|Home | About | Journals | Submit | Contact Us | Français|
Arenaviruses can cause lethal hemorrhagic fevers in humans with few preventative and therapeutic measures. The arenaviral glycoprotein stable signal peptide (SSP) is unique among signal peptides in that it is an integral component of the mature glycoprotein complex (GPC) and plays important roles not only in GPC expression and processing but also in the membrane fusion process during viral entry. Using the Pichinde virus (PICV) reverse genetics system, we analyzed the effects of alanine substitutions at many conserved residues within the SSP on viral replication in cell culture and in a guinea pig infection model. Our data showed that the K33A, F49A, and C57A mutations abolished GPC-mediated cell entry and therefore could not allow for the generation of viable recombinant viruses, demonstrating that these residues are essential for the PICV life cycle. The G2A mutation caused a marked reduction of cell entry at the membrane fusion step, and while this mutant virus was viable, it was significantly attenuated in vitro and in vivo. The N20A mutation also reduced membrane fusion activity and viral virulence in guinea pigs, but it did not significantly affect cell entry or viral growth in cell culture. Two other mutations (N37A and R55A) did not affect membrane fusion or viral growth in vitro but significantly reduced viral virulence in vivo. Taken together, our data suggest that the GPC SSP plays an essential role in mediating viral entry and also contributes to viral virulence in vivo.
IMPORTANCE Several arenaviruses, such as Lassa fever virus, can cause severe and lethal hemorrhagic fever diseases with high mortality and morbidity, and no FDA-approved vaccines or therapies are currently available. Viral entry into cells is mediated by arenavirus GPC that consists of an SSP, the receptor-binding GP1, and transmembrane GP2 protein subunits. Using a reverse genetics system of a prototypic arenavirus, Pichinde virus (PICV), we have shown for the first time in the context of virus infections of cell culture and of guinea pigs that the SSP plays an essential role in mediating the membrane fusion step as well as in other yet-to-be-determined processes during viral infection. Our study provides important insights into the biological roles of GPC SSP and implicates it as a good target for the development of antivirals against deadly human arenavirus pathogens.
Arenaviruses can cause neurologic or hemorrhagic fever diseases in humans, and few preventive or therapeutic measures are available (1, 2). The most significant arenavirus pathogen is Lassa fever virus (LASV), which causes infections that are endemic to many countries in West Africa, with an estimated 500,000 cases resulting in ~5,000 deaths annually (3). Except for Junin virus (JUNV), which has a vaccine, Candid#1, used only in Argentina, no licensed vaccine for human use is currently available for any other arenaviruses. Therapeutic options are limited and depend mainly on supportive care. Ribavirin, a nonspecific antiviral compound, has shown some efficacy only if it is administered at an early stage of viral infection when the symptoms are insidious (4). Therefore, development of effective vaccines and antivirals against arenavirus pathogens is urgently needed.
Arenaviruses are enveloped, bisegmented ambisense RNA viruses (5). Each viral genomic segment encodes two viral gene products. The large RNA segment (L) encodes viral RNA-dependent RNA polymerase L protein and the multifunctional protein Z. The small RNA segment (S) encodes nucleoprotein (NP) and glycoprotein (GPC). GPC is expressed as a single polypeptide with a long stable signal peptide (SSP) and is posttranslationally processed by the S1P cellular protease into the receptor-binding GP1 and transmembrane GP2 subunits (6, 7). The 58-amino-acid (aa) conserved SSP contains an N-terminal G2 myristoylation site, two hydrophobic domains (h1 and h2) separated by an ectodomain, and a C-terminal SP cleavage site (Fig. 1A) (8), although the exact boundaries of these domains have not been well defined (9, 10). SSP is an essential component of the envelope glycoprotein complex and plays important roles not only in regulating the intracellular trafficking and proteolytic maturation of GPC but also in the pH-induced membrane fusion reaction (11, 12). However, the biological roles of SSP in arenavirus infection in vitro and in vivo have not been investigated.
Pichinde virus (PICV) is a nonpathogenic biosafety level 2 (BSL2) agent, and its infection of guinea pigs has been used as a safe, convenient, and economic small-animal model for Lassa fever (13, 14). A low-passaged strain of PICV (i.e., P2) causes a brief febrile reaction in guinea pigs, while a high-passaged strain (i.e., P18) causes a severe Lassa fever-like disease in guinea pigs with high mortality (3, 13,–17). We have developed infectious clones for both the nonvirulent P2 strain and the virulent P18 strain of PICV (18) and characterized the virulence determinants at the molecular level (19,–21).
In this study, we investigate the biological roles of the conserved residues of the SSP in vitro and in vivo by using the PICV reverse genetics system. Many of these residues have been characterized previously in the context of GPC expression and pseudotyped viruses (8, 11, 22,–28); however, their biological roles have not been investigated. Here, we show that multiple SSP conserved residues are essential or critical for viral infection of cell culture and/or guinea pigs by participating in the membrane fusion reaction, and that two residues, N37 and R55, are required for viral virulence by a yet-to-be-determined mechanism.
293T, BHK21, and A549 cells were maintained in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum (FBS; Sigma) and 50 mg penicillin and streptomycin. BSRT7-5 cells, which stably express the T7 RNA polymerase, were obtained from K. Conzelmann (Ludwig-Maximilians-Universität, Germany) and cultured in DMEM supplemented with 10% FBS, 1 mg Geneticin per ml, and 50 mg penicillin and streptomycin per ml. Vero cells were maintained in Eagle's medium (MEM) with 10% FBS plus 50 mg/ml penicillin and streptomycin.
Mutations in the SSP gene were created in the pCAGGS-PICV-GPC plasmid by overlapping PCR mutagenesis reactions, and the mutations were confirmed by sequencing. Plasmids used to rescue the PICV SSP mutant viruses were described previously (19). A plasmid that contains the T7 promoter-directed firefly luciferase gene was obtained from J. Nunberg (University of Montana, USA). Plasmid transfection into 293T cells was performed by using the calcium phosphate method. Plasmid transfection into BSRT7-5 cells was performed by using the Lipofectamine 2000 reagent per the manufacturer's protocol (Invitrogen).
293T cells were lysed at 48 h posttransfection by lysis buffer (50 mM Tris-HCl, 150 mM NaCl, 1% NP-40, pH 7.4) with protease inhibitor cocktail (Roche). Lysates were assayed by SDS-PAGE and transferred to nitrocellulose membranes (NC membrane). NC membrane was blocked with 5% nonfat powdered milk in Tris-HCl buffer (TBS) for 1 h at room temperature. The membrane was incubated with anti-PICV guinea pig serum at 4°C overnight, washed three times with TBS (50 mM Tris-HCl, pH 7.4), probed with IRDye 800CW-labeled donkey anti-guinea pig antibody, and assayed by an Odyssey infrared imaging system (LI-COR Biosciences).
For enhanced green fluorescent protein (eGFP) syncytium formation assay, 293T cells were transfected with the eGFP plasmid together with the pCAGGS vector or individual GPC constructs. Thirty-six hours posttransfection, cells were treated with pH-adjusted medium (pH 5) for 5 min and then replaced with fresh complete DMEM for an additional 12 h of culturing. The cell cultures were visualized by fluorescence microscopy.
For luciferase-based fusion assay, 293T cells were transfected with the T7 promoter-driven firefly luciferase reporter plasmid together with either pCAGGS empty vector or individual GPC constructs. Twenty-four hours posttransfection, the transfected 293T cells were detached by trypsinization and mixed with BSR-T7 cells that constitutively express T7 RNA polymerase. At 36 h posttransfection, cells were treated with pH-adjusted medium (pH 5) for 5 min and then replaced with fresh complete DMEM for an additional 12 h of culturing. Cells were lysed 12 h later and analyzed by the luciferase assay kit (Promega).
293T cells were transfected with either pCAGGS empty vector or individual GPC constructs. Forty-eight hours posttransfection, cells were fixed with 4% paraformaldehyde (PFA) for 5 min, washed once with PBS, and blocked with 1% bovine serum albumin (BSA) for 1 h at 4°C. Cells were then incubated with guinea pig anti-PICV serum overnight at 4°C, washed with PBS three times, and incubated with Alexa Fluor 488-labeled goat anti-guinea pig antibody for 1 h at ambient temperature. Cells were washed three times with PBS and analyzed for GPC-positive cells by flow cytometry.
As described previously (20), 293T cells were transfected with three plasmids, pCMVRΔ8.91, a plasmid containing the HIV genome except for the packaging signal and envelope gene, pHR′GFP, a plasmid containing a GFP expression cassette and HIV packaging signal, and a pCAGGS expression vector expressing either wild-type (WT) PICV GPC or SSP mutant, using the Lipofectamine2000 (Invitrogen) transfection reagent. Supernatants of the transfected cells were collected at different time points up to 96 h posttransfection and filtered through a 0.45-μm filter. After purification through sucrose gradient ultracentrifugation, the pseudotyped VLPs were resuspended in DMEM, aliquoted, and stored at −80°C. The amount of VLPs was normalized by the HIV-1 p24 levels as determined by Western blotting using an anti-HIV-1 p24 antibody, which was kindly provided by E. Hunter (Emory University). For cell entry assay, respective GPC-pseudotyped VLPs were used to transduce 293T cells in the presence of Polybrene at 8 μg/ml. Twenty-four hours postransduction, GFP-positive cells were analyzed by flow cytometry.
WT and SSP mutant viruses were rescued as previously described (19). The L plasmid together with the S plasmid containing either WT or mutant SSP was transfected into BSRT7-5 cells. At various time points posttransfection, virus supernatants were collected for plaque assaying to determine whether any infectious virus particles were generated. Individual plaques were then picked and amplified once in BHK-21 cells to prepare virus stocks. Each viral mutant was confirmed by sequence analysis after reverse transcription-PCR (RT-PCR). For viral growth curve analysis, BHK-21 cells were infected with either WT or mutant virus at a multiplicity of infection (MOI) of 0.01, each in triplicate. At different time points postinfection, supernatants were collected and analyzed for infectious viral particles by plaque assaying in Vero cells as previously described (18).
Outbred male Hartley guinea pigs of 350 to 400 g were purchased from Charles River Laboratories and acclimatized for 3 days before experiments. Animals were administered intraperitoneally (i.p.) with PBS or 10,000 PFU of the respective virus. Body weights and rectal temperatures were monitored daily for up to 18 days. Animals were defined as having reached terminal points either when their body weight decreased by >30% compared to a nomogram or if the rectal temperature fell below 38°C in addition to body weight loss. Statistical analyses of the survival curves were performed using the log-rank (Mantel-Cox) χ2 test and/or the Fisher's exact test. The statistical significance of duration of fever was analyzed using the Student t test.
Sequence alignment of SSPs among 22 different arenaviruses identified 8 residues that are either completely or highly conserved (Fig. 1A): two at the N-terminal end (G2 and Q3), one within the h1 domain (N20), one in the loop region (K33), two within the h2 domain (N37 and F49), and two at the C-terminal end (R55 and C57). We introduced alanine substitutions at each of these conserved residues in the PICV GPC expression vector pCAGGS-PICV-GPC. We first examined whether each GPC mutant was expressed and processed properly after transfecting 1 μg of each plasmid into 293T cells and conducting a Western blot analysis with guinea pig anti-PICV antiserum, which detects both the polyprotein precursor GPC and the cleaved products GP1/GP2 but not SSP. The cleavage efficiency of each construct was calculated by dividing the amount of the cleaved products (GP1/GP2 subunits) by the total amount of the proteins (GPC precursor and GP1/GP2). Mutants G2A, Q3A, K33A, N37A, and R55A showed levels of expression and cleavage efficiency similar to that of the WT GPC (Fig. 1B, left). While N20A GPC precursor was expressed at a relatively high level, the GP1/GP2 cleavage was significantly reduced (Fig. 1B, left). Both F49A and C57A GPC precursors were expressed at much lower levels, and their GP1/GP2 processing was undetectable (Fig. 1B, left). Even after transfecting 4× more of the F49A or C57A plasmid (at 4 μg each), we could not detect any significant GP1/GP2 cleavage products (Fig. 1B, right), suggesting that these two mutations severely impaired GPC processing. In summary, among all conserved residues within the SSP, residues N20, F49, and C57 play an important role in regulating GP1/GP2 cleavage activity.
We next detected GPC assembly at the cell membrane by staining the transfected 293T cells with anti-PICV antiserum and fluorescein isothiocyanate (FITC)-conjugated secondary antibody without permeabilization. Representative histograms of empty vector- and WT GPC-expressing cells are shown in Fig. 1C (left). Using this method, we showed that, with the exception of F49A and C57A, there was no statistical significance in the surface expression levels of GPCs between WT and GPC mutants (Fig. 1C, right). In particular, N20A had a level of surface expression similar to that of the WT (Fig. 1C), even though it was defective in GP1/GP2 cleavage activity (Fig. 1B). For F49A and C57A, the reduced surface levels (~25% of WT) (Fig. 1C) most likely were due to the diminished total protein expression levels (~20% of WT) (Fig. 1B) and not because of any block(s) in the intracellular trafficking. The differential effects of F49A and C57A on GPC trafficking between our results of PICV GPC and other published studies on JUNV and lymphocytic choriomeningitis (LCMV) GPC (24, 29) may reflect the potential topological and/or biological differences between the GPCs. In summary, our data suggest that most (six out of eight) of the SSP conserved residues do not affect the intracellular transport or surface expression of the glycoprotein complex.
To evaluate the effects of individual SSP mutation on the GPC-mediated membrane fusion activity, we conducted both syncytium formation and luciferase-based (luc-based) fusion assays. For syncytium formation assay, 293T cells were transfected with a GFP expression plasmid, which allows for easy detection of the transfected cells, together with either an empty vector control or the individual GPC plasmid. Cells were exposed to a low-pH environment and observed for syncytium formation under fluorescence microscopy. Compared to the vector control (negative), WT GPC expression led to efficient formation of syncytia as evidenced by enlarged cells (Fig. 2A). Syncytium formation was also observed for Q3A, N37A, and R55A mutants (albeit to different levels) but not for G2A, N20A, K33A, F49A, and C57A mutants. To quantify the levels of GPC-mediated fusion activities, we used a luc-based fusion assay as described previously (11). GPC-mediated membrane fusion allows for the transfer of the T7 RNA polymerase from BSRT7-5 cells to transcribe the T7 promoter-driven luc reporter gene from the transfected 293T cells. Luc activity was measured for each of the GPC mutant constructs and normalized to the WT GPC control, which is set as 100% fusion efficiency (Fig. 2B). Consistent with the syncytium formations shown in Fig. 2A, mutants Q3A, N37A, and R55A induced relatively high levels of luc activities as a result of GPC-mediated membrane fusion activities (Fig. 2B). The Q3A and R55A mutants appeared to be even more fusogenic than the WT, the reason(s) for which are unknown. In contrast, N20A had a significantly lower level of fusogenic activity, whereas G2A, K33A, F49A, and C57A did not exhibit any obvious fusogenic activity in this assay. Our data are consistent with previous studies, which have demonstrated that K33A, F49A, and C57A lead to a nearly complete loss of fusogenic activity of JUNV GPC (22, 24, 25). K33 has been shown to play a critical role in mediating membrane fusion (8, 22, 23). As F49A and C57A are defective in GPC processing (Fig. 1B), the uncleaved GP1-GP2 precursor may not be able to expose the fusion peptide at the N-terminal GP2 subunit under low pH, which may explain the lack of fusogenic activity of the two mutants.
To compare the cell entry efficiency mediated by different GPC mutants, we generated lentivirus-like particles (VLPs) that contain WT or mutant GPC proteins using an established three-plasmid HIV-based system (30). Using the levels of HIV-1 p24 capsid protein on the VLPs for normalization purposes (Fig. 3A), similar amounts of GPC-pseudotyped VLPs were used to transduce 293T cells. It is worth noting that F49A and C57A VLPs contained much less GP1/GP2 (Fig. 3A), consistent with their defective GPC processing (Fig. 1B). As the VLPs encode the eGFP gene, its expression in the VLP-transduced cells was quantitatively evaluated by flow cytometry (Fig. 3B). The G3A and N20A GPC mutants showed levels of transduction efficiency (15 to 20%) similar to that of the WT, whereas the G2A mutation reduced the entry efficiency by 50% and the K33A, F59A, and C57A mutations almost completely abolished cell entry. In contrast, the N37A and R55A mutants led to even higher efficiency in cell entry than the WT GPC, the molecular mechanism(s) for which are unknown. In summary, using the GPC-pseudotyped VLP system, we have shown that some of the SSP mutations (i.e., K33A, F49A, C57A, and G2A) significantly impair GPC-mediated cell entry.
To examine the biological roles of the SSP conserved residues in the arenavirus replication cycle, we created the same series of alanine substitutions in the GPC gene of the S segment of the PICV (P18 strain) reverse genetics system (18). Together with the L plasmid, each S plasmid expressing WT or mutant SSP was transfected into BSRT7-5 cells in order to rescue recombinant viruses. We successfully rescued and amplified recombinant viruses with the individual mutations G2A, Q3A, N20A, N37A, and R55A but failed to rescue viruses with K33A, F49A, or C57A mutations even after repeated attempts (Fig. 4A), suggesting that K33, F49, and C57 are essential for the PICV life cycle. We sequenced the viable mutant viruses after passaging four times in cell culture and did not detect WT reversion.
We then compared viral growth ability between the WT and viable mutant viruses in BHK-21 cells at an MOI of 0.01 (Fig. 4B, left). Four mutants (Q3A, N20A, N37A, and R55A) showed growth kinetics similar to those of the WT, while G2A grew 0.5 to 1 log less than the WT, suggesting that the G2 residue plays an important role in viral replication. Similar results were obtained in human lung epithelial A549 cells (Fig. 4B, right). As N20A mutation significantly reduced the GPC processing in a protein expression system (Fig. 1B) but did not seem to affect viral growth in vitro (Fig. 4B), we wondered whether N20A indeed impaired the GPC processing in the context of viral infection. We infected BHK-21 cells with WT or N20A virus at an MOI of 10 for 24 h and analyzed cell lysates by Western blotting using the anti-PICV antisera, anti-NP antibody, and anti-actin antibody, respectively (Fig. 4C). The WT and N20A mutant showed similar levels of NP, consistent with their similar virus replication levels. The cleavage efficiency, which was calculated as the percentage of the cleaved products (GP1/GP2 subunits) over the total amount of the proteins (GPC precursor and GP1/GP2), was much lower in the N20A- than that in the WT-infected cells (Fig. 4C). Our results demonstrate that N20A significantly reduces the GPC processing in both the protein expression system (Fig. 1B) and viral infection (Fig. 4C).
We determined the degrees of virulence of the recombinant SSP mutants in an established guinea pig model (13, 17, 18). Consistent with our previous observations (18), all animals (n = 6) infected with the WT virus experienced a long duration of fever (average of 9 days) and significant body weight loss, and they reached terminal points by 14 days postinfection (Fig. 5). In contrast, all animals (n = 3) infected with G2A or N20A mutant virus survived the infection with no obvious body weight loss compared to the PBS control group (Fig. 5A and andB).B). The G2A-infected animals also experienced significantly fewer days of fever (average of 3.8 days) (Fig. 5C) than the WT-infected animals. Similarly, the N20A virus infection led to, on average, a shorter duration of fever than WT virus infection, although the difference is not statistically significant due to the relatively small number of infected animals (Fig. 5C). In summary, both the G2A and N20A mutant viruses almost completely lost virulence in vivo (Fig. 5A).
Infection with the Q3A mutant virus (n = 7) led to ~90% mortality, substantial loss of body weight, and prolonged fever in guinea pigs without significant difference from the WT virus infection (Fig. 5), demonstrating that the Q3A mutation does not greatly attenuate the virus in vivo. On the other hand, animals (n = 6) infected with either the N37A or R55A mutant virus experienced prolonged fever and variable degrees of body weight loss, with a survival rate of 30% to 50% that is significantly higher than that for animals infected with the WT virus (Fig. 5). Collectively, our data suggest that both the N37A and R55A mutant viruses are attenuated in vivo. It is worth mentioning that, for all mutant viral infections, animals that survived the infection did not contain any trace of infectious virus in the serum while all moribund animals had high viremia as determined by plaque assay, correlating with the previous reports that arenavirus virulence is associated with high levels of viral replication in vivo (3, 21, 31). We also sequenced all mutant viruses isolated from animals and did not detect any WT revertants.
In this study, we have determined the biological roles of 8 completely conserved residues of GPC SSP among all known arenaviruses using a PICV reverse genetics system. Table 1 shows the effects of alanine substitution of each conserved SSP residue on the expression and function of GPC as well as on viral infections in vitro and in vivo. Except for N20A, F49A, and C57A, other SSP mutations do not affect the levels of GPC expression, GP1/GP2 cleavage, or surface assembly of the glycoprotein complex (Fig. 1). However, multiple SSP conserved residues are essential for viral infection by participating in viral entry, while two residues (N37 and R55) are important for viral virulence by a yet-to-be-determined mechanism(s).
K33, F49, and C57 are required for viral entry and are indispensable for the PICV life cycle. Our results correlate well with prior studies of other arenaviruses, such as JUNV and LCMV (8). The K33 residue has been shown to play an essential role in GPC fusion by interacting with GP2 to prime the complex for pH-induced membrane fusion (8, 22, 23). The F49 residue is located within the recently identified FILL sorting signal motif that is required for glycoprotein processing and surface expression (8, 24). The C57 residue of GP2, along with its H459, C467, and C469 residues, form a zinc-binding center, which provides the structural basis for association with SSP in the glycoprotein complex in order to modulate optimal membrane fusion activity (25, 26).
Another residue critical for viral fusion is the G2 myristoylation site. The myristoylation-defective G2A mutation almost completely abolishes the fusion activity according to the cell-based fusion assay (Fig. 2). It is, however, worth mentioning that the fusion assay cannot fully reproduce the fusion activity of GPC during viral entry, and therefore the result does not exclude the possibility that G2A mutation retains a low level of fusogenic activity. Indeed, we show that G2A GPC-pseudotyped VLPs can still mediate cell entry, albeit at a reduced level (Fig. 3B). As such, the G2A mutant virus can be rescued but is significantly attenuated in vitro and in vivo (Fig. 4 and and5).5). Previous studies of JUNV and LCMV GPCs have also shown that the G2A mutation can markedly reduce the membrane fusion activity (8, 28). The molecular mechanism by which the G2 myristoylation site is involved in viral membrane fusion remains unknown. Myristoylation plays an important role in membrane targeting and protein-protein interactions (27). We and others have shown that GPC with the G2A substitution is still able to express properly and to assemble at the plasma membrane (Fig. 1C) (8, 28), suggesting that this G2 myristoylation site is not essential for membrane targeting but may be involved in the SSP-GP2 interaction in order to mediate proper membrane fusion.
Consistent with previous studies (8, 11), the N20A mutation reduces the membrane fusion activity by ~75% (Fig. 2B). The N20 residue is located within SSP transmembrane domain 1 (TM1), which has been suggested to directly interact with the GP2 TM domain to affect GPC-mediated membrane fusion (11). The reduced fusion activity of the N20A mutation does not significantly affect VLP-mediated cell entry (Fig. 2B) or viral growth in vitro (Fig. 4B) but markedly reduces viral virulence in vivo (Fig. 5), suggesting that suboptimal membrane fusion activity could have a major impact on viral infection in vivo.
The N37A and R55A mutations do not reduce the membrane fusion activity (Fig. 2B) but rather appear to increase the levels of VLP-mediated cell entry efficacy (Fig. 3B). While these mutant viruses grow similarly to the WT virus in cell culture (Fig. 4B), they showed reduced degrees of virulence in vivo (Fig. 5A and andB).B). It is possible that the seemingly insignificant effects observed in cell culturing conditions can sometimes be amplified in vivo. The exact molecular mechanism(s) by which N37 and R55 contribute to viral infectivity in vivo remain to be determined, but they do not seem to participate in membrane fusion and/or cell entry.
In summary, our data suggest that arenavirus SSP plays an essential role in membrane fusion and that SSP also contributes to viral virulence by a potentially fusion-independent mechanism. Knowledge gained from this study should aid a structure-directed approach toward the design of antivirals against SSP as a potential target and the development of attenuated viruses for use as vaccines against deadly human hemorrhagic fever-causing arenaviruses.
We thank J. Aronson (University of Texas Medical Branch) for providing the stock P2 and P18 viruses, K. Conzelmann (Ludwig-Maximilians-Universität, Germany) for the BSRT7-5 cells, J. Nunberg (University of Montana, USA) for the T7 promoter-directed firefly luciferase plasmid, and E. Hunter (Emory University, USA) for anti-p24 antibody.