The Nedd4-Type Rsp5p Ubiquitin Ligase Inhibits Tombusvirus Replication by Regulating Degradation of the p92 Replication Protein and Decreasing the Activity of the Tombusvirus Replicase

Recent in vitro proteomics screens revealed that many host proteins could interact with the replication proteins of Tomato bushy stunt virus (TBSV), which is a small, plus-stranded RNA virus (Z. Li, D. Barajas, T. Panavas, D. A. Herbst, and P. D. Nagy, J. Virol. 82:6911-6926, 2008). To further our understanding of the roles of host factors in TBSV replication, we have tested the effect of Rsp5p, which is a member of the Nedd4 family of E3 ubiquitin ligases. The full-length Rsp5p, via its WW domain, is shown to interact with p33 and the central portion of p92^pol replication proteins. We find that overexpression of Rsp5p inhibits TBSV replication in Saccharomyces cerevisiae yeast, while downregulation of Rsp5p leads to increased TBSV accumulation. The inhibition is caused by Rsp5p-guided degradation of p92^pol, while the negative effect on the p33 level is less pronounced. Interestingly, recombinant Rsp5p also inhibits TBSV RNA replication in a cell-free replication assay, likely due to its ability to bind to p33 and p92^pol. We show that the WW domain of Rsp5p, which is involved in protein interactions, is responsible for inhibition of TBSV replication, whereas the HECT domain, involved in protein ubiquitination, is not necessary for Rsp5p-mediated inhibition of viral replication. Overall, our data suggest that direct binding between Rsp5p and p92^pol reduces the stability of p92^pol, with consequent inhibition of TBSV replicase activity.Various interactions with their host cells are critical for plus-stranded (+)RNA viruses as they attempt to utilize the host translation machinery to produce viral proteins, gain access to the resources of the host cells, co-opt host proteins, and subvert host membranes (1, 17). Additional levels of interaction between virus and host reflect antiviral responses which may involve innate immunity, as well as other antiviral processes and factors. On-going research with several model viruses is striving to map all the interactions between viruses and hosts and characterize the functions of the co-opted host factors. In this regard, recent research has led to the identification of a large number of host proteins which affect the replication of various (+)RNA viruses and minus-stranded RNA viruses (4, 5, 9, 11, 22, 35, 39). The roles and functions of most of the host proteins identified as being involved in RNA virus replication, however, are currently unknown.Tombusviruses, such as Tomato bushy stunt virus (TBSV), are among the most advanced model systems in relation to the identification of host factors affecting (+)RNA virus replication. The TBSV genome codes for only five proteins, two of which are the replication proteins translated directly from the genomic RNA (45). One of these replication proteins is the abundant p33 replication cofactor; the other is the RNA-dependent RNA polymerase (RdRp) p92^pol. Due to the overlapping expression strategy, p33 is identical with the N-terminal portion of the larger p92^pol protein (Fig. ​(Fig.1A).1A). Both replication proteins contain an RNA-binding motif (arginine-proline-rich motif), phosphorylation sites that affect RNA binding by the p33 protein, a p33-p33/p92 interaction domain, and two transmembrane domains (Fig. ​(Fig.1A)1A) (18, 19, 32, 36, 37). Three short stretches of amino acids in p33 and p92^pol are involved in binding to the Pex19p host protein that facilitates the transportation of p33 and p92^pol from the cytosol to the cytosolic surface of the peroxisomes, the site of replicase complex formation and viral RNA replication (25). The essential nature of the above-named domains for obtaining functional replicase complexes suggests that multiple dynamic protein-protein, protein-RNA, and protein-membrane interactions must be required for robust tombusvirus replication.Open in a separate windowFIG. 1.Binding of Rsp5p to TBSV p33 and p92 proteins in vitro. (A) Schematic representation of viral proteins and their derivatives used in the binding assay. The various domains include the transmembrane domain (TMD), arginine-proline-rich RNA-binding domain (RPR), phosphorylated serine and threonine (P), and S1 and S2 subdomains involved in p33-p33/p92 interaction. The two RNA-binding regions in p92 are shown with boxes. (B) Affinity binding (pulldown) assay to detect interaction between GST-six-His-Rsp5p and the MBP-tagged viral proteins. The MBP-tagged viral proteins and MBP produced in E. coli were immobilized on amylose affinity columns. Then, GST-six-His-tagged Rsp5p expressed in E. coli was passed through the amylose affinity columns with immobilized MBP-tagged proteins. The affinity-bound proteins were specifically eluted with maltose from the columns. The eluted proteins were analyzed by Western blotting with anti-six-His antibody to detect the amount of GST-six-His-Rsp5p specifically bound to MBP-tagged viral proteins. (C) The amounts of MBP-tagged proteins eluted from the columns were analyzed by Coomassie blue staining of SDS-PAGE gels. (D) SDS-PAGE analysis of in vitro ubiquitination of replication protein p33 by purified recombinant Rsp5p. The components in the assays are indicated at the top. The ubiquitin-MBP-p33 product, detected by anti-six-His antibody, is marked by an arrowhead. Ub, ubiquitin; +, present; −, absent.In order to identify host genes involved in tombusvirus replication and recombination, systematic genome-wide screens that covered 95% of the host genes were performed in the model host Saccharomyces cerevisiae yeast (9, 22, 34, 35). These screens led to the identification of over 150 host genes, although the functions of these genes in TBSV replication are largely unknown. In addition, proteomics analysis of the highly purified tombusvirus replicase, as well as the use of yeast protein arrays containing ∼4,100 purified proteins to identify host proteins interacting with p33 and/or p92^pol, led to the identification of ∼60 pertinent yeast proteins (12, 33). Current efforts are focused on characterizing the functions of key host proteins in TBSV replication.Most of the host factors identified facilitate tombusvirus replication, though some are inhibitory. The list of characterized host factors includes heat shock protein 70 (Hsp70), which is required for the assembly of the viral replicase in vitro, as well as for membrane insertion and intracellular targeting of the viral replication proteins in vivo (29, 43). Another important host protein is GAPDH (glyceraldehyde-3-phosphate dehydrogenase), which affects plus-strand synthesis (42). The functions of two other host factors that are also present in the replicase complex, namely, Cdc34p E2 ubiquitin-conjugating enzyme, which ubiquitinates p33 replication protein in vitro, and translation elongation factor 1A (eEF1A), which binds to a 3′ cis-acting regulatory element in the TBSV (+)RNA, are not yet characterized with respect to their roles in viral replication (12, 13). Downregulation of all four of the above-described host factors inhibited TBSV accumulation in the yeast model host and in plants (12, 13, 33, 42, 43), suggesting that they are significant players in TBSV replication.In order to further the understanding of host factor roles in viral RNA replication, this paper addresses the effect of Rsp5p E3 ubiquitin ligase on TBSV accumulation. Rsp5p was selected since we have previously found an interaction between p33 and Rsp5p, based on the yeast protein array (12). Also, p33 is mono- and biubiquitinated in yeast cells (12), and Rsp5p is known to ubiquitinate select host proteins (3). These features of Rsp5p suggest its relevance to TBSV replication. Indeed, we found that Rsp5p inhibits TBSV replication when overexpressed in yeast cells, whereas its downregulation leads to increased TBSV accumulation. The inhibition is primarily caused by Rsp5p-mediated selective degradation of p92^pol. Its negative effect on the level of p33 is substantially less. However, the inhibitory function of Rsp5p is more complex, since the purified recombinant Rsp5p also inhibited RNA replication in a cell-free TBSV replication assay, likely due to the ability of Rsp5p to bind to both p33 and p92^pol. Surprisingly, the inhibitory function of Rsp5p is not caused by the HECT domain, which is involved in protein ubiquitination, but by its WW domain, which is involved in protein interactions. The observations suggest that direct binding between Rsp5p and p33 and, more importantly, p92^pol is likely involved in the inhibition of TBSV replication.