C-Terminal DxD-Containing Sequences within Paramyxovirus Nucleocapsid Proteins Determine Matrix Protein Compatibility and Can Direct Foreign Proteins into Budding Particles

ABSTRACT

Paramyxovirus particles are formed by a budding process coordinated by viral matrix (M) proteins. M proteins coalesce at sites underlying infected cell membranes and induce other viral components, including viral glycoproteins and viral ribonucleoprotein complexes (vRNPs), to assemble at these locations from which particles bud. M proteins interact with the nucleocapsid (NP or N) components of vRNPs, and these interactions enable production of infectious, genome-containing virions. For the paramyxoviruses parainfluenza virus 5 (PIV5) and mumps virus, M-NP interaction also contributes to efficient production of virus-like particles (VLPs) in transfected cells. A DLD sequence near the C-terminal end of PIV5 NP protein was previously found to be necessary for M-NP interaction and efficient VLP production. Here, we demonstrate that 15-residue-long, DLD-containing sequences derived from either the PIV5 or Nipah virus nucleocapsid protein C-terminal ends are sufficient to direct packaging of a foreign protein, Renilla luciferase, into budding VLPs. Mumps virus NP protein harbors DWD in place of the DLD sequence found in PIV5 NP protein, and consequently, PIV5 NP protein is incompatible with mumps virus M protein. A single amino acid change converting DLD to DWD within PIV5 NP protein induced compatibility between these proteins and allowed efficient production of mumps VLPs. Our data suggest a model in which paramyxoviruses share an overall common strategy for directing M-NP interactions but with important variations contained within DLD-like sequences that play key roles in defining M/NP protein compatibilities.

IMPORTANCE Paramyxoviruses are responsible for a wide range of diseases that affect both humans and animals. Paramyxovirus pathogens include measles virus, mumps virus, human respiratory syncytial virus, and the zoonotic paramyxoviruses Nipah virus and Hendra virus. Infectivity of paramyxovirus particles depends on matrix-nucleocapsid protein interactions which enable efficient packaging of encapsidated viral RNA genomes into budding virions. In this study, we have defined regions near the C-terminal ends of paramyxovirus nucleocapsid proteins that are important for matrix protein interaction and that are sufficient to direct a foreign protein into budding particles. These results advance our basic understanding of paramyxovirus genome packaging interactions and also have implications for the potential use of virus-like particles as protein delivery tools.

INTRODUCTION

The paramyxoviruses comprise a group of enveloped viruses that harbor nonsegmented, negative-sense RNA genomes (1). Included among the paramyxoviruses are a number of human and animal pathogens, including measles virus, mumps virus, Nipah virus, respiratory syncytial virus (RSV), and Newcastle disease virus (NDV). Paramyxovirus infections are spread via particles which bud from plasma membranes of infected cells. Formation of these particles is driven by the viral matrix (M) proteins which can self-assemble to form ordered yet flexible arrays (2, 3) that likely play key roles in generating the membrane curvature required for budding. M proteins also organize the particle assembly process by interacting with the viral glycoproteins via their cytoplasmic tails and also with the viral ribonucleoprotein (vRNP) complexes via the nucleocapsid (N or NP) proteins (reviewed in references 4 and 5). These interactions bring together and concentrate all of the viral structural components onto specific sites underlying infected cell plasma membranes, enabling infectious virions to subsequently bud from these locations.

For many paramyxoviruses, expression of M protein in the absence of any other viral components is sufficient to induce the assembly and release of virus-like particles (VLPs) from transfected cells. M proteins of Sendai virus (6, 7), measles virus (8, 9), Nipah virus (10, 11), Hendra virus (12), Newcastle disease virus (13), and human parainfluenza virus 1 (HPIV1) (14) are all capable of directing VLP production and release from transfected cells when expressed alone. In these cases, additional viral components, including the viral glycoproteins and the nucleocapsid-like structures that form upon expression of paramyxovirus N/NP proteins, can be efficiently packaged into the VLPs if they are coexpressed along with the M proteins (4). For other paramyxoviruses, including mumps virus (15) and parainfluenza virus 5 (PIV5) (16), the viral M proteins do not induce significant VLP production when expressed alone in transfected cells. In these cases, coexpression of M proteins together with viral glycoproteins and NP proteins is necessary for VLP production to occur. Such an arrangement could in theory provide a benefit to viruses by preventing the release of empty virions that lack vRNPs. Other negative-strand RNA (nsRNA) viruses, including Ebolavirus (17) and Tacaribe virus (18), for which enhancements to VLP production were observed upon coexpression of the viral nucleocapsid proteins may employ similar strategies.

Paramyxovirus N/NP proteins function to bind and encapsidate viral genomic and antigenomic RNAs, forming helical nucleocapsid structures that serve as the templates for the viral polymerase (19). Encapsidation, which is directed by the RNA-binding, N-terminal core regions of the N/NP proteins, also protects viral RNAs from RNase digestion and impairs recognition of viral RNAs by host innate immune responses (1). Flexibility between domains of the N/NP structures is thought to allow for presentation of the RNA bases and access by the viral polymerase only when needed (20–22). The C-terminal tail regions of paramyxovirus N/NP proteins are dispensable for RNA binding and instead function to direct interactions with a variety of viral and host proteins, including viral M proteins (23–25) and viral P proteins (26–28), although P protein binding and polymerase docking can in some cases be mediated instead by the N-terminal core region of N (29–31).

Interactions between matrix and nucleocapsid proteins of negative-strand RNA viruses are universally important for generation of infectious, genome-containing virus particles (32), but the details of these interactions are poorly understood. Studies with measles virus have defined a region very close to the C-terminal end of N protein that is necessary for M protein binding (24). For PIV5, the sequence DLD near the C-terminal end of NP protein is critical for its virus assembly functions (25). Mutations to DLD abolished particle formation function and disrupted M-NP interaction. Building on these observations, we showed in this study that the DLD-containing C-terminal residues of PIV5 and Nipah virus nucleocapsid proteins are sufficient to direct a foreign protein into budding PIV5 or Nipah VLPs. Moreover, we demonstrated that DLD-like sequences can act as the key determinants that define compatibilities between M and NP proteins of different paramyxoviruses. Our data suggest a model in which paramyxoviruses share an overall common strategy for directing M-NP interactions, but with important variations, controlled by DLD-like sequences, that play key roles in determining M/NP compatibilities.

MATERIALS AND METHODS

Plasmids.

The plasmids pCAGGS-PIV5 M, pCAGGS-PIV5 NP, and pCAGGS-PIV5 HN have been described before (16), as have plasmids pCAGGS-MuV M, pCAGGS-MuV NP, pCAGGS-MuV F, pCAGGS-NiV M, and pCAGGS-NiV N (15). Site-directed mutants of PIV5 and mumps virus NP genes were generated by PCR mutagenesis of the wild-type (wt) sequences, and the resulting cDNAs were subcloned into eukaryotic expression vector pCAGGS (33). cDNA corresponding to Renilla luciferase (RLuc) was obtained by PCR amplification of its coding sequence from plasmid pSMG-RLuc, which was a kind gift of Biao He, University of Georgia, Athens, GA. This sequence was modified using PCR to incorporate C-terminal sequences derived from PIV5 NP or Nipah virus N, with an additional double glycine (-GG-) linker added between the luciferase and virus-derived sequences. The resulting cDNAs were subcloned into the pCAGGS vector. pCAGGS-RLuc-NP15 was further modified by PCR mutagenesis to remove the sequence encoding the C-terminal four amino acid residues to generate pCAGGS-RLuc-NP15Δ4, and pCAGGS-RLuc-N15 was further modified by PCR mutagenesis to generate a set of alanine substitution mutants. DNA sequencing of the entire genes was carried out to verify their identities (Genomics Core Facility, Pennsylvania State University).

Antibodies.

PIV5 M and NP proteins were detected using monoclonal antibodies M-f and NP-a (34), which were both kind gifts of Richard Randall, St. Andrews University, St. Andrews, Scotland, United Kingdom. PIV5 HN protein was detected using the polyclonal antibody SDS-HN, a kind gift of Robert Lamb, Northwestern University, Evanston, IL. Mumps virus M, NP, and F proteins were detected using polyclonal anti-peptide antibodies that have been described before (15). Nipah virus M and N proteins were detected using polyclonal antibodies raised against the corresponding full-length recombinant proteins (15). Renilla luciferase was detected using a polyclonal antibody purchased from MBL International (Woburn, MA).

Membrane coflotation assays to measure M-NP protein interactions.

293T cells in 6-cm-diameter dishes, grown in Dulbecco's modified Eagle medium (DMEM) supplemented with 10% fetal bovine serum, were transfected with pCAGGS plasmids encoding M and/or NP proteins (1.2 μg and 800 ng of plasmid DNA, respectively). Lipofectamine-Plus reagents in Opti-MEM were used for transfection (Invitrogen, Carlsbad, CA). Cells were harvested at 24 h posttransfection (p.t.) and resuspended in a hypotonic solution (25 mM NaCl; 50 mM Na₂HPO₄ [pH 7.4]; 1 mM phenylmethyl-sulfonyl fluoride). The cell suspension was incubated for 20 min with rocking at 4°C. Cells were then disrupted by passaging them through a 23-gauge needle 20 times, and these cellular extracts were centrifuged at 1,500 × g for 5 min at 4°C to remove nuclei and cell debris. The resulting supernatants were mixed with 1.5 ml of 80% sucrose–NTE (0.1 M NaCl; 0.01 M Tris-HCl [pH 7.4]; 0.001 M EDTA). A sucrose gradient was then formed by overlaying these samples with 50% sucrose (2.4 ml) and 10% sucrose (0.6 ml) solutions in NTE. Samples were subjected to ultracentrifugation at 160,000 × g for 4 h in a Beckman SW55Ti rotor. Four fractions were collected from the top of each gradient, and 2% of each fraction was loaded on 10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) gels and resolved. Proteins were subjected to immunoblotting with antibodies specific to the viral M and/or NP proteins. Protein bands were detected and quantified using a Fuji FLA-7000 laser scanner (FujiFilm Medical Systems, Stamford, CT). The amount of NP protein in membrane fractions was quantified by measuring the NP protein in the top two fractions and dividing by the sum of the NP protein in all four fractions.

Measurements of VLP production.

293T cells in 6-cm-diameter dishes at about 70% to 80% confluence were grown in DMEM supplemented with 10% fetal bovine serum and were transfected with pCAGGS plasmids encoding PIV5, mumps virus, or Nipah virus proteins for production of PIV5, mumps virus, or Nipah virus-like particles (VLPs). Lipofectamine-Plus reagents in Opti-MEM were used for plasmid transfection. Plasmid quantities per dish were as follows: pCAGGS-PIV5 M, 400 ng; pCAGGS-MuV M, 400 ng; pCAGGS-NiV M, 400 ng; pCAGGS-PIV5 NP and derivatives, 100 ng; pCAGGS-MuV NP and derivatives, 100 ng; pCAGGS-NiV N, 100 ng; pCAGGS-RLuc, 100 ng; pCAGGS-RLuc-PIV5 NP/NiV N fusions, 100 ng; pCAGGS-PIV5 HN, 1.5 μg; pCAGGS-MuV F, 100 ng. To keep total plasmid amounts equal during transfection, an empty pCAGGS plasmid that does not encode any viral protein was included as necessary.

At 24 h p.t., the culture medium was replaced with DMEM containing 2% fetal bovine serum or, for metabolic labeling experiments, with DMEM containing one-tenth the normal amount of cysteine and methionine, along with 37 μCi of [³⁵S] Promix/ml (PerkinElmer, Waltham, MA). After an additional 16 to 18 h, cells and media were harvested. First, culture media were centrifuged at 8,000 × g for 2 min to remove cell debris. The supernatants were then layered onto 20% sucrose cushions (4 ml in NTE). Samples were centrifuged at 140,000 × g for 1.5 h, after which pellets containing VLPs were resuspended in 0.9 ml of 1× phosphate-buffered saline (PBS) (0.13 M NaCl; 2.6 mM KCl; 1.4 mM KH₂PO₄; 8.0 mM Na₂HPO₄·7H₂O; pH 7.4) and mixed with 2.4 ml of 80% sucrose–NTE. Layers of 50% sucrose–NTE (3.6 ml) and 10% sucrose–NTE (0.6 ml) were applied to the tops of the gradients, and these were then centrifuged at 140,000 × g for 3 h. A 4-ml volume was collected from the top of each gradient, and the VLPs contained in this fraction were pelleted by centrifugation at 190,000 × g for 1.5 h. VLP pellets were then resuspended in SDS-PAGE loading buffer containing 2.5% (wt/vol) dithiothreitol.

To prepare cell lysates, one-third of the cells from each sample were lysed with 0.1 ml of SDS-PAGE loading buffer. The lysates were centrifuged through QIAshredder homogenizers (Qiagen, Germantown, MD) to break up cell debris. Cell lysates and purified VLPs were fractionated by SDS-PAGE using 10% gels, and proteins were detected by immunoblotting using antibodies specific to the viral proteins and/or Renilla luciferase. Imaging and quantification were performed using a Fuji FLA-7000 laser scanner. PIV5 and mumps VLP production was measured by calculating the amount of PIV5 M protein or mumps virus M protein in VLPs, normalized to the amount of M protein present in cell lysates. Luciferase protein incorporation into Nipah VLPs was calculated as the amount of luciferase protein in particles divided by the amount of M protein in particles.

Luciferase activity measurements.

PIV5 VLPs were generated as described above. VLP pellets were suspended in 100 μl of passive lysis buffer (Promega, Madison, WI). One-third of the cells were also lysed in 100 μl passive lysis buffer. A 5-μl volume of cell lysate and 20 μl of lysed VLPs were subjected to Renilla luciferase assay, per the instructions of the manufacturer. Renilla luciferase activity was measured using a Veritas microplate luminometer (Turner BioSystems, Sunnyvale, CA). Levels of enzymatically active Renilla luciferase incorporated into VLPs were determined by calculating the luciferase activity in VLPs divided by the luciferase activity in the corresponding cell lysate fraction, normalized to the value obtained with RLuc-NP30.

Amino acid sequence comparisons.

To compare C termini of paramyxovirus nucleoprotein sequences, data were derived from GenBank files with the following accession numbers: PIV5, GenBank accession no. AF052755; mumps virus, JN012242; Nipah virus, AF212302; human parainfluenza virus type 2 (HPIV2), M55320; measles virus, AB016162; Sendai virus, M30202; Newcastle disease virus (NDV), AF064091; Hendra virus, AAC83187; human respiratory syncytial virus (HRSV), AEO45904.

RESULTS

Manipulation of genome packaging interactions to direct a foreign protein into PIV5 VLPs.

We recently identified a DLD sequence near the C-terminal end of PIV5 NP protein that is critical for M binding and for efficient VLP production (25) (Fig. 1A). To further investigate the role that these residues play in virus assembly, we transplanted segments from the C-terminal end of PIV5 NP onto the C-terminal end of Renilla luciferase (RLuc), as illustrated in Fig. 1A. The luciferase proteins were expressed together with PIV5 M and HN proteins in 293T cells for VLP production. The unmodified RLuc reporter protein completely lacked VLP assembly functions, as expression of RLuc together with PIV5 M and HN proteins led to poor VLP production, similar to that observed when M and HN proteins were expressed alone (Fig. 1B, lanes 1 and 3). In contrast, expression of M and HN proteins together with PIV5 NP protein led to highly efficient VLP production (Fig. 1B, lane 2), consistent with earlier findings (25). VLP production was quantified based on the amount of viral M protein detected in sucrose gradient-purified VLPs, normalized to the amount of M detected in cell lysate fractions (Fig. 1C). Fusion of either 5 residues or 10 residues from NP to the C-terminal end of RLuc had little impact on VLP production. However, when 15 or more residues were appended to RLuc, the modified RLuc gained the ability to stimulate VLP production (Fig. 1B and C). RLuc-NP15 expression led to VLP production that was about 60% of that observed with the authentic viral NP protein, and expression of each of RLuc-NP30 and RLuc-NP50 led to VLP production that was roughly equivalent to that observed with NP. Moreover, substantial quantities of modified RLuc were found within the purified VLP preparations (Fig. 1B). To more directly assess the incorporation efficiency, VLPs were produced in cells that were metabolically labeled with ³⁵S amino acids. VLPs were purified and loaded directly on SDS gels, and proteins were detected using a phosphorimager for visualization of VLP polypeptide composition (Fig. 1D). The result indicated that RLuc-NP30 was abundantly packaged into VLPs (1.1:1 ratio of RLuc-NP30/M, taking into account the numbers of methionine and cysteine residues present in the respective proteins).

FIG 1.

A total of 15 C-terminal residues of PIV5 NP protein are sufficient to trigger VLP production and to direct a foreign protein into budding particles. (A) (Top panel) C-terminal amino acid sequences of three paramyxovirus NP/N proteins. The DLD sequence critical for virus assembly functions of PIV5 NP protein is highlighted in bold. (Middle panel) Schematic illustration of Renilla luciferase, appended with residues derived from a paramyxovirus NP protein, being packaged into budding VLPs using the same interactions that would normally direct the packaging of vRNPs into virions. (Bottom panel) Illustration of Renilla luciferase proteins appended with residues derived from PIV5 NP protein. In the cases of RLuc-NP15 and RLuc-NP15Δ4, the NP-derived amino acid sequences are shown in full. (B) 293T cells were transfected to produce PIV5 M and HN proteins together with the indicated luciferase-NP fusions. Viral proteins from cell lysates and from sucrose gradient-purified VLPs were detected by immunoblotting. (C) Relative levels of efficiency of VLP production were calculated as the amount of M protein detected in VLPs divided by the amount of M protein detected in cell lysates, normalized to the value obtained with NP protein. Error bars indicate standard deviations (n = 3). **, P < 0.005. (D) 293T cells were transfected to produce PIV5 M and HN proteins together with either unmodified luciferase or RLuc-NP30. Cells were metabolically labeled with ³⁵S, and sucrose-gradient purified VLPs were loaded directly onto SDS gels. VLP-derived proteins were detected using a phosphorimager. (E) VLPs were generated and purified as described for Fig. 1B. Enzymatically active luciferase contained within VLP and cell lysate fractions was measured using a luminometer. Values were calculated as luciferase activity of VLPs divided by luciferase activity of cell lysate, normalized to the value obtained with RLuc-NP30 (n = 2). (F) 293T cells were transfected to produce PIV5 M and HN proteins together with the indicated RLuc-NP fusions. Viral proteins from cell lysates and from sucrose gradient-purified VLPs were detected by immunoblotting.

Additional experiments were carried out to test if the modified RLuc retains its enzymatic activity even after it has been packaged into VLPs. Cells were transfected to produce VLPs as before, and the amounts of enzymatically active luciferase in the VLP and cell lysate fractions were measured using a luminometer (Fig. 1E). Consistent with the results obtained by Western blot detection shown in Fig. 1B and C, we found that expression of RLuc-NP15 or RLuc-NP30 (but not unmodified RLuc) together with the viral M and HN proteins led to substantial release of VLPs containing enzymatically active luciferase (Fig. 1E). Luminometer-based VLP quantification revealed a difference of more than 15-fold between VLP production in the presence of RLuc-NP30 and VLP production in the presence of unmodified RLuc. In addition to providing an alternative method of VLP quantification, these results also suggest that biologically active foreign proteins can be made to efficiently incorporate into PIV5 VLPs through the addition of short sequences to their C-terminal ends.

To confirm that incorporation of modified RLuc into PIV5 VLPs is dependent on the DLD residues near the C-terminal end of the NP-derived transplanted sequence, RLuc-NP15 was further modified to create RLuc-NP15Δ4, in which the C-terminal 4 residues (including DLD) have been removed (Fig. 1A). In contrast to RLuc-NP15, RLuc-NP15Δ4 failed to stimulate VLP production and was poorly incorporated into the few VLPs that were made, similarly to unmodified RLuc (Fig. 1F). Hence, the DLD sequence near the C-terminal end of PIV5 NP is critical for VLP assembly functions, both in its natural context (25) and in the context of a foreign RLuc reporter protein.

PIV5 NP protein and mumps virus M protein can be engineered for compatibility.

In contrast to the PIV5 nucleocapsid protein, which harbors a DLD sequence near its C-terminal end, the NP protein of mumps virus lacks DLD near its C-terminal end (Fig. 2). This raised the possibility that PIV5 and mumps virus might be incompatible with respect to M-NP interactions, despite being very closely related viruses overall (both are within the Rubulavirus genus of the paramyxoviruses). To test this, mumps virus M, F, and NP proteins were expressed together in 293T cells for VLP production. Consistent with earlier results (15), VLP production was efficient when all three of these proteins were coexpressed but was poor (more than 10-fold reduced) when NP expression was omitted (Fig. 3A and B). When PIV5 NP protein was expressed together with the mumps virus M and F proteins, VLP production was markedly impaired (Fig. 3A and B). This result suggests that the PIV5 NP protein is incompatible with the mumps virus M protein for particle assembly. To further explore the parameters that govern compatibilities between these proteins, a series of chimeric NP proteins was generated as illustrated in Fig. 3C. These proteins are based on PIV5 NP protein, but the C-terminal ends were progressively replaced with sequence derived from the mumps virus NP protein. None of these alterations had significant effects on NP protein expression levels (Fig. 3D). The chimeric NP proteins were expressed together with mumps virus M and F proteins, and VLP production was measured (Fig. 3D and E). Chimera M1, which affects just a single amino acid residue at the very C-terminal end of NP protein, did not restore mumps VLP production. However, chimeras M3, M6, M7, M8, and M9 all restored mumps VLP production to levels similar to that obtained with authentic mumps virus NP protein (Fig. 3D and E). The M1 and M3 chimeras are markedly different in terms of mumps VLP production function (compare lanes 4 and 6) and yet differ from one another by only a single amino acid residue substitution. This change affects the DLD sequence of PIV5 NP, converting it to the DWD sequence that is found at the corresponding position in mumps virus NP. To determine if this single amino acid residue change is sufficient to induce M/NP compatibility, PIV5 NP L507W was generated by site-directed mutagenesis. This protein was fully functional for mumps VLP production (Fig. 3D, lane 5, and E). Hence, a single amino acid change converting DLD to DWD induced PIV5 NP protein to gain compatibility with mumps virus proteins for VLP production.

FIG 2.

C-terminal ends of paramyxovirus N/NP proteins. The upper portion includes sequences derived from paramyxoviruses within the Rubulavirus genus, while the lower portion includes sequences derived from paramyxoviruses outside the Rubulavirus genus. DLD-like sequences are highlighted in bold. MuV, mumps virus; NiV, Nipah virus; SeV, Sendai virus; MeV, measles virus.

FIG 3.

DLD and DWD sequences define compatibilities between PIV5 and mumps virus M/NP protein pairs. (A) 293T cells were transfected to produce mumps virus M and F proteins together with the indicated NP proteins for mumps VLP production. Viral proteins from cell lysates and from sucrose gradient-purified VLPs were detected by immunoblotting. (B) Relative levels of efficiency of VLP production calculated as described for Fig. 1C. Error bars indicate standard deviations (n = 3). (C) Schematic illustrating chimeric NP proteins. Substitutions shown in bold progressively convert the C-terminal end of PIV5 NP to match the sequence of mumps virus NP. (D) 293T cells were transfected to produce mumps virus M and F proteins together with the indicated NP protein chimeras for mumps VLP production. Viral proteins from cell lysates and from sucrose gradient-purified VLPs were detected by immunoblotting. (E) Relative levels of efficiency of VLP production calculated as described for Fig. 1C. Error bars indicate standard deviations (n = 3).

Compatibilities between PIV5 and mumps virus M/NP proteins were further analyzed using a membrane coflotation assay. This assay is based on the observation that paramyxovirus M proteins intrinsically bind to cellular membranes, whereas NP proteins do not. M-NP protein interaction indirectly recruits NP protein to membranes, allowing coflotation of NP with membranes on sucrose gradients. This assay has been used previously to monitor M-NP protein interactions of measles virus (9) and PIV5 (25). As anticipated based on these earlier studies, we found that mumps virus NP protein expressed alone did not float with membranes, whereas a significant fraction of M protein expressed alone floated with membranes (Fig. 4). Coexpression of M with NP induced significant coflotation of NP (about 47% recovered in membrane-bound fractions). However, coexpression of mumps virus M with PIV5 NP resulted in a substantially lower level of NP coflotation (about 27% recovered in membrane-bound fractions), consistent with the incompatibility that was observed between these proteins in VLP production experiments. The L507W single amino acid change significantly improved compatibility with mumps virus M protein (about 42% recovered in membrane-bound fractions). Together, these results indicate that PIV5 NP protein is incompatible with mumps virus M protein both for binding and for VLP production and that compatibility can be induced through a single amino acid change that converts the C-terminal DLD sequence to DWD.

FIG 4.

Compatibilities of PIV5 and mumps virus M/NP protein pairs measured by membrane coflotation analysis. (A) The indicated pairs of M and NP proteins were expressed in 293T cells, and detergent-free lysates were fractionated on sucrose flotation gradients. Viral proteins were visualized by immunoblotting. (B) The fraction of membrane-associated NP was calculated as the amount of NP in the top two fractions divided by the total NP protein signal in all four fractions. Error bars indicate standard deviations (n = 3). **, P < 0.005.

We next considered the possibility that the L507W change to PIV5 NP, while creating new compatibility with mumps virus M, might at the same time disrupt the existing compatibility with its natural binding partner, PIV5 M. To test this, PIV5 NP L507W was analyzed for PIV5 M interaction using coflotation (Fig. 5A and B) and for the ability to induce production of PIV5 VLPs (Fig. 5C and D). The results indicate that PIV5 NP L507W functions just as well as wt PIV5 NP, both for PIV5 M interaction and for PIV5 VLP production. Hence, this altered NP protein gained compatibility with mumps virus M without losing its original compatibility with PIV5 M. We also tested if the L507W change altered compatibility with the Nipah virus M protein. Nipah virus M protein is compatible with the wt PIV5 NP protein both for binding (Fig. 5A and B) and for incorporation into budding VLPs (Fig. 5E). The L507W change to PIV5 NP protein did not impair this compatibility (Fig. 5A, B, and E). These results indicate that the DWD-containing PIV5 L507W protein is compatible with the M proteins of at least three different paramyxoviruses—PIV5, mumps virus, and Nipah virus.

FIG 5.

PIV5 NP L507W protein retains its compatibilities with PIV5 and Nipah virus M proteins. (A) The indicated pairs of M and NP proteins were expressed in 293T cells, and detergent-free lysates were fractionated on sucrose flotation gradients. Viral proteins were visualized by immunoblotting. (B) The fraction of membrane-associated NP was calculated as the amount of NP in the top two fractions divided by the total NP protein signal in all four fractions. Error bars indicate standard deviations (n = 3). (C) 293T cells were transfected to produce PIV5 M and HN proteins together with either PIV5 NP or PIV5 NP L507W. Viral proteins from cell lysates and from sucrose gradient-purified VLPs were detected by immunoblotting. (D) Relative levels of efficiency of VLP production calculated as described for Fig. 1C. Error bars indicate standard deviations (n = 3). (E) 293T cells were transfected to produce Nipah virus M protein together with either PIV5 NP or PIV5 NP L507W. Viral proteins from cell lysates and from sucrose gradient-purified VLPs were detected by immunoblotting.

Alterations affecting the C-terminal end of mumps virus NP protein impair interaction with M protein and block VLP production.

The findings on M/NP compatibility described above suggested a critical role for the DWD residues near the C-terminal end of mumps virus NP protein. To directly test the importance of these residues within the context of full-length mumps virus NP protein, site-directed mutagenesis of mumps virus NP protein was carried out as illustrated in Fig. 6A. We found that mumps VLP production was severely impaired when residue D546 or residue W547 was changed to alanine (Fig. 6B and C). A similar negative effect on VLP production was also observed for NPΔ4, in which the four C-terminal residues (including D546 and W547) have been removed (Fig. 6B and C). Residue W547 was also changed to leucine, creating a DLD sequence in place of DWD. This severely impaired VLP production, reinforcing the observation that DLD sequences are incompatible with mumps virus M protein (Fig. 6B, lane 5, and C). In each of these cases, the low level of VLP production was similar to that of the negative control (absence of NP expression), indicating a nearly complete loss of VLP production function. On the other hand, alanine substitutions that targeted the adjacent D548 and E549 residues resulted in no significant impairments to VLP production (Fig. 6B and C). The altered mumps virus NP proteins were also tested for the ability to assemble together with mumps virus M protein in membrane coflotation experiments (Fig. 6D and E). We found that the same mutations that disrupted VLP production also impaired membrane coflotation with M protein. Together, these results indicate that virus assembly functions of mumps virus NP protein are critically dependent on residues D546 and W547 within its C-terminal DWD motif.

FIG 6.

The DWD-containing C-terminal region of mumps virus NP protein is important for its VLP production function. (A) Schematic illustrating mutations made to mumps virus NP protein. (B) 293T cells were transfected to produce mumps virus M and F proteins together with the indicated NP protein mutants for VLP production. Viral proteins from cell lysates and from sucrose gradient-purified VLPs were detected by immunoblotting. (C) Relative levels of efficiency of VLP production calculated as described for Fig. 1C. Error bars indicate standard deviations (n = 3). (D) The indicated pairs of M and NP proteins were expressed in 293T cells, and detergent-free lysates were fractionated on sucrose flotation gradients. Viral proteins were visualized by immunoblotting. (E) The fraction of membrane-associated NP was calculated as the amount of NP in the top two fractions divided by the total NP protein signal in all four fractions. Error bars indicate standard deviations (n = 3).

Manipulation of genome packaging interactions to direct a foreign protein into Nipah VLPs.

Sequences within the Nipah virus N protein that function to coordinate with M protein during virus assembly have not been defined. Here, we hypothesized on the basis of analogy with PIV5 that such sequences might be located near the C-terminal end of N protein. To test this hypothesis, we transplanted segments from the C-terminal end of Nipah virus N onto the C-terminal end of RLuc. The luciferase proteins were expressed together with Nipah virus M protein, and incorporation of RLuc into the budding VLPs was measured (Fig. 7). Note that in the case of Nipah virus, VLP production does not depend on expression of N protein or glycoproteins—VLPs are produced efficiently upon expression of Nipah virus M protein alone (10, 11). However, N protein was incorporated into the M-containing VLPs when it was coexpressed (11) (Fig. 7A, lane 2). We found that unmodified RLuc was incorporated poorly into M-VLPs (Fig. 7A, lane 3). Fusion of 5 residues derived from N to the C-terminal end of RLuc did not improve its incorporation into VLPs. However, when 10 residues were appended to RLuc, incorporation into VLPs improved, and when 15 or 30 residues were appended to RLuc, VLP incorporation was improved still further, to a level that was approximately 2.5 times greater than that observed with the unmodified RLuc control (Fig. 7A and B). RLuc-N50 was incorporated somewhat less efficiently into VLPs than RLuc-N15 or RLuc-N30, but its incorporation was still approximately 1.5-fold higher than that observed with the unmodified RLuc control. These findings with respect to Nipah virus parallel those shown in Fig. 1 with PIV5, even though these two viruses come from separate genera within the Paramyxoviridae and have M proteins with relatively poor homology (less than 25% amino acid identity). Overall, our results support the general idea that foreign proteins can be engineered for packaging into budding paramyxovirus VLPs through addition of small (10-to-15-residue) appendages to their C-terminal ends.

FIG 7.

A total of 15 C-terminal residues of Nipah virus N protein are sufficient to direct a foreign protein into budding particles. (A) 293T cells were transfected to produce Nipah virus M protein together with the indicated Renilla luciferase-Nipah N fusions. Viral proteins from cell lysates and from sucrose gradient-purified VLPs were detected by immunoblotting. (B) Relative efficiencies of luciferase incorporation into VLPs were calculated as the amount of luciferase detected in VLPs divided by the amount of M protein detected in VLPs, normalized to the value obtained with RLuc-N15. Error bars indicate standard deviations (n = 3). **, P < 0.001.

Amino acid residues 523 to 528 of N protein are important for RLuc packaging into Nipah VLPs.

The C-terminal 15 residues of Nipah virus N protein were targeted by alanine-scanning mutagenesis to determine which of these residues are necessary for efficient direction of the RLuc reporter into Nipah VLPs (Fig. 8A). The altered RLuc proteins were coexpressed together with Nipah virus M protein, and RLuc incorporation into VLPs was measured. Substitutions at any of the positions from residue 523 to residue 528 caused a significant reduction in luciferase incorporation (Fig. 8B and C). The most severe defects were associated with changes at positions N523 and D524. In those cases, luciferase incorporation into VLPs was similar to that observed with the unmodified luciferase control (Fig. 8B and C). Alanine substitutions targeting the surrounding residues, outside the sequence 523-NDLDFV-528, had little impact on luciferase incorporation into VLPs.

FIG 8.

Amino acid residues 523-NDLDFV-528 within Nipah virus N protein are important for the ability to direct a foreign protein into Nipah VLPs. (A) Schematic illustrating variations on RLuc-N15, generated by site-directed mutagenesis. Only the 15 residues derived from Nipah virus N protein are shown. (B) 293T cells were transfected to produce Nipah virus M protein together with the indicated variants of RLuc-N15. Viral proteins from cell lysates and from sucrose gradient-purified VLPs were detected by immunoblotting. The asterisk denotes the position of RLuc-N15 variants that could be detected using a polyclonal antibody raised against Nipah virus N protein. WB, Western blot. (C) Relative efficiencies of luciferase incorporation into VLPs were calculated as the amount of luciferase detected in VLPs divided by the amount of M protein detected in VLPs, normalized to the value obtained with RLuc-N15. Error bars indicate standard deviations (n = 3). *, P < 0.05; **, P < 0.005.

During the course of these experiments, we noticed that RLuc-N15, RLuc-N30, and RLuc-N50 (but not RLuc-N5 or RLuc-N10) could in some cases be detected on immunoblots using an N-specific polyclonal antibody (15) that had been raised against purified N protein expressed in Escherichia coli (Fig. 8B, marked with asterisks, and data not shown). Detection was impaired or lost when any of residues N523, D524, L525, or D526 was changed to alanine (Fig. 8B, top panels). These are the same residues that were the most critical for RLuc incorporation into VLPs. Our findings suggest that these residues, in addition to directing VLP incorporation, also define an epitope that is recognized by a polyclonal antibody raised to full-length N protein.

DISCUSSION

During assembly of negative-strand RNA viruses, viral genomes, in the form of vRNPs, are packaged into budding particles via M-NP interactions (32, 35). Hence, these interactions are fundamentally important for the production of genome-containing, infectious viral particles. Here, we have defined regions near the C-terminal ends of paramyxovirus NP proteins that direct their virus assembly functions. A 15-residue DLD-containing sequence derived from the C-terminal end of PIV5 NP protein was capable of directing a foreign protein (Renilla luciferase) into PIV5 VLPs. Likewise, a 15-residue DLD-containing sequence derived from the C-terminal end of Nipah virus N protein was sufficient to direct Renilla luciferase into Nipah VLPs. Other paramyxoviruses harbor similar DLD-like sequences near their C-terminal ends as well (illustrated in Fig. 2). For example, the HPIV2 NP protein contains DFD specifically in place of the DLD found in PIV5. NDV and HRSV N proteins have DND sequences near the C-terminal ends. Measles virus N protein contains the sequence DRDLLD at the C-terminal end, and alterations that affect this sequence have been found to disrupt virus assembly functions (24). The C-terminal portion of Sendai virus NP protein has also been implicated in virus assembly (23). The mumps virus NP protein harbors DWD in place of the DLD sequence found in PIV5 NP. We found this DWD sequence to be critical for efficient mumps VLP production. Interestingly, DWD and DLD sequences did not function equivalently for VLP production but were instead the key determinants that defined compatibilities between PIV5 and mumps virus M/NP protein pairs. Mumps VLP production was efficient only in the presence of DWD-containing NP proteins, such as wt mumps NP protein or PIV5 NP L507W that was engineered to contain DWD in place of DLD. Mumps VLP production was poor in the presence of DLD-containing NP proteins, such as wt PIV5 NP protein and mumps NP W547L that was engineered to contain DLD in place of DWD. Based on these collective findings, we propose that paramyxoviruses share an overall common strategy for directing M-NP interactions, but with important variations, controlled by DLD-like sequences, that play key roles in defining M/NP compatibilities.

For a subset of the nsRNA viruses, including PIV5 (16), mumps virus (15), Ebolavirus (17), and Tacaribe virus (18), matrix-nucleocapsid protein interactions appear to function not only as a means of recruiting vRNPs into particles but also as a signal to trigger particle release itself. The mechanism(s) by which nucleocapsid proteins enhance particle production in these cases is not clear. It is possible that a nucleocapsid requirement for the completion of particle assembly could benefit these viruses by minimizing the release of noninfectious, empty virions that lack viral genomes. Here, we found that luciferase proteins appended with short NP-derived, DLD-containing sequences completely replaced the requirement for NP protein in PIV5 VLP production. Hence, we believe that the mechanism for enhanced PIV5 particle release in the presence of NP protein has nothing to do with assembly of M protein onto encapsidated RNAs, as the appended sequences are all derived from the C-terminal tail region of NP and not from the N-terminal core which is responsible for RNA binding and encapsidation (20). It is possible that NP stimulates PIV5 particle production through direct occupation of a binding pocket on M protein by the DLD-containing sequence, which could then induce M protein to enter a budding-active state. An alternative possibility is that one or more host factors, essential for particle budding, could be recruited to PIV5 assembly sites via NP protein.

We found that a luciferase reporter protein could be induced to package into Nipah VLPs if it was appended with a 10-to-15-residue sequence derived from the C-terminal end of Nipah virus N protein. Here, the sequence 523-NDLDFV-528 within this region was critical for directing luciferase into the budding Nipah VLPs. However, mutagenesis of these residues in the context of full-length N protein did not substantially impair N protein incorporation into VLPs (data not shown). This is in contrast to the DLD-containing sequence near the C-terminal end of PIV5 NP protein, which not only was sufficient to direct a luciferase reporter into VLPs (Fig. 1C) but also was necessary for VLP assembly in the context of full-length NP protein (25). Likewise, the mumps DWD sequence was critical for mumps VLP release in the context of the full-length mumps virus NP protein (Fig. 6). It is possible that multiple, redundant interactions drive incorporation of Nipah virus N protein into budding particles, with one interaction based on 523-NDLDFV-528 near the C-terminal end and one or more additional interactions directed by sequences located elsewhere within N protein.

The ability to manipulate viral M-NP protein interactions could prove useful for future development of VLP-based protein delivery tools. In this scenario, foreign proteins of interest would be tagged to induce their interaction with M protein and subsequent incorporation into fusion-competent VLPs, which would then deliver the contents to target cells. Although incorporation of target proteins into VLPs has been demonstrated in the past, the approaches used typically require direct fusion of the target protein amino acid sequence to the viral Gag or M protein that directs particle budding (36–38). Here, we have instead achieved efficient incorporation of an enzymatically active foreign protein into paramyxovirus VLPs by harnessing the M-NP protein interactions that normally direct viral RNPs into budding virions. This approach is highly flexible, as it requires no modification at all to the viral matrix protein component, and the target protein in this case was modified only through the addition of a 15-amino-acid NP-derived binding sequence to its C-terminal end. Further modification of this approach to include the paramyxovirus fusion and attachment glycoproteins would in theory result in particles capable of transmitting the foreign proteins to target cells that would be similar to the “infectious” paramyxovirus VLPs that have been studied in the past and that are capable of delivering their NP-encapsidated minigenome cargos (39–43). Such VLP-based delivery vehicles could provide a highly flexible and safe platform for therapeutic delivery of functional proteins or toxins to cells.