associated with HR by autoactivation, whereas its ortholog from S. lycopersicum, Sw5a^S does not initiate HR upon infection with TSWV. However, reversion of Q599R in Sw5a^S led to gain of function and triggered HR (27, 28). In our study, SlMyb33 was found to regulate the expression of SlSw5a, which has the highest similarity with Sw5a^S (SI Appendix, Fig. S11A), and ideally, it should not initiate HR. aa sequence analysis revealed that single-aa reversion (Q599R) was naturally present in SlSw5a identified in both selected tomato cultivars (SI Appendix, Fig. S11B). The role of Sw5a in HR production was depicted by the occurrence of characteristic HR lesions, reduced levels of CAT and APX activities, and increased MDA accumulation in resistant cultivar after ToLCNDV infection (Fig. 4F and SI Appendix, Fig. S13). On the contrary, a significant alteration in the activity of these enzymes in SlMyb33- and SlSw5a-silenced resistant cultivar could be correlated to reduced HR (Fig. 4F and SI Appendix, Fig. S13). Furthermore, cv. MM containing Q599 was found to be susceptible to ToLCNDV infection, and no visible HR lesions were observed (SI Appendix, Fig. S12), which demonstrated that the single mutation in Sw5a was sufficient for the gain of function. These observations provide clear evidence that SlSw5a acts as a molecular switch to activate HR response, providing the resistant phenotype to cv. H-88-78-1.

The majority of R genes detect the pathogen-specific Avr (19, 33, 34). We identified that AC4 of ToLCNDV acts as Avr determinant for Sw5a by providing multiple in vivo and in vitro evidences demonstrating strong interaction between SlSw5a and AC4 (Fig. 5 B–D and SI Appendix, Fig. S14). Moreover, transient overexpression of SlSw5a with AC4 successfully developed the HR (SI Appendix, Fig. S15) evident by DAB staining and APX, CAT, and MDA assays, suggesting this interaction is essential to regulate the enzymatic activity of ROS-scavenging enzymes and initiate HR and cell death to restrict virus spread in tomato upon ToLCNDV infection.

AC4 is a symptom determinant involved in the movement of the virus and replication of the viral genome. Recently, C4 protein from Tomato yellow leaf curl virus was shown to inhibit the cell-to-cell spread of RNA interference (RNAi), thus regulating the sRNA-mediated defense in the host plant (47). Also, the C4 protein of Tomato leaf curl Yunnan virus was found to induce the transition of cell from G1 to S phases and preventing programmed cell death (48, 49). We modeled and experimentally demonstrated that SlSw5a–AC4 interaction takes place in the membrane, and the C-terminal portion of the AC4 protein–containing “RTSK” motifs is a prerequisite for the successful interaction between them and initiation of HR (Fig. 5 E and F). Resistance breakage affects yield and might cause epidemics. Therefore, it is necessary to continuously search for active source of resistance. Our findings show that in RTSK motif, S48 is the most important residue for AC4 interaction with Sw5a (SI Appendix, Fig. S18). Considering the high conservation of S48 in AC4 protein among 161 reported ToLCNDV strains (SI Appendix, Fig. S17), it is highly probable that Sw5a could be used as resistance source to develop ToLCNDV-resistant cultivars in tomato.

Among the five Sw5 paralogs, Sw5b was found to interact with movement protein, NSm, of TSWV and provide resistance against this virus (27, 28). Transient overexpression of NSm in leaves of cv. H-88-87-1 and cv. Punjab Chhuhara did not induce HR (SI Appendix, Fig. S19), depicting that these cultivars are not resistant to TSWV infection. Furthermore, this finding is consistent with a previous study which showed that introducing Q599R mutation in Sw5a^S NB-ARC was insufficient for NSm-dependent HR (27). The chimeric Sw5a^S protein–containing Q599R reversion and LRR domain from Sw5b were found to activate NSm-dependent HR (27). The LRR domain of Sw5a identified in our study shows 98% similarity to Sw5a^S and 92% similarity to Sw5b (SI Appendix, Fig. S19A), thus justifying the absence of HR in cv. H-88-78-1 in the presence of NSm. Altogether, our data support the idea that Sw5a (Q599R) revertant provides resistance against ToLCNDV and not against TSWV.

Through this study, we provide insights into R gene–Avr interaction and downstream HR activation, leading to defense against ToLCNDV infection. AC4 protein is an essential component to initiate HR response, and coexpression of SlSw5a with AC4 influences the increased ROS levels. This may direct the lipid peroxidation process, leading to the development of clear and concomitant HR, thus restricting the spread of virus. The interaction between AC4 and Sw5a, and subsequent HR initiation, highly depend on S48 within the RTSK motif (of AC4) since mutation of this aa causes the interaction to be lost, resulting in a weakened defense response against the virus. This immune process is governed by two steps of transcriptional regulation: 1) positively by SlMyb33-mediated activation of SlSw5a, a key component of R–Avr gene interaction and 2) negatively by sly-miR159–mediated suppression of SlMyb33 (Fig. 6). We proposed a model suggesting that resistant cultivar expression of sly-miR159 is down-regulated, leading to enhanced accumulation of SlMyb33 and SlSw5a, which provides resistance against ToLCNDV by inducing HR (Fig. 6). On the other hand, in the susceptible cultivar Punjab Chhuhara, enhanced expression of sly-miR159 inhibits the expression of SlMyb33. This restricts the binding of SlMyb33 on the promoter of SlSw5a, therefore suppressing the expression of SlSw5a and HR, Thus, depicting that presence of a functional R gene does not necessarily imply downstream resistance response. Furthermore, the involvement of both positive and negative form of gene expression adds to the complexity in activation of ToLCNDV-induced HR.

Fig. 6.

Schematic representation of sly-miR159-SlMyb33 complex in providing resistance against ToLCNDV. In the resistant cultivar H-88-78-1, sly-miR159 gets down-regulated upon virus infection (3), therefore restricting the degradation of SlMyb33 mRNA and causing increased expression of SlMyb33 (1). SlMyb33 further binds onto the promoter region of SlSw5a and induces the expression (2). Moreover, SlSw5a interacts with virus protein AC4 (pathogenesis determinant) and triggers HR response in the ToLCNDV-infiltrated regions and limits the spread of virus (4). On the contrary, in the susceptible cultivar Punjab Chhuhara, up-regulation of sly-miR159 inhibits the expression of SlMyb33. This constrains the binding of SlMyb33 on the promoter of SlSw5a, which suppresses the expression of SlSw5a and HR. The most important aa S48, which is critical for interaction, is highlighted in red.

To conclude, the outcomes presented here on the crucial role of miRNA in the regulation of gene expression against virus resistance could have wide-ranging implications in tomato breeding programs. The role of other uncharacterized miRNAs identified in this study will improve our understanding of plant–virus interaction. Also, our study highlights that miR159 shows differential expression in the resistant and susceptible cultivar during viral infection. Several studies have depicted the regulation of miRNAs through epigenetic modifications as well as small regulatory peptides during development and stress adaptations (50–55). Further characterization of the triggers that lead to the perturbation of miR159 during ToLCNDV infection will enhance our knowledge with respect to its regulation. Altogether, considering SlSw5a (Q599R revertant) role against ToLCNDV and SlSw5b in resistance against broad-spectrum orthotospovirus, these two homologs could be used for gene pyramiding through breeding for achieving broad-spectrum and durable disease resistance against viruses causing yield loss in tomato. Furthermore, the pyramiding strategy might avoid the evolution of virus strains providing enhanced durable resistance.

Materials and Methods

Detailed descriptions of additional experimental methods are provided in SI Appendix, SI Materials and Methods.

Plant Materials, Agroinoculation, Sample Collection, and Marker Assays.

Two tomato cultivars differing in their resistance level to ToLCNDV were used in the present study (cv. H-88-78-1, resistant; cv. Punjab Chhuhara, susceptible). Germination, agroinoculation with infectious ToLCNDV clones, sample collection, and marker assays were performed as described in SI Appendix, SI Materials and Methods.

Identification and Expression Profiling of miRNAs.

Total RNA was isolated using TRIzol reagent (Sigma) and eight sRNA libraries were constructed using two biological replicates of tomato cultivars (cv. H-88-78-1, resistant; cv. Punjab Chhuhara, susceptible) under mock-treated [cv. H-88-78-1 (HM); cv. Punjab Chhuhara (PM]) and virus-infected [cv. H-88–78-1 (HI); cv. Punjab Chhuhara (PI]) conditions and sequenced on a HiSEq. 2500 (Illumina) sequencer. Short read processing, quality assessment, and alignment are performed as described in SI Appendix, SI Materials and Methods. Low–molecular weight RNA was extracted, transferred to nylon membrane, and probed with γ-P32-ATP–labeled antisense-miRNAs oligonucleotide (small nuclear RNA; snRNA U6 was used as loading control). Image scanning and expression profiling were performed according to the procedures described in SI Appendix, SI Materials and Methods.

RLM-RACE and Quantification of Gene Expression.

mRNA isolated from total RNA using High Pure RNA Isolation Kit (Roche) was used for RLM-RACE using GeneRacer Kit (Invitrogen), followed by sequencing as per the procedures described in SI Appendix, SI Materials and Methods. The qRT-PCR was performed to quantify the gene expression (SI Appendix, Table S8). Each experiment was carried out in three biological replicates. The amount of target genes transcript was analyzed (2^−∆∆CT method) through normalization with internal control alpha-tubulin.

STTM.

STTM vectors were constructed by the method in SI Appendix, SI Materials and Methods. The constructs transformed into Agrobacterium GV3101 were transiently transformed in both tomato cultivars.

Y1H Assay and ChIP Assay.

Matchmaker Gold Yeast One-Hybrid Library Screening System (Clontech) was used for performing Y1H assay following the procedures described in SI Appendix, SI Materials and Methods. For ChIP assay, the leaves of cv. H-88-78-1 were agroinfiltrated with 35S:SlMyb33-cMyc, and transiently expressed SlMyb33-cMyc leaves (∼2 g) were subjected to cross-linking. Downstream processing followed by DNA precipitation was carried out as described in SI Appendix, SI Materials and Methods. The DNA obtained from two biological replicates of each input control samples (IP_rep1 and IP_rep2) and anti-cMyc antibody immunoprecipitated samples (Myb33_rep1 and Myb33_rep2) were processed and sequenced on Illumina HiSeq platform as described in SI Appendix, SI Materials and Methods.

Recombinant Protein Expression, Purification, and EMSA.

GST-tagged SlMyb33 protein was expressed in Escherichia coli BL21 (DE3) strain. Fusion protein was resuspended in IBS buffer (GE biosciences) and then incubated with Glutathione Sepharose 4B beads. This recombinant protein was eluted with elution buffer, and obtained eluants were dialyzed in dialysis buffer as described in SI Appendix, SI Materials and Methods. Binding of SlMyb33 protein onto 102-bp α-P32-dCTP–labeled fragment of SlSw5a promoter was confirmed using electrophoretic mobility shift assay according to the procedure described in SI Appendix, SI Materials and Methods. Competition assay was performed with a surplus of unlabeled probe as described in SI Appendix, SI Materials and Methods.

VIGS and Infectivity Assays.

TRV-based VIGS of SlMyb33 (Solyc01g009070) and SlSw5a (Solyc09g098130) and infectivity assay was performed according to the procedure described in SI Appendix, SI Materials and Methods.

Histochemical Staining and Enzyme Assay.

Cell death was assessed by staining tomato leaves with trypan blue solution. Detection of H₂O₂ in control and silenced tomato leaves was performed by staining procedure using DAB, as described in SI Appendix, SI Materials and Methods. Leaf samples were used to determine CAT and APX activity as described in SI Appendix, SI Materials and Methods. For measurement of lipid peroxidation, MDA content was measured by the reaction with Thiobarbituric acid, according to the protocol described in SI Appendix, SI Materials and Methods.

Y2H Assays.

Matchmaker Gold Yeast Two-Hybrid System (Clontech) was used for performing Y2H assays following the procedures described in SI Appendix, SI Materials and Methods.

GST-Pulldown Assays.

The fusion proteins GST-AC4 and His-SlSw5a were expressed in BL21 (DE3) E. coli cells. GST or GST-AC4–bound glutathione beads were incubated with the prey proteins His-SlSw5a for 2 to 4 h at 4 °C followed by washing and elution as described in SI Appendix, SI Materials and Methods. Anti-GST antibody was used for detecting the bait protein as the input control.

Localization, BiFC Assays, and Co-IP Assays.

For these assays, SlSw5a was fused with N-terminal region of YFP (pSYPNE-35S: Sw5-YFP^N) and cMyc as affinity tag whereas AC4 with C-terminal region of YFP (pSPYCE-35S; AC4-YFP^C) and HA tag. AC4 and NSm were cloned in pGWB6 for localization assay. BiFC and localization assays were performed using laser confocal scanning microscope (Leica Microsystems) as described in SI Appendix, SI Materials and Methods. For Co-IP assay, N. benthamiana leaves were coinfiltrated with SlSw5a-cMyc and AC4-HA constructs, and immunoprecipitation was performed as described in SI Appendix, SI Materials and Methods. The interaction between SlSw5a and AC4 was analyzed by immunoblots using anti-cMyc antibody (1:2,000 dilution).

SDM and Deletion Constructs.

The PCR for SDM was performed and amplified products were digested with restriction enzyme Dpn I and transformed into E. coli as described in SI Appendix, SI Materials and Methods. For deletion constructs of AC4, N- (1 to 78 nucleotides) and C-Terminal (79 to 174 nucleotides) regions of AC4 were generated as described in SI Appendix, SI Materials and Methods. The mutation was validated in all the constructs through DNA sequencing analysis.

Statistical Analysis.

The data obtained is a mean value of independent experiments. The error bars shown in the graph are mean values of SE (± SE). The significance of variation between mean values of mock and ToLCNDV-treated samples (*P < 0.05; **P < 0.01; ***P < 0.001) was calculated using the unpaired Student’s t test via Graphpad InStat software (https://graphpad.com:443/quickcalcs/ttest1.cfm).

Data Availability

The sRNA sequencing data and ChIP-seq libraries have been deposited in the National Center for Biotechnology Information Sequence Read Archive (NCBI SRA) database under BioProject PRJNA644048 (56) and BioProject PRJNA642674 (57), respectively.