Viral Silencing Suppressors (2015) PDF

Virology 479-480 (2015) 85–103 Contents lists available at ScienceDirect Virology...

Virology 479-480 (2015) 85–103 Contents lists available at ScienceDirect Virology journal homepage: www.elsevier.com/locate/yviro Review viral silencing suppressors: Tools forged to fine-tune host-pathogen coexistence Tibor Csorba n, Levente Kontra, József Burgyán n National Agricultural Research and Innovation Center – Agricultural Biotechnology Center (NARIC – ABC), Gödöllő, Szent-Györgyi A.u.4., Pest 2100, Hungary art ic l e i nf o a b s t r a c t Article history: RNA silencing is a homology-dependent gene inactivation mechanism that regulates a wide range of Received 18 December 2014 biological processes including antiviral defense. To deal with host antiviral responses viruses evolved Returned to author for revisions mechanisms to avoid or counteract this, most notably through expression of viral suppressors of RNA 31 January 2015 silencing. Besides working as silencing suppressors, these proteins may also fulfill other functions during Accepted 16 February 2015 Available online 9 March 2015 infection. In many cases the interplay between the suppressor function and other “unrelated” functions remains elusive. We will present host factors implicated in antiviral pathways and summarize the Keywords: current status of knowledge about the diverse viral suppressors’ strategies acting at various steps of Plants antiviral silencing in plants. Besides, we will consider the multi-functionality of these versatile proteins Viruses and related biochemical processes in which they may be involved in fine-tuning the plant-virus Antiviral RNA interference interaction. Finally, we will present the current applications and discuss perspectives of the use of these Viral suppressors of silencing Defense proteins in molecular biology and biotechnology. Pathogenesis & 2015 Elsevier Inc. All rights reserved. Contents Plant viruses............................................................................................................ 86 RNA silencing pathways in plants........................................................................................... 86 Antiviral silencing host factors.............................................................................................. 86 Initiation of antiviral silencing.......................................................................................... 86 Effector phase of antiviral silencing...................................................................................... 90 Amplification of antiviral silencing....................................................................................... 90 Actions of viral suppressors of RNA silencing.................................................................................. 91 Blocking initiation of antiviral response................................................................................... 91 Arrest of functional RISC assembly through AGO interaction.................................................................. 92 Inactivation of programmed antiviral RISC complex......................................................................... 92 VSR activities downstream of RISC and RITS............................................................................... 93 Modulation of AGO1 homeostasis....................................................................................... 93 Plant RDR-based activity suppression.................................................................................... 93 Targeting multiple steps of antiviral pathways............................................................................. 93 VSR interactions with host factors....................................................................................... 93 Driving factors in VSRs' evolution....................................................................................... 94 The control of pathogen impact on host...................................................................................... 94 Limitation of VSRs’ suppressor strength................................................................................... 94 VSRs as links between RNA-based and protein-based immunity............................................................... 95 Connecting antiviral silencing to hormone signaling........................................................................ 95 VSRs as tools............................................................................................................ 96 Unraveling molecular basis of silencing itself.............................................................................. 96 n Corresponding authors. Tel.: þ 36 28526100; fax: þ 36 28526101. E-mail addresses: [email protected] (T. Csorba), [email protected] (J. Burgyán). http://dx.doi.org/10.1016/j.virol.2015.02.028 0042-6822/& 2015 Elsevier Inc. All rights reserved. 86 T. Csorba et al. / Virology 479-480 (2015) 85–103 VSRs employed in molecular research.................................................................................... 96 VSRs as biotechnological and medical tools................................................................................ 96 Conclusions and perspectives............................................................................................... 96 Acknowledgments........................................................................................................ 97 References.............................................................................................................. 97 Plant viruses or DNA methylation, resulting in transcriptional gene silencing (TGS) of the homologous gene (Creamer and Partridge, 2011). In plants and Plant viruses are amongst the most important pathogens caus- worms the effector step can result in amplification of silencing ing huge economic losses worldwide by reducing crop quality and response involving RNA-dependent RNA polimerases (RDRs) proteins quantity. A better understanding of the viral infection processes and (Mourrain et al., 2000; Dalmay et al., 2000; Sijen et al., 2001; Vaistij plant defense strategies is important for crop improvement. et al., 2002; Voinnet et al., 1998). Amplification of RNA silencing has Based on their genome organization viruses can be classified into been implicated in the spread of an RNA silencing signal, a non-cell- positive-sense-, negative-sense-, double-stranded-RNA viruses and autonomous process (Kalantidis et al., 2008; Schwach et al., 2005). single-stranded or double-stranded DNA viruses (Hull, 2002). Differ- The best studied plant model, Arabidopsis thaliana genome ence in the genome organization implies difference in the replication encodes 4 members of DCLs (DCL1-4) (Bologna and Voinnet, strategy. Generally, the genetic information embedded into the viral 2014), five DRBs (HYL1/DRB1, DRB2, 3, 4, 5) (Hiraguri et al., RNA or DNA encode for a surprisingly restricted number of proteins 2005), 10 AGOs (AGO1-10) (Mallory and Vaucheret, 2010) and that coordinate the infection process. Viral proteins interact with host 6 RDRs (RDR1, 2, 3a, 3b, 3c and 6) (Wassenegger and Krczal, 2006). factors to manipulate biochemical events and molecular interactions These proteins have partially redundant roles and combine with required for the virus replication and movement. Viruses can spread each other to result in divers classes of small RNAs and different within the plants through plasmodesmata on short distance (cell-to- effector outputs of the RNA silencing pathways. The small RNA cell movement) or through phloem (systemic movement). During classes identified in plants include microRNAs (miRNAs), trans- host-pathogen co-evolution a set of complex interactions involving acting small interfering RNAs (ta-siRNAs), natural-antisense RNAs virus attack and host defense has been developed. These include (nat-siRNAs), repeat-associated siRNAs (ra-siRNAs), viral siRNAs hypersensitive reaction (HR) (Mandadi and Scholthof, 2013), systemic (vsiRNAs) and virus-activated siRNAs (vasiRNAs). These classes acquired resistance (SAR) (Kachroo and Robin, 2013), activation of possess specialized roles during development, stress responses, ubiquitin/26S proteasome system (UPS) (Dielen et al., 2010) or RNA heterochromatic silencing, viral infection and host-pathogen inter- silencing (RNA interference, RNAi) (Pumplin and Voinnet, 2013). play, respectively (Bologna and Voinnet, 2014). RNA silencing pathways in plants Antiviral silencing host factors RNA silencing is a fundamental genetic regulatory mechanism Initiation of antiviral silencing conserved in eukaryotic organisms. RNAi can act at transcriptional (Transcriptional Gene Silencing, TGS) or at post-transcriptional The hallmark of the antiviral silencing response is the Dicer- levels (Post-Transcriptional Gene Silencing, PTGS), and has many dependent production of viral siRNAs (vsiRNAs) (Hamilton and diverse roles including developmental regulation, stress response Baulcombe, 1999). In uninfected cells long dsRNA is not detectable, or defense against invading nucleic acids like transposons or however upon virus infection, viral dsRNA molecules of different viruses. The antiviral function of RNA silencing was demonstrated sources becomes available. Highly structured regions of viral single- in plants and invertebrates (Bronkhorst and van Rij, 2014; Pumplin stranded RNAs (ssRNA), replicative intermediates (RI) or overlap- and Voinnet, 2013), however recent reports have further provided ping bidirectional read-through transcripts from DNA virus genome evidence for a similar function in mammals (Cullen et al., 2013; Li may all contribute to vsiRNA production (Aregger et al., 2012; et al., 2013; Maillard et al., 2013). Blevins et al., 2011; Donaire et al., 2008; Molnar et al., 2005). Mechanistically, the RNA silencing process consists of initiation Although the viral dsRNA structures are likely accessible to all of the phase, effector phase and amplification phase. During silencing initia- DCLs, a strong hierarchy exists between them regarding vsiRNA tion double-stranded RNAs (dsRNA) of different origins are processed production: RNA virus infections are mainly affected by DCL4, while by an RNase III type enzyme Dicer (DCR, in plants DICER-LIKE proteins, DCL2 becomes critical in a dcl4 mutant background (Deleris et al., DCLs) into short, 21–24 nt long, small RNA (sRNA) duplexes (Bernstein 2006; Donaire et al., 2008; Garcia-Ruiz et al., 2010; Qu et al., 2008) et al., 2001; Hamilton and Baulcombe, 1999; Hutvagner et al., 2001). (Fig. 1). DCL3 has only a minor role against RNA viruses (Qu et al., DICERs require DOUBLE-STRANDED RNA BINDING (DRB) proteins for 2008; Raja et al., 2008). Recent report suggests additional functional accurate sRNA production (Eamens et al., 2012a,b; Hiraguri et al., diversity between DCL4 and DCL2, such as that DCL2 stimulates 2005). The sRNAs are stabilized at their 30 end by the HUA Enhancer 1 transitivity and secondary siRNA production, while DCL4 is suffi- (HEN1)-dependent methylation (a process found in plants and flies so cient for silencing on its own (Parent et al., 2015). The fact that far (Boutet et al., 2003; Yang et al., 2006) and exported from nucleus silencing suppressors of RNA viruses interfere with DCL3 pathway by HASTY (HST) (Park et al., 2005; Peragine et al., 2004) to be loaded suggests it could have important antiviral gene regulatory functions onto Argonaute proteins (Fagard et al., 2000; Hammond et al., 2001; (Azevedo et al., 2010; Hamera et al., 2012; Lacombe et al., 2010). Liu et al., 2004), the effectors of the RNA-Induced Silencing Complex DCL3 is essential against DNA viruses (Akbergenov et al., 2006) and (RISC) (Lee et al., 2004; Pham et al., 2004; Tomari et al., 2004) or RNA works presumably by inducing DNA methylation (Blevins et al., Induced Transcriptional Silencing complex (RITS) (Ekwall, 2004). 2006; Raja et al., 2014) (Fig. 1). Finally, DCL1 acts as a positive Guided by the sRNA sequence, RISC induces slicing or translational regulator in the production of vsiRNAs by making viral dsRNAs repression of its target RNAs (during PTGS) in a sequence-specific available to other DCLs both in RNA and DNA virus infections manner (Kim et al., 2014), whereas RITS complex causes histone and/ (Blevins et al., 2006; Moissiard and Voinnet, 2006) but also as a  T. Csorba et al. / Virology 479-480 (2015) 85–103 87 Fig. 1. The model of antiviral RNA silencing and the diverse suppressor strategies that interfer with the pathway. Silencing response against RNA viruses is initiated by DCL4 and DCL2 through production of 21–22 nt viral siRNAs and to a lesser extent DCL3 (DCL1 that also has roles in antiviral defense was omitted for simplicity, please see main text). Viral siRNAs are methylated by HEN1 and subsequently loaded into AGOs. AGO1 and AGO2 are the two main effectors, but AGO5 and 7 may also have antiviral roles. Loaded-AGO containing RISC complexes complete the effector step of post-transcriptional gene silencing (PTGS) through viral RNA cleavage or translational inhibition (left side). AGO-sliced products, aberrant RNAs or aborted viral transcripts serve as templates for RDR complexes to amplify the antiviral response. SGS3 and SDE5 cofactors are required for RDR6 activity. (RDR1 and RDR2 also have antiviral functions, omitted for simplicity). RDR activities may be primed by DCL2-derived 22 nt vsiRNAs (middle). In DNA virus infections DCL4 and DCL3-derived viral siRNA initiate PTGS or transcriptional gene silencing (TGS) respectively. (DCL1 recognizes highly folded structures like the CaMV 35S leader, omitted for simplicity). DCL3-generated 24 nt vsiRNA following the HEN1-methylation are loaded into AGO4. AGO4-RITS causes the hypermethylation of viral genome to complete TGS (right side). DNA-virus transcripts are poor templates for RDR-based amplification process. The diverse viral suppressors inhibiting PTGS and/ or TGS are shown in black boxes along the sides of the figure. The dashed lines point to the interference place where the VSRs interact with the antiviral pathway. negative regulator limiting DCL4 and DCL3 through miRNA pathway defense against RNA viruses Turnip crinkle virus (TCV) (Curtin et al., (Azevedo et al., 2010; Qu et al., 2008). DCLs interact with DRBs to 2008) or Turnip yellows mosaic virus (TYMV) (Jakubiec et al., 2012). produce small RNAs. DCL4 partner DRB4 is important in antiviral Evidence show that Cauliflower mosaic virus (CaMV) suppressor P6 88 T. Csorba et al. / Virology 479-480 (2015) 85–103 Table 1 The summary of identified VSRs, their mode of suppression and alternative functions. For suppressors encoded by different viruses within the genus and having the same name (e.g. cucumoviral 2b, potyviral HC-Pro etc.) the alternative functions are displayed in a combined manner. For details on each individual function please see the references within. Genus Species VSR Function Reference Phytoreo RDV PNS10 siRNA binding, RDR6 downregulation Ren et al. (2010) and Cao et al. (2005) virus RGDV PNS11 Unknown, miRNA pathway interference Shen et al. (2012) and Liu et al. (2008) RGDV PNS12 Unknown (nucl. localization) Guo et al. (2011) and Wu et al. (2011) Oryza RRSV PNS6 Unknown Wu et al. (2010) virus Tospo TSWV, GBNV NSs siRNA, miRNA and dsRNA binding, avirulance factor Zhai et al. (2014), Ronde et al. (2014), Schnettler et al. virus (2010) and Goswami et al. (2012) Cucumo CMV, TAV 2b sRNA binding, AGO1 binding, AGO4 binding, aphid Nemes et al. (2014), Du et al. (2014), Gonzalez et al. (2012), virus interaction, RDR6 downregulation, AGO1 downregulation Duan et al. (2012), Diaz-Pendon et al. (2007), Lewsey et al. via miR168 upregulation, downregulation of AGOs and (2007), Ahn et al. (2010), Chen et al. (2008), Zhang et al. DCL1, impairing 50 secondary siRNA genesis, cell-to-cell (2006), Feng et al. (2013), Gonzalez et al. (2012), Hamera movement, interaction with CAT3, induction of host drought et al. (2012), Westwood et al. (2013a,b), Ziebell et al. resistance, SA/JA pathway disruption, HR elicitor (2011), Zhang et al. (2008), Inaba et al. (2011), Lewsey et al. (2010), Várallyay and Havelda (2013), Zhou et al. (2014), Ji and Ding (2001) and Li et al. (1999) Ilarvirus AV-2 2b Unknown (no local suppression) Shimuran et al. (2013) Como CPMV CP (S) Unknown Canizares et al. (2004) virus Nepo ToRSV CP Mediating AGO1 degradation Karran and Sanfacon (2014) virus Rymo AgMV, HoMV Hc-Pro unknown Young et al. (2012) virus Poty TEV, ZYMV, PVY, Hc-Pro ds-siRNA binding, blocking HEN1 methyltransferase, HEN1 Sahana et al. (2014), Torres-Barcelo et al. (2010), –Ruiz et virus TuMV, SCMV, SMV, binding, blocking primary siRNA biogenesis by RAV2 al (2010), Torres-Barcelo et al. (2008), Goto et al (2007), TVY, PVA, PRSV interaction, RDR6 downregulation, impairs 30 secondary Shiboleth et al. (2007), Yu et al. (2006), Merai et al. (2006), siRNA genesis, JA pathway disruption, SAHH-interaction, Dunoyer et al. (2004), Llave et al. (2000), Mallory et al. Hip2 interaction, PaCRT interaction, ferredoxin-5 (2001), Lozsa et al. (2008), Ebhardt et al. (2005), Jamous interaction, NtMinD interaction, interfering with miRNA et al. (2011), Endres et al. (2010), Zhang et al. (2008), Seo pathway, interference with host gene expression levels, et al. (2010), Westwood et al. (2014), Canizares et al. AGO1 downregulation via miR168 upregulation (2013), Haikonen et al. (2013a), Shen et al. (2010), Cheng et al. (2008), Jin et al. (2007), Chapman et al. (2004), Kasschau et al. (2003), Soitamo et al. (2011) and Várallyay and Havelda (2013) PPV HcPro-P1 Unknown Valli et al. (2006) PVA VPg SGS3 interaction Rajamäki et al. (2014) and Rajamäki and Valkonen (2009) Tritimo WSMV, ONMV P1 Unknown Young et al. (2012) viruses Poace TriMV, SCSMV P1 Unknown Tatineni et al. (2012) virus Ipomo SPMMV P1 AGO binding Szabo et al. (2012) and Giner et al. (2010) virus CVYV P1b 21 and 22 nt siRNA binding, protease activity, putativ Zn- Valli et al. (2011) and Valli et al. (2008) finger Tombus CymRSV, CIRV, TBSV P19 ds-sRNA binding, interfering with sRNA 3' methylation, Law et al. (2013), Rawlings et al. (2011), Cheng et al. virus mir168 upregulation mediated AGO1 downregulation, HR (2009), Xia et al. (2009), Koukiekoloa et al. (2007), Merai elicitor, Hin19 interaction, ALY interaction et al. (2006), Lakatos et al. (2006), Omarov et al. (2006), Havelda et al. (2005), Dunoyer et al. (2004), Silhavy et al. (2002), Ye et al. (2003), Vargason et al. (2003), Lozsa et al. (2008), Yu et al. (2006), Chapman et al. (2004), Varallyay et al. (2010, 2014), Várallyay and Havelda (2013), Angel and Schoelz (2013), Hsieh et al. (2009), Park et al. (2004) and Uhrig et al. (2004) CNV P20 Unknown Hao et al. (2011) Aureus PoLV P14 dsRNA binding Merai et al. (2005, 2006) virus Carmo TCV P38 AGO1 and 2 binding, DCL1 upregulation to antagonize DCL4 Azevedo et al. (2010), Jin and Zhu (2010), Várallyay and virus and DCL 3, dsRNA binding, blocking primary siRNA Havelda (2013), Zhang J. et al. (2012), Merai et al. (2006), biogenesis by RAV2 interaction, TIP-interaction, DRB-HRT Endres et al. (2010), Donze et al. (2014), Choi et al. (2004), mediated HR elicitor, AGO1 downregulation via miR168 Ren et al. (2000), Zhu et al. (2013, 2014), Jeong et al. upregulation (2008) and Pérez-Cañamás and Hernández (2015) PFVB, HCRSV, PLPV P37 siRNA binding Martinez-Turino and Hernandez (2009), Meng et al. (2006) and Pérez-Cañamás and Hernández (2015) MNSV P7B Unknown, movement Protein Genovés et al. (2011) and Genoves et al. (2006) MNSV P42 Unknown Genoves et al. (2006) Diantho RCNMV replication Unknown (DCL1 dependent), miRNA pathway interference Takeda et al. (2005) virus RCNMV MP Unknown Powers et al. (2008) Clostero BYV P21 ds-sRNA binding, blocking HEN1 methyltransferase Merai et al. (2006), Yu et al. (2006), Reed et al. (2003) and virus Chapman et al. (2004) CTV P20 Unknown Lu et al. (2004)  T. Csorba et al. / Virology 479-480 (2015) 85–103 89 Table 1 (continued ) Genus Species VSR Function Reference CTV P23 Unknown (nucleolar localization) Ruiz-Ruiz et al. (2013) and Lu et al. (2004) CTV CP Unknown Lu et al. (2004) Crini SPCSV, SPVD RNase3 Endonuclease activity Cuellar et al. (2009) and Kreuze et al. (2002, 2005) virus SPCSV P22 Unknown Kreuze et al. (2005) CYSDV P25 Unknown (down stream of siRNA biogenesis) Kataya et al. (2009) ToCV P22 Unknown (local silencing) Canizares et al. (2008) ToCV CP Unknown, SAHH-interaction Canizares et al. (2008, 2013) ToCV Cpm Unknown Canizares et al. (2008) Ampelo GLRaV-3 P19,7 siRNA and miRNA pathway interference Gouveia and Nolasco (2012) and Gouveia et al. (2012) virus Polero BWYV, PLRV, CYDV, P0 Destabilizing AGOs, HR elicitor Hendelman et al. (2013), Fusaro et al. (2012), Derrien et al. virus BMYV, TuYV, MABYV, (2012), Kozlowska-Makulska et al. (2010), Han et al. ScYLV, CLRDV, CABYV (2010), Csorba et al. (2010), Mangwende et al. (2009), Bortolamiol et al. (2007), Pazhouhandeh et al. (2006), Delfosse et al. (2014) and Wang et al. (2014a,b) Enamo PEMV-1 P0 Destabilizing AGO1 Fusaro et al. (2012) virus Tymo TYMV P69 Unknown (inhibition of DNA methylation, miRNA and DCL1 Chen et al. (2004) virus upregulation) CVB P12 Unknown (ZF domain dependent but NLS independent) Lukhovitskaya et al. (2014) Potex PVX, PlAMV, AV3, P25 AGO1 degradation, coaggregation with SGS3/RDR6, CAT1 Chiu et al. (2010), Yan et al. (2012), Senshu et al. (2009), virus WclMV, TVX, PepMV (TGBp1) interaction, X-Body organization Okano et al. (2014), Mathioudakis et al. (2013) and Tilsner et al. (2012) PepMV CP Unknown (blocks systemic signaling) Mathioudakis et al. (2014) Carla PVM TGBp1 Unknown (no local suppression) Senshu et al. (2011) virus PVM, SPCFV, PlAMV CRP Unknown (local and systemic silencing) Deng et al. (2014), Senshu et al. (2011) and Okano et al. (NaBp) (2014) Tricho ACLSV P50 Unknown (no local suppression) Yaegashi et al. (2007, 2008) virus Vitivirus GVA P10 siRNA binding Zhou et al. (2006) Citri CLBV MP Unknown Renovell et al. (2012) virus Tobamo TMV, YoMV(ORMV) P126 Unknown, interference with HEN1-mediated methylation, Wang et al. (2012), Vogler et al. (2007), Ding et al. (2004) virus accumulation of novel miRNA-like sRNAs and Hu et al. (2011) ToMV P130 siRNA binding Kubota et al. (2003) TMV P122 siRNA and miRNA binding, AGO1 downregulation via Csorba et al. (2007) and Várallyay and Havelda (2013) miR168 upregulation Tenui RHBV, RSV NS3 Binding to RNA/RNA or RNA/DNA duplex, larger than 9 nt Hemmes et al. (2007), Shen et al. (2010) and Xiong et al. virus and long ssRNA (2009) RSV P2 Interaction with SGS3 Du et al. (2011) Tobra TRV 16K Downstream of dsRNA biogenesis Andika et al. (2012), Ghazala et al. (2008), Martínez-Priego virus et al. (2008) and Reavy et al. (2004) TRV 29K Unknown Deng et al. (2013) PepRSV 12K Unknown Jaubert et al. (2011) Furo SBWMV, CWMV 19K Unknown, interaction with N-ext/CP Sun et al. (2013a,b) and Te et al. (2005) virus Peclu PCV P15 siRNA binding, miRNA pathway interference Merai et al. (2006), Dunoyer et al. (2002, 2004) virus Beny BNYVV, BSBMV P14 Unknown Andika et al. (2012), Guilley et al. (2009), Chiba et al. virus (2013) and Kozlowska-Makulska et al. (2010) BdMoV P13 Unknown Andika et al. (2012) and Guilley et al. (2009) BNYVV p31 Unknown, interaction with PR-10 Rahim et al. (2007) and Wu et al. (2014) Hordei BSMV, PSLV ɣB siRNA binding Merai et al. (2006) and Yelina et al. (2002) virus Sobemo CfMV, RYMV P1 Unknown (siRNA binding ruled out), dcl4 dep.21 nt siRNA Lacombe et al. (2010), Gillet et al. (2013), Sarmientoa et al. virus delocalization, reduction of 24 nt siRNAs, Zn-finger like (2007) and Weinheimer et al. (2010) binding CfMV CP Unknown, nuclear localization Olspert et al. (2010, 2014) Nucleo RYSV P6 Blocking RDR6-mediated secondary siRNA biogenesis Guo et al. (2013) rhabdo virus Curto BCTV L2 Methylation interference by ADK inhibition Buchmann et al. (2009), Raja et al. (2008), Yang et al. virus (2007), Wang et al. (2003, 2005) and Hao et al. (2003) Begomo ACMV AC4 ss-sRNA binding Chellappan et al. (2005) virus CaLCuV, MYMV, ACMV AL2/ AC2 Methylation interference by ADK inhibition, tran activation of host genes, interference with host gene expression levels 90 T. Csorba et al. / Virology 479-480 (2015) 85–103 Table 1 (continued ) Genus Species VSR Function Reference Buchmann et al. (2009), Raja et al. (2008), Yang et al. (2007), Wang et al. (2003, 2005), Hao et al. (2003), Trinks et al. (2005) and Soitamo et al. (2012) EACMCV AV2 Unknown (downstream of siRNA biogenesis) Chowda-Reddy et al. (2008) BSCTV, TYLCV-C, C2 Methylation interference by SAMDC1 degradation Zhang et al. (2011), Wezel et al. (2002), Sharma et al. AYVV, ToLCJAV (2010) and Kon et al. (2007) AYVV, CLCuMV,TYLCV, V2 Interaction with SGS3, HR elicitor, interaction with PLCPs Zhang J. et al. (2012), Glick et al. (2008), Sharma et al. ToLCJV-A (2010), Sharma and Ikegami (2010), Bar-Ziv et al. (2012), Zrachya et al. (2007) and Amin et al. (2011) AYVV C4 Unknown Sharma et al. (2010) ToLCJB, ToLCJAV, βC1 Unknown, rgs-CAM mediated RDR6 disruption, SAHH Sharma et al. (2011), Kon et al. (2007), Cui et al. (2005), TYLCCNV, TYLCCNV- inhibition Einia et al. (2012), Li et al. (2014) and Yang et al. (2011), Y10, TbCSV-Y35, Amin et al. (2011) CLCuMV, GDARSLA, GmusSLA Alpha-Rep Unknown Nawaz-ul-Rehman et al. (2010) Mastre WDV Rep siRNA binding Wang et al. (2014a,b) and Liu et al. (2014) virus WDV RepA Unknown Liu et al. (2014) Caulimo CaMV P6 (TAV) DRB4 inactivation, interference with NPR1 (SA/JA crosstalk Laird et al. (2013), Haas et al. (2008), Shivaprasad et al. virus regulator), cell-to-cell movement, translational trans (2008), Love et al. (2007, 2012) and Rodriguez et al. (2014) activation CaMV Decoy Overloading DCL capacity Blevins et al. (2011) RNA protein binds to and inhibit DRB4 activity strongly suggesting for an was shown to be important in defense against CMV, TCV and antiviral role of DRB4 upon DNA virus infections (discussed latter). Potato virus X (PVX) viruses in A. thaliana (Harvey et al., 2011; DCL3 cooperates with DRB3 (and AGO4) in antiviral defense Jaubert et al., 2011) and against Tomato bushy stunt virus (TBSV) through genome methylation against DNA viruses Cabbage leaf curl infection in Nicotiana benthamiana (Scholthof et al., 2011). In this virus (CaLCuV) and Beet curly top virus (BCTV) (Raja et al., 2014). scenario AGO1 is both a sensor of the infection that activates HEN1 is also required for the antiviral defense through methylation antiviral pathways such as AGO2 activity but also a direct effector and stabilization of vsiRNAs (Vogler et al., 2007). Indeed, Hen1 of silencing. The phenotype of ago1ago2 double mutant indicates mutants are more susceptible to Cucumber mosaic virus (CMV) and that the two proteins act in a synergistic manner and have non- TCV virus infections (Boutet et al., 2003; Zhang J. et al., 2012). overlapping functions, as supported by their phylogenetic distance (Mallory and Vaucheret, 2010; Wang et al., 2011). Besides AGO1 Effector phase of antiviral silencing and AGO2, AGO5 and AGO7 were also proposed to possess antiviral activities against RNA viruses (Qu et al., 2008; Takeda In theory the initiation phase of silencing could be enough to et al., 2008). AGO7 seems to work as a surrogate of AGO1 but with limit virus replication and spread through processing of the viral a preference for the less structured RNA targets (Qu et al., 2008; RNA into vsiRNAs. Dicing per se, however, is not sufficient to Takeda et al., 2008). The nuclear localized AGO4 has been shown restrict or limit the viral infection (Wang et al., 2011), suggesting to possess important antiviral functions against geminiviruses. that the DCLs’ substrates may be the byproducts or aborted dcl3, drb3 and ago4 mutants fail to hypermethylate the viral transcripts of the viral replication process. It has been shown that genome that is required for host recovery (Raja et al., 2014). the effector step of silencing involving AGO-dependent activity is Besides, AGO4 was proposed be important in transcriptional required to restrict virus replication and spread (Wang et al., 2011). regulation of host transcriptional response during CMV virus Indeed, AGO proteins are essential in antiviral defense against both infection (Hamera et al., 2012). The knowledge about RISC cofac- RNA and DNA viruses (Azevedo et al., 2010; Pantaleo et al., 2007; tors that cooperate with AGOs in plants is very limited. Heat shock Qu et al., 2008; Raja et al., 2008, 2014; Carbonell et al., 2012; protein 70 and 90 (HSP70, HSP90) have been found to be Harvey et al., 2011; Wang et al., 2011). sRNA loading into AGOs is important players in AGO loading by using an in vitro cell-free governed mostly by their 50 terminal nucleotides but length, system that recapitulates the loading process (Iki et al., 2010). thermodynamical properties of sRNA duplex ends and duplex Further understanding of RISC components, assembly and function structure are also important factors (Mi et al., 2008; Schuck may be helped by this and similar in vitro systems (Iki et al., 2010; et al., 2013; Schwarz et al., 2003; Zhang et al., 2014). ago1 and Schuck et al., 2013). ago2 mutants are hypersusceptible to virus infections like CMV, Turnip mosaic virus (TuMV) or TCV (Carbonell et al., 2012; Harvey Amplification of antiviral silencing et al., 2011; Morel et al., 2002). AGO1 participate in removal of viral RNA through slicing activity (Carbonell et al., 2012), although In many plant-virus combinations RDR activities contribute to translational repression activity was also found to play a role the amplification of the antiviral response in order to achieve a (Ghoshal and Sanfacon, 2014) (Fig. 1). It was shown that during robust defense response (Bologna and Voinnet, 2014; Wassenegger RNA virus infections AGO1 homeostasis (Mallory et al., 2008) is and Krczal, 2006) (Fig. 1). RDR polymerase activity is stimulated by disrupted and AGO1 protein levels are decreased probably through the presence of aberrant RNAs lacking bona fide features like cap or translational repression of AGO1 mRNA by miR168 activity polyA tail (Gazzani et al., 2004; Moreno et al., 2013). In some host- (Varallyay et al., 2010). As AGO1 is the negative regulator of virus interactions (like during VSR-deficient CMV infection) sliced AGO2 through miR403 action in absence of AGO1 activity AGO2 products of AGO1 or AGO2 were not required for secondary vsiRNA levels are elevated (Azevedo et al., 2010; Harvey et al., 2011). AGO2 production (Wang et al., 2011) suggesting that the process may be therefore emerges as a second layer in antiviral pathways. AGO2 different from that of ta-siRNA biogenesis. RDR-derived dsRNAs are  T. Csorba et al. / Virology 479-480 (2015) 85–103 91 processed by DCL4 and DCL2 into 21–22 nt long vsiRNAs, respec- its helicase activity. Furthermore, SDE3 could increase the efficiency tively. Both 21 and 22 nt long vsiRNA are effective in antiviral of AGO-targeting of ssRNA targets (Garcia et al., 2012). response as has been shown in case of many virus infections (CMV, Oilseed rape mosaic virus (ORMV), TCV, Tobacco rattle virus (TRV), CaLCuV, CaMV) (Xie et al., 2004; Bouché et al., 2006; Deleris et al., Actions of viral suppressors of RNA silencing 2006; Blevins et al., 2006; Donaire et al., 2008). DCL2 activity becomes more pronounced under higher temperature (Zhang X. Some viruses avoid RNA silencing by replicating within well- et al., 2012b). However, Wang et al. (2011) have found that the 21- defined subcellular compartments/structures like ER spherules but not 22-nucleotide long vsiRNAs guide efficient silencing (Schwartz et al., 2002) or by replicating and moving fast enough to through AGO1 and AGO2 effectors in CMV infection although outrun the mobile silencing signal. However, the most common way AGO1 efficiently incorporates both of them. 22 nt long vsiRNAs to protect viral genome against RNA silencing-mediated inactivation is contribute to secondary siRNA production, as was shown for 22 nt to encode proteins that act as suppressors of RNA silencing (viral long miRNAs and ta-siRNAs or mediate systemic silencing (Garcia- suppressors of RNA silencing, VSRs) (Lakatos et al., 2006). In fact, the Ruiz et al., 2010; Wang et al., 2011). It is assumed that the secondary strongest support of RNA silencing having antiviral roles was the vsiRNA are able to move on short and long distances within the discovery of VSRs. In an early work Ding and coworkers identified the plant to immunize distant tissues ahead of the viral infection, cucumoviral 2b protein as responsible for induction of a non- however this awaits experimental validation. The requirement of conventional virus synergistic disease. Although at that time it was RDR6-activity for systemic movement of a silencing signal suggest- thought that 2b facilitate hypervirulence through its movement ing that RNA silencing amplification has antiviral roles in uninfected protein function (Ding et al., 1996), latter it was demonstrated to be distant tissues. RDR1, RDR2 and RDR6/SDE1/SGS2 were all found to a potent VSR (Brigneti et al., 1998). Another hint regarding viral be crucial factors in secondary vsiRNA production during RNA virus suppression of plant RNA-based defense was the synergistic interac- (PVX, CMV, Tobacco mosaic virus (TMV), Sugarcane mosaic virus tion of potyviruses with other viruses that relayed on potyviral Hc-Pro (SCMV), TuMV, TRV infections (Diaz-Pendon et al., 2007; Donaire protein activity (Pruss et al., 1997). Whilst Hc-Pro was shown to block et al., 2008; Garcia-Ruiz et al., 2010; Qu et al., 2008; Schwach et al., PTGS at tissue level, CMV 2b was shown to prevent systemic silencing 2005). Host RDR involvement in secondary vsiRNA production (Brigneti et al., 1998). Since these first observations more than a could not be verified in certain virus-host interactions. Indeed, decade ago, numerous VSRs were discovered and characterized upon tombusvirus infections the overwhelming part of vsiRNA (Table 1). Available evidences suggest that most viruses encode at derive from the positive RNA strand of the virus genome suggesting least one VSR that in most cases is essential for successful viral life that they are primary DCL products (Aregger et al., 2012; Blevins cycle. Although plant virus’ VSRs are more studied, silencing suppres- et al., 2011; Donaire et al., 2008; Molnar et al., 2005; Szittya et al., sion has been documented on insect and fungus-infecting viruses as 2010). In case of DNA viruses, viral transcripts appear to be poor well (Bronkhorst and van Rij, 2014). The extraordinary diversity in templates of RDRs. Majority of viral siRNAs accumulating during sequence and structure of VSRs within and across kingdoms indicates CaLCuV geminivirus infection were RDR1/2/6-independent primary that they have evolved independently. Diversity of VSRs implies siRNAs (Aregger et al., 2012). diverse mechanistic activities, and indeed, VSRs were shown to block RDR6 activity is facilitated by protein cofactors like Suppressor of virtually all steps of RNA silencing (Fig. 1) such as dicing, effector GENE SILENCINIG 3 (SGS3) (Mourrain et al., 2000), SILENCING assembly, targeting, amplification, transcriptional regulation of endo- DEFECTIVE 5 (SDE5) (Hernandez-Pinzon et al., 2007) and SILEN- genous factors that control RNA silencing and its connections with CING DEFECTIVE 3 (SDE3) (Dalmay et al., 2001). SGS3, a plant protein-based immunity and hormone signaling. specific protein, was found associated with RISC complex and to be important in ta-siRNA biogenesis (Allen et al., 2005; Yoshikawa Blocking initiation of antiviral response et al., 2005). It was proposed that SGS3 stabilize the RISC-cleavage product following slicing and enhance its conversion into dsRNA by One strategy used by VSRs is to hinder mounting of antiviral RDR6 activity (Yoshikawa et al., 2013). Elimination of SGS3 not only silencing by blocking the silencing initiation step. This can be abolishes ta-siRNA biogenesis but also leads to enhanced suscept- achieved through multiple ways like dicer protein or co-factor activity ibility to infection, at least in certain virus-host combinations. sgs3 inhibition, dsRNA/siRNA-sequestration or AGO protein destabilization mutants have severe symptoms when challenged with CMV but do prior of RISC assembly. A widespread suppressor strategy is the ds- not show any difference in their viral symptoms during TuMV or siRNA sequestration that is used by several VSRs encoded by diverse Turnip vein-clearing virus (TVCV) infections (Adenot et al., 2006; virus genera (P19, Hc-Pro, P21, p15, p130/p126/p122, γB, NS3, Pns10, Yoshikawa et al., 2013). SGS3 was shown to be required for CaLCuV NSs etc.) (Csorba et al., 2007; Harries et al., 2008; Hemmes et al., virus induced VIGS of endogenous genes and further suggested that 2007; Kubota et al., 2003; Lakatos et al., 2006; Merai et al., 2005, may be involved in the antiviral response against DNA viruses 2006; Silhavy et al., 2002) (Table 1 and Fig. 1). Probably the most (Muangsan et al., 2004). This is supported by the fact that Tomato characterized siRNA binder is the tombusviral p19 protein (Silhavy yellow leaf curl virus (TYLCV) encodes a VSR to compromise its et al., 2002). Crystallographic studies have shown that p19 head-to- activity (discussed below) (Glick et al., 2008; Fukunaga and Doudna, tail homodimer acts as a molecular caliper to size-select and 2009). SDE5 is an RNA trafficking protein homologue of human sequester siRNA duplexes in a sequence-independent manner mRNA export factor. It was proposed that SDE5 acts together with (Silhavy et al., 2002; Vargason et al., 2003). P19 siRNA sequestration RDR6 in converting specific ssRNAs into dsRNA. sde5 mutant plants prevents RISC assembly as shown by the heterologous in vitro are hypersusceptible to CMV but not TuMV infection (Hernandez- Drosophila embryo extract system (Lakatos et al., 2006). Similarly Pinzon et al., 2007). The amplification process is facilitated by the p19 sequestration of miRNAs is efficient to prevent RISC-loading in SDE3 an RNA-helicase like protein. SDE3 was shown to associate to p19-transgenic A. thaliana (Schott et al., 2012) and N. benthamiana AGOs through its GW motifs (Garcia et al., 2012). sde3 mutant plants plants (Kontra and Burgyan unpublished results). However, during are affected in defense response against CMV or PVX but not TRV authentic virus infections miRNA sequestration by p19 is not efficient infections (Dalmay et al., 2001). SDE3 activity occurs downstream to (Lozsa et al., 2008), suggesting that, miRNA binding by P19 may RDR6 and requires AGO1 and AGO2 activities (Garcia et al., 2012). depend on spatial and temporal co-expression of miRNA duplex and SDE3 was proposed therefore to facilitate the amplification process the virus encoded suppressor protein (Lozsa et al., 2008; Schott et al., by unwinding a fraction of RDR6-sythetized dsRNA products using 2012). A consequence of siRNA binding by VSRs is the block of HEN1- 92 T. Csorba et al. / Virology 479-480 (2015) 85–103 dependent methylation of sRNAs (Csorba et al., 2007; Lozsa et al., Inactivation of programmed antiviral RISC complex 2008; Vogler et al., 2007), however this also depends on the co- expression of sRNA and the suppressor (Lozsa et al., 2008). In addition VSRs may mimic cellular protein cofactors to inactivate pro- to blocking silencing activation at cellular level, p19 interferes with grammed RISC. In Sweet potato mild mottle ipomovirus (SPMMV) the the systemic spread of mobile silencing signal as well (Dunoyer et al., role of suppressor is played by P1 protein, a serine protease, in spite 2010; Molnar et al., 2010). This latter characteristic is an excellent of the presence of HC-Pro (the suppressor of Potyviruses). P1 indicator whether a VSR acts indeed as a bona fide siRNA sequester interacts directly with siRNA and/or miRNA-loaded AGO1 present during authentic viral infection. Some VSRs bind dsRNAs in a size- in the high molecular weight holo-RISC but not minimal-RISC independent manner: Pothos latent aureusvirus (PolV) P14, TCV p38 through GW/WG-motifs (AGO-hook). The conserved GW/WG-motif and CMV 2b (Deleris et al., 2006; Goto et al., 2007; Merai et al., 2005) containing protein family (GW182 family) has been shown to bind have been all described to bind dsRNA and suggested therefore to to AGOs and to be required for diverse RISC function (Eulalio et al., block vsiRNA maturation. In turn nuclear localized P6 suppressor of 2009). Site-directed mutagenesis in the P1 protein proved that the CaMV diminish dicing efficiency through protein-protein interaction. conserved AGO-hook motifs (three GW/WG domains) located at the The two importin-alpha dependent nuclear localization signals of P6 N-terminal part of P1 are absolutely necessary for both binding and are mandatory for CaMV infectivity. P6 genetically and physically suppression of AGO1 function (Giner et al., 2010). SPMMV P1 not interacts with the nuclear DRB4, a cofactor required for DCL4- just inhibits de novo RISC assembly but also block si/miRNA-loaded dependent vsiRNA processing (Haas et al., 2008). (TCV p38, CMV 2b RISC activity. The role GW/WG-motif in silencing suppression was and CaMV P6 are discussed latter). Similar strategy was described in further evidenced by the restoration of naturally inactive P1 protein insect infecting viruses: Flock house virus, Nodamura virus and Wuhan suppressor activity through the introduction of two additional GW/ nodavirus encoded B2, Drosophila C virus (DCV) encoded canonical WG motifs (Szabo et al., 2012). dsRBD, the VP3 proteins of Drosophila X virus or the dsDNA virus Coat protein of TCV also known as p38 may suppress silencing Invertebrate iridescent virus 6 (IIV-6)-encoded 340R suppressors can at multiple levels including vsiRNA generation and assembled RISC block Dicer-2 activity and prevent RISC loading via size-independent activity block. Since p38 possesses dsRNA-binding activity, and in dsRNA/siRNA-binding (Bronkhorst and van Rij, 2014). its presence siRNAs are undetectable it was proposed that p38 acts A completely different strategy (to siRNA-binding) but with to suppress Dicer's activity (Qu et al., 2003). Genetic evidences very similar outcome is used by Sweet potato chlorotic stunt supported the role of p38 in inhibiting DCL4 but not DCL2 (Deleris crinivirus (SPCSV) suppressor RNase3. In this case the build-up of et al., 2006). Although 22 nt vsiRNA were produced, they were an efficient antiviral silencing complex is prevented by endonu- inactive in the presence of p38 because Δp38 mutant TCV was not clease activity of RNase3 that cleaves the 21  24 nt vsiRNAs into able to infect systemically the single dcl, dcl2dcl3 or dcl3dcl4 14 bp products rendering them inactive (Cuellar et al., 2009; double mutant plants except dcl2dcl4 mutant (Deleris et al., Kreuze et al., 2005). Insect infecting Heliothis virescens ascovirus- 2006). In a later study however, p38 suppressor impact on DCL4 3e orf27 (RNase3 enzyme) suppressor works in analogous manner: was attributed to an indirect effect of AGO1-mediated DCL-home- it competes with Dcr-2 for dsRNAs and degrades dsRNAs and ostasis and has been shown that p38 blocks AGO1 but not AGO4 siRNAs (Hussain et al., 2010). activity through its GW-motif binding. DCL2-derived 22 nt vsiRNA production, HEN1 and AGO2-activity pathway is crucial against TCV infection at elevated temperature (Zhang X. et al., 2012). Of Arrest of functional RISC assembly through AGO interaction note, p38 is also capable to bind and inactivate AGO2 (Zhang X. et al., 2012). P38 interacts with unloaded AGOs. Site-directed The arrest in the assembly of a functional RISC can be carried out mutagenesis (GW-to-GA) in the p38 proved that GW motif is also through direct binding the protein component of minimal RISC, absolutely required for both binding and suppression of AGO1 AGO protein. P0 protein (Mayo and Ziegler-Graff, 1996; Sadowy et al., function. GA mutation was enough to abolish TCV virulence that is 2001) the suppressor of Poleroviruses does not possess RNA binding restored in ago1 hypomorph plants showing GW motif function- activity (Csorba et al., 2010; Zhang et al., 2006), instead interacts with ality during authentic virus infection (Azevedo et al., 2010). The E3-ligase S-phase kinase regulated protein 1 (SKP1) through its F-box coat protein (p37) of Pelargonium line pattern virus (PLPV) is a GW- motif. P0 was shown to enhance the degradation of multiple AGOs containing protein that also functions as a VSR. It was shown that (AGO1, 2, 4–6, 9) before holo-RISC assembly. Mutations in the F-box P37 mainly suppress silencing through siRNA sequestration. Muta- motif abolished P0 suppressor activity (Baumberger et al., 2007; tions within its GW-motif concurrently affected p37 localization, Bortolamiol et al., 2007; Csorba et al., 2010; Derrien et al., 2012; its interaction with AGO1 and its sRNA-binding ability. Further- Pazhouhandeh et al., 2006). These suggested a model where P0 more, binding assays have shown that also in case of TCV p38, the destabilizes AGOs through the proteasome pathway. Instead, it was GW-mutation abolished p38 sRNA and long dsRNA-binding capa- shown that P0-mediated AGO destabilization is not sensitive to city (Pérez-Cañamás and Hernández, 2015). The overlapping proteasome inhibitor and that AGO degradation occurs through multiple functions of p38 may lead to misinterpretation of autophagy pathway (Baumberger et al., 2007; Bortolamiol et al., experimental data. However, the data may suggest that these 2007; Csorba et al., 2010; Derrien et al., 2012; Pazhouhandeh et al., functions could cooperate: VSR interaction to AGO could enhance 2006). Massive accumulation of K48-linked ubiquitinated host pro- sRNA duplex sequestration in order to prevent RISC programming. teins accumulated in the presence of P0, including the N-terminal Silencing effector complex activity block can be achieved also cleavage product of AGO1. P0 therefore may have a role in promoting through targeting holo-RISC's RNA component, the guide RNA. degradation of AGOs and other host proteins through ubiquitination- African cassava mosaic virus (ACMV) encoded AC4 was shown to proteasome pathways (Csorba et al., 2010). Tomato ringspot virus bind to the ss-sRNAs but not dsRNA forms in vitro. Transgenic (ToRSV) suppressor CP binds to AGO1 to suppress its translational expression of AC4 correlated with decreased accumulation of inhibitory activity and to enhance AGO1 degradation through autop- miRNAs and upregulation of target mRNAs. AC4 acts downstream hagy (Karran and Sanfacon, 2014). It was shown that PVX p25 of the unwinding process: to bind mature miRNAs presumably physically interacts with AGOs (AGO1, 2, 3 4 but not AGO 5 and loaded into AGO protein (Chellappan et al., 2005; Xiong et al., 2009; AGO9) to promote their destabilization in a proteasome-dependent Zhou et al., 2006). Rice stripe virus suppressor NS3 is able to manner. Consistently with these, plants treated with proteasome suppress and reverse GFP silencing and also prevent long distance inhibitor were less susceptible to PVX (Chiu et al., 2010). spread of silencing signal. NS3 was found to bind to various RNA  T. Csorba et al. / Virology 479-480 (2015) 85–103 93 forms like ss-siRNA, ds-siRNA or long ssRNA but not long dsRNA suppressor of Tomato yellows leaf curl China virus (TYLCCNV) DNA (Chellappan et al., 2005; Xiong et al., 2009; Zhou et al., 2006). satellite interacts with the endogenous suppressor of silencing Grapevine virus A p10 suppressor was also suggested to act through calmodulin-like protein (rgsCAM) in N. benthamiana to repress RNA sequestration: recombinant p10 was able to bind to both ss- RDR6 expression and secondary siRNA production (Li et al., 2014). and ds-si/miRNA species (Chellappan et al., 2005; Xiong et al., 2009; Sugarcane mosaic virus (SCMV) encoded HC-Pro and TAV 2b were Zhou et al., 2006). shown to downregulate RDR6 mRNA in N. benthamiana transient assay (Zhang et al., 2008). Pns10 suppressor of Rice dwarf phytoreo- VSR activities downstream of RISC and RITS virus (RDV) downregulate RDR6 to enhance viral invasion of shoot apices (Ren et al., 2010). RDR6-based activity suppression therefore is VSRs may inactivate host defense downstream to RISC or RITS. a widely used strategy that effectively blocks antiviral silencing but Several DNA viruses encode VSRs that have been described to alter has only a limited impact on endogenous silencing pathways (dis- DNA/histone methylation, the effector step of TGS. Tomato golden cussed latter). mosaic virus (TGMV) suppressor AL2 and Beet curly top virus (BCTV) suppressor L2 inhibit adenosine kinase (ADK) activity that plays Targeting multiple steps of antiviral pathways crucial role in adenosine and methyl-cycle maintenance or cytokinin regulation. In the presence of AL2 and L2 global reduction in cytosine VSRs may act at multiple points to suppress silencing (Fig. 1). methylation was observed that leads to inactivation and reversal of There is substantial evidence that the cucumoviral 2b protein can antiviral silencing (Buchmann et al., 2009; Wang et al., 2003, 2005). interfere with silencing at many steps. 2b prevents the spread of In vitro methylated TGMV cannot replicate in protoplasts suggesting long-range silencing signal (Guo and Ding, 2002). N-terminal that viral genome methylation is a bona fide defense against domain of 2b contains a dsRNA-binding domain that exhibits high geminiviruses that is worth to be suppressed (Bisaro, 2006). Simi- affinity for short and long dsRNAs (Duan et al., 2012). CM95R larly, Tomato yellow leaf curl China virus (TYLCCNV) another Begomo- strain of CMV and the related Tomato aspermy virus (TAV) 2b were virus encodes ßC1 that interacts and inhibit activity of S-adenosyl- also shown to bind siRNAs and ds-miRNAs in vivo and in vitro homocystein-hydrolase (SAHH) that is involved in methyl-cycle and (Chen et al., 2008; Gonzalez et al., 2012; Goto et al., 2007). 2b (Fny therefore indirectly affects TGS (Yang et al., 2011). and SD strains) was found to interact with AGO through the PAZ- and partly PIWI domains and blocks AGO1/RISC slicer activity Modulation of AGO1 homeostasis (Duan et al., 2012; Zhang et al., 2006). FnyCMV 2b transgenic expression phenocopies ago1 mutant plants (Duan et al., 2012; Similarly to the aforementioned examples of DNA virus Zhang et al., 2006). Additionally, CMV 2b encoded by the SD strain encoded VSRs, RNA virus-encoded VSRs were also described to alters RdDM pathway as well. 2b facilitates cytosine methylation modulate host gene expression on transcriptional level to their through the transport of siRNAs into the nucleus (Kanazawa et al., benefit. During tombusviral infection AGO1 transcription is 2011). 2b interacts both with AGO4-related siRNAs and with AGO4 induced as part of the host antiviral arsenal. AGO1 homeostasis protein through PAZ and PIWI domains. Interaction of 2b with in plants depends on the miR168-guided AGO1 mRNA cleavage AGO4 reduces AGO4 access to endogenous target loci and conse- and translational inhibition (Rhoades et al., 2002). To counteract quently modulates endogenous transcription to create a favorable AGO1-based defense, the virus promotes miR168 transcriptional cellular niche for CMV proliferation (Duan et al., 2012; Gonzalez induction that results in miR168-guided AGO1 down-regulation. et al., 2010, 2012; Hamera et al., 2012). The miR168 accumulation spatially correlates with the virus localization and depends on its p19 VSR (Varallyay et al., 2010) VSR interactions with host factors (Fig. 1). Similarly to p19 all VSRs, which are very heterogeneous in protein sequence but bind vsiRNA, promote miR168 transcrip- There are emerging evidences that besides the “canonical” tional induction and AGO1 down-regulation suggesting that VSR- block of RNA silencing (through ds-, si-, mi-ssRNA-binding e.g. siRNA complexes are effectors and recognized by the plant p19, RNase3 etc. or manipulating silencing-related protein activ- surveillance system (Várallyay and Havelda, 2013). ities via direct/ indirect interactions e.g. P0, V2, P1 etc.) some suppressors may target endogenous regulators of the silencing to Plant RDR-based activity suppression modulate host defense. Potyviral helper-component protease (HC- Pro) is a multifunctional protein involved in many aspects of virus Host RDRs (RDR1, 2 and 6) contribute to amplification of RNA infection (Anandalakshmi et al., 1998; Carrington et al., 1989; Guo silencing and spread of a systemic signal by synthesis of vsiRNAs et al., 2011; Kasschau et al., 1997; Lakatos et al., 2006; Mallory (Schwach et al., 2005). Interestingly, plant RDR1 itself was suggested et al., 2001). Tobacco etch virus (TEV) HC-Pro suppresses silencing to have adverse functions. RDR1 is an antagonist of RDR6-mediated through vsiRNA-sequestration (Lakatos et al., 2006) and interferes sense-PTGS silencing therefore behaves as an endogenous silencing with vsiRNA methylation (Lozsa et al., 2008). Zucchini yellows suppressor (Ying et al., 2010). Suppression of RDR activities may mosaic virus (ZYMV) Hc-Pro interacts with HEN1 directly in in vitro constitute a target point for VSRs since it dampens cell-autonomous assays (Jamous et al., 2011). TEV HC-Pro was found to interact with silencing amplification and systemic signal movement in distant rgsCAM in a yeast two-hybrid system. RgsCAM itself is a host tissues to facilitate the virus replication and spread. It was shown suppressor of RNA silencing (Li et al., 2014; Anandalakshmi et al., that V2 protein of Tomato yellow leaf curl virus (TYLCV) directly 2000; Endres et al., 2010). Interestingly, geminivirus AL2 protein interacts with SGS3, the cofactor of RDR6, to block silencing ampli- induces expression and interacts with rgsCAM. Overexpression of fication (Glick et al., 2008). Another in vitro study has shown that V2 rgsCAM leads to increased virus susceptibility (Yong Chung et al., competes with SGS3 for dsRNA having 50 overhang ends that may be 2014). Contradictory, in another study it was shown that rgsCAM an RDR6/SGS3 intermediate in vsiRNA amplification (Fukunaga and counteracts HC-Pro through binding to its positively charged Doudna, 2009; Kumakura et al., 2009). This may suggest that a 50 - dsRNA-binding surface. RgsCAM binding prevents HC-Pro ds- overhanged intermediate may be the RDR6-complex template. Simi- siRNA binding activity and promotes its degradation through lar structures may be present on viral RNAs. Similarly, potexviral autophagy pathway (Nakahara et al., 2012). Multiple host inter- TRIPLE GENE BOX PROTEIN1 (TGBp1) was also shown to inhibit actors may further modulate activities of HC-Pro. It was proposed RDR6/SGS3-dependent dsRNA synthesis (Okano et al., 2014). βC1 that RAV2, a transcription factor is required for suppression of 94 T. Csorba et al. / Virology 479-480 (2015) 85–103 silencing mediated by TuMV HC-Pro. RAV2 downstream targets The control of pathogen impact on host include FIERY1 an endogenous silencing suppressor in Arabidopsis (Gy et al., 2007) and CML38, the likely homologue of rgsCAM in Antiviral and endogenous silencing pathways share common Arabidopsis (Anandalakshmi et al., 2000; Endres et al., 2010). elements. The ability of viruses to block antiviral silencing may Genome expression analysis reveals that RNA silencing related have an impact on endogenous silencing pathway that results in genes were unaltered in HC-Pro transgenic lines, instead RAV2 was alteration in short RNAs expression profile/activity and changes in required for HC-Pro-mediated induction of stress and defense- gene expression both in a direct and in an indirect manner. related genes (Endres et al., 2010). The unrelated Carmoviral p38 vsiRNA-binding VSRs can bind endogenous si- and miRNAs that had a very similar effect on RAV2-mediated changes. These could result in alteration of their downstream targets as was findings may suggest that RAV2 is a cross-talk point between shown previously (Chapman et al., 2004; Kasschau et al., 2003; pathogen stress-defense and silencing pathways that is used by Lozsa et al., 2008). In case of miRNAs that target RNA silencing VSRs to manipulate host reactions. Papaya ringspot virus (PRSV) target components an unpredicted number of genes will be altered HC-Pro interacts with papaya calreticulin to modulate host indirectly (e.g. miR162-mediated DCL1 negative feed-back loop, defense to virus infection through calcium signaling (Shen et al., DCL1-dependent suppression of DCL3 and DCL4, miR168 and 2010a; Shen et al., 2010b). HC-Pro of Potato virus A (PVA), Potato AGO1 mRNA-derived siRNA control of AGO1, miR403 control of virus Y (PVY) and TEV interacts also with microtubule-associated AGO2) (Allen et al., 2005; Mallory and Vaucheret, 2009; Qu et al., protein (HIP2) through its highly variable region (HVR). HIP2 2008; Rajagopalan et al., 2006; Vaucheret et al., 2006; Xie et al., depletion reduces virus accumulation, whereas mutations affect- 2005). The situation is similar in case of AGO-targeting VSRs (P0, ing HC-Pro HVR domain induces necrosis and ethylene- and P1, P38) (Azevedo et al., 2010; Baumberger et al., 2007; Derrien et jasmonic acid-mediated systemic induction of host pathogen- al., 2012; Giner et al., 2010). An elegant demonstration of this was related defense genes (Haikonen et al., 2013a,b). described for TCV p38. AGO1 quenching by p38 had a profound Another example of how the endogenous factors may be used impact on DCLs’ homeostasis uncovering the strong interconnec- to modulate silencing efficiency is provided by Red clover necrotic tion of the silencing components into a functional network mosaic virus (RCNMV). RCNMV recruits DCL enzymes into its (Azevedo et al., 2010). VSRs' presence therefore may have a big replication complex and therefore deprives them from the silen- impact and lead to an altered developmental program of host cing machinery. dcl1 mutant plants show reduced susceptibility to organism and symptom development. In support of VSRs as RCNMV (Takeda et al., 2005). A similar strategy is used by CaMV contributors to the viral symptoms, VSR-transgenic lines were pararetrovirus: massive amounts of vsiRNAs derive from the 35S created and analyzed. In many cases the VSR-expressing trans- leader sequence produced by all four DCLs. These do not restrict genic plants display phenotypes similar to viral infections viral replication but instead may serve as decoy RNAs to divert the (Dunoyer et al., 2004; Jay et al., 2011; Kasschau et al., 2003; effectors of the silencing machinery from important viral features Lewsey et al., 2007; Zhang et al., 2006). However, transgenic like promoter and coding sequences (Blevins et al., 2011) (Fig. 1). expression of VSR does not recapitulate the expression pattern in time and space of an authentic viral infection, therefore conclu- sions need to be drawn very carefully. Driving factors in VSRs' evolution Limitation of VSRs’ suppressor strength The high diversity in structure and function, the various position Some of the VSRs differentially impact the antiviral and the of their gene-code within the viral genome, the alternative expres- endogenous pathways. For example vsiRNA binding Tombusviral sional strategies like transcriptional read-through, leaky ribosomal p19 protein binds more efficiently to free ds-siRNA forms. It was scanning, proteolytic maturation and being often encoded by out- shown previously that p19 blocks very efficiently HEN1-dependent of-frame ORFs within conserved viral genes suggests that VSRs are methylation of vsiRNA but not miR159 (Lozsa et al., 2008). Further- of recent evolutionary origin (Ding and Voinnet, 2007). Therefore, in more, vsiRNA but not miR159 are bound into p19-dimer:siRNA most cases, the suppressor function of VRSs may have evolved after nucleoprotein complex whereas miR159 incorporates efficiently to the ancient role as replicase, coat protein, movement protein, AGO/RISC complex (Varallyay et al., 2010). The miR168, a particular protease, transcriptional regulator etc. or co-evolved with these to miRNA that has a weak AGO-loading rate (Mallory and Vaucheret, combine within the suppressor role and other essential roles 2009), was also available for in vivo p19-binding (Varallyay et al., important for viral life cycle. The different VSRs can inhibit all steps 2010). All these suggests that p19 affects more potently vsiRNA- but of the antiviral RNA silencing pathway, including cell-autonomous at less extent miRNA-pathway (Lozsa, Kontra and Burgyán, unpub- and non-cell autonomous aspects of it. Using mathematical model- lished.). These differences may rise probably from the differences in ing of dynamics of suppression has been shown that the different the DCL1-dependent and DCL4/2-dependent si/miRNA-maturation strategies employed result in slightly different outcomes regarding pathways and distinct sub-cellular localization of silencing compo- suppression of antiviral silencing. Suppressors targeting effector nents. Conversely to p19, the TEV HC-Pro significantly affected 30 - step are more potent at single cell level whereas siRNA binding is methylation of both si- and miRNAs. miRNA 30 -methylation may more effective at tissue level (Groenenboom and Hogeweg, 2012). take place in the nucleus where HC-Pro could access them (Lozsa Besides this however, an important driving factor in the suppres- et al., 2008). Similarly, TMV suppressor localizes to the nucleus as sors’ evolution was probably the availability of ancient/original viral well (dos Reis Figueira et al., 2002). This may explain the more protein activities that could be selected from with a minimum efficient sequestration of ds-miRNAs as assessed by the miRNA-star number of changes to acquire an additional suppressor features. strand accumulation (Csorba et al., 2007). Therefore modulating Many of viral proteins have RNA binding capacity (like replicase, sub-cellular localization of VSR will result in differential impact on coat protein, movement protein). This may explain why an over- antiviral and endogenous pathways respectively and therefore may whelming number of suppressors act through RNA binding. Silen- constitute a regulatory point (Papp et al., 2003). cing functions could have become established in the cases when the P6 is mostly present in the cytoplasm consistent with its transacti- tradeoff between its positive effect on viral life cycle and negative vator function, but a small fraction of it is imported into the nucleus effects on host were worth it. The suppressors being too weak or (where it blocks DRB4). It was shown that cytoplasmic-nuclear too strong were out-selected through evolution. shuttling is prerequisite for VSR function and successful virus infection  T. Csorba et al. / Virology 479-480 (2015) 85–103 95 (Haas et al., 2008). Although P6 contains a dsRNA-binding domain, this Decreasing rather than eliminating specific antiviral silencing is not required for its suppressor function. In fact, perhaps surprisingly, activities may allow setting a balance between blocking antiviral the dsRNA-binding domain has an indirect negative effect on P6 pathway and host transcriptome control through silencing. CaMV P6 suppressor function, presumably because nucleoplasmic localization of protein was described as a symptom and host range determinant, a P6 is decreased by the dsRNA-binding domain interaction with translational transactivator and a silencing suppressor (Bonneville et ribosomal proteins L13 and L18 in the nucleolus and consequently al., 1989; Haas et al., 2008; Love et al., 2007). P6 suppressor blocks P6 localizes more effectively to the nucleoplasm. Sub-cellular localiza- DRB4 function that is required for DCL4 function therefore only tion of P6 therefore alters its impact on host silencing. Similarly, diminishes (but not blocks) DCL4 activity. This maybe sufficient enhanced nuclear targeting of CMV Fny strain 2b by addition of a enough to dampen antiviral silencing but in the same time allow host nuclear localization signal compromises its suppressor activity but in gene control by the 21 nt 35S-derived vsiRNA. It was shown previously the same time enhances 2b-mediated pathogenicity and CMV viru- that several CaMV vsiRNAs exhibit near-complete complementarity to lence, making the virus necrotic and accentuating the disease symp- many endogenous targets (Moissiard and Voinnet, 2006). Besides, toms in Arabidopsis. Enhanced virulence was unrelated to effects of DRB4 targeting may allow accumulation of 24 nt vsiRNA that are nuclear localized 2b on vsi- or miRNA-regulated target functions (Du required for heterochromatin silencing of CaMV minichromosomes et al., 2014). Consequently, CMV Q strain 2b mild suppressor activity (Al-Kaff et al., 1998). could be explained by its preferential accumulation in the nucleus in contrary to cytoplasmic Fny 2b. VSRs as links between RNA-based and protein-based immunity The suppression of antiviral pathways could additionally also protect the host endogenous targets from the vsiRNA-“off-target” Alteration of silencing pathways (an RNA-based immunity) in the effects. Recently it was shown that in the absence of 2b suppressor of presence of the VSR and/or viral infection triggers the protein-based CMV virus-activated siRNAs (vasiRNAs) are produced from endogen- immunity in host as part of the counter-counter defense response. ous transcripts in a DCL4- and RDR1-dependent manner. AGO2 is Resistance (R) genes present in the plant genome convey disease essential for the silencing activity of vasiRNAs (Cao et al., 2014). resistance against pathogens by producing R proteins and their action Another VSR strength-limitation strategy is exemplified by the are main part of the protein-based immunity arsenal. Conserved Poleroviruses. Poleroviral P0 protein expression is restricted by its miRNA family controls a plethora of R genes (Li et al., 2012; suboptimal translation initiation codon, even though the use of Shivaprasad et al., 2012; Zhai et al., 2011). It is assumed that R genes such a weak suppressor-strategy results in phloem restriction of are silenced in the absence of the pathogen in order to minimize the the virus. Attempts to restore optimal P0 translation lead to cost for the plants and prevent autoimmunity reactions (Tian et al., secondary mutations indirectly affecting P0 (Pfeffer et al., 2002). 2003). It was found that the NBS-LRR genes (the main class of R This shows that P0's impact on endogenous pathways is so proteins with nucleotide binding site (NBS) and leucine-rich repeat devastating that is unfavorable for the virus. (LRR) motifs are silenced in a siRNA-regulated cascade similarly to ta- Suppressor strength of HC-Pro was tested on a collection of point siRNA biogenesis scheme: RDR6-dependent secondary siRNAs are mutants in a GFP co-agroinfiltration assay and parallel the virulence produced following the original 22 nt miRNA-mediated cleavage on a of mutant alleles encoding TEV viruses in vivo (Torres-Barcelo et al., R-gene transcript. The secondary siRNA may target other R-genes. 2008). Both hyposuppressor and hypersuppressor HC-Pro alleles When tomato plants were infected with viruses (TCV, CMV, TRV) (or were found suggesting that the wild type protein has intermediate bacteria) accumulation of miR482 was reduced. In the absence of suppressor activity strength. Interestingly, while viruses carrying miR482 activity the resistance R gene targets are released and con- hyposuppressor alleles induced milder symptoms and accumulated sequently R gene products accumulate to enhance immunity of the to a lower level, hypersuppressor mutant viruses produced symp- plants (Shivaprasad et al., 2012). In tobacco the R gene (TIR-NB-LRR toms and accumulated to levels characteristic to wild type virus immune receptor N) that confers resistance against TMV stays under (Torres-Barcelo et al., 2008). This suggests that the suppressor nta-miR6019 and nta-miR6020 regulation. It was also shown that strength could have been optimized by natural selection. miRNA repression of N attenuates resistance to TMV (Li et al., 2012). Since antiviral pathways strongly relay on amplification and Eight miRNA families were identified from Solanaceae (tobacco, tomato systemic spread of silencing, those suppressors that target amplifica- and potato) that may target R genes (Li et al., 2012). In summary tion of silencing (e.g. V2 or TGBp1 inhibition of SGS3) will have a miRNA-regulated R genes participate in a non-race-specific immunity definite impact to fight off antiviral pathway and promote virus spread mechanism where the miRNAs are the sensors of the infection. It is (Fukunaga and Doudna, 2009; Glick et al., 2008; Kumakura et al., supposed that release of R-gene based defense may be the cause of the 2009; Okano et al., 2014). In certain cases however, in order to keep a inhibitory action of pathogen-encoded suppressors of silencing (VSRs tight control and spare the host, virus’ systemic spread is limited by and bacterial suppressor proteins, BSR) on miRNA activity during inf- enhancing systemic silencing signal by the pathogen itself. Rice yellow ection, however this assumption needs to be experimentally tested in mottle virus (RYMV) P1 was described originally as the movement the future. It was shown, however, that in specific Nicotiana species the protein of the virus. Recently it was shown that P1 suppresses local presence of p19 activates extreme resistance (ER) to protect tissues transgenic silencing and inhibits DCL4-dependent endogenous siRNA against TBSV and siRNA binding of p19 was necessary for ER pathways in Nicotiana plants and its expression caused similar (Sansregret et al., 2013). TAV 2b suppressor was found to elicit developmental defects in rice as dcl4 mutation. In the same time P1 hypersensitive reaction (HR) when expressed in tobacco from the facilitated the short- and long distance spread of silencing signal in unrelated TMV genome (Li et al., 1999). Whether the R proteins are the Nicotiana plants. P1 therefore may have a dual role, both to suppress monitors of the VSR presence or activity remains to be established. and to elicit silencing of host (Lacombe et al., 2010). Although the precise mechanism of RYMV P1 activities are not known and need to Connecting antiviral silencing to hormone signaling be confirmed during authentic virus infections, it seems that P1 plays a fundamental role for an efficient infection process where the virus Several studies have shown that antiviral silencing might be preserves the integrity of the host. Similarly to RYMV movement connected to signal transduction pathways responsible for induction protein P1, the TMV MP (30kDa) has been shown to support the of salicylic acid (SA)-mediated resistance (Alamillo et al., 2006; Ji and spread of antiviral silencing signal (Vogler et al., 2008) and therefore Ding, 2001). SA is a plant hormone that is involved in local and counter-balance the activity of p130 the suppressor of TMV to systemic antiviral defense responses including systemic acquired attenuate pathogen impact (Kubota et al., 2003). resistance (SAR). SA induces expression of key antiviral silencing 96 T. Csorba et al. / Virology 479-480 (2015) 85–103 factor RDR1 (Liao et al., 2013; Xie et al., 2001). In turn, RDR1 affects VSRs employed in molecular research many jasmonic acid (JA)-regulated genes (Pandey et al., 2008). JA was implicated as a defense-related hormone (Lewsey et al., 2010). VSRs Apart of these, VSRs can be also employed as molecular biology seem to interfere with hormone signaling-based responses, although tools in research. P19 was used to probe the si- and miRNA content the precise mechanisms are elusive. It was shown that CaMV P6 of cellular milieu following immunoprecipitations. The use of P19 suppressor inhibits SA-induced gene expression through NPR1 (Laird provided sensitive detection without the need for amplification et al., 2013; Love et al., 2012). Upon CMV but not CMVΔ2b (Qavi et al., 2010). A surface plasmon resonance (SPR)-based suppressor mutant strain infection the SA levels accumulated miRNA sensing method was developed where the RNA probes (Lewsey et al., 2010; Zhou et al., 2014). 2b was shown to have a dual are immobilized on gold and p19 recognize and binds the miRNA: action: it inhibited expression of few SA-regulated genes but in the probe duplexes. This allowed detection of miRNA at nanomolar same time enhanced the effect of SA on others. Remarkably, 2b range. Coupled to immunoassay p19 aided miRNA detection at changes the expression of 90% of jasmonic acid (JA)-regulated genes femtomole sensitivity (Nasheri et al., 2011). Similarly P19 was used (Lewsey et al., 2010). 2b protein is known to affect RDR1 activity in protein-facilitated affinity capillary electrophoresis for miRNA (Diaz-Pendon et al., 2007) that may explain partly the 2b effect on JA- detection in blood serum (Berezovski and Khan, 2013). These high- signaling. HC-Pro was also shown to interfere with JA-regulated sensitivity methods are very powerful especially because there is transcript expression (Endres et al., 2010). TCV p38 has been shown no amplification bias in contrary to the PCR-based techniques. to interact with a NAC transcription factor called TIP in Arabidopsis. It was suggested that p38 interaction with TIP alters defense signaling VSRs as biotechnological and medical tools to favor enhanced TCV invasion (Donze et al., 2014). VSRs therefore emerge as regulators of hormone-based signaling to create favorable VSRs also emerge as important and useful tools for biotechnol- conditions for the virus. Although at the moment the complex ogy and medicine. RNA silencing impairs or limits the use of interplay between the RNA silencing and SA-mediated defense is transgenic plants in numerous biotechnological applications. VSRs elusive VSRs might be important coordinators of this crosstalk during can be employed efficiently to limit transgene silencing and attain infection. consistently high-level expression of diverse products like vac- cines and pharmaceuticals, high-nutritive foods, high-value pro- ducts etc. (Kanagarajan et al., 2012; Naim et al., 2012). VSRs as tools Unraveling molecular basis of silencing itself Conclusions and perspectives Due to their versatility and availability with a wide range of VSRs regulate the multiple layers of the complex defense, counter- actions covering theoretically every aspect of RNA si