ST-246 prevents the formation of wrapped viral particles
ST-246 targets VV p37 and has been shown to inhibit plaque formation in vitro and prevent systemic viral spread in vivo indicating that the compound inhibits the release of extracellular virus [24]. To determine if the block in EEV production was due to a lack of IEV production, VV was propagated in the presence and absence of ST-246, and virions were radiolabeled with tritiated thymidine and fractionated by equilibrium centrifugation (Fig. 1A). In the absence of compound, radiolabeled material from infected-cell lysates was partitioned into three distinct peaks of radioactivity, while radiolabeled material from the culture medium migrated as a single peak. Based upon the solution density of the peak fractions and immunoblot blot analysis, peaks from lysate samples were designated IMV, CEV and IEV, respectively, while the peak from the culture medium samples was designated EEV. In the presence of ST-246, only a single peak of radioactivity was observed from the cell associated virus fractions that migrated at the same density as IMV particles and no peak of radioactivity could be detected in virus from the culture medium. These results show that ST-246 inhibits IEV and CEV formation but not IMV formation during VV replication.
Immunoblot analysis of fractions from the equilibrium centrifugation experiment was conducted to confirm the identity of each type of viral particle in peak fractions. Immunoblots were probed with anti-L4 antiserum to detect viral cores and anti-B5 antiserum which is a component specific to wrapped viral particles (Fig. 1B). In the absence of ST-246, the L4 protein was detected in all fractions with a density between 1.168 g/ml and 1.300 g/ml, which is consistent with this protein being a component of all forms of VV particles. B5 protein was detected in fractions with a density between 1.168 g/ml and 1.250 g/ml, suggesting that these fractions contained IEV and CEV particles. In the presence of ST-246, L4 protein was detected in fractions with a density between 1.250 g/ml and 1.300 g/ml, while B5 protein was not detected in any of the fractions. Taken together, these results demonstrate that ST-246 inhibits wrapped virus formation and prevents extracellular virus production.
p37 co-precipitates with TIP47 in infected cells
Plaque formation requires p37 activity that facilitates wrapping of IMV in virus-modified membranes derived from a post TGN-compartment to produce egress competent viral particles. A conserved YW motif (positions 253 and 254) was identified in the p37 protein that was identical to a motif found in HIV Env, the same motif shown to mediate the association of Env with TIP47 [22]. In order to establish whether or not p37 physically interacts with TIP47, virion-free membrane fractions enriched for LE-proteins were isolated from BSC-40 cells infected with vvF13LGFP as described in the Material and Methods and resuspended in a buffered solution containing a non-ionic detergent. The fractions were then subjected to immunoprecipitation using polyclonal antibodies to either TIP47 (Fig. 2A, lanes 1 and 2), GFP (Fig. 2A, lanes 5 and 6), or TGN46 as a negative control (Fig. 2A, lanes 3, 4, 7, and 8). The immunoprecipitated proteins (Fig. 2A, lanes 1, 3, 5 and 7) or supernatants (unbound fraction) from the immunoprecipitation reaction mixtures (Fig. 2A, lanes 2, 4, 6 and 8) were resolved by SDS-PAGE, transferred to a nitrocellulose membrane and probed with either a monoclonal antibody to the GFP portion of the p37-GFP fusion protein (left panel) or a polyclonal antibody to TIP47 (right panel). p37-GFP co-precipitated with TIP47 (lane 1) and was also detected in the unbound fraction (lane 2). Likewise, TIP47 co-precipitated with p37-GFP (lane 5), however, TIP47 was not detected in the unbound fraction (lane 6) most likely because TIP47 is primarily a soluble cytosolic protein and would not normally copurify with membrane fractions unless tethered to a specific membrane associated cargo protein. A non-specific band of approximately 55 kDa was also detected in the blot probed with antibody to GFP and likely represents the heavy chain portion of the TGN46 antibody employed in the immunprecipitation reaction (Fig. 2, lane 3). p37 was found in the unbound fraction (Fig. 2, lane 4) when polyclonal antibody specific for TGN46 was used in the immunoprecipitation reactions while TIP47 could not be detected in the bound (Fig. 2, lane 7) or unbound fractions (Fig. 2 lane 8). Taken together, these results suggest that a physical interaction exists between p37 and TIP47. While interaction of membrane proteins in the presence of non-ionic detergent suggests direct physical interaction, additional experiments are required to prove specificity of this interaction.
To demonstrate antibody specificity, total cell lysates were prepared from BSC40 cells that were either mock-infected or infected with vvWR, vvGFP, vvΔF13LGFP, or vvF13LGFP. A portion of the lysate was immunoprecipitated with a polyclonal antibody to GFP. Immunoblots of the lysate material pre- and post-immunoprecipitation were probed with a monoclonal antibody to GFP (Fig. 2B). A band of approximately 66 kDa was observed on the immunoblots in samples (Fig. 2B, lane 3 and 8) infected with vvF13L-GFP migrating at the predicted size of the p37-GFP fusion protein. A band of approximately 27 kDa was observed in samples (Fig. 2B lane 4,5,9, and 10) infected with vvGFP and vvΔF13LGFP corresponding to the size of GFP. No bands were observed in mock-infected samples (Fig. 2B lanes 1 and 6) or samples infected with vvWR (Fig. 2B, lanes 2 and 7). These results suggest that the both monoclonal and polyclonal antibodies specific for GFP do not cross react with other viral and cellular proteins.
A conserved YW motif within p37 is required for complimentation of an F13L-deletion virus
To determine whether the YW motif in p37 (Fig. 3A) is required for interaction with TIP47, PCR-generated DNA fragments of F13L containing alanine substitutions within the YW motif were used to complement plaque formation of an F13L-deleted virus. Co-immunoprecipitation was then conducted to assess the ability of p37 expressed from the mutated F13L alleles to associate with TIP47 in virion-free membrane fractions obtained from infected cells. Changing either Y or W to A at positions 253 and 254, respectively, or both residues simultaneously (data not shown) prevented co-immunoprecipitation of p37 with TIP47 (Fig. 3C). In contrast, changing Y at position 253 to phenylalanine to maintain the aromatic character of the motif, or changing amino acids surrounding the YW motif to A (residues 251, 252, and 255 – Fig. 3A) reduced the association of p37 with TIP47 to varying degrees (Fig. 3C). Introduction of these mutations did not alter the subcellular localization of F13L in infected cells (see Additional File 1).
VV recombinants containing deletions in F13L produced normal levels of IMV particles but failed to form plaques within a 3-day incubation period at 37°C. After a 7-day incubation period plaques generated by F13L-deleted recombinants were comparable in size to plaques generated by wild-type virus after a 1-day incubation period. This strong plaquing phenotype was used to measure the ability of mutated F13L alleles to complement plaque formation of an F13L-deleted virus. Consistent with the co-immunoprecipitation data presented in Fig. 3C, p37 variants containing alanine substitutions at either Y253 or W254 respectively, were unable to complement plaque formation of an F13L-deleted virus (Fig. 3B). Alanine substitutions at positions surrounding the YW motif did not block complementation of the F13L-deleted virus (Fig. 3B) suggesting that the YW motif and not amino acids surrounding this motif are required for extra cellular virus formation. Complementing plaque formation by transfecting PCR products encoding p37 could occur through recombination of the F13L alleles into the viral genome or by transient expression of active p37. This distinction, however, does not affect the interpretation of the experiment. Taken together, these observations suggest that the aromatic nature of the YW motif is required for plaque formation and that the YW motif contributes to the interaction of p37 with TIP47 in membrane fractions of infected cells.
Confocal laser scanning microscopy was conducted to rule out the possibility that the p37 expressed from the Y253A and W254A mutants was mislocalized and no longer capable of complementing plaque formation of the F13L-deleted virus. BSC-40 cells were infected with VV, strain WR, at an MOI of 1 and then transfected with 200 ng of PCR-generated DNA encoding wild-type or mutant p37-GFP alleles. Approximately 50 to 100 infected and transfected cells per mutant construct were evaluated for altered intracellular localization of p37-GFP. The fluorescence profile obtained from the p37-GFP fusion protein was evaluated against a TGN marker and found to be indistinguishable from wild type p37 suggesting that alanine substitution in and around the YW motif does not alter the sub-cellular localization of p37 (Fig. S1).
ST-246 inhibits association of p37-GFP with LE proteins in membrane fractions from infected cells
Based on the observations that the interaction of p37 with TIP47 may be involved in plaque formation and that ST-246, which targets p37 activity, inhibited wrapped virus formation, the potential association of p37 with host proteins known to be involved in Rab9-dependent vesicle formation was evaluated. BSC-40 cells were mock-infected or infected in the presence or absence of ST-246 with vvF13LGFP, a recombinant virus that expresses a p37-GFP fusion protein. The isolated membrane fraction was immunoprecipitated in hypotonic buffer with anti-GFP to pull out the p37-GFP-containing complexes. Immunoblot analysis of the precipitated proteins using antibodies specific for Rab9 or the TGN resident protein, p230, was used to measure the association of these proteins with p37. A similar experiment was performed to measure the association of CI-MPR with p37-GFP and Rab9. In the absence of ST-246, immunoprecipitation of p37-GFP co-precipitated p230 and Rab9 protein (Fig. 4A, Lane 1). Likewise, immunoprecipitation of CI-MPR co-precipitated p37-GFP and Rab9 (Fig. 4B, Lane 1). In the presence of ST-246, the amount of p230 co-precipitating with p37-GFP remained constant while the amount of Rab9 associated with p37-GFP was undetectable (Fig. 4A, Lane 2). Similarly, the amount of both p37-GFP and Rab9 co-precipitating with CI-MPR was reduced in cells treated with ST-246 (Fig. 4B, Lane 2). ST-246 did not affect the co-precipitation of Rab9 with CI-MPR in samples from mock-infected cells (Fig. 4C, Lanes 1 and 2). Taken together, these results imply that ST-246 inhibits the interaction of p37-GFP with cellular proteins associated with Rab9-dependent vesicle formation but not proteins associated with the TGN in membrane fractions from infected cells. Moreover, ST-246 does not alter the association of CI-MPR with Rab9 in uninfected cells suggesting that the inhibitory effects of ST-246 are virus-specific and not the result of an interference with Rab9-dependent transport processes in uninfected cells.
ST-246 inhibits the co-localization of p37-GFP with LE proteins in the context of viral infection
To support the results of the co-immunoprecipitation experiments, BSC-40 cells were infected with vvF13LGFP in the presence and absence of ST-246 and the subcellular localization of p37-GFP protein was determined by immunofluorescence analysis and confocal microscopy. The results of these studies are presented in Fig. 5 and consist of images that are representative of a minimum of 50 infected cells. In the absence of ST-246, a discrete distribution of p37-GFP was observed, with no signal in the nucleus (Fig. 5A). This signal co-localized with cellular markers specific for GM130 (cis-Golgi) [25], p230 (TGN)[26], and CI-MPR (LE) [27] as shown by the pattern of yellow in the merged images (Fig. 5 and data not shown). These observations are in agreement with previous reports by other groups [3, 5]. In the presence of ST-246, however, p37-GFP co-localized with p230, with very little, if any, co-localization with CI-MPR. Consistent with the co-immunoprecipitation studies (Fig. 4), these observations suggest that ST-246 inhibits the intracellular trafficking of the p37 protein to LE compartments in infected cells. Thin-slice analysis (Z-stacks) by confocal microscopy further supported the above observation (data not shown).
Since the effect of ST-246 on co-localization of p37-GFP with LE proteins was observed in the context of virus infection, it was important to determine whether this effect required other viral proteins or was specific for p37. Vero cells were transfected with the plasmid pSI-F13L-GFP expressing p37-GFP in the presence or absence of ST-246 and analyzed by immunofluorescence and confocal microscopy (Fig. 5B). The pattern of p37-GFP signal was not affected by the presence of ST-246 and p37-GFP appeared to co-localize with p230 and CI-MPR in the presence and absence of ST-246 as measured by the intensity of the yellow color of the merged images. These results suggest that ST-246 inhibits a virus-specific activity required for co-localization of p37 with LE proteins and is consistent with the observation that functional p37 is required for co-localization of the EEV-specific envelope protein, B5, to post-Golgi vesicles [11].
ST-246 reduces p37-GFP association with vaccinia virus B5 protein and LE proteins in membrane fractions from infected cells
The co-localization of p37 with B5 protein in infected cells is thought to be an indicator of envelope precursor formation [11]. To examine whether p37-GFP co-fractionated with B5 and LE marker proteins, isolated membrane fractions were subjected to co-immunoprecipitation in hypotonic buffer with anti-GFP polyclonal antibody. Immunoblot analysis of the isolated membrane fraction prior to immunoprecipitation (input) showed the presence of the host factors CI-MPR, Rab9 and TIP47 as well as the viral IEV component, B5 (Fig. 6, left panel), in the presence and absence of ST-246. Immunoprecipitation of the membrane fraction using antibodies to p37-GFP in the absence of ST-246 co-precipitated CI-MPR, Rab9, TIP47 and B5 (Fig. 6, right panel). In the presence of ST-246, the amount of CI-MPR, Rab9 and TIP47 co-precipitating with p37-GFP was greatly reduced and B5 was undetectable. Although the membrane fractions were cleared of intact virion particles by velocity sedimentation as described the Materials and Methods, it is likely that other viral components, particularly structural proteins, which tend to be highly expressed, are present in the membrane fraction but not associated with the p37-containing membranes. To investigate this possibility, the isolated membrane fraction was probed with antibodies specific for the viral core proteins L4 and A27 both pre- and post- immunoprecipitation. The A27 and L4 proteins were detected in the input material (Fig. 6, left panel) in the presence and absence of ST-246 but not in the complex immunoprecipitated with antibody to p37-GFP (Fig. 6, right panel). This result suggests that these fractions contain envelope precursors and viral core proteins, but not intact IEV particles. These results suggest association of LE proteins with p37- and B5-containing membrane fractions is sensitive to inhibition by ST-246.
p37 expressed from an ST-246-resistant VV variant interacts with Rab9 and B5 in the presence and absence of compound
To confirm that inhibition of the association of LE proteins with p37-containing membrane fractions by ST-246 is mediated through p37, BSC-40 cells were infected with wild type vvF13LGFP or an ST-246 resistant (ST-246R) variant, which contains an Asp to Tyr change within the p37 protein. This amino acid change increases the concentration of compound required to inhibit 50% of the virus-induced cytopathic effects (EC50) by more than 1000-fold. The EC50 of ST-246 for wild-type VV is 50 nM while the EC50 for the ST-246R variant is > 50 μM. Immunoblot analysis of proteins immunoprecipitated with anti-GFP polyclonal antibody from membrane fraction from wild type virus-infected cells showed an association of p37 with Rab9 and B5 in the absence of ST-246. In the presence of ST-246, these interactions were nearly undetectable (Fig. 7A, left). In contrast, immunoblot analysis of immunoprecipitated proteins from membrane fractions of ST-246R variant virus-infected cells showed an association of Rab9 and B5 with p37 in the presence and absence of ST-246 (Fig. 7A, right). Densitometric analysis of the immunoblots presented in Fig. 7A shows that the amount of Rab9 and B5 that co-precipitate with p37-GFP in the presence of ST-246 was greatly reduced in wild-type VV-infected cells, compared to ST-246R virus-infected cells (Fig. 7B).
To support the co-immunoprecipitation studies described above, the intracellular localization of p37, p230, and CI-MPR within cells infected by the ST-246R variant in the presence and absence of ST-246 was examined by confocal microscopy (Fig. 7C). The images presented are representative of 50–100 infected cells throughout the entire sample and were confirmed by thin-slice (Z-stack) analysis. As observed in previous experiments (Fig. 4 and data not shown), p37 co-localized with p230 in the presence and absence of ST-246 (indicated by the pattern of yellow in the merged images) (Fig. 7C, left). Moreover, unlike p37 expressed from wild type vaccinia virus infected cells, the p37 expressed from the ST-246R variant co-localized with CI-MPR in the presence and absence of ST-246 (Fig. 7C, right). Taken together, these results suggest that ST-246 inhibits the formation of a membrane complex containing p37 and LE proteins.