Tetramerization of Phosphoprotein Is Essential for Respiratory Syncytial Virus Budding while Its N-Terminal Region Mediates Direct Interactions with the Matrix Protein

Human respiratory syncytial virus (RSV) is the leading cause of infantile bronchiolitis in the developed world and of childhood deaths in resource-poor settings. It is a major unmet target for vaccines and antiviral drugs.

ABSTRACT

It was shown previously that the matrix (M), phosphoprotein (P), and fusion (F) proteins of respiratory syncytial virus (RSV) are sufficient to produce virus-like particles (VLPs) that resemble the RSV infection-induced virions. However, the exact mechanism and interactions among the three proteins are not known. This work examines the interaction between P and M during RSV assembly and budding. We show that M interacts with P in the absence of other viral proteins in cells by using a split Nano luciferase assay. By using recombinant proteins, we demonstrate a direct interaction between M and P. By using nuclear magnetic resonance (NMR), we identify three novel M interaction sites on P, namely, site I in the α_N2 region, site II in the 115 to 125 region, and the oligomerization domain (OD). We show that the OD, and likely the tetrameric structural organization of P, is required for virus-like filament formation and VLP release. Although sites I and II are not required for VLP formation, they appear to modulate P levels in RSV VLPs.

IMPORTANCE Human respiratory syncytial virus (RSV) is the leading cause of infantile bronchiolitis in the developed world and of childhood deaths in resource-poor settings. It is a major unmet target for vaccines and antiviral drugs. The lack of knowledge of the RSV budding mechanism presents a continuing challenge for virus-like particle (VLP) production for vaccine purposes. We show that direct interaction between P and M modulates RSV VLP budding. This further emphasizes P as a central regulator of the RSV life cycle, as an essential actor for transcription and replication early during infection and as a mediator for assembly and budding in the later stages for virus production.

INTRODUCTION

Human respiratory syncytial virus (RSV) is the most frequent cause of infantile bronchiolitis and pneumonia worldwide (1). In France, 460,000 infants are infected each year, of which, ∼30% develop lower respiratory infections and 4.8% to 6.7% are hospitalized, representing 45% of the young children admitted to the hospital (2). The enormous burden of RSV makes it a major unmet target for a vaccine and antiviral drug therapy. However, despite >60 years of research since its discovery, there is still no vaccine available, and RSV therapy remains mainly supportive. The current standard of care consists of prophylactic treatment of at-risk infants with palivizumab (Synagis), a monoclonal antibody. However, its limited efficacy (approximately 50%), and high cost (€5,000 per treatment) limits its use to preterm infants. As a result, 60% of at-risk children remain untreated, and no efficient therapy is available to treat the adult population (3). There are currently 39 vaccines under development (4). One of the strategies for RSV vaccine development is based on virus-like particles (VLPs). However, all anti-RSV VLP vaccines currently in preclinical development are using foreign viral systems incorporating the RSV glycoproteins (5, 6). This is mostly due to the inefficiency of RSV VLP production (most of the virus is cell associated in cell culture [7]) and to insufficient understanding of RSV particle assembly and budding. RSV VLPs, if they can be produced at sufficient levels, will accurately mimic the viral morphology and structure. Identification of the minimal players involved in particle assembly and budding is an important step in understanding the mechanism behind RSV particle formation. The knowledge can be then used for large-scale VLP and attenuated virus production for vaccination purposes.

RSV belongs to the Pneumoviridae family in the order Mononegavirales (8). It primarily infects epithelial cells of the respiratory tract and replicates in the cytoplasm. It is an enveloped, nonsegmented, negative-strand RNA virus. The viral genome is encapsidated by the nucleoprotein (N), forming a ribonucleoprotein (RNP) complex, which constitutes the template for the viral polymerase. It was recently shown that the replication and transcription steps of RSV take place in virus-induced cytoplasmic inclusions called inclusion bodies (IBs), where all the proteins of the polymerase complex, i.e., the viral polymerase (L), its main cofactor the P protein, the RNPs, and the transcription factor M2-1, concentrate (9). It is noteworthy that pseudo-IBs, similar to those observed in RSV-infected cells, can be observed upon coexpression of only N and P (10, 11). We recently showed that the formation of these pseudo-IBs depends on a liquid-liquid phase separation induced by the N-P interaction (11). Once neosynthesized, RNPs have to be exported from IBs to the plasma membrane, where RSV virions assemble and bud, forming elongated membrane filaments (12). According to the common paradigm, RSV assembles on the plasma membrane, and infectious viral particles are mainly filamentous (13, 14). However, recent data suggest that viral filaments are produced and loaded with genomic RNA prior to insertion into the plasma membrane. According to this model, vesicles with RSV glycoproteins recycle from the plasma membrane and merge with intracellular vesicles, called assembly granules, containing the RNPs (15, 16).

Regardless of the cellular location, the minimal RSV viral proteins required for efficient filament formation and budding of VLPs are P, M, and the F protein, more specifically, its cytoplasmic tail (FCT) (17, 18). The atomic structure of the external part of F glycoprotein (excluding the transmembrane and cytoplasmic parts) has been resolved (19–21), but little is known about the FCT structure and its function in RSV assembly. M, a key structural protein, directs assembly and budding, probably by interacting with FCT on the one hand and with P associated with RNP on the other hand (22–24). M is required for filament elongation and maturation and, possibly, for transport of the RNP from IBs to the sites of budding (25). M was shown to localize to IBs where, presumably, the first interaction between M and the RNPs occurs. Some early reports have shown that M localization to IBs is mediated by interaction with M2-1 (26, 27). However, more recent work has demonstrated that M is found in IBs when expressed with the N and P proteins alone (17). As N is not required for RSV virus-like filament formation, M was suggested to interact with P. However, the exact mechanism of interactions between these proteins remains largely unknown. Structural data published previously by our group showed that M forms dimers and that the switch from M dimers to higher-order oligomers triggers assembly of viral filaments and virus production (28). Based on M structure, a long patch of positively charged surface spanning the entire monomeric protein was suggested to drive the interaction with a negatively charged membrane (29).

Functional and structural data are available for the P protein, which is a multifunctional protein capable of interacting with multiple partners. Recent studies allowed better characterization of its interactions and functions within the viral polymerase complex. P forms tetramers of elongated shape composed of a central oligomerization domain (OD), mapped to residues N131 to T151 (30–34), and of N- and C-terminal intrinsically disordered regions (IDRs). Structural study of P in solution by nuclear magnetic resonance (NMR) gave insight into the secondary structure propensity of these IDRs, forming almost stable helices in the C-terminal region and extremely transient helices in the N-terminal region (35). Residues 1 to 29 in the N-terminal region confer a chaperone function to P, by binding monomeric and RNA-free N (N⁰) and by maintaining N⁰ unassembled (36). P C-terminal residues 232 to 241 were shown to bind RNA-bound N assembled as rings mimicking the RNP (37). Very recently, the structure of the L protein bound to tetrameric P solved by cryo-electron microscopy revealed that each of the four P monomers adopts a distinct conformation upon binding to L, the entire L-binding region on P spanning residues 130 to 228. This includes the OD and the major part of the C-terminal domain (32, 38, 39). As part of viral transcription regulation, the P region spanning residues 98 to 109 was shown to be the binding site for the RSV transcription antitermination factor M2-1 (40). This interaction is involved in the recruitment of M2-1 to IBs (41). P also plays a pivotal role in M2-1 dephosphorylation mediated by the host protein PP1, which binds to P through an RVxF-like motif located at P residues 82 to 87 (41). Dephosphorylated M2-1 is recruited to specific regions in IBs, called IB-associated granules (IBAGs), where viral mRNAs are concentrated, before trafficking back to the cytoplasm. This cycle is essential for RSV transcription and translation (41).

Recently, a P region encompassing residues 39 to 57 was found to be critical for VLP formation (42). In this study, it was also observed that the OD of P had no significant contribution for VLP assembly and budding. Additionally, phosphomimetic substitutions in P region 39–57 inhibited VLP formation, suggesting that this region needs to be unphosphorylated for VLP production (42). P region 110–120 was also shown to be required for efficient virus budding, affecting the final step of filament scission and virus release (42, 43). However, until now, no direct interaction between P and M has been shown. The lack of structural information on the putative M-P complex makes it difficult to speculate about the mechanism of budding. Thus, identifying specific RSV M-P protein-protein interactions has the potential to break new ground in our understanding of the mechanism behind RSV particle formation. This can further benefit RSV VLP-based vaccine research.

(This article was submitted to an online preprint archive [44].)

RESULTS

Specific localization of M depends on expression of F, P, and N proteins.

As shown previously, RSV VLPs can be generated independently of viral infection by transfecting cells with plasmids encoding the RSV M, P, N, and F proteins (18, 28). Although N is not required for RSV filament formation (17), it localizes with M in filaments when present and is required, together with P, for the formation of pseudo-IBs (18, 28). In a first attempt to study the colocalization of M with N and P proteins in pseudo-IBs and in virus-like filaments at the plasma membrane, and to confirm the minimal requirement for filament formation in our system, we used a transfection-based assay.

BEAS-2B cells were transfected to express M, N, P, and F or various combinations of these four proteins. The intracellular localization of RSV proteins and the formation of pseudo-IBs and virus-like filaments were determined by confocal imaging after staining in parallel with either anti-M and anti-N antibodies (Fig. 1A) or anti-P and anti-N antibodies (Fig. 1B). In the presence of M, N, P, and F, the formation of pseudo-IBs and of virus-like filaments was detected by immunostaining of M, N, and P. Colocalization of M, N, and P was detected on filaments at the plasma membrane as well as in pseudo-IBs, as shown in zoomed merged images (Fig. 1A and B, top rows, positive control). When F was absent, M, P, and N localized in pseudo-IBs (Fig. 1A and B, second rows). M, but not P or N, was also found in small spikes, most probably at the plasma membrane. In the absence of P or N, no pseudo-IBs formed, as expected (Fig. 1A and B, third and bottom rows). In the absence of P, M was found only as small spikes, whereas N spread all throughout the cytoplasm (Fig. 1A and B, third rows). Finally, as previously reported, large filaments containing M and P formed in the absence of N (Fig. 1A and B, bottom rows). Altogether, our results confirm previous observations showing that only M, P, and F are required for virus-like filaments formation and that M colocalizes with N and P within pseudo-IBs and in virus-like filaments (17, 18, 26, 28).

FIG 1.

Specific localization of M depends on expression of F, P, and N proteins. BEAS-2B cells were cotransfected with pcDNA3.1 plasmids expressing RSV M, P, N, and F or a different combination of three proteins. Twenty-four hours posttransfection, cells were fixed, immunostained with anti-M (green) (A) and anti-N (red) (B) or anti-P (green) and anti-N (red) primary antibodies followed by Alexa Fluor secondary antibodies, and analyzed by confocal microscopy. Scale bars, 10 μm. Merged images are zoomed in 3×.

M interacts with P in cells.

As M localization in virus-like filaments seems to depend on P, we then studied whether M can interact with P in cells in the absence of other viral proteins. For that purpose, we used the Nano Luciferase (NanoLuc) interaction assay based on the split NanoLuc reporter (45). In this system, the 114 or the 11S NanoLuc fragments were fused to the C termini of viral proteins (Fig. 2A). Analysis of the lysates of transfected 293T cells by Western blotting using anti-P, anti-M, anti-N, or anti-M2-1 antibody confirmed protein expression with the fused NanoLuc fragments (Fig. 2B). In addition to the expected bands, a weak higher migrating band, probably unspecific, was detected in P-114 and P-11S samples, whereas a double band was detected in the M2-1-11S sample, corresponding most probably to the phosphorylated and unphosphorylated forms of M2-1 (41). To investigate the P-M interaction, combinations of two constructs were transfected into 293T cells. Twenty-four hours posttransfection, cells were lysed, luciferase substrate was added, and the luminescence, proportional to the strength of the interaction, was measured (Fig. 2C). As P is known to form tetramers (30–35), P/P interaction was used as positive control. As shown in Fig. 2C, transfection of P-114/P-11S resulted in a high luminescence signal, indicating a strong interaction. We also used the P-N interaction as a control: when coexpressing P-114 and N-11S, positive but relatively low luminescence was observed. Here, as the NanoLuc 114 subunit was cloned at the C terminus of P protein and thus blocked the interaction between the C terminus of P and RNA-bound N, the luminescence signal corresponded to the P-N⁰ interaction, which was previously shown to be rather weak, in the micromolar range (36). When M-114 was coexpressed with P-11S, the luminescence signal was similar to the P-N⁰ signal, suggesting that M interacts with P with a similar affinity as N⁰. In contrast, expression of M-114 with M2-1-11S did not produce luminescence, suggesting that M and M2-1 did not interact when coexpressed in cells in our system. However, for this negative result, we cannot exclude that the C-terminal tag blocks protein-protein interactions that occur via the C terminus of one or both proteins.

FIG 2.

M interacts with P in cells. Protein-protein interactions were measured using the NanoLuc assay. (A) Scheme of the RSV protein constructs fused with NanoLuc 114 or 11S subunit and pair combinations used in panel C. (B) 293T cells were transfected with plasmids encoding P, N, M, or M2-1 fused to 114 or 11S NanoLuc subunit. Cells were lysed 24 h posttransfection, and cell lysates were then subjected to Western analysis using anti-P, anti-M, anti-N, or anti-M2-1 polyclonal antibody. Size markers are shown on the left side of each gel. (C) 293T cells were transfected with pairs of constructs, combined as shown in the graph. P/P and P/N were used as positive controls. Cells were lysed 24 h posttransfection, and luminescence was measured using a Tecan Infinite 200 plate reader. The NLR is the ratio between actual read and negative controls (each protein with the empty NanoLuc vector). The graph is representative of four independent experiments, each conducted in three technical repeats. Data represent the means and error bars represent standard deviations across 4 independent biological replicates.

M directly interacts with P via its N-terminal region and OD.

Next, we investigated whether M directly interacts with P in vitro using recombinant proteins and determined the P region involved in the interaction. Based on structural data available for isolated P (30, 35), we generated rational deletions of each subdomain of P (Fig. 3A). While P fragments lacking the OD (P_ΔOD, P_1–126, and P_161–241) were shown to be monomeric by NMR, P fragments containing the OD (full-length P, P_OD, P_1–163, and P_127–241) appeared to be associated via the OD (35). The purity and the size of all the P fragments were controlled by SDS-PAGE (Fig. 3B). M and full-length P or P fragments were coincubated, and the formation of complexes was analyzed by native agarose gel electrophoresis. Of note, in native gels, proteins migrate according to the combination of their size, shape, and charges. This explains why P_ΔOD migrates at a higher apparent molecular weight than full-length P, as the global charges are −20.7 for monomeric P_ΔOD and −86.8 for tetrameric P at pH 7.4. When M was incubated with P fragments, shifts were observed only with full-length P and P_1–163 (Fig. 3C), indicating that the N-terminal domain of P (residues 1 to 131) and the OD (residues 131 to 151) are together required for a stable P-M interaction in this system. Additionally, no shift was observed either for P_ΔOD or for P_1–126 when incubated in the presence of M. This suggests that either the M-binding site is at least partly on the OD of P or that M interaction requires tetrameric P. Since no shift was observed for P_OD, P_127–241, and P_161–241 when incubated with M, these results showed that the OD or the C-terminal region of P on their own or together were not sufficient for a detectable P-M interaction in vitro.

FIG 3.

M directly interacts with P via its N-terminal region and OD. (A) Scheme of the secondary structure of P protein (α_N1, α_N2, α_C1, and α_C2 denote transient helices detected in isolated P [35]) and the P constructs used. P_1–163 and P_1–161 constructs were used for bacterial and mammalian expression, respectively. (B) SDS-PAGE and Coomassie blue staining of purified recombinant M, P, and P fragments. (C) M protein was coincubated with P or P fragments for 30 min prior to analysis of formation of complexes by band shift on native agarose gel. Arrowheads indicate complex formation. (D) BEAS-2B cells were cotransfected with pcDNA3.1 plasmids expressing RSV M, F, N, and P WT or P deletion mutants. Cells were fixed, immunostained with anti-M (green) and anti-P (red) antibodies followed by Alexa Fluor secondary antibodies, and analyzed by confocal microscopy. Scale bar, 10 μm. Merged images are zoomed in 3×.

Based on these results, we next wanted to assess if a P fragment containing the N-terminal region and the OD was sufficient to interact with M and to induce membrane filaments in cells. We used the P_1–161 fragment, shorter than P_1–163 by two residues, which are outside the OD on the C-terminal side. Again, we used the filament formation assay. BEAS-2B cells were transfected to express M, N, F, and P, P_1–161, or P_1–126 constructs, and the formation of RSV virus-like filaments was determined by confocal imaging after staining with anti-M and anti-P primary antibodies (Fig. 3D). As previously shown (Fig. 1), in the presence of M, N, F, and P, the formation of pseudo-IBs and virus-like filaments was observed, and M colocalized with P in both structures (Fig. 3D, top row, positive control). Colocalization was not seen in all pseudo-IBs, but this could reflect different maturation states of pseudo-IBs. In the absence of P or when P_1–126 was expressed, neither IBs nor virus-like filaments were detected, and M was found in small spikes (Fig. 3D, second and bottom rows). Virus-like filaments were detected when the P_1–161 construct was expressed, and M-P_1–161 colocalization occurred in virus-like filaments (Fig. 3D, third row). These results are in agreement with those obtained with the gel shift assay using recombinant proteins (Fig. 3C) and confirmed that the P_1–161 fragment is competent for M binding and for filament formation. It is noteworthy that under these conditions, no pseudo-IBs were detected, because the interaction between the C terminus of P (missing in the P_1–161 construct) and N is critical for their formation (46). Altogether, our results show that the P_1–161 fragment is sufficient to interact with M and to induce the formation of virus-like filaments in cells in the presence of M, N, and F.

Identification of M interaction sites on P by NMR.

Next, we sought for a method to identify the M interaction site on P, in a residue-specific manner, without resorting to mutations or internal deletions in potential regions of interest. We have previously used NMR to localize molecular recognition features (MoRFs) on P, taking advantage of the intrinsically disordered nature of this protein (35). Perturbations in amide ¹H-¹⁵N correlation spectra, either of intensities (Fig. 4A and 5) or chemical shifts, were used to map regions that sense direct interaction and/or conformational changes due to binding of a partner.

FIG 4.

Observation of a direct interaction between RSV P1-163 and M-Y229A mutant by NMR. Superimposed 2D ¹H-¹⁵N BEST-TROSY HSQC spectra (A) and 1D ¹H spectra (B) with water suppression of 25 μM ¹⁵N-labeled P_1–163 alone (red) and with 4 molar equivalents of M-Y229A (black). Acquisition was performed in 50 mM Na phosphate (pH 6.7), 150 mM NaCl buffer at 800-MHz ¹H frequency and a temperature of 288 K. Residue-specific assignment of each 2D peak is indicated by the P residue number and the amino acid type. Signals with significant intensity decrease are annotated in color. Arrows indicate specific M ¹H NMR signals, in methyl (−1 to 1 ppm) and aromatic (6 to 7 ppm) proton regions. (C) Intensity ratios (I/I₀), represented as bar diagrams, were measured for each peak in the HSQC spectra of 25 μM ¹⁵N-labeled P_1–163 in the absence and in the presence of M-Y229A, at P/M molar ratios of 1:2 and 1:4. Signals from the α-helical oligomerization domain (α_OD, hatched area) are broadened beyond detection. Colored background indicates the localization of P-specific regions: transient helices α_N1 and α_N2 (also termed site I), extended region β_N, and site II upstream of α_OD. (D) Intensity ratios (I/I₀) measured from HSQC spectra of 40 μM P_1–163 in the absence and in the presence of 150 μM M-Y229A, using two different salt concentrations, 150 mM and 300 mM, as indicated.

FIG 5.

Localization of RSV M-Y229A interaction regions on P by NMR using fragments of P_1–163. Superimposed 2D ¹H-¹⁵N BEST-TROSY HSQC spectra of ¹⁵N-labeled N-terminal P fragments. Samples were in 50 mM Na phosphate (pH 6.7), 150 mM NaCl buffer. Spectra were recorded at 800-MHz ¹H frequency. Plots showing 25 μM ¹⁵N-labeled P_1–126 alone (green) and after addition of 4 molar equivalents of M-Y229A (black) at a temperature of 288 K (A), 50 μM ¹⁵N-labeled P_113–163 alone (pink) and with 2 molar equivalents of M-Y229A (black), at 288 K and 298 K (B), and 25 μM ¹⁵N-labeled P_120–160 alone (blue) and in the presence of 2 molar equivalents of M-Y229A (black), at 288 K (C). (D) Intensity ratios (I/I₀) were determined from the HSQC spectra of P_1–126 and P_113–163 with and without M-Y229A. Signals from the OD (α_OD, hatched area), which are broadened beyond detection at 288K for P_113–163, become visible at 298 K. Other areas are highlighted using the same color code as in Fig. 4.

To control the oligomerization state of M, we used the M-Y229A mutant, which was shown to form dimers but is less prone to form higher-order oligomers than wild-type (WT) M (28). No aggregation or self-assembly of M-Y229A was observed; the samples stayed clear, and an M-specific signal was detected in ¹H NMR spectra, also in the presence of P (Fig. 4B). Of note, since M-Y229A is a dimeric 2 × 29-kDa folded protein, M signals are much broader than P signals that stem from the intrinsically disordered N-terminal region of P. This was also verified by size exclusion chromatography (SEC) analysis made after NMR measurements (data not shown). A buffer with reduced ionic strength (150 mM NaCl compared to 300 mM, the usual M storage buffer [13]) was used to enhance potential electrostatic interactions between M and P. The temperature was set to 288 K to increase the stability of M in this buffer and also to reduce NMR signal broadening of solvent-exposed amide protons due to exchange with water. The final concentration of M was set to 100 μM. Concentrations and molar ratios are given for protomers, independently of the oligomerization state of the proteins. Since the C-terminal part of P seemed to be dispensable for interaction with M and for virus-like filament formation (Fig. 3), we performed experiments with ¹⁵N-labeled P fragments devoid of this part rather than with full-length P: this reduces signal overlap in ¹H-¹⁵N correlation spectra.

Incubation of ¹⁵N-labeled P_1–163 with M-Y229A at a P/M molar ratio of 1:4 (25 μM P and 100 μM M) resulted in significant intensity decrease of several P_1–163 amide signals in two-dimensional (2D) ¹H-¹⁵N correlation NMR spectra (Fig. 4A). These perturbations took place in two proximal regions located immediately upstream of the OD: in the α_N2 region (residues 98 to 111) and in a stretch spanning residues 115 to 125 (Fig. 4C). The first corresponds to the extremely transient helix α_N2, previously identified as the binding site of RSV M2-1 protein (35, 41). Due to specific dynamics in the α-helical coiled-coil OD, amide ¹H-¹⁵N signals are broadened beyond detection at 288 K for P fragments that contain the OD flanked by N- and/or C-terminal extensions (35). Hence, M-binding to the OD could not be assessed using P_1–163.

A concentration effect for P_1–163 was evidenced by comparing the intensity ratio I/I₀ measured with two different P/M molar ratios (Fig. 4C). A maximal effect, where signals disappear completely, could not be reached, because M could not be concentrated >100 μM without starting to aggregate and because the concentration of ¹⁵N-labeled P had to be kept >10 μM due to the sensitivity limitation of NMR. Altogether, the NMR experiments with P_1–163 show that an interaction takes place between P_1–163 and M-Y229A, which involves a region immediately upstream of the OD. This interaction is rather weak. A set of NMR data with P_1–163 and M-Y229A was also acquired in a buffer with 300 mM NaCl and compared to that with 150 mM NaCl (Fig. 4D). Overall intensity perturbation was larger with the lower salt concentration, indicating that M-Y229A binding is stronger under these conditions and suggesting that there may be an electrostatic contribution to the P-M interaction.

To assess the role of the OD, we used the P_1–126 fragment, which is devoid of it. When incubated with 4 molar equivalents of M-Y229A, no significant effect was observed in the intensities of the ¹H-¹⁵N correlation spectrum (Fig. 5A and D, top), indicating that the OD is required for M binding. These results correlate well with the band shift experiments for the M/P complex using native gel analysis (Fig. 3C). To further characterize the binding site on P_1–163, we performed NMR interaction experiments using the shorter P_113–163 fragment, which contains only the OD flanked by residues 115 to 130 and 152 to 163 (Fig. 5B and D, bottom). In the presence of M-Y229A, NMR signal intensities decreased by 30% upstream and downstream of the OD. A stronger effect was again observed for residues 115 to 125 (Fig. 5B and D, bottom), suggesting that they play a particular role in M binding. As for P_1–163, the OD could not be observed for the P_113–163 fragment at 288 K.

However, these signals became visible at a higher temperature for P_113–163 (Fig. 5B and D, bottom). The intensity perturbation pattern of P_113–163 at 298 K suggests that the OD region is affected by the presence of M-Y229A. For P_120–160, the smallest fragment used in this study, all amide signals were observed, even at 288 K (Fig. 5C). M-Y229A induces a weak global reduction in intensity that may point to weak binding to P_120–160. Taken together, these results suggest that M binding to P is achieved through multiple contact sites that are located in α_N2 (site I), in the 115–125 region (site II), and potentially in the OD. Binding to any of these sites appears be rather weak, and M binding to the two N-terminal P regions requires tetrameric P.

Validation of M-binding sites on P using P deletion mutants.

In order to validate the NMR data showing multiple M binding sites, we then performed band shift assays with different fragments of P. We generated P deletions using P_1–163 as a template and deleted a different part of the N-terminal domain for each construct. The OD was present in all constructs to keep the P protein as a tetramer (Fig. 6A). Each construct contained different M-binding domains. P_40–163 contained the region spanning residues 39 to 57, previously reported to affect RSV assembly and potentially involved in M interaction (42), as well as all binding sites identified by NMR (Fig. 4 and 5). P_80–163 contained the two potential binding sites I and II identified by NMR. P_113–163 contained only site II, and P_120–60 lacked both sites, containing only the OD. We used both WT M and M-Y229A in order to ascertain that the NMR results were not biased by the Y229A M mutation. M proteins, full-length P, and P fragments were purified as described above and analyzed by SDS-PAGE (Fig. 6B). M and P constructs were coincubated before analysis of complexes by native agarose gel electrophoresis (Fig. 6C). Shifts were observed with full-length P and P_1–163, as shown in Fig. 3C. P_40–163 and P_80–163 fragments (containing both sites I and II and OD) also induced a shift. No band shift was detected when M was incubated with P_113–163 (containing only site II and OD) or with P_120–160. Similar shifts were observed for M WT and M Y229A. This is in line with our NMR results (Fig. 4 and 5), indicating that M binding by P is achieved through multiple sites. Interactions seen by NMR with P_113–126 or P_120–160 were not strong enough to induce a band shift when analyzed by native gel electrophoresis.

FIG 6.

Validation of novel M-binding sites on P using P deletion mutants. (A) Scheme of the P protein secondary structure and the P constructs used. (B) His-tagged and size exclusion chromatography-purified M, P, or P fragments purified on glutathione beads were analyzed using SDS-PAGE and Coomassie blue staining. (C) Purified M WT (top) or M Y229A (bottom) and P or P deletion constructs were incubated 30 min prior to analysis of formation of complexes by band shift on native agarose gel. Arrowheads indicate complex formation. (D) 293T cells were transfected with pairs of M and P constructs fused to 114 or 11S NanoLuc subunit, combined as shown in the graph. P/P was used as positive control. Cells were lysed 24 h posttransfection, and luminescence was measured using a Tecan Infinite 200 plate reader. The NLR is the ratio between actual read and negative controls (each protein with the empty NanoLuc vector). The graph is representative of four independent experiments, each conducted with three technical repeats. Data represent the means and error bars represent standard deviations across 4 independent biological replicates. *, P < 0.05 (unpaired two-tailed t test). (E) Same cellular lysates used in panel D were then subjected to Western analysis using anti-P or anti-M polyclonal antibody. Size markers are shown on the left side of each gel.

In order to verify which of the identified specific M/P binding regions is relevant for interaction between M and P in cells, we used again the NanoLuc interaction assay based on the split Nano luciferase reporter. The 114 or the 11S Nano luciferase fragments were fused to the C termini of M and P proteins (Fig. 2A). P_WT and four P deletion mutants were analyzed: P_Δ39–57 (lacking the region previously reported to be involved in M/P interaction [42]), P_Δ93–126 (lacking sites I and II as identified by NMR), P_{Δ39–57,Δ93–126} (lacking the three regions), and P_ΔOD. Combinations of two proteins were transfected into 293T cells. Twenty-four hours posttransfection, cells were lysed, luciferase substrate was added, and the luminescence, proportional to the strength of the interaction, was measured (Fig. 6D). P/P interaction was again used as positive control. As shown in Fig. 6D, transfection of P-114/P-11S resulted in a high luminescence signal, indicating a strong interaction. Transfection of M-114 and P-11S resulted in positive signals (comparable to those in Fig. 2C). Transfection of P_Δ39–57-11S, P_Δ93–126-11S, or P_{Δ39–57,Δ93–126}-11S and M-114 all resulted in luminescence comparable to that when transfection of P-11S was used. In contrast, when M-114 was transfected with P_ΔOD-11S, the luminescence signal was almost completely lost. These results indicate that, in transfected cells, P OD deletion and, as a consequence, abrogated tetramerization resulted in a significant loss of interaction with M, whereas deletion of amino acids 39 to 57, 93 to 126, or both regions did not. Of note, analysis of the lysates of transfected 293T cells by Western blotting using anti-P or anti-M antibody confirmed correct protein expression with the fused NanoLuc fragments (Fig. 6E).

Functional implications of P regions displaying direct M binding.

Next, we asked whether the identified M-binding domains on P were of functional significance. We analyzed virus-like filament formation using full-length P or deletion constructs based on NMR results and on previous publications. Specifically, we checked full-length P, P_1–161, P_ΔOD, P_Δ39–57, P_Δ93–110 (deleted of M-binding site I), P_Δ113–126 (deleted of M-binding site II), and P_Δ93–126 (lacking both sites I and II as identified by NMR) (Fig. 7A). BEAS-2B cells were transfected to express M, N, F, and P_WT or various P deletion constructs, and the formation of RSV virus-like filaments was determined by confocal imaging and immunofluorescence after staining with anti-M and anti-P primary antibodies (Fig. 7B). Transfecting M, N, F, and P resulted in pseudo-IBs, filament formation, and colocalization of M and P proteins in both (positive control, Fig. 7B, top left). Transfecting M, N, F, and P_ΔOD did not produce any virus-like filaments or pseudo-IBs, and M and P did not colocalize (Fig. 7B, bottom left). Cells expressing P_Δ39–57 failed to produce virus-like filaments, and although P was found in pseudo-IBs, M was not recruited to IBs (Fig. 7B, top right). Transfection of P_Δ93–126 resulted in virus-like filament formation (Fig. 7B, bottom right), and both M and P proteins were found in virus-like filaments, similarly to P_WT, showing that sites I and II are not critical for these processes.

FIG 7.

Functional analysis of P regions displaying direct M binding. (A) Scheme of the P protein secondary structure and the P constructs used. (B) BEAS-2B cells were cotransfected with pcDNA3.1 plasmids expressing RSV M, N, F, and PWT or P deletion mutants. Cells were fixed, permeabilized at 24 h posttransfection, immunostained with anti-M and anti-P primary antibodies followed by Alexa Fluor secondary antibodies, and analyzed by confocal microscopy. Scale bar, 10 μm. Merged images are zoomed in 3×. (C) HEp-2 cells were cotransfected with pcDNA3.1 plasmids expressing RSV M, F, and P_WT (left lane, positive control) or pcDNA3.1 plasmids carrying RSV M, F, and an empty pcDNA3.1 vector (second lane, negative control) or with the indicated RSV P mutant constructs (third to eighth lanes). At 48 h posttransfection, VLPs (top) were isolated from the supernatant by pelleting of the clean supernatant through a sucrose cushion. Cell lysates (bottom) were generated using RIPA buffer. VLPs and cell lysates were then subjected to Western analysis using anti-P or anti-M polyclonal antibody. The amounts of M and P protein in VLPs were quantified using ImageJ software and are presented as percentage (%) of M and P protein released when P_WT was used (100%). The graph is representative of four independent experiments. Data represent the mean percentages and error bars represent standard deviations across 4 independent biological replicates. ***, P < 0.0001 (unpaired two-tailed t test).

Finally, we asked whether P deletion mutants could negatively affect VLP release. HEp-2 cells were transfected to express M, F, and various P constructs. The N protein was not included in the assay, since it is not required for VLP production (Fig. 1) (17). Cell lysates (soluble fractions) and the VLPs released into the medium were analyzed by Western blotting using anti-P and anti-M antibodies (Fig. 7C, top). All P mutants were correctly expressed and were well detected by the anti-P antibody, as shown by the bands for the cell lysates. Quantification of relative M and P levels in the VLPs, revealed by Western blotting, is shown in Fig. 7C, bottom. When P_WT was expressed, VLP release was validated using anti-M antibodies, and P was detected in VLPs using anti-P antibodies (positive control). The absence of P prevented VLP release (negative control). The P_1–161 construct resulted in VLP release comparable to that for P_WT. Expression of P_ΔOD significantly reduced VLP release, as almost no M was detected. P_Δ39–57 abolished VLP release, as no M and P were detected. This is in agreement with our virus-like filament formation assay (Fig. 7B), where neither P_ΔOD nor P_Δ39–57 induced filament formation. Transfection of P_Δ93–110, P_Δ113–126, or P_Δ93–126 mutants did not abolish VLP release, as M was detected in the VLP fraction at comparable levels to those for P_WT with P_Δ93–126 or slightly reduced with P_Δ93–110 and P_Δ113–126. P_Δ93–110 incorporation into released VLPs was comparable to that of P_1–161, i.e., with only a 25% to 30% reduction relative to that for P_WT. In contrast, transfection of P_Δ113–126 and, even more, of P_Δ93–126 significantly reduced detection of these mutants in released VLPs. This suggests that P_Δ113–126 and P_Δ93–126 mutants still supported VLP formation, in contrast to P_ΔOD and P_Δ39–57, but could be more easily displaced from VLPs than P_Δ93–110, P_1–161, or P_WT. This effect appears to be linked to the deletion of site II. It was not evidenced in the virus-like filament formation assay carried out with P_Δ93–126 (Fig. 7B), most likely because this assay examined filaments formation in situ rather than VLPs that underwent a purification process. Additionally, the discrepancy can be attributed to the fact that the immunofluorescence method does not quantitate and the difference in P levels was not detected.

DISCUSSION

RSV M makes a direct interaction with RSV P.

It was reported earlier that three RSV proteins, P, M, and F, are sufficient for virus-like filament formation and VLP release (17). Our results confirm this minimal requirement (Fig. 1). A question that remained open was which interactions M needed to be involved in to promote assembly and budding of viral particles. For Paramyxoviridae, it was shown that M proteins organize viral assembly by bridging between the glycoproteins and the RNPs and that specific M-N interactions were required for the RNP to be packed into viral pseudoparticles (47). The interaction between the RNP and M proteins occurs most probably first in IBs and/or assembly granules (16), where M recruits the RNP, before they are transported to viral filaments formed on the cellular membrane.

Here, we showed that M colocalized with P in pseudo-IBs, which are formed when N and P are present, as well as in virus-like filaments, even in the absence of N (Fig. 1). Our NanoLuc results showed that M interacts with P in the absence of other viral proteins in cells (Fig. 2). Gel shift assays performed with purified recombinant M and P fragments (Fig. 3C) confirmed M and P directly interact. Moreover, our data indicated that the P domain responsible for this M interaction is located in fragment P_1–163, comprising the N-terminal domain and the OD of P (Fig. 3C and 6C). When M was coexpressed with P_1–161, no pseudo-IBs were formed, as expected, since P_1–161 lacks the C terminus of P needed for interaction with RNA-N complexes and formation of IBs (11, 37, 46). However, virus-like filaments were formed, where M and P_1-161 colocalized. This argues against the requirement of preliminary IB formation before virus-like filament assembly, as these were observed in cells transfected with RSV M, P, and F only. Furthermore, an interaction between RSV M and P strongly suggests that for RSV assembly, bridging between M and the RNP might be mediated by P.

P displays multiple direct contact sites for M.

According to our NMR data (Fig. 4 and 5), M binding to P could be achieved through multiple contact sites that are located within three regions, the α_N2 region (site I), the 115–125 region (site II), and the OD (Fig. 5). It must be noted that NMR experiments were carried out with the M-Y229A mutant, which stays dimeric in solution. The interactions observed by NMR therefore represent a P-M complex formed with dimeric M. The NMR results were corroborated in vitro by native gel interaction assays using both M WT and M-Y229A and N-terminally truncated P fragments (Fig. 3 and 6). Interestingly, neither sites I and II together but without the OD (P₁₋₁₂₆ compared to P₈₀₋₁₆₃ or P₁₋₁₆₃) nor the OD without sites I and II (P_120−160 compared to P₈₀₋₁₆₃) seemed to be sufficient to bind M (Fig. 3C, 4, 5, and 6C). The split luciferase assay shed more light on the complexity of M-P interactions. The deletion of the internal N-terminal region spanning site I and II did not impair M binding in cellulo, in contrast to the internal deletion of the OD, which completely impaired the M-P interaction under the same conditions (Fig. 6D). This could indicate that other short intrinsically disordered N-terminal motifs, similar in sequence to sites I or II, could complement the M-binding site, as they are closer to the OD than in P_WT due the 30-residue-long truncation, or that cellular proteins help stabilize the M-P complex. Interestingly, as illustrated by the P_Δ39–57 deletion mutant, highlighting the essential role of the 39–57 region (42), the ability to form an M-P complex in cellulo is not correlated with virus-like filament formation or P-M colocalization on these virus-like filaments (Fig. 6D and 7B).

Although P OD is not sufficient to bind M, it seems to play a major role, possibly because of the higher level of structural organization provided by tetramerization. We previously showed that transient intraprotomer interactions take place in P (35), and the cryo-electron microscopy (cryo-EM) structures of Pneumoviridae L-P complexes show that the four protomers can engage each in different interactions in a single complex (32, 38, 39). This suggests that the P-M interactions are complex and probably cooperative and that M could recognize one or several sites only formed when four P protomers are present.

As can be seen from Pneumoviridae P protein sequence alignment (Fig. 8), the N-terminal P region contains conserved motifs, in particular, the N⁰-binding motif, the RVxF-like motif, which is the binding site for cellular PP1 phosphatase (41), and α_N2, which is the binding site for M2-1 protein. This region also contains conserved clusters of negatively charged residues, notably, just upstream and inside binding site II. As indicated by reduced M-P binding in high-salt buffer, these residues could be important for interaction with M, which displays a large positively charged surface patch (29). Finally, the differences observed between in vitro and in cellulo experiments could also depend on posttranslational modifications of the proteins.

FIG 8.

Sequence alignment of the central part of Pneumoviridae phosphoproteins. Sequence alignment for human (strains A and B), bovine, and ovine respiratory syncytial virus, murine pneumonia virus (MPV), and human and avian metapneumovirus P proteins was generated using T-coffee suite programs (57). UniProt accession numbers are indicated next to each sequence, and residue numbers are given for each sequence. Complete conservation is indicated with white font on a red background, and relative conservation is indicated by red font. Conserved motifs and structure elements are annotated for human RSV (hRSV) P and boxed throughout the aligned sequences. The boundaries of the OD are from the cryo-EM structures of L-P complexes of hRSV and hMPV (32, 58). The “RVxF”-like motif that binds PP1 and the α_N2 M2-1 binding site (41) is boxed in green and magenta, respectively. M-binding sites I and M-binding site II, as identified by NMR, are indicated or boxed in blue, respectively. Identified phosphorylation sites in the hRSV P (43, 53) are indicated by asterisks above the sequence.

Functional relevance of direct binding between M and P.

Our virus-like filament formation and VLP assays shed light on the functional implications of these interactions. Whereas the OD, and likely tetrameric organization provided by the OD, was clearly needed for VLP formation, deletion of sites I or II individually, or both at the same time, did not prevent virus-like filament production (Fig. 7B). However, reduced levels of P were detected in released VLPs, when both sites I and II were simultaneously deleted (Fig. 7C). This could be attributed to detachment of mutated P from released VLPs during the purification process or to defects in P incorporation when the final VLPs are formed rather than defective recruitment to virus-like filaments on the plasma membrane, since these structures formed and colocalization between M and P was detected (Fig. 7B). This could also raise the question about the integrity of the VLPs that were still released.

M-binding sites I and II thus rather appear to modulate P binding to M and to affect the final organization of the VLPs. The newly identified sites I and II also stand in contrast to the previously reported functional importance of the P region spanning residues 39 to 57 (42). In agreement with previously published data, we confirm that P_Δ39–57 completely prevented filament formation and VLP budding (Fig. 7B and C). However, no perturbations were detected in the NMR signals in the 39–57 region of P in the presence of M-Y229A (Fig. 4). The P-M interaction was still detected when residues 39 to 57 were deleted in transfected cells (Fig. 6D), indicating that this is probably not a direct binding region for M, at least not in its dimeric form. Dephosphorylation of the Ser/Thr-rich 39–57 region of P was reported to be required for VLP formation (42). This was the case for recombinant P protein produced in Escherichia coli. As shown by our immunofluorescence (IF) confocal microscopy experiments (Fig. 7B), M failed to colocalize to pseudo-IBs when P_Δ39–57 was expressed. This could indicate that a third, possibly cellular, factor is required for M recruitment to IBs. Some early reports have shown that M localization to IBs is mediated by an interaction with M2-1 (26, 27), and M interaction with M2-1 in cells was shown by co-immunoprecipitation (co-IP) (48). Cryo-EM analysis of RSV filamentous particles showed that M2-1 is located between the M layer and RNP (14, 49, 50). However, our work here shows that M2-1 is not required for M localization into pseudo-IBs (Fig. 1). Moreover, M and M2-1 did not interact in our NanoLuc assay (Fig. 2).

In conclusion, our virus-like filament formation and VLP assays show a strong functional relevance of P region 39–57, which is essential for M localization to IBs and budding process but most probably not due to direct P-M binding. In addition, we show the functional importance of the tetramerization of P, occurring via the short OD region. The OD is probably part of the M interaction site with P but must be complemented by the N-terminal region of P to bind M (Fig. 7).

Possible role for RSV P as a switch between transcription and budding.

Although the functional relevance of binding site I (α_N2 region) could not be fully assessed in the framework of this study, it is striking that it completely overlaps the binding site of transcription antitermination factor M2-1, which is an essential factor for efficient RSV transcription (51) (Fig. 8). It is likely that RNA synthesis processes are frozen just before viral budding. However, such a mechanism has not been clearly investigated to our knowledge. It is established that P plays a critical role in viral transcription, viral RNA synthesis, and budding. In particular, interaction of P with M2-1 is critical for M2-1 recruitment to IBs (52). The overlap of binding sites for two essential proteins driving transcription (M2-1) on the one hand and virus assembly (M) on the other hand could be part of a switch mechanism, where both proteins compete for this site. Interestingly, binding sites I and II contain four reported phosphorylation sites on T108 (53), S116, S117, and S119 (43), which are conserved among RSV P proteins (Fig. 8). Previously published data showed that P phosphorylation on residue T108 abolished P-M2-1 interaction (53). Phosphomimetic mutants of P serines 116, 117, and 119 (inside binding site II) were reported to significantly prevent virus budding (43). Moreover, P lacking residues 110 to 120 reduced budding, which was completely restored when these residues were added to P (42). P is known to be highly phosphorylated in purified virions, and the timing of appearance of phosphorylated P was shown to correspond to the release of RSV virions (54). We therefore suggest that the phosphorylation state of P sites I and II could regulate the switch between RSV transcription/replication and assembly.

In summary, our work further confirms that RSV P protein is a multifunctional protein playing different roles depending on its interactions with other viral proteins. We confirm here that P plays a key role in RSV assembly. We also bring evidence for a direct interaction between M and P, P OD being required for M-P interaction and sites I and II modulating P binding to M in VLPs by mechanisms yet to uncover.

MATERIALS AND METHODS

Plasmid constructs.

pcDNA3.1 codon-optimized plasmids for mammalian expression encoding the RSV A2 M, P, N, F, and M2-1 proteins were a gift from Marty Moore, Emory University (55). Commercially made pciNanoLuc 114 and 11S vectors (GeneCust) were used to clone the RSV A2 codon-optimized M, P, N, and M2-1 constructs using standard PCR, digestion, and ligation techniques. pcDNA P_1–126 and P_1–161 deletion mutants of P were obtained by introducing stop codons at the appropriate sites in the coding sequences. All pcDNA3.1 P deletion mutants in the full-length construct were generated by using the Q5 site-directed mutagenesis kit (New England Biolabs), according to the manufacturer’s recommendations. pciNanoLuc 11S P_Δ39–57, P_Δ93–126, P_{Δ39–57,Δ93–126}, and P_ΔOD mutants were obtained by recloning the relevant constructs from pcDNA vector using standard PCR, digestion, and ligation techniques. For expression and purification of recombinant P proteins, the previously described pGEX-P and pGEX-P_1–126, pGEX-P_1–163, pGEX-P_ΔOD, pGEX-P_127–241, and pGEX-P_161–41 plasmids were used (36). pGEX-P_OD, pGEX-P_40–163, pGEX-P_80–163, pGEX-P_113-163, and pGEX-P_120–160 were generated using standard PCR, digestion, and ligation techniques and introducing codons at the appropriate sites in the coding sequences. It is important to note that pcDNA P_1–161 and pGEX P_1–163 were used for cellular expression and bacterial expression, respectively. For expression and purification of recombinant M protein, the previously described pCDF-M and pCDF-M_Y229A were used (28).

Cell culture.

HEp-2 (ATCC CCL-23) and 293T cells were maintained in Dulbecco modified Eagle medium (DMED; Eurobio) supplemented with 10% fetal calf serum (FCS; Eurobio), 1% l-glutamine, and 1% penicillin-streptomycin. The transformed human bronchial epithelial cell line (BEAS-2B) (ATCC CRL-9609) was maintained in Roswell Park Memorial Institute (RPMI) 1640 medium (Eurobio) supplemented with 10% FCS (Eurobio), 1% l-glutamine, and 1% penicillin-streptomycin. The cells were grown at 37°C in 5% CO₂.

Bacterial expression and purification of recombinant proteins.

For M expression (WT and Y229A mutant), E. coli Rosetta 2 bacteria transformed with the pCDF-M plasmid were grown from fresh starter cultures in Luria-Bertani (LB) broth for 5 h at 32°C, followed by induction with 0.4 mM isopropyl-β-d-thiogalactopyranoside (IPTG) for 4 h at 25°C. Cells were lysed by sonication (4 times for 20 s each time) and lysozyme (1 mg/ml; Sigma) in 50 mM NaH₂PO₄-Na₂HPO₄, 300 mM NaCl, pH 7.4, plus protease inhibitors (Roche), RNase (12 g/ml, Sigma), and 0.25% CHAPS {3-[(3-cholamidopropyl)-dimethylammonio]-1-propanesulfonate}. Lysates were clarified by centrifugation (23,425 × g, 30 min, 4°C), and the soluble His₆-M protein was purified on a nickel Sepharose column (HiTrap, 5 ml, IMAC HP; GE Healthcare). The bound protein was washed extensively with loading buffer plus 25 mM imidazole and eluted with a 25 to 250 mM imidazole gradient. M was concentrated to 2 ml using Vivaspin20 columns (Sartorius Stedim Biotec) and purified on a HiLoad 10/600 Superdex S200 column (GE Healthcare) in 50 mM NaH₂PO₄-Na₂HPO₄, 300 mM NaCl, pH 7.4. The M peak was concentrated to 3 mg/ml using Vivaspin4 columns. The quality of protein samples was assessed by SDS-PAGE. Protein concentration was determined by measuring absorbance at 280 nm. For NMR interaction experiments, a fresh preparation of M was dialyzed into NMR buffer.

For P expression, E. coli BL21(DE3) bacteria transformed with pGEX-P-derived plasmids were grown at 37°C for 8 h in LB broth. Protein expression was induced by adding one volume of fresh LB medium with 0.4 mM IPTG for 16 h at 28°C. Bacterial pellets were resuspended in lysis buffer (20 mM Tris-HCl [pH 7.4], 60 mM NaCl, 1 mM EDTA, 1 mg/ml lysozyme, 1 mM dithiothreitol [DTT], 0,1% Triton X-100) supplemented with complete protease inhibitor cocktail (Roche) for 1 h on ice. Benzonase (Millipore) was then added, and the lysate was incubated for 1 h at room temperature (RT) under rotation. The lysates were centrifuged at 4°C for 30 min at 10,000 × g. Glutathione-Sepharose 4B beads (GE Healthcare) were added to the clarified supernatants, and the mixtures were incubated overnight at 4°C under rotation. The beads were washed with lysis buffer, washed three times with 1× phosphate-buffered saline (PBS), and then stored at 4°C in equal volumes of PBS.

For ¹⁵N-labeled P expression, bacteria were grown in minimal medium supplemented with ¹⁵NH₄Cl (Eurisotop) as a ¹⁵N source. The glutathione transferase (GST) tag was removed by thrombin (Millipore) cleavage, and the cleaved product was exchanged into NMR buffer (50 mM Na phosphate at pH 6.7, 150 mM NaCl). The quality of protein samples was assessed by SDS-PAGE. Protein concentration was determined by Bradford assay (Bio-Rad) and checked by measuring absorbance at 280 nm for fragments containing tyrosine residues. It is given as protomer concentration in the case of P tetramers.

NanoLuc interaction assay.

Constructs expressing the NanoLuc subunits 114S and 11S were used (45). 293T cells were seeded at a concentration of 3 × 10⁴ cells per well in 48-well plate. After 24 h, cells were cotransfected in triplicates with 0.4 μg of total DNA (0.2 μg of each plasmid) using Lipofectamine 2000 (Invitrogen). Twenty-four hours posttransfection, cells were washed with PBS and lysed for 1 h at room temperature using 50 μl NanoLuc lysis buffer (Promega). NanoLuc enzymatic activity was measured using the NanoLuc substrate (Promega). For each pair of plasmids, three normalized luminescence ratios (NLRs) were calculated as follows: the luminescence activity measured in cells transfected with the two plasmids (each viral protein fused to a different NanoLuc subunit) was divided by the sum of the luminescence activities measured in both control samples (each NanoLuc-fused viral protein transfected with a plasmid expressing only the NanoLuc subunit). Data represent the means and standard deviations (SDs) from 4 independent experiments, each performed in triplicates. Luminescence was measured using Infinite 200 Pro (Tecan, Männedorf, Switzerland).

NMR spectroscopy.

For NMR experiments, P and M solutions were mixed to obtain the desired molar ratio and concentrated to reach a concentration of 100 μM for M (WT and Y229A mutant).

NMR measurements were carried out in a Bruker Avance III spectrometer at a magnetic field of 18.8 T (800-MHz ¹H frequency) equipped with a cryogenic TCI probe. The magnetic field was locked with 7% ²H₂O. The temperature was set to 288 K or 298 K. ¹H-¹⁵N correlation spectra were acquired with a BEST-TROSY version. Spectra were processed with Topspin 4.0 (Bruker BioSpin) and analyzed with CCPNMR 2.4 (56) software.

¹H and ¹⁵N amide chemical shift assignments of full-length P and P_1–126 and P_1–163 fragments were performed as described previously (35). Amide chemical shift assignment of ¹⁵N-labeled P_113–160 and of ¹⁵N-labeled P_120-160 was performed by recording a ¹⁵N-separated nuclear Overhauser effect spectroscopy-heteronuclear single quantum coherence (NOESY-HSQC) spectrum with an 80-ms mixing time. Sequential information was retrieved through H_Ni-H_Ni-1, H_Ni-H_Ni+1, and H_Ni-H_αi-1 correlations.

Virus-like filament/particle formation.

Overnight cultures of BEAS-2B cells seeded at 4 × 10⁵ cells/well in 6-well plates (on a 16-mm micro-cover glass for immunostaining) were transfected with pcDNA3.1 codon-optimized plasmids (0.4 μg each) carrying the RSV A2 WT or deletion/mutant P protein along with pcDNA3.1 codon-optimized plasmids carrying RSV A2 M, N, and F using Lipofectamine 2000 (Invitrogen) according to the manufacturer’s recommendations. Cells were fixed at 24 h posttransfection, immunostained, and imaged as described below. For VLP formation, overnight cultures of HEp-2 cells seeded at 4 × 10⁵ cells/well in 6-well plates were transfected as described above. Released VLPs were harvested from the supernatant; the supernatant was clarified of cell debris by centrifugation (1,300 × g, 10 min, 4°C) and pelleted through a 20% sucrose cushion (13,500 × g, 90 min, 4°C). Cells were lysed in radioimmunoprecipitation assay (RIPA) buffer. Cellular lysates and VLP pellets were dissolved in Laemmli buffer and subjected to Western analysis.

Immunostaining and imaging.

Cells were fixed with 4% paraformaldehyde in PBS for 10 min, blocked with 3% bovine serum albumin (BSA) in 0.2% Triton X-100–PBS for 10 min, and immunostained with monoclonal anti-M (1:200; a gift from Mariethe Ehnlund, Karolinska Institute, Sweden), polyclonal anti-N (1:5,000 [30]), polyclonal anti-P (1:500 [30]), or monoclonal anti-P (1:100; a gift from Jose A. Melero, Madrid, Spain) antibodies, followed by species-specific secondary antibodies conjugated to Alexa Fluor 488 and Alexa Fluor 568 (1:1,000; Invitrogen). Images were obtained using the white light laser SP8 (Leica Microsystems, Wetzlar, Germany) confocal microscope at a nominal magnification of ×63. Images were acquired using the Leica Application Suite X (LAS X) software.

SDS-PAGE and Western analysis.

Protein samples were separated by electrophoresis on 12% polyacrylamide gels in Tris-glycine buffer. All samples were boiled for 3 min prior to electrophoresis. Proteins were then transferred to a nitrocellulose membrane (Roche Diagnostics). The blots were blocked with 5% nonfat milk in Tris-buffered saline (pH 7.4), followed by incubation with rabbit anti-P antiserum (1:5,000) (30), rabbit anti-N antiserum (1:5,000) (30), rabbit anti-M antiserum (1:1,000), or rabbit anti-M2-1 antiserum (1:2,000) (30) and horseradish peroxidase (HRP)-conjugated donkey anti-rabbit IgG (1:5,000) antibodies (P.A.R.I.S.). Western blots were developed using freshly prepared chemiluminescent substrate (100 mM Tris-HCl [pH 8.8], 1.25 mM luminol, 0.2 mM p-coumaric acid, 0.05% H₂O₂) and exposed using a Bio-Rad ChemiDoc Touch imaging system.

Generation of M antiserum.

Polyclonal anti-M serum was prepared by immunizing a rabbit three times at 2-week intervals using purified His fusion proteins (100 mg) for each immunization. The first and second immunizations were administered subcutaneously in 1 ml Freund’s complete and Freund’s incomplete adjuvant (Difco), respectively. The third immunization was performed intramuscularly in Freund’s incomplete adjuvant. Animals were bled 10 days after the third immunization.

Native gel.

Protein samples were separated by electrophoresis on 1% agarose gel in TBE (Tris-borate-EDTA) buffer. Fifty micrograms M and 10 μg P were incubated in PBS buffer (pH 7.4) for 20 min at RT. Samples were mixed with 50% sucrose and run for 2 h at 80 V, following by staining with amido black.