|Home | About | Journals | Submit | Contact Us | Français|
Temperature-sensitive (ts) mutants of simian rotavirus (RV) strain SA11 have been previously created to investigate the functions of viral proteins during replication. One mutant, SA11-tsC, has a mutation that maps to the gene encoding the VP1 polymerase and shows diminished growth and RNA synthesis at 39°C compared to that at 31°C. In the present study, we sequenced all 11 genes of SA11-tsC, confirming the presence of an L138P mutation in the VP1 N-terminal domain and identifying 52 additional mutations in four other viral proteins (VP4, VP7, NSP1, and NSP2). To investigate whether the L138P mutation induces a ts phenotype in VP1 outside the SA11-tsC genetic context, we employed ectopic expression systems. Specifically, we tested whether the L138P mutation affects the ability of VP1 to localize to viroplasms, which are the sites of RV RNA synthesis, by expressing the mutant form as a green fluorescent protein (GFP) fusion protein (VP1L138P-GFP) (i) in wild-type SA11-infected cells or (ii) in uninfected cells along with viroplasm-forming proteins NSP2 and NSP5. We found that VP1L138P-GFP localized to viroplasms and interacted with NSP2 and/or NSP5 at 31°C but not at 39°C. Next, we tested the enzymatic activity of a recombinant mutant polymerase (rVP1L138P) in vitro and found that it synthesized less RNA at 39°C than at 31°C, as well as less RNA than the control at all temperatures. Together, these results provide a mechanistic basis for the ts phenotype of SA11-tsC and raise important questions about the role of leucine 138 in supporting key protein interactions and the catalytic function of the VP1 polymerase.
IMPORTANCE RVs cause diarrhea in the young of many animal species, including humans. Despite their medical and economic importance, gaps in knowledge exist about how these viruses replicate inside host cells. Previously, a mutant simian RV (SA11-tsC) that replicates worse at higher temperatures was identified. This virus has an amino acid mutation in VP1, which is the enzyme responsible for copying the viral RNA genome. The mutation is located in a poorly understood region of the polymerase called the N-terminal domain. In this study, we determined that the mutation reduces the ability of VP1 to properly localize within infected cells at high temperatures, as well as reduced the ability of the enzyme to copy viral RNA in a test tube. The results of this study explain the temperature sensitivity of SA11-tsC and shed new light on functional protein-protein interaction sites of VP1.
Rotaviruses (RVs) are segmented, double-stranded RNA (dsRNA) viruses and gastrointestinal pathogens of many animal species (1). In unvaccinated children, RVs cause severe watery diarrhea and vomiting, killing an estimated 215,000 children each year (2). The RV virion is a nonenveloped icosahedral particle with three concentric protein layers (3, 4). The outermost layer is composed of two proteins (VP4 and VP7), whereas the middle layer and inner core shell are each made of a single protein (VP6 and VP2, respectively). Within the viral core reside the 11 dsRNA genome segments (i.e., genes) and several copies each of the viral RNA polymerase (VP1) and RNA capping enzyme (VP3). Altogether, the genome codes for six structural proteins (VP1 to VP4, VP6, and VP7), as well as five or six nonstructural proteins (NSP1 to NSP5 and NSP6 in some strains) that play various roles during intracellular viral replication.
Studies of RV protein function have been hampered by the lack of efficient reverse genetic approaches that allow for the targeted engineering of viral mutants. However, several temperature-sensitive (ts) mutant viruses have been created by using forward genetic approaches (5,–8). In particular, one panel of ts mutants was generated via chemical mutagenesis of prototypic laboratory strain SA11 (H96 derivative, referred to here as SA11-H96) (7). Reassortment analyses were used to map ts lesions of several mutants to individual genes, i.e., SA11-tsA (VP4), SA11-tsB (VP3), SA11-tsC (VP1), SA11-tsE (NSP2), SA11-tsF (VP2), and SA11-tsG (VP6) (9). SA11-tsC mapped to the gene encoding the VP1 polymerase, and this mutant shows diminished viral growth, RNA synthesis, protein synthesis, and virion morphogenesis at 39°C compared to that at 31°C (10,–12). By using a general anti-RV antiserum and epifluorescence microscopy, it was also shown that viral proteins were more diffusely distributed in the cytosol of SA11-tsC-infected cells than in that of the other ts mutants from this panel, which showed a more punctate localization pattern (10). The phenotype of SA11-tsC is less severe at 31°C, yet the virus still exhibits reduced RNA synthesis and diffuse protein localization even at the lower, more permissive temperature (10, 12). Sequencing of the VP1-coding gene from SA11-tsC identified a single leucine-to-proline change at position 138 (L138P) (13). It remains unknown whether the L138P lesion causes the VP1 protein to have a ts phenotype outside the genetic context of the virus. Moreover, the sequences of the other 10 genes of SA11-tsC have not yet been deduced, preventing an understanding of how additional mutations may contribute to its replication defect.
VP1 is a compact, globular 125-kDa protein and is organized as three distinct domains, an N-terminal domain (residues 1 to 332); a central polymerase domain with canonical fingers, palm, and thumb subdomains (residues 333 to 778); and a C-terminal “bracelet” domain (residues 779 to 1089) (Fig. 1) (14). Together, the N- and C-terminal domains enclose the central polymerase domain to create a cage-like enzyme with a buried active site. The N-terminal domain is not directly involved in catalysis, and such a domain is missing from all other viral RNA polymerases outside the Reoviridae family (15). This region of VP1 helps form a template RNA entry tunnel, and it may also aid in polymerase regulation by serving as a binding platform for other viral proteins. VP1 is known to engage at least three other viral proteins (VP2, NSP2, and NSP5) during the RV life cycle, though the interaction interfaces are poorly defined (16,–19). Binding of VP1 by NSP2 and/or NSP5 localizes the polymerase to viroplasms, sites of genome replication and early particle assembly in the cell cytosol (20, 21). Within the viroplasm, binding of VP1 by core shell protein VP2 triggers the polymerase to initiate genome replication (i.e., dsRNA synthesis) in the context of subviral assembly intermediates (18, 19, 22). Because the VP1 protein of SA11-tsC has an L138P mutation in the N-terminal domain, we wondered whether it affects the abilities of the polymerase to localize to viroplasms, interact with other viral proteins, and synthesize viral dsRNA in a temperature-dependent manner.
In this study, we sequenced all 11 genes of SA11-tsC, confirming the presence of the L138P mutation in the VP1 N-terminal domain and also identifying an additional 52 changes in four other viral proteins (VP4, VP7, NSP1, and NSP2). By employing VP1 ectopic expression systems, we show that the L138P mutation alone abrogates NSP2 and/or NSP5 interactions at high temperatures and prevents efficient localization of the polymerase to viroplasms. We further show that the L138P mutation alone reduces the enzymatic activity of the polymerase in vitro in a temperature-dependent manner. Altogether, our findings increase our mechanistic understanding of the basis for the ts phenotype of the SA11-tsC phenotype and raise interesting questions about the importance of leucine 138 and its involvement in polymerase regulation.
Previous studies have determined the VP1-coding gene sequence for SA11-tsC, revealing an L138P lesion in the N-terminal domain of the polymerase (13). However, the sequences of the other 10 genes of SA11-tsC had not yet been determined, limiting our understanding of additional mutations that may contribute to the replication-defective ts phenotype of this virus. Therefore, we employed reverse transcription (RT)-PCR and the Sanger approach to determine the sequences of all 11 SA11-tsC gene open reading frames (ORFs). To identify changes that may contribute to the SA11-tsC phenotype, we aligned the deduced ORF nucleotide and amino acid sequences with those of parental strain SA11-H96 (7) (Table 1). Altogether, we identified 219 nucleotide and 53 amino acid differences when comparing SA11-tsC to SA11-H96; the vast majority of these differences were in the NSP2-coding gene. BLAST analyses revealed that the SA11-tsC NSP2-coding gene is identical to that of O agent, an RV strain isolated in 1965 from a slaughterhouse (23). Indeed, an amino acid sequence alignment shows that SA11-tsC NSP2 shows no amino acid changes compared to O agent NSP2 but that this protein differs from that of the parental SA11-H96 strain at 45 residues (Fig. 2). BLAST analyses of the other 10 SA11-tsC genes verified that they are most closely related to those of parental strain SA11-H96. This result suggests that SA11-tsC is a monoreassortant with 10 genes derived from SA11-H96 and an NSP2-coding gene derived from O agent.
Not counting the reassorted NSP2-coding gene, there are total of 16 nucleotide and 8 amino acid differences between SA11-tsC and SA11-H96 (Table 1). The VP1-coding gene of SA11-tsC had only the single previously reported nucleotide change, which results in the L138P lesion in the N-terminal domain of the polymerase (13). The VP4-coding gene had eight nucleotide changes, five of which (T72M, P157S, A187G, Y332S, and V366M) caused amino acid changes in the protein. The VP7-coding gene had two nucleotide changes, one of which (T60A) caused an amino acid change in the protein. Likewise, two nucleotide changes were identified in the NSP1-coding gene, one of which (K39N) caused an amino acid change in the protein. Three nucleotide changes were identified in the VP3- and VP6-coding genes, but none of them resulted in changes in the viral proteins. No nucleotide changes were found in the VP2-, NSP3-, NSP4-, and NSP5/6-coding genes.
These results demonstrate that, in addition to the VP1 L138P lesion, the SA11-tsC virus has multiple differences from the parental strain in its genome, including 52 amino acid changes in four other viral proteins (Table 1). Because we currently lack reverse genetic technologies to recapitulate the VP1 L138P mutation in an isogenic SA11-H96 background, it is not possible to determine its contribution toward the ts phenotype in the context of an infectious virus. Therefore, in this study, we sought to investigate the effect of the L138P lesion alone on the temperature-dependent localization and enzymatic activity of the VP1 protein by using ectopic expression systems.
To test the effect of the L138P lesion on the ability of VP1 to localize to viroplasms at various temperatures, we first needed to establish an ectopic expression system that would allow us to track the mutant polymerase in wild-type (WT) virus-infected cells. Therefore, we engineered WT and mutant VP1 proteins with green fluorescent protein (GFP) fused to their C termini (VP1L138P-GFP and VP1WT-GFP, respectively) and determined their localization relative to that of unfused GFP by using immunofluorescence confocal microscopy. For these experiments, Cos-7 cells were used because they are permissive for RV infection and show high plasmid DNA transfection efficiencies. Strain SA11-4F was used as the WT RV because it infects Cos-7 cells significantly better than does SA11-H96 as a result of changes in its VP4 attachment protein (13). However, we note that the VP1 proteins of SA11-4F and SA11-H96 are 100% identical, eliminating any concerns about potential strain-specific differences in VP1-protein interactions.
To verify that VP1WT-GFP localized to viroplasms in infected cells and that unfused GFP did not, these control proteins were expressed in Cos-7 cells on glass coverslips for 48 h at 37°C (Fig. 3). The cells were then either mock infected or infected with SA11-4F and incubated for 8 h at 37°C. The confocal micrographs showed that both unfused GFP and VP1WT-GFP were diffusely distributed throughout the cytosol of mock-infected cells (Fig. 3A and andB).B). The localization pattern of unfused GFP was the same in infected and mock-infected cells, with the vast majority of the protein localizing in the cytosol and no visible enrichment in viroplasms, which were stained with antiserum against VP2 or NSP2 (Fig. 3A and data not shown). In contrast, nearly all of the visible VP1WT-GFP colocalized with viroplasms and very little diffuse protein was seen in the cytosol (Fig. 3B). These results suggest that ectopically expressed WT VP1 is capable of being recruited to/retained within viroplasms and that the C-terminal GFP tag does not interfere with any interactions that mediate its viroplasmic localization.
To determine the effect of the L138P lesion on VP1 localization, VP1L138P-GFP, VP1WT-GFP, or unfused GFP was expressed in Cos-7 cells on glass coverslips at 31°C, 37°C, or 39°C for 48 h (Fig. 4). Thereafter, cells were either mock infected (data not shown) or infected for 12 h at 31°C or for 8 h at 39°C. The percentage of expressed VP1WT-GFP, VP1L138P-GFP, or GFP that colocalized with viroplasms (i.e., anti-VP2 staining) was quantified for several individual cells and averaged (Fig. 5). The expression levels of VP1WT-GFP, VP1L138P-GFP, and GFP were equivalent at all of the temperatures tested, as determined by average signal intensities and Western blot analyses (data not shown). The results show that when the infections proceeded at 31°C, the average percent colocalization of VP1L138P-GFP with viroplasms was similar to that of VP1WT-GFP (Fig. 4A and andCC and Fig. 5). However, when the infections proceeded at 39°C, the average percent colocalization of VP1L138P-GFP with viroplasms was significantly less than that of VP1WT-GFP (Fig. 4D, ,B,B, and andFF and Fig. 5). Specifically, there was no visible enrichment of VP1L138P-GFP in viroplasms at this high temperature, and instead, the mutant protein was largely diffuse (Fig. 4D and andFF and Fig. 5). In fact, the average percentage of VP1L138P-GFP that localized to viroplasms at 39°C was indistinguishable from that of unfused GFP (Fig. 5). Together, these results suggest that the L138P lesion abrogates the ability of VP1 to efficiently be recruited to/retained within viroplasms in a temperature-dependent manner. Interestingly, we noticed that while the ability of VP1L138P-GFP to localize to viroplasms was dependent upon the infection temperature, it was independent of the temperature at which the protein was expressed (Fig. 4C to toFF and and5).5). For example, even following its expression at 31°C, the average percent colocalization of VP1L138P-GFP with viroplasms was still significantly reduced relative to that of VP1WT-GFP after 8 h of infection at 39°C (Fig. 4F and and5).5). This observation suggests that the increased infection temperature of 39°C can cause defects in VP1L138P-GFP, even following its expression (and presumably its proper folding) at the lower temperature of 31°C.
It has been previously demonstrated that when NSP2 and NSP5 are coexpressed in uninfected cells, they interact to form discrete structures resembling viroplasms (i.e., viroplasm-like structures) (20). It has also been shown that ectopically expressed WT VP1 is recruited to/retained within these inclusions via its interaction(s) with NSP2 and/or NSP5 (21). Therefore, we sought to investigate whether the defect in VP1L138P-GFP viroplasmic localization at 39°C was due to lack of interactions with NSP2 and/or NSP5 by testing its ability to localize to viroplasm-like structures. For these experiments, Cos-7 cells on glass coverslips were transfected with plasmids expressing VP1WT-GFP or VP1L138P-GFP along with WT NSP5 of strain SA11-4F and either WT NSP2 of strain SA11-4F (i.e., NSP2SA11) or WT NSP2 of strain SA11-tsC (i.e., NSP2Oagent) (Fig. 6). Following incubation of the cells at 31°C or 39°C for 48 h, protein localization was determined by immunostaining and confocal microscopy. Antiserum against NSP2 served as a marker for punctate viroplasm-like structures in cells expressing NSP2SA11 (Fig. 6A to toD).D). Antiserum against NSP5 served as a marker for punctate viroplasm-like structures in cells expressing genetically divergent NSP2Oagent (Fig. 6E and andF).F). No punctate structures were detected in cells expressing VP1 and NSP2 but lacking NSP5 (data not shown). The percentage of expressed VP1WT-GFP or VP1L138P-GFP that colocalized with NSP2 or NSP5 in viroplasm-like structures was quantified in several individual cells at both temperatures and averaged (Fig. 7).
Our results show that VP1WT-GFP was enriched in punctate inclusions and colocalized with NSP2 or NSP5 (i.e., viroplasm-like structures) at both 31°C and 39°C (Fig. 6A and andBB and and7).7). For VP1L138P-GFP, however, we observed a temperature-dependent difference in the viroplasm-like structure localization. Specifically, at 31°C, the average percent colocalization of VP1L138P-GFP with viroplasm-like structures was the same as that of VP1WT-GFP (Fig. 6C and andEE and and7).7). In contrast, at 39°C, much less VP1L138P-GFP than VP1WT-GFP colocalized with viroplasm-like structures (Fig. 6D and andFF and and7).7). Importantly, the localization pattern of VP1L138P-GFP was the same irrespective of which strain's NSP2 was used (i.e., NSP2SA11 or NSP2Oagent). These results suggest that the L138P lesion causes a temperature-dependent defect in VP1 that prevents it from binding to NSP2 and/or NSP5, which in turn prevents it from being recruited to/retained within viroplasms.
We next sought to determine if the L138P lesion affects the enzymatic activity of the polymerase. To test this, we expressed and purified the recombinant mutant protein (rVP1L138P) and assayed it for the ability to support VP2-dependent dsRNA synthesis in vitro at 31°C, 37°C, or 39°C compared to that of the WT control protein (rVP1WT) or a catalytically inactive mutant protein (rVP1D632A) (Fig. 8). We found that the solubility of rVP1L138P was indistinguishable from that of rVP1WT, even following heating at 39°C for 3 h, suggesting that the recombinant mutant polymerase was not grossly misfolded (Fig. 8A and data not shown). The amount of dsRNA synthesized by rVP1L138P after a 180-min incubation was significantly diminished relative to that of the control at all of the temperatures tested (Fig. 8B and andC).C). In particular, at 31°C, 37°C, and 39°C, the levels of RNA produced by rVP1L138P were ~73%, ~63% and ~50%, respectively, of those produced by rVP1WT (Fig. 8C). However, the mutant was more sensitive to higher temperatures, as the percentage of dsRNA made by rVP1L138P at 39°C was ~65% of that made at 31°C while that made by rVP1WT was ~93%, (Fig. 8C). To determine if the temperature-induced defect is reversible, rVP1L138P was incubated at 39°C for 60 min and then assayed for dsRNA synthesis at 31°C for 180 min (Fig. 8D). The results showed that rVP1L138P synthesized less dsRNA following a high-to-low temperature shift than when it was maintained at 31°C during the entire reaction (Fig. 8D). However, the levels of dsRNA were higher than when rVP1L138P was maintained at 39°C during the entire reaction (Fig. 8D). This result suggests that the temperature-induced defect in rVP1L138P was partially reversible or that it was reversible in some fraction of the protein population. Time course analyses indicated no obvious kinetic delay for rVP1L138P at 39°C, as the levels of dsRNA plateaued by 60 min and never reached those of rVP1WT even following longer incubation (Fig. 8E and andF).F). These results suggest that the L138P lesion in VP1 reduces its activity in a temperature-dependent manner, possibly because of local structural alterations in the protein.
To gain insight into the possible temperature-induced structural changes resulting from the VP1 L138P lesion, unrestrained molecular dynamics simulations were performed. Structures of WT and L138P mutant VP1 were each simulated for 20 ns at 312K to approximate the conditions under which the mutant shows reduced viroplasmic localization and enzymatic activity (i.e., 39°C). Over the last 5 to 10 ns of the simulation, both WT and L138P mutant VP1 had similar backbone root mean square deviations compared to the simulation starting structures (data not shown), suggesting that the mutation does not induce a globally destabilizing effect in VP1 at a high temperature. To identify putative locally destabilizing impacts of the L138P mutation, root mean square fluctuations (RMSFs) of the α carbons (WT VP1 versus L138P mutant VP1) were used to calculate B factors for each residue. We found small but significant differences in B factors at positions proximal to 138 (residues S137, L/P138, S140, and L141) (Fig. 9A and andB).B). More dramatic differences were detected in a modeled flexible loop (residues 346 to 358) of the polymerase domain, specifically for residues K348, E350, Y351, and D356 (Fig. 9A and andC)C) (14, 24). Thus, in silico analyses predict subtle, temperature-dependent changes in the region proximal to the L138P lesion of the VP1 N-terminal domain, as well as in a distal loop element of the polymerase domain.
Mutant viruses with ts phenotypes have been widely used to interrogate viral protein function and to engineer live-attenuated vaccines (25). Because their mutations are conditionally lethal only at specific temperatures, ts mutants are particularly useful when studying viral proteins that are essential for replication or those that mediate multiple, distinct steps of the replication cycle. For RV, several different panels of ts mutants have been generated and characterized to various degrees over the years (5,–8). The most well-studied of these were created by using the prototypic simian RV strain SA11-H96 in the early 1980s by Frank Ramig (7). These so-called “Baylor mutants” show diminished replication when infections proceed at the nonpermissive temperature of 39°C rather than the more permissive temperature of 31°C. By genetically crossing individual mutant clones with rhesus RV and screening for ts progeny, Ramig and colleagues mapped the ts lesions to individual viral genes (SA11-tsA, VP4; SA11-tsB, VP3; SA11-tsC, VP1; SA11-tsE, NSP2; SA11-tsF, VP2; SA11-tsG, VP6) (9). However, prior to this study, SA11-tsE was the only Baylor mutant whose ts phenotype had been biochemically validated at the protein level (26). Here, we provide similar biochemical validation for SA11-tsC and show that an L138P mutation in the VP1 protein is a true ts lesion that abrogates polymerase localization and enzymatic activity at 39°C.
SA11-tsC shows defects in growth, RNA synthesis, protein synthesis, virion morphogenesis, and protein localization at 39°C compared to 31°C (10,–12). Yet, even when incubated at the lower temperature of 31°C, the mutant still does not replicate as efficiently as the WT virus (12). This observation suggests that SA11-tsC also has a general replication defect in addition to a ts phenotype, and it is possible that the L138P mutation in VP1 is responsible for both effects. In support of this idea, we found that the in vitro enzymatic activity of the recombinant mutant protein rVP1L138P at 31°C was ~27% lower than that of the WT control rVP1WT. While modest, such a reduction in vitro could translate into larger defects in the infected cell, where multiple rounds of robust RNA synthesis are required for successful infection. Alternatively, it could be the combination of mutations in the SA11-tsC genome that accounts for its poor growth at low temperatures. For instance, we discovered that the SA11-tsC NSP2 gene exhibits 203 nucleotide changes (45 amino acid changes) compared to SA11-H96 but that it is 100% identical to the NSP2 gene of the O agent strain, which is a putative bovine RV that was isolated from slaughterhouse remnants (23). SA11-tsC likely acquired this NSP2 gene at some point by reassortment with O agent following unintentional culture contamination. We do not have O agent in our laboratory, suggesting that the contamination occurred during earlier SA11-tsC genesis or propagation, similar to what has been reported for other SA11 strains (13). In addition to the reassorted NSP2 gene, we also identified several noncoding changes throughout the genome, as well as five coding changes in VP4 (viral attachment protein) and one coding change each in VP7 (viral outer capsid protein) and NSP1 (viral innate immune antagonist). The five changes in VP4 (T72M, P157S, A187G, Y332S, and V366M) are found in other SA11 derivatives and are reported to correlate with differences in viral plaque size (27). For VP7, the single T60A change is reported in at least one other SA11 sequence in GenBank, but the consequence of this change is not known. The single K39N change in NSP1 is unique to SA11-tsC and is not reported for another other group A RV whose sequence is available. It is possible that these mutations together contribute to the general replication defect of SA11-tsC. However, reverse genetics would be required to formally test this notion and determine the effect of each mutation on viral replication. Nevertheless, the results of this study shed new light on the mechanistic basis of the phenotype of SA11-tsC.
Two distinct classes of ts mutant proteins (thermolabile and folding) have been described on the basis of how they behave following low-to-high temperature shift experiments (28). In particular, thermolabile mutants show functional defects at high temperatures, even when they are synthesized at lower, permissive temperatures. In contrast, folding mutants do not show defects when shifted to higher, nonpermissive temperatures if they are synthesized at lower, permissive temperatures. Our results suggest that the L138P lesion causes VP1 to behave as a thermolabile mutant as opposed to a folding mutant. For instance, even when VP1L138P-GFP was expressed at 31°C, a temperature that allows for proper folding, the mutant protein was still defective in the ability to localize to viroplasms during 39°C infections. Likewise, in vitro dsRNA synthesis by rVP1L138P was diminished at 39°C despite the expression of the protein at room temperature in insect cells. For thermolabile mutants, higher temperatures induce structural alterations of the protein, thereby abrogating its interactions and functions (28). Of course, the extent of the structural alterations (i.e., global versus local) impacts the conclusions that can be drawn from the loss-of-function phenotype. For RV VP1, we think that the L138P lesion causes local alterations at high temperatures as opposed to global changes in the overall protein fold. In support of this notion, we showed that when VP1L138P-GFP was expressed at the high temperature of 37°C, the mutant protein was able to localize to viroplasms with the same efficiency as VP1WT-GFP when shifted to 31°C during infection. This result suggests that the ability to target to viroplasms at the lower temperature was restored in some fraction of the mutant VP1 protein population. Likewise, following incubation at 39°C, we found no differences in the solubility of recombinant mutant protein rVP1L138P, and we showed that its activity was partially restored when it was shifted to 31°C. Because VP1 is a large, globular protein, it is unlikely that we would detect any reversibility of protein function at a low temperature if there were gross structural alterations. Even more, molecular dynamics simulations did not predict a global destabilization of the L138P mutant protein at high temperatures. Instead, the in silico analyses showed only subtle temperature-dependent structural changes in a region proximal to the ts lesion (i.e., around position 138) and in a flexible loop (residues 346 to 358) located in the polymerase domain (14, 24).
If it is the case that the mutation induces only local changes, then the results presented here may shed new light on the functional regions of VP1. The leucine at position 138 is conserved in all group A RVs, and it resides in the N-terminal domain of the protein. In the atomic structure of SA11-4F VP1, the N-terminal domain wraps around one side of the polymerase domain and creates a continuous surface between the finger and thumb subdomains, effectively closing the enzyme (14). While it is not directly involved in catalysis, the N-terminal domain is essential for VP1 activity because it helps create a channel for the entry of single-stranded RNA templates into the catalytic interior and is involved in sequence-specific recognition of viral RNAs (14, 29). The large surface-exposed area of the N-terminal domain may also provide a platform for the binding of regulatory cofactor proteins (e.g., VP2, NSP2, and NSP5), in turn orchestrating VP1 localization and function (15, 17). A leucine at position 138 in the N-terminal domain may be critical for maintaining such intra- and intermolecular protein interactions. In silico simulations of VP1 structural dynamics also revealed that the L138P mutation increases the movement of a highly flexible loop in the polymerase domain (residues 346 to 358), which may impede VP1-protein interaction and/or polymerase activity. Ongoing and future experiments in our laboratory are employing biochemical approaches (e.g., protease accessibility mapping, surface plasmon resonance, gel shift assays, etc.) to further investigate the effect of the L138P mutation on VP1 folding, structural dynamics, and protein-protein and protein-RNA interactions. We are also keen to elucidate the role, if any, the flexible loop plays in VP1 localization and activity.
In addition to enhancing a mechanistic understanding of the ts phenotype of SA11-tsC, this study also describes the development of a novel system to study VP1 localization determinants. More specifically, we show that VP1 can target to viroplasms when it is ectopically expressed as a GFP fusion protein. The WT control VP1WT-GFP was highly enriched in viroplasms in infected cells at all of the temperatures tested, suggesting (i) that the fusion protein does not drastically impede viroplasm formation or viral replication and (ii) that the GFP tag does not prevent interactions required for VP1 recruitment to or retention within viroplasms. It was previously shown by Contin et al. that SA11 VP1 with an N-terminal SV5 tag localizes to viroplasm-like structures formed of NSP2 and NSP5, as well as to structures formed of VP2 and NSP5 (21). We did not detect viroplasm-like structure formation when using VP2 and NSP5 (data not shown), perhaps because of several experimental differences between our study and the previous one (i.e., differences in the RV strain, cell type, fixation method, and protein expression conditions used). Nevertheless, our results confirm and extend the finding that ectopically expressed VP1 localizes to inclusions made of NSP2 and NSP5 (21). An advantage of tagging VP1 with GFP, as opposed to SV5, is that it affords an opportunity to perform live-cell imaging. Future experiments in our laboratory will employ VP1WT-GFP, VP1L138P-GFP, and other engineered mutant proteins to study the temporal dynamics of VP1 viroplasmic localization (timing of recruitment and retention) to identify cis-acting residues within VP1 critical for viroplasmic targeting and define the trans-acting factors (i.e., NSP2, NSP5, VP2, cellular protein, etc.) that regulate polymerase recruitment to and retention within viroplasms. We also seek to determine whether ectopically expressed VP1 can substitute for the virus-expressed protein, as such a transcomplementation assay would allow us to probe functional domains of the RV polymerase in the context of the infected cell.
In summary, the results of this study clarify the genetic identity of SA11-tsC and validate the contribution the L138P lesion to the ts phenotype of this mutant virus. These results also raise new questions about the importance of leucine 138 in VP1 and describe a novel assay system to investigate VP1 localization determinants. This work is significant because it informs an understanding of RV replication, which may in turn stimulate the development of novel therapeutics to prevent or treat viral disease.
Monkey kidney cell lines (MA104 and Cos-7) were obtained from the American Type Culture Collection and maintained as described by Arnold et al. (30) in either medium 199 (Life Technologies) or Dulbecco modified Eagle medium (Life Technologies) that was supplemented to contain 10% heat-inactivated fetal bovine serum, 100 U/ml penicillin, 100 μg/ml streptomycin, and 0.5 μg/ml amphotericin B (30). WT SA11-4F and SA11-tsC (clone 606) were generously provided by John Patton (University of Maryland, College Park) with permission from Frank Ramig (Baylor University). Infections with WT SA11-4F or SA11-tsC were performed as described previously (30). Guinea pig polyclonal antiserum against VP2 (anti-VP2-SM) was generated against the baculovirus-expressed recombinant protein (SA11-4F strain) by the Pocono Rabbit Farm & Laboratory, Inc. (Canadensis, PA). Guinea pig polyclonal antisera against NSP2 (anti-NSP2-53962) and NSP5 (anti-NSP5-53964) were generously provided by John Patton (University of Maryland, College Park).
Plaque-purified SA11-tsC was used to infect MA104 cells in a 75-cm2 culture flask (~1 × 107 cells) at 31°C for 72 h. Total RNA was extracted from the clarified cell supernatant with TRIzol LS Reagent (Life Technologies) in accordance with the manufacturer's protocol. RT-PCR was performed with the SuperScript One-Step RT-PCR System by using Platinum Taq DNA Polymerase (Life Technologies). Primers were designed on the basis of the published gene sequences of SA11-H96 (GenBank accession numbers DQ838610, DQ838615, DQ838620, DQ838625, DQ838630, DQ838635, DQ838637, DQ838640, DQ838645, DQ838650, and DQ841262). For the NSP2-coding gene, primers were also designed on the basis of the O agent strain (GenBank accession number DQ838599). In general, the primers were designed to amplify <1-kb regions of each viral gene. The sequences of all of the primers used in this study for RT-PCR and sequencing are available by request. RT-PCRs were performed in accordance with the manufacturer's instructions, with the exception that the RNA template was denatured at 95°C for 10 min in 50% dimethyl sulfoxide prior to being used. The amplified cDNA products were gel purified and cleaned with the QIAquick gel extraction kit (Qiagen) in accordance with the manufacturer's protocol prior to being sent to the Biocomplexity Institute of Virginia Tech (Blacksburg, VA) for Sanger sequencing.
Contigs were assembled de novo from the raw sequence data with Geneious Pro v6.8.1 (Biomatters, Inc.). For each of the 11 SA11-tsC genome segments, the entire ORF was sequenced; for some genes, portions of 5′ and 3′ untranslated regions were also sequenced. The ORF sequences were deposited in GenBank. The basic local alignment search tool (BLAST) was used to determine the likely parental origin of each of the 11 SA11-tsC genes (http://blast.ncbi.nlm.nih.gov/Blast.cgi). To identify nucleotide and amino acid differences, the deduced SA11-tsC gene sequences were aligned with those of strains SA11-H96 (GenBank accession numbers DQ838610, DQ838615, DQ838620, DQ838625, DQ838630, DQ838635, DQ838637, DQ838640, DQ838645, DQ838650, and DQ841262) and O agent (NSP2-coding gene only; GenBank accession number DQ838599).
The ORF sequences of WT or L138P mutant VP1 were cloned into the pEGFP-N1 vector (Clontech) to express the polymerases with enhanced GFP fused to their C termini (VP1WT-GFP or VP1L138P-GFP, respectively). The WT SA11-4F VP1 ORF was amplified from the pENTR-SA11 VP1 vector by PCR (18). Primer-generated restriction sites (5′ SacI and 3′ KpnI) were used to subclone the WT SA11-4F VP1 ORF into the pEGFP-N1 vector, thereby creating pEGFP-SA11-VP1WT, which expresses VP1WT-GFP. Outward PCR and site-directed mutagenesis were then used to engineer the L138P mutation into the SA11-4F VP1 ORF with pEGFP-SA11-VP1WT as the template, thereby creating pEGFP-SA11-VP1L138P, which expresses VP1L138P-GFP. Both expression vectors were sequenced to verify the integrity of the VP1 ORF, maintenance of the translational frame with GFP, and the absence of second-site mutations.
To express VP1WT-GFP, VP1L138P-GFP, or the unfused GFP control, Cos-7 cells grown on glass coverslips in 12-well plates (~0.5 × 106 cells/well) were transfected with 1 μg of plasmid DNA (pEGFP-SA11-VP1WT, pEGFP-SA11-VP1L138P, or the pEGFP-N1 empty vector, respectively) with Trans-IT-LT1 (Mirus) in accordance with the manufacturer's instructions. The transfected cells were incubated for 48 h at 31°C, 37°C, or 39°C, depending upon the experiment, to allow for protein expression. Thereafter, the transfected cells were either mock infected or infected with SA11-4F at a multiplicity of infection (MOI) of 10 PFU/cell. Infections proceeded at 31°C, 37°C, or 39°C, depending upon the experiment. At 8 to 12 h postinfection, the cells on coverslips were washed once with Dulbecco's phosphate-buffered saline (DPBS) prior to being fixed and permeabilized by incubation in 100% methanol for 5 min at room temperature. The coverslips were then stored in DPBS at 4°C for <3 days prior to immunostaining.
The NSP2 ORFs of strains SA11-4F and SA11-tsC were amplified from viral RNA by RT-PCR and subcloned into the pCI mammalian expression vector (Promega) at the 5′ XhoI and 3′ XbaI restriction sites, yielding pCI-NSP2SA11 and pCI-NSP2Oagent. The NSP5 ORF of strain SA11-4F was similarly subcloned into pCI except at the 5′ XhoI and 3′ NotI restriction sites, yielding pCI-NSP5. Expression vectors were sequenced to verify the integrity of the cloned ORFs. Cos-7 cells on glass coverslips were cotransfected with pCI vector pCI-NSP5 along with either pCI-NSP2SA11 or pCI-NSP2Oagent to express NSP5 along with either NSP2SA11 or NSP2Oagent. To test whether VP1WT-GFP or VP1L138P-GFP could localize to viroplasm-like structures, the respective expression vectors (pEGFP-SA11-VP1WT and pEGFP-SA11-VP1L138P) or that expressing unfused GFP (pEGFP-N1) were included in the transfection mixture. In all experiments, Cos-7 cells grown on glass coverslips in 12-well plates (~0.5 × 106 cells/well) were transfected with a total of 1 μg of plasmid DNA with Trans-IT-LT1 (Mirus) in accordance with the manufacturer's instructions. The transfected cells were incubated for 48 h at 31°C or 39°C, depending upon the experiment, to allow for protein expression. The cells on coverslips were washed once with DPBS prior to being fixed and permeabilized by incubation in 100% methanol for 5 min at room temperature. The coverslips were then stored in DPBS at 4°C for <3 days prior to immunostaining.
Coverslips were blocked overnight at 4°C with rocking with a solution of 5% bovine serum albumin (BSA) in DPBST (1.6 mM KH2PO4, 2.7 mM KCl, 8.1 mM N2HPO4, 137.1 mM NaCl, 0.1% Triton X-100). Following three 10-min washes with DPBST, primary guinea pig antiserum was added at a 1:1,000 dilution in a solution of 2.5% BSA in DPBST for 1 h of incubation at room temperature with rocking. Following three 10-min washes with DPBST, the Alexa Fluor 546 goat anti-guinea pig IgG (Life Technologies) secondary antibody was added at a 1:10,000 dilution in 2.5% BSA in DPBST for 30 min of incubation at room temperature with rocking. Following three 10-min washes with DPBST, the stained coverslips were inverted and mounted onto glass microscope slides with the ProLong Diamond Antifade Mountant with 4′,6-diamidino-2-phenylindole (DAPI; Life Technologies). The slides were cured for >24 h at room temperature, protected from light, prior to imaging.
Imaging was performed with a Zeiss LSM 880 inverted laser-scanning confocal microscope with the 63× oil immersion objective and the 405-nm, 488-nm, and 561-nm laser lines. The following settings were used for all images: pinhole = 1 AU, digital offset = 0, and digital gain = 1.0. The amplitude and master gain were optimized for each image to prevent pixel saturation. Quantification of unmodified images was performed with ImageJ (31). Specifically, binary masks of images were used to independently calculate the green and red signals in each cell. The percentage of the total green signal that colocalized with the red signal in each cell was determined by subtracting the masks. Averages were obtained with four to six cells per condition tested. Two-sample t tests of the mean were performed for each data set by using Smith's Statistical Package, version 2.80. P values of <0.01 were considered statistically significant. Representative confocal micrographs of each condition were minimally modified for Hue/Saturation and Brightness/Contrast with Adobe Photoshop (ver. 13.0).
The BaculoDirect expression system (Life Technologies) was used in accordance with the manufacturer's protocol to create His-tagged recombinant VP1 containing the L138P lesion (rVP1L138P). Briefly, outward PCR and site-directed mutagenesis were used to engineer the L138P mutation into the SA11-4F VP1 ORF of the pENTR-SA11 VP1 vector (18). The entire VP1L138P ORF was then transferred into the BaculoDirect C-Term linear DNA by recombination with LR Clonase II. Spodoptera frugiperda (Sf9) cells were maintained at 28°C in complete Grace's medium (Life Technologies) supplemented to contain 10% heat-inactivated fetal bovine serum, 100 U/ml penicillin, 100 μg/ml streptomycin, 0.5 μg/ml amphotericin B, and 1% Pluronic F-68 (Life Technologies). The recombinant baculovirus DNA was transfected into Sf9 cells with the Cellfectin reagent (Life Technologies), and recombinant baculovirus was harvested from medium containing 100 μM ganciclovir. Baculoviruses expressing His-tagged WT SA11-4F VP1 (rVP1WT), catalytically inactive His-tagged SA11-4F VP1 (rVP1D632A), and WT SA11-4F VP2 (rVP2) were described previously (18, 32). Protein expression and purification were performed as described previously by McDonald and Patton (18). Purified proteins were assessed for quality and relative quantity by electrophoresis versus protein standards in sodium dodecyl sulfate (SDS)-polyacrylamide gels and GelCode Blue staining (Thermo Scientific). The Precision Plus Protein Kaleidoscope Prestained Protein Standards (Bio-Rad) were used as molecular weight markers. Proteins were stored at 4°C in low-salt buffer (2 mM Tris-HCl [pH 7.5], 0.5 mM EDTA, 0.5 mM dithiothreitol [DTT]) containing 1× complete protease inhibitors (Roche). rVP1 proteins were used within 3 days of their purification.
The activity of rVP1WT, rVP1L138P, and rVP1D632A was determined with an in vitro dsRNA synthesis assay (18, 33). Each 20-μl reaction mixture contained 2 pmol of rVP1; 20 pmol of rVP2; 4 pmol of SA11 gene 8 +RNA; 50 mM Tris-HCl (pH 7.5); 1 μl of 30% polyethylene glycol 8000; 20 mM Mg acetate; 1.6 mM Mn acetate; 2.5 mM DTT; 1.25 mM (each) ATP, CTP, and UTP; 5 mM GTP; 1 μl of RNasin (Promega); and 1 μCi of [α-32P]UTP (3,000 Ci/mmol; PerkinElmer). Reactions proceeded at 31°C, 37°C, or 39°C for the times indicated. The [32P]-labeled dsRNA products of the reaction were electrophoresed in 12% SDS-polyacrylamide gels. The gels were dried onto filter paper, and dsRNA bands were visualized and quantitated with a GE Healthcare Storm 860 phosphorimager. One- or two-sample t tests of the mean were performed for each replicase assay data set with Smith's Statistical Package, version 2.80. P values of <0.01 were considered statistically significant.
Molecular dynamics simulations were performed with GROMACS v5.1.3 on modified atomic models of strain SA11-4F VP1 (PDB accession no. 2R7Q) (14, 34). Specifically, to create WT VP1 for simulations, we needed to model a flexible loop (residues 346 to 358) that is missing from the SA11-4F structure. This loop was modeled with the Modeler loop-modeling tool embedded in the program UCSF Chimera v1.11 (35, 36). The model with the most favorable zDOPE score was chosen for further computations, and the resulting structure is consistent with that of a previous study that used strain Bristol VP1 (24). To create the L138P mutant VP1 structure for simulations, Chimera's Rotamers tool was used to replace the native leucine residue at position 138 in the WT VP1 structure with a proline residue. The proline rotamer from the Dunbrack backbone-dependent rotamer library with the highest probability was chosen for the L138P mutant VP1 model. The PDB files of both structures are available upon request. Before the simulations were performed, the complete WT and L138P mutant VP1 structures were explicitly solvated with a three-point water model (TIP3P) in a rhombic dodecahedron water box (solute-box distance of 1.0 nm) under periodic boundary conditions, with charges neutralized by chloride ions. The AMBER99SB-ILDN force field was used for all simulations (37). Starting structures were energy minimized until a convergence at a maximum force (Fmax) of <1,000 kJ/mol/nm. A 100-ps position-restrained NVT (amount of substance [N], volume [V], and temperature [T]) equilibration simulation was run for water relaxation at 312K with a modified Berendsen (velocity rescaling) thermostat, followed by a 100-ps NPT equilibration simulation with the Parrinello-Rahman barostat for pressure coupling. After equilibration, an unrestrained 20-ns NPT (amount of substance [N], pressure [P], and temperature [T]) molecular dynamics simulation was run at 312K. Three trajectories initiated with different random seeds were run for each protein structure. The RMSF of all of the α carbons from each of the three trajectories was calculated by using the gmx rmsf command in GROMACS. B factors for each residue were calculated from the RMSF values with the equation B factor = (8Π2/3) × (RMSF)2.
We thank members of the McDonald laboratory for intellectual and technical support on this. We also thank John Patton (University of Maryland, College Park) for the generous donation of viruses and antiserum. Finally, we express our gratitude to John Chappell, Michael Fox, and Jamie Smyth (Virginia Tech Carilion) for assistance with confocal microscopy and image quantitation.
This work was supported through start-up funding from the Virginia Tech Carilion Research Institute and through grants from the National Institutes of Health (R01-AI116815 and R21-AI119588). A.O.M. was also supported by the Biomedical and Veterinary Sciences Graduate Program of the Virginia-Maryland College of Veterinary Medicine and the Virginia Tech Carilion Medical Scholars fund.