Host factors against plant viruses

Summary Plant virus genome replication and movement is dependent on host resources and factors. However, plants respond to virus infection through several mechanisms, such as autophagy, ubiquitination, mRNA decay and gene silencing, that target viral components. Viral factors work in synchrony with pro‐viral host factors during the infection cycle and are targeted by antiviral responses. Accordingly, establishment of virus infection is genetically determined by the availability of the pro‐viral factors necessary for genome replication and movement, and by the balance between plant defence and viral suppression of defence responses. Sequential requirement of pro‐viral factors and the antagonistic activity of antiviral factors suggest a two‐step model to explain plant–virus interactions. At each step of the infection process, host factors with antiviral activity have been identified. Here we review our current understanding of host factors with antiviral activity against plant viruses.


I N T RO D U C T I O N
Infection of a plant by a virus initiates in a single cell. Viral proteins are synthesized by the host cell before genome replication, virion formation and movement to a neighbouring cell. The cycle is repeated at every newly infected cell (Nelson and Citovsky, 2005). Using the vascular system, plant viruses move long distances to infect tissues away from the initial site of infection, such as roots and young leaves (Heinlein, 2015;Wan et al., 2015). Multiple genetic analyses have shown that the entire infection cycle, including virus replication and movement, is genetically determined by viral and cellular factors that synchronize their activities in time and space (Fig. 1A) (Diaz et al., 2015;Hofius et al., 2007;Laliberte and Zheng, 2014;Li et al., 2016;Sasvari et al., 2018;Zhang et al., 2018).
Plant-virus combinations could result in an incompatible or compatible interaction. Incompatible interactions occur between a virus and a non-host plant, are characterized by the absence of virus infection and may be explained by the lack of cellular factors essential for the virus to replicate or move, antiviral defence or a combination ( Fig. 1B) Lellis et al., 2002). In contrast, compatible interactions occur between a virus and a susceptible host, are characterized by the establishment of virus infection and indicate the presence of pro-viral cellular factors and resources necessary for virus infection and movement. Infection may spread through the entire plant, parts of the plant or be limited to the vascular system or the initially infected organ (Calvo et al., 2014a(Calvo et al., , 2014bLv et al., 2017;Otulak-Koziel et al., 2018).
In susceptible hosts the absence of critical pro-viral host factors results in the absence of infection and reduced virus replication, movement or both (Hofius et al., 2007;Lellis et al., 2002;Wang and Nagy, 2008). Accordingly, the absence of pro-viral factors may turn a susceptible host into a non-host, as is the case with resistant cultivars, landraces or ecotypes within a susceptible species (Hashimoto et al., 2016;Lellis et al., 2002). Because their presence conditions susceptibility, while their absence results in immunity or resistance, several terms have been used to describe host genes with pro-viral activity, such as loss of susceptibility, recessive resistance or susceptibility genes Hashimoto et al., 2016).
Interestingly, susceptible hosts harbour factors with antiviral activity (Diaz-Pendon et al., 2007;Kushner et al., 2003;Panavas et al., 2005;Scholthof et al., 2011). To establish infection, viruses escape from or suppress antiviral defence activated by viral proteins or nucleic acids, particularly RNA ( Fig. 1A) (Garcia and Pallas, 2015;Gorovits et al., 2016). With or without a hypersensitive reaction, the defence response restricts essential parts of the infection cycle, such as viral RNA translation, virus replication or movement, resulting in reduced virus accumulation and/or a delay in the establishment of systemic infection. Symptoms may or may not develop (Donze et al., 2014;Korner et al., 2018).
Multiple genetic analyses have shown that the outcome of plant-virus interactions is genetically determined by viral factors, host factors and their interaction (Fig. 1A) (Chisholm et al., 2001;Jaubert et al., 2011;Kushner et al., 2003;Lellis et al., 2002;Panavas et al., 2005). Consistent with this model, for all parts of the infection cycle, at least one host gene with antiviral activity has been identified (Table 1). Here I review our current understanding of host factors with antiviral activity against plant viruses. Their antagonistic activity is presented following sequential parts of the infection cycle.

V I R A L D E T E R M I N A N T S O F I N F EC T I O N
RNA translation, genome replication, and virion formation and movement are core parts of the infection cycle of a plant by a virus (Ahlquist, 2006;Nelson and Citovsky, 2005). To accomplish these tasks, plant viruses encode replication, capsid, movement and auxiliary proteins. Additionally, to condition a cellular environment that is conducive to virus replication and movement, viral proteins target key components of antiviral immunity. Viral factors that determine the extent of infection and disease severity are considered pathogenicity determinants. Gene silencing suppressors are remarkable pathogenicity determinants of plant viruses. Virus-encoded suppressors interfere with antiviral defence mechanisms mediated by gene silencing (Anandalakshmi et al., 1998;Diaz-Pendon et al., 2007;Garcia-Ruiz et al., 2015;Jaubert et al., 2011). Interestingly, the activity of several gene-silencing suppressors promotes infection of heterologous viruses when expressed in cis-, transor synergistic co-infections Gupta et al., 2018;Maliogka et al., 2012) (Fig. 2).

H OST D E T E R M I N A N T S O F I N F EC T I O N
Genome-wide screens of Saccharomyces cerevisiae (yeast) replicating brome mosaic virus (BMV) (Kushner et al., 2003) or tomato bushy stunt virus (TBSV) (Panavas et al., 2005) showed that a compatible host contains both pro-viral and antiviral factors that affect virus replication at the cellular level. Mutagenesis and genetic analyses in Arabidopsis thaliana (Arabidopsis) allowed identification of pro-viral and antiviral factors that affect virus replication at the cellular level, cell-to-cell and systemic movement (Diaz-Pendon et al., 2007;Garcia-Ruiz et al., 2010;Guo et al., 2017;Lellis et al., 2002;Zhang et al., 2019). Accordingly, viruses need pro-viral host factors and are targeted by antiviral host factors. Pro-viral host factors are necessary for essential steps of the infection cycle and work in synchrony with viral factors. In contrast, antiviral defence is mediated by host factors that target viral nucleic acids or proteins by multiple mechanisms such as autophagy, proteasome degradation, RNA decay and gene silencing (Fig. 1A). Their expression is dependent on the host RNA translation machinery. Their activity requires host factors (pro-viral) and resources. Antiviral immunity consists of host factors that target viral proteins or nucleic acids to restrict virus infection. (B) A two-step model in plant-virus interactions. Compatibility is determined by the availability of pro-viral host factors. Susceptibility is determined by the balance between antiviral defence and suppression of antiviral defence.     M o l ec u l a r P l a n t Pat h o lo gy (2019) 20(11), 1588-1601

CO M PAT I B I L I T Y A N D SU SC E P T I B I L I T Y I N PL A N T-V I RU S I N T E R AC T I O N S
The role of host and viral factors, and their availability in the cell during the establishment of infection, suggests a two-step model to explain plant-virus interactions (Fig. 1B). Compatibility is determined by the availability of pro-viral host factors. The absence of one or more pro-viral factors results in incompatibility in a host plant (Lellis et al., 2002). Viral proteins must be translated by the host translational machinery before replication can occur (Ahlquist, 2006;Machado et al., 2017;Miller et al., 2016). This feature makes viral RNA translation a critical determinant of the outcome in plant-virus interactions. An example is translation initiation factor eIF(iso)4E and potyviruses. Arabidopsis is susceptible to tobacco etch virus (TEV) and turnip mosaic virus (TuMV). However, mutant plants lacking eIF(iso)4E are immune to TEV and TuMV (Lellis et al., 2002). Similarly, down-regulation of eIF(iso)4E in plum confers resistance to plum pox virus (PPV) (Wang et al., 2013). The mechanism is likely mediated by the lack of RNA translation to form potyviral polyproteins and possibly the lack of cell-to-cell movement (Contreras-Paredes et al., 2013;Lellis et al., 2002). Similar effects have been detected for several plant species and their corresponding potyviruses (Sanfacon, 2015). In compatible plant-virus combinations, susceptibility is determined by the balance between antiviral defence and  Yeast: Saccharomyces cerevisiae.  suppression of antiviral defence (Fig. 1B). Defence responses may prevent infection from spreading to the entire plant, determining different levels of susceptibility to virus infection. If infection is stopped early, before the formation of local foci, the plant phenotype may be the same as that of a non-host (Garcia-Ruiz et al., 2015;Lellis et al., 2002;Qu et al., 2008). A genetic analysis of infection of Arabidopsis by green-fluorescence protein-tagged TuMV (TuMV-GFP) illustrates a two-step model in plant-virus interactions ( Fig. 2A). Suppressor-deficient (mutant helper component proteinase HC-Pro) TuMV-AS9-GFP cannot infect wild-type plants or dcl2-1 mutants. Infection is halted at the cellular level by gene silencing. However, suppressor-deficient TuMV-AS9-GFP is able to infect dcl4-2 mutants, which lack the contribution of Dicer-like protein 4 (DCL4) to gene silencing. Visible infection foci form and the virus moves systemically into cauline leaves without reaching the inflorescence. Interestingly, suppressor-deficient TuMV-AS9-GFP is able to establish local and systemic infection of cauline leaves and inflorescence in dcl2-1 dcl4-2 double mutants. Thus, in the absence of the HC-Pro silencing suppression activity, gene silencing restricts infection in a tissue-specific manner. In contrast, TuMV-GFP establishes local and systemic infection, including the inflorescence, of Arabidopsis plants, wild-type or mutants. Accordingly, the antiviral role of gene silencing is defeated by the TuMVencoded silencing suppressor HC-Pro (Garcia-Ruiz et al., 2010).
Knockdown of RDR6 by RNA interference in N. benthamiana (rdr6i) (Schwach et al., 2005) rescued local and systemic infection by TuMV-AS9-GFP. In an alternative approach, expression in cis of the NSs protein from tomato spotted wilt virus (TSWV)  supported the establishment of local and systemic infection by TuMV-AS9-GFP (Fig. 2B). These observations show that tissue-specific restriction of virus infection is determined by the balance between gene silencing and gene silencing suppression (Garcia-Ruiz et al., 2015, 2018Schwach et al., 2005).

H OST G E N E S W I T H A N T I V I R A L AC T I V I T Y
Host factors with antiviral activity (Fig. 2) limit virus accumulation, movement or both, resulting in a virus-resistant or tolerant phenotype that normally displays symptoms less severe than susceptible plants (Diaz-Pendon et al., 2007;Garcia-Ruiz et al., 2010;Huh et al., 2013). For each part of the infection cycle, at least one host gene with antiviral activity has been identified (Table 1). Representative host factors are described below.

Viral RNA translation
Translation of viral proteins from genomic RNA, subgenomic RNA or mRNA is dependent on cellular factors and the protein translation machinery. Being a critical step that determines the availability of viral proteins, both host and viral factors regulate translation (Ahlquist, 2006;Miller et al., 2016;Sanfacon, 2015). In Arabidopsis, a leucine-rich repeat receptor-like kinase (NIK1) is a master regulator of translation. As a defence mechanism, using an NIK1-dependent pathway, plants downregulate translation upon begomovirus infection. This effect results in a reduction in virus replication and accumulation. Remarkably, begomoviral nuclear shuttle protein (NSP) inactivates NIK1 to up-regulate translation and promote susceptibility (Zorzatto et al., 2015).
Using nucleotide sequences as recognition signatures, Arabidopsis Pumilio RNA binding protein 5 (APUM5) binds cucumber mosaic virus (CMV) and TuMV mRNA to inhibit translation. Accordingly, mutant plants lacking APUM5 accumulate CMV and TuMV to higher levels than plants harbouring wild-type APUM5 (Huh et al., 2013).

Virus replication complex formation
After reaching the nucleus of infected cells, DNA viruses form minichromosomes that are replicated by cellular DNAdependent DNA polymerases (Ceniceros-Ojeda et al., 2016). In contrast, on cellular membranes, RNA viruses induce the formation of vesicles that contain RNA-dependent RNA polymerases and genomic RNA, and are the sites of replication. Several cellular proteins that antagonize the formation of viral replication compartments have been identified and characterized (Table 1).
Phospholipids are crucial membrane components. Phosphatidic acid phosphohydrolase 1 (PAH1) limits phospholipid synthesis. Genetic analyses in yeast and N. benthamiana showed that PAH1 negatively regulates BMV and TBSV replication complex formation, resulting in reduced virus replication at the cellular level and reduced accumulation in plants (Chuang et al., 2014;Zhang et al., 2018).

Accumulation or activity of replication proteins
Virus-encoded RNA-dependent RNA polymerases replicate the genome of RNA viruses and, if present, transcribe subgenomic RNAs that are essential for gene expression. Viral RNA-dependent RNA polymerases contain a conserved GDD motif  and are targeted for degradation by autophagy protein 6 (ATG6 or Beclin1). Beclin1 is a core component of autophagy, interacting with and triggering degradation of the RNA-dependent RNA polymerase (NIb) of several potyviruses, including TuMV, PPV, soybean mosaic virus (SMV) and TEV. Beclin1 also triggers degradation of the RNA-dependent RNA polymerases of cucumber green mottle mosaic virus and pepino mosaic virus . Additionally, in pepper (Capsicum annum), pathogenesisrelated protein 4c (Pvr4c) interacts with NIb and triggers cell death upon infection by pepper mottle virus or potato virus Y (PVY) (Kim et al., 2015).

RNA replication
Glycolytic enzyme glyceraldehyde 3-phosphate dehydrogenase (GAPDH) binds the 3′ UTR of bamboo mosaic virus (BaMV) and satellite BaMV RNAs. This interaction down-regulates BaMV negative-strand RNA synthesis. Accordingly, GAPDH knockdown in N. benthamiana enhanced accumulation of BaMV. In contrast, GAPDH overexpression reduced BaMV accumulation (Prasanth et al., 2011). GAPDH has the opposite effect on TBSV replication. GAPDH preferentially binds to the 3′ end of negative-strand TBSV RNA, retaining it in the replication complex to promote positive-strand RNA synthesis. This activity results in asymmetric RNA replication characterized by higher synthesis and accumulation of positive-over negative-strand genomic RNA, which is normal in TBSV replication (Wang and Nagy, 2008)  Viruses express their genes through mRNA (Ahlquist, 2006). After translation, cellular and viral mRNAs are deadenylated, decapped and cleaved 5′ to 3′ by exoribonuclease 4 (XRN4) through the decapping-dependent RNA decay pathway (Thran et al., 2012;Tsuzuki et al., 2017). Recent genetic analyses showed that RNA decay has an antiviral role that limits virus accumulation and may contribute to plant recovery from virus-induced symptoms Ma et al., 2015;Moon and Wilusz, 2013;Tsuzuki et al., 2017). RNA decay and RNA silencing seem to act in coordination to suppress virus infection, and their activities partially overlap Peng et al., 2011).
Consistent with the antiviral role of RNA decay, potyviral HC-Pro and genome-linked protein (VPg) are silencing suppressor proteins that interfere with both gene silencing and mRNA decay. Interference with mRNA decay occurs through interactions with XRN4 and decapping protein 2 (DCP2), respectively, two core components of the 5′ to 3′ RNA decay pathway .

Virus movement
Plant viruses move cell to cell as virions or nucleoprotein complexes through plasmodesmata. As a critical component of this process, plant viruses encode movement proteins that increase the plasmodesmata size exclusion limit and/or form microtubules (Taliansky et al., 2008). Several host factors that antagonize virus movement have been identified. They target viral proteins and RNA, or trigger cell death (Table 1).
Plants encode mechanosensitive ion channels that regulate ion movement across cells. In Arabidopsis, ESC1 encodes a piezo protein that functions as a mechanosensitive Ca 2+ permeable channel and limits systemic infection of CMV and TuMV (Zhang et al., 2019).
Systemic movement of TEV, and some isolates of PPV and lettuce mosaic virus (Decroocq et al., 2009), is restricted by restricted-TEV-movement (RTM) genes RTM1, RTM2 and RTM3. These genes are expressed in phloem sieve elements and interact with the viral coat protein (Chisholm et al., 2001). Interestingly, resistance-breaking isolates had mutations in the N-terminus of the coat protein (Decroocq et al., 2009). BTR1 is a ribonucleoprotein K-homology RNA-binding protein that binds ToMV genomic RNA and restricts cell-to-cell movement (Fujisaki and Ishikawa, 2008). In potato (S. tuberosum) plants the Ny-1 gene confers resistance to PVY by triggering cell death at the infection sites, limiting cell-to-cell movement (Lukan et al., 2018).

Gene silencing
In plants, gene silencing is an essential mechanism of antiviral defence. Gene silencing targets viral RNA for degradation or translational repression. The result is restriction of virus replication and movement, and recovery from virus-induced symptoms (Korner et al., 2018;Szittya and Burgyan, 2013). Gene silencing targets DNA and RNA viruses, satellite RNA viruses and viroids (Blevins et al., 2006;Diaz-Pendon et al., 2007;Minoia et al., 2014;Shimura et al., 2011). All viruses express their genes and/or replicate their genome through an RNA intermediate (Ahlquist, 2006). This feature exposes viruses to gene silencing.
The core components of gene silencing include Dicer-Like (DCL), Argonaute (AGO), double-stranded RNA binding (DRB) and RNA-dependent-RNA-polymerase (RDR) proteins. These proteins are conserved across plants (Incarbone and Dunoyer, 2013;Szittya and Burgyan, 2013;Zvereva and Pooggin, 2012). A signature feature of antiviral gene silencing is the accumulation of virus-derived small interfering RNAs (siRNAs) in infected plants. Viral RNA is processed by DCL proteins into siRNAs that are 21 to 24 nucleotides long. Virus-derived siRNAs are loaded into AGO proteins, programming them for specific slicing or translational repression of viral RNA (Garcia-Ruiz et al., 2015;Karran and Sanfacon, 2014;Schuck et al., 2013). Accordingly, viral RNA is targeted by both DCL and AGO proteins. Antiviral gene silencing might be triggered by viral RNA replication intermediates, self-complementary sequences forming hairpin structures in viral single-stranded RNA, and by products of overlapping transcription (Pantaleo et al., 2007;Szittya and Burgyan, 2013). DCL proteins process double-stranded virus RNA into primary virus-derived siRNAs that are necessary but not sufficient to prevent virus infection. Establishment of an antiviral state requires silencing amplification by plant RNA-dependent RNA polymerases that synthesize double-stranded RNA from single-stranded viral RNA (Diaz-Pendon et al., 2007;Garcia-Ruiz et al., 2010).
Virus-derived siRNA profiling has demonstrated that, in compatible plant-virus interactions, the entire genome of positive-strand and negative-strand RNA viruses is targeted by gene silencing in both monocot and dicot plants (Donaire et al., 2009;Garcia-Ruiz et al., 2015;Margaria et al., 2015;Tatineni et al., 2014;Wang et al., 2011;Xia et al., 2016). However, gene silencing is not enough to restrict virus infection. That is due to the inhibitory activity of virus-encoded gene silencing suppressors. Suppressors condition susceptibility, promote virus replication and movement, and promote symptom development by interfering with endogenous and antiviral gene silencing (Burgyan and Havelda, 2011; Kasschau et al., 2003). The mechanisms of silencing suppression include triggering the degradation of core components of gene silencing such as DCL, AGO, RDR6 and suppressor of gene silencing 3 (SGS3) proteins, and binding of both virus-derived and cellular siRNAs including micro-RNAs (miRNAs) (Burgyan and Havelda, 2011; Garcia-Ruiz et al., 2015;Del Toro et al., 2017). These effects prevent the biogenesis and/or activity of virus-derived and cellular siRNAs. Plant development and response to abiotic and biotic stress is in part regulated by miRNAs and other siRNAs. Accordingly, virus-encoded gene silencing suppressors are determinants of symptom development Kasschau et al., 2003). Furthermore, viral silencing suppressors impact virus accumulation and the spatial distribution of virus infection , and are frequently determinants of host range (Garcia-Ruiz et al., 2015;Jaubert et al., 2011;Li et al., 2014).
The antiviral role of gene silencing was unambiguously demonstrated using viruses lacking gene silencing suppressors (Diaz-Pendon et al., 2007;Garcia-Ruiz et al., 2010Pantaleo et al., 2007;Qu et al., 2008). Turnip crinkle virus (TCV), TBSV, CMV and TuMV accumulate to similar levels in wild-type plants and in mutant plants lacking core components of the silencing machinery. However, suppressor-deficient viruses cannot infect wild-type plants. Instead, suppressor-deficient viruses can only infect plants lacking core gene silencing components (Fig. 2). These genetic systems have been used to identify and characterize components of gene silencing (Diaz-Pendon et al., 2007;Garcia-Ruiz et al., 2010Pantaleo et al., 2007;Qu et al., 2008).
As illustrated by genetic analyses using suppressor-deficient TuMV-AS9-GFP, antiviral gene silencing restricts virus infection and movement in a tissue-specific manner. In Arabidopsis plants lacking DCL4, AGO2, RDR1 or RDR6, TuMV-AS9-GFP established local infection and moved systemically into non-inoculated leaves, without reaching the inflorescence. Systemic infection of the inflorescence only occurred in the absence of both DCL2 and DCL4, or RDR1 and RDR6, or AGO1, AGO2 and AGO10 (Garcia-Ruiz et al., 2010. To prevent their inhibitory effect on gene silencing, several plant factors target virus-encoded silencing suppressors or regulate expression of gene silencing components (Table 1), as illustrated by the following examples. In Nicotiana tabacum, a calmodulin-like protein (rgs-CaM) binds to and, via autophagy directs degradation of, 2b, the silencing suppressor in CMV (Jeon et al., 2017). Autophagy cargo receptor NBR1 targets potyviral HC-Pro for degradation, thus affecting silencing suppression and reducing accumulation of TuMV and watermelon mosaic virus (WMV). Interestingly, TuMV VPg and 6K2 prevent NBR1dependent degradation of HC-Pro (Hafren et al., 2018).
In petunia (Petunia hybrida), PhOBF1, a leucine transcription factor, is up-regulated by tobacco rattle virus (TRV) infection. PhOBF1 is a positive regulator of salicylic acid biosynthesis and of core components of gene silencing: DCL, AGO and RDRs. Thus, PhOBF1 enhances antiviral responses to TRV (Sun et al., 2017). In tomato, the Ty-1 gene encodes an RNA-dependent RNA polymerase that confers resistance to geminiviruses by enhancing transcriptional gene silencing (Butterbach et al., 2014).

Virion assembly and disassembly
In BMV, the negative-strand RNA core promoter consists of a short stem with a three-nucleotide loop that forms a clamp adenine motif. An array of 5000 yeast proteins was screened for proteins that bind the clamp adenine motif. Pseudouridine synthase 4 (PUS4) was identified. Functional characterization in N. benthamiana showed that PUS4 binding to BMV positive-strand RNA prevented encapsidation, resulting in a slight reduction in viral RNA accumulation and a drastic reduction in BMV systemic movement (Zhu et al., 2007).

Host factors that condition virus resistance by undetermined mechanisms
For a growing number of plant-virus combinations, reduced virus accumulation has been observed in the presence of genes with antiviral activity although the part of the infection cycle that is affected has not been identified (Table 1). In these cases, virus infection triggers a hypersensitive response that results in the formation of necrotic lesions. Cell death might reduce virus movement and confine the virus to the infection sites and surrounding cells, but is not sufficient to prevent virus movement out of the cell death zone (Lukan et al., 2018).
The following plant-virus combinations are examples of host factors that condition virus resistance by undetermined mechanisms. The N resistance gene from Nicotiana glutinosa was introduced into N. tabacum and confers resistance to TMV (Levy et al., 2004). The N resistance protein is a receptor that contains three essential domains: a Toll-interleukin-1 (TIR), a nucleotide-binding site (NBS) and a leucine-rich repeat (LRR) (Dinesh-Kumar et al., 2000). Transcription and alternative splicing of the N gene is stimulated by TMV infection (Levy et al., 2004), the protein coded by the N gene recognizes the helicase domain in TMV replication protein 126-kD, and triggers a hypersensitive response visible as local necrotic lesions. As a result, TMV infection is restricted to cells surrounding the entry site (Abbink et al., 1998;Levy et al., 2004;Padgett et al., 1997).
The arginine-rich cyclin 1 (RCY1) gene in Arabidopsis recognizes the coat protein in CMV strain Y and triggers local cell death (Takahashi et al., 2001). Similarly, TSWV infection triggers cell death in plants carrying the Tsw and Sw5b genes, which recognize the NSs or NSm proteins, respectively (Mariana et al., 2014). Likewise, the Tm-2 2 gene in tomato encodes a leucine-rich protein that interacts with the movement protein and confers resistance to tobamoviruses, including TMV. The response is mediated by a hypersensitive response and localized cell death . In soybean (Glycine max), the Rsv1 gene confers resistance to SMV strain N. SMV strain G7 is not affected. Both P3 and HC-Pro mediate Rsv1dependent restriction of SMV strain N (Eggenberger et al., 2008). A component of the oxygen-evolving complex pathosystem II, PSBP, interacts with alfalfa mosaic virus coat protein and delays activation of antiviral responses mediated by reactive oxygen species (Balasubramaniam et al., 2014).

I D E N T I F I C AT I O N O F H OST FAC TO R S W I T H A N T I V I R A L AC T I V I T Y
Host factors with antiviral activity have been identified and characterized using several experimental approaches (Table 1). For a small number of cases, natural genetic variation in plant populations has been used to identify, map and clone genes with antiviral activity (Lukan et al., 2018;Maiti et al., 2012;Szajko et al., 2008). However, experimental model systems based on yeast, Arabidopsis and N. benthamiana have contributed most of the genes with antiviral activity known to date (Table 1). These experimental systems allow systematic genetic analysis of virus-host interactions. Using yeast, genome-wide screens have been conducted for BMV and TBSV. Based on the Arabidopsis mutant collection, multiple screens have been done for gene families such as DCL, AGO, RDRs, RNA decay or autophagy mutants (Blevins et al., 2006;Diaz-Pendon et al., 2007;Garcia-Ruiz et al., 2015;Jaubert et al., 2011;Qu et al., 2008).
Based on the concept that host and viral factors colocalize and may interact, yeast-two hybrid assays and cell fractionation or immunoprecipitation followed by mass spectrometry has led to the identification of several genes with antiviral activity (Fujisaki and Ishikawa, 2008;Ishibashi et al., 2007). These assays involved a virus natural host or an experimental host (Table 1).
Functional characterization of the genes identified has been done using loss-of-function or gain-of-function mutants in Arabidopsis, N. benthamiana or natural hosts. Additionally, virus-induced gene silencing (VIGS) has been widely used to down-regulate genes in N. benthamiana. In both Arabidopsis and N. benthamiana, transient or transgenic expression has been used to validate the antiviral activity of a growing number of genes (Haxim et al., 2017;Jaag and Nagy, 2009;Sun et al., 2017).

FAC TO R S E SS E N T I A L A N D N O N -E SS E N T I A L F O R H OST SU RV I VA L
Genes with antiviral activity might be essential or non-essential for plant survival. Non-essential genes affect virus replication or movement without affecting the host (Table 1). However, essential genes cannot be removed from the host. AGO1 participates in antiviral defence (Garcia-Ruiz et al., 2015;Qu et al., 2008;Wang et al., 2011) and is essential for miRNAdependent regulation of gene expression and development. Accordingly, ago1 null mutants show severe developmental phonotypes and are sterile. Hypomorphic ago1 mutant alleles retain part of their activity and have been used to genetically characterize the role of AGO1 in antiviral defence (Morel et al., 2002). In contrast, DCL2 and DCL4 are non-essential, are redundant to each other, and single and double mutants show only mild leaf malformation (Diaz-Pendon et al., 2007;Garcia-Ruiz et al., 2010).

CO N C LU S I O N S
The infection cycle of a plant by a virus is genetically determined by viral factors, cellular factors and their interaction. Viruses use cellular factors and resources to replicate and move. Viral protein or nucleic acids are targeted by antiviral immunity (Fig. 1A). A two-step model for plant-virus interactions explains plant susceptibility to viruses (Fig. 1B). Initially, establishment of infection is determined by the level of plant-virus compatibility. Incompatibility might result from the lack of pro-viral factors, while compatibility is determined by the availability of pro-viral host factors. Subsequently, in compatible plant-virus combinations susceptibility is determined by the balance between antiviral defence and suppression of antiviral defence (Fig. 1B). Strong antiviral defence may stop infection at any point before spreading to the entire plant. This range results in plants with different levels of susceptibility. The lowest level of susceptibility, resulting from arrest of infection at the initially infected cell, is difficult to distinguish from an incompatible interaction that occurs in a non-host (Fig. 1B). Susceptible hosts harbour both pro-viral factors and factors with antiviral activity (Table 1). Their functional characterization has improved our understanding of the mechanisms of virus pathogenicity and antiviral defence in plant-virus interactions.

F U T U R E D I R EC T I O N S
Host genes with antiviral activity provide an interesting option to develop genetic resistance to viruses in crops. However, viruses have a remarkable ability to mutate and are rapidly evolving (Duffy, 2018). Virus-resistant plants select for variants capable of breaking genetic resistance. An example is the emergence of tomato brown rugose fruit virus (ToBRV), described in 2016. ToBRV originated from a recombination event between TMV and tomato mild mottle virus (ToMMV) (Salem et al., 2016). Interestingly, within a year, a ToBRV isolate that broke the Tm-2 2 -dependent resistance was identified (Luria et al., 2017).
A complementary or alternative approach to the deployment of genes with antiviral activity is the identification, characterization and deployment of pro-viral factors that determine susceptibility to plant viruses .
To date, genome-wide screens and genetic analysis have been done mainly in model viruses using heterologous hosts and/or model plant systems (Table 1). Current advances in genome editing (Zhe et al., 2018) make it possible to implement genetic analysis in crop plants. Genome editing in combination with epitope-tagging of viral proteins, either individually or in the context of an infectious clone, make it currently possible to identify and characterize host genes with antiviral activity and pro-viral genes crop plants. Thus, it is safe to predict that viruses causing devastating diseases in staple crops will receive more attention in the near future. This includes orthotospoviruses, potyviruses, tobamoviruses and geminiviruses. Prime examples are the causal agents of maize lethal necrosis, maize chlorotic mottle virus and sugarcane mosaic virus (Wamaitha et al., 2018), and recently described ToBRV (Luria et al., 2017;Salem et al., 2016).

AC K N OW L E D G E M E N T S
This research was supported by NIH grant R01GM120108 to Hernan Garcia-Ruiz and by the Nebraska Agricultural Experiment Station with funding from the Hatch Act (Accession Number 1007272) through the USDA National Institute of Food and Agriculture. Open access costs were provided by the same grant.

CO N F L I C T S O F I N T E R E ST
The author declares no conflict of interest. The founding sponsors had no role in the design of the study, in the collection, analyses or interpretation of data, in the writing of the manuscript or in the decision to publish the results.