Structural Basis of Human Parainfluenza Virus 3 Unassembled Nucleoprotein in Complex with Its Viral Chaperone

ABSTRACT

Human parainfluenza virus 3 (HPIV3) belongs to the Paramyxoviridae, causing annual worldwide epidemics of respiratory diseases, especially in newborns and infants. The core components consist of just three viral proteins: nucleoprotein (N), phosphoprotein (P), and RNA polymerase (L), playing essential roles in replication and transcription of HPIV3 as well as other paramyxoviruses. Viral genome encapsidated by N is as a template and recognized by RNA-dependent RNA polymerase complex composed of L and P. The offspring RNA also needs to assemble with N to form nucleocapsids. The N is one of the most abundant viral proteins in infected cells and chaperoned in the RNA-free form (N⁰) by P before encapsidation. In this study, we presented the structure of unassembled HPIV3 N⁰ in complex with the N-terminal portion of the P, revealing the molecular details of the N⁰ and the conserved N⁰-P interaction. Combined with biological experiments, we showed that the P binds to the C-terminal domain of N⁰ mainly by hydrophobic interaction and maintains the unassembled conformation of N by interfering with the formation of N-RNA oligomers, which might be a target for drug development. Based on the complex structure, we developed a method to obtain the monomeric N⁰. Furthermore, we designed a P-derived fusion peptide with 10-fold higher affinity, which hijacked the N and interfered with the binding of the N to RNA significantly. Finally, we proposed a model of conformational transition of N from the unassembled state to the assembled state, which helped to further understand viral replication.

IMPORTANCE Human parainfluenza virus 3 (HPIV3) causes annual epidemics of respiratory diseases, especially in newborns and infants. For the replication of HPIV3 and other paramyxoviruses, only three viral proteins are required: phosphoprotein (P), RNA polymerase (L), and nucleoprotein (N). Here, we report the crystal structure of the complex of N and its chaperone P. We describe in detail how P acts as a chaperone to maintain the unassembled conformation of N. Our analysis indicated that the interaction between P and N is conserved and mediated by hydrophobicity, which can be used as a target for drug development. We obtained a high-affinity P-derived peptide inhibitor, specifically targeted N, and greatly interfered with the binding of the N to RNA, thereby inhibiting viral encapsidation and replication. In summary, our results provide new insights into the paramyxovirus genome replication and nucleocapsid assembly and lay the basis for drug development.

INTRODUCTION

Human parainfluenza virus 3 (HPIV3, also known as human respirovirus 3) belongs to the genus Respirovirus, a member of the family Paramyxoviridae that contains several highly pathogenic viruses, such as measles virus (MeV), Nipah virus (NiV), Hendra virus (HeV), Newcastle disease virus (NDV), and mumps virus (MuV) (1, 2). HPIV3, breaking out repeatedly during annual spring to early summer, is often associated with pneumonia, bronchiolitis, and bronchitis and causes a range of acute respiratory tract infections, particularly in newborns and infants (3, 4).

Like other members in the order Mononegavirales, HPIV3 is enveloped on the outside and contains an ∼17,000-nucleotide (nt)-long single-stranded genomic RNA inside, which is encapsidated by a nucleoprotein (N) and forms the highly ordered helical nucleocapsids (NCs). The helical NC is a template for viral transcription and replication (5). The phosphoprotein (P), RNA polymerase (L), and NC form the ribonucleoprotein (RNP) complex, which is necessary for viral transcription and replication. As the largest protein of paramyxoviruses, the L contains all the polymerase activities and forms an RNA-dependent RNA polymerase (RdRp) complex with P (6). The P acts as a bridge to anchor the L to NC.

Like the Ns of other viruses in the family Paramyxoviridae, HPIV3 N is composed of an ordered N_CORE region and an intrinsically disordered N_TAIL region. The N_CORE contains a globular N-terminal domain (NTD) and a C-terminal domain (CTD), flanked by NT_ARM and CT_ARM at its amino and carboxyl termini, respectively (7). Several structures of N of paramyxoviruses indicate that N has assembled or unassembled conformations. When expressed recombinantly, N binds to cellular RNAs to form NC assembly in a closed state. The RNA-binding site is on the groove between the NTD and CTD, and the NT_ARM and the CT_ARM are involved in binding to the neighboring N protomers (8). The N_TAIL, due to its flexibility, is on the outside of NC and does not participate in NC assembly, but it can associate with P and adjust the looseness of NC and may further regulate the binding of the RdRp complex to RNA in NC (9–11). For paramyxoviruses, the RNA-binding groove faces the outside of the NC and determines the number of bases and the twist of RNA wrapped in each N (12–15). The P of paramyxoviruses has a modular structure containing three main stable domains: the N⁰ (unassembled form)-binding domain at the amino terminus, the multimerization domain (MD) that binds with L to form L-P complex (16–21), and the C-terminal domain (XD) that binds with the N_TAIL serving as a bridge between NC and L (10, 22, 23). The binding and release of XD and N_TAIL presumably make the RdRp complex move on the NC and synthesize RNA continuously (24, 25). The N⁰-binding domain of P acts as a specific chaperone, keeping N in an RNA-free conformation N⁰ and preventing the polymerization of N (26–28). Then, N⁰ is carried to the replication center by P, which is conducive to the encapsidation of synthetic RNA and avoids the degradation of RNA by host nucleases (12). However, the chaperoning mechanism of P to N and the conformational changes of N in genus Respirovirus are still a lack of knowledge.

To investigate the chaperone effect of P on N of HPIV3 and the mechanism of NC assembling, herein we reconstituted the N⁰-P complex and characterized its crystal structure at 3.23 Å. The complex is uniform both in solution and in crystal. The structure revealed that N⁰ is in an open conformation. The region (residues 22 to 42) of P is sufficient to bind with N⁰, and the interface between P and N⁰ is mainly formed by hydrophobic residues, with the binding affinity reaching a two-digit nanomolar level. Our structure-based mutagenesis experiments showed that the chaperone effect of P on N would be attenuated if conserved hydrophobic residues were mutated. Therefore, the interaction of N⁰-P can be used as a target for the development of antiviral drugs. We also developed a P-derived fusion peptide, which hijacked the N with 10-times higher affinity and interfered with the binding of the N to RNA, as a promising antiviral peptide inhibitor. Our data also contribute to a better understanding of the formation of NC. Based on structural analyses and bioassays, we suggested a reasonable model about the conformational transformation of N-binding RNA during viral replication.

RESULTS

The amino terminus of the HPIV3 P chaperones the N to form an unassembled N⁰-P complex.

The interaction of N and P has been shown to play a critical role in the replication cycle of the virus. Previous studies indicated that the amino terminus of P can prevent premature N polymerization (29, 30). We expressed full-length N (residues 1 to 515), N_TAIL removed (residues 1 to 403), and NT_ARM or CT_ARM truncated (residues 1 to 374 or 29 to 374) proteins in Escherichia coli, but the proteins formed inclusion bodies, which prevented us from obtaining significant amounts of soluble proteins. To improve the solubility of the proteins, we fused a dual His₆ maltose-binding protein (MBP) tag to the N terminus of the proteins. The purified fusion proteins were soluble but formed aggregations that were eluted in the excluded volume of a Superdex 200 Increase 10/300 GL column (Fig. 1A and B). Therefore, we reconstituted several variants of N⁰-P complex: peptide P_1-42 encompassing the N⁰-binding region was fused to the C terminus of N proteins that truncated at the N_TAIL, NT_ARM, or CT_ARM. The chimeras were purified using immobilized metal affinity chromatography, followed by anion exchange and size-exclusion chromatography (SEC), which indicated that the chimeras were monomeric in solution (Fig. 1A and B). The purified N⁰-P complexes migrated as a single band on SDS-PAGE. Because the absorbance ratio of the N_29-374⁰-P_1-42 complex at 260 and 280 nm was 0.56 (A_{260 nm}/A_{280 nm}), the protein sample might not contain nucleic acid (Fig. 1C). Further, by small-angle X-ray scattering (SAXS) experiments (Table 1), the molecular weight of the reconstituted N_29-374⁰-P_1-42 complex was consistent with the monomeric result obtained from SEC. Together, these results verify that the first 42 residues of HPIV3 P maintain N⁰-P complex monodisperse, soluble, and free of nucleic acid.

FIG 1.

Structure of HPIV3 N⁰-P complex. (A) Schematic representation of HPIV3 N and P constructs used in this study. NTD, N-terminal domain of N_CORE; CTD, C-terminal domain of N_CORE; NT_ARM, N-terminal arm of N; CT_ARM, C-terminal arm of N; MD, multimerization domain of P; XD, X domain of P. (B) The size-exclusion chromatogram (SEC) of various HPIV3 N⁰-P complexes. The samples were run on Superdex 200 Increase 10/300 GL. N_1-515, full-length protein; N_1-403, N_TAIL is truncated; N_1-374, both CT_ARM and N_TAIL are truncated; N_29-374, N_TAIL, NT_ARM, and CT_ARM are truncated. Due to the low solubility of individual expression, N_1-515 and N_29-374 were fused to a dual His₆ maltose-binding protein (MBP) tag at the N terminus. N_1-515 and N_29-374 proteins were still aggregated after the tag was cut. (C) SEC of HPIV3 N_29-374⁰-P_1-42 chimera. The right panel shows Coomassie blue-stained SDS-PAGE of protein from SEC peak. Lane M, molecular mass markers (kDa). Elution volumes of the protein standards on a Superdex 200 Increase 10/300 GL column are marked at the top of the figure, respectively. (D) View in cartoon representation of the HPIV3 N⁰-P structure. The N⁰ and P peptides are shown in slate and orange, respectively. Secondary structure elements are labeled. (E) Electrostatic surface potential of HPIV3 N⁰-P complex at ±5 kT/e: red (acidic), white (neutral), and blue (basic). (F) The “omit” F_o − F_c differential electron density map of P in the N⁰-P complex, contoured at 2.5 σ. (G) Analysis of the N_29-374⁰-P_1-42 complex by SAXS. Comparison of small-angle X-ray scattering (SAXS) experimental data (black) and calculated scattering profiles (blue) from the crystal structure of HPIV3 N_29-374⁰-P_1-42. The theoretical curve was calculated and fit to the experimental data using CRYSOL (62). (H) Ab initio model of the complex. The crystal structure of the complex fits well with the ab initio models calculated from SAXS data. OD_{260 nm}, optical density at 260 nm.

TABLE 1.

Statistics of SAXS analysis

Parameter	Value
Data collection
Wavelength (Å)	1.033
Beam size (μm)	320 × 43
Exposure time (s)	1 per frame
Camera length (m)	2.68
Temp (K)	283
q-measurement range (Å−1)	0.007 to 0.355
Concentrations (mg/mL)	1.0
No. collected images	20
Structural parameters
I(0) (cm−1) (from P(r))	14.7 ± 0.1
Rg (Å) (from P(r))	27.9 ± 0.2
I(0) (cm−1) (from Guinier)	15.2 ± 0.27
Rg (Å) (from Guinier)	29.0 ± 0.7
Dmax (Å)	86.0
Porod vol Vp (Å3)	49,300.0
Molecular mass (kDa)
Calculated from Porod vol Vp	40.9
Calculated from protein sequence	42.0
Calculated from program SAXS MoW	43.9
Software employed
Primary data reduction	BioXTAS-RAW (58)
Data processing	GNOM (59)
Ab initio analysis	DAMMIN (60)
Computation of model scattering	CRYSOL (62)
Validation and averaging	DAMAVER (61) DAMSEL (61) DAMFILT (61) SUPCOMB (63)

Crystal structure of the HPIV3 N⁰-P complex.

To explore the relationship between function and structure of N and P, we expressed and purified several HPIV3 N⁰-P complexes. After extensive crystallization trials, we luckily obtained crystals of N_29-374⁰-P_1-42 complex that could be used for diffraction. The HPIV3 N⁰-P complex was crystallized in the space group I422. We determined the structure at a resolution of 3.23 Å by molecular replacement using the MeV N-RNA complex structure (Protein Data Bank [PDB] code 6H5Q) (31) as a template (Fig. 1D; Table 2). The final structural model contained a chimeric molecule in the asymmetric unit. The amino acid sequence of the N_29-374⁰-P_1-42 chimera structure was traced from P residues 20 to 42 and from N residues 29 to 374 except residues 136 to 153. The data collection and model refinement statistics are summarized in Table 2.

TABLE 2.

Data collection and refinement statistics of HPIV3 N_29-374⁰-P_1-42

Parameter	Value
Data collection
Wavelength	0.979
Space group	I 4 2 2
Cell dimensions
a, b, c (Å)	117.1, 117.1, 143.2
α, β, γ (°)	90, 90, 90
Resolution (Å)	50.0 to 3.23 (3.29 to 3.23)a
Rmerge (%)	9.8 (61.3)
Mean I/σ (I)	18.1 (1.8)
CC1/2	1.0 (0.7)
Completeness (%)	99.8 (99.5)
Total no. of reflections	56,267
Unique reflections	8,248 (395)
Multiplicity	6.8 (5.3)
Refinement
Resolution (Å)	50.0 to 3.23 (3.29 to 3.23)
No. of reflections	7,630 (399)
Rwork (%)b	21.9
Rfree (%)c	23.8
No. of atoms	2,708
Protein	2,654
Water	54
Avg B-factors (Å2)
Protein	51.25
Water	26.66
RMSD
Bond lengths (Å)	0.008
Bond angles (°)	1.411
Ramachandran plot (%)
Favored regions	98.8
Allowed regions	1.2
Disallowed regions	0

^aStatistics for the highest-resolution shell are shown in parentheses.

^bR_work = Σ(‖Fp(obs)| − |Fp(calc)‖)/Σ|Fp(obs)|.

^cR_free is an R factor for a preselected subset (5%) of reflections that was not included in refinement.

The overall structure of HPIV3 N_29-374⁰-P_1-42 adopts a bean-like architecture with approximate dimensions of 85 Å in height and 43 Å in width (Fig. 1D). The HPIV3 N_CORE mainly contains two major domains, the NTD (residues 29 to 263) and the CTD (residues 264 to 374), both of which are predominantly α-helices. The NTD domain is formed by α1 helix to α10 helix, two 3₁₀ helices η1 and η2, and three β-sheets β1 to β3 (Fig. 1D). The CTD domain is formed by α11 helix to α14 helix and three 3₁₀ helices, η3 to η5. Electrostatic analysis revealed that the electrostatic surface charge is distributed asymmetrically with negative charge clusters at the NTD and CTD domains, defining a basic hinge region that provides a putative site for RNA binding (Fig. 1E). P_20-42 forms a helix about 24 Å long, binding to the groove formed by the α11, η3, and α12 helices of the N (Fig. 1D and F). SAXS experiments reveal that the complex conformation in the crystal is consistent with that in the solution (Fig. 1G and H).

The residues 22 to 42 of P is sufficient to interact with N.

The residues 1 to 19 of P were not traceable in the electron density map, but proteins from several crystals were analyzed by SDS-PAGE, and only one band was found at 44 kDa (Fig. 2A), consistent with the chimeric molecular weight, indicating no obvious degradation during crystallizing. Thus, the region (residues 1 to 19) of P may be flexible. In addition, the primary sequence alignment of P in respiroviruses implies that the conserved region bound to N may be located at residues 22 to 42 of HPIV3 P (Fig. 2B). To evaluate the role of the region (residues 1 to 21) of P in N⁰ binding, we performed pulldown assays of N_29-374 when coexpressed with glutathione S-transferase (GST)-tagged P_1-42 and its truncations (Fig. 2C). P_1-42 and truncation P_22-42 could efficiently bind N_29-374, while the P_1-21 truncation could not. The differences in affinity require further determination. To obtain RNA-free and P-free N⁰, we designed and expressed a mutated chimeric protein N-TEV-P_1-42 (L29E) (His₆ tag at its C terminus) in E. coli. After being cleaved by tobacco etch virus (TEV) protease, the protein was further purified using immobilized metal affinity chromatography, followed by anion exchange and SEC (Fig. 2D). Microscale thermophoresis (MST) binding experiments showed that residues 1 to 21 of P had little effect on the affinity between N_29-374 and P_1-42 (Fig. 2E). Thus, residues 22 to 42 of P are sufficient to interact with N.

FIG 2.

The P_22-42 is enough to bind N⁰. (A) SDS-PAGE of HPIV3 N⁰-P complex in solution (lane 2) and crystal (lane 3). Lane M, molecular mass markers (kDa). (B) Alignment of the P peptide of typical respiroviruses. Asterisks represent residues binding to N⁰. The identical residues are highlighted in red, and similar residues are in red type. The following viral proteins were used: HPIV1 (human respirovirus 1) P (UniProtKB-P32530), SeV P (P14252), PRV1 (porcine respirovirus 1) P (S5LVM8), BPIV3 (bovine respirovirus 3) P (P06163), and HPIV3 P (P06162). (C) Coomassie blue-stained SDS-PAGE of GST pulldown of N_29-374 by GST-P_1-42 or its mutants when coexpressed in E. coli. Lane M, molecular mass markers (kDa). Note that some mutations affect electrophoretic migration rates in SDS-PAGE (7). (D) SEC of HPIV3 N_29-374⁰ (right panel) Coomassie blue-stained SDS-PAGE of protein from the SEC peak (lane N_29-374⁰). Lane M, molecular mass markers (kDa). The lane labeled “No cut” shows crudely purified protein that has not been cleaved by TEV enzyme. (E) Measurement of the binding affinity of the P peptide or mutations with N_29-374⁰ by microscale thermophoresis (MST). The experiments were repeated three times.

The interaction between N⁰ and the amino terminus of P is mainly hydrophobic.

The crystal structure shows that the P_20-42 docks to the hydrophobic pocket formed by the helices α11, η3, and α12 of the CTD domain of N⁰ (Fig. 3A and B) (32). The buried surface area in the crystallized complex is about 800 Ų, and the interface is mostly composed of hydrophobic residues (Fig. 3B and C). Most residues of P_20-42, such as Ile-25, Leu-29, Ile-31, Ile-32, and Ile-35, form hydrophobic interaction with N⁰, and two residues of P_20-42 form hydrogen bonds with N⁰. The hydrophobic interface was also determined by screening dissociation of N⁰ from P_1-42 under different conditions (Fig. 3D). The screening conditions involve hydrophobic and ionic interactions. NaCl concentrations ranging from 0.5 to 3 M did not cause the release of N⁰, whereas conditions containing Nonidet P-40 (NP-40) did. Conditions containing urea of different concentrations led to denaturing and releasing the complex from agarose beads (Fig. 3D). These results are consistent with those of MeV (7). Thus, hydrophobic interactions make a major contribution to N⁰-P complex formation. We also coexpressed N_29-374 and GST-P_1-42, as well as its mutants. Wild-type P_1-42 efficiently bound N_29-374, whereas mutants at the core hydrophobic interface with negatively charged residues (I25E or L29E) greatly reduced their binding to N (Fig. 2C, lanes 5 and 6). Consistently, MST binding experiments also showed that these two mutations in P significantly reduced the affinity for N (Fig. 2E).

FIG 3.

Hydrophobic interaction between N⁰ and P. (A) The hydrophobic surface of N⁰ and P peptide (purple cartoon). (B) Closeup view of the binding interface between N⁰ and P peptide. The interaction residues are shown as sticks, respectively. Oxygen is shown in red, and nitrogen is shown in blue. Hydrogen bonds are represented by black dotted lines. (C) Details of the interactions between the N⁰ and P peptides were calculated by Ligplot, setting hydrogen-bonds parameters, maximum distances D-A of 3.5 Å and H-A of 2.7 Å, and hydrophobic contacts parameters. Maximum and minimum distances were 3.9 and 2.9 Å, respectively (32). (D) The effect of buffer components on the binding of N to P. SDS-PAGE profile of proteins released from GST-P_1-42 and N_29-374 complex bound to GST beads, when subjected to different conditions. The dosages (final concentrations) are listed at the top. Nonidet P-40 (NP-40) caused N⁰ to release from GST-P_1-42. Urea induced the denaturation of the complex to release from agarose beads.

Unique features of HPIV3 N⁰-P structure.

When we prepared the manuscript, several closed states of Sendai virus (SeV) N-RNA structure were published online (33). The sequence identity of the N between HPIV3 and SeV is 59%, which allows us to use the helical NC of SeV as the NC of HPIV3 to compare and analyze the assembled and unassembled conformations of HPIV3 N. The structures of HPIV3 unassembled N⁰ and SeV assembled N in NC are similar with the root mean square deviation (RMSD) of 2.0 Å (HPIV3 N⁰ as the reference, all residue coverage). The RMSD values are 1.1 Å for the NTD domain and 1.7 Å for the CTD domain (HPIV3 NTD and CTD as the reference, all residue coverage), but there is an ∼10° relative rotation of these two domains around the hinge region (Fig. 4A). Theoretically, the RNA-binding groove of N in the HPIV3 N⁰-P is open with the CTD domain lift up and has a larger cavity volume, readying to bind single-stranded RNA (ssRNA). However, the binding positions of P and ssRNA are opposite at the N, so P does not directly block the RNA-binding site. The N has no sequence specificity for RNA binding, and the two motifs located in the NTD and CTD domains (HELIX-N and HELIX-C) form a tweezers-like structure, which is critical for binding RNA. In these two motifs of SeV N, the residues (K180, K190, N194, Y260, H321, Y350, R354, and Y356) that form hydrogen bonds/salt bridges with the RNA phosphate backbone are almost conserved in the HPIV3 N⁰. However, due to the conformational difference and the deflection of the side chains of these residues, none of these residues on HPIV3 N⁰ is in the right position to bind RNA (Fig. 4B and C).

FIG 4.

Structural superposition of HPIV3 N⁰-P and SeV N-RNA shows conformational changes. (A) Structural comparisons of HPIV3 N⁰-P and SeV N-RNA complex (PDB code 6M7D) (33). HPIV3 N⁰ (slate), HPIV3 P (orange), SeV N (pink), and RNA (gray) were bound to SeV N. The key motifs HELIX-N (motif of NTD domain) and HELIX-C (motif of CTD domain) involved in RNA binding were framed by black dotted lines. Superposing at their NTD domains shows their respective CTDs at an angle of about 10°. (B) Closeup view of RNA-binding groove of SeV N in its closed conformation. Conserved residues (K180, K190, N194, Y260, H321, Y350, R354, and Y356) involved in RNA binding are shown as sticks, and hydrogen bonds are shown as yellow dotted lines. (C) The closeup view of RNA-binding groove of HPIV3 N⁰. The residues that correspond to those in panel B are shown. Structure superimpositions were carried out using DALI (http://ekhidna2.biocenter.helsinki.fi/dali/). Using the HPIV3 N⁰ as the reference, an RMSD value of 2.0 Å was obtained over all residue coverage.

The assembled (N-RNA) and unassembled (N⁰-P) structures of the N of some viruses in the order Mononegavirales have been resolved (2, 7, 12, 15, 31, 34–37). The structures of N in unassembled and assembled states of other viruses (MeV, parainfluenza virus 5 [PIV5], human metapneumovirus [HMPV], and Ebola virus [EBoV]) were compared with HPIV3 N⁰, respectively (Fig. 5). The NTD and CTD domains of N at the unassembled state of all these viruses are relatively rotated and far away from each other; in particular, the CTD and NTD of MeV N have about 40° relative rotation in the open and closed conformations (7). However, the DALI server calculation indicated that the HPIV3 N⁰ structure in the open state is similar to the assembled N of SeV and those viruses, revealing that HPIV3 N might not undergo significant conformational changes during NC assembly and has a unique mechanism (Fig. 4 and 5) (38).

FIG 5.

The structures of N from different viruses in Mononegavirales. The upper and lower panels show the unassembled and assembled Ns, respectively. The NTD domain and CTD domain of N are shown in light orange and dark salmon, respectively. P is shown in slate), NT_ARM of N(i + 1) is shown in green, and RNA is shown in gray. The following viral structures were used: PDB code PIV5 (Paramyxoviridae), unassembled (5WKN), assembled (4XJN); MeV (Paramyxoviridae), unassembled (5E4V), assembled (6H5Q); HMPV (Pneumoviridae), unassembled (5FVD), assembled (5FVC); EBoV (Filoviridae), unassembled (6OAF), assembled (5Z9W) (2, 7, 12, 31, 35–37). Structure superimpositions were carried out using DALI. Using the HPIV3 N⁰ as the reference, the RMSD value and residue coverage are marked.

The mechanism of HPIV3 N binding to RNA.

To further understand the binding of HPIV3 N to RNA, a model for the assembled state of HPIV3 NC was built according to the helical NC of SeV (Fig. 6A). The comparison between N⁰-P and N-RNA showed that the P peptide and NT_ARM of N_i+1 protomer competed for the same binding site (Fig. 6A). NT_ARM provides several important binding sites for the assembly of adjacent N. In the SeV NC-like structure, the binding interface area between NT_ARM of N_i+1 protomer and adjacent N protomer exceeds 1,000 Ų by several hydrogen bonds, salt bridges, and hydrophobic interactions. Therefore, the P peptide maintains the monomeric state of N by interfering with the interaction between N protomers. However, in the N-RNA closed state, the nucleic acid induces the conformational changes of the N protomers and the association between the N protomers, making them arrange in an orderly manner along with the RNA. Nuclear magnetic resonance (NMR) experiment for N⁰-P of NiV showed that the flexible resonances of NT_ARM and CT_ARM disappeared, but the resonances of P could be detected upon the addition of RNA (39). Therefore, we hypothesized that P might be replaced by NT_ARM of adjacent N within the coparticipation of NT_ARM and CT_ARM of adjacent Ns during NC assembly. To verify our hypothesis, we tried electrophoretic mobility shift assay (EMSA) experiments. The results showed when different truncations of the N in the unassembled state of N⁰-P were incubated with RNA, all the fragments containing the NT_ARM bound RNA (Fig. 6B). Fragment N_1-403 containing both NT_ARM and CT_ARM had a higher binding ability than fragment N_1-374 containing NT_ARM alone (Fig. 6B). Fragment N_29-403 that contained only CT_ARM could bind to RNA weakly when the P peptide was removed (Fig. 6C). However, fragment N_29-374, in the absence of both NT_ARM and CT_ARM, hardly bound to RNA even if P was removed (Fig. 6C). In addition, N protected RNA from degradation by nucleases (Fig. 6D). Moreover, competitive binding experiments further illustrated that the addition of RNA does cause P peptide to dissociate from N (Fig. 6E). Therefore, in the process of N binding to RNA, NT_ARM is the main competitor of P peptide, and CT_ARM will further promote the position exchange of P with NT_ARM. Because the binding sites of P and RNA on N are different, RNA-binding or P-binding switches the closed or open conformations conveniently.

FIG 6.

The mechanism of HPIV3 N binding to RNA. (A) Superposition of HPIV3 N⁰-P with a SeV N protomer in helical NC. SeV N is in wheat, NT_ARM is in cyan, CT_ARM is in dark teal, RNA is in purple, HPIV3 N⁰ is in slate, and HPIV3 P peptide is in hot pink. The inset shows the localization of the SeV three N protomers within the NC. Structure superimpositions were carried out using PyMOL. Using the HPIV3 N⁰-P as the reference, the RMSD value between SeV N-RNA (PDB code 6M7D) and HPIV3 N⁰ is 2.0 Å over all residue coverage. (B) EMSA of different HPIV3 N truncations in N⁰-P with ssRNA. The amino terminus of N and P are fused with the MBP and glutathione S-transferase (GST) tags, respectively. Lane 1 shows ssRNA. Lanes 2 to 6 show ssRNA incubated with N_29-403⁰/P_1-42 (lane 2), N_29-374⁰/P_1-42 (lane 3), N_1-403⁰/P_1-42 (lane 4), N_1-374⁰/P_1-42 (lane 5), and MBP tag and GST-P_1-42 (lane 6). Lanes 7 and 8 show only N_1-403⁰/P_1-42 and N_1-374⁰/P_1-42. The bound and free ssRNAs were labeled on the side. The ssRNA was detected by the FAM signal. The experiments were repeated three times. (C) EMSA of HPIV3 N⁰ truncations with ssRNA. Lane 1, ssRNA only. Lanes 2 to 5, ssRNA incubated with N_29-374⁰ (lane 2), N_29-403⁰ (lane 3), N_29-403⁰/P_22-42 (lane 4), and N_1-403⁰/P_22-42 (lane 5). (D) RNA or N-RNA complex was digested with benzonase nuclease and analyzed by a 20% polyacrylamide-8 M urea gel. N-RNA complex was obtained by incubating N_1-403⁰/P_1-42 complex with ssRNA. The experiments were repeated three times. (E) Affinity competitions of ssRNA and P for N. The coexpressed MBP-N_1-403 and GST-P_22-42 were incubated and bound with the GST beads, and then increasing concentrations of ssRNA were added. Lane M, prestained protein ladder.

P-derived peptide as an antirespirovirus inhibitor.

The presence of high conservation of N and P in the genus Respirovirus could facilitate the development of inhibitors to treat infections. To increase the binding affinity, we tested several mutations of P but failed (data not shown). Therefore, we tried other kinds of mutations. Because the P peptide and the CT_ARM of adjacent N_i-1 protomer within the NC are relatively close in three-dimensional structure (Fig. 6A) and the CT_ARM participates in binding N to stabilize the NC, we hypothesize that the fusion of the CT_ARM to the P peptide might increase the binding affinity to N. To verify this hypothesis, we designed and purified a peptide (named PARM) in which the P_1-42 peptide and the CT_ARM were connected by a (GS)₄ linker and observed an ∼10-fold increased binding affinity to N (K_D, PARM: 4.0 ± 1.3 nM versus P: 43.0 ± 3.9 nM) (Fig. 7A and B). We next assessed the effect of the PARM in the process of N binding to RNA. In the presence of PARM, the binding ability of N to RNA was significantly reduced (Fig. 7C). For the virus, the appropriately medium affinity between N and P is very important for NC assembly, because P needs to depart from N at the right time under the combined action of adjacent N protein and RNA. However, the PARM has greatly strengthened its connection with N and inhibits the binding of N to RNA; that is, the PARM hijacks N. Because of the essential role of N in viral RNA transcription and replication, the PARM with affinity in the low nanomolar range might be used as an antiviral peptide inhibitor targeting N.

FIG 7.

P-derived peptide targets N stronger than P peptide. (A) Schematic representation of the peptide PARM. Sequence alignment of the conserved CT_ARM of the genus Respirovirus. The identical residues are background-shaded red in blue squares, and similar residues are red text in blue squares. The following vrial proteins were used: HPIV1 N (UniProtKB-P26590), SeV N (Q07097), PRV1 N (S5LSJ0), BPIV3 N (P06161), and HPIV3 N (P06159). (B) Measurement of the binding affinity of the PARM and P_1-42 peptide with N_29-374⁰ by MST. The experiments were repeated three times. (C) EMSA of HPIV3 N_1-403/P_1-42 (abbreviated as “P”) and N_1-403/PARM with ssRNA. PARM bound N stronger than P, inhibiting the binding of RNA to N. The concentration of ssRNA was kept constant at 0.1 μM, while the concentration of protein increased gradually.

DISCUSSION

In our structure, the P locates at the CTD domain of N⁰, which is consistent with all published N⁰-P structures. We provided evidence demonstrating that residues 22 to 42 of HPIV3 P are enough to maintain the N at an unassembled conformation. Although we were not able to get the structure of HPIV3 P_1-19 due to the lack of density map, via MST and GST pulldown experiments we proved that the region has little effect on N⁰ binding (Fig. 2). In the structure of the HPIV3 N⁰-P complex, N⁰ is unassembled and in an open conformation, suggesting that the conformational change induced in HPIV3 N upon binding RNA might be more subtle than is the case in MeV, PIV5, EBoV, and HMPV (Fig. 5).

The chaperone functions of P.

Based on our solved structures, when the key hydrophobic residues of P in the interface are mutated to negatively charged ones, the affinity of P mutants to N⁰ would be significantly reduced, indicating the compromised chaperone function of these P mutants. It came to us that we can obtain a P-free N⁰ with the help of this weak chaperone function. By coexpressing with these P mutants, the RNA-free and P-free HPIV3 N⁰ was successfully obtained, and thus the affinity value of N⁰ binding to P was measured at the first time (Fig. 2D and E). Because the interacting residues between P and N⁰ are conserved in the family Paramyxoviridae, this method might be applied to other paramyxoviruses.

The chaperone P is very important for virus replication. Because the RdRp complex cannot recognize the naked RNA and the nascent RNA needs to be protected from degradation by the host, the P carries the N⁰ to the replication center, encapsulating sense RNA and offspring RNA (12). The P and the NT_ARM of the N_i+1 protomer in NC might share a partially overlapping binding site at the CTD domain of N (Fig. 6A). In the process of N⁰-P binding to RNA, N will undergo a conformational change. The HELIX-C, which is a loop and very flexible originally, might be induced by RNA to form a helix to stabilize RNA, and the side chains of the HELIX-N will rearrange (Fig. 4). Therefore, we suggest a reasonable model for the conformational transition of N-binding RNA during viral replication (Fig. 8).

FIG 8.

Schematic model of nucleocapsid (NC) assembly. The L (light blue) and P (pink) proteins form RdRp (RNA-dependent RNA polymerase) complex. The C-terminal domain of P binds to template (solenoid shape), and the N-terminal domain of P binds to N⁰ (wheat). RNA is shown in purple. RdRp, RNA-dependent RNA polymerase.

In this model, RNA is synthesized by the L-P complex and extended in the direction of 5′-end to 3′-end. The N⁰ is carried by the P and binds to RNA from the 5 '-end, but only one N⁰ protein is not enough to bind RNA. The P continues to carry more N⁰ to the replication center. When there are more than two N⁰, the CT_ARM of N_i-1 binds to the CTD domain of N_i, and the NT_ARM of N_i exchanges with the P on N_i-1, and then the P is released. In the participation of RNA, the N_i-1 grasps the RNA through conformational changes of these two motifs (HELIX-C and HELIX-N) at the hinge region and becomes closed (Fig. 8). Once N_i-1 completes the binding of RNA, it will promote the same conformational change of its tightly bound N_i. The next P will begin to be dissociated under the spatial conflict of CT_ARM, and more RNA will be bound. The binding process of N_i will be greatly strengthened by the NT_ARM of N_i+1 when N_i+1 is carried to the RNA synthesis site. This triggers a subsequent chain effect, leading to the extension of the nucleocapsid assembly (Fig. 8). Our model is supported by the following phenomena. First, the P of SeV that belongs to the same genus as HPIV3 is a tetramer (16, 20). Therefore, the P can carry four Ns to the replication site at the initial stage of NC assembly. Second, MeV N bound 6-nt ssRNA to form NC, but one 6-nt ssRNA was distributed in grooves of two Ns in NC, indicating that two Ns are jointly involved in the binding of the ssRNA (31).

The N⁰-P interface as an antiviral target.

The interaction of N⁰ and P is very important for the life cycle of the virus. Disruption of the interaction between N⁰ and P could lead to failure of viral RNA encapsidation and degradation of viral RNA, inhibiting viral growth (12, 40). The importance of the N in viral transcription and replication makes it an attractive target for developing broad-spectrum antiviral inhibitors. In fact, the N of other negative-strand viruses as a drug target has been maturely studied. For example, small-molecule replication inhibitors were developed by targeting the nucleoprotein (NP) of the influenza virus, triggering NP aggregation and inhibiting its nuclear accumulation (41, 42). The amino terminus of P of NiV, RSV, and EBoV binding to N⁰ as a peptide inhibitor can effectively inhibit virus replication in cells (40, 43, 44). The modification of RSV P containing three domains (oligomerization, L-binding, and nucleocapsid binding) has maximum inhibitory activity (45). In response to the N and P interaction of EBoV and RSV, a range of small-molecule inhibitors have been identified by high-throughput screens (46, 47). Thus, the development of small-molecule inhibitors or peptide drugs to interfere with the binding of N and P provides a new strategy for antiviral therapy. Here, the P-derived PARM peptide inhibitor we designed is also aimed at the N⁰-P interface (Fig. 7). It was verified in vitro to interfere with the assembly of N and RNA by hijacking the N. Furthermore, PARM can be fused with a nuclease to design a peptide drug conjugate (PDC), like capsid-nuclease fusion proteins (48, 49). The PDC may not only inhibit the binding of N to RNA but also carry the nuclease to the virus replication center to degrade the virus RNA by using the targeting effect of PARM to N, causing a more in-depth antiviral effect (48, 49). Due to the high conservation of P and N of the genus Respirovirus, the PARM may have broad-spectrum antiviral activity, which requires further in vivo experiments to illustrate.

MATERIALS AND METHODS

Cloning, expression, and purification.

All constructs were derived from reverse-transcribed N gene (GenBank accession number D10025.1) (N45D mutation) and P gene (GenBank accession number M14932.1) of HPIV3 strain NIH 47885 (a gift from Mingzhou Chen). Various fragments of genes for N or P were PCR-amplified and subcloned into pET28a (for N-terminal His₆ tag or dual His₆-MBP fusion tag) or pGEX-6P-1 (for N-terminal GST-tag) vectors via the BamH I/XhoI restriction sites with a tobacco etch virus (TEV) protease cleavage site as a previous publication for recombinant protein expressions (50).

For the N_29-374⁰-P_1-42 chimera connected by a (GS)₄ linker, the N_29-374 and P_1-42 coding sequences were amplified by PCR to generate megaprimers with overlapping sequences. Then, the megaprimers were annealed and extended. The product was subcloned into the bacterial pET28a expression vector, which contained a TEV protease cleavage site following the N-terminal His₆ tag and a stop codon introduced at the 3′-end. Mutations in the P constructs were introduced by site-directed mutagenesis.

Proteins were expressed in E. coli Rosetta (DE3) (Merck Millipore). Cells were grown at 37°C until an optical density at 600 nm (OD₆₀₀) of 0.8 to 1.0. Expression was induced by supplementing 0.2 mM isopropyl-β-d-1-thiogalactopyranoside (IPTG) and was allowed to proceed for 12 h at 16°C. The cells were collected by low-speed centrifugation at 4,000 × g for 10 min and resuspended in lysis buffer (20 mM Bis-tris propane, pH 8.4, 1 M NaCl, 5% [vol/vol] glycerol) supplemented with 1 mM phenylmethylsulfonyl fluoride (PMSF). The cells were lysed by sonication and centrifuged at 40,200 × g for 45 min to separate the cell debris. The supernatant was collected and applied to a nickel-nitrilotriacetic acid (NTA) column or GST affinity column (GE Healthcare). The column was washed three column volumes, and the protein was eluted by using imidazole in concentrations ranging from 20 to 300 mM in buffer (20 mM Bis-tris propane, pH 8.4, 1 M NaCl) or using 10 mM reduced glutathione (GSH). The eluted protein fractions were concentrated with Millipore Amicon Ultra 30-kDa-cutoff spin concentrators to remove imidazole or GSH. The N-terminal His₆ tag was then cleaved by TEV protease overnight at 4°C, and the protein was further purified by anion exchange (Q Sepharose, GE Healthcare). The elution was then applied to a size-exclusion chromatography (Superdex 200 Increase 10/300 GL, GE Healthcare). The major peak fractions were pooled and concentrated as described above. All the protein purification procedures were performed at 4°C.

For the coexpression of the N and P constructs, pET28a-N and pGEX-6P-1-P plasmids were cotransformed into BL21(DE3) cells (Merck Millipore). Protein expression and purification process were similar to that described above. Briefly, after cell growth, expression, and ultrasonic disruption, the supernatant was collected and applied to a GST affinity column (GE Healthcare). Then, the elution was further purified by ion-exchange chromatography followed by SEC in buffer (20 mM Bis-tris propane, pH 8.4, 1 M NaCl).

Only inclusion bodies or aggregations were formed when N proteins were expressed alone. To obtain the homogeneous N⁰ protein, the mutant P_1-42 (L29E), which interacted with N weakly and worked as a compromised chaperone, was fused to N. The pET28a-N-TEV-P_1-42 (L29E) (a His₆ tag at the C terminus) plasmids were constructed and transformed into BL21(DE3) cells. After induction by IPTG, the cells were collected, resuspended in lysis buffer (20 mM Bis-tris propane, pH 8.4, 1 M NaCl, 5% [vol/vol] glycerol, 1 mM PMSF), and then lysed by sonication. The protein was purified in three steps. The cleared supernatant, supplemented with 20 mM imidazole, was loaded to an NTA resin. The resin was washed with three column volumes buffer (20 mM Bis-tris propane, pH 8.4, 1 M NaCl, 20 mM imidazole) and eluted in 5 column volumes elution buffer (20 mM Bis-tris propane, pH 8.4, 1 M NaCl, 400 mM imidazole). After being cleaved by TEV protease, the eluent was applied to an NTA resin, and the flowthrough (contains only N⁰) was concentrated with Millipore Amicon Ultra 30-kDa-cutoff spin concentrators and loaded on ion exchange (Q Sepharose, GE Healthcare), followed by Superdex Increase 200 10/300 GL column. Thus, we purified the N_29-374⁰ and N_29-403⁰. However, the N⁰ truncations containing NT_ARM (residues 1 to 28) led to the formation of inclusion bodies or aggregations, with almost no homogeneous proteins obtained, which could not be purified a step further.

Crystallization and structure determination.

The purified protein was concentrated to about 10 mg/mL. Crystals of the N_29-374⁰-P_1-42 chimeric protein were grown by sitting-drop vapor diffusion (4°C) by mixing 1 μl of protein with 1 μl of reservoir (15% [wt/vol] polyethylene glycol 3350, 0.1 M HEPES, pH 7.0, 0.1 M magnesium chloride, 5 mM nickel (II) chloride) after 2 weeks, but the diffraction was poor. The diffraction qualities were not improved after extensive optimization by various concentrations and species of precipitants, buffer, salt, additives, and detergents. After the long-standing trials, final crystals were grown by seeding at 4°C with a solution containing 20% (wt/vol) polyethylene glycol 6000, 0.1 M cacodylic acid sodium, pH 6.5, 20 mM nickel (II) chloride. All crystals were cryoprotected in mother liquor containing 25% (wt/vol) glycerol and flash-frozen in liquid nitrogen. The diffraction data were collected on Beamlines BL17U1 and BL19U1 at the Shanghai Synchrotron Radiation Facility. All of the data were processed using HKL2000 (51).

The structure was determined by molecular replacement (MR) using BALBES (52) in the CCP4, and the best solution stopped at R_work/R_free values of 0.374/0.450 using MeV N (PDB code 6H5Q) (31) as a search model. The model was then manually built using the program COOT (53) and refined using REFMAC5 (54) in iterative cycles. The structure was finally refined to 3.23 Å in the space group I422, with an R_work of 21.9% and an R_free of 23.8%. Coordinates and structure factor amplitudes have been deposited in the PDB under accession number 7EV8. The data collection and processing statistics are summarized in Table 2. To confirm the conformation of NTD relative to CTD, we split the N structure of 5E4V (7) into two domains, NTD and CTD, and deleted the flexible loop of NTD. The CTD domain structures both with and without the P structure were tried. The modified structures were used as search models, but the results of MR from BALBES were still very bad. We then used PHASER-MR (full-featured) in PHENIX (55) to MR and set ensemble1 with NTD and ensemble2 with CTD, and the strategy was set to first search ensemble1 and then search ensemble2. The initial solution was seemingly rational, with log-likelihood gain/translation function Z-score (LLG/TFZ) values of 159/11.9, and the primitive density map had a relatively clear outline. We manually adjusted the initial structure in COOT and optimized it with REFMAC5; the R_work/R_free values were 0.413/0.456. After more than 90 cycles of manual rebuilding and refining, the best solution stopped at R_work/R_free values of 0.217/0.256. All structural figures were prepared by using the PyMOL program (56).

N⁰-P heterocomplex interaction experiments.

To detect the interaction between different P constructs and N_29-374⁰, all of the P constructs were coexpressed with N_29-374. The cells were lysed by sonication in buffer (20 mM HEPES, pH 7.5, 300 mM NaCl, 5% [vol/vol] glycerol, 1 mM PMSF). The cleared lysates were incubated with glutathione beads for 2 h at 4°C. The beads were washed three times and then eluted with 10 mM reduced glutathione. Protein elution was analyzed by SDS-PAGE.

For the dissociation of N_29-374⁰-P_1-42 in different solutions, pET28a-N_29-374 and pGEX-6P-1-P_1-42 were cotransformed and coexpressed, and then the complex was bound to glutathione beads (Thermo Scientific). To test the stability of the complex, the beads were eluted with different buffers: 20 mM HEPES, pH 7.5 supplemented with one of the following: NaCl at 0.5, 1, 2, or 3 M; NP-40 at 0.1, 0.5, 1, or 2%; or urea at 0.5, 1, 2, or 4 M. The supernatant of the released proteins was analyzed by SDS-PAGE.

For the competition between RNA and P for N, about 200 nM MBP-N_1-403 and GST-P_22-42 complex was bound to glutathione beads in a 500-μl mixture buffer (containing 20 mM HEPES, pH 7.5, 100 mM NaCl, 2 mM EDTA). After three times washes with the mixture buffer, the beads were incubated with the increasing concentrations (0, 20, 100, 200, 500, and 1,000 nM) of single-stranded RNA (sequence: 5′-[6-carboxyfluorescein (FAM)]-AAAAAAAAAAAAAAAAAAAAAAAAAAAAAA-3′; the poly(A) sequence was used because N is not specific to RNA sequence) in a 200-μl mixture buffer overnight at 4°C, separately. The beads were washed three times and then analyzed by SDS-PAGE.

EMSAs.

About 0.2 μM ssRNA (sequence is shown above) was incubated with 2 μM different MBP-N constructs and GST-P_1-42 complex in a 20-μl reaction mixture (containing 20 mM HEPES, pH 7.5, 100 mM NaCl, 2 mM EDTA, 6% [vol/vol] glycerol) overnight at room temperature, and 4 μl of the mixture was then separated on a 7% native polyacrylamide gel (80:1) in 0.5 × TBE buffer (45 mM Tris, pH 8.0, 45 mM boric acid, 1 mM EDTA). The FAM-ssRNA signal was detected by 488-nm excitation.

Benzonase digestions.

About 0.2 μM ssRNA (used in EMSA) was incubated with 2 μM MBP-N_1-403 and GST-P_1-42 complex overnight at room temperature to form an N-RNA complex. The ssRNA and N-RNA complex were digested by benzonase (final concentration, 0.01 mg/mL) for 1 h at 30°C. Digested samples were visualized by running a 20% polyacrylamide–8 M urea gel, and the FAM-ssRNA signal was detected by 488-nm excitation.

Microscale thermophoresis.

The binding affinities between P_1-42 and N_29-374⁰ were measured using the Monolith NT.115 (NanoTemper Technologies). The N_29-374⁰ protein was fluorescently labeled according to the manufacturer's procedure of RED fluorescent dye NT-647. Wild-type and mutated GST-P_1-42 proteins were mixed with labeled N_29-374⁰ in a buffer containing 20 mM HEPES, pH 7.3, 150 mM NaCl, 1% (wt/vol) bovine serum albumin (BSA). After incubation for 10 min, the samples were loaded into silica capillaries (NanoTemper Technologies), and temperature-induced fluorescence changes were measured at 24°C by using 20% light-emitting diode (LED) power and 40% MST power. Data analysis was performed by using NTAnalysis software (NanoTemper Technologies). There is 95% confidence, and the K_d value is within the given range. The values of faction bound are transformed from the original MST data “Fnorm.” The experiments were repeated three times.

Small-angle X-ray scattering.

Small-angle X-ray scattering (SAXS) data were collected at Beamline BL19U2 at the Shanghai Synchrotron Radiation Facility (SSRF) following previously published methods (57). N_29-374⁰-P_1-42 was subjected to SEC in the buffer (20 mM HEPES, pH 7.3, 500 mM NaCl, 2 mM dithiothreitol). The scattering data from the buffer alone were measured before and after each sample measurement. The average of the scattering data was used for background subtraction using RAW (58). The ATSAS packages were used for subsequent data processing (59). Ab initio low-resolution bead models of the N_29-374⁰-P_1-42 were generated from the scattering curves using the program DAMMIN (60). Twenty independent models were aligned, averaged, and filtered with SUPCOMB, DAMFILT, DAMSEL, and DAMAVER (61). The fitting of the theoretical scattering curves between models and the experimental data were calculated using CRYSOL (62). All figures were generated with PyMOL (56). The final statistics for data collection and scattering-derived parameters are summarized in Table 1.

Data availability.

Final refined coordinates and structure factors have been deposited in the Protein Data Bank (PDB) under accession code 7EV8.