A ToLCNDV isolated from Pakistan, belonging to ToLCNDV-India strain group (here designated as ToLCNDV-In) and the Italian isolate 45/16, belonging to subgroup I of ToLCNDV-ES strain group (here designated as ToLCNDV-ES), were used in the present study (Parrella et al., 2018; Vo et al., 2022a).

Infectious clones were successfully generated in a previous study (Vo et al., 2022a). Nicotiana benthamiana and three different tomato cultivars including ‘San Pedro’ from Italy, ‘Seogwang’ from Korea, and ‘Moneymaker’ were cultivated in a growth chamber with 16 h light/8 h dark cycles at 22–28 °C

Subgenomic components were swapped to form two combinations, ToLCNDV-In DNA-A + ToLCNDV-ES DNA-B and ToLCNDV-ES DNA-A + ToLCNDV-In DNA-B (hereafter designate as ToLCNDV-In-A, ToLCNDV-In-B, ToLCNDV-ES-A and ToLCNDV-ES-B respectively) using an approach similar to other agroinoculations (Vo et al., 2022a). Seven chimeric sequences corresponding to six ORF and the IR of DNA-A components were synthesized by swapping the ToLCNDV-ES sequence on the ToLCNDV-In backbone (Macrogen, Korea) based on the wild-type sequence of the two ToLCNDV isolates with tandem repeat constructs (Figure S1). Plasmids containing swapped sequences were digested with HindIII/SpeI and ligated into the pCAMBIA 1303 vector to generate recombinant plasmids. After transformation of Agrobacterium tumefaciens strain GV3101 using the freeze-thaw transformation method (Weigel and Glazebrook, 2006), the infectivity of all the chimeric clones was tested on N.benthamiana and tomato plants.

Site-directed mutants of CP were produced using Q5^® Site-Directed Mutagenesis Kit (New England Biolabs, Ipswich, MA, USA) following the manufacturer’s instructions. Based on the different amino acids in the V1 sequence between the two isolates using multiple alignments, 16-point mutant clones were generated using the designed primer sets (Table S1) with extracted viral DNA from infected leaves as a template. All mutant sequences were confirmed by sequencing, which were then transformed into Agrobacterium to generate infectious clones using the same methodology.

A.tumefaciens harboring each viral clone was grown in 20 mL of LB broth medium containing kanamycin (50 µg/mL), rifampicin (50 µg/mL), and gentamycin (50 µg/mL) for 24 h at 28°C with shaking until the optical density (OD) reached 1.0 at 600 nm. Activated cells were pelleted by centrifuging and resuspended in infiltration buffer (10 mM MgCl₂, 10 mM MES, 200 µM acetosyringone). A. tumefaciens strains containing the recombinant plasmids ToLCNDV-In-A/ToLCNDV-In-B, ToLCNDV-ES-A/ToLCNDV-ES-B, ToLCNDV-In-A/ToLCNDV-ES-B, and ToLCNDV-In-B/ToLCNDV-ES-A were mixed in a 1:1 ratio, and the strains containing chimeric and site-directed mutant clones were mixed in equal proportions. Cell suspensions were used for the infectivity assay of three-week-old tomato plants by pinpricking the main apical shoot.

Inoculated leaf samples were collected 21 days post-inoculation (dpi) to confirm the viral infection. Genomic DNA was extracted and used for PCR amplification with specific primer sets, as described previously (Vo et al., 2022a). To confirm viral replication, Southern blotting was performed as previously described (Southern, 2006). Briefly, after electrophoresis, DNA was transferred to nylon membranes and hybridized with a probe containing [α³²P]-dCTP at 65°C for 16 h. The membrane was washed and exposed to an X-ray film.

Quantitative PCR (qPCR) was performed on the agroinoculated plants at 21 dpi to determine the viral titer. Total DNA from different tissues, including new leaves, stems, and roots of the three plants, was extracted and used as a template for qPCR following our previous study (Vo et al., 2022b).

To identify the interactions between the CPs of each isolate and the host factors, yeast two-hybrid screening of the tomato cDNA library was conducted. For cDNA library construction, total RNA from tomato leaves was isolated using the RNeasy Plant Mini Kit (Qiagen, Hilden, Germany) according to the manufacturer’s instructions. The integrity of the total RNA was analyzed using 1% agarose gel electrophoresis and the concentration was determined using a BioTek Epoch Microplate spectrophotometer (BioTek, Winoosku, VT, USA). Then, this RNA was used to construct the Y2H tomato cDNA library by Panbionet (Pohang, Korea). The CP (V1) encoding gene of each isolate was introduced into the pGBKT7 (referred to as In-V1 and ES-V1) vector, which contained a GAL4 DNA-binding domain, kanamycin resistance for selection in Escherichia coli, and a TRP1 nutritional marker for selection in yeast. Two plasmids, pGBKT7-In-V1 and pGBKT7-ES-V1, were generated as baits in the GAL4-based two-hybrid system for CPs of ToLCNDV-ES and ToLCNDV-In isolates, respectively. Bait vectors were checked for toxicity and autoactivation in yeast before processing the hybrids. The bait was then hybridized with a cDNA library to identify CP-interaction proteins in tomatoes by Panbionet.

To confirm the interaction between each ToLCNDV CP and the host protein, yeast two-hybrid assays were conducted using a modified lithium acetate yeast transformation (Thompson et al., 1998). Interaction of different host genes including Cathepsin B-like protease 3 (XM_004233173), NAD(P)H-quinone oxidoreductase subunit M (XM_004237313), Ribulose bisphosphate carboxylase small chain 3B (NM_001309210), protein phosphatase 2C (NM_001247571), LOW PSII ACCUMULATION (XM_004247389), NAD(P)H-quinone oxidoreductase subunit M (XM_004237313), and RING finger protein 44- like (XM_004243217) was examined with In-V1 and ES-V1. Prey, including host genes, were cloned into the pGADT7 vector and co-transformed with bait vector pGBKT7 carrying different CPs into Saccharomyces cerevisiae strain Y2H Gold. The transformed yeast cells were plated on three nutrient dropout media: SD/-Trp/-Leu (DDO), SD/-Trp/-Leu/-His (TDO), and SD/-Trp/-Leu/-Ade/-His (QDO). The plasmids pGBKT7-53 (Gal4 DNA-BD fused with murine 53) and pGBKT7-Lam (Gal4 BD fused with lamin) were co-transformed with pGADT7-T encoding the Gal4 AD fused with the SV40 large T-antigen and used as positive and negative controls, respectively.

For RT-qPCR, cDNA was synthesized from 1µg of total RNA using Oligo dT primers and Moloney murine leukemia virus (MMLV) reverse transcriptase (Bioneer, Daejeon, Korea). The reaction was performed using TB Green® Premix Ex Taq™ II (Tli RNaseH Plus; TaKaRa Bio, Shiga, Japan) with the primer sets (Table S2). Forty cycles of PCR were performed using a Rotor-Gene Q thermocycler (Qiagen) under the following conditions: 10 s denaturation at 95°C, 15 s annealing at 60°C, and 20 s polymerization at 72°C. Elongation 1α (EF1α) gene was used for internal normalization and each reaction was replicated three times. Data analyses were conducted using the 2^−ΔΔCt method (Livak and Schmittgen, 2001). Statistical analyses were performed using the t-test in the GraphPad Prism software (GraphPad Software, Boston, MA, USA).

Infectivity assays revealed that the ToLCNDV-In infectious clone could infect tomatoes, whereas the ToLCNDV-ES clone was not adapted to this crop (Figure 1). Typical symptoms, including leaf curl, mosaic, and stunting, were induced in ToLCNDV-In-infected tomatoes, whereas a normal phenotype was observed in ToLCNDV-ES-inoculated plants for all the three tomato cultivars (Moneymaker, Seogwang, and San Pedro) (Figure 1A). The PCR detection results were consistent with the appearance of symptoms in the inoculated tomatoes (Figure 1B).

To determine the genomic region responsible for ToLCNDV pathogenicity in tomato, the subgenomes of the two isolates were swapped by mixing DNA components as follows: ToLCNDV-ES-A + ToLCNDV-In-B (A_ESB_In) and ToLCNDV-In-A + ToLCNDV-ES-B (A_InB_ES) and processed, and original ToLCNDV-ES-A/B (A_ESB_ES) and ToLCNDV-In-A/B (A_InB_In) were used as positive and negative control, respectively. The results revealed that only ToLCNDV-In-A supplemented with either ToLCNDV-In-B or ToLCNDV-ES-B could infect tomatoes (Table 1). Infected plants exhibited a typical disease phenotype at 21 dpi (Figure 1A). Furthermore, the viral genome was detected in all symptomatic plants inoculated with A_InB_In and A_InB_ES using PCR amplification with specific primers. No infection was observed with A_ESB_ES or A_ESB_In (Figure 1B). Virulence was only observed when tomatoes were co-inoculated with ToLCNDV-In-A, demonstrating that the DNA-A component of ToLCNDV-In was responsible for infectivity in tomatoes.

To identify the gene(s) that determine tomato infection by ToLCNDV-In, we synthesized seven chimeric constructs by substituting genomic regions containing C1 (Rep), C2 (TrAP), C3 (REn), C4, V1 (CP), V2, and IR of ToLCNDV-In into the ToLCNDV-ES backbone (abbreviated as In_IRES, In_C1ES, In_C2ES, In_C3ES, In_C4ES, In_V1ES, and In_V2ES). After three weeks of inoculation, chimeric construct In_V1ES induced moderate leaf curl symptoms, consistent with the PCR amplification results, whereas visible symptoms and viral DNA were not detected in all other constructs in tomatoes (Figure 2A; Table 1). A chimeric clone harboring ToLCNDV-ES V1 in the ToLCNDV-In backbone (ES_V1In) was constructed to confirm the role of CP in ToLCNDV infectivity. There was no virus accumulation in symptomless ES_V1In-inoculated plants, as confirmed by PCR (Figure 2B). According to the Southern hybridization results, open circular and super-coiled forms (double–stranded DNA), as well as a single-stranded form of both DNA components indicated that the chimeric In_V1ES clone can replicate in tomato (Figure 2C). Taken together, these results suggest that the CP is responsible for the infectivity in tomatoes of the ToLCNDV-In and of its chimeric construct In_V1ES.

To determine the role of different amino acids in the CPs in pathogenicity in tomato, eight site-directed mutant constructs were generated. These included ES-V1^S2A (serine residue at amino acid 2 was changed to alanine), ES-V1^T29A (threonine residue at amino acid 29 was changed to alanine), ES-V1^S32V (serine residue was changed to valine at amino acid 32), ES-V1^R42K (arginine residue was changed to lysine at amino acid 42), ES-V1^W143R (residue tryptophan at amino acid 143 was changed to arginine), ES-V1^T148S, (serine replaced threonine at amino acid 148), ES-V1^S193C (cysteine residue replaced serine at amino acid 193), ES-V1^R194K (arginine residue was changed to lysine at amino acid 194) based on alignment of V1 full length of ToLCNDV-ES. At contrary, mutants In-V1^A2S, In-V1^A29T, In-V1^V32S, In-V1^K42R, In-V1^R143W, In-V1^S148T, In-V1^C193S, and In-V1^K194Rwere generated based on ToLCNDV-In V1 sequence and their infectious clones were constructed (Figure S2). All mutant infectious clones were agroinoculated into tomato plants to confirm their infectivity with the wild-type ToLCNDV clones. A symptomless phenotype was observed in all plants inoculated with either mutant or wild-type ToLCNDV-ES clone after three weeks of inoculation (Figure 3A; Table S3). Nevertheless, viral DNA was only detected in ES-V1^W143R-inoculated plants. Three different tissues, upper leaves, stems, and roots of asymptomatic infected plants, were examined for viral presence to determine the movement of the ToLCNDV mutant clone (Figure 3B). PCR results showed that this mutant clone was present in all tested tissues, indicating that ToLCNDV-ES with a single mutation at amino acid 143 position of the CP could replicate and spread systemic infection to other tissues (Figure 3C). In addition, qPCR was performed to quantify the relative viral accumulation of the mutant clone in various tomato tissues after infection. No viral DNA accumulated in the mock-treated plants, whereas there was a high level of mutant ES-V1^W143R in all infected tissues, especially in the roots (Figure 3D). In addition, all mutant clones of different amino acids, including amino acid 143 on the ToLCNDV-In CP, showed infectivity and induced typical symptoms in tomatoes. The detectable PCR results (data not shown) and symptomatic phenotype of inoculated plants indicated that the mutation at the amino acid 143 position did not have a major effect on ToLCNDV-In pathogenicity but was essential for ToLCNDV-ES infectivity in tomato, even when no disease symptoms were induced.

In previous reports, some ToLCNDVs isolated from Spain also showed weak adaptation to asymptomatic tomatoes, even though a few plants showed slight vein yellowing only in the lower leaves (Fortes et al., 2016; Yamamoto et al., 2021). Comparison of amino acid sequences among CPs of Spanish ToLCNDV and the two isolates in this study revealed that both isolates from Mediterranean only differ in amino acid 143, suggesting that this amino acid plays an important role in ToLCNDV-ES’s ability to infect tomatoes.

Yeast two-hybrid screening was conducted to determine the interaction between the CP of each ToLCNDV isolate and host proteins. In the screening, only 26 of 53 and 10 of 25 candidates truly interacted with ToLCNDV-In CP and ToLCNDV-ES CP, respectively. After sequencing, five host proteins including Cathepsin B-like protease 3, NAD(P)H-quinone oxidoreductase subunit M,Ribulose bisphosphate carboxylase small chain 3B (RBCS4), Protein phosphatase 2C (DIG3), and Protein LOW PSII ACCUMULATION 1 interacted with ToLCNDV-In CP, whereas CP of ToLCNDV-ES interacted with NAD(P)H-quinone oxidoreductase subunit M protein, RING finger protein 44- like, GAGA- binding transcriptional activator, protein CROWDED NUCLEI 4, 1-D-deoxyxylulose 5-phosphate synthase, and thaumatin-like protein 1b (Table 2). Both baits interacted with the same NAD(P)H-quinone oxidoreductase subunit M protein, a flavoenzyme that serves as a quinone reductase, in connection with the conjugation reactions of hydroquinones involved in detoxification pathways (Sellés Vidal et al., 2018).

To verify the interaction between the two CPs and host genes derived from the screening list, baits containing ToLCNDV-In CP and ToLCNDV-ES CP were co-transfected with the prey vector harboring the host genes in the S. cerevisiae strain Y2H Gold. Transfected yeasts were grown under different nutrient deficiencies (DDO, TDO, and QDO) with serial dilutions to confirm the interactions. After three days of incubation, the results showed that all captured host proteins interacted with ToLCNDV-In CP and ToLCNDV-ES, except for Ring finger protein 44-like. Interestingly, Ring-finger protein 44- like interacted only with ToLCNDV-ES CP and did not show any binding with ToLCNDV-In CP bait (Figure 4A). The yeast co-transfected with ES CP bait and ring finger prey grew on three nutrient-deficient nutrient media, while yeast co-transfected with ToLCNDV-In CP did not produce any colonies on DDO and QDO media, indicating that there was no interaction between ToLCNDV-In CP and this host protein.

We compared the Ring-finger protein 44-like transcript levels in tomato leaves inoculated with ToLCNDV-In, ToLCNDV-ES and A. tumefaciens GV3101 (mock) using RT-qPCR. The relative expression of this protein in ToLCNDV-inoculated leaves was significantly higher than that in GV3101-infiltrated leaves at 21 dpi as shown in Figure 4B, suggesting that the expression of Ring-finger protein 44-like was induced by virus inoculation. However, no significant difference in Ring protein expression between the two ToLCNDV clone-inoculated plants suggests that the infectivity of ToLCNDV in tomato is not related only to the accumulation of Ring-finger protein 44-like. We also assessed the expression of five genes in the upstream and downstream regions under these two conditions (Figure 5; Table S4). RT-qPCR showed increase in expression of cytochrome P450 CYP72A219-like in the upstream region and ATP-dependent zinc metalloprotease FTSH2 in the downstream region with viral infection compared to their expression in mock plants. Expression of upstream genes increased the most in the ToLCNDV-ES-inoculated group. This gene is involved in growth and defense mechanisms by inducing the biosynthesis of defense compounds such as jasmonic acid and flavonoids. These factors also potentially affect ToLCNDV pathogenicity in tomatoes.