Interaction of Human Parainfluenza Virus Type 3 Nucleoprotein with Matrix Protein Mediates Internal Viral Protein Assembly

ABSTRACT

Human parainfluenza virus type 3 (HPIV3) belongs to the Paramyxoviridae family. Its three internal viral proteins, the nucleoprotein (N), the phosphoprotein (P), and the polymerase (L), form the ribonucleoprotein (RNP) complex, which encapsidates the viral genome and associates with the matrix protein (M) for virion assembly. We previously showed that the M protein expressed alone is sufficient to assemble and release virus-like particles (VLPs) and a mutant with the L305A point mutation in the M protein (M_L305A) has a VLP formation ability similar to that of wild-type M protein. In addition, recombinant HPIV3 (rHPIV3) containing the M_L305A mutation (rHPIV3-M_L305A) could be successfully recovered. In the present study, we found that the titer of rHPIV3-M_L305A was at least 10-fold lower than the titer of rHPIV3. Using VLP incorporation and coimmunoprecipitation assays, we found that VLPs expressing the M protein (M-VLPs) can efficiently incorporate N and P via an N-M or P-M interaction and M_L305A-VLPs had an ability to incorporate P via a P-M interaction similar to that of M-VLPs but were unable to incorporate N and no longer interacted with N. Furthermore, we found that the incorporation of P into M_L305A-VLPs but not M-VLPs was inhibited in the presence of N. In addition, we provide evidence that the C-terminal region of P is involved in its interaction with both N and M and N binding to the C-terminal region of P inhibits the incorporation of P into M_L305A-VLPs. Our findings provide new molecular details to support the idea that the N-M interaction and not the P-M interaction is critical for packaging N and P into infectious viral particles.

IMPORTANCE Human parainfluenza virus type 3 (HPIV3) is a nonsegmented, negative-sense, single-stranded RNA virus that belongs to the Paramyxoviridae family and can cause lower respiratory tract infections in infants and young children as well as elderly or immunocompromised individuals. However, no effective vaccine has been developed or licensed. We used virus-like particle (VLP) incorporation and coimmunoprecipitation assays to determine how the M protein assembles internal viral proteins. We demonstrate that both nucleoprotein (N) and phosphoprotein (P) can incorporate into M-VLPs and N inhibits the M-P interaction via the binding of N to the C terminus of P. We also provide additional evidence that the N-M interaction but not the P-M interaction is critical for the regulation of HPIV3 assembly. Our studies provide a more complete characterization of HPIV3 virion assembly and substantiation that N interaction with M regulates internal viral organization.

INTRODUCTION

Human parainfluenza virus type 3 (HPIV3) is a negative-strand RNA virus (NSV) that belongs to the Paramyxoviridae family and often causes lower respiratory tract infections in infants and young children. The HPIV3 genome consists of 6 open reading frames that encode 6 structural proteins: the nucleoprotein (N), phosphoprotein (P), RNA-dependent RNA polymerase (L), matrix protein (M), and two glycoproteins, hemagglutinin/neuraminidase (HN) and the fusion protein (F). N, P, and L encapsulate the viral RNA to form a helical assembly termed the ribonucleoprotein (RNP) complex, which is the minimum structure required for viral transcription and replication. HN is involved in viral attachment to the host cell, while F is required for fusion with the host cell plasma membrane. The M protein binds directly to the viral envelope. For most NSVs, the M protein is the primary force driving viral assembly, and the budding and formation of virus-like particles (VLPs) are critically dependent on the presence of viral M proteins (1–6). In some NSVs, M-protein expression alone is sufficient for the formation and release of VLPs, such as the M proteins of human parainfluenza virus type 1 (7), Sendai virus (8), respiratory syncytial virus (RSV) (9), measles virus (10), Nipah virus (5), Newcastle disease virus (4), vesicular stomatitis virus (11), Ebola virus (12), and influenza A virus (13). In contrast, the M proteins of other NSVs, such as mumps virus (6) and parainfluenza virus type 5 (PIV5) (14), require an accessory protein, e.g., F or N, for maximum VLP release efficiency. Previously, we also demonstrated that the M protein of human parainfluenza virus type 3 (HPIV3) alone in mammalian cells could lead to the formation and release of enveloped VLPs (15), which are morphologically similar to virions.

In general, in the process of virion assembly, the M protein links internal viral proteins through interaction with RNPs and envelope glycoproteins via their cytoplasmic tails. However, the mechanism by which RNPs are recruited to budding sites and incorporated into viral particles seems to be different for different NSVs and is not well understood. A recent study showed that, even within the Paramyxoviridae family, the architecture of the various virions is often different (16). This may be due to the differences in RNP assembly into virions. For RSV, a transcription antiterminator, M2-1, mediates the association of RNPs with the M protein and is required for the incorporation of RNPs into virions (17), and further structural analysis showed that M2-1 is located between the RNP and M in isolated viral particles (18). However, for viruses belonging to Paramyxovirinae subfamilies and some other enveloped viruses, such as retroviruses and filoviruses, N has been described to be a mediator of virion assembly and budding (6, 14, 19–21). At least two reports suggested that N of influenza virus plays a critical role in virion assembly, possibly through its interaction with M1. One report showed that a temperature-sensitive mutation at N residue 239 results in the production of abnormally shaped virions without affecting virus RNA synthesis (22). Another report showed that three N residues, 214, 217, and 253, which localize at the potential M1 interaction sites on the viral RNP, play a critical role in virion morphology (23).

In addition, interactions between M and the glycoprotein (HN and F) cytoplasmic tails of parainfluenza viruses have been thought to play a critical role in virion assembly at specific locations on plasma membranes (24–27), but the contribution of HN and F in virion formation may differ between viruses. For PIV5, the deletion of the HN cytoplasmic tail resulted in a marked defect in viral budding and release, whereas the F cytoplasmic tail was dispensable for normal viral budding (26). In contrast, for Sendai virus, HN was not necessary for virion budding, and the loss of the cytoplasmic tail of F protein resulted in a reduction in the levels of accumulation of M and RNP at the plasma membrane and overall virion production (28–30). It is possible that RNP accumulation at assembly sites requires a stable interaction with the M protein at the plasma membrane, which could be provided by the specific interaction of M with the F cytoplasmic tail. So far, no direct evidence has shown that the F cytoplasmic tail is associated with the RNP.

In this study, to further understand the mechanism of HPIV3 assembly and release, we characterized the requirements for internal viral protein assembly using VLP and coimmunoprecipitation assays. Specifically, we sought to determine how N and P are assembled into viral VLPs, whether the P-M interaction is involved in HPIV3 assembly, and how the N-P interaction inhibits the incorporation of P into VLPs expressing the M protein (M-VLPs). Our findings suggest a direct role of N in linking the viral N-P complex to the M protein, thereby promoting internal viral protein incorporation into virions.

MATERIALS AND METHODS

Cells and virus.

Cells (293T and LLC-MK2 [MK2] cells) were maintained in Dulbecco's modified Eagle's medium (Gibco) supplemented with 10% fetal bovine serum (FBS; HyClone). Wild-type HPIV3 (NIH 47885) and recombinant HPIV3 (rHPIV3) expressing the M protein with the L305A point mutation (M_L305A) were propagated in MK2 cells by inoculation at a multiplicity of infection (MOI) of 0.1.

Antibodies.

Mouse monoclonal anti-Myc antibody and anti-green fluorescent protein (anti-GFP) antibody were purchased from Santa Cruz Biotechnology. Mouse antihemagglutinin (anti-HA) and anti-Flag monoclonal antibodies were from Sigma. A mouse monoclonal antibody to the anti-HN protein (anti-HN) of parainfluenza virus type 3 was from Abcam.

Plasmid constructs.

The plasmids carrying HA-M, HA-M_L305A, HA-P, Myc-M, Myc-M_L305A, Myc-N, and N-Flag have been described previously (15). pOCUS-HPIV3 was used as a template for PCR amplification in the genetic manipulations. cDNA encoding mutant N_L478A with a Flag epitope tag at the C terminus was cloned into pCAGGS by using site-specific mutagenesis. PCR products encoding P mutants with deletions of 20, 40, 60, 80, or 100 amino acids at the C terminus (PΔC20, PΔC40, PΔC60, PΔC80, and PΔC100, respectively) and an HA tag at the N terminus were also cloned into pCAGGS. The P mutant with a 500-amino-acid deletion at the N terminus (PΔN500) was fused with glutathione S-transferase (GST) and cloned into pCAGGS. The sequences of all the aforementioned constructs were verified via DNA sequencing.

Determination of virus growth curves.

Monolayers of MK2 cells in 6-well plates were grown to 50 to 60% confluence and infected with wild-type HPIV3 or HPIV3 expressing M_L305A (HPIV3-M_L305A) at an MOI of 0.001. Then, the infection medium was removed and replaced with fresh medium containing 4% FBS, and the supernatant was sampled every 12 h for 4 days prior to determination of the infectious virus titer, which was performed by standard plaque assay on MK2 cells as described previously (15). After the viral titers were calculated, growth curves were determined by using the software GraphPad Prism (version 5.0).

Ultrathin sectioning and transmission electron microscopy.

In 6-well plates, 293T cells were grown to 40 to 50% confluence and transfected with the plasmids indicated below; then, the culture medium was replaced with fresh medium containing 10% FBS at 24 h posttransfection. At 48 h posttransfection, the cells were fixed with 2.5% glutaraldehyde and 4% paraformaldehyde in 0.1 M sodium phosphate buffer (pH 7.4) for 2 h at room temperature. The cells were harvested and fixed with 2.5% glutaraldehyde on ice for 2 h and then fixed in 2% osmium tetroxide. Then, the cells were dehydrated with sequential washes in 50%, 70%, 90%, 95%, and 100% ethanol. Areas containing cells were block mounted and thinly sliced. Final samples were absorbed onto a carbon-coated copper grid negatively stained with 1% phosphotungstic acid (pH 7.0) and then analyzed on a transmission electron microscope.

VLP budding assays.

For the preparation of VLPs to be examined, 293T cells in 6-well plates were grown to 40 to 50% confluence and transfected with a pCAGGS plasmid carrying M, M_L305A, N, or P by using calcium phosphate transfection. The plasmid amount was equalized by the amount of empty pCAGGS plasmid. At 48 h posttransfection, the supernatant was harvested and clarified by centrifugation at 13,000 rpm for 3 min and then pelleted through a 20% sucrose cushion and subsequently ultracentrifuged at 35,000 rpm for 2 h at 4°C on a P55 ST2 rotor (Hitachi). The VLPs pelleted at the bottom of the tubes were resuspended in 40 μl cold sodium chloride-Tris-EDTA buffer. Cells were scraped into cold phosphate-buffered saline (PBS), pelleted by centrifugation at 13,000 rpm for 1 min, lysed in cold TNE buffer (50 mM Tris-HCl [pH 7.4], 150 mM NaCl, 2 mM EDTA [pH 8.0], 0.1% 2-mercaptoethanol, protease inhibitor cocktail) for 30 min on ice, and then centrifuged at 13,000 rpm for 30 min at 4°C. VLP pellets and cell lysates were mixed with 5× sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis (PAGE) protein-loading buffer (250 mM Tris-HCl [pH 6.8], 10% SDS, 50% glycerol, 0.3% bromophenol blue, 2.5% 2-mercaptoethanol), incubated at 100°C for 10 min, and subjected to Western blot analysis.

Western blot analysis.

The prepared samples were separated by 10% or 12% SDS-PAGE and then electroblotted onto a nitrocellulose membrane. The membranes were blocked with skim milk in phosphate-buffered saline with Tween 20 (1/1,000 Tween 20) for 30 min at room temperature and subsequently incubated with primary antibodies for 1 h and secondary antibodies for 1 h. The primary antibodies used were as follows: anti-Myc (1:2,500), anti-HA (1:10,000), anti-GFP (1:2,500), anti-Flag (1:10,000), and anti-HN (1:2,500). The secondary antibodies, used at a 1:5,000 dilution, were goat anti-rabbit IgG and goat anti-mouse IgG.

Protease protection assay.

The aforementioned prepared VLPs were divided into four aliquots and subjected to the following treatments: (i) no treatment, (ii) treatment with Triton X-100 to a final concentration of 1%, (iii) treatment with tosylsulfonyl phenylalanyl chloromethyl ketone-treated trypsin (Sigma-Aldrich) to a final concentration of 1 μg/ml, or (iv) treatment with Triton X-100 plus trypsin. All the samples were incubated at 37°C for 30 min and then mixed with 5× SDS-PAGE protein-loading buffer, incubated at 100°C for 10 min, and subjected to Western blot analysis.

Coimmunoprecipitation assay.

In 6-well plates, 293T cells were grown to 40 to 50% confluence and transfected with the indicated plasmids. At 48 h posttransfection, cell lysates were prepared in 350 μl TNE buffer as described above in “VLP budding assays”; 40 μl lysates was removed for input analysis, and the remaining cell lysates were then incubated with relevant antibodies (anti-c-Myc tag affinity gel [BioLegend] or anti-Flag affinity gel [Sigma]) overnight at 4°C with gentle rotation. The beads were collected by centrifugation at 5,000 rpm for 2 min at 4°C and washed 3 times with TNE buffer. The beads were then mixed with 2× SDS-PAGE loading buffer, incubated at 100°C for 10 min, and subjected to Western blot analysis.

Immunofluorescence analysis.

HeLa cells grown on coverslips in 12-well plates were transfected with pCAGGS-N-Flag or pCAGGS-HA-P alone or together with pCAGGS-Myc-M or pCAGGS-Myc-M_L305A using the Lipofectamine 2000 reagent (Invitrogen). At 24 h posttransfection, the cells were washed three times with cold PBS, fixed with 4% paraformaldehyde for 20 min, permeabilized with 0.2% Triton X-100 for 20 min, and blocked by 3% bovine serum albumin (BSA) for 30 min, and then the cells were incubated with rabbit polyclonal anti-Myc antibody (1:200; Santa Cruz), mouse monoclonal anti-HA antibody (1:2,000; Sigma), or mouse monoclonal anti-Flag antibody (1:2,000) for 1 h. After these incubations, the cells were washed three times with 1% BSA and incubated with goat anti-rabbit IgG rhodamine (1:100; Thermo) or goat anti-mouse IgG fluorescein (1:200; Thermo) secondary antibody for 1 h. The cells were then washed three times with cold PBS and stained with DAPI (4′,6-diamidino-2-phenylindole) for 5 min. Confocal images were collected to visualize the location of N-M and N-M_L305A or P-M and P-M_L305A.

RESULTS

Recombinant HPIV3-M_L305A shows lower titers than wild-type HPIV3.

Our previous study showed that expression of the M protein alone is sufficient to form and release VLPs, a mutant with a point mutation of the HPIV3 M protein (M_L302A) is deficient in VLP production, and recombinant HPIV3 expressing M_L302A could not be rescued; in contrast, a mutant with another point mutation of the HPIV3 M protein (M_L305A) could release VLPs as efficiently as wild-type M and recombinant HPIV3 expressing M_L305A could be readily rescued (15). However, to our surprise, when we performed a growth curve assay with HPIV3-M_L305A and compared its growth to that of wild-type HPIV3, HPIV3-M_L305A showed markedly slower replication kinetics in MK2 cells, with virus titers being up to 12-fold lower for HPIV3-M_L305A than for wild-type HPIV3 (Fig. 1A), although HN protein expression in the cell lysates at different time points appeared to be comparable between wild-type HPIV3 and HPIV3-M_L305A (Fig. 1B).

FIG 1.

Recombinant HPIV3 expressing M_L305A showed lower titers than wild-type HPIV3, but M_L305A was still able to form authentic VLPs. (A) Growth curve of the recombinant wild-type (WT) HPIV3 and HPIV3-M_L305A. MK2 cells were infected as described in Materials and Methods, and the supernatant was harvested every 12 h (as indicated on the x axis) for the plaque assay to determine virus growth curves. Standard errors were calculated from two independent experiments. (B) HN expression from cell lysates infected with recombinant wild-type HPIV3 and HPIV3-M_L305A. MK2 cells were infected as described in Materials and Methods, and the cells was pelleted every 12 h (as indicated by the numbers beneath the lanes) for the HN expression assay. (C) Representative transmission electron microscopy images of ultrathin sections of 293T cells. 293T cells were transfected with M, M_L305A, or M_L302A. After 48 h, the cells were treated as described in Materials and Methods and then visualized by transmission electron microscopy. Arrows, the VLPs outside the sliced cells. (D) Protease protection assay for M- and M_L305A-VLPs. The indicated purified VLP samples were treated as described in Materials and Methods and then analyzed via Western blotting using anti-HA antibody. WB, Western blotting; GAPDH, glyceraldehyde-3-phosphate dehydrogenase.

To determine whether the nature of the VLPs formed by M_L305A is indeed the same as that of the VLPs formed by wild-type M protein, we analyzed the morphology of VLPs by using ultrathin sectioning and transmission electron microscopy and found that the VLPs released by M_L305A were morphologically similar to those released by wild-type M protein (Fig. 1C). Furthermore, a protease protection assay showed that VLPs released by M_L305A were also enclosed by a lipid bilayer because it could be digested only when treated with trypsin plus Triton X-100 (Fig. 1D, lane 4). These data suggest that M_L305A alone is indeed fully functional in producing authentic VLPs, as wild-type M protein and the decrease in the HPIV3-M_L305A titer might be caused during the virion assembly and budding process.

The M protein can specifically incorporate N or P into M-VLPs.

Since the M protein of paramyxoviruses is the primary driving force of the virion assembly process and plays a linking role by interacting with both viral RNPs and glycoproteins (31) and earlier studies showed that the incorporation of RNPs into viral particles is determined by the interaction of the M protein with the RNPs (32), we first sought to determine whether N or P could be incorporated into M-VLPs. To this end, we coexpressed N or P with the M protein in 293T cells and performed the VLP budding assay. We also included GFP as a negative control since GFP does not interact with M and exists only in lysates and not in the pellet fraction of the culture medium if lysis of cells does not occur. When expressed alone, neither N nor P was detected in the pellet fraction of the culture medium (Fig. 2A and B, top), but N and P could be detected in the pellet fraction of the culture medium when coexpressed with M, whereas GFP could not be detected (Fig. 2A and B, bottom, lanes 4), in spite of its expression in lysates, suggesting that N and P in the pellet fraction of the culture medium were not released by lysis of cells and the M protein could specifically incorporate the N or P protein into VLPs. To confirm that N or P was indeed incorporated into M-VLPs, we performed a protease protection assay, and the fractions from either N-M-transfected or P-M-transfected cell medium were treated as described in the legend to Fig. 1D. The results showed that N and P together with the M protein were all protected from proteolysis by trypsin treatment alone (Fig. 2C and D, lanes 3), but the addition of Triton X-100 disrupted the lipid envelope of VLPs, resulting in the degradation of the N, P, and M proteins in the presence of trypsin (Fig. 2C and D). These data suggest that N or P is specifically incorporated into the M-VLPs.

FIG 2.

M could specifically incorporate N or P into M-VLPs. (A, B) VLP budding assays for the cotransfection of N and M (A) or P and M (B); GFP was included as a negative control. 293T cells were transfected with the indicated plasmids for 48 h. The cell lysates and VLPs were prepared as described in Materials and Methods, and then the samples were subjected to Western blot analysis by using anti-Myc, anti-Flag, anti-HA, or anti-GFP antibodies. Arrows, the bands of the detected proteins. (C, D) Protease protection assays for the N-M and P-M VLPs. 293T cells were cotransfected with either N plus M (C) or P plus M (D). The VLP samples in the supernatant were divided into four aliquots, treated as described in Materials and Methods, and then analyzed via Western blot analysis with anti-Flag, anti-HA, or anti-Myc antibodies. Arrows, the bands of the detected proteins.

M_L305A can incorporate P but is deficient in incorporating N into M_L305A-VLPs.

Next, we sought to determine whether M_L305A could also incorporate N or P as efficiently as wild-type M protein. To this end, we coexpressed N or P with M_L305A in 293T cells and performed the VLP assay as described in the legends to Fig. 2A and B. The results showed that when N was coexpressed with the M protein, N could be detected in the VLPs (Fig. 3A, bottom right, lane 2), which is consistent with the results in Fig. 2A; conversely, when coexpressed with M_L305A, N could barely be detected in VLPs (Fig. 3A, bottom right, lane 3), even though M_L305A had a VLP formation ability similar to that of wild-type M protein. However, when coexpressed with either M or M_L305A, P was readily detected in both M- and M_L305A-VLPs at a similar intensity (Fig. 3B, bottom right, lanes 2 and 3). We also assessed the ability of the M protein and M_L305A to incorporate HN into VLPs, and the results showed that both the M protein and M_L305A can incorporate the HN protein into VLPs at the same level (Fig. 3C). Furthermore, we found that M_L305A also has an oligomerization ability similar to that of M in a coimmunoprecipitation assay (Fig. 3D). When these data are taken together, these results suggest that M_L305A maintained the ability to effectively incorporate P into VLPs, but its ability to incorporate N into VLPs was severely weakened.

FIG 3.

M_L305A is able to incorporate P but not N into VLPs. (A to C) VLP budding assays for the cotransfection of N and M/M_L305A (A), P and M/M_L305A (B), or HN and M/M_L305A (C). 293T cells were transfected with HN and M/M_L305A. At 48 h posttransfection, VLP budding assays were performed as described in the legend to Fig. 2. (D) M_L305A interacts with both M and itself. 293T cells were transfected with plasmids carrying the indicated proteins. At 48 h posttransfection, immunoprecipitation (IP) assays were performed with anti-Myc antibodies. Western blotting was performed with anti-Myc, anti-HA, and anti-HN antibodies. Arrows, the bands of the detected proteins.

M_L305A maintained its interaction with P but not with N.

To determine whether the inability of M_L305A to incorporate N into VLPs is caused by the lack of interaction between M_L305A and N, we examined the interaction of N with M or M_L305A by performing a coimmunoprecipitation assay. A plasmid carrying Myc-tagged M or M_L305A was transfected either individually or jointly with a plasmid carrying N-Flag into 293T cells, immunoprecipitation was performed by using an anti-Flag affinity gel, and Western blotting was carried out with anti-Myc and anti-Flag monoclonal antibodies. Wild-type M protein was efficiently coimmunoprecipitated with N-Flag, whereas the coimmunoprecipitation ability of M_L305A was severely weakened (Fig. 4A, top, lanes 2 and 4); furthermore, immunofluorescence results also showed that the colocalization ability of M_L305A with N was dramatically decreased (Fig. 4B), suggesting that M interacts with N, but the M_L305A interaction with N was severely weakened. Next, the interaction of HA-P with Myc-tagged M or M_L305A was also examined via coimmunoprecipitation assay using an anti-Myc affinity gel, and the results showed that HA-P was coimmunoprecipitated either by Myc-M or by Myc-M_L305A at similar levels (Fig. 4C, top, lanes 2 and 3), and immunofluorescence assay results also showed that both M and M_L305A similarly colocalized with P (Fig. 4D), suggesting that both M and M_L305A can interact with P. As expected, the ability of M or M_L305A to interact with N and P correlated with their ability to incorporate N and P into VLPs. So far, we can conclude that the lower titers of HPIV3-M_L305A were caused by the inability of M_L305A to interact with N and to incorporate N into viral particles.

FIG 4.

M_L305A maintained its interaction with P but not N. (A, C) Coimmunoprecipitation assays for N and M/M_L305A (A) or P and M/M_L305A (C). 293T cells were transfected with plasmids carrying the indicated proteins. At 48 h posttransfection, the cells were harvested and subjected to a coimmunoprecipitation assay as described in Materials and Methods. (A) Immunoprecipitation was performed using an anti-Flag affinity gel, and lysates were detected via Western blotting using anti-Myc and anti-Flag monoclonal antibodies. (C) Immunoprecipitation was performed using an anti-Myc affinity gel, and lysates were detected via Western blotting using anti-HA and anti-Myc monoclonal antibodies. Arrows, the bands of the detected proteins. (B, D) Colocalization assays for N and M/M_L305A (B) or P and M/M_L305A (D). HeLa cells were transfected with plasmids carrying the indicated proteins. At 24 h posttransfection, cells were fixed and stained with anti-Flag, anti-Myc, and anti-HA antibodies and visualized via confocal microscopy.

N inhibits the incorporation of P into M_L305A-VLPs but not M-VLPs.

Having found that wild-type M-VLPs could incorporate either N or P, whereas M_L305A-VLPs incorporated P but not N, we next sought to determine what would happen when the N, P, and M proteins were expressed together. Both N and P were efficiently incorporated into M-VLPs when N, P, and M were expressed together (Fig. 5A, right, lane 4). When N, P, and M_L305A were coexpressed, not only was it found that N could not be incorporated into M_L305A-VLPs, but also the incorporation of P into M_L305A-VLPs was found to be severely inhibited, despite the efficient formation and release of M_L305A-VLPs (Fig. 5, right, lane 5), which prompted a key question: how does N inhibit the incorporation of P into M_L305A-VLPs but not into M-VLPs? To answer this question, we used a point mutant of N (N_L478A), which was unable to interact with P in our previous study (33), to further examine whether N_L478A also inhibits the incorporation of P into M_L305A-VLPs. N or N_L478A was coexpressed with P and M_L305A. We found that N_L478A was unable to inhibit the incorporation of P into M_L305A-VLPs (Fig. 5, middle right, lane 7). Of note, similar to the findings for N, N_L478A could still be incorporated into M-VLPs other than those induced by M_L305A (Fig. 5, bottom right, lanes 6 and 7). Taken together, these data lead to the following conclusions: (i) N inhibits the incorporation of P into M_L305A-VLPs but not M-VLPs; (ii) N_L478A is unable to interact with P, and therefore, N_L478A is also unable to inhibit the incorporation of P into M_L305A-VLPs; and (iii) M assembles the N-P complex into viral particles via an N-M interaction but not an P-M interaction because the N-P interaction inhibits the P-M interaction.

FIG 5.

N but not N_L478A inhibits the incorporation of P into M_L305A. The results of VLP budding assays for the cotransfection of N or N_L478A and P and of N or N_L478A and M/M_L305A are shown. 293T cells were transfected with plasmids carrying the indicated proteins for 48 h. The VLP budding assays were performed as described in the legend to Fig. 2. Then, samples were analyzed via Western blotting with anti-HA, anti-Myc, and anti-Flag antibodies. Arrows, the bands of the detected proteins.

Both M and N interact with the C terminus of P.

Having established that the N-P interaction inhibits the P-M_L305A interaction, we sought to identify the functional domains within P that are critical for the incorporation of P into M-VLPs and to determine whether N interacts with P and inhibits the incorporation of P into M_L305A-VLPs via this domain. To this end, we constructed a series of mutants with mutations in the C terminus of P, coexpressed each of the mutants with M in 293T cells, and performed VLP budding assays. The PΔC20, PΔC40, PΔC60, and PΔC80 mutants with deletions at the C terminus of P could be incorporated into M-VLPs as efficiently as wild-type P (Fig. 6A, bottom right, lanes 3 to 6); however, the PΔC100 mutant was barely detected in the M-VLPs (Fig. 6A, bottom right, lane 7), suggesting that the residues between residues 80 and 100 of the C terminus of P are critical for its interaction with M and for the incorporation of P into M-VLPs. To exclude the possibility that the failure to incorporate PΔC100 into M-VLPs was due to a conformational alteration of P, we took advantage of another construct, PΔN500, which expresses only 102 residues of the C terminus of P. Since the PΔN500 mutant showed very poor expression of P and GST-fused PΔN500 dramatically enhanced its expression, we used this construct for VLP assays, and the results showed that GST-PΔN500 could be specifically incorporated into M-VLPs as efficiently as wild-type P (Fig. 6B, bottom right, lanes 2 and 6), whereas GST alone could not be incorporated into M-VLPs (Fig. 6B, bottom right, lane 4). Next, by using coimmunoprecipitation assays, we examined the interaction of GST-PΔN500 with M or M_L305A. The results showed that GST-PΔN500 was coimmunoprecipitated by both Myc-M and Myc-M_L305A (Fig. 6C, top right, and D, top, lanes 4), whereas GST alone was not coimmunoprecipitated by Myc-M or Myc-M_L305A (Fig. 6C, top right, and D, top, lanes 2), suggesting that the C terminus of P is indispensable and sufficient to interact with the M protein and to incorporate P into M-VLPs.

FIG 6.

Both M_L305A and N can interact with the 100 residues in the C terminus of P. (A) VLP budding assays for the cotransfection of M and mutant P. 293T cells were transfected with plasmids carrying the indicated proteins. At 48 h posttransfection, VLP budding assays were performed as described in the legend to Fig. 2. Then, samples were analyzed via Western blotting with anti-Myc and anti-HA antibodies. (B) VLP budding assays for the cotransfection of M and PΔN500. Plasmids carrying the indicated proteins were transfected into 293T cells for 48 h. VLP budding assays were performed as described in the legend to Fig. 2. Then, samples were analyzed via Western blotting with anti-Myc and anti-HA antibodies. (C, D) Coimmunoprecipitation assays for M and PΔN500 (C) or M_L305A and PΔN500 (D). 293T cells were transfected with the plasmids carrying the indicated proteins. At 48 h posttransfection, the cells were processed as described in the legend to Fig. 4, immunoprecipitation was performed using an anti-Myc affinity gel, and lysates were detected via Western blotting using anti-HA and anti-Myc monoclonal antibodies. (E) Coimmunoprecipitation assays for N and PΔN500. Cells were transfected and processed as described in the legend to panel C, except that immunoprecipitation was performed using an anti-Flag affinity gel and lysates were detected via Western blotting using anti-HA and anti-Flag monoclonal antibodies. Arrows, the bands of the detected proteins.

Next, we sought to determine whether N interacts with P via the C terminus of P; thus, we assessed the interaction between GST-PΔN500 and N. GST-PΔN500 was expressed alone or together with N-Flag in 293T cells, and immunoprecipitation was performed by using an anti-Flag affinity gel. The results revealed that, as a positive control, P interacted with N (Fig. 6E, top, lane 2), as did GST-PΔN500 (Fig. 6E, top, lane 6), whereas GST alone failed to interact with N-Flag (Fig. 6E, top, lane 4). Taken together, our results show that both N and M interact with P via the C terminus of P.

N inhibits incorporation of PΔN500 into M_L305A-VLPs, while N_L478A does not.

Having confirmed that both N and M interact with P via the C terminus of P and that N inhibits the incorporation of P into M_L305A-VLPs, we sought to determine whether N also inhibits the incorporation of GST-PΔN500 into M_L305A-VLPs. GST-PΔN500 was coexpressed with M_L305A or M with or without N, and VLP budding assays were performed. GST-PΔN500 was readily detected in both M_L305A- and M-VLPs (Fig. 7A, top right, lanes 2 and 3) in the absence of N, whereas when N was present, GST-PΔN500 was unable to be incorporated into M_L305A-VLPs but could still be incorporated into M-VLPs (Fig. 7A, top right, lanes 4 and 5). As expected, N_L478A lost the ability to inhibit the incorporation of GST-PΔN500 into M_L305A-VLPs (Fig. 7B, top right, lane 5). Taken together, these data show that N but not N_L478A indeed inhibits the incorporation of P into M_L305A-VLPs via interaction with the C terminus of P.

FIG 7.

N but not N_L478A inhibits the incorporation of M_L305A into PΔN500. (A) VLP budding assays for the cotransfection of N, PΔN500, and M/M_L305A. (B) VLP budding assays for the cotransfection of N_L478A, PΔN500, and M/M_L305A. 293T cells were transfected with plasmids carrying the indicated proteins for 48 h. The VLP budding assays were performed as described in the legend to Fig. 2, and then samples were analyzed via Western blotting with anti-HA, anti-Flag, and anti-Myc antibodies. Arrows, the bands of the detected proteins.

DISCUSSION

In this study, we found that recombinant HPIV3-M_L305A shows lower titers than wild-type HPIV3 (Fig. 1A). Subsequently, we found that the levels of expression of the M protein and M_L305A and release of VLPs are similar (Fig. 1C and D). Therefore, we hypothesized that M_L305A has an effect on virion assembly. To prove this, we first demonstrated that the M protein of HPIV3 interacts with both N and P and can incorporate both N and P into M-VLPs (Fig. 2 and 4). Because N also interacts with P to form the N-P complex (33), it is unclear whether N or P mediates the interaction of the N-P complex with M in the virion assembly process. Since M_L305A might be assembly defective, we sought to determine whether M_L305A is defective in its interaction with N or P, or both. Our subsequent results showed that M_L305A still interacted with P and could incorporate P into M_L305A-VLPs but was unable to interact with N and incorporate N into M_L305A-VLPs (Fig. 3A and B). The finding that M_L305A interacts with P but fails to interact with N provides a rationale for assessing how the M protein assembles internal viral proteins into particles. On the basis of the result that M_L305A interacts with P and incorporates P but not N into M_L305A-VLPs, it is reasonable to speculate that one of two scenarios occurs when N, P, and M_L305A are expressed together: (i) if P mediates the N-P complex interaction with the M protein, N is also incorporated into M_L305A-VLPs, which suggests that P is a key mediator of the incorporation of the N-P complex into viral particles; (ii) in contrast, if N mediates the N-P complex interaction with the M protein, neither N nor P is incorporated into M_L305A-VLPs, which suggests that N is a key mediator of the incorporation of the N-P complex into viral particles. Our results clearly demonstrate that N inhibits the incorporation of P into M_L305A-VLPs via the formation of the N-P complex, but N_L478, a point mutant of N which lost the ability to interact with P for the formation of the N-P complex (33), has no effect on the incorporation of P into M_L305A-VLPs (Fig. 5), suggesting that N is the critical mediator of the incorporation of the N-P complex into viral particles.

The results from the aforementioned experiments prompted an interesting question: how does the N-P interaction inhibit the incorporation of P into M_L305A-VLPs? We used mutational analysis to identify regions in P that are relevant for the incorporation of P into M-VLPs and the P-M interaction. Our data first showed that the PΔC100 mutant abrogates the incorporation of P into M-VLPs (Fig. 6A). Then, we demonstrated that the expression of GST-PΔN500 is efficient for both the interaction between GST-PΔN500 and M and the recruitment of PΔN500 into M-VLPs (Fig. 6B and C). Interestingly, PΔN500 is also required and sufficient for the interaction of P with N (Fig. 6E). N inhibited the incorporation of not only P but also GST-GST-PΔN500 into M_L305A-VLPs (Fig. 7A). One possibility is that N binds to PΔN500 with a higher infinity and competitively prevents M from binding to PΔN500, but we still cannot exclude the possibility that the binding of N to PΔN500 results in structural changes in P or PΔN500 so that it can no longer be recognized by the M protein. It seems impossible that the formation of the N-P complex would result in structural changes in N since the N-P complex can be incorporated into M-VLPs (Fig. 5, right, lane 4).

In general, the finding that N regulates N-P complex assembly via direct interaction with a single amino acid residue in the M protein is remarkable for the study of the enveloped virion assembly process. Despite advancements in our understanding of viral particle assembly, the factors that drive the incorporation of viral RNPs into virions are still incompletely defined, and N has been shown to be indirectly involved in the virion assembly process for several enveloped viruses. For example, Ebola virus VLP formation is driven by VP40 but is significantly enhanced by the presence of N (21), and the N of Marburg virus was found to increase VP40-induced VLP release by recruiting Tsg101 via its PSAP motif (20), possibly to compensate for the absence of a PT/SAP motif in VP40 itself. In our study, although the expression of N had no effect on the formation and release of M-VLPs (data not shown), it could be incorporated into M-VLPs and mediate virion assembly.

Although we found that the interaction of N and the M protein plays a critical role in N-P complex recruitment to viral particles, resulting in the incorporation of internal viral proteins into the virions, we still cannot exclude the possibility that a cellular protein(s) is also involved in this assembly process by interacting with N or M, or both. Recently, a role for the viral N, in cooperation with the M proteins and the ESCRT pathway, has been suggested for viral particle formation of Marburg virus, HIV-1, and Mopeia virus (20, 34, 35). Therefore, additional cellular factors may associate with N through residue L305 in the M protein for efficient N-P complex incorporation into virions. A more detailed analysis of the association of N and the M protein with the components of the ESCRT pathway and its associated proteins will potentially reveal the role of such cellular determinants in N-P complex assembly. In addition, the M proteins of several NSVs, including influenza virus and human RSV, have been reported to interact directly with either RNA or N (36, 37). Further study is required to determine whether the M protein also packages the N-P complex into viral particles through its interaction with viral RNA.

In addition, the domain within N that mediates the interaction of N with the M protein remains unconfirmed. Previous studies have suggested that the C terminus of N of paramyxoviruses may be involved in its binding to P and the M protein (38–40). Whether the HPIV3 C-terminal domain interacts with both P and the M protein is not clear. Since N_L478A, which fails to bind to P, can still be incorporated into M-VLPs (Fig. 5B, bottom right, lane 6), it seems that N interacts with both P and M via different regions. Further investigations into essential regions within N that regulate the N-M interaction might provide important insights into the mechanism of the incorporation of HPIV3 internal proteins into virions.

Of note, although the N-P interaction inhibits the P-M interaction and N is a critical regulator of the interaction of the N-P complex with M, the possibility that P plays an indirect role in the correct assembly of virions cannot be excluded. By using cryo-electron microscopy and tomography to visualize Ebola VLPs, which were produced with different combinations of viral proteins, Bharat et al. found that when N was coexpressed with VP40, VP40 could condense a loosely coiled helix formed by the binding of N to RNA. When N was coexpressed with VP40, VP24, and VP35, the helix was further rigidified and the helix inside the VLPs was indistinguishable from that in Ebola virus virions (41). Therefore, for HPIV3, it is also possible that the loose helix formed by the binding of N to RNA is condensed by the binding of M to N, which further requires the rigidification of the condensed coils into a tight helix by the binding of P to N. Therefore, it is reasonable to speculate that VLPs formed by N, P, and M are much more compact than those formed by N and M.

In summary, we have demonstrated that the interaction of N but not P with the M protein of HPIV3 mediates internal viral protein assembly and a critical residue, L305, in the M protein is required for the N-M interaction. Our results will strengthen the understanding of the mechanisms that drive HPIV3 internal viral protein assembly and could lead to the development of novel targeted antiviral drugs for therapeutic use against HPIV3 infection.