SCE1, the SUMO-Conjugating Enzyme in Plants That Interacts with NIb, the RNA-Dependent RNA Polymerase of Turnip Mosaic Virus, Is Required for Viral Infection

Abstract

SUMOylation, which is catalyzed by small ubiquitin-like modifier (SUMO) enzymes, is a transient, reversible posttranslational protein modification that regulates diverse cellular processes. Potyviruses, the largest group of known plant viruses, comprise many agriculturally important viruses, such as Turnip mosaic virus (TuMV). The potyviral genome encodes 11 mature proteins. To investigate if SUMOylation plays a role in potyvirus infection, a yeast two-hybrid screen was performed to examine possible interactions of each of the 11 viral proteins of TuMV with AtSCE1, the only SUMO-conjugating enzyme in Arabidopsis thaliana homologous to the key SUMO-conjugating enzyme E2 in mammalian cells or Ubc9 in yeast. A positive reaction was found between AtSCE1 and NIb, the potyviral RNA-dependent RNA polymerase. Further bimolecular fluorescence complementation (BiFC) and fluorescence resonance energy transfer (FRET) assays revealed that the NIb and AtSCE1 interaction occurred in both the cytoplasm and nuclei of epidermal cells of Nicotiana benthamiana. The interaction motif was mapped to a region encompassing NIb amino acids 171 to 300 which contains a potential negatively charged amino acid-dependent SUMOylation motif (NDSM). An Escherichia coli SUMOylation assay showed that NIb can be SUMOylated and that the lysine residue (K172) in the motif is a potent SUMOylation site. A TuMV infectious clone with an arginine (R) substitution mutation at K172 compromised TuMV infectivity in plants. In comparison with wild-type Arabidopsis plants, sce1 knockdown plants exhibited increased resistance to TuMV as well as a nonrelated RNA virus. To the best of our knowledge, this is the first report showing that the host SUMO modification system plays an essential role in infection by plant RNA viruses.

INTRODUCTION

Posttranslational modifications (PTMs) of proteins play crucial roles in a variety of biological processes, such as cellular differentiation, development, and rapid and specific responses to endogenous or exogenous stimuli. Besides phosphorylation, glycosylation, and ubiquitination, SUMOylation has emerged as a central theme of PTMs in diversifying proteome activity (1). SUMOylation is a highly dynamic process by which small ubiquitin-like modifiers (SUMOs), a group of small proteins, are conjugated to a wide range of cellular proteins (2). As is generally known, the reversible process of SUMOylation is carried out through a multistep enzymatic reaction (1). At first, SUMO hydrolase removes some C-terminal amino acids (aa) of SUMO, exposing two glycine residues to facilitate SUMO conjugation. The SUMO molecule is then adenylated and covalently linked to a SUMO-activating E1 enzyme. Subsequently, SUMO is transferred to the SUMO-conjugating E2 enzyme SCE1 (Ubc9 in yeast). Finally, SCE1 directs the attachment of the small SUMO protein to the target proteins via an isopeptide bond, a process catalyzed by E3 ligase. SCE1 can directly recognize and SUMOylate Lys residues embedded in a SUMOylation consensus motif, ψ-K-X-E/D, where ψ represents a hydrophobic amino acid residue, including Leu, Ile, Val, or Phe (3, 4).

An increasing body of evidence suggests that SUMOylation regulates diverse cellular processes, such as transcriptional regulation, nuclear-cytoplasmic shuttling, RNA transport, and protein-protein interactions (3–5). SUMOylation has been implicated in infections by various intracellular pathogens (6–10). As cell invaders and usurpers of the cellular machinery, viruses have evolved to manipulate the conserved host cell SUMOylation system for productive infection (11, 12). A number of viral proteins, almost all from mammalian viruses (either DNA or RNA viruses), have been shown to interact with SCE1 (13–15). Although in vivo SUMOylation is not detectable for all the SCE1-interacting viral proteins, some are indeed SUMOylated (16–18). Regardless of being SUMOylated or not, such interactions have been suggested to modify the SUMOylation status of host proteins by de novo SUMOylation or by enhancing deSUMOylation and/or to directly regulate the function of the SCE1-interacting viral proteins (11, 13). Despite the importance of SUMOylation in host-pathogen interactions, very little has been studied regarding its involvement in infections by plant viruses. So far, the only published data related to plant viruses have come from studies of Tomato golden mosaic virus (TGMV), a geminivirus. In this case, the interaction between SCE1 and the viral replicase protein AL1 is essential for the infectivity of this DNA virus (19, 20).

To explore if SUMOylation also plays a role in infections by plant RNA viruses, potyviruses were selected for this study. Potyviruses constitute the largest plant virus family, including many agriculturally important viruses, such as Turnip mosaic virus (TuMV), Tobacco etch virus (TEV), and Soybean mosaic virus (SMV). The potyviral genome is a monopartite, positive-sense single-stranded RNA (ssRNA) of about 10 kb that has a viral protein genome (VPg) covalently linked to its 5′ end and a poly(A) tail at the 3′ end. The viral genome, which replicates in the cytoplasm, contains a single open reading frame (ORF) encoding a large polyprotein of about 350 kDa and a truncated frameshift peptide, which together are processed into 11 protein end products (21, 22). To study if SUMOylation is involved in potyviral infection, we screened for positive interactions between each of the 11 potyviral proteins and SCE1, the key SUMOylation enzyme in plants. We found that SCE1 from Arabidopsis or Nicotiana benthamiana is an interaction partner of NIb (the nuclear inclusion “b” protein). We further mapped the SCE1 binding region to the central region of NIb, which contains a potential negatively charged amino acid-dependent SUMOylation motif (NDSM). We demonstrated that NIb could be SUMOylated. A point mutation in the potent SUMOylation motif affected TuMV infectivity. Moreover, SCE1-silenced plants exhibited resistance to TuMV. Our data suggest that SCE1 is an essential host factor for TuMV infection.

MATERIALS AND METHODS

Plasmid construction.

Gateway technology (Invitrogen, Burlington, Ontario, Canada) was used to generate all the vectors used in this work, unless stated otherwise. Gene sequences were amplified by PCR using Phusion DNA polymerase (NEB, Pickering, Ontario, Canada). The full-length NIb coding sequences of TuMV (nucleotides [nt] 7208 to 8758; GenBank accession no. NC002509), TEV (nt 6981 to 8516; GenBank accession no. NC001555), and SMV (nt 6984 to 8534; GenBank accession no. EU871724) were obtained by PCR amplification of vectors pCambiaTunos/GFP (23), p35STEV (24), and SMV-L (25), respectively, with relevant primer pairs (data not shown). The cDNAs of AtSCE1 (GenBank accession no. NM115649) and NbSCE1 (GenBank accession no. AJ580839) were amplified from Arabidopsis thaliana and N. benthamiana leaves, respectively, by use of relevant primer pairs (data not shown). The resulting DNA fragments were purified and transferred by recombination into the entry vector pDONR221 (Invitrogen), using BP clonase II (Invitrogen) following the supplier's recommendations. Insertions in the resulting pDONR clones were verified by DNA sequencing.

Forward primers NIb-F, NIb-148F, and NIb-171F (data not shown) were designed to amplify regions of TuMV NIb starting at amino acids 148, 171, 301, and 422. Reverse primers NIb-147R, NIb-300R, NIb-422R, and NIb-R were designed to amplify regions of NIb ending at amino acids 148, 300, and 422. The amplified fragments, NIb_1-147, NIb_148-422, NIb_171-300, and NIb_301-517, were cloned into plasmid pDONR221. An NIb single amino acid substitution mutant (K172R) was generated using primers NIb-K172R-F and -R by overlap PCR technology. The amplified fragments were cloned into the pDONR221 vector and then verified by DNA sequencing.

To construct vectors for the yeast two-hybrid assay, the aforementioned pDONR221 constructs were ligated with a modified pGADT7gateway or pGBKT7gateway vector (26) to yield pGAD-NIb, pGAD-TEV NIb, pGAD-SMV NIb, pGAD-NIb_1–147, pGAD-NIb_148–422, pGAD-NIb_171–300, pGAD-NIb_301–517, pGBK-AtSCE1, and pGBK-NbSCE1. The vectors used as two-hybrid assay controls were pGBKT7-p53 (Clontech, Mountain View, CA), encoding the murine p53 protein fused to the GAL4 DNA binding domain (DBD); pGBKT7-Lam (Clontech), encoding the GAL4 DBD fused with human lamin C; and pGADT7-T (Clontech), encoding a fusion of the simian virus 40 (SV40) large T antigen and the GAL4 activation domain.

For bimolecular fluorescence complementation (BiFC) experiments, AtSCE1, NIb, and NIb_171–300 were ligated with a 35S-YNgateway or 35S-YCgateway vector (26) to yield NIb-YC, NIb_171–300-YC, and AtSCE1-YN. Vectors CI-YN and P3NPIPO-YC (27) were used as positive controls.

For transient expression analysis in plant cells, NIb and AtSCE1 were ligated with pEarley101 or pEarley102 to produce plant expression vectors for transient expression of NIb-yellow fluorescent protein (NIb-YFP), NIb_171–300-YFP, and AtSCE1-cyan fluorescent protein (AtSCE1-CFP) in plants. The GUS gene was retrieved from plasmid pENTR-GUS (Invitrogen) and ligated with pEarley101 to yield GUS-YFP as a control.

To silence SCE1 by use of a Tobacco rattle virus (TRV)-based virus-induced gene silencing (VIGS) system, a 228-bp AtSCE1 cDNA fragment was amplified by PCR using primers SCE1-189EcoRIF and SCE1-416BamHIR (primer sequences are available upon request). The amplified fragment was digested with EcoRI and BamHI and cloned into the corresponding sites of the pTRV2 vector (28) to generate vector TRV-SCE1.

Yeast two-hybrid screen.

Yeast two-hybrid assays were performed following the Clontech yeast protocol handbook. In brief, yeast cells (strain AH109) were plated onto a selective medium lacking tryptophan and leucine (SD−Trp−Leu) and a selective medium lacking tryptophan, leucine, histidine, and adenine (SD−Trp−Leu−His−Ade). Quantitative β-galactosidase assays were performed on liquid cultures, using ONPG (o-nitrophenyl-β-d-galactopyranoside; Rockland, Kirkland, Quebec, Canada) as the substrate. Proteins were extracted using a total protein extraction kit as instructed (Dualsystems Biotech, Schlieren, Switzerland). Immunoblotting of full-length NIb or NIb truncation proteins extracted from 6 ml of yeast cells at an optical density at 600 nm (OD₆₀₀) of 0.6 was visualized with an enhanced chemiluminescence (ECL) detection system (Amersham Pharmacia Biotech, Baie d'Urfe, Quebec, Canada). The primary antibody was an anti-hemagglutinin (anti-HA) polyclonal antibody (Sigma, St. Louis, MO).

Confocal microscopy.

For BiFC assay, reconstitution of YFP or CFP was determined by transient coexpression of the selected protein pairs in N. benthamiana leaves that were infiltrated with Agrobacterium tumefaciens GV3101 carrying the corresponding binary plasmids. Yellow or cyan fluorescence was analyzed as described previously (27). For fluorescence resonance energy transfer (FRET) analysis, 4-week-old N. benthamiana leaves were coinfiltrated with YFP and CFP fusions. After 48 h of incubation, leaf epidermal cells exhibiting coexpression of both fluorescent proteins were bleached five times in the acceptor YFP channel with a 514-nm argon laser. Before and after photobleaching, CFP fluorescence intensity was monitored by confocal microscopy (TCS SP5 2; Leica), and the FRET efficiency was calculated as follows: E = [(CFP signal after photobleaching − CFP signal before photobleaching)/CFP signal after photobleaching] × 100 (29).

SUMOylation assays in Escherichia coli.

The Duet cloning system (Novagen, Mississauga, Canada) was used to reconstitute the Arabidopsis SUMO pathway in E. coli. Putative substrates were cloned into pET32b-GW as fusions with N-terminal Trx and 6-His tags (30). E. coli BL21(DE Star) cells (Novagen) transformed with both pCDF-SAE and pACYC-SCE-SUMO or pACYC-SCE(C94S)-SUMO were kindly provided by Nabil Elrouby and George Coupland (30). pET32b-GW containing the test substrate was then introduced into these strains. To introduce expression of all genes and SUMOylation of the substrate, cultures were grown at 37°C for 2 h, induced with 1 mM IPTG (isopropyl-β-d-thiogalactopyranoside; Invitrogen), and incubated at 28°C overnight and then at 25°C for 2 h. Proteins were separated by 7.5% SDS-PAGE and blotted onto polyvinylidene difluoride (PVDF) membranes (Bio-Rad) by standard methods. Anti-Trx (Sigma-Aldrich) was the primary antibody, and a goat anti-rabbit antibody conjugated to horseradish peroxidase (Sigma-Aldrich) was the secondary antibody. The blots were treated with ECL detection reagents (GE Healthcare) and exposed to films for visualization.

Plant growth, TRV-based VIGS, T-DNA mutant analysis, and virus inoculation.

Arabidopsis plants were grown in pots at 23°C in a growth chamber under a 16-h–8-h photoperiod with 60% humidity. For VIGS assays, wild-type Arabidopsis seedlings (ecotype Col-0) were used approximately 15 to 17 days after seed germination. pTRV1 and pTRV-SCE1 were introduced into A. tumefaciens strain GV3101. Arabidopsis plants coinfiltrated with Agrobacterium cells harboring pTRV1 and pTRV2 empty vector or pTRV2-PDS were used as controls. A 5-ml culture was grown overnight at 28°C in 50 mg/liter rifamycin, 100 mg/liter kanamycin, and 25 mg/liter gentamicin. A. tumefaciens cells were harvested and resuspended in infiltration medium (10 mM MgCl₂, 10 mM MES [morpholineethanesulfonic acid], pH 5.6, and 150 μM acetosyringone), adjusted to an OD₆₀₀ of 1.5, and left at room temperature for 3 to 4 h. All Arabidopsis inoculations were performed using 1:1 (vol/vol) mixtures of two GV3101 agrobacterial cultures, one containing the pTRV1 vector and the second containing a pTRV2-derived vector. Two entire leaves of each three-leaf-stage seedling were infiltrated with a needleless 1-ml syringe. At 12 days post-VIGS treatment, the plants were challenged with TuMV by mechanic inoculation (31). Each silencing experiment was repeated at least three times, and each treatment included at least three independent plants.

T-DNA insertion lines (SALK_006164 and SALK_022200) were obtained from the Arabidopsis Biological Resource Centre (http://www.biosci.ohio-state.edu/pcmb/Facilities/abrc/abrchome.htm). T-DNA mutant analysis was carried out as described previously (31). Wild-type and sce1 mutant Arabidopsis plants were inoculated with TuMV and Tobacco mosaic virus (TMV) (32) mechanically or via agroinfiltration. The inoculation assay was repeated at least three times, and each treatment included three pots (each consisting of four plants).

SUMOylation in wild-type Arabidopsis and its sce1 mutants.

Antibodies against SUMO1 (ab5316) and SCE1 (ab98965) were purchased from Abcam (Toronto, Canada). A monoclonal anti-actin (plant) antibody was purchased from Sigma. Goat anti-rabbit or anti-mouse secondary immunoglobulins conjugated to horseradish peroxidase were purchased from Sigma. Wild-type Arabidopsis, sce1 T-DNA mutant, TRV-SCE1-silenced, and TRV empty vector-inoculated seedlings were subjected to a 37°C heat shock for 30 min and then returned to 24°C to recover. Leaf samples were homogenized directly in SDS-PAGE sample buffer, and the extracts were subjected to immunoblot analysis with anti-SUMO1 and anti-SCE1 antibodies. An anti-actin blot was used to confirm equal protein loading. Extracts were rapidly heated and then centrifuged at 14,000 × g for 15 min. Supernatants were used directly for SDS-PAGE (12%) and immunoblot analysis.

Bioinformatic analysis and structural modeling.

A three-dimensional (3D) model of TuMV NIb was obtained using the Web server I-TASSER (http://zhanglab.ccmb.med.umich.edu/I-TASSER/; Center for Computational Medicine and Bioinformatics, University of Michigan, MI) (33), and the structures were constructed using PyMol (34).

TAS ELISA, quantitative reverse transcription-PCR (qRT-PCR), and SCE1 detection.

Triple-antibody sandwich enzyme-linked immunosorbent assay (TAS ELISA) was conducted using a TuMV TAS ELISA kit (Agdia, Elkhart, IN). Absorbances were recorded at 405 nm with an iMark microplate reader (Bio-Rad, Mississauga, Ontario, Canada).

For qRT-PCR, total RNA was isolated from Arabidopsis leaf tissues by use of TRIzol reagent following the supplier's protocol (Invitrogen). One microgram of RNA pretreated with DNase I was used as a template for first-strand cDNA synthesis, using Superscript III reverse transcriptase and an oligo(dT)_12–18 primer (Invitrogen). Primer pairs (Actin-F/R, TuMV CP-rtF/rtR, TRV CP-rtF/rtR, and SCE1-73F/174R) for real-time PCR were designed based on online software (http://quantprime.mpimp-golm.mpg.de/). Real-time PCR amplification, validation experiments, and calculations of relative differences in expression levels were performed per the manufacturer's instructions (Bio-Rad). Briefly, real-time PCR amplification was performed with 10 μl of reaction solution, containing 40 ng of cDNA, 5 μM (each) primers, and 1× SsoFast EvaGreen Supermix (Bio-Rad). Relative transcript abundances were calculated using Bio-Rad CFX Manager software (version 1.6), with Actin as a reference.

Proteins were extracted from Arabidopsis leaf tissues by use of 4× denaturing SDS-PAGE loading buffer. An anti-AtSCE1 antibody (Abcam, Cambridge, MA) was used to detect the SCE1 protein.

Construction of infectious TuMV clone mutants.

PCR-based overlap extension mutagenesis was used for site-directed mutation (35). To generate a unique enzyme sequence, a PCR product was synthesized with primers TuMV-SnaBI-F and TuMV-MluI-R in the presence of pCambiaTunos/GFP. The amplified fragment was ligated into the pGEM-T Easy vector (Promega), digested with SalI and SacII, and then ligated into the corresponding sites of the pBSK vector to generate pBSK-SnaBI-MluI. The inserted sequence was confirmed by DNA sequencing.

The NIb mutant NIb172R, described above, was digested with NdeI and MluI and ligated into similarly digested pBSK-SnaBI-MluI to generate pBSK-SnaBI-MluI-172R. The resulting vector was digested with SnaBI and MluI and ligated into similarly digested pCambiaTunos/GFP to generate TuMV172R. Agrobacterium infiltration was used to inoculate N. benthamiana plants. At 8 days postinoculation (dpi), pictures were taken under a handheld UV lamp (36), and RNAs or proteins were extracted from infiltrated leaves (local) and systemic leaves (systemic) of three plants per pot as a pool, with three repeats (pools). The experiment was performed three times. Immunoblotting of the CP protein expressed in N. benthamiana was monitored with a nitroblue–5-bromo-4-chloro-3-indolylphosphate (NBT-BCIP) staining system (Roche, Mississauga, Ontario, Canada), using polyclonal anti-TuMV antibodies (Agdia).

RESULTS

The SUMO-conjugating enzyme (SCE1) in plants is an interacting partner of the potyviral RdRp NIb.

To investigate whether SUMOylation plays a role in potyviral infection and to identify putative SUMO substrates from potyviruses, we performed a yeast two-hybrid screen to examine possible interactions of each of the 11 proteins of TuMV with AtSCE1, the only Arabidopsis SUMO-conjugating enzyme. Cotransformants were selected and plated on selective media to detect activation of the reporter genes HIS3 and ADE2. A positive reaction was found between AtSCE1 and NIb, the potyviral RNA-dependent RNA polymerase (RdRp) (Fig. 1A).

Fig 1.

NIb, the RNA-dependent RNA polymerase of potyviruses, interacts with SCE1, the key SUMO-conjugating enzyme. (A) Yeast two-hybrid assay of protein-protein interaction between NIb and SCE1. Yeast cotransformants were grown on selective medium. Measurements of β-galactosidase activity in total protein extracts (the amount that hydrolyzes 1 μmol of ONPG to o-nitrophenol and d-galactose per min per cell) were performed, and the error bars correspond to standard errors. Bars: 1, pGAD-T plus pGBK-p53 (positive control); 2, pGAD-T plus pGBK-Lam (negative control); 3, TuMV NIb plus AtSCE1; 4, TuMV NIb plus NbSCE1; 5, TEV NIb plus AtSCE1; 6, TEV NIb plus NbSCE1; 7, SMV NIb plus AtSCE1; 8, SMV NIb plus NbSCE1. (B) BiFC assay of NIb and AtSCE1 interactions in N. benthamiana. Three-week-old seedling leaves were coagroinfiltrated with constructs expressing NIb and AtSCE1 fused to the C- or N-terminal half of YFP. Reconstituted YFP fluorescence was monitored 2 days after infiltration, using a confocal microscope (frame I). Negative controls coexpressing YN and NIb or AtSCE1-YN and YC are shown in frames II and III. Coexpression of TuMV CI-YN and P3NPIPO-YC served as a positive control. DIC, differential interference contrast.

To determine whether the AtSCE1 protein is also able to interact with the NIb proteins from other potyviruses, we fused the GAL4 activation domain to the NIb coding sequences of TEV and SMV and evaluated their possible interactions with AtSCE1 in yeast. Cotransformants were isolated and plated on different selective media to detect activation of the reporter genes HIS3 and ADE2. As shown (Fig. 1A), the NIbs from TEV and SMV also interacted with AtSCE1, as only the cells in which the reported genes were activated were able to grow in a medium lacking histidine and adenine. Since Arabidopsis is not a susceptible host of SMV, we also cloned SCE1 from N. benthamiana (designated NbSCE1) and evaluated the interaction of NbSCE1 with the NIbs of TuMV, TEV, and SMV. The results showed that all three NIbs tested (from TuMV, TEV, and SMV) could interact with NbSCE1 (Fig. 1A).

The interaction of AtSCE1 or NbSCE1 with each of the three NIbs was quantified by measuring β-galactosidase activity in AH109, since activation of the GAL4-responsive promoter drives the transcription of the reporter gene lacZ. The β-galactosidase activity in yeast cells with any combination of SCE1 and NIb was significantly higher than that in the negative control (Fig. 1A). Among the six combinations, coexpression of TuMV NIb and AtSCE1 or NbSCE1 led to a remarkably high β-galactosidase activity (Fig. 1A). TuMV NIb was therefore chosen for subsequent characterization.

To confirm the interaction between NIb and AtSCE1 in planta, a BiFC assay was carried out. AtSCE1-YN and NIb-YC were transiently coexpressed in N. benthamiana by agroinfiltration. Yellow fluorescence was clearly observed in the nucleus and cytoplasm (Fig. 1B, frame I). No BiFC fluorescence was detected in the negative controls coexpressing YN and NIb-YC or coexpressing AtSE1-YN and YC (Fig. 1B, frames II and III). As expected, yellow fluorescence in cells coexpressing CI-YN and P3NPIPO-YC as a positive control was evident (27) (Fig. 1B, frame IV). These data suggest that AtSCE1 does interact with TuMV NIb in the plant cell.

We further coexpressed NIb-YFP and AtSCE1-CFP in N. benthamiana leaf epidermal cells and examined the physical interaction between NIb and SCE1 by using the FRET technique, which permits high-spatial-resolution assays of protein-protein interactions in living cells. Consistent with the above observations, SCE1-CFP and NIb-YFP were colocalized in the cytoplasm and nucleus (Fig. 2A). No obviously different distribution patterns of NIb-YFP were evident when NIb-YFP was expressed alone (Fig. 2B) or in the presence of AtSCE1 (Fig. 2A). Analyses of the change in the cyan fluorescence signal intensity after photobleaching of the yellow fluorescence revealed that the cyan fluorescence signal intensity increased, with a FRET efficiency of 21.85% for seven different regions of interest (ROIs) with two repeats (Fig. 2C), whereas the FRET efficiency of a negative control employing SCE1-CFP and GUS-YFP was negligible (Fig. 2C). These data further confirm that TuMV NIb interacts with AtSCE1 in planta.

Fig 2.

Colocalization of NIb and AtSCE1 in planta. (A) Transient coexpression of NIb and AtSCE1 in N. benthamiana. Fluorescence was monitored 2 days after infiltration, using a confocal microscope. (B) Transient expression of NIb alone in N. benthamiana. Fluorescence was monitored 2 days after infiltration, using a confocal microscope. (C) Quantification of FRET efficiency after acceptor photobleaching. FRET efficiency was calculated with the following formula: FRET efficiency = [(CFP signal after photobleaching − CFP signal before photobleaching)/CFP signal after photobleaching] × 100. AtSCE1-CFP plus GUS-YFP was used as a negative control. The error bar indicates the standard deviation (SD) for 7 independent FRET analyses in two independent experiments.

The SCE1 binding domain of NIb is mapped to its middle region, where amino acid sequences are highly conserved among potyviral NIbs.

To map the AtSCE1 binding region in the NIb protein, yeast two-hybrid assays were used to analyze the interactions between AtSCE1 and different regions of NIb. Based on conserved domain prediction (http://www.ncbi.nlm.nih.gov), NIb could be divided into three fragments: N-terminal (NIb_1–147), central (NIb_148–422), and C-terminal (NIb_423–517) fragments. Weak interactions existed between AtSCE1 and the N- or C-terminal fragment (aa 1 to 147 or aa 301 to 517) (Fig. 3A). In contrast, the central region of NIb (aa 148 to 422) showed a strong interaction with AtSCE1, with a strength similar to that with full-length NIb (aa 1 to 517) (Fig. 3A). Deletion of the first 23 and last 122 amino acids of the central region did not affect the binding ability of the central core fragment (NIb_171–300) to AtSCE1 (Fig. 3A), suggesting that NIb_171–300 is the SCE1 interaction domain of NIb. Western blotting using anti-HA antibodies detected comparable levels of NIb or its deletion fusions with the GAL4 activation domain and an HA epitope tag in yeast (data not shown), suggesting that the variations in the interaction efficiency were not due to differences in expression levels of the fusion proteins.

Fig 3.

SUMOylation of NIb in E. coli. (A) The central region of NIb interacts with AtSCE1 in the yeast two-hybrid assay. Truncated NIb proteins are indicated by the positions of the first and last amino acid residues. (B) Sequence alignment of the SCE1-interacting region (NIb_171–300) of TuMV NIb (GenBank accession no. NC002509) and the corresponding regions of PPV (Plum pox virus; GenBank accession no. P13529), PVA (Potato A virus; GenBank accession no. Q85197), PVY (Potato Y virus; GenBank accession no. S70722), SMV (Soybean mosaic virus; GenBank accession no. Q90069), and TEV (Tobacco etch virus; GenBank accession no. P04517). Identical amino acids are highlighted in yellow. The conserved SUMOylation motif (LKAE) is marked by stars. (C) SUMOylation of NIb or mutant NIb K172R in E. coli with a reconstituted SUMO conjugation pathway. Western blot analysis shows that both NIb and the NIb K172R mutant are SUMOylated by SUMO3, but with a lower level of SUMOylation for the K172 mutant. The arrow points to the SUMOylated form; the asterisk indicates the regular form.

To further verify the interaction between NIb_171–300 and AtSCE1 in planta, we carried out a BiFC assay and a FRET analysis. YFP fluorescence was observed in N. benthamiana cells coexpressing NIb_171–300-YC and AtSCE1-YN (Fig. 4A). Consistently, NIb_171–300-YFP and AtSCE1-CFP were colocalized to the cytoplasm and the nucleus (Fig. 4B). After photobleaching of the yellow fluorescence, the cyan fluorescence signal intensity increased, with a FRET efficiency of 18.86% for seven different ROIs with two repeats (Fig. 4C). These data confirm that the AtSCE1 binding domain alone can efficiently interact with AtSCE1 in planta.

Fig 4.

NIb_171–300 interacts with AtSCE1 in planta. (A) BiFC assay of NIb_171–300 and AtSCE1 interactions in N. benthamiana. Reconstituted YFP fluorescence was monitored 2 days after infiltration, using a confocal microscope. (B) Colocalization of NIb_171–300-YFP and AtSCE1-CFP. (C) Quantification of FRET efficiency after acceptor photobleaching. FRET efficiency was calculated as described in the legend to Fig. 2. AtSCE1-CFP plus GUS-YFP was used as a negative control. Error bars indicate SD for seven independent FRET analyses in two independent experiments.

Amino acid sequence alignments of NIbs from six different potyviruses suggested that the identified AtSCE1 binding domain of NIb was highly conserved among all the potyviruses analyzed, including TuMV, Plum pox virus (PPV), Potato A virus (PVA), Potato Y virus (PVY), SMV, and TEV (Fig. 3B), indicating that the NIb-SCE1 interaction is likely a general event for all potyviruses.

NIb is SUMOylated in E. coli.

Because AtSCE1 is the only Arabidopsis SUMO-conjugating enzyme, its interacting partner may be a substrate of the enzyme. To detect the SUMOylated form of NIb, an NIb-His fusion protein was transiently expressed in plants. However, possibly due to the reversible nature of the SUMOylation process or to only a very small amount of viral protein being SUMOylated, as suggested previously (20), the SUMOylated form of NIb was not detected by immunoblotting (data not shown). An E. coli strain system that was reconstituted with the SUMO conjugation pathway (30) was employed for SUMOylation assay. In this system, expression of proteins and SUMOylation of the substrate take place in bacterial cells (30). Immunoblot analysis showed that NIb was SUMOylated, as a higher-molecular-weight form of NIb was evident in the proteins extracted from E. coli containing the functional SUMO conjugation pathway (NIb, SUMO3, E1, and E2) but was not detectable when E2 was replaced with a dysfunctional E2 mutant, i.e., E2(C94S) (Fig. 3C). These results suggest that NIb can be SUMOylated.

The SCE1 binding domain contains a potential SUMOylation motif in which a point mutation (K172R) inhibits viral accumulation in plants.

Published data suggest that SUMOylation substrates interact with SCE1 primarily through the SUMOylation motif ψ-K-X-D/E (37, 38), and SCE1 binds preferentially to hydrophobic regions containing LK and/or KL dipeptides (39). Analyses of the central NIb region by use of a SUMOylation prediction program (SUMOsp 2.0 [http://SUMOsp.biocuckoo.org]) identified a putative SUMOylation motif, ¹⁷¹LKAE¹⁷⁴, characteristic of a negatively charged amino acid-dependent SUMOylation motif [NDSM; Ψ-K-X-E-X_(2–5)-E/D-X₍₂₎-E/D, where K is the putative amino acid for SUMOylation] (40) (Fig. 5A). In agreement with this prediction, the 3D structure of TuMV NIb, based on protein structure and function predictions obtained with a widely used online platform (33), was a donut-like shape, and K172 was positioned on the inner surface of the structure (Fig. 6).

Fig 5.

K172R mutation inhibits viral accumulation. (A) Sequences of SUMO sites within NIb, predicted using the NDSM motif (51). (B) Immunoblot analysis of protein extracts from N. benthamiana infiltrated leaves (Loc) or systemically infected leaves (Sys), using an anti-TuMV monoclonal antibody to detect TuMV. (C) N. benthamiana plants agroinfiltrated with wild-type TuMV (TuMV WT) tagged by GFP or with its NIb 172R mutant (TuMV172R) were examined for viral accumulation at 8 dpi under a handheld long-wavelength UV illuminator. Infiltrated leaves are indicated by an arrow. Weak GFP fluorescence was evident on local leaves of the plants infiltrated with TuMV172R. (D) Relative TuMV RNA accumulation levels were determined by real-time RT-PCR using TuMV CP-specific primers. Samples were collected from infiltrated leaves (Loc) or systemically infected leaves (Sys) at 8 dpi. The real-time RT-PCR data were normalized against the internal control and further converted into fold changes based on the positive control, WT-Loc. WT-Loc, leaf of Arabidopsis inoculated with wild-type TuMV; WT-Sys, systemic leaves of Arabidopsis inoculated with wild-type TuMV; K172R-Loc, leaf of Arabidopsis inoculated with the TuMV172R mutant; K172R-Sys, systemic leaves of Arabidopsis inoculated with the TuMV172R mutant; Mock-Loc, leaf of Arabidopsis inoculated with buffer; Mock-Sys, systemic leaves of Arabidopsis inoculated with buffer.

Fig 6.

Three-dimensional model of NIb from TuMV, in cartoon style (A) or surface style (B), based on protein structure and function predictions obtained using the online I-TASSER server (http://zhanglab.ccmb.med.umich.edu/I-TASSER/; University of Michigan, MI), with structures constructed using PyMol (33, 34). The lysine 172 residue analyzed in this work is indicated in red.

To determine whether K172 is important for viral infection or symptom development, K172 in NIb was replaced with the amino acid arginine (R). The point mutation was introduced into plasmid pCambiaTunos/GFP to generate plasmid TuMV172R. N. benthamiana or Arabidopsis seedlings inoculated with pCambiaTunos/GFP (wild-type TuMV) developed typical TuMV symptoms, such as mosaicism, leaf curling, lumpy, reduced apical dominance, and stunted growth, at 6 to 8 dpi (Fig. 7). In contrast, plants inoculated with the TuMV172R mutant did not produce detectable symptoms at 8 dpi, except for very mild symptoms in infiltrated leaves (Fig. 7). No detectable symptoms in systemic leaves were observed over an extended observation period (4 weeks). Western blotting was conducted to detect the presence of viral CP protein in these plants. TuMV CP accumulated in both the local and systemic leaves in plants inoculated with wild-type TuMV (Fig. 5B). However, in plants inoculated with the TuMV172R mutant, very little CP was detected in the local leaves, and no CP was detectable in the systemic leaves (Fig. 5B). Under a handheld UV lamp, green fluorescent protein (GFP) fluorescence was clearly evident in both the local and systemic leaves of plants inoculated with pCambiaTunos/GFP (GFP-tagged wild-type TuMV) (Fig. 5C). In plants inoculated with the K172R mutant, weak GFP fluorescence was detectable only in small regions of the inoculated leaves, not in the systemic leaves or buffer-infiltrated mock (control) plants (Fig. 5C). We examined TuMV RNA levels in the inoculated and systemic leaves at 8 dpi, using real-time RT-PCR. The results showed that TuMV RNA accumulated in both the inoculated and systemic leaves in plants inoculated with the wild-type virus (Fig. 5D). In plants challenged with the K172R mutant, very little virus was found in the inoculated leaves, and no virus was detected in the systemic leaves (Fig. 5D). Taken together, these results suggest that the K172 residue is a SUMOylation site and is important for TuMV infection.

Fig 7.

Phenotypes of N. benthamiana (A) and Arabidopsis (B) plants inoculated with wild-type TuMV and TuMV172R. Pictures were taken at 12 days postinoculation.

K172 is a potent SUMOylation site, and NIb has additional SUMOylation sites.

To determine whether K172 is important for SUMOylation, E. coli SUMOylation assays were carried out. The results showed that the K172R mutation did not abolish but did affect NIb SUMOylation, as indicated by the fact that the SUMOylated form of NIb K172R was less intensive than that of NIb in three independent experiments (Fig. 3C). Therefore, K172 is a potent SUMOylation site, and NIb has additional SUMOylation sites.

If the K172R mutant can be SUMOylated, it should interact with SCE1. To confirm this, a yeast two-hybrid assay was used to examine the interaction between the NIb K172R mutant and AtSCE1. Indeed, AtSCE1 and the K172R mutant strongly interacted with each other (Fig. 8).

Fig 8.

NIb172R interacts with AtSCE1 in a yeast two-hybrid assay, as does wild-type NIb. pGAD-T+pGBK-Lam, negative control; pGAD-T+pGBK-p53, positive control.

NIb is the potyviral RdRp that binds to RNA. To determine if the K172R mutant can also bind to ssRNA, we conducted a gel electrophoretic mobility shift assay. Our results showed that not only NIb but also the K172R mutant could bind to ssRNA and form complexes, though their efficiencies varied slightly from experiment to experiment (data not shown).

Knockdown of SCE1 expression reduces SUMOylation of cellular proteins and inhibits TuMV and TRV infections in Arabidopsis.

To address whether SCE1 is required for TuMV infection, we initially chose the T-DNA mutant approach. Unfortunately, as SCE1 is essential for Arabidopsis development, no sce1 knockout homozygous mutants are obtainable (41). We analyzed a T-DNA knockdown line of SCE1, SALK_006164 (sce1-4), that has a T-DNA left-border sequence upstream of the 5′-untranslated region (5′-UTR) and located 87 nt upstream of the SCE1 gene start codon (41). Immunoblot analysis confirmed that the amount of SCE1 protein was remarkably reduced in the homozygous T-DNA line compared to that in wild-type plants (Fig. 9B). To check if knockdown of SCE1 affects SUMOylation of cellular proteins, Arabidopsis plants were heat shocked to induce SUMOylation. Immunoblot analyses showed that the amounts of SUMOylated cellular proteins in the sce1-4 mutant plants were much lower than those in the wild-type plants (Fig. 9C).

Fig 9.

Knockdown of SCE1 reduces the cellular SUMOylation capacity and inhibits TuMV infection in Arabidopsis. (A) Phenotypes of wild-type and sce1-4 mutant plants at 15 dpi. (B) SCE1 protein levels after different treatments, determined by Western blotting using anti-SCE1 antibodies. (C) SUMOylation in Arabidopsis. Plants were heat shocked for 30 min at 37°C, allowed to recover at 24°C for 30 min, and then subjected to protein extraction. Western blotting was carried out with anti-SUMO1 serum. (D) ELISA analysis of wild-type and sce1-4 mutant Arabidopsis plants infiltrated with TuMV at 15 dpi. (A, B, and D) WT+Mock, wild-type Arabidopsis infiltrated with buffer; sce1-4+Mock, sce1-4 T-DNA knockdown mutant infiltrated with buffer; sce1-4+TuMV, sce1-4 T-DNA knockdown mutant infiltrated with the GFP-tagged TuMV infectious clone; WT+TuMV, wild-type Arabidopsis infiltrated with the GFP-tagged TuMV infectious clone. (C) WT, wild-type Arabidopsis; sce1-4, sce1-4 T-DNA mutant; TRV-SCE1, wild-type Arabidopsis in which SCE1 was silenced by coinfiltration of pTRV1 and pTRV2-SCE1; TRV, wild-type Arabidopsis infiltrated with pTRV1 and pTRV2 (empty silencing vectors) as a negative control.

The sce1 knockdown mutants and wild-type plants were mechanically inoculated with TuMV. ELISA results revealed that in the new upper leaves of T-DNA knockdown mutants, TuMV accumulated about 50% as much as that in the wild-type plants at 15 dpi (Fig. 9D). Severe disease symptoms such as stunted growth were found in the infected wild-type plants (Fig. 9A). In contrast, the sce1 mutant plants inoculated with TuMV showed weaker disease symptoms. Similar results (data not shown) were also obtained with another T-DNA knockdown line of SCE1, SALK_022200 (sce1-7), which also has a T-DNA left-border sequence upstream of the 5′-UTR and located 92 nt upstream of the SCE1 gene start codon (41).

An alternative approach to using T-DNA mutants for characterization of gene functions is VIGS. We employed a TRV-based VIGS system to silence AtSCE1 in Arabidopsis. In plants treated with a positive control (TRV-PDS), the typical white color resulting from silencing of the PDS gene was visible in about 50% of VIGS-treated plants at 12 dpi (data not shown), suggesting that under our experimental conditions, VIGS was effective in about 50% of plants. Thus, all of the infiltrated plants were mechanically inoculated with TuMV. TAS ELISA was performed to detect TuMV in these plants 15 days after inoculation with TuMV. It was found that all control plants coinfiltrated with pTRV1 and pTRV2 were susceptible to TuMV. In contrast, about 50% of the plants infiltrated with pTRV1 and TRV-SCE1 were resistant to TuMV. We monitored the level of VIGS by real-time RT-PCR and found that low SCE1 transcript levels in TRV-SCE1-treated plants were correlated with high resistance to TuMV (Fig. 10C). Western blotting further confirmed that the AtSCE1 protein level in SCE1-downregulated plants showing TuMV resistance was much lower than that in the wild-type plants, which were highly susceptible to TuMV (Fig. 10B). Consistently, SUMOylation of cellular proteins in the VIGS-induced sce1-silenced plants in response to heat shock treatment was also clearly reduced (Fig. 9C). Phenotypes of the wild-type and SCE1-downregulated plants inoculated with TuMV were also examined. Inoculation of TuMV into the wild-type or empty TRV vector-treated plants induced severe TuMV symptoms, such as mosaicism and necrosis on leaves, severe growth retardation, reduced apical dominance, curled bolts, and the typical inflorescence stunting (Fig. 10A). However, SCE1-downregulated plants inoculated with TuMV did not show such typical TuMV symptoms and exhibited only a slightly smaller stature than that of the mock-treated wild-type plants (Fig. 10A).

Fig 10.

SCE1-silenced plants exhibit strong resistance to TuMV. (A) Phenotypes of silenced plants inoculated by TuMV at 15 dpi. TuMV was inoculated at 12 days postcoinfiltration of pTRV1 and pTRV2 or TRV-SCE1. (B) Western blot analysis of SCE1 proteins from wild-type Arabidopsis (WT) and sce1-silenced Arabidopsis (TRV-SCE1) inoculated with TuMV or buffer (mock). (C) Relative TuMV accumulation and AtSCE1 expression quantities by real-time RT-PCR. RNA was extracted at 15 dpi for real-time RT-PCR analysis. (A to C) WT+Mock, wild-type Arabidopsis pretreated with buffer and then inoculated with buffer; pTRV+TuMV, wild-type Arabidopsis preinoculated with the empty TRV VIGS vector and then inoculated with TuMV; TRV-SCE1+TuMV, wild-type Arabidopsis preinoculated with the TRV-SCE1 VIGS vector to silence SCE1 and then inoculated with TuMV; WT+TuMV, wild-type Arabidopsis pretreated with buffer and then inoculated with TuMV. (D) Relative fold changes in TMV accumulation and SCE1 gene expression. Arabidopsis sce1-4 T-DNA mutant, sce1-silenced (TRV-SCE1), and wild-type plants were inoculated with TMV. RNA extraction was performed at 10 dpi for real-time RT-PCR analysis. sce1-4+TMV, sce1-4 T-DNA mutant agroinfiltrated with TMV; TRV-SCE1+TMV, wild-type Arabidopsis preinoculated with the TRV-SCE1 VIGS vector to silence SCE1 and then inoculated with TMV; WT+Mock, wild-type Arabidopsis inoculated with buffer; WT+TMV, wild-type Arabidopsis agroinfiltrated with TMV.

To further check if knockdown of SCE1 expression would affect infection with a nonpotyvirus, sce1-4 mutants were challenged with TMV. To our surprise, sce1-4 mutants showed strong resistance to TMV (Fig. 10D). Taken together, the above data suggest that SCE1 is required for TMV and TuMV infection.

DISCUSSION

SCE1, a conjugating enzyme, interacts physically with almost all known SUMO acceptors to catalyze the formation of an isopeptide bond between the ε-amino group of the lysine residue of the substrate and the C-terminal glycine residue of SUMO (42, 43). AtSCE1 is the only E2 enzyme required for SUMO modification in Arabidopsis (44, 45). In this work, we identified AtSCE1 as a host protein interacting with the NIb protein of TuMV (Fig. 1 and 2), and TuMV NIb could be SUMOylated (Fig. 3C). Not only AtSCE1 but also its counterpart from N. benthamiana (NbSCE1) can interact with TuMV NIb. These two SCE1s also interact with the NIb proteins from two additional potyviruses, TEV and SMV, and all NIbs contain a highly conserved SUMOylation motif, suggesting that binding to SCE1 and subsequent SUMOylation are likely conserved features of the potyvirus replication protein. Moreover, we showed that viral RNA accumulation was significantly reduced in sce1 knockdown Arabidopsis mutants (Fig. 9D) and SCE1-silenced Arabidopsis (Fig. 10C). Moreover, sce1 knockdown Arabidopsis mutants showed strong resistance to TMV (Fig. 10D), which is not related to TuMV. The molecular mechanism underlying this unexpected resistance to TMV is being explored. Nevertheless, these data suggest that SCE1 is a host factor essential for TuMV infection.

SCE1 can directly SUMOylate Lys residues embedded in the SUMOylation consensus motif of the substrate. The AtSCE1 binding domain of NIb contains the residue K172, positioned in the RdRp functional structure (Fig. 3) and predicted to be the SUMO attachment site (Fig. 5A). When this residue was replaced with arginine, the virus mutant had significantly reduced infectivity (Fig. 5D). However, the NIb K172R mutant could bind to AtSCE1 (Fig. 7) and be SUMOylated, though not as efficiently as the wild type (Fig. 3C). This phenomenon was frequently observed for proteins containing multiple SUMOylation sites, possibly owing to the fact that each SUMOylation site may be conjugated by a different number of SUMO molecules (20). Indeed, analysis of the entire NIb by use of the SUMOylation prediction program revealed additional potential SUMOylation motifs (data not shown). It would be interesting to investigate if these sites are also important for SUMOylation and TuMV infection.

Protein-protein interactions and posttranslational modifications may lead to translocation of the involved proteins from one subcellular compartment to another or to functional changes. SUMO targets and the SUMO-conjugating enzyme Ubc9 (SCE1) participate in nucleocytoplasmic transport (46, 47). NIb and SCE1 are known nuclear proteins (41, 48, 49). In TuMV-infected cells, NIb is recruited into the cytoplasmic membrane-bound vesicles that house the viral replication complex (VRC) through its interaction with 6K-VPg-Pro (27, 48, 50, 51). Therefore, it is possible that nucleocytoplasmic transport of the NIb-SCE1 complex or the SUMOylated form of NIb may facilitate the recruitment of NIb into the VRC. This assumption is supported by findings showing that NIb is involved in nuclear translocation activities (49, 52).

It is also possible that NIb-SCE1 interaction and/or subsequent SUMOylation of NIb may directly regulate the function of NIb. NIb deletion analysis mapped the AtSCE1 binding site to the middle of the viral protein, in the region spanning residues 148 to 422 (Fig. 3). This region constitutes a typical RdRp structure encoded in the genomes of all RNA viruses without a DNA stage. RdRp catalyzes synthesis of the RNA strand complementary to a given RNA template. Two-hybrid analysis of truncated proteins further narrowed down the AtSCE1 binding domain to residues 171 to 300. This domain was conserved among all the potyviruses analyzed in this study, with 66.9% to 86.2% amino acid sequence identities (Fig. 3B). Therefore, it is reasonable to speculate that this region is very important for NIb's replication function and that interaction with AtSCE1 and subsequent SUMOylation may optimize its conformation for the RdRp activity.

One more speculation is that interactions of SCE1 and viral proteins may impair the SUMOylation pathway in infected cells to generate a cellular environment more favorable for viral multiplication (13). Indeed, a recent report identified 238 Arabidopsis proteins as potential SUMO substrates (30), and expression of SCE1-interacting AL1 did modify the SUMOylation state of certain host proteins (20). It remains to be determined if expression of NIb can also disturb the SUMOylation profiles of host proteins.

As indicated earlier, SUMO is conjugated to Lys residues. The conjugation of ubiquitin to cellular proteins, a posttranslational modification referred to as ubiquitination, also occurs at Lys residues (53). In contrast to SUMOylation, in ubiquitination the proteins attached to ubiquitin are transferred to the proteasome for degradation. Thus, ubiquitination and SUMOylation may compete for the Lys residues of NIb. As a result, the fate and functions of ubiquitinated and SUMOylated NIbs are expected to be completely different. Ubiquitination has also been implicated in plant virus infections (53). The possible involvement of ubiquitination in regulating NIb function is certainly worth investigating.

In recent years, the roles of SUMOylation in fundamental biology, biochemistry, and host-microbe interactions have just begun to be understood. Our data represent the first report that SCE1 is an essential host factor for a plant RNA virus and that SUMOylation may play a crucial role for the virus to establish infection. However, the exact role of the interactions of SCE1 with a viral protein(s) and/or subsequent SUMOylation of the viral protein(s) in infections by viruses remains a challenging question to be answered.