Background
Sobemovirus is a small genus of plant viruses with single-stranded positive-sense RNA genome (for review [
1]). Sobemoviruses have a viral protein genome linked (VPg) covalently attached to the 5′ end of genomic and subgenomic RNA. The 5′ proximal open reading frame 1 (ORF1) encodes the P1 protein, while the viral capsid or coat protein (CP) is expressed from the 3′ proximal ORF3 via subgenomic RNA. The central part of the genome encodes the viral polyproteins P2a and P2ab, the latter is translated through a-1 programmed frameshift mechanism [
2].
Sobemoviral proteins, and CP among them, are multifunctional. CPs primary function is structural, being the building block for the virion. Virions of sobemoviruses have an icosahedral structure according to T = 3 symmetry comprising of 180 CP monomers [
3]. Though the primary sequences of sobemoviral CPs are poorly conserved, their three-dimensional structures are highly similar. Sobemoviral CPs contain two domains–the N-terminal R (random) and C-terminal S (shell) domain which is responsible for subunit-subunit contacts in the virion [
3‐
8]. The N-terminal part of all sobemoviral CPs is rich in basic amino acids and contains an arginine-rich region, which is necessary for CP-RNA interactions and RNA encapsidation [
9‐
11]. It has also been shown that the N-terminal part of
Southern cowpea mosaic virus (SCPMV) CP interacts with membranes [
12] and that the N-terminus of
Cocksfoot mottle virus (CfMV) CP contains functional nuclear localisation signals [
13]. CPs have also been reported to be involved in virus movement. For instance, CPs of
Rice yellow mottle virus (RYMV) and SCPMV are required for cell-to-cell movement, but are not necessary for virus replication [
14,
15]. Moreover, long distance movement of RYMV and SCPMV has been proposed to be dependent on viral particle formation [
16,
17]. The CP of
Sesbania mosaic virus (SeMV) has been shown to interact with P1, which is a suppressor of RNA silencing and has been implicated in virus movement [
18‐
20].
In this study we analyse the role of CfMV CP in suppression of RNA silencing and virus movement. We show that the CP of CfMV acts as an RNA silencing suppressor and that CP is dispensable for cell-to-cell as well as systemic movement of CfMV in oats, wheat and barley.
Discussion
It was previously reported that P1 of CfMV is indispensable for virus movement and accumulation in oats [
19] and that it is a suppressor of RNA silencing [
20]. However, it still remains unclear whether P1 facilitates virus spread as an RNA silencing suppressor or as a member of the movement RNP complex. Here we demonstrate that CP is also an RNA silencing suppressor in 16c
N. benthamiana. It is not a surprise that CfMV encodes two suppressors of RNA silencing, since more than one suppressor has been identified for other viruses as well [
21,
22]. Both suppressors, P1 and CP, are able to interfere with the RNA silencing mechanism independently and a strong synergistic effect was not observed. It seems plausible that CfMV CP contributes to virus spread through enhancement of accumulation. Such claim is supported by experiments with RYMV, which accumulated to higher levels in transgenic plants expressing RYMV CP compared to control plants, indicating that transgenic CP further enhanced virus infection and accumulation [
23]. The authors speculated that CP may enhance viral accumulation by influencing replication or host susceptibility or, alternatively, by suppressing RNA silencing.
The arginine-rich region of the CP of sobemoviruses has been studied before. The CP of mutants R3L and R5X localises to cytoplasm and nucleus whereas the wt CP localises almost exclusively to the nucleus [
13]. Results obtained here, with viruses containing the same mutations in CP, demonstrate that these mutations have no deleterious effect on virus cell-to-cell and systemic movement as well as on mechanical transmission in oats. The majority of CP-EGFP fusion protein localises to the cell nucleus when expressed independently from the rest of the virus genome [
13]. Here we observed that when CP-EGFP was expressed together with the rest of the viral proteins, the fluorescence did not accumulate in the nucleus, but remained evenly distributed between cytoplasm and nucleus. Most probably the CP was interacting with other viral and/or host proteins or with full-length genomic RNA and therefore was not accumulating in the nucleus any more.
Despite these possible interactions, CfMV CP null mutant (noCP) was capable of cell-to-cell as well as systemic movement in all three tested hosts. The fact that CfMV CP is not strictly needed for the infection was further demonstrated by experiments with viruses expressing CP-EGFP and CPdelta-EGFP. As expected, the virus replicated in single epidermal cells and migrated to the mesophyll, beneath these epidermal cells, from where it presumably entered vascular tissue. Spontaneous CfMV recombination mutants detected in plants inoculated with mutants CP-EGFP/CPdelta-EGFP that lacked most of the CP cistron (group I, deleted nt 3254-3864), produced systemic infection as well as symptoms, again demonstrating the dispensability of CfMV CP for cell-to-cell and systemic movement.
In contrast, it has been previously documented that sobemoviruses require P1 as well as CP for systemic movement [
14‐
17]. An RYMV CP mutant failed to infect rice plants systemically, but accumulated in the inoculated leaves, indicating cell-to-cell movement [
14], whereas SCPMV CP initiation codon mutant, analogous to CfMV noCP, was undetectable even in inoculated leaves, but replicated in protoplasts [
15]. The CP of
Turnip rosette virus, another sobemovirus, has been reported to facilitate long distance movement of red clover necrotic mosaic dianthovirus [
24]
. Based on protein interaction studies of P1 with either CP or native virions of SeMV, closely related to SCPMV, Chowdhury and Savithri [
18] have proposed a model for CP involvement in SeMV movement. We conclude that CfMV in general utilises a different movement strategy from SCPMV and RYMV.
Interestingly, we were unable to transmit noCP mechanically, whereas R3L and R5X were transmittable. This demonstrates that CP is necessary for an efficient sap-transmission and that mutations in the arginine-rich region do not affect the viral RNA-CP complex involved in this transmission. The mutant noCP should be, in theory, transmissible as well, since viral RNA is all what is needed for initiating the infection. Nevertheless, we have observed that mechanical inoculation with in vitro synthesised CfMV RNA is less efficient than biolistic inoculation with the same RNA (our unpublished observations). When compared to the R3L and R5X (that were transmissible), mutant noCP did not have significantly lower viral RNA levels in the systemically infected leaves which were used as the source for sap-inoculation (data not shown). Most likely the formation of virus particles is necessary for the efficient CfMV transmission.
Mutant R5X failed to produce systemic infection in mechanically inoculated plants. We propose that this deficiency is due to the mutation in RdRp yielding a less fit virus.
CPdelta-EGFP produced a considerably weaker EGFP signal in infected cells than CP-EGFP. It is possible that the EGFP sequence was just lost more rapidly than in case of CP-EGFP or that CPdelta-EGFP had considerably lower accumulation. It seems that, if available, the virus usually maintains as much of the CP sequence as possible. Foreign EGFP sequence seems to be the trigger for recombination because we have not encountered recombinant viruses in plants infected with mutant noCP or with the wt virus.
It is interesting that the mutants, which had lost the EGFP and CP coding sequences, all contained almost the entire 3′ UTR. The 3′ UTR starts at nt 3858 and the recombinant mutants had retained the sequence starting from nt 3865 or 3869. This leads us to speculate that the 3′ UTR might contain sequences or structural elements important for the transport, as the full-length 3′ UTR is not needed for replication, translation and accumulation in oats (Olspert, unpublished results). Another explanation is that nt 3865-3869 simply contain a hot-spot for recombination. These two hypotheses are, of course, not mutually exclusive.
Conclusions
Sobemoviral P1 is known to be a suppressor of RNA silencing. This study demonstrates that, at least for CfMV, CP is a second suppressor.
The viral RNA genome is usually transported from cell-to-cell and systemically through the vasculature either as virions or as some other form of RNP complex. Trafficking as virions can be now ruled out for CfMV, at least in the tested host plants, since all the experiments reported here indicate that CP is not needed for cell-to-cell and systemic movement.
Altogether we have demonstrated that in different pathosystems individual sobemoviruses can exploit alternative cell-to-cell and long distance movement strategies. It would be interesting to determine whether the CfMV P1 RNA silencing suppressor activity can be uncoupled from the movement function and to determine the composition of the movement RNP complex. The RNA silencing suppression mechanism by which CP contributes to higher virus accumulation of CfMV deserves further investigation as well.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
AO conceived the study, constructed CP mutants and carried out all experiments involving these mutants. KK performed the experiments involving EGFP expressing CfMV viruses. CS and JG carried out the RNA silencing suppression experiments. AO, KK and CS drafted the manuscript. ET conceived the study and helped to draft the manuscript. All authors read and approved the final manuscript.