Domains of the cucumber mosaic virus 2b silencing suppressor protein affecting inhibition of salicylic acid-induced resistance and priming of salicylic acid accumulation during infection

The cucumber mosaic virus (CMV) 2b silencing suppressor protein allows the virus to overcome resistance to replication and local movement in inoculated leaves of plants treated with salicylic acid (SA), a resistance-inducing plant hormone. In Arabidopsis thaliana plants systemically infected with CMV, the 2b protein also primes the induction of SA biosynthesis during this compatible interaction. We found that CMV infection of susceptible tobacco (Nicotiana tabacum) also induced SA accumulation. Utilization of mutant 2b proteins expressed during infection of tobacco showed that the N- and C-terminal domains, which had previously been implicated in regulation of symptom induction, were both required for subversion of SA-induced resistance, while all mutants tested except those affecting the putative phosphorylation domain had lost the ability to prime SA accumulation and expression of the SA-induced marker gene PR-1.

The cucumber mosaic virus (CMV) 2b silencing suppressor protein allows the virus to overcome resistance to replication and local movement in inoculated leaves of plants treated with salicylic acid (SA), a resistance-inducing plant hormone. In Arabidopsis thaliana plants systemically infected with CMV, the 2b protein also primes the induction of SA biosynthesis during this compatible interaction. We found that CMV infection of susceptible tobacco (Nicotiana tabacum) also induced SA accumulation. Utilization of mutant 2b proteins expressed during infection of tobacco showed that the N-and C-terminal domains, which had previously been implicated in regulation of symptom induction, were both required for subversion of SA-induced resistance, while all mutants tested except those affecting the putative phosphorylation domain had lost the ability to prime SA accumulation and expression of the SA-induced marker gene PR-1.
Salicylic acid (SA) is required for elicitor-triggered immunity and establishment of systemic acquired resistance against a wide range of pathogens, including viruses (Palukaitis & Carr, 2008). SA induces several antiviral mechanisms, some of which may involve RNA silencing but none of which are well understood (Lee et al., 2011;Lewsey & Carr, 2009). In Arabidopsis thaliana and tobacco (Nicotiana tabacum) resistance to systemic movement of cucumber mosaic virus (CMV) is induced by treatment with SA or its synthetic analogues (Mayers et al., 2005;Naylor et al., 1998;Smith-Becker et al., 2003), or by inducing endogenous SA biosynthesis by pre-infection with an incompatible pathogen (Bergstrom et al., 1982;Naylor et al., 1998).
Increased SA biosynthesis has been associated with incompatible plant-pathogen interactions, especially those involving hypersensitive-type resistance and host cell death in the infection zone (Malamy et al., 1990;Métraux et al., 1990), or with abortive defence reactions involving systemic necrosis . However, SA accumulation can occur during some non-necrotizing, compatible plant-virus interactions (Love et al., 2005;Whitham et al., 2003), including infection by CMV of A. thaliana (ecotype Col-0). Expression of ICS1 (a gene encoding a key SA biosynthetic enzyme; Wildermuth et al., 2001), SA accumulation and expression of pathogenesis-related (PR)-1 (a highly SAresponsive gene) (Cutt et al., 1988) were all increased in leaves systemically infected with CMV (Lewsey et al., 2010a). Experiments using a mutant lacking the 2b gene (CMVD2b; Ryabov et al., 2001) or plants constitutively expressing a 2b transgene showed that the 2b protein is necessary but not sufficient for inducing SA accumulation. This suggested that the 2b protein facilitates or primes elicitation of SA biosynthesis by other viral gene product(s) (Lewsey et al., 2010a). Contrastingly, the CMV P6 silencing suppressor inhibits SA-induced gene expression (Laird et al., 2013).
We investigated various 2b protein domains for roles in evasion of SA-induced resistance or in priming of CMVinduced SA accumulation. We used CMV strain Fny (Roossinck & Palukaitis, 1990). This strain induces severe symptoms in tobacco and A. thaliana and expression of its 2b protein in transgenic plants generates strong developmental phenotypes . Experiments were carried out using WT CMV, CMVD2b, eight CMV variants with a range of site-directed mutations in various 2b protein functional domains (González et al., 2010;, and an additional site-directed mutant harbouring a deletion in the 2b sequence encoding amino acids 62-65 (mutant DGSEL), corresponding to four conserved amino acids within the domain identified by Ye et al. (2009) as conditioning strong RNA silencing suppression by the 2b protein of the SD strain of CMV, although mutation of this sequence did not abolish silencing suppression by the Fny-CMV 2b protein. This is in a region thought to interact with the host silencing factor Argonaute 1 (González et al., 2012). As noted previously, because of the overlap of the CMV 2a and 2b ORFs in CMV RNA 2, mutations in the 2b gene affect the 2a protein sequence (Du et al., 2008;. the effects of some of the CMV 2b mutants (with the exception of CMVD3T, which affects a sequence beyond the 2a/2b gene overlap), we believe this is unlikely. This is because no evidence from previous studies using 2btransgenic plants or CMVD2b variants in which there are no deletions in the 2a reading frame (Ji & Ding, 2001;Lewsey et al., 2010b) suggests any relationship(s) between the 2a protein, SA biosynthesis, or the subversion of SA-induced resistance.
Four-week-old CMV-susceptible tobacco plants ('Xanthi') were sprayed with 2 mM SA or with water [containing 0.11 % (v/v) ethanol] as the control treatment daily for 4 days prior to inoculation on lower leaves with virus. The 2b protein allows CMV accumulation to reach the same level in SA-treated as in control-treated inoculated leaf tissue (Ji & Ding, 2001;Lewsey & Carr, 2009). This was examined in greater detail by imaging infection sites on SA-treated and control leaves inoculated with previously described viral constructs expressing the GFP: CMV-GFP and CMVD2b-GFP (Soards et al., 2002) at 4 days post-inoculation (Fig. 1a). As seen previously, cell-to-cell movement of CMV-GFP was unaffected by SA and, in the absence of SA, CMVD2b-GFP moved preferentially through mesophyll rather than through epidermal cells Soards et al., 2002) (Fig. 1a). CMVD2b-GFP infection sites were difficult to locate on inoculated leaves of SA-treated plants and in those sites observed, GFP was localized to single epidermal cells (Fig. 1a). The behaviour of CMVD2b-GFP appeared similar to that of a GFP-expressing tobacco mosaic virus; cell-to-cell movement of which was inhibited in SA-treated tobacco .
Due to the difficulty of imaging CMVD2b-GFP in SAtreated tobacco plants, immunoblotting was used to detect virus accumulation in most experiments. A suspension of 5 mg ml 21 of purified virions of WT or mutant CMV was mechanically inoculated onto leaves as previously described (Naylor et al., 1998: Soards et al., 2002. WT and mutant CMV genomic RNAs were generated by in vitro transcription. Tobacco plants inoculated with these infectious RNAs were used to purify virions (Westwood et al., 2013). Four days later, protein was extracted from inoculated leaves and analysed for virus accumulation, using CMV coat protein (CP) accumulation as a proxy, by SDS-PAGE and immunoblotting with anti-CMV CP . Each mutant was examined at least three times for the effects of SA on its accumulation in inoculated leaves.
In line with previous results (Ji & Ding, 2001;Lewsey & Carr, 2009) CMV accumulation in the inoculated leaves was unaffected by pre-treatment with SA but accumulation of CMVD2b was markedly inhibited (Fig. 1b). CMV variants carrying mutations in one or both elements of the bipartite, arginine-rich nuclear localization sequence (NLS) that overlaps the key RNA binding domain needed for silencing suppression (González et al., 2012) (mutants DNLS1, DNLS2 and DNLS 1+2, respectively) were, like WT CMV, able to accumulate to readily detectable levels in SA-treated tissue (Fig. 1b). Similarly, mutant viruses possessing 2b proteins with alanine to serine substitutions in one of two putative phosphorylation sites (mutants S40A and S42A), or deletion of both sites (DKSPSE), as well as the DGSEL mutant were able to accumulate in SA-treated leaves (Fig. 1b). However, SA treatment did inhibit accumulation of the CMV mutants D5T, in which the RNA sequence encoding the N-terminal 17 aa of the 2b protein was deleted, and D3T, which expresses a 2b protein lacking its 16 C-terminal residues (Fig. 1b). Paradoxically, although the 2b protein permits CMV to evade SA-induced resistance to local replication and movement (Ji & Ding, 2001;Lewsey & Carr, 2009;Naylor et al., 1998), CMV infection increased expression of SA-regulated genes including PR-1 in A. thaliana (Whitham et al., 2003). These gene expression changes occurred as a result of increased SA accumulation, which is facilitated but not directly elicited by the 2b protein (Lewsey et al., 2010a). We used immunoblotting (with an anti-PR1 serum crossreacting with PRs 1a, 1b and 1c; Carr et al., 1985) to examine accumulation of WT or mutant variants of CMV. Consistent with previous results in A. thaliana (Lewsey et al., 2010a) we found that at 4 days post-inoculation PR1 accumulation was elevated in CMV-inoculated but not in CMVD2b-inoculated leaves of tobacco plants (Fig. 2a). Of the mutants tested, only CMVS40A, CMVS42A and CMVDKSPSE retained the ability to elicit PR1 accumulation to levels comparable to, or slightly higher than, WT CMV (Fig. 2a).
PR1 accumulation is considered a reliable indicator for increased SA levels in tobacco (Malamy et al., 1990). To check that this assumption is correct for CMV-infected tobacco, leaf extracts were analysed using high performance liquid chromatography (Surplus et al., 1998). Analysis using a selection of CMV mutants indicated that PR1 accumulation correlated broadly with SA level (Fig. 2b). The highest SA levels were seen in leaves inoculated with the mutant CMVS40A (Fig. 2b). Interestingly, some PR1 accumulated in leaves infected with CMVDNLS2 ( Fig. 2a) but these leaves showed no apparent increase in SA (Fig.  2b), suggesting the existence of a weak, SA-independent PR-1-inducing mechan.
The mechanism by which the 2b protein facilitates SA accumulation is different from that by which the 2b protein of the HL strain induces increased reactive oxygen species production and necrosis in A. thaliana, since this function was not abolished by the deletion of the Cterminal 12-33 aa (Inaba et al., 2011). Based on previous analysis of 2b protein domain function (González et al., 2010), it appears that the ability of the 2b protein to prime SA and PR1 accumulation is unrelated to silencing suppression. Both NLS domains and the KSPSE domain are required for silencing suppression although mutations of individual phosphorylation sites in the KSPSE domain (S40A and S42A) do not abolish RNA silencing (González et al., 2012). SA and PR1 levels increased in leaves inoculated with CMV variants carrying the S40A and the S42A mutations, as well as the KSPSE mutation. Using GFP-2b protein fusions, González and colleagues (2010) showed that WT 2b and the S40A and S42A mutants accumulate in nuclei and nucleoli to similar extents, whereas DKSPSE mutant 2b protein localizes predominantly to the nucleus and nucleolus and is absent from the cytoplasm. Overall, the data suggest that several 2b protein domains are required for priming of SA biosynthesis. Results obtained with mutations in the putative phosphorylation domain suggest that priming of SA accumulation by the 2b protein requires its accumulation within the nuclear or nucleolar compartments and that the presence of a functional phosphorylation domain limits the extent to which the 2b protein primes SA biosynthesis (Fig. 3).
With respect to the roles of the N-and C-terminal domains in evasion of SA-induced resistance (Fig. 3), previous analyses showed that these domains, respectively, positively and negatively regulate symptom induction by CMV . Thus, deletion of the RNA sequence encoding the C-terminal domain of the 2b protein resulted in the mutant CMVD3T, which induced more severe symptoms than WT CMV in three host species (tobacco, Nicotiana benthamiana and A. thaliana), whilst deletion of the N-terminal domain (CMVD5T) ameliorated symptoms (Lewsey et al., , 2010b. Mutation of 2b protein domains identified as affecting the binding of shortinterfering RNAs and suppression of antiviral RNA silencing (mutants DNLS1, DNLS2, DNLS1+2 and DKSPSE; González et al., 2010, 2012) did not compromise the ability of the virus to evade SA-induced resistance. This finding is consistent with previous data suggesting that RNA silencing does not play an indispensable role in SA-induced resistance to CMV (Lewsey & Carr, 2009). The mode(s) of action through which the C-and N-terminal regions of the 2b protein affect symptom expression remain unknown. The C-terminal domain is relatively unstructured although it appears to have Mg 2+ -binding properties (Gellért et al., 2012). Other experiments in vitro and in yeast, respectively, indicated that the 2b protein N-and C-terminal domains have DNA binding properties and that the C terminus has transcriptional activation activity (Ham et al., 1999;Sueda et al., 2010). However, if the C-terminal domain of the 2b protein was exerting its effects on SA-induced resistance by interacting with host DNA sequences, then mutating the NLS domain of the protein would have compromised the ability of the DNLS1, DNLS2 and DNLS1+2 mutants to overcome SA-induced resistance. Since this did not occur, it suggests that the unknown cellular target(s) for the 2b protein (in its role as a suppressor of SA-induced resistance) lie outside of the nucleus. Interestingly, this contrasts with our analysis of 2b-primed SA synthesis, which suggests that for this role the 2b protein has a nuclear or nucleolar target.