The carboxyl segment of the mumps virus V protein associates with Stat proteins in vitro via a tryptophan-rich motif

(1)

Article

Reference

The carboxyl segment of the mumps virus V protein associates with Stat proteins in vitro via a tryptophan-rich motif

NISHIO, Machiko, et al.

Abstract

Viruses of the Paramyxovirinae, similar to other viruses, have evolved specific proteins that interdict IFN action as part of a general strategy to counteract host innate immunity. In many (but not all) cases, this interdiction is accompanied by a lowering of the intracellular levels of the STAT proteins. Among rubulaviruses, there is a notable variation in how they interfere with IFN action. Whereas SV41, SV5, and MuV all act by lowering Stat1, hPIV2 acts by lowering Stat2. Here, we show that the mumps and hPIV2 V proteins both form a complex with several Stat proteins in a mixed-extract assay. This suggests that the specific degradation of these Stat proteins is not determined by complex formation, but presumably at some later stage of the degradation pathway. V/Stat complex formation requires a specific carboxyl segment of V.

However, a previously unrecognized trp-rich motif, rather than the Zn(++)-binding cys-cluster of this segment, appears to be required for V/Stat interaction. The C protein of Sendai (respiro-) virus, another P gene encoded protein, also forms a complex with Stat1, and prebinding of MuV V to Stat1 [...]

NISHIO, Machiko, et al . The carboxyl segment of the mumps virus V protein associates with Stat proteins in vitro via a tryptophan-rich motif. Virology , 2002, vol. 300, no. 1, p. 92-9

DOI : 10.1006/viro.2002.1509 PMID : 12202209

Available at:

http://archive-ouverte.unige.ch/unige:38134

Disclaimer: layout of this document may differ from the published version.

1 / 1

(2)

The Carboxyl Segment of the Mumps Virus V Protein Associates with Stat Proteins in Vitro Via a Tryptophan-Rich Motif

Machiko Nishio,¹ Dominique Garcin, Viviane Simonet, and Daniel Kolakofsky²

Department of Genetics and Microbiology, University of Geneva School of Medicine, CMU, 9 Ave de Champel, CH1211 Geneva, Switzerland Received January 29, 2002; returned to author for revision March 21, 2002; accepted April 18, 2002

Viruses of theParamyxovirinae, similar to other viruses, have evolvedspecific proteins that interdict IFN action as part of a general strategy to counteract host innate immunity. In many (but not all) cases, this interdiction is accompanied by a lowering of the intracellular levels of the STAT proteins. Among rubulaviruses, there is a notable variation in how they interfere with IFN action. Whereas SV41, SV5, andMuV all act by lowering Stat1, hPIV2 acts by lowering Stat2. Here, we show that the mumps andhPIV2 V proteins both form a complex with several Stat proteins in a mixed-extract assay. This suggests that the specific degradation of these Stat proteins is not determined by complex formation, but presumably at some later stage of the degradation pathway. V/Stat complex formation requires a specific carboxyl segment of V. However, a previously unrecognizedtrp-rich motif, rather than the Zn^⫹⫹-binding cys-cluster of this segment, appears to be required for V/Stat interaction. The C protein of Sendai (respiro-) virus, another P gene encoded protein, also forms a complex with Stat1, and prebinding of MuV V to Stat1 prevents the subsequent binding of SeV C. Our results suggest that rubulavirus V proteins may be relatedto both the C andthe V proteins of respiroviruses. © 2002 Elsevier Science (USA)

INTRODUCTION

The Paramyxovirinaeare a large subfamily of important human and animal pathogens that are currently classified in three genera; respiro-, morbilli-, and rubulaviruses. This classification reflects in part the genetic organization of their P genes (Fig. 1). Paramyxovirus P genes express not only P proteins that are essential subunits of the viral RNA polymerase, they also express

“accessory” proteins such as C and V that are not expressed by all P genes of this subfamily. These accessory proteins are not essential for virus replication in some cell culture conditions, and they appear to act primarily to counteract the innate defenses of host cells (for review, see Lamband Kolakofsky, 2001).

All viruses of the Paramyxovirinae (with the notable exception of hPIV1) contain an mRNA-editing site at which G residues are inserted into the P mRNA in a programmed manner during its synthesis (Fig. 1; dotted vertical line) (Cattaneoet al., 1989; Thomaset al.,1988;

Vidal et al., 1990). The number of G residues inserted varies between the different viruses and appears to be matched to the organization of their alternate ORFs. The editing site separates each P gene into two segments, and it generates three different mRNAs in which the three possible ORFs downstream of the editing site (re-

gardless of their size) are fused to a unique upstream ORF (Fig. 1). These mRNAs are translated into three polypeptides with common N-terminal segments and different C-terminal segments, called P (polymerase cofactor module), V, and W/D/I. In respiroviruses and morbilliviruses, additional P gene complexity is provided by the presence of two (overlapping) ORFs upstream of the editing site (a longer ORF that is in acidic in nature, and a shorter basic C ORF), and only the longer ORF extends to this site. For Sendai virus and hPIV1, the C ORF that terminates before the editing site can also code for a nested set of up to four polypeptides, due to the use of four ribosomal initiation sites. For rubulaviruses, there is but a single ORF upstream of the editing site. The P genes of the Paramyxovirinae are thus composed of a series of modules that are combined in various ways via mRNA editing, or as additional modules expressed via unusual translational initiation (Boecket al.,1992; Curran and Kolakofsky, 1988; Latorreet al.,1998). However, the P protein itself is the only P gene product that is essential for viral RNA synthesis (Curranet al.,1995; Murphy and Parks, 1997). For SeV, C and V have negative effects on viral RNA synthesis when this property is examined in isolation (Curranet al.,1991, 1992). However, at least the C proteins act in a positive manner in promoting early steps in intracellular replication, when they are present at low concentrations (Latorreet al., 1998).

Viruses of the Paramyxovirinae, similar to other vi- ruses, have evolved specific proteins that interdict IFN signaling as part of a general strategy to counteract host innate immunity. In many (but not all) cases, this inter-

1On sabbatical leave from Dept. of Microbiology, Mie University School of Medicine, Tsu-Shi, Mie-Ken 514-8507, Japan.

2To whom correspondence and reprint requests should be addressed.

Fax: (41 22) 702 5702. E-mail: Daniel.Kolakofsky@Medecine.unige.ch.

doi:10.1006/viro.2002.1509

0042-6822/02 $35.00

92

(3)

diction is accompanied by a lowering of the intracellular levels of the STAT proteins. This has been noted for SeV and hPIV3, two respiroviruses (Didcock et al., 1999a;

Garcin et al., 2000, 2002), and for four rubulaviruses, mumps virus (MuV) (Kubotaet al.,2001), hPIV2 (Younget al.,2000; Nishioet al., 2001), SV41 (Nishioet al.,2001), and SV5 (Didcocket al.,1999). For SeV, the C proteins are responsible for this effect, as their expression by themselves is sufficient to interdict IFN signaling via C/Stat complex formation (Garcinet al.,2000; Katoet al.,2001;

Takeuchi et al., 2001). The C proteins have also been shown to induce Stat1 ubiquitination and degradation, but only the longer C proteins are competent for this task (Garcinet al.,2002). The proteasome may play a role in this degradation, as drugs such as MG132 and lactacys-

tin can prevent C-induced Stat1 instability, as well as the ability of SeV infection to counteract the antiviral (VSV) state. For the rubulaviruses that are thought not to encode C proteins (see below), it is the P gene encoded V protein that is responsible for this action, and loss of Stat1 is similarly prevented by MG132 and lactacystin.

These drugs also prevent the virus-induced anti-IFN effects (Didcock et al., 1999). Remarkably, respiroviruses and rubulaviruses appear to use a different P gene encoded protein to interdict IFN action, at least in part, by inducing Stat degradation.

Among rubulaviruses, there is also a notable variation in how they interfere with IFN action. Whereas SV41, SV5, and MuV all act by lowering Stat1, hPIV2 acts by lowering Stat2, and this difference is mirrored in how these rubulaviruses counteract IFN␣and IFN␥signaling (Young et al., 2000; Nishio et al., 2001; Parisien et al., 2001). Al- though V/Stat complex formation has not as yet been directly demonstrated for a rubulavirus V protein, we hoped this could be carried out by anin vitroapproach that demonstrated SeV C protein/Stat1 complex formation (Garcin et al.,2002). If so, we could then examine where this specificity for Stat degradation was located, by constructing chimeric V proteins consisting of the amino-terminal and carboxyl segments of MuV and hPIV2 proteins fused at the editing site. This article reports that the MuV and hPIV2 V proteins both form a complex with both Stat1 and Stat2in vitro. This suggests that the specific degradation of these Stat proteins is not determined by complex formation, but presumably at some later stage of the degradation pathway. Moreover, MuV V protein may interact with a similar surface of Stat1 as the SeV C protein, as prebinding of V prevents the subsequent binding of C.

RESULTS

SeV C protein/Stat1 complexes can be coimmunopre- cipitated from infected human and murine cell extracts (Takeuchiet al.,2001; Garcinet al.,2002). This complex can also be formedin vitroby mixing MEF cell extracts that are naturally rich in Stat1 with extracts of BSR T7 cells (Buchholzet al.,1999), in which the C proteins were expressed by transfection, and which naturally contain low levels of Stat1 and are IFN-incompetent, i.e., they do not respond to IFN treatment (data not shown). Complex formationin vitrois rapid, appears not to require energy, and is specific in that a natural-point mutation in C (F170S) responsible for the loss of anti-IFN activity also abrogates C/Stat1 complex formation (Garcin et al., 2002). We therefore first examined whether the MEF extracts also contained detectable amounts of Stat2 and Stat3 (the latter as a prospective negative control) by Western blotting with anti-Stat1, 2, or 3. These MEFs appear to be in a permanent antiviral state. Highly IFN- sensitive viruses such as VSV or SeV-C (F170S) do not

FIG. 1.Schematic diagram of the modular P genes of paramyxoviruses and their genetic organization. For the respiroviruses and morbilliviruses,the faithful transcript or unedited mRNA (top left) is translated to yield the P protein (N-assembly⫹polymerase cofactor modules). The addition of one G nucleotide at the editing site during mRNA synthesis produces an mRNA that encodes the V protein (N-assembly

⫹V ORF). Addition of two G nucleotides at the editing site produces an mRNA that encodes the W or D proteins (N-assembly⫹W/D ORF) depending on the virus. The respiroviruses and morbilliviruses P mRNAs also encode the C protein ORF at the 5⬘end of the mRNA independent of editing status. The unedited version of the Rubulavirus “P” mRNA encodes the V protein (C-like⫹V ORF). The cysteine-rich domain of the V protein is indicated by Zn (Zn^⫹⫹-binding domain). Addition of two G nucleotides at the rubulavirus-editing site produces an mRNA that encodes the essential rubulavirus P protein. The RNA editing site, at which pseudotemplated nucleotides are added to the mRNAs, is indicated by the vertical dotted line. STAT-BD indicates the possible presence of a Stat protein-binding domain.

93 CARBOXYL SEGMENT OF THE MUMPS VIRUS V PROTEIN

(4)

grow in these cultures (Garcin et al., 2000), and their supernatants contain high levels of IFN (D. Garcin, un- published results). These MEFs might then contain ele- vated levels of many ISGs, including several Stat proteins. Lanes “none” of Fig. 2 show that these extracts contained detectable amounts of Stat2 and Stat3 as well as Stat1. To determine whether our antibodies were act- ing in a specific manner, i.e., that they did not cross- select the other Stat proteins, MEF extracts were first precipitated with anti-Stat1, 2, or 3, and the individual precipitates were then Western blotted with anti-Stat1, 2, or 3. All three anti-Stat antibodies were found to precip- itate only their cognate Stat proteins (Fig. 2). Preexisting Stat protein complexes that may be present in the MEF extracts (such as Stat1/Stat2 and Stat1/Stat3 het- erodimers) are presumably below the level of detection with this method.

The MuV V, I (or NS2), and P proteins were then expressed by transfection in BSR T7 cells (Fig. 3, top panel, “extract”) and their extracts were mixed with MEF extracts. Possible complexes with Stat1, 2, or 3 were precipitated with their respective antibodies, and these precipitates were then Western blotted with an antibody (T61) to the common amino-terminal segment of V, NS2, and P. Unexpectedly, all three anti-Stat antibodies were found to coprecipitate V, whereas NS2 and P/Stat complexes were not detected. Moreover, the reciprocal pre- cipitation of the mixed extracts with T61 followed by Western blotting with anti-Stat1 and anti-Stat3 also showed specific V/Stat complex formation (Fig. 3). A similar experiment was then carried out with the hPIV2 V

and P proteins, using mAb84-1A to the common amino- terminal domain of hPIV2 V and P. As before, all three anti-Stat antibodies coprecipitated the V protein, but not the P protein (Fig. 4). Thus, both the MuV and the hPIV2 V proteins form complexes in vitro with all three Stat proteins examined.

The carboxyl, cys-rich, Zn^⫹⫹-binding segment of the V proteins

The carboxyl segment of the V ORF is by far the best conserved of the various P gene ORFs (Fig. 5A). There are seven perfectly conserved cysteine residues, and this domain binds two atoms of Zn^2⫹(Liston and Briedis, 1994; Paterson, 1995). Similar to the carboxyl pol-cofactor module of the P protein, the cys-rich V ORF is found only as the carboxyl segment of a fusion protein; this module is never expressed by itself. As rubulavirus P proteins share the amino-terminal segments of their V proteins

FIG. 3.MuV V protein associates with Stat proteins in the mixed- extract assay. Cytoplasmic extracts of MEFs were first mixed with those of BSR T7 cells in which the MuV NS2, V, and P proteins were expressed by transfection (see Materials and Methods), or mock- transfected extracts as a negative control (extract, top panel, left). The mixed extracts were then precipitated with anti-Stat1, anti-Stat2, or anti-Stat3 as indicated, and replicate blots were probed with an anti- peptide antiserum (T61) to the common amino-terminal region of NS2, V, and P. Alternatively, the mixed extracts were precipitated with T61 and replicate blots were probed with anti-Stat1, anti-Stat2, or anti-Stat3 (right side). The anti-Stat2 blot did not reveal a band and is not shown.

FIG. 2.Specific detection of Stat1, 2, and 3 in MEF extracts with antibodies. Cytoplasmic extracts of MEFs were precipitated with anti- Stat1, anti-Stat2, or anti-Stat3, with the aid of Staph-A beads. After washing (see Materials and Methods), the proteins remaining on the beads were solubilized in protein sample buffer, separated by SDS–

PAGE, and transferred to a nylon membrane, and replicate blots were probed with anti-Stat1, anti-Stat2, or anti-Stat3, as indicated. Ten per- cent of the samples used for immunoprecipitation were separated directly on the gel and not immunoprecipitated (lanes “none”). None lanes indicate the relative levels of Stat proteins in the starting extract and also serve as electrophoretic mobility markers.

(5)

but do not prevent IFN signaling, the carboxyl segment of the V protein is thought to be primarily responsible for this effect (Didcocket al., 1999). To examine the importance of the Zn^⫹⫹-binding cys cluster for MuV V protein/

Stat interactionin vitro, two pairs of closely spaced cys residues (C²⁰/C²⁴and C³⁶/C³⁸) were individually mutated to ala (residues numbered from start of the expressed V ORF, Fig. 5A). These mutant V proteins were then tested for their interaction with either Stat1 or Stat2 by coimmunoprecipitation. The C20A/C24A and C36A/C38/A double mutants consistently interacted well with both Stat1 and 2 (Fig. 5B, lanes MuV-C20A/C24A and lanes MuV-C36A/

C38A, and data not shown). Thus, either a completely intact cys cluster is not critical for Stat1 or Stat2 interaction, and/or the mixed-extract assay is particularly per- missive for this complex formation. As the MuV NS2 and P proteins do not associate with Stat proteinsin vitro, we further examined the specificity of this complex formation using chimeric MuV V proteins (fused at the editing site). To examine whether any paramyxovirus carboxyl segment will suffice for Stat1 binding, the MuV V ORF was replaced with that of SeV (the SeV V protein does not appear to bind Stat1; Fig. 6A). The chimeric MuV/SeV V protein was found not to interact with Stat1, 2, or 3 (lane SeV, Fig. 5B). MuV V/Stat interactionin vitrothus requires a specific carboxyl segment.

The above results suggest that residues in the carboxyl segment other than those conserved for Zn^⫹⫹- binding are important for Stat interaction. The V ORF can itself be divided into two parts; the downstream cys-rich

FIG. 4.hPIV2 V protein associates with Stat proteins in the mixed- extract assay. A similar experiment to that shown in Fig. 3, except that the hPIV2 V and P proteins were expressed in BSR T7 cells, and mAb 84-1A to the common region of hPIV2 V and P was employed.

FIG. 5.Interaction of mutant V proteins and Stat1, 2, and 3. The amino acid sequences of the expressed SeV, MuV, and hPIV2 V ORFs, numbered from their beginning downstream of the editing site, are shown in A. (B) All the V proteins tested contained the MuV amino-terminal segment fused to the wild-type V ORFs of MuV (MuV) or SeV (SeV), or their mutated V ORFs as indicated, using the numbering shown in A. The V proteins were assayed for their ability to associate with Stat1, 2, and 3 in MEF extracts as for Fig. 3. The top “extract” panels show the amounts of the various V proteins present in 10% of the starting BSR T7 extracts used for the immunoprecipitates (lower panels). (C) A short region of homology found in the SeV C proteins and the amino-terminal segments of the MuV and hPIV2 V proteins. The aa are numbered from the amino-terminus. The upward arrow indicates the critical F170 of the SeV C proteins.

(6)

Zn^⫹⫹-binding domain that begins with the conserved tetrapeptide ¹⁹WCNP²² (where cys to ala double mutations have little or no effect on Stat interaction) and the region upstream that extends to the fusion site, beginning with the conserved tetrapeptide¹HRRE⁴. Within this upstream region, there are two trp residues that are notable (W⁵and W⁹); they are partially conserved, and trp residues are normally very rare. W9 is conserved among all rubulaviruses (and morbilliviruses), but this position is occupied by a charged residue in the respiroviruses.

Position 5 is always occupied by an aromatic residue, but W, Y, and H are all found here. Together with¹⁹WCNP²², this region vaguely resembles the well- studied WW protein domain that is notorious for protein/

protein interaction (Sudol and Hunter, 2000). WW domains contain two strictly conserved trp residues that are spaced 20–22 aa apart, and there is a highly conserved proline that follows the downstream trp by three residues (Freemont, 1993; Jianget al.,2001). W⁹and W¹⁹

are only 10 residues apart, but W¹⁹is indeed followed by P²².

To examine whether W⁵ and W⁹ of the carboxyl segment of MuV V are important for V/Stat1interaction, both trp residues were mutated to those found at the same location in the SeV V protein, i.e., W⁵ to H and W⁹ to E.

This MuV V(W5H/W9E) protein was found to be inactive in Stat complex formation (lanes MuV-W5H/W9E, Fig. 5B).

Thus, in contrast to mutations of the cys cluster, mutation of the upstream trp residues inactivated Stat interaction, and at least one of these two trp are thus essential for this interactionin vitro. To examine whether the absence of these two W by themselves could possibly account for the inability of the MuV/SeV V protein chimeras to interact with Stat1 in this assay, we constructed the reciprocal mutant MuV/SeV (H5W/E9W). These mutations that restore the W⁵/W⁹/¹⁹WCNP²²motif to the chimeric MuV/

SeV V protein also restore the ability of this chimera to interact with Stat (lanes SeV-H5W/E9W, Fig. 5B). Thus, in contrast to the conserved cys-rich Zn^⫹⫹-binding module, the W⁵/W⁹/¹⁹WCNP²²motif of MuV V protein appears to be critical for Stat complex formationin vitro.

Although the available evidence indicates that the carboxyl segment of V is primarily responsible for V/Stat interaction, we nevertheless also examined a single residue in the amino-terminal segment that could be in- volved in this interaction. There is circumstantial evidence that the amino-terminal modules of the rubulavirus V proteins are related to the respirovirus C proteins;

they are all ca. 170 residues long and are basic (pI ⬃ 10.5) and there is a small region of apparent sequence conservation that includes F170 of SeV C that is critical for its interaction with Stat1 (Fig. 5C). However, when the MuV V (F139S) protein was tested, this phe-to-ser mutation did not abolish V interaction with Stat1 (lanes MuV- F139S, Fig. 5B).

MuV V andSeV C proteins may interact with the same region of Stat1

As both the MuV V proteins and the SeV C proteins coimmunoprecipitate with Stat1 in our mixed-extract assay, we next determined whether this interaction was mutually exclusive, i.e., whether these two proteins could compete for interaction with a limiting amount of Stat1.

This experiment was carried out with GFP-C fusion proteins whose binding to Stat1 is also sensitive to the F170S mutation (Fig. 6A, top). From numerous assays, the SeV C proteins appear to bind Stat1 as avidly as the MuV V protein. A strong test of competitive binding was therefore to determine whether the MuV V protein could prevent subsequent C protein interaction with Stat1. MEF extracts were mixed with BSR T7 extracts containing either none, a 1⫻, or a 2⫻amount of MuV V protein for 1 h at 20°C to allow V/Stat1 complex formation. Other BSR T7 extracts containing GFP-C (or GFP-C^F170S as a

FIG. 6.Prebinding of MuV V protein to Stat1 prevents subsequent binding of SeV C protein. (A) Cytoplasmic extracts of MEFs were mixed with those of BSR T7 cells in which GFP fused to the amino-terminus of the wild-type SeV C protein (GFP-C) or the F170S mutant (GFP- C(F170S)), or the HA-tagged SeV V proteins were expressed by transfection. The mixed extracts were precipitated with anti-GFP plus anti- HA, and the precipitates were Western blotted with anti-Stat1. (B) Cytoplasmic extracts of MEFs were mixed with those of BSR T7 cells containing various relative amounts of the MuV V protein as indicated (0⫽nontransfected BSR T7 extract; 2⫻ ⫽transfected extract, and 1⫻

⫽an equal mixture of the two). After 1 h at 20°C, other BSR T7 extracts containing either GFP-C or GFP-C(F170S) were added, and after a further 1 h the mixed extracts were precipitated with anti-GFP and the amount of Stat1 present was estimated by Western blotting. Lane 5 shows the amount of Stat1 proteins present in 10% of the starting BSR T7 extract used for the immunoprecipitates in Lanes 1–4. The GFP-C and GFP-C(F170S) BSR T7 extracts contained equivalent amounts of GFP fusion protein (not shown).

(7)

negative control) were then added and after a further 1 h, the samples were precipitated with anti-GFP and the amounts of Stat1 recovered were estimated by Western blotting. As shown in Fig. 6B, increasing amounts of the MuV V protein containing BSR T7 extract increasingly interfered with the subsequent C protein interaction with Stat1. The MuV V and SeV C proteins thus appear to interact with a similar region of Stat1, as the prebinding of one prevents the subsequent binding of the other. At this stage, however, one cannot rule out the possibility that that this interference is indirect, i.e., that V binding to Stat1 induces a conformational change in Stat1 that is responsible for the loss of subsequent C/Stat interaction.

DISCUSSION

SV5, SV41, and MuV infections all lower Stat1 levels, whereas hPIV2 lowers Stat2, and this difference is con- sistent with how these rubulavirus infections interfere with IFN␣and IFN␥ signaling. However, when the MuV and hPIV2 V proteins were examined for V/Stat complex formation by coimmunoprecipitation, both V proteins unexpectedly associated with both Stat1 and Stat2. More- over, both V proteins also associated with Stat3, which was included in the study as prospective negative control. The association of the MuV V protein with the three Stat proteins nevertheless appears to be specific, in that it requires two trp residues of a trp-rich motif (W⁵/W⁹/ WCNP²²) that vaguely resembles a WW protein/protein interaction domain (Sudol and Hunter, 2000). In contrast, V/Stat complex formation in vitro does not appear to require an intact Zn^⫹⫹-binding cys cluster that vaguely resembles a RING domain (both contain seven cys that coordinate two Zn^⫹⫹ atoms) that is also notorious as a protein/protein interaction domain (Zheng et al., 2000).

Many RING finger proteins have recently been found to be components of ubiquitin-ligase complexes and to function by controlled degradation of specific target protein (Freemont, 1993, 2000). There may be two different protein interaction domains in the carboxyl segment of the rubulavirus V proteins: one that mediates Stat protein interaction and the other that mediates Stat degradation.

For SeV, a respirovirus, the V protein does not counteract IFN signaling (Garcinet al.,1999), nor interact with Stat1 in vitro(Fig. 6A). SeV V nevertheless counteracts some aspect of the innate immune response during infection of mice (Katoet al., 1997; Delendaet al., 1998), and muta- tional analysis has shown that the cys cluster is clearly critical for this effect (Huang et al., 2000). The SV5 V protein cys cluster is similarly thought to be required for V interaction with UV/DBP and possible cell-cycle control (Lin and Lamb, 2000). It will now be of interest to examine the relative importance of the trp- and cys-rich motifs of the rubulavirus V proteins in preventing IFN signaling and Stat protein degradation.

Although MuV V/Stat1 interaction in vitro requires a

specific carboxyl segment, i.e., one that contains the W⁵/W⁹/WCNP²²motif, we were surprised that the V protein appeared to interact equally with Stat1, 2, and 3. The SeV C proteins also associate with Stat1 in the mixed- extract assay, but their association with Stat2 or Stat3, in contrast, is just above the level of detection in some experiments, and absent in others (data not shown).

Stat1, 2, and 3 are part of a family of seven latent mammalian transcription factors that are structurally and functionally related (Darnell, 1997). They are activated by phosphorylation of a single tyrosine in the carboxyl end of the molecule, homo- or heterodimerized via reciprocal SH2 domains, and migrate to the nucleus where they recognize specific elements in the promoters of genes and activate transcription. Stat proteins are subsequently thought to be inactivated by dephosphorylation and returned to the cytoplasm; their activity is not normally thought to be controlled by degradation (Haspel et al., 1996). The SH2 domains of Stat1, 2, and 3 are also functionally compatible, as Stat1 can heterodimerize with both Stat2 and Stat3 (Bromberg, 2001). It is not impossi- ble that common features of Stat1, 2, and 3 are recog- nized by the rubulavirus V proteins, at least in the mixed- extract assay. Nmi-1 and CBP/p300, two cellular regula- tors of transcription, bind to the same region of most Stats (Horvath, 2000). If the amino-terminal segment of the rubulavirus V proteins also binds to Stat (the carboxyl segment is never expressed by itself), then V might interact with Stat via both amino-terminal and carboxyl segments. However, whether these interactions also take place in vivo remains to be determined, e.g., whether there is evidence of impaired Stat3 signaling during rubulavirus infections.

If the MuV and hPIV2 V proteins associate with both Stat1 and Stat2, then why does only the former only lower Stat1 levels and vice-versa? How this might occur is unknown, but there is a precedent close at hand that may be germane. All four SeV C proteins bind Stat1 and interdict IFN signaling in reporter gene assays. However, only the longer C proteins prevent the IFN-mediated antiviral state and virus-induced apoptosis during infection (Garcin et al., 2002). Moreover, only the longer C proteins induce Stat1 ubiquitination and instability during infection (Garcin et al., 2002). This suggests that the binding of Stat1 and its turnover are separable events for SeV C protein, and this may also be the case for the rubulavirus V proteins. As the determining step in protein degradation/ubiquitination is generally its association with an ubiquitin ligase (Hershko and Ciechanover, 1998), we suggested that only the longer C proteins might act as connectors to a cellular ubiquitin ligase, presumably via their amino-terminal extensions. By anal- ogy, it is possible that the MuV and hPIV2 V proteins bind both Stat1 and Stat2, but only the MuV V/Stat1 interaction leads to further interaction with an ubiquitin ligase, and vice-versa, thus accounting for the specificity of Stat

(8)

degradation. Part of this hypothetical ubiquitin–ligase connection site might also be located in the amino- terminal segment, as a mutation in this region of the SV5 V protein (N100D) determines anti-IFN activity in murine cells (Younget al.,2001).

As mentioned above, respirovirus C proteins and the amino-terminal segments of rubulavirus V proteins (up to the editing site) are similar in size and overall charge, and there is also very limited sequence conservation.

However, there are no obvious similarities between these related C-like polypeptides and the carboxyl segment of V. Nevertheless, the MuV V protein and SeV C proteins both bind to Stat1 and the binding of MuV V to Stat1 can interfere with the subsequent binding of SeV C.

There are three possible explanations for this interference: (1) SeV C and the carboxyl segment of V bind to different surfaces of Stat1, but V interaction alters subsequent C interaction allosterically. (2) C and V bind to the same surface of Stat1, even though these polypeptides appear to be unrelated structurally. (3) It is also possible that the carboxyl segment of V binds to a site adjacent to that bound by SeV C, and that the “C-like”

amino-terminal segment of V, once tethered to Stat1 via the carboxyl segment, interacts with the adjacent SeV C-binding site and interferes with subsequent SeV C association with Stat1. This latter possibility is appealing, as the presence of Stat binding domains on both the amino-terminal and the carboxyl segments of the rubulavirus V proteins, together with the advent of mRNA editing, would have facilitated the evolution of these paramyxoviruses into their present genera. As MuV NS2 does not associate with Stat, V/Stat interactionin vitro, similar to anti-IFN signaling activityin vivo, requires the carboxyl segment of V. A recombinant hPIV2 in which the carboxyl segment of V is closed by a stop codon (and thus expresses NS2 instead of V) is hypersensitive to IFN action (Kawanoet al.,2001).

MATERIALS AND METHODS Expression plasmids and transient transfections.

The MuV (Belfast strain) V, NS2, and P genes were obtained from Bert Rima (Belfast) and the previously described hPIV2 V and P genes (Nishioet al.,1996) were cloned into the NcoI (start codon) and EcoRI sites of pTM1, which contains a T7 promoter and an EMCV IRES (Mosset al.,1990). Mutant and chimeric V proteins were prepared by PCR mutagenesis, and all constructs were confirmed by sequencing. For transfections, 200,000 BSR T7 cells (that express T7 RNAP) (Buchholzet al., 1999) were plated in 3.5-cm six-well plates 20 h before transfection with 3 ␮g pTM1-V DNA and 5 ␮l of Fugene (Roche) according to the manufacturer’s instructions. Af- ter 24 h, the cells were harvested and treated as described below.

Cell extracts

MEFs in 10-cm dishes were scraped into their medium and pelleted by centrifugation (5 min at 5000 rpm in Eppendorf). Cell extracts were prepared by vortexing the cell pellet with 600␮l lysis buffer containing 0.5% NP-40, 50 mM Tris–Cl, pH 7.4, 150 to 250 mM NaCl, 10 mM EDTA, 3␮g/ml aprotinin, and 1 mM AEBSF. The disrupted cells were then centrifuged at 13,000gfor 2 min and the cytoplasmic supernatant was used for immunoprecipitation.

Antibodies and coimmunoprecipitation

Equal volumes (150␮l) of MEF and BSR T7 cytoplasmic extracts were combined and incubated for 1 h at 20°C. One of the following antibodies (1 ␮l or 1/300 dilution) was then added; a mouse monoclonal antibody to Stat1 carboxyl-terminus (Transduction Laboratories, S21120); Stat2 (Santa Cruz, sc-1668), Stat3 (Santa Cruz, sc-7179), a mAb(84-1A) to the amino-terminal region of hPIV2 P protein (Nishioet al., 1997), a rabbit polyclonal antibody (T61) to a peptide from the amino-terminal region of MuV P protein (K. Takeuchi, Tokyo), rabbit polyclonal antibody to GFP (Clontech), or anti-HA. After a further 1 h at 20°C, protein A-Sepharose beads were added and the mixture was kept at 4°C overnight. The beads were then washed three times in lysis buffer containing 250 mM NaCl, and the remaining proteins were solubilized in protein sample buffer and separated by SDS–PAGE (10 or 15%). The gels were then blotted (semidry) onto nylon membranes and detected using one of the various primary antibodies listed above. The sec- ondary antibodies used were alkaline phosphatase con- jugated goat antibodies specific for either rabbit or mouse immunoglobulin G (Bio-Rad). The immobilized proteins were detected by light-enhanced chemilumines- cence (Bio-Rad).

Note added in proof. Parisienet al.(J. Virol.76,4190–4198, 2002) have recently reported that the SV5 and hPIV2 V protein both bind Stat1 and Stat2, and that the selective degradation of each Stat protein is dependent on the presence of the other Stat protein.

REFERENCES

Boeck, R., Curran, J., Matsuoka, Y., Compans, R. W., and Kolakofsky, D.

(1992). The parainfluenza virus type 1 P/C gene uses a very efficient GUG codon to start its C⬘protein.J. Virol.66,1765–1768.

Bromberg, J. F. (2001). Activation of STAT proteins and growth control.

Bioessays23,161–169.

Buchholz, U. J., Finke, S., and Conzelmann, K. K. (1999). Generation of bovine respiratory syncytial virus (BRSV) from cDNA: BRSV NS2 is not essential for virus replication in tissue culture, and the human RSV leader region acts as a functional BRSV genome promoter.

J. Virol.73,251–259.

Cattaneo, R., Kaelin, K., Baczko, K., and Billeter, M. A. (1989). Measles virus editing provides an additional cysteine-rich protein.Cell56, 759–764.

Curran, J., Boeck, R., and Kolakofsky, D. (1991). The Sendai virus P gene

(9)

expresses both an essential protein and an inhibitor of RNA synthesis by shuffling modules via mRNA editing.EMBO J.10,3079–3085.

Curran, J., and Kolakofsky, D. (1988). Ribosomal initiation from an ACG codon in the Sendai virus P/C mRNA.EMBO J.7,245–251.

Curran, J., Marq, J.-B., and Kolakofsky, D. (1995). An N-terminal domain of the Sendai paramyxovirus P protein acts as a chaperone for the NP protein during the nascent chain assembly step of genome replication.J. Virol.69,849–855.

Curran, J., Marq, J.-B., and Kolakofsky, D. (1992). The Sendai virus nonstructural C proteins specifically inhibit viral mRNA synthesis.

Virology189,647–656.

Darnell, J. E. J. (1997). STATs and gene regulation.Science277,1630–

1635.

Delenda, C., Taylor, G., Hausmann, S., Garcin, D., and Kolakofsky, D.

(1998). Sendai viruses with altered P, V, and W protein expression.

Virology242,327–337.

Didcock, L., Young, D. F., Goodbourn, S., and Randall, R. E. (1999a).

Sendai virus and simian virus 5 block activation of interferon-respon- sive genes: Importance for virus pathogenesis. J. Virol.73,3125–

3133.

Didcock, L., Young, D. F., Goodbourn, S., and Randall, R. E. (1999b). The V protein of simian virus 5 inhibits interferon signalling by targeting STAT1 for proteasome-mediated degradation.J. Virol.73,9928–9933.

Freemont, P. S. (1993). The RING finger. A novel protein sequence motif related to the zinc finger.Ann. NY Acad. Sci. USA684,174–192.

Freemont, P. S. (2000). RING for destruction?Curr. Biol.10, R84–R87.

Garcin, D., Curran, J., and Kolakofsky, D. (2000). Sendai virus C proteins must interact directly with cellular components to interfere with interferon action.J. Virol.74,8823–8830.

Garcin, D., Latorre, P., and Kolakofsky, D. (1999). Sendai virus C proteins counteract the interferon-mediated induction of an antiviral state.

J. Virol.73,6559–6565.

Garcin, D., Marq, J. B., Strahle, L., Le Mercier, P., and Kolakofsky, D.

(2002). All four Sendai virus C proteins bind Stat1, but only the larger forms also induce its mono-ubiquitination and degradation.Virology 295,256–265.

Haspel, R. L., Salditt-Georgieff, M., and Darnell, J. E. J. (1996). The rapid inactivation of nuclear tyrosine phosphorylated Stat1 depends upon a protein tyrosine phosphatase.EMBO J.15,6262–6268.

Hershko, A., and Ciechanover, A. (1998). The ubiquitin system.Annu.

Rev. Biochem.67,425–479.

Horvath, C. M. (2000). STAT proteins and transcriptional responses to extracellular signals.Trends Biochem. Sci.25,496–502.

Huang, C., Kiyotani, K., Fujii, Y., Fukuhara, N., Kato, A., Nagai, Y., Yoshida, T., and Sakaguchi, T. (2000). Involvement of the zinc-binding capacity of Sendai virus V protein in viral pathogenesis.J. Virol.74, 7834–7841.

Jiang, X., Kowalski, J., and Kelly, J. W. (2001). Increasing protein stability using a rational approach combining sequence homology and structural alignment: Stabilizing the WW domain.Protein Sci.10,1454–

1465.

Kato, A., Kiyotani, K., Sakai, K., Yoshida, T., and Nagai, Y. (1997). The paramyxovirus, Sendai virus, V protein encodes a luxury function required for viral pathogenesis.EMBO J.16,578–685.

Kato, A., Ohnishi, Y., Kohase, M., Saito, S., Tashiro, M., and Nagai, Y.

(2001). Y2, the smallest of the Sendai virus C proteins, is fully capable of both counteracting the antiviral action of interferons and inhibiting viral RNA synthesis.J. Virol.75,3802–3810.

Kawano, M., Kaito, M., Kozuka, Y., Komada, H., Noda, N., Nanba, K., Tsurudome, M., Ito, M., Nishio, M., and Ito, Y. (2001). Recovery of infectious human parainfluenza type 2 virus from cDNA clones and properties of the defective virus without V-specific cysteine-rich domain.Virology284,99–112.

Kubota, T., Yokosawa, N., Yokota, S., and Fujii, N. (2001). C terminal cys-rich region of mumps virus structural V protein correlates with

block of interferon alpha and gamma signal transduction pathway through decrease of STAT 1-alpha.Biochem. Biophys. Res. Commun.

283,255–259.

Lamb, R. A., and Kolakofsky, D. (2001). In “Fields Virology” (D. M. Knipe and P. M. Howley, Eds.) 4th ed., pp. 1305–1340. Lippincott Williams &

Wilkins, Philadelphia.

Latorre, P., Cadd, T., Itoh, M., Curran, J., and Kolakofsky, D. (1998). The various Sendai virus C proteins are not functionally equivalent, and exert both positive and negative effects on viral RNA accumulation during the course of infection.J. Virol.72,5984–5993.

Latorre, P., Kolakofsky, D., and Curran, J. (1998). Sendai virus Y proteins are initiated by a ribosomal shunt.Mol. Cell. Biol.18,5021–5031.

Lin, G. Y., and Lamb, R. A. (2000). The paramyxovirus simian virus 5 V protein slows progression of the cell cycle.J. Virol.74,9152–9166.

Liston, P., and Briedis, D. J. (1994). Measles virus V protein binds zinc.

Virology198,399–404.

Moss, B., Elroy-Stein, O., Mizukami, T., Alexander, W. A., and Fuerst, T. R. (1990). New mammalian expression vectors.Nature348,91–92.

Murphy, S. K., and Parks, G. D. (1997). Genome nucleotide lengths that are divisible by six are not essential but enhance replication of defective interfering RNAs of the paramyxovirus Simian virus 5.

Virology232,145–157.

Nishio, M., Tsurudome, M., Ito, M., Kawano, M., Komada, H., and Ito, Y.

(2001). High resistance of human parainfluenza type 2 virus protein- expressing cells to the antiviral and anti-cell proliferative activities of alpha/beta interferons: Cysteine-rich V-specific domain is required for high resistance to the interferons.J. Virol.75,9165–9176.

Nishio, M., Tsurudome, M., Ito, M., Watanabe, N., Kawano, M., Komada, H., and Ito, Y. (1997). Human parainfluenza virus type 2 phosphoprotein: Mapping of monoclonal antibody epitopes and location of the multimerization domain.J. Gen. Virol.78,1303–1308.

Nishio, M., Tsurudome, M., Kawano, M., Watanabe, N., Ohgimoto, S., Ito, M., Komada, H., and Ito, Y. (1996). Interaction between nucleocapsid protein (NP) and phosphoprotein (P) of human parainfluenza virus type 2: One of the two NP binding sites on P is essential for granule formation.J. Gen. Virol.77,2457–2463.

Parisien, J. P., Lau, J. F., Rodriguez, J. J., Sullivan, B. M., Moscona, A., Parks, G. D., Lamb, R. A., and Horvath, C. M. (2001). The V protein of human parainfluenza virus 2 antagonizes type I interferon responses by destabilizing signal transducer and activator of transcription 2.

Virology283,230–239.

Paterson, R. G. (1995). The paramyxovirus SV5 V protein binds two atoms of zinc and is a structural component of virions.Virology208, 121–131.

Sudol, M., and Hunter, T. (2000). New wrinkles for an old domain.Cell 103,1001–1004.

Takeuchi, K., Komatsu, T., Yokoo, J., Kato, A., Shioda, T., Nagai, Y., and Gotoh, B. (2001). Sendai virus C protein physically associates with Stat1.Genes Cells6,545–557.

Thomas, S. M., Lamb, R. A., and Paterson, R. G. (1988). Two mRNAs that differ by two nontemplated nucleotides encode the amino coterminal proteins P and V of the paramyxovirus SV5.Cell54,891–902.

Vidal, S., Curran, J., and Kolakofsky, D. (1990). A stuttering model for paramyxovirus P mRNA editing.EMBO J.9,2017–2022.

Young, D. F., Chatziandreou, N., He, B., Goodbourn, S., Lamb, R. A., and Randall, R. E. (2001). Single amino acid substitution in the V protein of simian virus 5 differentiates its ability to block interferon signaling in human and murine cells.J. Virol.75,3363–3370.

Young, D. F., Didcock, L., Goodbourn, S., and Randall, R. E. (2000).

Paramyxoviridae use distinct virus-specific mechanisms to circum- vent the interferon response.Virology269,383–390.

Zheng, N., Wang, P., Jeffrey, P. D., and Pavletich, N. P. (2000). Structure of a c-Cbl-UbcH7 complex: RING domain function in ubiquitin-protein ligases.Cell102,533–539.