Structural Insight into the Interaction of Sendai Virus C Protein with Alix To Stimulate Viral Budding

ABSTRACT

Sendai virus (SeV), belonging to the Respirovirus genus of the family Paramyxoviridae, harbors an accessory protein, named C protein, which facilitates viral pathogenicity in mice. In addition, the C protein is known to stimulate the budding of virus-like particles by binding to the host ALG-2 interacting protein X (Alix), a component of the endosomal sorting complexes required for transport (ESCRT) machinery. However, small interfering RNA (siRNA)-mediated gene knockdown studies suggested that neither Alix nor C protein is related to SeV budding. In the present study, we determined the crystal structure of a complex comprising the C-terminal half of the C protein (Y3) and the Bro1 domain of Alix at a resolution of 2.2 Å to investigate the role of the complex in SeV budding. The structure revealed that a novel consensus sequence, LXXW, which is conserved among Respirovirus C proteins, is important for Alix binding. SeV possessing a mutated C protein with reduced Alix-binding affinity showed impaired virus production, which correlated with the binding affinity. Infectivity analysis showed a 160-fold reduction at 12 h postinfection compared with nonmutated virus, while C protein competes with CHMP4, one subunit of the ESCRT-III complex, for binding to Alix. All together, these results highlight the critical role of C protein in SeV budding.

IMPORTANCE Human parainfluenza virus type I (hPIV1) is a respiratory pathogen affecting young children, immunocompromised patients, and the elderly, with no available vaccines or antiviral drugs. Sendai virus (SeV), a murine counterpart of hPIV1, has been studied extensively to determine the molecular and biological properties of hPIV1. These viruses possess a multifunctional accessory protein, C protein, which is essential for stimulating viral reproduction, but its role in budding remains controversial. In the present study, the crystal structure of the C-terminal half of the SeV C protein associated with the Bro1 domain of Alix, a component of cell membrane modulating machinery ESCRT, was elucidated. Based on the structure, we designed mutant C proteins with different binding affinities to Alix and showed that the interaction between C and Alix is vital for viral budding. These findings provide new insights into the development of new antiviral drugs against hPIV1.

INTRODUCTION

Acute respiratory infection, caused primarily by virus infection, is one of the leading causes of mortality in children under the age of 5, especially in developing countries (1). Although there are vaccines and effective antivirals for infection with certain viruses, such as influenza virus, there are none for others, such as parainfluenza virus. Human parainfluenza virus type I (hPIV1), which belongs to the Respirovirus genus in the family Paramyxoviridae, is one of the major causes of croup in the younger population and poses a threat of occasional outbreak in immunocompromised patients and the elderly. Therefore, antiviral agents against hPIV1 remain needed. A recent study has highlighted the challenges in vaccine development against hPIV1 (2). On the other hand, it is important to note that hPIV1 is significantly attenuated by introducing the loss-of-function mutation into its C protein (3), whose exact function remains to be elucidated.

Sendai virus (SeV), a pneumotropic virus of rodents and a murine counterpart of hPIV1, is the most extensively studied member of the Respirovirus genus to determine the molecular and biological properties of hPIV1 and to develop an effective antiviral treatment against it (4–6). SeV also expresses the C proteins, which are translated from the P and V mRNAs in a coding frame different from that of the P and V proteins. The C proteins comprise a nested set of four independently initiated and carboxy-coterminal proteins, namely C [amino acids (aa) 1 to 204], Y1 (aa 24 to 204), Y2 (aa 30 to 204), and C′ (with a 11-aa addition to the N terminus of C), where C is the major protein expressed in infected cells (7, 8). The C proteins are categorized as nonessential accessory proteins but contribute greatly to virus replication in vitro and are indispensable for the in vivo multiplication and pathogenesis of the infection they cause (9). In fact, a mutated virus lacking all four components of the C proteins generated by multiple-site mutagenesis was reported to rapidly produce the C protein-producing revertant viruses during serial passages in embryonated chicken eggs (10).

The C protein inhibits the signal transduction of interferons (IFNs) by associating with the signal transducer and activator of transcription 1 (STAT1) (11, 12). A previous study has demonstrated that the C-terminal half of C (aa 99 to 204), designated Y3, can bind to the dimeric structure formed by two N-terminal domains of STAT1 (STAT1ND), thus elucidating the mechanism underlying the inhibition of IFN-γ signal transduction (13). In addition, it has been suggested that Y3 can bind to the heterodimeric structure formed by STAT1ND and STAT2ND, thereby inhibiting IFN-α/β signal transduction (14). The C protein also regulates viral RNA synthesis to suppress the production of the IFN-inducing abnormal RNA species (15–18) and to control viral genome polarity (19, 20), possibly by interacting with the L protein, the viral RNA polymerase. In addition, the C protein regulates the formation of viral particles (9, 21) through the interaction with the ALG-2 interacting protein X (Alix), which is a component of a membrane pinching machinery, endosomal sorting complex required for transport (ESCRT) (22). The interaction of C or C′ with Alix, unlike that of Y1 and Y2, which lack the membrane-targeting sequence, has been shown to facilitate the formation of virus-like particles (VLPs) (23, 24) and virus production (22). However, Gosselin-Grenet et al. showed that Alix, vacuolar protein sorting-associated protein 4 (VPS4), an essential ATPase working in the ESCRT system (25), and C protein are not associated with SeV budding (26). Therefore, there remains ambiguity regarding the role of the interaction between Alix and C protein and the involvement of ESCRT in SeV budding. Although generation of recombinant SeV possessing mutated C proteins that lack Alix-binding ability can provide the answers, this may also cause a mutation in the overlapped P and V proteins. Therefore, elucidation of the role of C in SeV budding is much more challenging.

In the present study, we aimed to determine the role of C protein based on the elucidated crystal structure of a complex between Y3 and the Bro1 domain of Alix. Based on the structure, mutations causing the amino acid substitution in the C protein that affects the Alix-binding ability but no alterations in the amino acid sequences of P and V proteins were designed. Recombinant SeV possessing the mutated C proteins makes it possible to understand the role of the association between C protein and Alix during SeV budding.

RESULTS

Association of Y3 with the Bro1 domain.

A previous study showed that the C protein can bind to an N-terminal region (aa 1 to 423) of Alix containing the Bro1 domain (Fig. 1A) (27). Y3, the C-terminal half of the C protein (Fig. 1B), may be suitable for cocrystallization with Alix, since it lacks the N-terminal region, which is predicted to be disordered by use of the program XtalPred (28). To determine whether Y3 binds to the Bro1 domain, Y3 and Bro1 were prepared using the Escherichia coli expression system (Fig. 1C). Size exclusion chromatography analysis showed that Y3 coeluted with the Bro1 domain at approximately 11 ml of elution volume, whereas Y3 alone eluted at approximately 14.5 ml of elution volume (Fig. 1D), suggesting that Y3 contains a region responsible for the association with the Bro1 domain. Y3 was therefore used for further investigations.

FIG 1.

Binding analysis between C and Alix. (A) Linear representation of the domains of human Alix. PRR, proline-rich region. Amino acid residue numbers are shown. (B) Schematic diagram of constructs of C′, C, Y1, Y2, and Y3. Amino acids are numbered with the starting amino acid of the C protein, the major component, as 1. (C) SDS-PAGE analysis of Bro1 and Y3 used for the following size exclusion chromatography. (D) Analytical size exclusion chromatograms of Bro1 in the presence (gray line) and absence (black line) of Y3. Bro1 (20 μM) was preincubated with or without an equimolar concentration of Y3 for 30 min, followed by size exclusion chromatography. A chromatogram of Y3 alone is shown as a dashed line. Proteins in fractions eluted from the size exclusion chromatograms were separated by SDS-PAGE, followed by Coomassie blue staining.

Crystal structure of the Y3:Bro1 complex.

The crystal structure of the Y3:Bro1 complex was determined at a resolution of 2.2 Å (Fig. 2A). An asymmetric unit in the crystal is composed of one heterodimer generated by Y3 and Bro1. Almost all residues in the Bro1 domain were observed in the electron density map. However, the electron density of the residues from Asp¹⁰⁰ to Ser¹⁰⁸, which is part of the loop between the β1- and β2-strands protruding into the solvent region, is not clearly observed. Meanwhile, the electron density of the residues in Y3 is clear, except for the residues from Arg¹⁶¹ to Lys¹⁶⁸ in the loop between the α4- and α5-helices. In addition, the electron density of the eight successive histidine residues fused at the N-terminal residue of Y3 (Met⁹⁹) was clearly observed. These histidine residues, which are the N-terminal part of the α1-helix, interact with an adjacent Y3:Bro1 complex in the crystal, suggesting that the intermolecular interaction helped to crystallize the Y3:Bro1 complex.

FIG 2.

Crystal structure of Bro1-bound Y3. (A) Heterodimeric structure of Y3 and Bro1 shown in a ribbon representation. Y3 and Bro1 are colored in red and blue, respectively. (B) Superposition of Bro1-bound Y3 (colored in green) on STAT1ND-bound Y3 (colored in white). Structural superposition was performed using the “align” command in PyMOL (63). (C) Contact areas on the molecular surface of Y3. The Bro1 contact area is colored in magenta, while the STAT1ND contact area is colored in green.

The Bro1-associated Y3 can be well superimposed on the STAT1ND-associated Y3 except for the α4- and α5-helices, with a root mean square deviation (RMSD) of 0.431 Å for 79 Cα atoms, indicating that Y3 does not change conformation on binding to Alix or STAT1 (Fig. 2B). The α4- and α5-helices of Y3, which have no contact with the Bro1 domain, are likely to be pushed to the current positions by the N-terminal part of the α1-helix of Y3. Mapping of the Bro1 and STAT1ND contact areas on the molecular surface of Y3 shows no overlapping (Fig. 2C). The contact area for the Bro1 domain (620 Ų) is significantly smaller than that for the STAT1ND homodimer (1,150 Ų), suggesting that the C protein preferably binds to the STAT1 dimer rather than Alix.

To investigate this hypothesis, we conducted an affinity analysis using fluorescence resonance energy transfer (FRET). As described previously (13), if enhanced cyan fluorescent protein (ECFP) fused to one protein is close to the enhanced yellow fluorescent protein (EYFP) fused to its binding partner (within approximatly100 Å), emission from ECFP at 475 nm, which is generated after excitation of the ECFP chromophore at 435 nm, is decreased, while emission from EYFP at 527 nm is increased. Prior to affinity analysis, we purified an ECFP-fused Bro1 and an EYFP-fused Y3 (Fig. 3A), which was followed by gel filtration chromatography analysis to confirm the complex formation. It was found that ECFP-fused Bro1 can bind to EYFP-fused Y3, while no association was observed between ECFP-fused Bro1 and unfused EYFP (Fig. 3B and C). For EYFP-fused Y3 alone, a peak of the EYFP-fused Y3 oligomer was observed at approximately 11 ml of elution, near the exclusion limit of the chromatography column (Fig. 3B). Since the peak area was unchanged in the presence of ECFP-Bro1, ECFP-Bro1 is thought to bind to EYFP-fused Y3, forming a monomer. By using the increase in FRET as an indicator, a titration experiment in which EYFP-fused Y3 was successively added to a solution of ECFP-fused Bro1 at the known concentration was conducted. FRET was increased when using EYFP-fused Y3, while no FRET was observed when using unfused EYFP (Fig. 3D and E). In contrast to our expectation, affinity analysis showed that the dissociation constant (K_d) between Y3 and the Bro1 domain was 0.31 ± 0.03 μM (Fig. 3F), which is comparable to that between the Y3 and the STAT1ND dimer (K_d1 = 0.71 ± 0.24 μM and K_d2 = 0.44 ± 0.08 μM [K_d1 and K_d2 are the dissociation constants for the first and the second EYFP-fused Y3 for each of the two Y3-binding sites in the ECFP-fused STAT1ND dimer, respectively]) (13).

FIG 3.

Affinity analysis. (A) SDS-PAGE analysis of ECFP-fused Bro1, EYFP-fused Y3, and unfused EYFP used for size exclusion chromatography. (B and C) Analytical size exclusion chromatograms of ECFP-fused Bro1 in the presence of EYFP-fused Y3 (B) or unfused EYFP (C). Bro1 (5 μM) was preincubated with (gray line) or without (black line) unfused EYFP or EYFP-fused Y3 (10 μM) for 30 min, followed by size exclusion chromatography. Unfused EYFP or EYFP-fused Y3 alone is shown as a dashed line. Proteins in fractions eluted from the size exclusion chromatograms were separated by SDS-PAGE, followed by Coomassie blue staining. (D and E) Fluorescence emission spectra of ECFP-fused Bro1 mixed with EYFP-fused Y3 (D) or unfused EYFP (E). The spectra of ECFP-fused Bro1 (0.7 μM) in the absence of EYFP-fused Y3 or unfused EYFP, which was excited at 435 nm, are shown by a black line, while the spectra in the presence of EYFP-fused Y3 or unfused EYFP (0.13, 0.27, 0.53, 1.1, or 2.1 μM) are shown by a gray line. Two arrows indicate decreased intensity at 475 nm and increased intensity at 527 nm. RFU, relative fluorescence units. (F) Affinity analysis between EYFP-fused Y3 and ECFP-fused Bro1. The quenched fluorescence intensity (ΔI) of ECFP-fused Bro1 at 475 nm, which is caused by the association of EYFP-fused Y3, is divided by the total concentration of ECFP-fused Bro1, and each value is plotted against the concentration of free EYFP-fused Y3. The gray curve represents the best fit to the data.

Interaction between Y3 and Bro1.

The crystal structure of the Y3:Bro1 complex shows that Y3 associates with a hydrophobic patch on the molecular surface of the Bro1 domain in Alix (Fig. 4A). Specifically, Leu¹²² and Trp¹²⁵ of Y3 interact with a hydrophobic patch formed by Phe¹⁹⁹, Met²⁰⁸, Ile²¹², Leu²¹⁶, and Leu³³⁷ of Bro1 (Fig. 4B). In addition, the indole nitrogen of Trp¹²⁵ in Y3 forms a hydrogen bond with the Oδ1 atom of Asp¹⁴³ in the Bro1 domain, the Oδ2 atom of which forms a salt bridge with the Nζ atom of intramolecular Lys²⁰². Glu¹¹⁴ and Glu¹¹⁵ of Y3, which are positioned at the N-terminal part of the α2-helix in Y3, are likely to form salt bridges with Lys²⁰⁹ and Lys²¹⁵ of the Bro1 domain, respectively.

FIG 4.

Comparison between the binding interface of the Y3:Bro1 complex and that of the CHMP4B:Bro1 complex. (A) The C-terminal helix of CHMP4B (aa 205 to 224) (colored in green; PDB code 3C3Q) is superimposed on the Y3:Bro1 complex (colored in red and white, respectively). (B and C) Binding interface between Bro1 and Y3 (B) or between Bro1 and CHMP4B (C). The dotted lines represent the hydrogen bonds. (D) Sequence alignment of the C-terminal helix of CHMP4A–C and the corresponding regions of SeV C. Conserved residues are colored in red. Sequence alignment was performed by using ClustalW (64).

Similar intermolecular interactions are observed in a complex between one of the charged multivesicular body protein 4 (CHMP4) isoforms, CHMP4B, and Bro1 (Fig. 4A and C). CHMP4, which consists of three isoforms (CHMP4A to -C [CHMP4A–C]), is responsible for the membrane remodeling in the ESCRT system (29). Particularly, residues corresponding to Glu¹¹⁴, Leu¹²², and Trp¹²⁵ in Y3 are conserved in CHMP4A–C (Fig. 4D).

Affinity analysis.

The residues important for the interaction between the C protein and Alix were investigated using the various mutants created by introducing a point mutation into EYFP-fused Y3. Each mutant, in which Leu¹²² or Trp¹²⁵ of Y3 was replaced by Ala, showed significantly reduced affinity to Bro1 (Table 1). On the other hand, the amino acid sequences of the P and V proteins remained unaltered when the mutation, in which Trp¹²⁵ of the C protein is replaced with Leu or Ser, was introduced into the SeV genome. Like the W125A mutant of Y3, the W125L and W125S mutants significantly reduced the affinity to the Bro1 domain (Table 1). When Thr¹¹⁹ of Y3 was replaced by Ala, the K_d value was increased by approximately 16 times (5.1 ± 0.5 μM) (Table 1). The residues corresponding to Thr¹¹⁹ in the C protein are Leu, Met, and Ile in CHMP4A–C, respectively (Fig. 4D) (29). Introduction of the hydrophobic residue at the Thr¹¹⁹ position in C may increase the affinity to the Bro1 domain. The T119I mutation in the C protein results in no alteration in the amino acid sequences of the P and V proteins. However, affinity analysis showed that Y3 and the T119I mutant have comparable affinities to the Bro1 domain (Table 1), suggesting that the Cγ atom of Thr¹¹⁹ creates a hydrophobic interaction with Bro1.

TABLE 1.

Affinity parameters of Y3 and its mutants

EYFP-Y3 variant	ECFP-Bro1 Kd (μM)	ECFP-STAT1NDa
		Kd1 (μM)	Kd2 (μM)
Wild type	0.31 ± 0.03	0.71 ± 0.24	0.44 ± 0.08
E114A	6.7 ± 0.6	0.10 ± 0.02	0.14 ± 0.01
T119A	5.1 ± 0.5	0.31 ± 0.14	0.46 ± 0.04
T119I	0.61 ± 0.05	0.36 ± 0.09	0.15 ± 0.02
Y121A	6.4 ± 0.3	0.19 ± 0.07	0.52 ± 0.05
L122A	230 ± 10	0.12 ± 0.05	0.22 ± 0.02
W125A	200 ± 10	0.19 ± 0.05	0.18 ± 0.01
W125L	220 ± 10	1.5 ± 0.4	0.12 ± 0.02
W125S	140 ± 10	0.67 ± 0.26	0.04 ± 0.01
E128A	3.2 ± 0.2	0.68 ± 0.20	0.87 ± 0.07
M150Ab	0.86 ± 0.07

^aThe dissociation constants for the first and second EYFP-Y3 to each of the two Y3-binding sites in the ECFP-STAT1ND homodimer were defined as K_d1 and K_d2, respectively.

^bThe dissociation constant of the M150A variant for the ECFP-STAT1ND homodimer, 15 μM, was obtained as a simple 1:1 interaction.

Each of the Y3 mutants in which Glu¹¹⁴, Tyr¹²¹, or Glu¹²⁸ was replaced by Ala showed reduced affinity to Bro1 (Table 1), illustrating the contributions of the hydrophilic interactions to forming the complex between Y3 and Bro1. However, based on the change in binding affinity of the mutants, the contribution of the hydrophilic residues to the binding to Bro1 seems to be smaller than that of the hydrophobic ones (Leu¹²² and Trp¹²⁵).

Alix is a binding partner of the C proteins from hPIV1 (30) and SeV. To determine whether Leu¹²² and Trp¹²⁵ of hPIV1 C, which correspond to Leu¹²² and Trp¹²⁵ of SeV C, respectively, are important for Alix binding, size exclusion chromatography analysis was conducted using EYFP-fused hPIV1 C and ECFP-fused Bro1. By using nickel affinity chromatography, the wild type and its mutants (L122A and W125A) of EYFP-fused hPIV1 C were purified together with impurities seen on the gel (Fig. 5A). Size exclusion chromatography was performed without removing any impurities, since these fusion proteins were degraded during an additional purification. When EYFP-fused hPIV1 C protein alone was applied to size exclusion chromatography, only a peak corresponding to the impurities was observed (Fig. 5B). Meanwhile, the elution peak of the ECFP-fused Bro1 was found at approximately 11 ml of elution volume (Fig. 5B). In the case of the EYFP-fused hPIV1 C mixed with ECFP-fused Bro1, a new peak emerged at an elution volume of approximately 10 ml (Fig. 5B). SDS-PAGE analysis showed that the new peak contained both proteins, indicating that EYFP-fused hPIV1 C protein interacts with ECFP-fused Bro1. The hPIV1 C protein seems to be an intrinsically disordered protein, whose binding to a partner leads to a structural transition to a folded form, as described previously (31).

FIG 5.

Interaction analysis between Bro1 and hPIV1 C. (A) SDS-PAGE analysis of ECFP-fused Bro1 and EYFP-fused hPIV1 C, and EYFP used for the size exclusion chromatography. (B to D) Analytical size exclusion chromatograms of ECFP-fused Bro1 in the presence (gray line) and absence (black line) of EYFP-fused hPIV1 C (B) and the L122A (C) and W125A (D) mutants. Bro1 (5 μM) was preincubated with or without the EYFP-fused hPIV1 C (10 μM) for 30 min, followed by size exclusion chromatography. EYFP-fused hPIV1 C alone is shown as a dashed line. Proteins in fractions eluted from the size exclusion chromatograms were separated by SDS-PAGE, followed by Coomassie blue staining.

When using the hPIV1 C protein mutant in which Leu¹²² or Trp¹²⁵ is replaced by Ala, a slight peak of the mutant alone was observed (Fig. 5C and D), unlike in the case of the wild-type hPIV1 C. When using a mixture of ECFP-fused Bro1 and the L122A or W125A mutant hPIV1 C, a new peak was not observed (Fig. 5C and D), indicating that Leu¹²² and Trp¹²⁵ of hPIV1 C are important for the binding to Alix. In addition, the novel consensus sequence, LXXW, which is necessary for the association of the C proteins from SeV and hPIV1 with Alix, is conserved among the known Respirovirus C proteins on the basis of the sequence alignment (Fig. 6).

FIG 6.

Sequence alignment among the Respirovirus C proteins. Amino acids are numbered with the starting amino acid of the C protein as 1. Identical residues are indicated by asterisks. Residues in the other Respirovirus C proteins corresponding to Leu¹²² or Trp¹²⁵ of SeV C are indicated by light blue highlighting. The membrane targeting sequence in SeV C is underlined. Sequence alignment was performed by using ClustalW (64).

All Y3 mutants described above can bind to STAT1ND with similar or slightly stronger affinities than the wild type. On the other hand, the replacement of Met¹⁵⁰ with Ala in C has been reported to cause decreased affinity to STAT1 (13). A Y3 mutant in which Met¹⁵⁰ was replaced by Ala hardly interacted with STAT1ND, while retaining a strong affinity to the Bro1 domain (Table 1). It is concluded that a point mutation introduced on the molecular surface of Y3, which contacts one target, has little to no effect on the affinity to the other target. A coimmunoprecipitation assay was conducted to confirm the effect of mutations introduced into the full-length C protein on the binding ability to Alix. A FLAG-tagged mutant of C (T119I, W125L, or W125S) was transiently coexpressed with hemagglutinin (HA)-tagged Alix in 293T cells, and the proteins were immunoprecipitated with anti-FLAG antibody (Fig. 7A). HA-tagged Alix was found to be coprecipitated with the wild-type C protein and the T119I mutant but not with the W125S or W125L mutant.

FIG 7.

Coimmunoprecipitation analysis between C and Alix or STAT1. (A) Coimmunoprecipitation of FLAG-tagged C with Alix. 293T cells were transfected with an expression plasmid for HA-tagged Alix together with that for FLAG-tagged C. Proteins in the cell lysates were immunoprecipitated with an anti-FLAG antibody and analyzed by Western blotting using anti-HA and anti-FLAG antibodies. (B) Coimmunoprecipitation of C in SeV-infected cells with Alix. LLC-MK2 cells were transfected with the expression vector for HA-tagged Alix, followed by infection with SeV. Proteins in the cell lysates were immunoprecipitated with anti-HA antibody and analyzed by Western blotting. (C) Coimmunoprecipitation of C in SeV-infected cells with STAT1. LLC-MK2 cells were infected with SeV. Proteins in the cell lysates were immunoprecipitated with anti-STAT1 antibody and analyzed by Western blotting.

We then aimed to generate SeV with mutant C proteins. Although SeV with the T191I or W125L mutant C protein was generated, we could not obtain SeV possessing the W125S mutant C protein. To confirm the interaction of C with Alix in SeV-infected cells, LLC-MK2 cells, which were transfected with the expressing plasmid for the HA-tagged Alix, were infected with the parental or mutant SeV at an input multiplicity of infection (MOI) of 15. At 48 h postinfection, cells were lysed, and HA-tagged Alix was immunoprecipitated with anti-HA antibody. Western blot analysis showed that the wild-type and T119I mutant C proteins were coimmunoprecipitated with HA-tagged Alix, but the W125L mutant C protein was not (Fig. 7B). As expected from the results of affinity analysis, wild-type, T119I mutant, or W125L mutant C was coimmunoprecipitated with endogenous STAT1 using an anti-STAT1 antibody (Fig. 7C).

Effect of mutant C protein on the signal transduction of IFN-α/β.

IFN-β, which has antiviral activity, is encoded by the IFNB1 gene in rhesus monkey. To determine the relative expression level of IFNB1 mRNA, LLC-MK2 cells were infected with one of the recombinant SeVs at an input MOI of 10, and the total RNA was extracted at 19 h postinfection, followed by quantitative real-time PCR (qRT-PCR) analysis (Fig. 8A). The expression level of IFNB1 mRNA in the cells after infection with wild-type or T119I mutant SeV was as low as that in the mock-infected cells, suggesting that the antiviral state was not triggered by the infection. Although the expression level of IFNB1 mRNA increased after infection with W125L mutant SeV in comparison to that in mock-infected cells, the increase was only about 1.6-fold.

FIG 8.

Suppression of antiviral response in SeV-infected cells. (A) Expression of IFN-β gene (IFNB1) in LLC-MK2 cells triggered by mock or SeV infection. At 19 h postinfection, total RNA was extracted from the cells, and qRT-PCR was performed to estimate the relative amounts of IFNB1 mRNA. Amounts of GAPDH mRNA were used for calibration, and the value of the mock-infected sample was set to 1. (B) Reporter assay for IFN-β signal transduction in SeV-infected LLC-MK2 cells. To estimate the strength of the response to IFN-β, LLC-MK2 cells were transfected with pISRE-EGFP, followed by infection with the indicated SeV. At 7 h postinfection, IFN-β (1,000 U ml⁻¹) was added to the culture medium. The fluorescence intensity of EGFP in each cell is shown in the graph. The value in the IFN-β-treated, mock-infected cells was set to 1. (C) Expression of IFIT2, IFIT5, and OAS1 genes in LLC-MK2 cells triggered by mock or SeV infection. Total RNA was extracted from the cell pellets, followed by qRT-PCR to estimate the relative amounts of mRNA of the IFIT2, IFIT5, and OAS1 genes. The value in the mock-infected cells was set to 1. (D) Relative expression levels of EGFP in LLC-MK2 cells infected with SeV and rVSV-EGFP. Mock- or SeV-infected LLC-MK2 cells were superinfected with rVSV-EGFP at 16 h postinfection. As a negative control, mock-infected cells were treated with IFN-α before the superinfection. The expression levels of EGFP and SeV-derived proteins were estimated by Western blotting. The relative expression levels of EGFP among samples are shown in the bar graph. Signal intensity of actin was used for the calibration among samples. The EGFP expression level in mock-infected cells was set to 1. All measured values in the bar graph are represented by black triangles. An error bar indicates the standard deviation, which was calculated from the data of at least three experiments. The P value was calculated on the basis of Student’s t test.

The strength of the signal transduction by IFN-β in the cells can be estimated by the activity of the interferon-sensitive response element (ISRE) promoter (32). In this study, LLC-MK2 cells were transfected with pISRE-EGFP, a reporter plasmid for measuring ISRE promoter activity, followed by infection with SeV at an input MOI of 15. At 12 h after stimulation with IFN-β, the cell was lysed, and the fluorescent intensity was measured to estimate the expression level of EGFP in the cells. The signal transduction of IFN-β in the cells infected with each recombinant SeV was significantly suppressed in comparison to that in mock-infected cells (Fig. 8B). Especially, T119I and W125L mutations in the C protein did not significantly alter the signal transduction, indicating the association of the mutated C proteins with the STAT1:STAT2 heterodimer.

IFIT2 (33), IFIT5 (34), and OAS1 (35) genes are part of the IFN-stimulated genes involved in host protection against virus pathogenesis. To estimate the relative mRNA expression levels in response to infection with SeV at an input MOI of 10, total RNA was extracted from the cells at 19 h postinfection, followed by qRT-PCR analysis (Fig. 8C). The expression of IFIT2 mRNA was elevated 3-fold by infection with wild-type SeV, although it was much lower than that in the IFN-β-stimulated mock-infected cells. Similarly, the expression levels of IFIT5 and OAS1 mRNAs in the cells infected with wild-type SeV were not significantly changed compared with those in the mock-infected cells, but they were greatly lower than those in the IFN-β-stimulated mock-infected cells. In this case also, T119I and W125L mutations in the C protein did not significantly alter the expression levels of IFIT2, IFIT5, and OAS1 mRNAs in response to infection with SeV.

Additionally, the antiviral state was estimated by superinfection with recombinant vesicular stomatitis virus expressing EGFP (rVSV-EGFP), and no significant differences were found in the expression level of EGFP triggered by mock infection or infection with wild-type, T119I mutant, or W125L mutant SeV (Fig. 8D), suggesting that the antiviral state is not triggered by SeV infection. These recombinant SeVs therefore seem to be appropriate for analyzing the role of the association between C and Alix in the budding of SeV.

Virus particle formation by the recombinant SeV possessing the C mutants.

By using recombinant SeV possessing the C mutants, the relationship between Alix-binding ability and viral infectivity was investigated. LLC-MK2 cells were infected with SeV at an input MOI of 10, and a part of the culture medium was collected at a given time to measure the infectivity and the hemagglutination activity. Infectivity of W125L mutant SeV at 12 h postinfection was significantly lower than that of wild-type SeV by approximately 160-fold and that of T119I mutant SeV by approximately 60-fold (Fig. 9A). However, infectivity of W125L mutant SeV was increased at 24 h and 36 h postinfection, and the ratio to that of wild-type or T119I mutant SeV was reduced to approximately 5-fold. The amounts of SeV nucleoprotein (N) in the culture medium at 24 h postinfection were analyzed using Western blotting (Fig. 9B), indicating that the release of virus after infection with W125L mutant SeV was lower than that after infection with parental or T119I mutant SeV. At 36 h postinfection, the amounts of released N protein in the culture medium of cells infected with W125L mutant SeV were rarely different from those of the parent or T119I mutant SeV, suggesting that cellular N proteins may partially leak out of dead or dying cells into the medium (Fig. 9A and B). In addition, the hemagglutination titers of the culture medium were found to be positively correlated with infectivity (Fig. 9C).

FIG 9.

Virus growth of SeV mutants and release of virus particles. Infectivity (A) and hemagglutination activity (C) of the culture medium of SeV-infected cells are shown. LLC-MK2 cells were infected with SeV at a high MOI, and the medium was harvested at the indicated times to measure the infectivity and hemagglutination activity. (B) A part of the medium was collected at 24 h and 36 h postinfection, followed by Western blotting to confirm the accumulation of the released viral proteins. Similarly, proteins in the cells at 36 h postinfection were analyzed by Western blotting. CIU, cell infectious units; HAU, hemagglutination titer units. **, P < 0.05, compared with wild-type SeV. (D) Metabolic labeling of SeV mutants. LLC-MK2 cells were infected with SeV at a high MOI, followed by metabolic labeling with ³⁵S-labeled Cys and Met. Labeled proteins in the medium and cells were immunoprecipitated using anti-SeV or anti-C antisera. (E) Budding ratio of SeV calculated by the expression level of labeled N protein in the medium. The value of the SeV C′/C(-)-infected sample was set to 1. All measured values are represented by black triangles in the bar graph. Standard deviations were calculated from the data of at least three independent experiments. P values in panels A and E were calculated on the basis of Student’s t test.

The infectivity of the W125L mutant SeV at 12 h postinfection was as low as that of SeV C′/C(-) (Fig. 9A), which lacks the two larger C protein subpopulations, C′ and C, but expresses the other two smaller C protein subpopulations, Y1 and Y2 (12, 17). However, the infectivity of W125L mutant SeV was slightly increased compared to that of SeV C′/C(-) with a P of <0.05 at 24 h and a P of <0.1 at 36 h postinfection (Fig. 9A). The amounts of N protein released into the culture medium showed the same tendency (Fig. 9B). The reduced affinity between W125L mutant C protein and Alix may be partially compensated by the accumulation of the mutated C protein in the infected cell. SeV C′/C(-) has a higher hemagglutination titer than W125L mutant SeV, despite lower infectivity (Fig. 9A and C). The reason is currently unknown.

The budding ability of SeV is more precisely evaluated by the amount of the viral proteins released in a short time period during early infection than by the amount of the accumulated ones, because of minimization of the leak of viral protein from dead or dying cells (Fig. 9B). Therefore, after the proteins synthesized in the SeV-infected cells were labeled with ³⁵S-containing Cys and Met, the radioactivity of the proteins in the cell or culture medium, which were immunoprecipitated with anti-SeV or anti-C antiserum, was detected (Fig. 9D). While the labeled viral proteins were found in the medium of cells infected by wild-type or T119I mutant SeV, they were only slightly detected in the medium of cells infected by W125L mutant SeV or SeV C′/C(-). We estimated the budding rates of SeV by the ratio of the radioactivity of N protein in the medium to that in the cell. The budding ratios of W125L mutant SeV and SeV C′/C(-) were calculated to be significantly lower than that of wild-type or T119I mutant SeV (Fig. 9E). In addition, SeV matrix (M) protein, which plays a crucial role in virion assembly and budding, seemed to be accumulated in the cells infected by W125L mutant SeV or SeV C′/C(-), perhaps due to the low budding efficiency (Fig. 9D). The low budding efficiency of the W125L mutant SeV suggests that the interaction of C with Alix significantly affects budding ability. Meanwhile, the low budding rate of SeV C′/C(-), which possesses the C derivatives with Alix-binding ability, may be due to the lack of the N-terminal region in the C protein.

Competition between Y3 and CHMP4 for binding to the Bro1 domain.

Based on the crystal structure of the Y3:Bro1 complex, Y3 is likely to compete with CHMP4 for binding to the Bro1 domain. This hypothesis was investigated using the C-terminal regions of CHMP4 isoforms, aa 205 to 222 of CHMP4A, aa 205 to 224 of CHMP4B, and aa 216 to 233 of CHMP4C, which are responsible for the association with Bro1 (29). For convenience, each Bro1-binding region of the CHMP4 isoforms was fused with ProS2 at the N terminus.

For the competition analysis, Y3:Bro1 complex, unfused ProS2, ProS2-fused CHMP4A_205–222, ProS2-fused CHMP4B_205–224, and ProS2-fused CHMP4C_216–233 were prepared (Fig. 10A). The purified Y3:Bro1 complex was mixed with an excess amount of ProS2-fused CHMP4A_205–222, followed by size exclusion chromatography. The results showed an emergence of a new peak at 14 ml of elution volume (Fig. 10B). SDS-PAGE analysis demonstrated that the new peak consists of ProS2-fused CHMP4A_205–222 and the Bro1 domain, while unbound Y3 was eluted in the range of 15 to 18 ml of elution volume (Fig. 10B). Similar results were obtained when using ProS2-fused CHMP4B_205–224 or ProS2-fused CHMP4C_216–233, instead of ProS2-fused CHMP4A_205–222 (Fig. 10C and D). However, when using unfused ProS2, size exclusion chromatography showed no emergence of a new peak, indicating that unfused ProS2 cannot interact with Bro1 (Fig. 10E). These results demonstrate that CHMP4A–C can associate with the Bro1 domain in competition with Y3.

FIG 10.

Binding of CHMP4 to Alix in competition with C. (A) SDS-PAGE analysis of Y3:Bro1 complex, ProS2, and ProS2-fused C-terminal region of CHMP4A–C used for size exclusion chromatography. (B to E) Analytical size exclusion chromatograms of Y3:Bro1 complex in the absence (dashed line) or presence (gray line) of the ProS2-fused C-terminal region of CHMP4 isoforms or ProS2. Y3:Bro1 complex (1.5 μM) was preincubated with or without a 10-fold excess of CHMP4A_205–222 (B), CHMP4B_205–224 (C), CHMP4C_216–233 (D), or ProS2 (E) for 30 min, followed by size exclusion chromatography. Chromatograms of one of the ProS2-fused C-terminal regions of CHMP4 isoforms alone or ProS2 are shown as black lines. Proteins in fractions eluted from the size exclusion chromatograms were separated by SDS-PAGE, followed by Coomassie blue staining. (F) Affinity analysis between EYFP-fused Y3 and ECFP-fused Bro1 in competition with ProS2-fused CHMP4B_205–224. The quenched fluorescence intensity (ΔI) of ECFP-fused Bro1 at 475 nm, which is caused by the association with EYFP-fused Y3, is divided by the concentration of total ECFP-fused Bro1, and each value is plotted against the concentration of free EYFP-fused Y3. Affinity analysis was performed in the absence (light gray) or presence of 30 μM (dark gray) or 50 μM (black) of ProS2-fused CHMP4B_205–224. The curves represent the best fits to the 1:1 competitive binding model.

Additionally, affinity analysis between EYFP-fused Y3 and ECFP-fused Bro1 was performed in the presence of ProS2-fused CHMP4B_205–224. CHMP4B was selected from CHMP4A–C for the competition analysis, since it is a major binding partner for Alix (35). The data obtained were a good fit in the 1:1 competitive binding model (Fig. 10F). The dissociation constant between ECFP-fused Bro1 and ProS2-fused CHMP4B_205–224 was estimated to be 49 ± 1 μM, which is similar to the value previously obtained by the Biacore analysis (29). Conversely, the dissociation constant between EYFP-fused Y3 and ECFP-fused Bro1 was 0.29 ± 0.01 μM, indicating that Alix associates more strongly with C than CHMP4.

DISCUSSION

Several enveloped viruses use the host’s ESCRT machinery for budding. In the family Paramyxoviridae, PIV5 (36) and Nipah virus (37), which belong to the Rubulavirus and Henipavirus genera, respectively, have been investigated to see whether the ESCRT system is used for the budding process. Of particular note is that the Nipah virus C protein stimulates the ESCRT pathway for efficient viral release, possibly by interacting with the tumor susceptibility gene 101 protein (Tsg101), one of the ESCRT factors. Conversely, the release of VLPs from cells infected by measles virus, which belongs to the genus Morbillivirus, is independent of the ESCRT pathway (38). Therefore, the budding mechanism of paramyxoviruses may differ among genera.

The budding of hPIV1, which, like SeV, belongs to the genus Respirovirus, is thought to be dependent on the ESCRT pathway (30), while the release of the SeV M protein may be caused by the ESCRT-dependent budding process (23). In addition, C-deficient SeV was shown to produce virus particles with highly decreased infectivity (9). However, Gosselin-Grenet et al. have reported that the budding of SeV infectious particles was unaffected by siRNA-based gene knockdown of Alix, VPS4A, or the C protein (26). These inconsistencies may be attributed to the failure of the siRNA technique to induce a sufficient loss of gene expression to affect the budding process or that the gene knockdown of Alix or VPS4A was complemented by other ESCRT-related proteins. In this study, the effect of the interaction between C protein and Alix on SeV budding was analyzed by viral infection experiments with SeV generated based on the crystal structure of the Y3:Bro1 complex. The mutation causing the decrease in binding affinity between C protein and Alix resulted in significantly low infectivity (Fig. 9), demonstrating that C protein enhances SeV budding through binding to Alix.

An ESCRT-independent mechanism is used for the budding of several viruses (39). In this case, the virus recruits the necessary components for the budding process to the budding site on the cell membrane (37–43). Since the C protein of SeV has a membrane-targeting sequence (aa 1 to 23) responsible for cell membrane localization (Fig. 6) (31, 44), it may have a role in recruiting the necessary components to the budding site. In fact, C protein in SeV-infected cells is known to be colocalized with the M protein and hemagglutinin-neuraminidase (HN), which are membrane compartments of the viral particle, and the P protein (21). The M protein of SeV, a central player in viral assembly and budding, is also known to interact with Alix (27, 45). The binding of C protein to Alix seems to have a role in the recruitment of Alix near the cell membrane (23). In addition, the interaction between C and Alix may stimulate the recruitment of the M protein to the budding site through interactions with Alix. That is, C protein has a role in recruiting the proteins necessary for budding on the cell membrane by using the N-terminal membrane-targeting sequence and the C-terminal Alix-binding sequence. The decreased infectivity of SeV C′/C(-) possessing Y1 and Y2, which lack the membrane-targeting sequence of C, and that of the SeV variant possessing W125L mutant C, which has a lowered binding affinity to Alix, support this hypothesis. The C proteins in hPIV1 and hPIV3, which are the other members of the genus Respirovirus, also possess an Alix-binding motif. The hPIV1 C protein is thought to be recruited to the budding site on the cell membrane by interacting with Alix (30), while the hPIV3 C protein is localized on the cell membrane by using its own N-terminal region (46). The C proteins in the genus Respirovirus may commonly have a role in the enhancement of the viral budding process.

It remains unclear whether SeV budding is dependent on the ESCRT system. Although CHMP4 is an essential factor for the budding of VLPs formed by the M protein (23), the C protein competes with all known CHMP4 isoforms, CHMP4A–C, for binding to Alix. Given that the affinity between C protein and Alix is significantly stronger than that between CHMP4 and Alix, at the budding site on the cell membrane, the C protein bound to Alix may be replaced by CHMP4, followed by the recruitment of other ESCRT-related proteins to the budding site. This would eventually result in the release of SeV particles from the cell membrane via an VPS4-dependent membrane scission process.

The present study also showed that C protein interacts with Bro1 and STAT1ND with comparable affinities, although the binding areas for Bro1 and STAT1ND on the molecular surface of Y3 are significantly different. The Y3 mutant, in which Met¹⁵⁰ is replaced with Ala, has a considerably reduced affinity to STAT1 (Table 1). However, the interaction between the mutated C protein and STAT1 was detectable when using a coimmunoprecipitation assay. Currently, a point mutation responsible for the complete loss of binding ability of the C protein to STAT1 has not been identified, suggesting that the extended binding area contributes to the strong affinity to STAT1.

Since a recent study has highlighted the difficulties in development of vaccines against hPIV1 (2), antiviral agents remain essential. The interaction between C and Alix is a potentially effective target for the development of anti-hPIV1 drugs. Moreover, such drugs targeting C may suppress an emergence of resistant mutants, because the mutation in C also influences the amino acid sequence of the P and V proteins essential for the virus. The crystal structure of the Y3:Bro1 complex elucidated herein can help in the structure-based design of the protein interaction inhibitors. The C protein alone was difficult to crystallize due to its low solubility. Since the binding area for Alix in the C protein does not overlap with that for STAT1ND (Fig. 2C), a compound inhibiting the interaction between C protein and Alix may be cocrystallized with a Y3:STAT1ND complex.

In this study, we have demonstrated that the C protein is associated with Alix in a manner similar to that of CHMP4. CHMP4 interacts with the other two Bro1 domain-containing ESCRT-related factors, BRO1 domain- and CAAX motif-containing protein (BROX) (47) and His domain-containing protein tyrosine phosphatase (HD-PTP) (48), as well as Alix in the same manner. The Gag protein of the human immunodeficiency viruses can bind to the Bro1 domains of Alix (49, 50) and BROX (50, 51). Although the effect of the interaction between the Gag protein and Alix has been extensively studied, that between the Gag protein and BROX remains to be investigated. Based on the increased infection of W125L mutant SeV at a later stage (Fig. 9A), SeV may have an alternative budding mechanism that requires no interaction between the C protein and Alix. It will be interesting to explore whether the SeV C protein interacts with BROX or HD-PTP and if this interaction is involved in the viral budding mechanism.

MATERIALS AND METHODS

Cells, viruses, and antibodies.

293T cells (human embryonic kidney cells expressing simian virus 40 T antigen) were provided by the RIKEN Cell Bank (RCB2202). Details of LLC-MK2 cells (macaque kidney-derived cells) were described in a previous study (52). These cells were propagated in Dulbecco’s modified minimum essential medium (DMEM) supplemented with 10% fetal bovine serum (Biological Industries, Kibbutz, Israel) and 100 U ml⁻¹ penicillin–100 μg ml⁻¹ streptomycin (Thermo Fisher Scientific).

SeV Z strain was propagated in 10-day-old embryonated chicken eggs, and its infectivity, which was expressed in cell infectious units per milliliter, was measured with the fluorescent infectious focus assay using LLC-MK2 cells as described previously (52). Recombinant vaccinia virus vTF7.3, which contains the bacteriophage T7 RNA polymerase gene (53), was provided by Bernard Moss (National Institutes of Health, Bethesda, MD). Details of recombinant vesicular stomatitis virus expressing enhanced green fluorescent protein (rVSV-EGFP) were described previously (17).

Mouse monoclonal antibodies against HA tag (G036; Abcam), FLAG tag (M2; Sigma-Aldrich), STAT1 (610185; BD Transduction Laboratories), and actin (MAB1501; Chemicon International) and rabbit polyclonal antibody against GFP (sc-8334; Santa Cruz Biotechnology) were used according to the manufacturers’ instructions. Anti-SeV particle rabbit antiserum (52) was used in this study. Anti-SeV C rabbit antiserum was generated by using purified Y3 mutant as an antigen, with Trp¹²⁵ replaced by Ala to increase its protein solubility. Reaction strengths of the antiserum against the C protein variants are different. The antiserum reacts with W125L mutant C protein more strongly than the wild type, while it reacts with T119I mutant C more weakly than the wild type.

Plasmid construction.

The plasmids for the expression of N-terminally HA-tagged Alix and N-terminally FLAG-tagged SeV C were constructed by using pCAGGS vector, as described previously (13, 22). Amino acid residues of the C protein and its derivatives are numbered with the N-terminal methionine residue of the C protein as 1. T119I, W125L, and W125S mutants of C were generated by introducing mutations into the cDNA of C using the KOD-plus-mutagenesis kit (Toyobo), and the mutated genes were inserted into the pCAGGS vector. The mutations were confirmed by DNA sequencing analysis. Construction of pISRE-EGFP, which possesses an ISRE element upstream of the EGFP gene, was described previously (32). Construction of pSeV(+) containing a cDNA copy of the full-length SeV antigenome, as well as that of pGEM-N, pGEM-P, and pGEM-L, was also described previously (54).

For the expression of Bro1 in E. coli, a DNA fragment encoding the Bro1 domain (aa 1 to 359 in Alix), which was amplified with a forward primer, 5′-GGAATTCCATATGGCGACATTCATCTCGG-3′ (underline indicating an NdeI site), and a reverse primer, 5′-CGGGATCCTTACTGCTGTGGATAGTAAG −3′ (underline indicating a BamHI site), was cloned into the NdeI-BamHI sites in the modified pCold ProS2 vector (TaKaRa), in which an additional sequence corresponding to two histidine residues and an NdeI restriction enzyme site was inserted just after the His tag sequence using an AMAP multisite-directed mutagenesis kit (Amalgaam, Japan). Construction of the expression plasmid of Y3 (aa 99 to 204 in the C protein) has been described previously (13). Alix-binding regions of CHMP4A–C, each N terminus of which is fused with ProS2 via a short linker containing the recognition site of tobacco etch virus protease, were also expressed in E. coli. A DNA fragment encoding CHMP4A_205–222 was amplified by PCR with a forward primer, 5′-CGGTGGGAGCGGTGAAAACCTGTATTTTCAGGGAGGTCCCAAAGTGGATGAAGATGAAGA-3′, and a reverse primer, 5′-GGATCCCTCGAGCTAGGATACCCACTCAGCCAACTGCTT-3′. A DNA fragment encoding CHMP4B_205–224 was amplified with a forward primer, 5′-CGGTGGGAGCGGTGAAAACCTGTATTTTCAGGGAGGTAAGAAGAAAGAAGAGGAGGACGA-3′, and a reverse primer, 5′-GGATCCCTCGAGCTATTACATGGATCCGGCCCAGTTCTCC-3′. A DNA fragment encoding CHMP4C_216–233 was amplified with a forward primer, 5′-CGGTGGGAGCGGTGAAAACCTGTATTTTCAGGGAGGTCAGAGGGCAGAAGAAGAGGATGA-3′, and a reverse primer, 5′-GGATCCCTCGAGCTAAGTAGCCCAAGCTGCCAATTGTTTGA-3′. Each of the three amplified fragments was assembled with the vector fragment, which was amplified with a forward primer, 5′-TAGCTCGAGGGATCCGAATTCAAGCTTG-3′, and a reverse primer, 5′-TCACCGCTCCCACCGCCGCCAGAACCGCCACCTCCTGAACCCGCGGACCTCGGCTGCACC-3′, and pCold ProS2 vector as a template DNA, using the NEBuilder HiFi DNA assembly kit (New England Biolabs, Japan). An original pCold ProS2 vector was used for the expression of ProS2. For the expression of the ECFP-fused Bro1, a DNA fragment encoding ECFP was amplified using a forward primer, 5′-GGAATTCCATATGGTGAGCAAGGGCGAGG-3′ (underline indicating an NdeI site), and a reverse primer, 5′-GGAATTCCATATGACCACTACCGCGTGGCACCAGACCCTTGTACAGCTCGTCCATGCCGAG-3′ (underline indicating an NdeI site), which includes the region encoding the thrombin recognition site. The fragment was cloned into the NdeI site of the Bro1 expression vector. Construction of plasmids for the expression of EYFP-fused Y3 and ECFP-fused STAT1ND has been described previously (13). E114A, T119A, T119I, Y121A, L122A, W125A, W125L, W125S, E128A, and M150A mutants of the EYFP-fused Y3 were generated by introducing mutations into the expression vector for the EYFP-fused Y3 by using a KOD-plus-mutagenesis kit. The mutations were confirmed by DNA sequencing analysis. For expression of hPIV1 C, the N terminus of which was fused with EYFP via a short linker containing the recognition site of tobacco etch virus protease, a DNA fragment encoding hPIV1 C was amplified with a forward primer, 5′-GCGGTAGTGGTCATATGATGCCTTCTTTTTTGAGAGGGATCCT-3′, and a reverse primer, 5′-GGATCCCTCGAGCTATTCTTGTACTATGTGTGCTGCTAGT-3′, followed by assembly with the vector fragment, which was amplified with a forward primer, 5′-TAGCTCGAGGGATCCGAATTCAAGCTTG-3′, and a reverse primer, 5′-GGAATTCCATATGACCACTACCGCGTGGCACCAGACCCTTGTACAGCTCGTCCATGCCGAG-3′, and the expression vector for EYFP-fused Y3 as a template DNA, using the NEBuilder HiFi DNA assembly kit. L122A and W125A mutants of the EYFP-fused hPIV1 C protein were generated by introducing mutations into the expression vector for EYFP-fused hPIV1 C using a KOD-plus-mutagenesis kit. For the expression of EYFP, a DNA fragment encoding EYFP was amplified with a forward primer, 5′-GGAATTCCATATGGTGAGCAAGGGCGAGG-3′ (underline indicating an NdeI site), and a reverse primer, 5′- GCTCTAGATTACTTGTACAGCTCGTCCATG-3′ (underline indicating an XbaI site). The fragment was cloned into the NdeI-XbaI sites in the modified pCold ProS2 vector.

Protein preparations.

Expression of Y3 and EYFP-fused Y3 variants was carried out in E. coli BL21(DE3)-CodonPlus RIL (Novagen) at 15°C for 24 h after induction with 0.2 mM IPTG (isopropyl-β-d-thiogalactopyranoside). Expression of Bro1, ECFP-fused Bro1, ECFP-fused STAT1ND, EYFP-fused hPIV1 C variants, EYFP, ProS2-fused CHMP4A_205–222, ProS2-fused CHMP4B_205–224, ProS2-fused CHMP4C_216–233, and ProS2 was carried out in E. coli BL21(DE3)pLysS (Novagen) in a similar manner. Y3 was purified by nickel affinity chromatography using His-Bind resin (Novagen) at 4°C, while the other proteins possessing a His tag sequence at the N terminus were purified by using Cosmogel His-Accept (Nacalai Tesque), according to each supplier’s instruction manual. Except for EYFP-fused hPIV1 C variants, proteins were further purified by size exclusion chromatography using a HiLoad 16/600 Superdex 75 prep grade column (GE Healthcare Life Science) and 20 mM Tris-HCl buffer (pH 8.0) containing 100 mM NaCl, 1 mM EDTA, 1 mM dithiothreitol (DTT), and 5% glycerol as an elution buffer. Nickel affinity chromatography was also used for preparation of the Y3:Bro1 complex after mixing the supernatant from Bro1-expressed E. coli with the excess supernatant from Y3-expressed E. coli. The complex was then separated from the uncomplexed Y3 by gel filtration chromatography using a HiLoad 16/600 Superdex 75 prep grade column equilibrated with 20 mM Tris-HCl buffer (pH 7.5) containing 100 mM NaCl, 1 mM EDTA, and 1 mM DTT.

Protein concentrations of Y3, Bro1, Y3:Bro1 complex, ProS2-fused CHMP4A_205–222, ProS2-fused CHMP4B_205–224, ProS2-fused CHMP4C_216–233, and ProS2 were determined by measuring the absorbance at 280 nm using the molar extinction coefficients of 22,100, 23,400, 45,800, 15,900, 15,900, 15,900, and 8,940 M⁻¹ cm⁻¹, respectively. Protein concentrations of ECFP-fused Bro1 and ECFP-fused STAT1ND were determined by measuring the absorbance at 435 nm using the molar extinction coefficient of 28,750 M⁻¹cm⁻¹. Protein concentrations of EYFP-fused Y3 and its variants, EYFP-fused hPIV1 C and its variants, and EYFP were determined by measuring the absorbance at 514 nm using the molar extinction coefficient of 83,400 M⁻¹cm⁻¹.

Size-exclusion chromatography analysis.

Prior to gel filtration analysis, Y3 (20 μM) was preincubated with Bro1 (20 μM) for 10 min on ice. A portion of the solution was injected into a Superdex 75 increase 10/300 GL gel filtration column (GE Healthcare) equilibrated with 20 mM Tris-HCl buffer (pH 8.0) containing 100 mM NaCl, 1 mM EDTA, 1 mM DTT, and 5% glycerol at 4°C at a flow rate of 0.5 ml min⁻¹. The Bro1 and Y3 contained in each fraction were resolved by SDS-PAGE and detected by Coomassie blue staining. Similarly, the purified ECFP-fused Bro1 (5 μM) was mixed with EYFP, EYFP-fused Y3, EYFP-fused hPIV1 C, or its variant (10 μM each), and the mixture was preincubated at 4°C for 30 min, followed by gel filtration analysis using the Superdex 75 increase 10/300 GL gel filtration column. The purified Y3:Bro1 complex (6 μM) was mixed with ProS2, ProS2-fused CHMP4A_205–222, ProS2-fused CHMP4B_205–224, or ProS2-fused CHMP4C_216–233 (60 μM each), and the mixture was preincubated at 4°C for 30 min. A portion of the solution was injected into a Superdex 200 increase 10/300 GL filtration column (GE Healthcare).

Crystallography.

Prior to crystallization, the Y3:Bro1 complex solution was dialyzed against 20 mM Tris-HCl buffer (pH 7.6) containing 100 mM NaCl, 1 mM DTT, and 1 mM EDTA and then concentrated to 10 mg ml⁻¹ using Amicon Ultra (Millipore). The crystals of the Y3:Bro1 complex were grown using the sitting-drop vapor diffusion method with a 1:1 (vol/vol) ratio of protein solution to precipitant solution. Small and plate-like crystals were formed within 1 week by using 200 mM malonate-NaOH buffer (pH 8.5) containing 10% (wt/vol) polyethylene glycol 3,350 (Sigma-Aldrich) as a precipitant solution. Crystals suitable for diffraction analysis were obtained by using the microseeding technique. Diffraction intensities of the crystals were measured using the synchrotron radiation from BL38B1 at SPring-8 (Harima, Japan) and a charge-coupled-device (CCD) camera equipped at the station. Data sets were processed with the XDS program (55), and reduction was performed with Aimless (56). The tertiary structure of the Y3:Bro1 complex was solved by the molecular replacement method by the program Molrep in the CCP4 program suite (57) using the atomic coordinates of the Bro1 domain of human Alix (PDB code 2OEV) as a search model. The model was built and refined using the programs Coot (58) and PHENIX (59), respectively. Based on the difference electron density map, a model of Y3 was built step by step. A subset of 5% of the reflections was used to monitor the free R factor (R_free) (60). Each refinement cycle included the refinement of the positional parameters and the individual isotropic B-factors as well as the revision of the model, which was visualized by the program Coot. Details of the data collection and refinement statistics are shown in Table 2.

TABLE 2.

Data collection and refinement statistics for the Y3:Bro1 domain

Parameter	Y3:Bro1 domain
Data collection
Beamline	BL38B1, SPring-8
Space group	P212121
Cell dimensions (Å)
a	50.3
b	102.7
c	103.6
Wavelength (Å)	1.00000
Resolution (Å)	46.26–2.20 (2.27–2.20)
Unique reflections	28,045 (2,383)
Redundancya	5.4 (5.5)
Completeness (%)a	100.0 (100.0)
Rmerge (%)a,b	6.2 (88.0)
I/σa	16.9 (2.1)
Refinement
Resolution (Å)	45.31–2.20 (2.26–2.20)
Used reflections	26,498 (2,019)
No. of atoms
Protein	3,762
Solvent	138
R (%)	20.5
Rfree (%)	26.2
RMSDc
Bond length (Å)	0.007
Bond angle (°)	1.478
Mean B-factor (Å2)
Protein
Bro1 domain	57.8
Y3	63.9
Solvent	53.1
Ramachandran plot (%)
Favored	96.34
Allowed	2.58
Disfavored	1.08

^aValues in parentheses are for the highest-resolution bin.

^bR_merge = Σ|I – 〈I〉|/ΣI, where I is the observed intensity and 〈I〉 is the mean value of I.

^cRoot mean square deviations (RMSD) are calculated by PHENIX (59).

Affinity analysis.

The binding affinity of EYFP-fused Y3 to ECFP-fused Bro1 or ECFP-fused STAT1ND was measured using FRET as previously described (13). Briefly, an initial solution was prepared at 25°C in a well of a 96-well plate (ProteoSave 96F; Sumilon) with a volume of 0.25 ml of 20 mM Tris-HCl buffer (pH 8.0) containing 100 mM NaCl, 1 mM EDTA, 1 mM DTT, 5% glycerol, 0.2 mg ml⁻¹ bovine serum albumin (BSA), and 0.2 μM ECFP-fused Bro1 or ECFP-fused STAT1ND. Five-microliter aliquots containing the given concentrations of EYFP-fused Y3 (0.05 to 100 μM) were successively added to the solution, followed by shaking using a plate shaker, extraction of a 5-μl aliquot of the reaction mixture, and recording the fluorescence intensity at 475 nm using an EnSpire 2300 plate reader (Perkin Elmer).

We defined the dissociation constant for the EYFP-fused Y3 to ECFP-fused Bro1 as K_d:

$$K_d=\frac{\left[X\right]\left[Y\right]}{\left[X\cdot Y\right]}$$ (1)

In this equation, X means ECFP-fused Bro1, while Y means EYFP-fused Y3. The quenched fluorescence intensity (ΔI) is represented by equation 2:

$$\Delta I=\frac{\Delta I_{\text{max}}\cdot {\left[X\right]}_t\cdot \left[Y\right]}{K_d+\left[Y\right]}$$ (2)

In this equation, [X]_t means the sum of concentrations of X complexed and uncomplexed with Y, ΔI_max means the maximum quenched intensity per molar of X, and [Y] means the concentration of the uncomplexed Y, which can be determined as a solution satisfying equations 3 and 4:

$${\left[X\right]}_t=\left[X\right]\left(1+\frac{\left[Y\right]}{K_d}\right)$$ (3)

$${\left[Y\right]}_t=\left[Y\right]\left(1+\frac{\left[X\right]}{K_d}\right)$$ (4)

With respect to the binding affinity of EYFP-fused Y3 to ECFP-fused STAT1ND, we defined the dissociation constants for the first and the second EYFP-fused Y3 for each of the two Y3-binding sites in the ECFP-fused STAT1ND dimer as K_d1 and K_d2, respectively:

$$K_{d1}=\frac{\left[Y\right]\left[Z_2\right]}{\left[Y\cdot Z_2\right]}$$ (5)

$$K_{d2}=\frac{\left[Y\cdot Z_2\right]\left[Y\right]}{\left[Y_2\cdot Z_2\right]}$$ (6)

In these equations, Z means ECFP-fused STAT1ND. Assuming that the quenching ratio after the binding of Y to Z₂ homodimer is the same as that after the binding of Y to Y·Z₂ heterotrimer, quenched fluorescence intensity (ΔI) is represented by equation 7:

$$\Delta I=\frac{\Delta I_{\mathit{max}}\cdot {\left[Z_2\right]}_t\cdot \left(K_{d2}\cdot \left[Y\right]+2{\left[Y\right]}^2\right)}{2K_{d1}\cdot K_{d2}+2K_{d2}\cdot \left[Y\right]+2{\left[Y\right]}^2}$$ (7)

In this equation, [Z₂]_t means the sum of concentrations of Z₂ homodimer complexed and uncomplexed with Y, ΔI_max means the maximum quenched intensity per molar of Z₂ homodimer, and [Y] means the concentration of the uncomplexed Y, which can be determined as a solution satisfying equations 8 and 9:

$${\left[Z_2\right]}_t=\left[Z_2\right]\left(1+\frac{\left[Y\right]}{K_{d1}}+\frac{{\left[Y\right]}^2}{K_{d1}\cdot K_{d2}}\right)$$ (8)

$${\left[Y\right]}_t=\left[Y\right]+\left[Z_2\right]\left(\frac{\left[Y\right]}{K_{d1}}+\frac{2{\left[Y\right]}^2}{K_{d1}\cdot K_{d2}}\right)$$ (9)

Affinity analysis between EYFP-fused Y3 and ECFP-fused Bro1 was also carried out in the presence of ProsS2-fused CHMP4B_205–224. The dissociation constant for the ProS2-fused CHMP4B_205–224 to ECFP-fused Bro1 was defined as K_d3:

$$K_{d3}=\frac{\left[X\right]\left[C\right]}{\left[X\cdot C\right]}$$ (10)

In this equation, C means ProS2-fused CHMP4B_205–224. The quenched fluorescence intensity (ΔI) is represented by equation 11:

$$\Delta I=\frac{{\left[X\right]}_t\cdot \left(\Delta I_{\text{max}1}\cdot K_{d3}\cdot \left[Y\right]+\Delta I_{\text{max2}}\cdot K_d\cdot \left[C\right]\right)}{K_d\cdot K_{d3}+K_{d3}\cdot \left[Y\right]+K_d\cdot \left[C\right]}$$ (11)

In this equation, [X]_t means the sum of concentrations of X complexed with Y or C and uncomplexed with Y and C. ΔI_max1 and ΔI_max2 mean the maximum quenched intensities per molar of X occurring with the association with Y and C, respectively. [Y] and [C] mean the concentrations of uncomplexed Y and C, respectively, which can be determined as a solution satisfying all of equations 4, 12, and 13:

$${\left[X\right]}_t=\left[X\right]\left(1+\frac{\left[Y\right]}{K_d}+\frac{\left[C\right]}{K_{d3}}\right)$$ (12)

$${\left[C\right]}_t=\left[C\right]\left(1+\frac{\left[X\right]}{K_{d3}}\right)$$ (13)

The K_d, K_d1, K_d2, and K_d3 values were determined by the nonlinear least-squares method together with ΔI_max, ΔI_max1, and ΔI_max2.

Generation of recombinant SeV possessing the mutated C proteins.

Mutations to change the amino acid sequence of the C protein were introduced into the SalI-EagI fragment in the SeV genome (positions 2074 to 2746), which corresponds with the overlapping region of the P, V, and C genes, without changing the amino acid sequences of the P and V proteins. The corresponding fragment in pSeV(+) containing the whole antigenomic cDNA of the SeV Z strain was replaced by the mutated one, resulting in the generation of a pSeV(+) variant expressing the T119I or W125L mutant of C. SeV was recovered from the cDNA as described previously, with some modifications (54). Briefly, 293T cells were infected with vTF7.3 and transfected with pSeV(+) variant and the plasmids encoding trans-acting proteins, pGEM-N, pGEM-P, and pGEM-L, with the aid of FuGENE HD transfection reagent (Promega). The cells were then maintained in DMEM containing cytosine β-d-arabinofuranoside (Ara-C; Sigma) to minimize the cytopathogenicity of vaccinia virus. After 2 days, the cells were suspended in 1 ml of Dulbecco’s phosphate-buffered saline (PBS) and disrupted by three cycles of freezing and thawing. The cell lysates were inoculated into 10-day-old embryonated chicken eggs.

Coimmunoprecipitation and Western blotting.

293T cells in a 12-well dish were transfected with a plasmid for the expression of HA-tagged Alix (0.5 μg) and one of the plasmids for the expression of C variants (0.5 μg) using FuGENE HD reagent (Promega). After 24 h, the cells were solubilized in 0.25 ml of NP-40 lysis buffer (0.5% NP-40, 50 mM Tris-HCl [pH 7.6], 50 mM NaCl, 1 mM EDTA, and cOmplete mini protease inhibitor cocktail [Roche Diagnostics]). Cell lysates were immunoprecipitated with anti-FLAG antibody together with protein G Sepharose (GE Healthcare), and the precipitates were washed three times with the NP-40 lysis buffer and once with wash II buffer (50 mM Tris-HCl [pH 7.4], 50 mM NaCl, 5 mM EDTA). The supernatant was analyzed by SDS-PAGE, followed by Western blotting using the anti-HA or anti-FLAG antibody. Protein bands were detected by using horseradish peroxidase (HRP)-conjugated anti-mouse IgG antibody and Luminata Forte Western HRP substrate (Millipore), followed by analysis using a LumiCube imaging analyzer (Liponics, Tokyo, Japan). A part of the cell lysate was also analyzed by Western blotting to confirm the expression of the protein.

For the immunoprecipitation analysis using SeV-infected cells, LLC-MK2 cells in a 6-well dish were transfected with a plasmid for the expression of HA-tagged Alix (1 μg) using FuGENE HD reagent. When the cells reached 100% confluence, they were incubated with 0.2 ml of the indicated SeV inoculum at an input MOI of 15. After adsorption for 1 h at 37°C with an occasional tilting, virus inoculum was removed and the cells were incubated in 2 ml of serum-free DMEM. After further incubation for 47 h, the cells were solubilized in 0.5 ml of NP-40 lysis buffer, and the cell lysates were immunoprecipitated with the anti-HA antibody together with protein G Sepharose, followed by SDS-PAGE and Western blotting using anti-HA antibody or anti-C antiserum. For the immunoprecipitation of STAT1 endogenously expressed in SeV-infected cells, confluent LLC-MK2 cells in a 12-well dish were incubated with 0.1 ml of the indicated SeV inoculum at an input MOI of 15. After 1 h of absorption, inoculum was removed, and the cells were incubated in 1 ml of serum-free DMEM. At 24 h postinfection, the cells were solubilized in 0.25 ml of NP-40 lysis buffer, and the cell lysates were immunoprecipitated with the anti-STAT1 antibody together with protein G Sepharose, followed by SDS-PAGE and Western blotting.

qRT-PCR.

Confluent LLC-MK2 cells in a 12-well dish were absorbed with SeV or its recombinants at an input MOI of 10. After adsorption for 1 h at 37°C with an occasional tilting, inocula were removed, and the cells were incubated with serum-free DMEM at 37°C. After further incubation for 18 h, total RNA was extracted from the cell pellets using a Maxwell RSC simplyRNA cell kit (Promega) in a Maxwell RSC instrument (Promega) according to the supplier’s instruction manuals. Quantitative real-time PCR (qRT-PCR) was performed using a One-Step TB Green PrimeScript RT-PCR kit II (TaKaRa) on a LightCycler 480 real-time instrument (Roche) and with each primer set for the rhesus macaque genes (Table 3). The primer set for the IFNB1 gene was described in a previous study (61), and those for the other genes were designed using Primer-BLAST (https://www.ncbi.nlm.nih.gov/tools/primer-blast/). The expression levels of mRNA were analyzed in triplicate by the comparative threshold cycle (C_T) method, using GAPDH mRNA as an endogenous control.

TABLE 3.

Sequences of qRT-PCR primers

Gene	Sequence for:
	Forward primer	Reverse primer
IFNB1	TGCCTCAAGGACAGGATGAAC	GCGTCCTCCTTCTGGAACTG
IFIT2	CTGCCGAACAGCTGAGAATTGC	TTAGTTGCCGTAGGCTGCTCTC
IFIT5	CTGCAGAGCGTTGCCATCAT	CATTCTAACTCCAACAGAATGGCCT
OAS1	GCTGAGGCCTGGCTGAATTAC	TGGATGTAACCATGTTGCTGATACA
GAPDH	AACAGCCTCAAGATCGTCAGCAAC	GTGGTCATGAGTCCTTCCACGATAC

Reporter assay for IFN-α/β signal transduction.

For IFN-α/β signal transduction, subconfluent LLC-MK2 cells in a 24-well dish were transfected with pISRE-EGFP (0.25 μg) using the FuGENE HD reagent, followed by SeV infection at an input MOI of 15 at 5 h posttransfection. After an additional 1 h of incubation at 37°C, inocula were removed and serum-free DMEM was added. The cells were further incubated at 37°C for 6 h, followed by the addition of IFN-β (1,000 U ml⁻¹). After further incubation for 12 h, cells were solubilized in 0.15 ml of NP-40 lysis buffer. An aliquot (0.1 ml) of the cell lysates was analyzed using an EnSpire 2300 plate reader to estimate the fluorescence intensity of EGFP at 509 nm after excitation at 488 nm. Standard deviations were calculated from the data of at least three experiments. A part of the cell lysates was also analyzed by Western blotting to confirm the expression of the protein.

Measurement of antiviral response in SeV-infected cells.

LLC-MK2 cells in a 12-well plate were infected with SeV at an input MOI of 10. After adsorption at 37°C for 1 h, inoculum was removed, and 1 ml of serum-free DMEM was added. After an additional incubation for 15 h, the cells were superinfected with rVSV-EGFP at an MOI of 10, followed by further incubation at 37°C for 9 h. The cells were then lysed by SDS sample buffer, followed by SDS-PAGE and Western blotting using anti-GFP antibody and anti-SeV antiserum. Protein bands were visualized and analyzed as described above.

Mock-infected LLC-MK2 cells were stimulated with IFN-α (20 U ml⁻¹). After an additional incubation for 15 h, the cells were superinfected with rVSV-EGFP at an MOI of 10 and further incubated at 37°C for 9 h, followed by preparation of the cell lysate to use for SDS-PAGE and Western blotting.

Metabolic labeling of SeV-infected cells.

LLC-MK2 cells in a 35-mm dish were washed with PBS and incubated with 0.2 ml of virus inoculum at an input MOI of 10. After adsorption for 1 h at 37°C with an occasional tilting, virus inoculum was removed and the cells were incubated in 1 ml of DMEM. At 5.5 h postinfection, the infected cells were washed with PBS and incubated in Cys- and Met-free DMEM (Thermo Fisher Scientific) for 30 min for the starvation of Cys and Met. The cells were then incubated in DMEM containing 3.7 MBq ml^{−1 35}S-labeled Cys and Met (Express protein labeling mix, [³⁵S]; PerkinElmer, Inc., Waltham, MA, USA) and a 1/50 concentration of cold Cys and Met for an additional 6 h. The culture medium was separated after low-speed centrifugation, and the supernatant was mixed with an equal volume of 2-fold-concentrated radioimmunoprecipitation assay buffer (10 mM Tris-HCl [pH 7.4], 1% Triton X-100, 1% [wt/vol] sodium deoxycholate, 0.1% [wt/vol] SDS, 150 mM NaCl, and cOmplete protease inhibitor cocktail). The cells were lysed in the radioimmunoprecipitation buffer and clarified by centrifugation with a microcentrifuge at maximum speed. The culture medium and cell lysates were immunoprecipitated with anti-SeV or anti-C antiserum together with protein A Sepharose (Pharmacia) and analyzed by SDS-PAGE as described previously (62).

Data availability.

The atomic coordinates and structure factors of the Y3:Bro1 complex have been deposited in the Protein Data Bank under accession code 6KP3.