|Home | About | Journals | Submit | Contact Us | Français|
The filovirus VP40 protein is capable of budding from mammalian cells in the form of virus-like particles (VLPs) that are morphologically indistinguishable from infectious virions. Ebola virus VP40 (eVP40) contains well-characterized overlapping L domains, which play a key role in mediating efficient virus egress. L domains represent only one component required for efficient budding and, therefore, there is a need to identify and characterize additional domains important for VP40 function. We demonstrate here that the 96LPLGVA101 sequence of eVP40 and the corresponding 84LPLGIM89 sequence of Marburg virus VP40 (mVP40) are critical for efficient release of VP40 VLPs. Indeed, deletion of these motifs essentially abolished the ability of eVP40 and mVP40 to bud as VLPs. To address the mechanism by which the 96LPLGVA101 motif of eVP40 contributes to egress, a series of point mutations were introduced into this motif. These mutants were then compared to the eVP40 wild type in a VLP budding assay to assess budding competency. Confocal microscopy and gel filtration analyses were performed to assess their pattern of intracellular localization and ability to oligomerize, respectively. Our results show that mutations disrupting the 96LPLGVA101 motif resulted in both altered patterns of intracellular localization and self-assembly compared to wild-type controls. Interestingly, coexpression of either Ebola virus GP-WT or mVP40-WT with eVP40-ΔLPLGVA failed to rescue the budding defective eVP40-ΔLPLGVA mutant into VLPs; however, coexpression of eVP40-WT with mVP40-ΔLPLGIM successfully rescued budding of mVP40-ΔLPLGIM into VLPs at mVP40-WT levels. In sum, our findings implicate the LPLGVA and LPLGIM motifs of eVP40 and mVP40, respectively, as being important for VP40 structure/stability and budding.
Ebola and Marburg viruses are members of the family Filoviridae. Filoviruses are filamentous, negative-sense, single-stranded RNA viruses that cause lethal hemorrhagic fevers in both humans and nonhuman primates (5). Filoviruses encode seven viral proteins including: NP (major nucleoprotein), VP35 (phosphoprotein), VP40 (matrix protein), GP (glycoprotein), VP30 (minor nucleoprotein), VP24 (secondary matrix protein), and L (RNA-dependent RNA polymerase) (2, 5, 10, 12, 45). Numerous studies have shown that expression of Ebola virus VP40 (eVP40) alone in mammalian cells leads to the production of virus-like particles (VLPs) with filamentous morphology which is indistinguishable from infectious Ebola virus particles (12, 17, 18, 25, 26, 27, 30, 31, 34, 49). Like many enveloped viruses such as rhabdovirus (11) and arenaviruses (44), Ebola virus encodes late-assembly or L domains, which are sequences required for the membrane fission event that separates viral and cellular membranes to release nascent virion particles (1, 5, 7, 10, 12, 18, 25, 27, 34). Thus far, four classes of L domains have been identified which were defined by their conserved amino acid core sequences: the Pro-Thr/Ser-Ala-Pro (PT/SAP) motif (25, 27), the Pro-Pro-x-Tyr (PPxY) motif (11, 12, 18, 19, 41, 53), the Tyr-x-x-Leu (YxxL) motif (3, 15, 27, 37), and the Phe-Pro-Ile-Val (FPIV) motif (39). Both PTAP and the PPxY motifs are essential for efficient particle release for eVP40 (25, 27, 48, 49), whereas mVP40 contains only a PPxY motif. L domains are believed to act as docking sites for the recruitment of cellular proteins involved in endocytic trafficking and multivesicular body biogenesis to facilitate virus-cell separation (8, 13, 14, 16, 28, 29, 33, 36, 43, 50, 51).
In addition to L domains, oligomerization, and plasma-membrane localization of VP40 are two functions of the protein that are critical for efficient budding of VLPs and virions. Specific sequences involved in self-assembly and membrane localization have yet to be defined precisely. However, recent reports have attempted to identify regions of VP40 that are important for its overall function in assembly and budding. For example, the amino acid region 212KLR214 located at the C-terminal region was found to be important for efficient release of eVP40 VLPs, with Leu213 being the most critical (30). Mutation of the 212KLR214 region resulted in altered patterns of cellular localization and oligomerization of eVP40 compared to those of the wild-type genotype (30). In addition, the proline at position 53 was also implicated as being essential for eVP40 VLP release and plasma-membrane localization (54).
In a more recent study, a YPLGVG motif within the M protein of Nipah virus (NiV) was shown to be important for stability, membrane binding, and budding of NiV VLPs (35). Whether this NiV M motif represents a new class of L domain remains to be determined. However, it is clear that this YPLGVG motif of NiV M is important for budding, perhaps involving a novel mechanism (35). Our rationale for investigating the corresponding, conserved motifs present within the Ebola and Marburg virus VP40 proteins was based primarily on these findings with NiV. In addition, Ebola virus VP40 motif maps close to the hinge region separating the N- and C-terminal domains of VP40 (4). Thus, the 96LPLGVA101 motif of eVP40 is predicted to be important for the overall stability and function of VP40 during egress. Findings presented here indicate that disruption of these filovirus VP40 motifs results in a severe defect in VLP budding, due in part to impairment in overall VP40 structure, stability and/or intracellular localization.
Human 293T cells used for transfection experiments were grown in Dulbecco modified Eagle medium (DMEM; Invitrogen/Life Technologies) supplemented with 10% fetal calf serum (HyClone) and 1× penicillin-streptomycin (Invitrogen/Life Technologies) at 5.0% CO2 at 37°C. pCAGGS vector and plasmid pCAGGS-VP40-WT have been described previously (25). All mutant constructs described below were generated in pGEM T-Easy vector (Promega) by using a QuikChange II site-direct mutagenesis kit (Stratagene) or overlapping PCR and then subcloned into pCAGGS by using the SphI and NheI restriction sites. Plasmids mVP40-WT fused with either Flag or hemagglutinin (HA) epitope tags were generously provided by Stephan Becker (Marburg, Germany). The mVP40-ΔLPLGIM plasmid containing the Flag tag was generated by overlapping PCR using the mVP40-WT plasmid as a template. Plasmid eVP40-ΔLPLGVA-Flag was constructed by overlapping PCR. All plasmids containing introduced mutations or deletions were verified by automated DNA sequencing. Antipeptide antiserum against Ebola virus VP40 was described previously (25, 32). Antiserum against the Flag epitope tag and the HA tag was purchased from Santa Cruz Biotechnology and was used according to the protocol of the supplier.
A functional VLP budding assay for VP40 was performed essentially as previously described (25, 30). Equivalent amounts of VLPs and cell lysates were analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), followed by Western blot analysis with appropriate antibodies.
Human 293T cells were grown on glass coverslips in six-well plates and transfected with the indicated plasmids. Cells were transfected with pCAGGS vector as a negative control. The cells were washed once with phosphate-buffered saline (PBS) at 24 h posttransfection and fixed in cold methanol-acetone (1:1 [vol/vol]) for 10 min at room temperature. Cells were washed three times with PBS and then stained with primary antibodies to eVP40 or mVP40, incubated at room temperature for 1 h, and then washed as described above. Secondary antibodies conjugated to either Alexa 594 goat against rabbit IgG (H+L) or Alexa 488 fluorophore donkey anti-mouse (both from Molecular Probes, Invitrogen, Carlsbad, CA) in PBS were added, followed by incubation in a dark place at room temperature for 1 h. Cells were washed four times, stained with DAPI (4′,6′-diamidino-2-phenylindole) for 10 min at room temperature, washed four times with PBS, and affixed to glass slides with Prolong Antifade (Molecular Probes, Invitrogen, Carlsbad, CA). Confocal imagines were obtained by using a Zeiss LSM-510 Meta confocal microscope.
Human 293T cells were transfected with the indicated plasmids. At 24 h posttransfection, the cells were washed once with PBS, lysed with PBS containing 1% Triton X-100, and incubated at 4°C for 10 min. Cell lysates were centrifuged at 2,000 rpm for 10 min at 4°C. Cell lysates were then filtered through a 0.22-μm-pore-size filter and separated on a Superdex-200 10/30 high-resolution, fast-protein liquid chromatography (FPLC) column (GE Healthcare). Eluted proteins were collected in 0.5-ml fractions and analyzed by SDS-PAGE and then by Western blotting with anti-eVP40 or anti-Flag antibody, as described above. The chromatogram plotting absorbance (280 nm) versus elution volume was generated with Unicorn software. The molecular mass standards depicted on the chromatogram had molecular masses of 670, 158, 44, 17, and 1.35 kDa.
Human 293T cells seeded in six-well plate were transfected with eVP40-WT, eVP40-ΔLPLGVA, mVP40-WT, and mVP40-ΔLPLGIM. At 12 h posttransfection, the cells were metabolically labeled with [35S]Met-Cys (Perkin-Elmer) at 150 μCi/ml for 30 min. The culture medium was then removed, washed three times with PBS, and replaced with fresh DMEM (with 4.5 g of d-glucose/liter but lacking glutamine, sodium pyruvate, l-methionine, and l-cysteine). Afterward, the cells were harvested at 0, 0.5, 1, 2, 4, and 8 h. Equal amounts of cell lysates were loaded on SDS-PAGE gels. The gels were fixed for 30 min in fixing buffer (40% methanol, 10% acetic acid), washed three times with washing buffer (10% acetic acid, 10% glycerol), dried, and then exposed to chemiluminescence film overnight and developed.
A YPLGVG motif was identified recently in the M protein of Nipah virus and shown to be important for efficient VLP budding and intracellular localization of NiV M protein (35). This motif was found to be highly conserved in the M proteins of morbilliviruses, but was not present in the M proteins of all paramyxoviruses or in the M proteins of rhabdoviruses (vesicular stomatitis virus and rabies virus) (Fig. (Fig.1).1). Interestingly, this motif is present in the VP40 matrix proteins of the Ebola and Marburg viruses (Fig. (Fig.1).1). Thus, we sought to determine whether the corresponding LPLGVA and LPLGIM motifs of Ebola and Marburg virus VP40 proteins, respectively, were important for efficient VLP egress. Toward this end, we generated deletion mutants for both eVP40 and mVP40 (Fig. (Fig.2A).2A). The eVP40-WT and mVP40-WT proteins served as positive controls for VLP budding, and the previously characterized eVP40-ΔPT/PY L-domain mutant served as a budding defective control (Fig. (Fig.2A).2A). Human 293T cells were transfected with the indicated plasmids, and the presence of the respective VP40 proteins in both cells and VLPs was assessed by Western blotting (Fig. (Fig.2B).2B). As expected, eVP40-WT and mVP40-WT were readily detected in both cell extracts and VLPs (Fig. (Fig.2B,2B, lanes 1 and 4). In addition, the eVP40-ΔPT/PY mutant lacking both l-domain motifs was detected in cell extracts but was severely defective in budding as a VLP (25) (Fig. (Fig.2B,2B, lane 3). Notably, both eVP40-ΔLPLGVA and mVP40-ΔLPLGIM were detected in cell extracts, but VLP budding mediated by these mutants was completely abrogated in repeated experiments (Fig. (Fig.2B,2B, lanes 2 and 5). These findings suggest that the Ebola virus LPLGVA and Marburg virus LPLGIM motifs are crucial for VLP budding as determined by a standard VLP budding assay.
Since disruption of the YPLGVG motif of NiV M protein resulted in altered intracellular localization patterns, we sought to determine whether localization of the eVP40-ΔLPLGVA and mVP40-ΔLPLGIM mutants was also distorted. Previous studies revealed that eVP40-WT localizes predominantly at the plasma membrane, as well as in the cytoplasm when expressed alone in mammalian cells (25, 30, 34). We utilized confocal microscopy to visualize the eVP40-ΔLPLGVA and mVP40-ΔLPLGIM mutants. Briefly, human 293T cells were transfected with eVP40-WT, eVP40-ΔLPLGVA, eVP40-ΔPT/PY, mVP40-WT, mVP40-ΔLPLGIM, or pCAGGS vector alone, and VP40 proteins were detected by immunofluorescence and visualized by confocal microscopy (Fig. (Fig.3A).3A). Consistent with previous findings (25, 30, 34), the confocal imaging of cells expressing eVP40-WT revealed a predominant pattern of plasma-membrane staining with some disperse cytoplasmic staining as well (Fig. (Fig.3A,3A, top row). Similarly, the intracellular staining pattern of eVP40-ΔPT/PY protein mimicked that of eVP40-WT (Fig. (Fig.3A,3A, second row). Again, this finding is consistent with previous work (12, 25) and, importantly, represents an example of an eVP40 deletion mutant whose pattern of intracellular localization remains consistent with that of eVP40-WT. In contrast, the localization pattern observed for eVP40-ΔLPLGVA differed dramatically from those of both eVP40-WT and eVP40-ΔPT/PY (Fig. (Fig.3A,3A, third row). Indeed, the eVP40-ΔLPLGVA mutant displayed virtually no plasma-membrane staining in repeated experiments but rather appeared as punctuate aggregates in close proximity to the nucleus (Fig. (Fig.3A,3A, third row). Similarly, the staining pattern of mVP40-ΔLPLGIM was also found to be more punctuate than that observed for mVP40-WT (Fig. (Fig.3B).3B). Taken together, these findings indicate that the eVP40-ΔLPLGVA and mVP40-ΔLPLGIM mutants exhibited impaired plasma-membrane localization with increased punctuate staining in the cytoplasm. These findings are consistent with the severely impaired budding phenotype observed for these mutants (Fig. (Fig.22).
To define further the residues within the LPLGVA motif of eVP40 that are important for VLP budding, we introduced alanine substitutions at the highly conserved amino acid positions, 96, 97, and 98 (Fig. (Fig.4A)4A) and ascertained the ability of each point mutant to bud using a VLP budding assay (Fig. (Fig.4B).4B). Our results indicated that the three point mutants were expressed to levels equal to those of eVP40-WT and eVP40-ΔLPLGVA in cell extracts (Fig. (Fig.4B,4B, cells); however, the levels of the point mutants observed in the corresponding VLPs varied (Fig. (Fig.4B,4B, VLPs). As expected, the eVP40-ΔLPLGVA was virtually undetectable in VLPs (Fig. (Fig.4B,4B, lane 2), whereas mutants L96A (lane 3) and P97A (lane 4) were each reduced by >50% compared to eVP40-WT (lane 1). The L98A mutant was capable of budding to a level equivalent to that of eVP40-WT (Fig. (Fig.4B,4B, VLPs, compare lanes 1 and 5). It should be noted that a fourth mutant of eVP40, in which the LPLGVA sequence was converted to the corresponding Marburg virus sequence (LPLGIM), was able to bud efficiently as a VLP to a level identical to that of eVP40-WT (Y. Liu and R. N. Harty, data not shown). These results suggest that conservation of leucine-96 and proline-97 of eVP40 is necessary for efficient egress of VLPs and that changing leucine-98 to alanine does not disrupt VLP budding. Lastly, the Marburg virus LPLGIM motif was able to functionally replace the corresponding LPLGVA sequence of Ebola virus VP40, indicating that valine-100 and alanine-101 can be replaced with isoleucine and methionine, respectively, without any loss of budding function (Liu and Harty, data not shown).
Next, we examined the intracellular localization patterns of the eVP40 mutants L96A, P97A, and L98A by confocal microscopy (Fig. (Fig.5).5). Our results showed that mutants L96A and P97A displayed an intermediate phenotype between that of eVP40-WT and eVP40-ΔLPLGVA. For example, some cells exhibited staining at the plasma-membrane (similar to eVP40-WT), whereas others exhibited a more punctuate aggregate pattern of staining (similar to eVP40-ΔLPLGVA) (Fig. (Fig.5).5). In contrast, the intracellular localization pattern for mutant L98A was predominantly at the plasma-membrane and virtually identical to that of eVP40-WT (Fig. (Fig.5).5). These findings correlate well with the results obtained in the VLP budding assay for each of the three eVP40 point mutants and suggest that L96 and P97 are required for optimal localization of eVP40.
In addition to L domains and plasma-membrane binding, efficient homo-oligomerization of VP40 is necessary to promote VLP egress. Thus, it was of interest to examine the oligomerization profiles of eVP40-ΔLPLGVA and mVP40-ΔLPLGIM mutants. Briefly, human 293T cells were transfected with eVP40-WT, eVP40-ΔLPLGVA, eVP40-ΔPT/PY, mVP40-ΔLPLGIM, and pCAGGS vector alone for 24 h. Cell lysates were clarified and filtered through a 0.22-μm-pore-size filter, and proteins were then separated on an FPLC column, along with internal molecular weight standards (Fig. (Fig.6A).6A). Eluted proteins were harvested in 0.5-ml samples and analyzed by Western blotting (Fig. (Fig.6B).6B). An expression control for each VP40 protein was harvested prior to gel filtration and detected by Western blotting (Fig. (Fig.6B,6B, lane CL).
As a positive control, eVP40-WT eluted from the FPLC column in two main peaks at elution volumes 8.5 to 9.5 ml and 14.5 to 15.5 ml, likely representing oligomeric and monomeric forms, respectively (Fig. (Fig.6B).6B). These findings are similar to those described previously for eVP40-WT (30), and identical oligomeric and monomeric peaks were observed for mVP40-WT as well (Liu and Harty, data not shown). The eVP40-ΔPT/PY l-domain mutant was also detected in two peaks, albeit narrower peaks than those observed for eVP40-WT. This finding is also consistent with previous data, which demonstrated that eVP40 l-domain deletion mutants are not defective in either membrane localization or homo-oligomerization (12). In contrast, both the eVP40-ΔLPLGVA and the mVP40-ΔLPLGIM mutants were detected in a single prominent peak at elution volumes of 8.5 to 10.0 ml, likely representing oligomeric forms of the proteins (Fig. (Fig.6B).6B). Indeed, both eVP40-ΔLPLGVA and mVP40-ΔLPLGIM were virtually undetectable in fractions predicted to harbor monomers of these mutants (Fig. (Fig.6B).6B). Thus, the LPLGVA and LPLGIM motifs may be important for the overall structure of eVP40 and mVP40 proteins, respectively.
To further examine whether the LPLGVA and LPLGIM motifs contribute to overall stability of eVP40 and mVP40, respectively, pulse-chase analysis was performed. Briefly, human 293T cells were transfected with plasmids encoding WT or deletion mutants of eVP40 or mVP40. Proteins were radiolabeled with [35S]methionine-cysteine for 30 min and then chased with cold medium for various times up to 8 h (Fig. (Fig.7).7). Total cell lysates were immunoprecipitated with anti-eVP40 antiserum or anti-Flag antiserum to detect eVP40 or mVP40, respectively (Fig. (Fig.77).
The four proteins were readily detected at the zero time point, and both eVP40-WT and mVP40-WT were found to be stable over the entire 8-h chase period (Fig. (Fig.7).7). In contrast, eVP40-ΔLPLGVA exhibited significant degradation between 2 and 4 h of chase, and mVP40-ΔLPLGIM exhibited significant degradation between 1 and 2 h of chase (Fig. (Fig.7).7). The levels of both eVP40-ΔLPLGVA and mVP40-ΔLPLGIM proteins were decreased by ~10-fold at the 8-h time point compared to those at the zero time point (Fig. (Fig.7).7). These data provide further evidence implicating the Ebola virus LPLGVA and Marburg virus LPLGIM motifs in providing optimal stability of VP40.
Coexpression of Ebola virus GP has been shown to enhance budding of both eVP40-WT VLPs (6, 20, 23), as well as eVP40 l-domain mutant VLPs (26). Thus, it was of interest to determine whether expression of Ebola virus GP could rescue budding of eVP40-ΔLPLGVA into VLPs. In addition, we sought to determine whether eVP40-WT could rescue budding of mVP40-ΔLPLGIM, and conversely, whether mVP40-WT could rescue budding of eVP40-ΔLPLGVA.
Briefly, human 293T cells were transfected with a constant amount of eVP40-ΔLPLGVA and increasing amount of Ebola virus GP as indicated (Fig. (Fig.8A).8A). Expression of the eVP40-ΔLPLGVA mutant and GP in appropriate samples was confirmed by Western blotting (Fig. (Fig.8A).8A). Interestingly, expression of increasing amounts of Ebola virus GP failed to rescue budding of eVP40-ΔLPLGVA into VLPs (Fig. (Fig.8A,8A, top panel). A similar approach was used to determine whether expression of either eVP40-WT or mVP40-WT could rescue budding of the corresponding heterologous VP40 deletion mutant (Fig. 8B and C). mVP40-WT (Fig. (Fig.8B)8B) and mVP40-ΔLPLGIM (Fig. (Fig.8C)8C) proteins were both Flag-tagged and were detected in VLPs and cell extracts using anti-Flag antiserum. As observed for expression of Ebola virus GP, expression of increasing amounts of mVP40-WT failed to rescue budding of eVP40-ΔLPLGVA into VLPs (Fig. (Fig.8B).8B). Surprisingly, expression of increasing amounts of eVP40-WT successfully rescued budding of mVP40-ΔLPLGIM into VLPs in a dose-dependent manner (Fig. (Fig.8C)8C) to levels comparable to that of mVP40-WT (Fig. (Fig.8C,8C, compare lanes 6 and 7). The fact that Ebola virus GP was unable to rescue budding of eVP40-ΔLPLGVA suggests that the LPLGVA motif of eVP40 is not functioning as a typical L domain (Fig. (Fig.8A).8A). In addition, the ability of eVP40-WT to rescue budding of mVP40-ΔLPLGIM suggests that the budding competent phenotype of eVP40-WT is dominant over the budding defective phenotype of mVP40-ΔLPLGIM (Fig. (Fig.8C).8C). Whether the heterologous VP40 proteins interact (oligomerize) to bud remains to be determined.
Homologous rescue experiments were performed for both eVP40 (Fig. (Fig.8D)8D) and mVP40 (Fig. (Fig.8E)8E) proteins to serve as controls for the above heterologous rescue experiments. As expected, budding of the eVP40-ΔLPLGVA mutant into VLPs was rescued by coexpression of eVP40-WT (Fig. (Fig.8D).8D). eVP40-WT was detected by using anti-VP40 monoclonal antibody, and eVP40-ΔLPLGVA was Flag tagged at its N terminus and detected by using anti-Flag antiserum (Fig. (Fig.8D).8D). No cross-reactivity between anti-Flag and anti-eVP40 antisera was observed (Fig. (Fig.8D).8D). Similarly, budding of the mVP40-ΔLPLGIM mutant into VLPs was rescued by coexpressing mVP40-WT (Fig. (Fig.8E).8E). Antisera specific for Flag and HA epitope tags were used to detect mVP40-ΔLPLGIM-Flag and mVP40-WT-HA proteins, respectively (Fig. (Fig.8E).8E). Interestingly, the stability of the mVP40-ΔLPLGIM-Flag mutant may be reduced in the presence of high levels of mVP40-WT (Fig. (Fig.8E,8E, last lane), whereas a similar phenomenon was not evident for Ebola virus proteins (Fig. (Fig.8D).8D). Taken together, these results indicated that expression of the homologous WT protein was sufficient to rescue the corresponding deletion mutant. Lastly, it should be noted that rescue of mVP40-ΔLPLGIM by eVP40-WT (Fig. (Fig.8C)8C) was dependent on the presence of a functional L domain within the eVP40-WT protein (Liu and Harty, data not shown).
The filovirus VP40 matrix protein is a major structural protein that plays a central role in virus assembly and budding. With the exception of L domains, additional functional elements of filovirus VP40 proteins remain to be identified. Based on previous structural studies on Ebola virus VP40 (4) and recent findings on important budding domains within the Nipah virus M protein (35), we sought to determine whether the conserved 96LPLGVA101 and 84LPLGIM89 motifs of eVP40 and mVP40, respectively, were important for VP40 function in budding. The corresponding YPLGVG sequence in Nipah virus M protein was found to play an important role in NiV VLP budding and M protein localization (35). To date, the YPLGVG motif of NiV M protein does not appear to function as a typical L domain but rather may play a role in assembly and/or budding via a novel mechanism (35).
In the present study, we generated 96LPLGVA101 and 84LPLGIM89 deletion mutants, as well as point mutants of the Ebola virus sequence, and compared the budding phenotypes of these mutants to those of WT VP40 proteins. The results from our VLP budding assays clearly indicated that both VP40 deletion mutants were severely defective in budding, even greater than previously characterized l-domain mutants (25). To begin to address the mechanism of this budding defect, we examined the intracellular localization pattern of the deletion and point mutants and compared them to those of eVP40-WT and mVP40-WT. Both the eVP40-ΔLPLGVA and the mVP40-ΔLPLGIM deletion mutants, as well as the L96A and P97A point mutants of eVP40, displayed intercellular localization patterns that differed from those observed for WT VP40 proteins. Since we did not generate point mutants within this motif of mVP40, we can only speculate that the phenotypes of point mutants of mVP40 would likely mimic those observed here for point mutants of eVP40. Thus, the LPLGVA and LPLGIM motifs likely play some direct or indirect role in localization of VP40 at the plasma-membrane required for efficient egress of VLPs. In contrast to these results, intracellular localization of the L98A point mutant of eVP40 remained virtually identical to that of eVP40-WT. Importantly, localization of the eVP40-ΔPT/PY l-domain deletion mutant, as expected, also mimicked that of eVP40-WT, indicating that simply deleting any short amino acid stretch of eVP40 does not result in altered localization patterns. These results are in agreement with those described for the YPLGVG mutants of NiV M protein and are similar to those previously described for the 212KLR214 mutants of eVP40 (30). However, the degree of non-plasma-membrane, punctate staining exhibited by the eVP40-ΔLPLGVA mutant appears to be more pronounced than that observed for the 212KLR214 mutants of eVP40 (30).
In addition to changes in intracellular localization patterns, the eVP40-ΔLPLGVA and mVP40-ΔLPLGIM mutants also exhibited altered elution patterns, as determined by gel-filtration analysis, and were less stable than their corresponding WT proteins as determined by pulse-chase analysis. Ebola virus VP40 is known to adopt at least four conformational states: a monomeric form (4), a dimeric form (9, 47), as well as two higher-order oligomeric ring-like structures that include hexamers and octamers (38, 40, 46). Proper oligomerization of eVP40 is an important step in the formation and subsequent budding of eVP40 VLPs (30, 34, 47, 48). The results from gel filtration and Western blot analyses revealed that monomeric forms of eVP40-ΔLPLGVA and mVP40-ΔLPLGIM were virtually undetectable (Fig. (Fig.6).6). A peak corresponding to higher-order oligomeric structures of eVP40-ΔLPLGVA and mVP40-ΔLPLGIM predominated. These findings with the eVP40-ΔLPLGVA and mVP40-ΔLPLGIM deletion mutants differed from that observed for the eVP40-ΔPT/PY l-domain deletion mutant, whose gel filtration profile closely resembled that of eVP40-WT (Fig. (Fig.6).6). The surprising inability to detect monomeric forms of the eVP40-ΔLPLGVA and mVP40-ΔLPLGIM mutants suggested that these motifs may play a role in overall protein structure and stability. Based on the known structure of eVP40, the LPLGVA motif does indeed map to a region predicted to be important for eVP40 structure and folding. The results from pulse-chase analysis demonstrated that both eVP40-ΔLPLGVA and mVP40-ΔLPLGIM deletion mutants were considerably less stable than their WT counterparts over an 8 h chase period (Fig. (Fig.77).
In an attempt to understand further how the filovirus LPLGVA and LPLGIM motifs contribute to budding, we attempted to rescue budding of the eVP40-ΔLPLGVA and mVP40-ΔLPLGIM deletion mutants by coexpressing either Ebola virus GP or WT VP40 proteins. If the Ebola virus LPLGVA motif functions as a typical L domain, then we would predict that coexpression of GP would enhance budding of eVP40-ΔLPLGVA VLPs as it did for Ebola virus VP40 PTAPPEY l-domain mutants (26). Expression of Ebola virus GP failed to rescue budding of eVP40-ΔLPLGVA, which was not surprising since this mutant exhibited phenotypic characteristics that differed significantly from those of eVP40-ΔPT/PY l-domain mutant.
A more intriguing finding was that eVP40-WT was able to rescue budding of mVP40-ΔLPLGIM to WT levels; however, we were unable to detect rescue of eVP40-ΔLPLGVA, even when mVP40-WT was coexpressed at high levels. One possible explanation as to why eVP40-WT could rescue the Marburg virus mutant, whereas the mVP40-WT could not rescue the Ebola virus mutant, is that the overlapping nature of the L domains of Ebola virus VP40 imparts stronger budding activity to eVP40 than the single PPPY l-domain motif does to mVP40. Indeed, others have noted the enhanced l-domain activity associated with the Ebola virus PTAPPEY sequences when these sequences are inserted into heterologous matrix proteins (42, 52). Alternatively, deletion of the LPLGVA motif from eVP40 may have more severe consequences on the structure of eVP40 than deletion of the LPLGIM motif from mVP40 has on the structure of mVP40. Finally, membrane trafficking and preferred sites of VLP budding for eVP40 and mVP40 may not be identical. Although much of the existing data suggest that eVP40 VLPs bud primarily from the plasma-membrane, there is evidence to suggest that budding of mVP40 VLPs may occur efficiently at internal endosomal membranes (21-24). Experiments are now under way to more precisely elucidate the mechanism of heterologous rescue and identify the domains required for possible interactions between eVP40 and mVP40 proteins (Liu and Harty, data not shown).
In sum, these filovirus motifs appear to be both structurally and functionally crucial for overall activity of VP40 in VLP assembly and/or budding. The fact that these motifs are fairly well conserved within the matrix proteins of a plethora of negative-sense RNA viruses strongly suggests that they maintain an important and conserved role in the assembly and budding pathways of these various human pathogens. Further investigations into the mechanism of action of these motifs may provide useful information toward the development of inhibitors of filovirus assembly and budding.
We thank members of the Harty and Sunyer laboratories for fruitful discussions and Jasmine Zhao of the Biomedical Imaging Core at the University of Pennsylvania for assistance with confocal microscopy. We also thank Chris Broder for critical reading of the manuscript and Stephan Becker for providing mVP40.
This study was supported in part by NIH grants AI077014 and AI072008 to R.N.H.
Published ahead of print on 23 December 2009.