C2 from Beet curly top virus promotes a cell environment suitable for efficient replication of geminiviruses, providing a novel mechanism of viral synergism

Authors

  • Zaira Caracuel,

    1. Instituto de Hortofruticultura Subtropical y Mediterránea ‘La Mayora’, Universidad de Málaga-Consejo Superior de Investigaciones Científicas (IHSM-UMA-CSIC), Departamento Biología Celular, Genética y Fisiología, Universidad de Málaga, Campus Teatinos, 29071 Málaga, Spain
    Search for more papers by this author
    • These authors contributed equally to this work.

  • Rosa Lozano-Durán,

    1. Instituto de Hortofruticultura Subtropical y Mediterránea ‘La Mayora’, Universidad de Málaga-Consejo Superior de Investigaciones Científicas (IHSM-UMA-CSIC), Departamento Biología Celular, Genética y Fisiología, Universidad de Málaga, Campus Teatinos, 29071 Málaga, Spain
    Search for more papers by this author
    • These authors contributed equally to this work.

  • Stéphanie Huguet,

    1. Unité de Recherche en Génomique Végétale (URGV), UMR INRA 1165 – Université d’Evry Val d’Essonne – ERL CNRS 8196, 2 rue G. Crémieux, CP 5708, F-91057 Evry Cedex, France
    Search for more papers by this author
  • Manuel Arroyo-Mateos,

    1. Instituto de Hortofruticultura Subtropical y Mediterránea ‘La Mayora’, Universidad de Málaga-Consejo Superior de Investigaciones Científicas (IHSM-UMA-CSIC), Departamento Biología Celular, Genética y Fisiología, Universidad de Málaga, Campus Teatinos, 29071 Málaga, Spain
    Search for more papers by this author
  • Edgar A. Rodríguez-Negrete,

    1. Instituto de Hortofruticultura Subtropical y Mediterránea ‘La Mayora’, Universidad de Málaga-Consejo Superior de Investigaciones Científicas (IHSM-UMA-CSIC), Departamento Biología Celular, Genética y Fisiología, Universidad de Málaga, Campus Teatinos, 29071 Málaga, Spain
    Search for more papers by this author
  • Eduardo R. Bejarano

    1. Instituto de Hortofruticultura Subtropical y Mediterránea ‘La Mayora’, Universidad de Málaga-Consejo Superior de Investigaciones Científicas (IHSM-UMA-CSIC), Departamento Biología Celular, Genética y Fisiología, Universidad de Málaga, Campus Teatinos, 29071 Málaga, Spain
    Search for more papers by this author

Author for correspondence:
Eduardo R. Bejarano
Tel: +34 952 131 677
Email: edu_rodri@uma.es

Summary

  • Geminiviruses are plant viruses with circular, single-stranded (ss) DNA genomes that infect a wide range of species and cause important losses in agriculture. Geminiviruses do not encode their own DNA polymerase, and rely on the host cell machinery for their replication.
  • Here, we identify a positive effect of the curtovirus Beet curly top virus (BCTV) on the begomovirus Tomato yellow leaf curl Sardinia virus (TYLCSV) infection in Nicotiana benthamiana plants.
  • Our results show that this positive effect is caused by the promotion of TYLCSV replication by BCTV C2. Transcriptomic analyses of plants expressing C2 unveil an up-regulation of cell cycle-related genes induced on cell cycle re-entry; experiments with two mutated versions of C2 indicate that this function resides in the N-terminal part of C2, which is also sufficient to enhance geminiviral replication. Moreover, C2 expression promotes the replication of other geminiviral species, but not of RNA viruses.
  • We conclude that BCTV C2 has a novel function in the promotion of viral replication, probably by restoring the DNA replication competency of the infected cells and thus creating a favourable cell environment for viral spread. Because C2 seems to have a broad impact on the replication of geminiviruses, this mechanism might have important epidemiological implications.

Introduction

Geminiviruses constitute a group of plant viruses with circular, single-stranded (ss) DNA genomes packaged within geminate particles that infect a wide range of plants and cause important losses in agriculture (Rojas et al., 2005; Seal et al., 2006). The family Geminiviridae is classified into four genera, Begomovirus, Curtovirus, Topocuvirus and Mastrevirus, based on their genomic organization, host range and insect vector (Fauquet et al., 2003, 2008). Whitefly (Bemisia tabaci)-transmitted geminiviruses, with either bipartite (like Tomato golden mosaic virus, TGMV) or monopartite (like Tomato yellow leaf curl Sardinia virus, TYLCSV) genomes, are included in the genus Begomovirus, whereas those having monopartite genomes, which are transmitted by leafhopper vectors and infect dicotyledonous plants, are included in the genus Curtovirus, with Beet curly top virus (BCTV) as the type species.

Over the past 20 yr, geminiviruses have emerged as serious constraints to the cultivation of a variety of vegetable crops in various parts of the world. The emergence of geminiviral diseases is often associated with the high genetic variation of the viral populations and the occurrence of mixed infections. Simultaneous infections with distinct viruses have been observed in nature, and may result in unpredictable effects, from disease amelioration to symptom synergy (Hammond et al., 1999). Although both synergism and interference have been reported previously for begomovirus and curtovirus species, most of the effort has been focused on the analysis of the synergism between species of the same genus (Fondong et al., 2000; Briddon & Markham, 2001; Pita et al., 2001; Morilla et al., 2004; Vanitharani et al., 2004; Alves-Junior et al., 2009; Renteria-Canett et al., 2011), the only exception being the synergism reported between geminiviruses and the +ssRNA viruses (Pohl & Wege, 2007; Wege & Siegmund, 2007; Sardo et al., 2011).

Monopartite begomovirus and curtovirus genomes encode six and seven open reading frames (ORFs), respectively. In both cases, the virion sense strand contains two ORFs (V2 and CP), and an extra one (V3) can be identified in the curtovirus genome; the complementary sense strand encodes four ORFs (Rep, C2/L2, C3/L3 and C4/L4). So far, the described functions of the viral proteins include cell cycle manipulation, suppression of host defence (including gene silencing), viral replication, and intra- as well as intercellular trafficking of nucleoprotein complexes (reviewed in Hanley-Bowdoin et al., 2000, 2004; Gutierrez et al., 2004).

Geminiviruses replicate in the nucleus of plant cells through double-stranded (ds) DNA intermediates that can form minichromosomes (Pilartz & Jeske, 1992). Two viral proteins have been found to be involved in viral replication: Rep (also called AL1, L1 and C1), which is the only viral protein essential for this process (Elmer et al., 1988a), and C3 (also called AL3, L3 and REn), which enhances viral DNA accumulation (Sunter et al., 1990; Stanley et al., 1992; Hormuzdi & Bisaro, 1995). Although geminiviruses do not encode their own DNA polymerases and rely on the host DNA replication machinery for their propagation, they are able to replicate in differentiated cells that no longer contain detectable levels of host DNA polymerases and associated factors. To overcome such restriction, geminiviruses must induce the accumulation of the DNA replication machinery in mature plant cells by reprogramming host gene expression. Evidence suggests that the cell cycle in infected plant cells is reprogrammed differently by different geminivirus genera. Some bipartite begomoviruses modify gene expression to allow DNA replication primarily by triggering endoreduplication (Bass et al., 2000; Ascencio-Ibanez et al., 2008), whereas curtoviruses are also able to induce extensive mitosis (Latham et al., 1997). To date, two geminiviral proteins have been shown to play a role in the induction of DNA replication competency in the infected cells: Rep and the curtovirus C4. Rep is a multifunctional protein that interferes with the host cell cycle through its interactions with the plant retinoblastoma-related protein (RBR), an activator of G1/S-phase transition in plant cells (Hanley-Bowdoin et al., 2004), and with essential components of the DNA replisome, such as the proliferating cell nuclear antigen (PCNA) (Castillo et al., 2003; Bagewadi et al., 2004), replication protein A (RPA32) (Singh et al., 2007) or replication factor C (RFC) (Luque et al., 2002). Expression of the curtovirus C4 induces a severe developmental phenotype and extensive cell division in all tissue types examined (Latham et al., 1997; Lai et al., 2008; Mills-Lujan & Deom, 2010). Although the mechanism of action of C4 is unclear, previous studies have suggested that it may regulate the expression of host genes directly or indirectly to control the cell cycle (Lai et al., 2008).

In this work, we identify a positive effect of the curtovirus BCTV on the begomovirus TYLCSV infection using Nicotiana benthamiana 2IRGFP plants (Morilla et al., 2006). Agroinfiltration experiments show that BCTV is able to enhance TYLCSV replication by a novel mechanism which is dependent on C2, but independent of Rep and C4; similar experiments reveal that this feature is not shared by the begomoviral C2. The analysis of plants expressing BCTV C2 unveils an up-regulation of the genes involved in DNA replication and control of the G2/M transition, which are known to be induced on cell cycle re-entry. These results strongly suggest that C2 triggers a re-activation of the cell cycle, restoring the DNA replication competency and therefore creating a cell environment favourable for geminiviral replication.

Materials and Methods

Details of the microorganisms and general methods, transient expression assays, plant materials, immunoblot analysis and PCR conditions are provided in Supporting Information Methods S1, Tables S1, S2.

Plasmids

The constructs expressing TYLCSV Rep (pACS1) and BCTV C2 have been described in Morilla et al. (2006) and Lozano-Duran et al. (2011), respectively. The binary plasmid to express BCTV V2 was obtained by cloning a PCR fragment, amplified with primers V2BCup and V2BClow, into the EcoRV site of pBSSKII+ to yield pV2BC. A BamHI–SalI fragment of pV2BC, containing the complete V2 ORF, was cloned into the BamHI–SalI site of pBINX1 (Sanchez-Duran et al., 2011) to yield pBIV2BC. The binary plasmid to express BCTV C4 was obtained by cloning a PCR fragment, amplified with primers C4BCup and C4BClow, into the EcoRV site of pBSSKII+ to yield pC4BC. An SmaI–KpnI fragment of pC4BC, containing the complete C4 ORF, was cloned into the SmaI–KpnI site of pBINX1 to yield pBIC4BC.

To yield the binary plasmid to express BCTV Rep (named pBINL1BCTV), a fragment of 1.1 kbp from pBIN1.2 (Briddon et al., 1989), containing the BCTV Rep ORF (positions 2906 to 1802), was PCR amplified with primers L1BCup and L1BClow, and cloned into the EcoRV site of pBSSKII to obtain pBL1BCTV. A 1.1-kbp KpnI–XbaI fragment of pBL1BCTV was cloned into pBINX1 to yield pBINL1BCTV. Infective BCTV clones for wild-type (pBIN1.2) and ΔC2 (L2-2), ΔV2 and ΔC4 (L4-1) mutants have been described in Briddon et al. (1989), Stanley & Latham (1992) and Stanley et al. (1992). BCTV C2 mutants C2-2 and C2-3 (L2-2 and L2-3, respectively) have been described in Hormuzdi & Bisaro (1995). Infective TGMV (pGA2 and pGB2) and TYLCSV (pTYA14) clones have been described in Briddon et al. (1989), Bejarano & Lichtnstein (1994) and Morilla et al. (2004).

To obtain the plasmid containing the BCTV 2IRGFP construct, an EcoRI–HindIII fragment from pBINGFP (Morilla et al., 2006), containing the 35S cauliflower mosaic virus (CaMV) promoter, the complete green fluorescent protein (GFP) ORF and the Nos terminator, was cloned into the EcoRI–Hind III site of pGreenII0229 (Hellens et al., 2000) to yield pGGFP. A 1.08-kbp fragment of pBIN1.2 (Briddon et al., 1989) containing the intergenic region (IR) of BCTV (positions 2535 throughout 2993, to 599) was amplified by PCR using primers OBCFHindIII and OBCRHindIII, and cloned into the EcoRV site of pBSSKII (Agilent Technologies, Santa Clara, CA, USA) to yield pBBCIR. A HindIII fragment from pBBCIR, containing the IR of BCTV, was cloned into the HindIII site of pGGFP to obtain pGIRGFP. An EcoRI fragment from pBBCIR, containing the IR of BCTV, was then cloned into the EcoRI site of pGIRGFP to yield pG2IRGFP. The orientation of the IR fragments was determined by PCR using primers pBINX2 and either oBCFHindIII (for the IR cloned into the HindIII site) or oBCFEcoRI (for the IR cloned into the EcoRI site). The CP promoter in the IRs is in the same direction as the 35S CaMV promoter. The binary vector containing the P19 protein of Tomato bushy stunt virus (pBIN61-P19) was kindly provided by Voinnet et al. (2003).

Viral infections

TYLCSV infections of Nicotiana benthamiana L. plants were performed by agroinoculation, as described previously (Elmer et al., 1988b). Seven plants were agroinoculated with pGreenTYA14 (binary vector containing a partial dimer of TYLCSV; Lozano-Duran et al., 2011) and/or pBIN1.2 (binary vector containing a partial dimer of BCTV; Briddon et al., 1989); as a control, two plants were mock inoculated with Agrobacterium tumefaciens harbouring the empty binary vector pGreenII0229 (Hellens et al., 2000) or pBIN (van Engelen et al., 1995). Symptoms were evaluated every week until 21 d post-inoculation (dpi), when samples were taken.

Viral DNA accumulation was quantified by quantitative real-time PCR or Southern blot hybridization. For hybridizations, 10 μg of total plant DNA were used. Membranes were hybridized with TYLCSV radiolabelled probes. Viral DNA accumulation was quantified by Phosphorimager analyses of DNA gel blots and normalized to genomic DNA. DNA probes are described in Table S3.

Viral replication assays were performed by agroinfiltration with Agrobacterium cultures containing infectious clones. For DNA viruses (TYLCSV and BCTV), the same infectious clones as employed in the infection assays were used. Infectious clones of Tobacco mosaic virus and Potato virus X expressing the green fluorescent protein (TMV-GFP and PVX-GFP) were kindly provided by Dr Peter Moffett, and have been described elsewhere (Peart et al., 2002). For Tobacco rattle virus (TRV), clones to deliver TRV RNA1 and RNA2 (pTV00 and pBINTRA6) were used (Ratcliff et al., 2001).

Microarray analysis

Three independent biological replicates were produced. For each biological repetition and each point, RNA samples were obtained by pooling RNAs from 15 plants. Samples were collected on plants at 1.02 developmental growth stages (Boyes et al., 2001), cultivated at 24°C with an 18-h light cycle. Total RNA was isolated from three replicates of control or transgenic Arabidopsis seedlings expressing BCTV C2 using TRIzol (Invitrogen), and subsequently cleaned using an RNeasy MinElute Cleanup Kit (Invitrogen). RNA quantity and quality were assessed with a Nanodrop ND-1000 spectrophotometer (Labtech, Ringmer, East Sussex, UK) and an Agilent 2100 bioanalyser (Agilent Technologies, Santa Clara, CA, USA), respectively.

Microarray hybridization was carried out at the Affymetrix platform at INRA-URGV (Evry, France) using an Affymetrix GeneChip® ATH1. One microgram of total RNA was used to synthesize biotin-labelled cRNAs with the One-Cycle cDNA Synthesis Kit (Affymetrix, Santa Clara, CA, USA). Superscript II reverse transcriptase and T7-oligo (dT) primers were used to synthesize the ss cDNA at 42°C during 1 h, followed by the synthesis of the ds cDNA using DNA ligase, DNA polymeraseI and RNaseH during 2 h at 16°C. Cleanup of the ds cDNA was performed with the Sample Cleanup Module (Affymetrix), followed by in vitro transcription (IVT) in the presence of biotin-labelled UTP using a GeneChip® IVT Labelling Kit (Affymetrix). The quantity of the cRNA labelled with RiboGreen® RNA Quantification Reagent (Turner Biosystems, Sunnyvale, CA, USA) was determined after cleanup by the Sample Cleanup Module (Affymetrix). Fragmentation of 15 μg of labelled cRNA was carried out for 35 min at 94°C, followed by hybridization for 16 h at 45°C to an Affymetrix GeneChip® Arabidopsis genome array representing 22 500 probe sets corresponding to 24 000 gene sequences. After hybridization, the arrays were washed with two different buffers (stringent, 6 × Saline-sodium phosphate-EDTA, 0.01% Tween-20; nonstringent, 100 mM Mes, 0.1 M [Na+], 0.01% Tween-20) and stained with a complex solution including streptavidin R–phycoerythrin conjugate (Invitrogen/Molecular Probes, Carlsbad, CA, USA) and anti-streptavidin biotinylated antibody (Vectors Laboratories, Burlingame, CA, USA). The washing and staining steps were performed in a GeneChip® Fluidics Station 450 (Affymetrix). The Affymetrix GeneChip® Genome Arrays were finally scanned with the GeneChip® Scanner 3000 7G piloted by the GeneChip® Operating Software (GCOS).

For the data analysis, the data were normalized using the GCRMA algorithm (Irizarry et al., 2003), available in the Bioconductor package (Gentleman & Carey, 2002). To determine the differentially expressed genes, we performed a usual two-group t-test that assumes equal variance between groups. The variance of the gene expression per group was a homoscedastic variance, in which genes displaying extremes of variance (too small or too large) were excluded. The raw P values were adjusted by the Bonferroni method, which controls the Family Wise Error Rate (FWER) (Ge et al., 2003). A gene was declared to be differentially expressed if the Bonferroni P value was < 0.05.

The raw CEL files were imported into R software for data analysis. All raw and normalized data are available through the CATdb database (AFFY_L2bctv_Ath; Gagnot et al., 2008) and from the Gene Expression Omnibus (GEO) repository at the National Center for Biotechnology Information (NCBI) (Barrett et al., 2007), accession number GSE24214.

Results

Synergistic infections of BCTV and TYLCSV in N. benthamiana

To analyse potential synergistic interactions between TYLCSV and other geminivirus species, we tested a co-infection between TYLCSV and BCTV in the model plant N. benthamiana. In comparison with the single infections, co-infection of N. benthamiana with BCTV and TYLCSV resulted in increased systemic symptoms at 21 dpi (Fig. 1a), indicating the existence of a positive synergistic relationship between the symptomatologies caused by these two geminivirus species. To study this interaction further, we took advantage of transgenic 2IRGFP N. benthamiana plants (Morilla et al., 2006). These plants harbour a 35S:GFP expression cassette between two direct repeats of the TYLCSV IR that allows the monitoring of TYLCSV replication in vivo. Following viral infection or heterologous expression of TYLCSV Rep protein, these plants display a Rep-dependent GFP overexpression driven by the generation of mGFP replicons (hereafter referred to as ‘mGFP’) (Morilla et al., 2006), which correlates with Rep activity. 2IRGFP plants were co-inoculated with TYLCSV and BCTV, or inoculated with either virus alone as a control. GFP expression was exhaustively monitored at different times post-infection, and samples were collected to determine the accumulation of viral DNA in the infected plants. When the plants were observed under UV light, we noticed that the extension of GFP expression in leaves was clearly higher in those plants co-infected with TYLCSV and BCTV than in those infected with TYLCSV alone, as the number of leaves showing GFP overexpression and the GFP-expressing area in each leaf were larger (Fig. 1b). However, the pattern of GFP overexpression remained unaltered in the presence of BCTV (green fluorescence was restricted to vascular bundles; Fig. 1c). As expected, the analysis of total DNA extracted from the infected plants proved that the increase in GFP correlated with a greater accumulation of TYLCSV DNA (Fig. 1d). No differences in the accumulation of BCTV DNA were detected in the presence of TYLCSV (Fig S1a). Taken together, these results suggest that BCTV promotes TYLCSV accumulation by increasing the number of leaves capable of sustaining its replication, and probably by enhancing the efficiency of the replication itself, but without altering its phloem-limited distribution.

Figure 1.

Synergism between Beet curly top virus (BCTV) and Tomato yellow leaf curl Sardinia virus (TYLCSV) in Nicotiana benthamiana 2IRGFP plants. (a) Representative systemic symptoms in plants inoculated with TYLCSV, BCTV or both. (b) Green fluorescent protein (GFP) expression in leaves from plants infected with TYLCSV or co-infected with BCTV and TYLCSV. Leaves were numbered from the inoculation point, leaf + 1 being the first leaf placed on the agroinfection site. (c) Details of GFP expression in the leaves from plants infected with TYLCSV or co-infected with BCTV and TYLCSV. (d) Relative amount of TYLCSV DNA in plants. Viral DNA was determined by quantitative real-time PCR of total DNA extracted from leaves + 8, + 9 and + 10. Values are the average of 10 plants from two independent experiments. Bars represent standard error. All photographs and samples were taken from plants at 4 wk (TYLCSV) and 3 wk (BCTV) post-infection. Asterisks indicate a statistically significant difference according to Student’s t-test (< 0.05).

The activity of TYLCSV Rep is enhanced by the curtovirus BCTV, but not by the begomoviruses TGMV or Tomato yellow leaf curl virus (TYLCV)

In order to confirm that BCTV promotes TYLCSV replication, we performed a leaf patch test in 2IRGFP plants, comparing the expression of GFP following the agroinfiltration of a binary vector to express TYLCSV Rep or the co-agroinfiltration of this binary vector and an infectious clone of BCTV. In this assay, A. tumefaciens strains harbouring a Rep expression construct or an infectious clone of BCTV were infiltrated alone or co-infiltrated in different patches of the same leaf. At 3 dpi, samples were harvested and photographs were taken. As shown in Fig. 2, co-infiltration with BCTV triggered a clear increase in the Rep-derived overexpression of GFP (Fig. 2a), which correlated with a strong increase in the Rep-assisted production of mGFP replicons (Fig. 2b). To determine whether other geminivirus species might exert a similar effect, we performed similar leaf patch tests using infectious clones of TGMV and TYLCV. Co-agroinfiltration with either TGMV or TYLCV led to a modest increase in the expression of GFP, which did not result from an enhanced production of mGFP replicons, as shown in Fig. 2(b), indicating the specificity of the effect of BCTV on TYLCSV replication. The presence of these viruses in the agroinfiltrated patches was confirmed in all cases using quantitative real-time PCR (Fig. 2c). Intriguingly, although TYLCSV Rep did not affect the accumulation of TGMV or TYLCV, it enhanced the accumulation of BCTV (Fig. 2c).

Figure 2.

Rep-dependent green fluorescent protein (GFP) expression and mGFP production are enhanced by Beet curly top virus (BCTV), but not Tomato golden mosaic virus (TGMV) or Tomato yellow leaf curl virus (TYLCV). (a) Leaves from transgenic 2IRGFP Nicotiana benthamiana plants were co-agroinfiltrated to express TYLCSV Rep, and either an infectious clone of geminiviruses BCTV, TGMV (DNA A and B) or TYLCV, or the binary vector pGA482 (−) as a negative control. (b) DNA was extracted from the agroinfiltrated leaf patches and used to detect mGFP by Southern blotting. Undigested DNA (4 μg per lane) was blotted and hybridized with a specific probe for mGFP. Similar results were obtained in three independent experiments. (c) Relative amount of BCTV, TGMV (DNA-A) or TYLCV DNA in agroinfiltrated leaf patches. Viral DNA was determined by quantitative real-time PCR of total DNA extracted from the infiltrated patches. Values are the average of 10 plants from two independent experiments. Bars represent standard error. Asterisks indicate a statistically significant difference according to Student’s t-test (P < 0.05).

BCTV C2, but not C4 or V2, enhances the activity of geminivirus Rep

As increasing experimental evidence has indicated the important roles of viral silencing suppressor proteins in the emergence of synergistic diseases (Vanitharani et al., 2005; Latham & Wilson, 2008), we decided to analyse whether any of the putative silencing suppressors encoded by BCTV were involved in the observed effect on TYLCSV replication. Although several reports have confirmed the gene silencing suppression activity of C2, recent data obtained in our laboratory have shown that C4 and V2 are also able to restrict post-transcriptional gene silencing (PTGS) in N. benthamiana (A. P. Luna et al., unpublished). Therefore, all three genes were included in this analysis: leaf patch tests were performed on 2IRGFP Nbenthamiana plants with BCTV mutants in C2 (BCTVΔC2), C4 (BCTVΔC4) or V2 (BCTVΔV2) genes (Stanley & Latham, 1992; Stanley et al., 1992). Although co-agroinfiltration with BCTV wild-type, BCTVΔC4 and BCTVΔV2 triggered a strong increase in fluorescence, co-agroinfiltration with BCTVΔC2 did not (Fig. 3a). The increase in green fluorescence correlated with an enhanced production of mGFP replicons (Fig. 3b), whereas the accumulation of TYLCSV Rep protein remained unaltered (Fig. 3c). As reported previously, no differences in DNA accumulation were detected between wild-type and mutant viruses (Fig. S1c) (Hormuzdi & Bisaro, 1995).

Figure 3.

Rep-dependent green fluorescent protein (GFP) expression and mGFP production are enhanced by both wild-type and C4 or V2 Beet curly top virus (BCTV) mutant viruses, but not by a C2 BCTV mutant virus. (a) Leaves from transgenic 2IRGFP Nicotiana benthamiana plants were co-agroinfiltrated to express Tomato yellow leaf curl Sardinia virus (TYLCSV) Rep, and either an infectious clone of BCTV wild-type (WT), C2 (ΔC2), C4 (ΔC4) or V2 (ΔV2) mutant, or the binary vector pGA482 (−) as a negative control. (b) DNA was extracted from the agroinfiltrated leaves and used to detect mGFP by Southern blotting. Undigested DNA (4 μg per lane) was blotted and hybridized with a specific probe for mGFP. (c) Immunoblot analysis of protein extracts from agroinfiltrated leaf patches to detect TYLCSV Rep. The molecular weight of the observed band is 48 kDa. Coomassie staining of the large subunit of Rubisco is shown as loading control. Similar results were obtained in three independent experiments.

The specific role of C2 in enhancing the production of mGFP replicons was confirmed using leaf patch tests on 2IRGFP plants with the three BCTV proteins (C2, C4 and V2) separately. As shown in Fig. 4(a), co-agroinfiltration with a binary vector expressing C2, but not C4 or V2, resulted in a notable increase in the Rep-dependent production of GFP. This increase correlated with an enhanced production of mGFP replicons (Fig. 4b), whereas the accumulation of Rep protein remained unaltered (Fig. 4c). Remarkably, TYLCSV C2 did not have such an effect (Fig. S2). Taken together, these results indicate that C2 is the BCTV protein responsible for the enhancement of TYLCSV Rep-derived production of mGFP replicons.

Figure 4.

Rep-dependent green fluorescent protein (GFP) expression and mGFP production are enhanced by C2, but not by C4 or V2. (a) Leaves from transgenic 2IRGFP Nicotiana benthamiana plants were co-agroinfiltrated to express Tomato yellow leaf curl Sardinia virus (TYLCSV) Rep, and either a binary plasmid containing expression constructs of Beet curly top virus (BCTV) C2, C4 or V2, or the binary vector pGA482 (−) as a negative control. (b) DNA was extracted from the agroinfiltrated leaves and used to detect mGFP by Southern blotting. Undigested DNA (4 μg per lane) was blotted and hybridized with a specific probe for mGFP. (c) Immunoblot analysis of protein extracts from agroinfiltrated leaf patches to detect TYLCSV Rep. The molecular weight of the observed band is 48 kDa. Coomassie staining of the large subunit of Rubisco is shown as loading control. Similar results were obtained in three independent experiments.

In order to test whether C2 could have a similar effect on the replication activity of BCTV Rep, we performed leaf patch tests by co-agroinfiltration of binary vectors expressing C2, C4 or V2 together with BCTV Rep and a BCTV 2IRGFP construct. This construct contains a 35S:GFP expression cassette between two direct repeats of BCTV IR; in the presence of BCTV Rep protein, there was a Rep-dependent production of mGFP replicons and a concomitant GFP overexpression, in a similar manner to that in the TYLCSV 2IRGFP transgenic plants/TYLCSV Rep system described previously (Fig. S3a). The co-infiltration of a binary vector expressing C2 in this system led to a higher production of GFP and mGFP replicons, whereas binary vectors expressing C4 and V2 did not have such an effect (Fig. S3a,b).

Suppression of PTGS is not sufficient to enhance the activity of geminivirus Rep

As BCTV C2 has been described to function as a suppressor of PTGS (Wang et al., 2005), it is possible that the promoting effect of C2 over viral Rep could be related to this activity. In order to determine whether this was the case, we decided to test the effect of the well-known silencing suppressor P19 from the tombusvirus Tomato bushy stunt virus (TBSV) on the Rep-assisted production of mGFP replicons in 2IRGFP plants. As shown in Fig. 5(a,b), co-infiltration with P19 triggered an increase in fluorescence which, unlike in the case of C2, did not correlate with an increased production of mGFP replicons. As expected from its silencing suppressor activity, however, co-expression of P19, but not C2, led to a higher accumulation of Rep (Fig. 5c). Moreover, the simultaneous infiltration with both C2 and P19 led to a further increase in fluorescence, suggesting that these two proteins have an additive effect (Fig. 5a). Taken together, these results suggest that the suppression of PTGS is not sufficient to boost the activity of geminivirus Rep.

Figure 5.

Rep-dependent mGFP production is enhanced by C2, but not by the silencing suppressor P19. (a) Leaves from transgenic 2IRGFP Nicotiana benthamiana plants were co-agroinfiltrated to express Tomato yellow leaf curl Sardinia virus (TYLCSV) Rep, and either a plasmid containing expression constructs of C2 or P19, or the binary vector pGA482 (−) as a negative control. (b) Total DNA was extracted from agroinfiltrated leaf patches and used to determine the relative amount of mGFP replicons by quantitative real-time PCR. The amplification of 25s rRNA was used as the internal control. Values are the mean of three independent experiments with three replicates each. Bars represent standard error. Asterisks indicate a statistically significant difference according to Student’s t-test (< 0.05). All values are normalized to the value in the mock sample, where a background can be detected because of the presence of the transgene. (c) Immunoblot analysis of protein extracts from agroinfiltrated leaf patches to detect TYLCSV Rep. The molecular weight of the observed band is 48 kDa. Coomassie staining of the large subunit of Rubisco is shown as loading control. Similar results were obtained in three independent experiments.

Geminivirus replication is enhanced by BCTV C2

To further confirm that the increase in TYLCSV replication observed in the co-infections is dependent on the effect of BCTV C2, we measured the accumulation of TYLCSV DNA and mGFP replicons in leaf patch tests of 2IRGFP plants agroinfiltrated with infectious clones of TYLCSV and BCTV wild-type or BCTVΔC2 mutant. As shown in Fig. 6, mutation in C2 abolished the enhancement in TYLCSV and mGFP accumulation, demonstrating that C2 is the viral protein responsible for the promotion of TYLCSV replication triggered by BCTV. No differences in DNA accumulation of BCTV were detected among the wild-type and mutant virus (Fig. S1b).

Figure 6.

Beet curly top virus (BCTV)-mediated induction of Tomato yellow leaf curl Sardinia virus (TYLCSV) replication is lost in a C2 BCTV mutant. (a) Leaves from transgenic 2IRGFP Nicotiana benthamiana plants were co-agroinfiltrated with an infectious clone of TYLCSV and either an infectious clone of wild-type (BCTV) or C2 mutant (ΔC2BCTV) BCTV, or the binary vector pGA482 (−) as a negative control. The relative accumulation of TLCSV DNA (b) and mGFP replicons (c) was determined by quantitative real-time PCR on total DNA extracted from agroinfiltrated leaf patches. Values are the mean of three independent experiments with three replicates each. Bars represent standard error. Asterisks indicate a statistically significant difference according to Student’s t-test (< 0.05). All values are normalized to the value in the mock sample, where a background can be detected because of the presence of the transgene.

Based on the observation that BCTV C2 promoted the activity of both BCTV and TYLCSV Rep proteins, and that BCTV enhanced the accumulation of TYLCSV in a C2-dependent manner, we hypothesized that the expression of C2 could have a similar impact on different geminiviruses. With the aim of determining the effect of BCTV C2 on geminivirus accumulation, we performed leaf patch tests by co-agroinfiltration of BCTV C2 together with infectious clones of a curtovirus (BCTV) or monopartite and bipartite begomoviruses (TYLCSV and TGMV) in wild-type N. benthamiana leaves. Co-infiltration with C2 led to a higher viral accumulation, between two- and 13-fold, in all cases (Fig. 7). However, as shown in Fig. S4, C2 did not exert a positive effect on the accumulation of RNA viruses. These results indicate that the mechanism by which BCTV C2 enhances geminiviral accumulation is not species specific, but impacts only DNA viruses.

Figure 7.

C2 induces replication of other geminiviruses. DNA was extracted from Nicotiana benthamiana leaves co-agroinfiltrated with an infectious clone of Tomato yellow leaf curl Sardinia virus (TYLCSV), Beet curly top virus (BCTV) or Tomato golden mosaic virus (TGMV) (DNA A and B) and either a binary plasmid to express BCTV C2 (C2) or the binary vector pGA482 (mock) as a negative control. Samples were collected at 7 d post-inoculation. Viral DNA accumulation was determined by quantitative real-time PCR using specific primers for each virus. Values are the mean of three independent experiments. Bars represent standard error. Asterisks indicate a statistically significant difference according to Student’s t-test (< 0.05).

Transcriptomic analysis reveals an induction of cell cycle genes by BCTV C2

Geminiviral replication relies heavily on the expression of host genes involved in DNA replication and/or cell cycle regulation. According to our results, it would be feasible that the effect of BCTV C2 on geminivirus replication might depend on the promotion of the expression of host genes favourable for geminivirus replication. With the aim of shedding light on the transcriptional changes triggered by BCTV C2, we performed a microarray analysis of transgenic Arabidopsis plants expressing this viral gene (Lozano-Duran et al., 2011). T2 transgenic C2 or control seedlings were grown on plates with kanamycin, samples were harvested and total RNA was extracted after 10 d; three biological and three technical replicates were used. The RNA obtained from each technical replicate in the different biological replicates was pooled and subsequently used for the microarray hybridizations (GEO accession number GSE24214).

As shown in Fig. 8(a), 444 genes were significantly up-regulated and 154 were significantly down-regulated in the transgenic C2 plants. To explore which functional categories were affected by the expression of BCTV C2, gene ontogeny (GO) functional enrichment analysis was performed using the VirtualPlant BioMaps tool (Katari et al., 2010); http://virtualplant.bio.nyu.edu/cgi-bin/vpweb2/). Several functional categories were over-represented in the subsets of up- or down-regulated genes, as shown in Fig. 8(a). Interestingly, among those biological processes over-represented in the subset of up-regulated genes, cell cycle-related categories (cell cycle, regulation of cell cycle, DNA packaging, nucleosome assembly, establishment and/or maintenance of chromatin architecture) can be found. The up-regulated genes in the C2 plants in these functional categories are listed in Table 1. As expected from the existence of nonoverlapping functions between curtovirus C2 and its begomovirus counterpart, the transcriptional changes driven by BCTV C2 were not equivalent to those triggered by TYLCSV C2 (R. Lozano-Durán et al., unpublished; GEO accession number GSE18667), and the intersections between the subsets of up- or down-regulated genes in these transgenic lines were minimal (Fig. 8b).

Figure 8.

Transcriptomic analysis of transgenic Arabidopsis plants expressing Beet curly top virus (BCTV) C2. (a) Functional categorization of differentially expressed genes (either up- or down-regulated) in transgenic Arabidopsis plants expressing BCTV C2. The total numbers of up- and down-regulated genes, as well as the total numbers of genes in each category, are indicated. Nonredundant gene ontology (GO) categories over-represented in each subset of genes are shown. (b) Venn diagrams depicting the intersection between up- and down-regulated genes in transgenic Arabidopsis plants expressing BCTV C2 or Tomato yellow leaf curl Sardinia virus (TYLCSV) C2.

Table 1.   Cell cycle-related genes up-regulated in the transgenic Arabidopsis L2 plants according to the microarray data
GeneFold P valuePeakCell cycle re-entryACC Arabidopsis thaliana
  1. The peak of expression in the cell cycle and the induction on cell cycle re-entry, as described by Menges et al. (2003), are indicated: M, mitosis; S, S phase; UR, up-regulation.

Cell cycle/cell cycle regulation
 FAR1 (alcohol-forming fatty acyl-CoA reductase 1)1.806.24E-4  At5g22500
 CYCB1;4 (cyclin B1;4)1.682.16E-2  At2g26760
 CYCB2;4 (cyclin B2;4)1.735.94E-3MURAt1g76310
 DMC1 (disruption of meiotic control 1)2.301.47E-10  At3g22880
 FAR4 (alcohol-forming fatty acyl-CoA reductase 4)2.301.17E-10  At3g44540
 MAD2 (mitotic arrest-deficient 2)2.202.63E-9MURAt3g25980
 CYCA1;1 (cyclin A1;1)2.106.54E-8MURAt1g44110
 CDKB2;1 (cyclin-dependent protein kinaseB2;1)2.132.30E-8MURAt1g76540
 CDKB2;2 (cyclin-dependent protein kinaseB2;2)1.682.57E-2MURAt1g20930
 CDKB1;2 (cyclin-dependent protein kinaseB1;2)1.728.20E-3  At2g38620
 Kinesin 12B/PAKRP1L (microtubule motor kinesin)1.691.48E-2MURAt3g23670
 ATMAP65-3 (microtubule-associated protein 65-3)2.394.89E-12  At5g51600
 CYCB1;5 (cyclin B1;5)1.959.02E-6MURAt1g34460
 CKS2 (CDK-subunit 2)2.043.34E-7  At2g27970
 CYCA2;4 (cyclin A2;4)1.974.47E-6  At1g80370
Nucleosome assembly/chromatin assembly or disassembly
 Histone H2A2.043.55E-7SURAt3g54560
 Histone H2A1.895.15E-5SURAt3g20670
 Histone H2B1.691.44E-2SURAt3g45980
 Histone H2B2.018.72E-7SURAt3g53650
 Histone H32.088.67E-8  At5g10400
 Histone H31.752.77E-3SURAt3g27360
 Histone H41.691.73E-2SURAt3g45930
 Histone H41.771.79E-3SURAt5g59690
 Histone H41.682.54E-2SURAt1g07820
 NFD4 (nucleosome/chromatin assembly factor D4)1.711.06E-2  At2g17560

To validate the microarray data, we measured the expression of several up-regulated cell cycle-related genes in the transgenic Arabidopsis plants expressing C2 by quantitative real-time PCR, including cyclin-dependent kinases (CDKs) (CDKB1;2, CDKB2;1), cyclines (CYCA1;1, CYCB1;4, CYCB1;5 and CYCB2;4) and histones (HISTONE H2A, HISTONE H2B, HISTONE H3 and HISTONE H4). Proliferating cellular nuclear antigen 1 (PCNA1) was included in the analysis as a control, as the expression of this gene, which is not altered according to the microarray data, is induced during geminiviral infection by a Rep-dependent mechanism (Nagar et al., 1995; Egelkrout et al., 2001); transgenic plants expressing TYLCSV C2 (Lozano-Duran et al., 2011) were used as a control. As shown in Fig. S5, the results confirmed the microarray data for all genes, except for At3g54560, which encodes a Histone H2A. The expression of most of the tested genes was approximately between two- and four-fold higher in plants expressing BCTV C2. Interestingly, no increase in the expression of these cell cycle-related genes was detected in plants expressing TYLCSV C2, with the exception of PCNA1, in which a modest increase could be detected. To determine whether C2 induced similar changes in N. benthamiana, we analysed the expression of three cell cycle-related genes (CDBK2, CYCA1 and CYCA2) in leaves agroinfiltrated with a binary vector to express C2 by semi-quantitative reverse transcription-PCR. As shown in Fig. S6, C2, but not C4 or V2, also induced the up-regulation of all three genes when transiently expressed in N. benthamiana. Most cell cycle-related genes up-regulated in the transgenic Arabidopsis plants expressing BCTV C2, depicted in Table 1, are up-regulated in cell cycle re-entry and present their maximum expression during mitosis or S-phase (Menges et al., 2003). These transgenic plants, however, did not display any obvious phenotype in plant growth or development (Lozano-Duran et al., 2011), indicating that aberrant cellular divisions probably do not occur.

The N-terminal region of C2 is sufficient to induce the expression of cell cycle genes and to promote the activity of geminivirus Rep

In order to gain an insight into the molecular mechanisms by which C2 promotes the induction of cell cycle genes and enhances geminiviral replication, we tested the effect of two different mutated versions of C2, C2-2 and C2-3 (Hormuzdi & Bisaro, 1995), on the activity of TYLCSV Rep and the expression of CDBK2, CYCA1 and CYCA2 in transgenic 2IRGFP N. benthamiana plants. Although the wild-type C2 protein is 173 amino acids in length, C2-2 and C2-3 mutated versions contain premature stop codons, generating truncated proteins of 72 and 13 amino acids, respectively (Fig. 9a). As shown in Fig. 9(b), C2-3 mutant protein did not lead to a higher GFP expression when co-expressed with TYLCSV Rep, but C2-2 did, which correlated with an enhanced production of mGFP replicons (Fig. 9c). Moreover, C2-2, but not C2-3, triggered an over-expression of CDKB2, CYCA1 and CYCA2, similar to that triggered by wild-type C2 (Fig. S7). Interestingly, these results suggest that the N-terminal region of C2, containing only the first four amino acids of the conserved zinc finger domain, is sufficient to trigger an induction of cell cycle genes and the promotion of the activity of geminivirus Rep.

Figure 9.

Rep-dependent green fluorescent protein (GFP) expression and mGFP production are enhanced by C2-2, but not by the C2-3 mutant. (a) Diagram of the Beet curly top virus (BCTV) C2 protein showing residue number, the relative position of the zinc finger-like motif and the stop codon in the C2-2 and C2-3 mutated versions. (b) Leaves from transgenic 2IRGFP Nicotiana benthamiana plants were co-agroinfiltrated to express Tomato yellow leaf curl Sardinia virus (TYLCSV) Rep (Rep), and either wild-type C2, the mutated versions (C2-2, C2-3) or the binary vector pGA482 (–) as a negative control. (c) Total DNA was extracted from agroinfiltrated leaf patches and used to determine the relative amount of mGFP replicons by quantitative real-time PCR. Values are the mean of three independent experiments with three replicates each. Bars represent standard error. Asterisks indicate a statistically significant difference according to Student’s t-test (< 0.05). All values are normalized to the value in the mock sample, where a background can be detected because of the presence of the transgene. (d) Total RNA extracted from the same leaves was used to determine the expression levels of CDKB2 and C2 in the agroinfiltrated leaves by semi-quantitative reverse transcription-PCR. E1Fα was used as the internal control. Results at 25 and 30 cycles (for CDKB2 and C2) and at 30 and 35 cycles (for E1Fα) are shown. (e) Total RNA extracted from the agroinfiltrated leaves was used as a template for PCR before cDNA synthesis. The absence of amplification confirms that there is no DNA contamination. Similar results were obtained in two independent experiments.

Discussion

BCTV exerts a C2-dependent synergistic effect over TYLCSV and other begomovirus species

Nicotiana benthamiana 2IRGFP plants are an excellent tool to detect positive or negative viral synergisms between TYLCSV and other viruses. Using these plants, we have been able to detect a synergistic effect exerted by a curtovirus, BCTV, over a begomovirus (TYLCSV) infection. Co-infection with both viruses results in more severe symptoms and increased TYLCSV DNA accumulation. Subsequent experiments in leaf patch tests, using BCTV C2, C4 and V2 mutants, have proven that this positive synergism on TYLCSV infection is a result of the promoting effect of BCTV C2 over TYLCSV replication, as wild-type BCTV as well as C4 and V2 mutants, but not the C2 mutant, enhanced TYLCSV accumulation. Because we did not perform similar experiments with BCTV mutants in Rep or C3 (as mutations in these genes abolish or reduce viral replication) or in the CP gene, we cannot completely rule out the possibility that any of these genes could, in addition to C2, contribute to this synergistic effect on TYLCSV.

As both viruses replicate in phloem cells and the co-infection of nuclei has been described in geminiviruses (Morilla et al., 2004), it is feasible that these two viruses share similar cell environments, allowing BCTV induction of TYLCSV replication in a natural mixed infection. The synergy described here might have epidemiological implications, given that C2 also exerts its positive effect on all geminivirus species analysed, suggesting a general mechanism to enhance the replication of DNA viruses (Fig. 7). Interestingly, Briddon & Markham (2001) reported a positive effect of BCTV over the spreading of the DNA-A of the begomovirus TGMV in N. benthamiana plants. Although they interpreted their observations as complementation in movement functions, they may have been caused by the replicational support described in this work.

BCTV C2, but not TYLCSV C2, promotes the activity of geminivirus Rep and enhances geminivirus accumulation

Our data show that the curtovirus BCTV, but not begomoviruses, increases the Rep-assisted accumulation of mGFP replicons in N. benthamiana 2IRGFP plants, suggesting that the ability to enhance TYLCSV replication is specific to curtoviruses. This result is supported by the fact that BCTV C2, but not TYLCSV C2, is able to increase the accumulation of mGFP. BCTV C2 shares some functions and sequence similarities with its begomoviral counterpart: (1) both proteins interact with and inactivate SNF1-related kinase (SnRK1) (Sunter et al., 2001; Hao et al., 2003) and the adenosine kinase (ADK) (Wang et al., 2003); (2) both proteins have a conserved zinc finger motif; (3) both proteins produce an enhanced susceptibility phenotype when expressed in transgenic plants (Sunter et al., 2001); and (4) both proteins function as gene silencing suppressors of PTGS and TGS (reviewed in Raja et al., 2010). However, they are also known to have functional differences, as BCTV C2 does not complement a mutation in the TGMV C2 gene (Stanley et al., 1992; Sunter et al., 1994; Hormuzdi & Bisaro, 1995; Baliji et al., 2004, 2007) and, unlike the begomoviral C2, BCTV C2 does not harbour a formal transcriptional activation domain, and thus is considered to lack the transcriptional activation activity required for the expression of late viral genes (Sunter & Bisaro, 1992; Hormuzdi & Bisaro, 1995). Strikingly, comparison between the transcriptome of plants expressing either TYLCSV C2 or BCTV C2 (Fig. 8) unveiled clear differences, confirming the functional divergence between these viral proteins. Although both induce changes in gene expression, the sets of genes altered in each case are mostly dissimilar, suggesting that C2 from begomoviruses and curtoviruses might use different mechanisms to reprogram gene expression in the host cell.

BCTV C2 has a novel function in the re-activation of the cell cycle

The assay used in this work, based on 2IRGFP constructs, allows us to study the replication efficiency of geminiviruses, which is dependent on both the activity of Rep and the availability of the required cell machinery. It has been suggested that geminiviral DNA replication is coupled to changes in the cell cycle stage of the host cell because of the requirement for cellular factors. Transcriptomic analysis of transgenic Arabidopsis plants expressing BCTV C2 unveiled an up-regulation of the genes involved in cell cycle regulation and chromatin structure. Most of these genes have been identified as being up-regulated on cell cycle re-entry from G1-phase after sucrose starvation (Menges et al., 2003). Bearing in mind that geminiviruses usually infect terminally differentiated cells, and thus must trigger the re-activation of the DNA replication machinery, the cell cycle re-entry might be an essential step in the infectious process. Taking together these results, it is tantalizing to speculate that C2 mediates the up-regulation of genes required to re-initiate the cell cycle.

So far, two geminiviral proteins have been implicated in the reprogramming of the cell cycle: Rep from mastreviruses and begomoviruses, through its interactions with RBR (Hanley-Bowdoin et al., 2004), and C4 from curtoviruses (Latham et al., 1997; Mills-Lujan & Deom, 2010). Our results raise the unexpected possibility that BCTV might rely on at least two different proteins (C2 and C4), in addition to Rep, to create a cellular environment favourable for viral replication. The fact that the positive effect of BCTV C2 on viral accumulation is specific to DNA viruses, as it does not trigger dramatic changes in the accumulation of RNA viruses, supports this idea. Although mutations in curtovirus C2 could lead to reduced viral accumulation, lower infectivity and/or a recovery phenotype (Hormuzdi & Bisaro, 1995; Baliji et al., 2007; Lozano-Durán & Bejarano, 2011), C2 is not essential for replication, suggesting that the gene induction triggered by C2 is not required, but favourable, for viral replication.

The use of two mutated versions of BCTV C2 allowed us to determine that the N-terminal region of C2, containing only the first 72 amino acids (C2-2), is sufficient to trigger both the promoting effect over the activity of TYLCSV Rep and the transcriptional changes in N. benthamiana. This region of C2 contains the first four amino acids of the zinc finger domain (C–X–C–X), which has been described in begomoviruses as being essential for the transactivation activity and to mediate homodimerization. However, BCTV C2 has not been found to form homodimers despite the fact that the zinc finger domain is conserved (Yang et al., 2007). The N-terminal region of begomovirus C2 also contains the nuclear localization signal (NLS) that is essential for targeting of the protein to the nucleus (Dong et al., 2003); although this signal is not strictly conserved in curtoviruses, BCTV C2, similar to begomovirus C2, contains a large percentage of positively charged amino acid residues in this region. To date, the subcellular localization of BCTV C2 has not been elucidated; it will be interesting to determine whether BCTV C2, like its begomoviral counterpart, is localized to the nucleus and, if this is the case, whether the diverging NLS is still functional, as suggested by the functionality of the mutant C2-2.

In this work, we describe a novel function for curtovirus C2 in promoting geminiviral replication, most likely through the re-activation of the cell cycle. This novel function resides in the N-terminal part of the protein and seems to have a broad impact on the replication of geminiviruses, which has important epidemiological implications. Future work will aim to dissect the molecular mechanisms underlying this novel role of C2.

Acknowledgements

We thank Mayte Duarte and Silvia Hernández for excellent technical assistance, Miguel A. Sánchez-Durán for generating the anti-Rep antibodies, Peter Moffett and Olivier Voinnet for sharing materials and Javier Ruiz-Albert for critical discussions. This research was supported by grants from the Spanish Ministerio de Ciencia y Tecnología (AGL2007-66062-C02-02/AGR and AGL2010-22287-C02-02) and Fondo Europeo de Desarrollo Regional (FEDER). Z. C. was awarded a Postdoctoral Fellowship JAE-Doc from Consejo Superior de Investigaciones Científicas (CSIC); R.L-D. and M.A-M. were awarded Predoctoral Fellowships from the Spanish Ministerio de Educación y Cultura.

Ancillary