SyYVCV βC1, a geminiviral pathogenicity determinant, undergoes SUMOylation in host plants
SyYVCV is a new monopartite Begomovirus recently characterized by our group, having a 2.7-kb DNA A component and a satellite DNA β of 1.3 kb length [33]. It causes vein clearing disease in its natural host and leaf curling in Nicotiana tabacum. SyYVCV βC1 is 118 aa long protein with multiple intrinsically disordered regions coded by the only ORF known in DNA β. Using bioinformatics tools (GPS SUMO, JASSA) [26, 34], we identified three putative SUMOylation sites spread throughout the length of the protein (Fig. 1a). Among the three predicted SUMOylation sites (Ss), Ss1 (Lysine, K18) and Ss2 (Lysine, K24) showed inverted SUMOylation consensus (D/EXKψ), whereas Ss3 (K83) was predicted to have a consensus site for SUMOylation (ψKXE/D) with a low score (Additional file 1: Fig. S1A). Both inverted and consensus sites usually get SUMOylated [35].
Among the three predicted sites, Ss1 (K18) was the most conserved SUMOylation site (77% of all βC1 entries from the nr Uniprot database) when compared between viruses that are associated with β DNA. This was followed by Ss3 (K83; 37%). Ss2 (K24) was least conserved among all predicted sites (10%), mostly restricted to SyYVCV and AYVV (Ageratum Yellow Vein Virus) cluster of viruses that result in vein clearing symptoms (Fig. 1a, Additional file 1: Fig. S1B and Additional file 2: Table S1).
To validate whether SyYVCV βC1 is a direct target for SUMO conjugation, an in vitro SUMOylation assay containing recombinant E. coli-purified MBP-βC1, SUMO-activating enzyme mixture (E1 homolog), Ubc9 (E2 homolog) and His-tagged SUMO1 (NbSUMO1-GG, C-terminal activated SUMO1 N. benthamiana) was performed. We observed multiple slow-migrating high molecular weight intermediate products in a denaturing PAGE gel when blotted with anti-AtSUMO1 antibody only in the presence of βC1 as substrate. These bands were detected only in the presence of Mg2+-ATP, and likely represented a SUMOylated βC1 (Fig. 1b). The SUMOylated products were a result of the direct reaction involving E1 and E2, since the absence of E1 and E2 in the reaction did not produce higher-order bands (Additional file 1: Fig. S2A). These results likely suggested the SUMOylation of βC1 in vitro in an E1, E2-catalysed ATP-dependent reaction. To further validate this interaction, we performed yeast two-hybrid (Y2H) assay with NbSUMO1 as prey and βC1 as bait. We observed the growth of yeast cells, suggesting an interaction between NbSUMO1 and βC1 (Additional file 1: Fig. S2B).
To understand the biological function of βC1 SUMOylation, we generated transgenic N. tabacum plants overexpressing βC1. As expected and observed in case of overexpression of many pathogenicity determinants, overexpression of βC1 induced symptoms similar to virus-infected plants [36] (Fig. 1c, d, e, and Additional file 1: Fig. S3). These plants exhibited abnormal phenotypes such as stunted growth, early flowering, pointed leaves, shorter internodes, branching and mosaic patches on the leaves. In addition to these phenotypes, the expression of βC1 induced exerted stigma phenotype and seed sterility. Although the expression of a DNA viral protein in plants was not previously reported to show exerted stigma phenotype, RNA viral proteins are known to produce such defects [37].
To substantiate the defects observed in transgenic plants overexpressing βC1 was because of the interactions mediated by βC1 SIM and SUMOylation motifs, we verified whether βC1 undergoes SUMOylation in planta, we performed co-immunoprecipitation (Co-IP) of βC1 from stable transgenic plants overexpressing βC1. βC1, but not control GFP expressing plants, showed high molecular weight bands when blotted with anti-GFP (Additional file 1: Fig. S4A) and anti-AtSUMO1 antibodies (Additional file 1: Fig. S4B). Since SUMOylation is a dynamic process and less than 1% of any substrate protein is SUMOylated at a given point in cells [21], we transiently overexpressed 3X Flag-tagged NbSUMO1 and GFP-tagged βC1 in N. tabacum to increase the chance of βC1 SUMOylation and its detection through WB (Fig. 1f). Co-IP of βC1 was performed, followed by western blot analysis with anti-FLAG antibody. Any signal from the pull-down products after blotting with anti-FLAG antibody essentially indicates an interaction of NbSUMO1 with βC1 (Fig. 1g). After Co-IP with anti-GFP and detection with anti-FLAG antibody, we observed signals ranging from 60 to 150 kDa only in βC1 pull-down products, but not in control, indicating the presence of SUMO-conjugated βC1 (Fig. 1g). SUMOylation of βC1 produced higher-order intermediates likely due to SUMOylation of multiple SUMO conjugation sites of βC1 as well as poly-SUMOylation of NbSUMO1 conjugated to βC1 [38].
We further validated these results by swapping tags and using a non-conjugable form of NbSUMO1 (NbSUMO1ΔGG) in the abovementioned assays. During the SUMOylation process, SUMO proteins undergo proteolysis at their C-terminus to expose their di-Glycine motifs. As a result, when we transiently overexpressed NbSUMO1ΔGG in plants, higher-order conjugation products were absent at the global level suggesting inefficiency of NbSUMO1ΔGG to undergo SUMOylation (Additional file 1: Fig. S4C). We used GFP-tagged NbSUMO1 and NbSUMO1ΔGG to validate that the higher-order bands observed are actual conjugated products of NbSUMO1 derived from the SUMOylation cascade. We used MBP-tagged βC1 as a substrate for detecting SUMOylation along with GFP-tagged NbSUMO1in conjugable and non-conjugable forms. After Co-IP with anti-MBP followed by blotting with anti-GFP and anti-MBP, we observed high molecular weight bands in βC1 when co-expressed with a conjugable form of NbSUMO1, but not with non-conjugable form (Fig. 1h).
NbSUMO1 is the only identified SUMO protein in Nicotiana sp., whereas the model plant Arabidopsis has 4 characterized SUMO proteins [39]. To further explore the possibility of interaction between βC1 and other SUMO proteins, we used Y2H assay. As SUMO1 and SUMO2 have highly redundant biochemical functions in Arabidopsis, we used only SUMO1, SUMO3 and SUMO5 of Arabidopsis as prey proteins fused to activation domain (AD) in a yeast two-hybrid screen with βC1 fused to the binding domain (BD). We observed a strong interaction of βC1 with AtSUMO3 and AtSUMO5 (Additional file 1: Fig. S5A). AtSUMO5 caused auto-activation when fused to AD domain alone; however, the strength of interaction with βC1 in the quadruple (-LWHA) knockout media was clearly observed. Auto-activation caused by AtSUMO5 was minimal in -LWHA, but in the presence of βC1, the growth of cells was enhanced suggesting interaction. In the case of AtSUMO3, a strong interaction was observed only with βC1, indicating βC1 might also interact with AtSUMO3. To distinguish between SUMOylation and SIM-mediated interactions, we used di-Glycine deleted AtSUMO3 and AtSUMO5. Interestingly, the deletion of the di-Glycine motif caused no difference in the interaction of βC1 with AtSUMO5, but completely abolished its interaction with AtSUMO3. These results suggest that other SUMO proteins might also interact with βC1 via SUMOylation or SIM-mediated interactions. In these assays, protein expression and stability of all proteins were verified. All other proteins except AtSUMO5 were expressing at almost equal levels in yeast cells, while AtSUMO5 was expressed at unusually high levels, and might be the reason for its auto-activation (Additional file 1: Fig. S5B) [40]. To further verify the biological significance of the observed interaction between βC1 and other SUMO proteins, we overexpressed GFP-tagged versions of both AtSUMO3 and AtSUMO5 along with βC1 and performed an IP with βC1 as previously described (Fig. 1f). We observed a very weak pull-down signal of AtSUMO3 and AtSUMO5 as compared to NbSUMO1 (Additional file 1: Fig. S5C). These experiments suggest that even though βC1 is able to interact with other SUMO proteins in yeast, in Nicotiana sp., βC1 majorly interacted with NbSUMO1. Together, these results strongly indicate that βC1 undergoes SUMOylation in plants and that it interacts with host SUMO proteins.
SUMOylation sites are essential for the stability of βC1 in host plants
Since there are three predicted SUMOylation sites in βC1 (Fig. 1a), we explored which among these predicted sites are necessary and sufficient for SUMOylation. We substituted lysine residues to arginine, which will abolish SUMO modification of the consensus sequence without leading to much structural disruption. Since Ss1 (K18) and Ss2 (K24) residues of βC1 are close to each other, we designed a double mutant K18, 24R (henceforth mK18, 24R) to cover both these sites, a single Ss3 (K83) mutant (mK83R) and a null mutant with all three predicted lysines mutated to arginines, i.e. K18R, K24R and K83R (mK18,24,83R) (Fig. 2a). We recombinantly expressed and purified the abovementioned mutants of βC1 from E. coli and performed an in vitro SUMOylation assay with NbSUMO1. We observed that all predicted lysines (K18, 24 and 83) underwent NbSUMO1 conjugation and mutating these sites to arginine in double mutant or in triple null mutant abolished NbSUMO1 modification in vitro (Fig. 2b). To further confirm the above observations, we also performed an in vitro SUMOylation assay with NbSUMO1 using short peptides covering the βC1 SUMOylation consensus lysine sites. Mutating SUMOylation sites inhibited NbSUMO1 conjugation in vitro (Additional file 1: Fig. S6A). These results indicate that there is a propensity of all 3 predicted sites of βC1 to undergo SUMOylation in vitro.
To further understand the importance of these SUMOylation sites of βC1, we generated transgenic plants overexpressing SUMOylation site mutants of βC1. Along with mK18, 24R, a Ss3 mutant was generated where the consensus lysine K83 was kept intact, while SUMOylation consensus sequence was removed (mS SMs3, Fig. 2a). Interestingly, transgenic plants overexpressing mK18, 24R double mutant was devoid of abnormal phenotypes and was similar to GFP overexpressing control plants (Fig. 2c), indicating that SUMOylation in K18, K24 residues of SyYVCV βC1 is essential for the development of symptoms in plants. mS SMs3 mutant overexpressing transgenic plants also showed recovery from severe phenotype to some extent (Fig. 2d and Additional file 1: Fig. S6B, C and D).
In order to understand why transgenic plants expressing mK18, 24R double mutant were asymptomatic, we performed detailed molecular analysis. Our immunoblot analysis of βC1 and SUMOylation-deficient mutant plants revealed a reduced protein accumulation of mK18, 24R double mutant as compared to WT βC1 or other mutants (Fig. 2e). SUMOylation-deficient double mutant mK18, 24R protein accumulated only 10% of WT βC1 (Fig. 3a and Additional file 1: Fig. S7A). mS SMs3 mutant protein levels were comparable to that of WT βC1(Fig. 2e). The transgenic plants expressing βC1 mutant showed disproportionate severity of symptoms, likely indicating a hierarchy in these sites to undergo PTMs and subsequent functions. Reduction in mK18, 24R protein level was not due to reduced transcription as seen in RT-PCR analysis (Fig. 2e). Further, to verify the observations of transgenic plants, we transiently overexpressed WT βC1, mK18, 24R and mK83R mutants in N. tabacum and performed an IP for βC1. Similar to transgenic plants, mK18, 24R mutant levels were significantly reduced in our IP analysis performed from transient overexpression of these proteins (mK18, 24R and mK83R) (Fig. 2f). To understand the SUMOylation status of these three lysines of βC1, we performed a Co-IP assay by co-expressing MBP-tagged βC1 or mK18, 24, 83 R along with GFP-NbSUMO1. We observed a significant decrease in SUMOylation of mK18, 24, 83R triple mutant, indicating that these sites are the major sites of NbSUMO1 conjugation (Fig. 2g).
N-terminal SUMOylation-deficient double mutant of βC1 is prone to enhanced degradation in plants
It was shown previously that geminiviral βC1 from diverse viruses undergo degradation in host cells. TYLCV βC1 undergoes ubiquitin-mediated proteasomal degradation, while CLCuMuV βC1 directly interacts with ATG8 which leads to an autophagy-mediated degradation [13, 14]. To check if mK18, 24R mutant of βC1 undergoes active degradation in the host, a time course protein analysis of βC1 and mK18, 24R mutant was performed (Additional file 1: Fig. S7B). SyYVCV βC1 protein levels were slightly reduced after 1 day post infiltration (DPI), whereas mK18, 24R accumulated only to half the level of βC1 at 1 DPI and was barely detectable at 2 DPI. To understand the cause of reduced accumulation associated with SUMOylation-deficient mK18, 24R mutant, we employed degradation pathway inhibitors to check if mK18, 24R SUMOylation-deficient mutant undergoes enhanced degradation in vivo. Upon treatment with MG132 (carbobenzoxy-l-leucyl-l-leucyl-l-leucinal), a potent reversible inhibitor of proteasomal activity, protein levels of βC1 increased substantially (Fig. 3b), indicating that it undergoes active proteasomal degradation in plants. Interestingly, we also observed a drastic increase in the levels of mK18, 24R mutant whose protein level stabilized more than WT βC1 upon MG132 treatment (Fig. 3b and c quantification). We further used a wide spectrum protease inhibitor N-ethylmalemide (NEM) that inhibits cysteine proteases and partially inhibits proteasome [41]. Upon treatment with NEM, mK18, 24R mutant protein levels increased 3- to 5-fold (Fig. 3d and e quantification), suggesting enhanced degradation of mK18,24R βC1 mutant in plants. These assays suggest that the N-terminal SUMOylation-deficient mutant of βC1 undergoes enhanced degradation mediated by the host protein degradation pathway.
As observed in protein expression analysis from transgenic plants expressing SUMOylation mutants of βC1, there appears to be a disparity between three SUMOylation motifs of βC1. To further understand the functional significance of these N- and C-terminal localized SUMOylation motifs of βC1, we performed transient overexpression assays for the abundance of mK18, 24R, mK83R and mK18,24, 83 R βC1 mutants. While the mK18, 24R mutant was unstable in transient assay as well as in transgenic plants, a triple SUMOylation site mutant (mK18, 24, 83R) was surprisingly stable (Fig. 3f). To pinpoint the exact SUMOylation motif involved in the stability of βC1, we generated individual K to R mutants of N-terminal SUMOylation motifs K18 and K24. Interestingly, none of the single mutants was unstable (Additional file 1: Fig. S7C). We speculated that removing all three SUMOylation sites either altered recognition of the protein or other steps in proteasome-mediated degradation, leading to the stability of the protein as may be the case of mK18, 24, 83R. To further confirm that the enhanced degradation of SUMOylation-deficient mK18, 24R mutant in plants is via the plant degradation pathway and not due to intrinsic instability of mutants, we expressed βC1, mK18, 24R and mK18, 24, 83R in a yeast WT strain (BY4741). All the mutants were as stable to levels comparable to WT βC1 protein (Additional file 1: Fig. S7D). Taken together, all these results confirm that SyYVCV βC1 undergoes rapid degradation in plants similar to other viral βC1 proteins observed previously. These results also indicate that due to loss of protective marks, mK18, K24R degradation is enhanced suggesting the significance of N-terminal SUMOylation motifs in the stability of βC1.
SUMO-interacting motifs (SIMs) of SyYVCV βC1 interact with NbSUMO1
In multiple proteins that undergo SUMOylation, a complementary stretch of SIM was routinely observed within the candidate protein. In SyYVCV βC1, along with three SUMOylation motifs, four SIMs were predicted using SIM prediction softwares JASSA and GPS SUMO. The N-terminal SIM (residue 14–17, SIM1) overlaps with the K18 SUMOylation motif consensus sequence. The second and the third predicted SIM sequences are in an overlapping stretch forming SIM2, 3 (residue 90–93 and 91–94) (Additional file 1: Fig. S8A). The last SIM (SIM4) (residue 101–104) is towards the extreme C-terminal end. It is important to note that SIM2, 3 and SIM4 are in close proximity to the third SUMOylation motif consensus lysine K83.
Plant SUMO proteins are diverse and form a distinct clade even though the SUMO proteins are highly conserved from yeast to mammals (Additional file 1: Fig. S8B). In Arabidopsis, eight SUMO coding genes are known, out of which four SUMO proteins (AtSUMO1, 2, 3 and 5) are known to express and being observed to be functionally active. As reported earlier, we observed differential tissue-specific expression of Arabidopsis SUMO proteins (Additional file 1: Fig. S9A). To identify the SUMO proteins potentially interacting with βC1 SIMs, we used NMR titration experiments. We recombinantly expressed and purified 15N-labelled SUMO 1, 2, 3 and 5 from Arabidopsis, and SUMO1 from N. benthamiana (Additional file 1: Fig. S9B). Interestingly, AtSUMO3 and 5 were insoluble in E. coli (Additional file 1: Fig. S9C) and, after refolding, exist as soluble higher-order multimers (Additional file 1: Fig. S9G and S9H) whereas NbSUMO1 and AtSUMO1 exist as monomers (Additional file 1: Fig. S9D, E and F). The physiological significance of this multimerization property of plant SUMO proteins is unknown; however, many studies have observed AtSUMO3 and AtSUMO5 localized as nuclear speckles [30]. NbSUMO1 and AtSUMO1 are structurally identical with a pairwise sequence identity of 97%. We used NbSUMO1 for screening multiple SIMs of SyYVCV βC1 through 15N-edited Heteronuclear Single Quantum Coherence (HSQC) experiments using SIM peptides (Additional file 1: Fig. S10A). Upon titration with the SIMs derived from βC1, NbSUMO1 showed interaction with SIM2, 3 and SIM4 (Fig. 4b, c, left panel). However, SIM1 did not show any interaction with NbSUMO1 (Fig. 4a, left panel). Based on the NMR analysis, a structural model of NbSUMO1 indicating the residues involved in interaction with βC1 SIMs were predicted (Fig. 4d). The chemical shifts of the backbone 1HN, 15N, 13Cα, 13Cβ and 13CO resonances of the NbSUMO1 were assigned by standard triple resonance NMR experiments (see the “Methods” section) (Additional file 1: Fig. S11A). The chemical shifts were used in a modelling software CS-ROSETTA [42] to obtain a structural model of NbSUMO1 (Fig. 4d, right panel). The chemical shift perturbations (CSPs) of SIMs were mapped on the NbSUMO1 structure to highlight the SUMO: SIM interface (Fig. 4d, middle panel). To understand the structural interaction and binding pocket involved in NbSUMO1: βC1 SIM interaction, we modelled them together based on the CSP data and human SUMO1/IE2-SIM structure (PDB id: 6K5T) in UCSF-Chimera (Fig. 4d, left panel). IE2 is a human cytomegalovirus protein [43]. The CSPs identified residues 30 to 50 as the region of NbSUMO1 binding βC1 SIM 2, 3 motif (Fig. 4b, right panel) and SIM 4motif (Fig. 4c, right panel and Additional file 1: Fig. S11B). To validate these interactions, we further used SIM mutants in our HSQC experiments. As expected, SIM2, 3 or SIM4 mutants did not interact with NbSUMO1 (Additional file 1: Fig. S11C, D and S11E).
We also verified NbSUMO1 interaction with βC1 using in vivo and in vitro pull-down assays. βC1 was able to pull-down GFP-tagged NbSUMO1 from plants as well as recombinantly purified NbSUMO1 in an in vitro pull-down assay. We purified SIM mutated βC1 proteins and used them as bait to pull-down NbSUMO1 (recombinantly expressed as 6X HIS NbSUMO1 in E. coli or transiently expressed in plants as GFP-NbSUMO1). As SIM2, 3 and SIM 4 can separately bind to NbSUMO1, mutating both SIM 2, 3 and SIM4 abolished non-covalent interactions with SyYVCV βC1 (Additional file 1: Fig. S12A and S12B).
SIMs of SyYVCV βC1 are essential for its function as a symptom determinant
SIMs play an important role in protein-protein interactions. To gain further insight into the functional significance of SIMs in βC1, we generated N. tabacum transgenic lines expressing βC1 mutants, where SIMs were mutated to structurally similar motif but without the potential to interact with SUMO (Additional file 1: Fig. S10B). The single mutant of either SIM2, 3 (mS SIM2, 3) or SIM4 (mS SIM4) reverted most of the symptoms observed in βC1 overexpressing transgenic lines (Fig. 5a). Although single SIM mutants of βC1 transgenic plants had reduced severity of the symptoms when compared to WT βC1, they still exhibited mild symptoms such as yellowing of leaves and enhanced branching (Fig. 5b). However, unlike WT βC1, transgenic lines stably expressing βC1 SIM mutants were fertile, with no defect in floral organs and produced viable seeds (Fig. 5b–d). To understand the basis for this phenotype reversal, we further quantified the level of protein expression of the SIM mutants in transgenic plants. Unlike mK18, 24 R mutant, single SIM mutants were stable in plants and maintained protein levels comparable to that of WT βC1 (Fig. 5e). We further validated the stability of SIM mutants by transiently overexpressing C-terminal SIM mutant proteins in plants followed by an IP analysis (Fig. 5f). To reinforce our observation, we transiently overexpressed βC1 with its SIM site mutated to structurally similar motif or to an alanine patch completely removing SIM potential (Fig. 5g and h quantification) (Additional file 1: Fig. S10B). Surprisingly, SIM mutants were much more stable than WT βC1, and mutating a single C-terminal SIM motif was sufficient to increase the stability of the protein considerably. These results suggest that SyYVCV βC1 SIM motifs also play an important role in its function and stability.
Since removing SIMs of βC1 led to an increase in protein stability, we mutated SIM 2, 3 and SIM 4 of βC1 N-terminal SUMOylation motif mutant (mK18, 24R). The N-terminal SUMOylation motif mutant (mK18, 24R) had reduced stability and underwent rapid degradation, whereas upon additional mutations in SIMs enhanced the stability of the protein (Additional file 1: Fig. S12C). We further performed a time course experiment to check for the stability of these SIM and SUMOylation motif mutants of βC1. Unlike SUMOylation motif mutant (mK18, 24R), protein levels of SIM mutated βC1 were much higher at both 1 and 3 DPI even more than WT βC1 levels (Additional file 1: Fig. S12D and 12E quantification).
To understand the reason behind increased stability of SIM mutants, we performed a pull-down experiment using C-terminal SIM mutant (mSIM2, 3, 4) and screened for poly-ubiquitination. WT βC1 exhibited poly-ubiquitination as expected and in accordance with previous studies [12, 13, 44]. MBP control showed very little poly-ubiquitination signal. Very interestingly, SIM mutant unlike WT βC1 did not accumulate poly-ubiquitin chains (Additional file 1: Fig. 12F), indicating that disruption of SIM sites is necessary and sufficient to block ubiquitination. Global ubiquitination was not reduced in any of these samples (Additional file 1: Fig. 12F, Input).
SUMOylation motifs and SIMs of SyYVCV βC1 are necessary for its function as a viral counter-defence protein
βC1 from multiple begomoviruses have been shown to act as host defence suppressors and mediators of viral replication in host plants. Local viral replication assay was performed to ascertain the role of βC1 in augmenting viral replication. Co-inoculation of infectious SyYVCV DNA-A partial dimer along with p35S: SyYVCV βC1 in N. tabacum leaves led to enhanced viral accumulation in local leaves (Additional file 1: Fig. S13A). However, mutating SUMOylation motifs of SyYVCV βC1 led to the complete abolishment of βC1 activity thereby decreasing the viral accumulation (Fig. 6a). Substitution of p35S: SyYVCV βC1 with SIM mutants also reduced viral replication in case of SIM2, 3 mutant (Fig. 6b). Similar results were obtained while infecting SyYVCV DNA-A partial dimer on SIM or SUMOylation motif mutant overexpressing transgenic plants (Additional file 1: Fig. S13A and B). We further performed viral replication assay with individual SIM mutants, and as expected, null mutants and structural mutants behaved similarly. Substitution of WT βC1 with p35S: mSIM 2, 3, 4 triple mutant in local viral replication assay resulted in the loss of βC1 function as a viral replication augmenting protein, resulting in much reduced viral accumulation (Additional file 1: Fig. S13C).
However, surprisingly, when WT βC1 was substituted with its SIM4 mutant, we observed enhanced viral accumulation (Fig. 6b). The exact mechanism behind this local enrichment of virus upon mutating βC1 SIM 4 is not clear. Most probably, differential functions of the two validated SIMs played a role in this disparity observed in viral replication assay. SIM4 of βC1 binds to NbSUMO1 with greater affinity than SIM2, 3 (Fig. 4c, right panel). Even transgenic lines expressing mutated SIM2, 3 or SIM4 motif showed distinct recovery phenotype as compared to WT βC1. Altogether, these results suggest that βC1 SIM and SUMOylation motifs are necessary for its function as a viral pathogenicity determinant protein.
SUMOylation motifs and SIMs of SyYVCV βC1 are also essential for systemic viral movement
In bipartite begomovirus with DNA-A and DNA-B, the B component codes for ΒC1 and BV1 that act as movement-associated proteins in the systemic spread of the virus. In case of monopartite viruses with a β satellite, the function of ΒC1 and BV1 is fulfilled by β satellite that codes for a single βC1 protein. To verify the function of SyYVCV βC1 as a viral movement protein (MP) and to determine the importance of its SIM and SUMOylation motifs in viral movement, we co-inoculated partial dimers of SyYVCV DNA-A along with SyYVCV DNA-β with its only ORF coding for WT βC1 or mSIM 2, 3, 4 (SIM) or mK18, 24, 83R (SUMO) motif mutants incorporated in viral genome, in 3-week-old N. benthamiana plants. As replication of DNA β is assisted by DNA-A coded Rep protein, we first verified that all DNA β dimers are able to replicate locally irrespective of their mutation in βC1 SIM or SUMOylation motifs (Fig. 6c). We observed classic viral symptom development (slight yellowing and curling of newly emerging leaves) only in plants inoculated with DNA-β coding for WT βC1 (Fig. 6d, DNA β plants 1 and 2). Upon Southern analysis to verify viral replication in newly emerging systemically infected leaves, we observed the presence of DNA-A replicative form (RF) in the presence of WT βC1 containing DNA-β (Fig. 6c). The symptoms in plants co-inoculated with DNA-A and DNA β were much prominent at 39 DPI which was also verified by the increased accumulation of DNA-A in systemic leaves. However, we did not observe detectable levels of DNA-A RFs in plants inoculated with DNA-β carrying mutated βC1 of either SIM2, 3, 4, or that of SUMOylation motifs such as SUMO K18,24,83R (Fig. 6c, top blot and Additional file 1: Fig. S13D) even at an earlier time point. These results clearly suggest the importance of SIM and SUMOylation motifs of βC1 in modulating systemic infection by facilitating viral movement.
SUMOylation of βC1 also affects its cellular localization
SUMOylation of specific proteins has been implicated in altering the intracellular localization. In order to explore this possibility and to identify mechanism for the perturbations in the functions of βC1 upon SIM and SUMOylation motif mutations, we transiently expressed GFP-tagged βC1 and its mutants in epidermal cells of N. benthamiana leaves and analysed their localization using confocal microscopy. WT βC1 was diffusely localized in the nucleus and prominently in the nucleolus matching previous observations for related homologs [45]. Surprisingly, we also observed βC1 localizing in the chloroplast, strongly overlapping with chlorophyll auto-fluorescence shown in red (Fig. 7a, 2nd row). Mutating N-terminal double SUMOylation motif mK18, 24R did not result in any qualitative defect in chloroplastic localization (Fig. 7a, 3rd row). However, interestingly, we observed that the Ss3 mutant mK83R showed defects in chloroplastic localization even though its nucleolar localization was unaffected (Fig. 7a, 4th row). The same was observed in the case of triple SUMOylation motif mutant mK18, 24, 83R, where chloroplastic localization was significantly reduced (Fig. 7a, 5th row). We hypothesize that altered localization of βC1 mutants might have abolished its ability to support movement and replication of SyYVCV. In contrast, mutating SIM of βC1 did not alter the localization. SIM mutants were localized similar to WT βC1 in the nucleus, nucleolus and chloroplast, indicating SIM-mediated interactions did not affect localization (Additional file 1: Fig. S14A).