Assembly of Nucleocapsids with Cytosolic and Membrane-Derived Matrix Proteins of Vesicular Stomatitis Virus

Virology 261, 295–308 (1999) Article ID viro.1999.9856, available online at http://www.idealibrary.com on

Assembly of Nucleocapsids with Cytosolic and Membrane-Derived Matrix Proteins of Vesicular Stomatitis Virus E. Alexander Flood 1 and Douglas S. Lyles Molecular Genetics Program and Department of Microbiology and Immunology, Wake Forest University School of Medicine, Winston-Salem, North Carolina 27157-1064 Received March 31, 1999; returned to author for revision May 19, 1999; accepted June 17, 1999 During budding of vesicular stomatitis virus (VSV), the viral matrix (M) protein binds the viral nucleocapsid to the host plasma membrane and condenses the nucleocapsid into the tightly coiled nucleocapsid–M protein (NCM) complex observed in virions. In infected cells, the viral M protein exists mostly as a soluble molecule in the cytoplasm, and a small amount is bound to the plasma membrane. Despite the high concentrations of M protein and intracellular nucleocapsids in the cytoplasm, they are not associated with each other except at the sites of budding. The experiments presented here address the question of why M protein and nucleocapsids associate with each other only at the plasma membrane but not in the cytoplasm of infected cells. An assay for exchange of soluble M protein into NCM complexes in vitro was used to show that both cytosolic and membrane-derived M proteins bound to virion NCM complexes with affinities similar to that observed for virion M protein, indicating that both cytosolic and membrane-derived M proteins are competent for virus assembly. However, neither cytosolic nor membrane-derived M protein bound to intracellular nucleocapsids with the same high affinity observed for virion NCM complexes. Cytosolic M protein was able to bind intracellular nucleocapsids, but with an affinity approximately eightfold less than that observed in virion NCM complexes. Membrane-derived M protein exhibited little or no binding activity for intracellular nucleocapsids. These data indicate that intracellular nucleocapsids, and not intracellular M proteins, need to undergo an assembly-initiating event in order to assemble into an NCM complex. Since neither membrane-derived nor cytosolic M protein could initiate high-affinity binding to intracellular nucleocapsids, the results suggest that another viral or host factor is required for assembly of the NCM complex observed in virions. © 1999 Academic Press

capsids incorporated into progeny virions as a result of their association with M protein. In virions, M protein is present in approximately a 2:1 molar ratio to the N protein and it condenses the nucleocapsid into a tightly coiled helical nucleocapsid–M protein (NCM) complex with bullet-like morphology (Barge et al., 1993; Lyles and McKenzie, 1998; Lyles et al., 1996b; Newcomb and Brown, 1981; Thomas et al., 1985). The mechanism of assembly of M protein with the nucleocapsid is the topic of this report. The majority of intracellular M protein in VSV-infected cells is present as a soluble molecule in the cytoplasm, although 10–20% of M protein is bound to the plasma membrane (Chong and Rose, 1993; Knipe et al., 1977). Cytosolic M protein is mostly monomeric (McCreedy et al., 1990) and enters rapidly into virions from a soluble pool (Knipe et al., 1977). Curiously, the cytoplasm of infected cells contains high concentrations of intracellular nucleocapsids and cytosolic M protein, yet they do not colocalize with each other (Lyles et al., 1988; Ohno and Ohtake, 1987; Ono et al., 1987). Possible reasons for the failure of cytoplasmic nucleocapsids to bind M protein include (1) either M protein or nucleocapsids require a maturation step to become competent for virus assembly, (2) there may be cytoplasmic inhibitors that prevent

INTRODUCTION Enveloped viruses acquire their lipid envelopes by budding through host cell membranes, which is mediated by interactions of viral membrane proteins with the viral nucleocapsid core (Garoff et al., 1998). In the case of vesicular stomatitis virus (VSV), the prototype member of the Rhabdovirus family, the viral matrix (M) protein has been shown to bind the nucleocapsid during virus assembly by budding from host plasma membranes (McCreedy and Lyles, 1989; Odenwald et al., 1986; Ohno and Ohtake, 1987). Viral nucleocapsids are assembled during the process of RNA replication in the cytoplasm of infected cells (Wertz et al., 1987). Both negative strand genomes and positive strand antigenomes are encapsidated by the major nucleocapsid (N) protein as they are synthesized. These nucleocapsids also contain the viral L and P protein subunits of the viral RNA polymerase. Intracellular nucleocapsids containing negative strand RNA serve three functions: they are (a) the templates for the transcription of viral mRNA, (b) the templates for replication of antigenomes, and (c) the source of nucleo1 To whom correspondence and reprint requests should be addressed. Fax: 336-716-7200. E-mail: [email protected].

295

0042-6822/99 $30.00 Copyright © 1999 by Academic Press All rights of reproduction in any form reserved.

296

FLOOD AND LYLES

binding, or (3) the cytoplasm may lack a factor that initiates assembly. Such an assembly-initiating factor would likely be located in the plasma membrane, since this is the only place in infected cells where binding of M protein to the nucleocapsid and condensation into an NCM complex has been observed (McCreedy and Lyles, 1989; Odenwald et al., 1986; Ohno and Ohtake, 1987; Ono et al., 1987). An attractive hypothesis is that membranebound M protein serves as a recognition site for binding intracellular nucleocapsids to the plasma membrane to initiate assembly. Alternatively, membrane-bound M protein may serve other functions in virus assembly that have been ascribed to M protein. For example, M protein contains an intrinsic budding activity; cells expressing M protein in the absence of other VSV proteins release vesicles containing M protein from the plasma membrane (Justice et al., 1995; Li et al., 1993; Lyles et al., 1996b). M protein also stabilizes trimer formation of the viral envelope glycoprotein (G protein) and facilitates organization of the G protein into patches in membranes (Luan and Glaser, 1994; Luan et al., 1995; Lyles et al., 1992; Reidler et al., 1981). The in vitro binding of M protein to nucleocapsids that results in a nucleocapsid structure similar to that observed in virions likely represents a native interaction in virions. When virions are treated with detergent at low ionic strength, M protein remains bound to nucleocapsids, yielding tightly coiled NCM complexes morphologically similar to those in native virions. Raising the ionic strength leads to dissociation of M protein from nucleocapsids, resulting in loosely coiled and flexible nucleocapsids (Barge et al., 1993; Newcomb and Brown, 1981). Binding and dissociation of M protein in NCM complexes have been analyzed biochemically by sedimentation and light scattering assays (Lyles and McKenzie, 1998). These analyses demonstrated that dissociation of M protein in NCM complexes consists of two phases, a reversible phase and an irreversible phase. Reversible binding of M protein to the NCM complex occurs at physiological ionic strength (e.g., 120 mM NaCl), where M protein partially dissociates from the NCM complex. The extent of dissociation depends on the concentration of M protein and nucleocapsids, as well as the ionic strength, and has an apparent equilibrium constant of 0.4 mM at 120 mM NaCl. The irreversible phase occurs at higher ionic strengths (e.g., $250 mM NaCl), which leads to nearly complete dissociation of M protein from the NCM complex. Nucleocapsids stripped of M protein at high ionic strength cannot bind M protein at physiological ionic strength (i.e., 120 mM NaCl). It was proposed that, in infected cells, assembly of M protein with nucleocapsids might also proceed through two steps, corresponding to the two phases observed during dissociation of M protein from virion NCM complexes. The first step in assembly of an NCM complex would be a

process that initiates high-affinity binding of M protein, corresponding to the process that is irreversibly inactivated by high ionic strength. The transition to high-affinity binding of M protein would be followed by a separate step in which most of the M protein is recruited into the NCM complex, corresponding to the reversible interaction at physiological ionic strength. We have developed an in vitro system capable of measuring the binding of M protein and nucleocapsids derived from either virions or from infected cells. In this report we determined the ability of intracellular M protein and nucleocapsids to assemble into NCM complexes. The results indicated that both cytosolic and membranederived M proteins were competent to assemble into virion NCM complexes with an affinity similar to that observed for virion M protein. However, neither cytosolic nor membrane-derived M proteins bound to intracellular nucleocapsids with the same affinity observed for virion NCM complexes. This indicated that intracellular nucleocapsids, and not the intracellular M proteins, need to undergo an assembly-initiating event in order to assemble into an NCM complex. These data suggest that neither membrane-bound nor cytosolic M protein alone can initiate assembly of NCM complexes. RESULTS Binding of M protein from cytosolic or membrane fractions of infected cells to NCM complexes An M protein exchange assay was used to determine if cytosolic M protein was competent to assemble into NCM complexes (Fig. 1A). We had previously demonstrated that M protein present in the NCM complex reversibly dissociates and reassociates with the nucleocapsid. Therefore, if exogenous radiolabeled M protein is mixed with NCM complexes, the radiolabeled M protein will exchange with the endogenous M protein (Lyles and McKenzie, 1998). The amount of radiolabeled M protein bound to the nucleocapsid reflects its binding affinity relative to the endogenous M protein. The binding affinity of M protein for NCM complexes is dependent on the salt concentration, with higher salt concentrations favoring dissociation of M protein (Lyles and McKenzie, 1998). The exchange of cytosolic M protein into virion NCM complexes was tested at physiologic ionic strength (120 mM NaCl) using protein concentrations (30–40 mg/ml of virion protein) that resulted in dissociation of 80% of endogenous M protein from the NCM complexes. A radiolabeled cytosolic fraction was prepared from VSV-infected baby hamster kidney (BHK) cells pulse labeled with [ 35S]methionine at 4 h postinfection. Radiolabeled cytosol was mixed with non-radiolabeled purified virions as a source of NCM complexes in buffer containing 120 mM NaCl (Fig. 1B). The detergent Triton X-100 was added to solubilize the virion envelope. The mixture

ASSEMBLY OF VSV NCM COMPLEXES

FIG. 1. M protein exchange assay. (A) Schematic representation of reversible binding of M protein in NCM complexes. Radiolabeled soluble M protein (closed circles) mixed with virion NCM complexes will exchange with the endogenous virion M protein (open circles) originally present in virion NCM complexes. Incubation at physiological ionic strength (120 mM NaCl) results in partial dissociation of NCM complexes. The amount of radiolabeled M protein bound to the NCM complex reflects its binding affinity relative to the endogenous virion M protein. (B) Binding of cytosolic M protein of VSV, Indiana serotype (VSV-IND), to NCM complexes from different VSV serotypes and different virus families at 120 mM NaCl. A radiolabeled cytosolic fraction isolated from VSV-IND-infected BHK cells was mixed with equivalent amounts (40 mg/ml) of non-radiolabeled purified VSV-IND (lanes 4, 5), VSV, New Jersey serotype (VSV-NJ) (lanes 6, 7), influenza virus (infl.-X47) (lanes 8, 9), and Sendai virus Z strain (Sen. Z) (lanes 10, 11) as sources of NCM complexes. Triton X-100 was added to solubilize the virion envelopes, and the mixtures were incubated at room temperature for 5 min and then were centrifuged. The supernatant (S) and pellet (P) fractions were analyzed by SDS–PAGE, phosphorescence images of which are shown. As a control, the radiolabeled cytosolic fraction isolated from VSV-infected cells was treated in the same manner but without the addition of virions as a source of nucleocapsids (lanes 2, 3). Radiolabeled purified virions (lane 1) were used as protein markers.

297

was incubated at room temperature for 5 min, which allows for complete equilibration of M protein in VSV NCM complexes (Lyles and McKenzie, 1998), and then was centrifuged to pellet the nucleocapsids and any M protein associated with the nucleocapsid. The supernatants and pellets were analyzed by SDS–PAGE and phosphorescence imaging. In these experiments the radiolabeled M protein was added in trace amounts (;100-fold less intracellular M protein than the endogenous virion M protein in the NCM complex), so that the distribution of endogenous M protein between bound and unbound states was not perturbed by the addition of exogenous M protein. As a control, the radiolabeled cytosolic fraction was treated in the same manner but without virions as a source of NCM complexes. This measured the amount of M protein that was detected in the pellet fraction for nonspecific reasons. As additional specificity controls, the radiolabeled cytosolic fraction prepared from cells infected with VSV-Indiana (VSV-IND) was mixed with equivalent amounts of purified VSV of another serotype, VSV-New Jersey (VSV-NJ), or with purified viruses of other families, influenza virus, and Sendai virus. A phosphorescence image from one such experiment is shown in Fig. 1B. When the radiolabeled cytosolic fraction from VSV-IND-infected cells was mixed with nonradiolabeled purified VSV-IND or -NJ serotypes, the cytosolic M protein was found in both the supernatant and the pellet fractions (Fig. 1B, lanes 4–7), reflecting the exchange of cytosolic M protein into NCM complexes that had partially dissociated at 120 mM NaCl. The heterotypic binding of cytosolic M protein of VSV-IND to VSV-NJ NCM complexes was expected based on the finding that the M proteins of VSV-IND and -NJ are able to complement temperature-sensitive M-gene mutants of both serotypes (Sun et al., 1994). In the absence of virions as a source of NCM complexes, only a trace quantity of M protein was detected in the pellet fraction (lane 3). Similarly, in the presence of influenza or Sendai virus NCM complexes, only trace quantities of M protein were detected in the pellet fraction (Fig. 1B, lanes 9 and 11). In additional experiments, no binding above background was detected with twice the concentration of influenza or Sendai viruses (80 mg/ml) or when membrane-derived M protein was mixed with influenza or Sendai viruses (data not shown). These results indicate that cytosolic M protein is able to exchange with VSV NCM complexes but not with those of influenza and Sendai viruses (Fig. 1B, lanes 5 and 7 versus 9 and 11). The cytosolic fraction also contains soluble N, P, and L viral proteins. These proteins also associated with VSV NCM complexes but not with those of influenza and Sendai viruses. It has not been determined if this binding represents an exchange of cytosolic N, P, and L proteins for proteins in the NCM complex, as is the case for M protein (Lyles and McKenzie, 1998).

298

FLOOD AND LYLES

It has been proposed that M protein on the cytoplasmic surface of host plasma membranes provides a binding site for intracellular nucleocapsids to initiate budding (Chong and Rose, 1993; McCreedy and Lyles, 1989). Thus, it might be expected that membrane-derived M protein would have a higher affinity than cytosolic M protein for binding nucleocapsids. The M protein exchange assay was used to compare the affinities of cytosolic versus membrane-derived M proteins for binding virion NCM complexes. These experiments were performed at two different ionic strengths, 120 mM NaCl (Fig. 2A) and 60 mM NaCl (Fig. 2B), since the affinity of M protein binding to NCM complexes is dependent on ionic strength (Lyles and McKenzie, 1998). When radiolabeled cytosolic or membrane-derived fractions from infected cells were mixed with non-radiolabeled purified virions at 120 mM NaCl, both cytosolic and membrane-derived M proteins were found in the pellet fraction (lanes 3 and 7), reflecting the exchange of intracellular M protein into NCM complexes. Similar results were obtained at 60 mM NaCl, except that a greater amount of both cytosolic and membrane-derived M protein bound to the NCM complex, reflecting the salt-dependent binding of M protein (compare lanes 5 and 7 of Figs. 2A and 2B). The membrane fraction also contained the viral G, N, P, and L proteins. G protein was detected in the supernatant as expected for a solubilized membrane protein (lanes 6 and 8). However, the N, P, and L proteins were detected in the pellet in both the presence (lane 7) and the absence (lane 9) of virion NCM complexes. These probably represent nucleocapsids associated with the plasma membrane in the process of virus assembly (McCreedy and Lyles, 1989; Odenwald et al., 1986), which became released from the membrane by the addition of detergent. These nucleocapsids released from membranes may bind a small amount of the solubilized membranederived M protein, as shown by the increased level of membrane-derived M protein detected in the pellet in the absence of virions compared to cytosolic M protein, particularly at 60 mM NaCl (lane 9 versus lane 5, Fig. 2B). Data from three independent experiments were quantitated, and the percentage of M protein in the pellet was calculated as the amount of M protein in the pellet divided by the sum of M protein in supernatant and pellet (Fig. 2C). Both cytosolic and membrane-derived M protein bound to NCM complexes to similar extents, approximately 55–65% at 60 mM NaCl (black bars) and approximately 20% at 120 mM NaCl (hatched bars). In the absence of virions as a source of NCM complexes there was less than 5% of M protein in the pellet fraction, except in the case of the membrane fraction at 60 mM NaCl, as noted above. These data indicate that both cytosolic and membrane-derived M proteins were competent to assemble into NCM complexes and that there was little or no difference between cytosolic and mem-

brane-derived M proteins in their ability to bind to NCM complexes. It is possible that infected cells contain activities that either promote or inhibit the association of M protein and nucleocapsids. We reasoned that any activities present in infected cells which affected the binding of M protein and nucleocapsids in the cytoplasm may also affect the binding of M protein to NCM complexes from virions. Thus, either an increase or a decrease in the extent of dissociation of virion M protein from NCM complexes due to the addition of a subcellular fraction would indicate the presence of an activity that affected the binding of M protein and nucleocapsids. To determine if the cytosolic and membrane fractions contained such an activity an assay identical to that in Fig. 2 was used, except that in these experiments purified virions were radiolabeled and the subcellular fractions were not radiolabeled. Purified 35S-labeled virions were mixed with non-radiolabeled cytosolic or membrane fractions isolated from infected cells; the membranes and virion envelopes were solubilized with detergent and the binding of M protein to NCM complexes was allowed to equilibrate and then was centrifuged. The resulting supernatant and pellet fractions were analyzed by SDS–PAGE and the amount of virion M protein that bound to the NCM complexes was quantitated by phosphorescence imaging. Figure 3 shows data from one such experiment performed at 120 mM (Fig. 3A) and 60 mM (Fig. 3B) NaCl, as well as quantitation of results from three independent experiments (Fig. 3C). When radiolabeled virions were mixed with non-radiolabeled cytosolic or membrane fractions from infected cells at 120 mM NaCl, M protein was found in both the supernatant and pellet fractions (lanes 1–4, Fig. 3A), reflecting the partial dissociation of endogenous virion M protein from the NCM complex. The distribution of M protein between supernatant and pellet fractions was similar to that obtained in the absence of either subcellular fraction (lanes 5 and 6), with approximately 20% of M protein bound to NCM complexes in either the presence or the absence of subcellular fractions (Fig. 3C, hatched bars). Similar results were obtained at 60 mM NaCl, except that less M protein dissociated from the NCM complex, resulting in approximately 75% of M protein detected in the pellet in both the presence and the absence of subcellular fractions (Fig. 3C, black bars). These results indicate that the cytosolic and membrane fractions isolated from VSV-infected BHK cells did not contain activities that affected the binding of virion M protein to NCM complexes. Also, neither subcellular fraction contained sufficient M protein to increase the binding of M protein by a mass action effect (100-fold less intracellular M protein than virion-derived M protein). The binding of cytosolic and membranederived M protein to NCM complexes (Fig. 2) was similar to the binding of endogenous virion M protein (Fig. 3),

ASSEMBLY OF VSV NCM COMPLEXES

299

although intracellular M proteins bound to NCM complexes slightly less than endogenous virion M protein. Overall, these results indicate that (1) M proteins from infected cell cytosolic or membrane-derived fractions are competent to assemble into NCM complexes, (2) intracellular M proteins bind with relative affinities similar to that of endogenous virion M protein, and (3) binding of M protein to NCM complexes is not affected by other components in the subcellular fractions. Composition and morphology of intracellular nucleocapsids and virion nucleocapsids stripped of M protein

FIG. 2. Binding of cytosolic and membrane-derived M protein to NCM complexes at 120 mM NaCl (A) or 60 mM NaCl (B) measured by M protein exchange assay described in Fig. 1. 35S-labeled cytosolic (lanes 2–5) or membrane-derived (lanes 6–9) fractions from infected cells were incubated in the presence (lanes 2, 3, 6, 7) or the absence (lanes 4, 5, 8, 9) of purified virions (30 mg/ml total protein). Triton X-100 was added to solubilize the virion envelopes and cellular membranes. The mixtures were incubated at room temperature for 5 min and then were centrifuged. Supernatant (S) and pellet (P) fractions were analyzed by SDS–PAGE, phosphorescence images of which are shown. Radiolabeled purified virions (lane 1) were used as protein markers. (C) The amount of radiolabeled M protein in each fraction was quantitated

Intracellular nucleocapsids are not observed to colocalize with M protein in the cytoplasm of infected cells (Lyles et al., 1988; Ohno and Ohtake, 1987; Ono et al., 1987). This suggests that intracellular nucleocapsids are not able to bind M protein. Therefore, the ability of intracellular nucleocapsids to bind cytosolic M protein was determined using the same conditions that demonstrated binding of M protein to virion NCM complexes. Intracellular nucleocapsids were isolated at 5 h postinfection from BHK cells lysed in buffer containing Triton X-100 and 100 mM NaCl, similar to previously described methods (Naeve et al., 1980). This salt concentration preserves the reversible binding and dissociation of M protein in NCM complexes from virions (Lyles and McKenzie, 1998). The intracellular nucleocapsids were isolated from the postnuclear lysate by sucrose gradient centrifugation in the presence of 100 mM NaCl and then dialyzed against 10 mM NaCl buffer. This method avoided the use of high ionic strength buffers, since virion NCM complexes treated with high ionic strength, e.g., $250 mM NaCl, lose their ability to bind M protein. The protein composition and morphology of virion nucleocapsids and intracellular nucleocapsids isolated at different ionic strengths were determined by SDS–PAGE and staining with Coomassie blue (Fig. 4A) or by negative stain electron microscopy (Figs. 4B and 4C). Virion nucleocapsids isolated at 250 mM NaCl contained the viral N, P, and L proteins but were largely devoid of M protein (lane 2, Fig. 4A). These nucleocapsids were termed stripped nucleocapsids. Intracellular nucleocapsids exhibited a protein profile similar to that of stripped nucleocapsids, although they were isolated at 100 mM NaCl. They contained the viral N, P, and L proteins and were also largely devoid of M protein (lane 3). When larger numbers of intracellular nucleocapsids were ana-

by phosphorescence imaging and is expressed as the percentage of M protein in the pellet. Data shown are the means 6SD from three independent experiments in the presence of 60 mM NaCl (solid bars) or 120 mM NaCl (hatched bars).

300

FLOOD AND LYLES

lyzed, host proteins and trace quantities of M protein were detected, consistent with a low level of contamination with other intracellular components (lanes 4–6). Electron microscopy of intracellular nucleocapsids, isolated at 10 mM NaCl, showed that they were highly extended and were often coiled around each other (Fig. 4B). Single nucleocapsid strands were visible in regions where nucleocapsids had become uncoiled (Fig. 4B, arrow). Stripped nucleocapsids isolated from virions appeared quite similar to intracellular nucleocapsids after dialysis against buffer containing 10 mM NaCl (Fig. 4C). These results indicate that although intracellular and stripped nucleocapsids were isolated at different salt concentrations they had similar protein compositions and morphologies. Binding of intracellular nucleocapsids to intracellular M protein

FIG. 3. Binding of virion M protein to NCM complexes at 120 mM (A) or 60 mM (B) NaCl in the presence of subcellular fractions from VSV-infected BHK cells. Purified 35S-labeled virions (40 mg/ml total protein) were mixed with non-radiolabeled cytosolic (lanes 1 and 2) or membrane-derived fractions (lanes 3 and 4) or in the absence of either subcellular fraction (lanes 5 and 6). Triton X-100 was added to solubilize the virion envelopes and cellular membranes. The mixtures were incubated at room temperature for 5 min and then were centrifuged. Supernatant (S) and pellet (P) fractions were analyzed by SDS–PAGE, phosphorescence images of which are shown. (C) The amount of

To determine if intracellular nucleocapsids could bind intracellular M protein, radiolabeled cytosolic or membrane-derived fractions isolated from VSV-infected BHK cells at 4.5 h postinfection were mixed with increasing amounts of intracellular nucleocapsids at 120 mM NaCl. Radiolabeled subcellular fractions were mixed with virion NCM complexes as a positive control and with stripped nucleocapsids as a negative control. The samples were allowed to equilibrate, then centrifuged, and the supernatant and pellet fractions were analyzed by SDS–PAGE and phosphorescence imaging (Fig. 5). In order to compare the binding of M protein to the three different nucleocapsid species, the concentration of nucleocapsid protein was normalized to the number of free binding sites for M protein present on each nucleocapsid species. Nucleocapsid protein concentrations were determined by quantitating the amount of N protein in each preparation by SDS–PAGE and Coomassie blue staining. Each N protein can bind two molecules of M protein (Lyles et al., 1996a; Thomas et al., 1985). Therefore, for intracellular or stripped nucleocapsids, an N protein concentration of 3 mM corresponded to 6 mM free M protein binding sites. For NCM complexes, the number of free M protein binding sites was calculated by subtracting the number of sites occupied by endogenous virion M protein from the total number of sites on the NCM complex at each protein concentration using a dissociation constant of 0.4 mM (120 mM NaCl) (Lyles and McKenzie, 1998). For example, at the highest concentration of nucleocapsids in Fig. 5, the N protein concentration was 3.9

radiolabeled M protein in each fraction was quantitated by phosphorescence imaging and is expressed as the percentage of M protein in the pellet. Data shown are the means 6 SD from three independent experiments in the presence of 60 mM NaCl (solid bars) or 120 mM NaCl (hatched bars).

ASSEMBLY OF VSV NCM COMPLEXES

FIG. 4. Analysis of intracellular and virion nucleocapsids. Intracellular nucleocapsids were isolated by sucrose gradient centrifugation in the presence of 100 mM NaCl. Virion nucleocapsids were stripped of M protein by incubation in the presence of 250 mM NaCl and isolated by sucrose gradient centrifugation. (A) SDS–PAGE and Coomassie blue staining of virions (1.8 mg N protein, lane 1), virion nucleocapsids (0.5 mg N protein, lane 2), and intracellular nucleocapsids (0.5, 1.0, 2.0, 4.0 mg N protein, lanes 3–6, respectively). Isolated intracellular (B) and virion (C) nucleocapsids were analyzed by negative stain electron microscopy after dialysis against 10 mM Tris, pH 7.5, 10 mM NaCl. Arrow indicates region of single nucleocapsid strand. Bar, 50 nm.

mM; therefore, the total concentration of M protein binding sites was 7.8 mM. However, 71% of these sites were occupied by endogenous virion M protein, thus reducing

301

the concentration of free M protein binding sites to 2.3 mM. Cytosolic M protein bound with high affinity to NCM complexes (Fig. 5A, black squares), as expected from the data in Fig. 2. Stripped nucleocapsids bound little if any M protein (open squares), since they have lost the ability to bind M protein after isolation at high ionic strength (Lyles and McKenzie, 1998). Intracellular nucleocapsids exhibited an affinity for binding cytosolic M protein that was intermediate between that of NCM complexes and intracellular nucleocapsids (open circles). The apparent affinity of intracellular nucleocapsids and cytosolic M protein was approximately eightfold less than that determined for NCM complexes (apparent K d 5 14 mM versus 1.7 mM, respectively). Similar results were obtained for binding of M protein isolated from virions to intracellular nucleocapsids (data not shown). These data show that cytosolic M protein bound to intracellular nucleocapsids, but the affinity was much lower than that of NCM complexes from virions. Binding of intracellular nucleocapsids to membranederived M protein was analyzed as described for cytosolic M protein. Membrane-derived M protein bound with high affinity to NCM complexes (Fig. 5B, solid squares) and exhibited little if any binding activity to stripped nucleocapsids (open squares). However, membrane-derived M protein also exhibited little binding activity for intracellular nucleocapsids (circles). These experiments failed to support the hypothesis that membrane-derived M protein may serve as a recognition site for binding intracellular nucleocapsids at the plasma membrane. It is possible that solubilizing the membranes with detergent inactivated any binding activity contained in the membrane fraction from infected cells. Therefore, binding of nucleocapsids to membranes from VSV-infected cells was analyzed using membranes that were not solubilized with detergent (Fig. 6). It has been reported that stripped nucleocapsids isolated from virions bind membranes containing M protein (Chong and Rose, 1993; Ogden et al., 1986). The same procedure used previously (Chong and Rose, 1993; Ogden et al., 1986) was used to measure membrane binding of radiolabeled stripped nucleocapsids mixed with membranes isolated from VSV-infected (squares) and mock-infected (diamonds) BHK cells or in the absence of membranes (circles). The samples were analyzed by flotation in a discontinuous sucrose gradient and gradient fractions were analyzed by SDS–PAGE. The amount of radiolabeled N protein was quantitated as a measure of the numbers of nucleocapsids in each fraction and was plotted as the percentage of total N protein present in the gradient. The majority of the nucleocapsids remained at the bottom of the gradient. The membranes floated to fractions 2 and 3 in these gradients. There was no reproducible difference in the amount of N protein in the

302

FLOOD AND LYLES

fractions containing membranes from infected versus uninfected cells. VSV-infected (squares) and mock-infected (diamonds) membranes contained 4% and 2% of the total N protein in the gradient, respectively. In the absence of membranes (circles), an equivalent amount of N protein was detected in these fractions (6% of total N protein), indicating that only background levels of N protein were detected in these fractions. Similar results were obtained over a pH range from 7.1 to 8.1, as well as in sucrose gradients without any NaCl in the gradient (data not shown). These experiments did not detect activities in membranes isolated from infected cells that promoted the association of nucleocapsids with the plasma membrane. Homogeneity of the interaction between intracellular nucleocapsids and M protein

FIG. 5. Binding of cytosolic and membrane-derived M protein to intracellular and virion nucleocapsids at 120 mM NaCl. Radiolabeled cytosolic (A) or membrane (B) fractions isolated from VSVinfected BHK cells were mixed with increasing amounts of purified virions as a source of NCM complexes (solid squares), intracellular nucleocapsids (circles), or virion nucleocapsids stripped of M protein (open squares). Triton X-100 was added to solubilize the virion envelopes. The mixtures were incubated at room temperature for 5 min and then were centrifuged. Supernatant and pellet fractions were analyzed by SDS–PAGE, and the amount of M protein was quantitated by phosphorescence imaging. Binding of M protein to the three different nucleocapsid species is shown plotted against the number of free binding sites for M protein present on each nucleocapsid species (calculated from the concentrations of N and M proteins). The binding data of cytosolic M protein (A) for NCM complexes are means from seven independent experiments; the standard deviations of each point were 66.7– 8.6%. Data for stripped nucleocapsids are means from four independent experiments; standard deviations of each point were 61.1– 2.6%. Data for intracellular nucleocapsids were obtained from three independent experiments, where each point represents the mean value from duplicate or triplicate samples within each experiment. The binding data of membrane-derived M protein (B) for NCM complexes are means from three independent experiments; the standard deviation of each point ranged from 60.5 to 6.7%, with one exception of 611.7%. Data for stripped nucleocapsids are means from three independent experiments; the standard deviation of each point ranged from 60.4 to 5.2%. Data for intracellular nucleocapsids were obtained from two independent experiments. Each point represents the mean value from triplicate samples within each experiment. The standard deviation of each point ranged from 62.2 to 6.8%.

The intermediate affinity of cytosolic M protein binding to intracellular nucleocapsids (Fig. 5A) may reflect the presence of a mixture of nucleocapsids containing varying affinities for M protein. For example, intracellular nucleocapsids destined for virus assembly may exhibit a high affinity for M protein, while those destined to remain intracellular may exhibit a low affinity. To determine if intracellular nucleocapsids were homogeneous or heterogeneous for binding M protein, intracellular nucleocapsids were sedimented through a sucrose gradient in the presence and the absence of M protein (Fig. 7). Intracellular nucleocapsids that bound M protein should

FIG. 6. Binding of virion nucleocapsids stripped of M protein to membranes analyzed by membrane flotation. 35S-labeled nucleocapsids (10 5 CPM) isolated from virions were mixed with membranes isolated from 3 3 10 7 VSV-infected cells (squares), mock-infected cells (diamonds), or without membranes (circles). The reactions were incubated for 1 h at 37°C, mixed with 66% sucrose, and overlaid with 40 and 10% (w/w) sucrose solutions. The samples were centrifuged for 16 h to float the membranes. Fractions were collected from the bottom and analyzed by SDS–PAGE. The amount of N protein in each fraction was quantitated by phosphorescence imaging and is plotted as the percentage of total N protein. The membranes were in fractions 2 and 3.

ASSEMBLY OF VSV NCM COMPLEXES

FIG. 7. Sedimentation of intracellular and virion nucleocapsids stripped of M protein in gradients containing M protein. Intracellular (A) and stripped (B) nucleocapsids were sedimented through 10–60% (w/v) sucrose gradients containing (squares) or lacking (circles) M protein throughout the gradient. The source of M protein in the gradients was a soluble extract obtained from virions treated with Triton X-100 at 250 mM NaCl. Sucrose gradients containing M protein throughout were generated by dissolving sucrose into the soluble extract containing M protein to yield 10 and 60% (w/v) sucrose solutions containing equal amounts of M protein with final concentrations of 125 mM NaCl, 0.25% Triton X-100, and approximately 0.2 mg/ml (7 mM) M protein. For each fraction, the amount of N protein was determined by SDS–PAGE and phosphorescence imaging. The top of the gradient is indicated; P indicates material recovered from the pellet fraction.

be separated from those that did not due to their differences in mass. Since the M protein–nucleocapsid interaction is dynamic, M protein was present throughout the gradient to stabilize formation of the M protein–nucleocapsid complex. The source of M protein in the gradient was a soluble extract from virions treated with Triton X-100 in the presence of 250 mM NaCl. This extract also contained the viral G protein and envelope lipids. The binding activity of M protein isolated at 250 mM NaCl is not inactivated, since it can still bind to NCM complexes from virions at 120 or 60 mM NaCl (Lyles and McKenzie,

303

1998). The soluble extract from virions was diluted to give a salt concentration in the gradient of 125 mM NaCl. Sedimentation of intracellular nucleocapsids in gradients containing M protein (Fig. 7A, squares) resulted in a single peak of slightly increased sedimentation velocity compared to intracellular nucleocapsids in gradients without M protein (circles). This result indicates that intracellular nucleocapsids are homogeneous for binding M protein since all the nucleocapsids were shifted to a faster sedimentation velocity. As a control, stripped nucleocapsids, which do not bind M protein, were sedimented in sucrose gradients in the presence and the absence of M protein. Stripped nucleocapsids sedimented to similar positions in each gradient as expected (Fig. 7B), since they do not bind M protein. The data in Fig. 7 are consistent with those in Fig. 5 in showing that intracellular nucleocapsids bind cytosolic M protein, while stripped nucleocapsids do not. Taken together, intracellular nucleocapsids behave as a homogeneous population that have a reduced affinity for M protein compared to NCM complexes from virions. In order to determine if the low-affinity binding of M protein to intracellular nucleocapsids affected their morphology, intracellular nucleocapsids were incubated in the presence or the absence of M protein and analyzed by negative stain electron microscopy (data not shown). Nucleocapsids treated at 120 mM NaCl appeared mostly as a loosely coiled helix with a diameter of 20.0 6 1.2 nm. This was markedly different from the highly extended structure visible in nucleocapsids prepared in 10 mM NaCl (Fig. 4C), although uncoiled regions of nucleocapsid with an extended structure were often apparent in samples treated with 120 mM NaCl. Addition of M protein to intracellular nucleocapsids did not change their morphology much beyond the changes induced by the higher salt concentration alone, although binding of M protein resulted in slightly more coiled regions with a diameter of 17.0 6 1.0 nm. These results are consistent with the idea that the relatively low-affinity binding of M protein to intracellular nucleocapsids is not sufficient to induce condensation of the intracellular nucleocapsids into a structure resembling the NCM complex observed in virions. DISCUSSION The experiments presented here address the question of why M protein and nucleocapsids assemble into NCM complexes at the plasma membrane but not in the cytoplasm of infected cells. One possibility was that cytosolic M protein cannot bind to nucleocapsids, while membrane-derived M protein can bind. However, both cytosolic and membrane-derived M proteins were competent to assemble into virion NCM complexes with affinities similar to that of virion M protein

304

FLOOD AND LYLES

(Figs. 2 and 3). Another possibility was that the cytosolic fraction from infected cells contained an inhibitor of M protein binding. However, neither cytosolic nor membrane-derived fractions from infected cells contained activities that affected binding of virion M protein to NCM complexes (Fig. 3). A third possibility is that intracellular nucleocapsids are not able to bind M protein. In support of this hypothesis, neither cytosolic nor membrane-derived M proteins bound to intracellular nucleocapsids with the same affinity observed for virion NCM complexes (Fig. 5). Therefore, intracellular nucleocapsids, and not intracellular M proteins, need to undergo a transition to high-affinity binding in order to assemble into NCM complexes observed in virions. Similar to results obtained with intracellular nucleocapsids, neither cytosolic nor membrane-derived M protein bound to virion nucleocapsids stripped of M protein at high ionic strength (Fig. 5). This result is consistent with previous data showing that dissociation of M protein from nucleocapsids at high ionic strength is irreversible (Lyles and McKenzie, 1998). At very low ionic strength (e.g., 10 mM NaCl), M protein will bind to virion nucleocapsids that have been stripped of M protein at high ionic strength (Lyles and McKenzie, 1998; Newcomb et al., 1982), and this binding is sufficient to condense nucleocapsids into the bullet-like shape observed in virion NCM complexes (Newcomb et al., 1982). However, this reassociation results in the formation of NCM complexes with strikingly different biochemical properties from those of NCM complexes assembled in vivo (Lyles and McKenzie, 1998). For example, binding is undetectable at 60 mM or 120 mM NaCl, and the rates of M protein dissociation upon shifting the salt concentration were approximately an order of magnitude faster than those of NCM complexes assembled in vivo. In order for M protein to be recruited into NCM complexes it may interact with other M protein molecules already bound to the nucleocapsid. This could explain why intracellular M proteins can bind virion NCM complexes, which contain endogenous M protein, while they cannot bind to virion nucleocapsids stripped of M protein. Aggregates of M protein have been shown to form at low ionic strength in the absence of nucleocapsids (Gaudin et al., 1995; McCreedy et al., 1990). However, self-association of M protein in the absence of nucleocapsids differs from M protein binding to NCM complexes in both its salt dependence and the kinetics of dissociation (Lyles and McKenzie, 1998). For example, M protein dissociates from aggregates much more slowly than it dissociates from NCM complexes. Also, at physiologic ionic strength, in the presence of Triton X-100, M protein remains soluble in the absence of NCM complexes (Figs. 1 and 2 and McCreedy et al., 1990), but binds efficiently to NCM complexes (Fig. 5 and Lyles and McKenzie, 1998).

Even though intracellular nucleocapsids did not bind M protein with high affinity, they did bind cytosolic and virion-derived M protein with an apparent affinity approximately eightfold less than that of virion NCM complexes (Figs. 5 and 7 and data not shown). This lower binding affinity could not be accounted for by heterogeneous affinities among intracellular nucleocapsids, since they sedimented as a homogeneous population in gradients containing M protein (Fig. 7). The morphology of intracellular nucleocapsids was not markedly affected by the low-affinity interaction with M protein. This contrasts with the high-affinity interaction with M protein in virions, which condenses the nucleocapsid into a bullet-shaped structure (Barge et al., 1993; Newcomb and Brown, 1981). Previous studies indicated that M protein and nucleocapsids did not colocalize with each other in the cytoplasm of infected cells and that condensed nucleocapsids characteristic of NCM complexes were observed only at the plasma membrane (Lyles et al., 1988; McCreedy and Lyles, 1989; Odenwald et al., 1986; Ohno and Ohtake, 1987; Ono et al., 1987). The low-affinity interaction observed here may lead to an association of M protein and nucleocapsids in the cytoplasm of infected cells that was not apparent by methods using microscopy. Such an interaction may play a role in the regulation of viral transcription by M protein (Carroll and Wagner, 1979; Clinton et al., 1978; Morita et al., 1987; Wilson and Lenard, 1981). It is possible that this low-affinity interaction of M protein with intracellular nucleocapsids is part of the virus assembly pathway. Alternatively, it might represent an in vitro interaction that does not occur in virions. In either case, these data indicate that in order for M protein to assemble into an NCM complex, intracellular nucleocapsids need to undergo a transition to high-affinity binding of M protein similar to the binding observed in NCM complexes from virions. Infected cells contain a mixture of nucleocapsids with negative strand genomes and positive strand antigenomes. The genome RNA contains a nucleotide sequence near the 59 terminus that provides a basis for selective budding of nucleocapsids with negative strand RNA genomes (Whelan and Wertz, 1999). The inability of intracellular nucleocapsids to bind M protein with high affinity could be explained if the intracellular nucleocapsids primarily contained positive strand antigenomes. However, at 3 h postinfection approximately 60% of intracellular nucleocapsids contain negative strand RNA and at 6 h postinfection approximately 80% of intracellular nucleocapsids contain negative strand RNA (Simonsen et al., 1979; Soria et al., 1974). In the experiments described here, intracellular nucleocapsids were radiolabeled from 3 to 5 h postinfection. Therefore, at least 60% of the radiolabeled nucleocapsids contain negative strand genomes. If M protein bound only to nucleocapsids with negative strand RNA, these nucleocapsids

ASSEMBLY OF VSV NCM COMPLEXES

should have been distinguished from nucleocapsids with positive strand RNA by their different sedimentation velocities in the presence of M protein (Fig. 7). However, intracellular nucleocapsids sedimented as a uniform population in the presence of M protein. Thus, even though greater than 60% of the intracellular nucleocapsids contain negative strand genomes, they still need to undergo a transition in order to bind M protein with high affinity. This raises the question of what is the nature of the transition that leads to assembly of NCM complexes of high affinity. The results suggest two possible models for assembly of NCM complexes based on whether the low-affinity interaction of cytosolic M protein with intracellular nucleocapsids is a precursor to the high-affinity interaction of M protein and nucleocapsids in virions. In one model, cytosolic M protein first binds to nucleocapsids with low affinity. A separate step converts the interaction to one of high affinity. An alternative model is that the low-affinity interaction is not part of the assembly pathway. In this model, assembly of NCM complexes is initiated by the transition to high-affinity binding of M protein followed by recruitment of additional M protein. For example, binding of the initial M protein to the nucleocapsid could be fundamentally different from binding of the M protein that is subsequently recruited into the NCM complex. Binding of the initial M protein(s) to the nucleocapsid could trigger a high-affinity state throughout the entire nucleocapsid such that additional M proteins are randomly recruited into the assembling NCM complex. Alternatively, once the initial M protein(s) binds to the nucleocapsid, additional M proteins could be recruited into the assembling NCM complex by polymerizing with the preexisting M proteins already associated, providing an ordered mechanism of assembly and disassembly similar to that observed in the helical tobacco mosaic virus (Butler, 1984). These proposed steps in assembly correspond with the salt-dependent disassembly of virion NCM complexes. The recruitment step in assembly corresponds to the reversible dissociation and reassociation of M protein in NCM complexes at physiological ionic strength. The initiation step in assembly corresponds to the process that is irreversibly inactivated by high ionic strength. It has been proposed that membrane-bound M protein initiates binding of intracellular nucleocapsids to the plasma membrane (Chong and Rose, 1993; McCreedy and Lyles, 1989). According to this model, membranederived M protein should either bind nucleocapsids with higher affinity than cytosolic M protein or be able to initiate binding to nucleocapsids devoid of M protein, while cytosolic M protein cannot. However, membranederived M protein did not bind to virion NCM complexes with an affinity higher than that of cytosolic M protein (Fig. 2) and exhibited little to no binding activity for either

305

intracellular nucleocapsids or virion nucleocapsids stripped of M protein (Figs. 5 and 6). The inability of membrane-derived M protein to bind nucleocapsids devoid of M protein disagrees with previous reports that demonstrated binding of virion nucleocapsids to M protein in reconstituted liposomes (Ogden et al., 1986) or to membranes isolated from transfected cells (Chong and Rose, 1993). We have tried to reproduce the conditions used in these previous reports but have not detected binding activity above background levels (Fig. 6 and unpublished data). The reasons for this discrepancy are not known. It is possible that the virion nucleocapsids used in the previous reports retained enough M protein to be similar to the partially dissociated NCM complexes analyzed here, which bind both cytosolic and membranederived M protein with high affinity. Since membrane-derived M protein cannot initiate assembly alone, it appears likely that an additional factor is required for nucleocapsids to undergo the transition to high-affinity binding. Such a factor may be of either viral or host origin. The participation of other viral components, such as G protein, in assembly of NCM complexes has not been ruled out. However, all of the viral proteins were present in the binding experiments using subcellular fractions from infected cells, yet M protein and intracellular nucleocapsids did not associate with high affinity. This makes it more likely that the activity of a host factor is necessary to assemble high-affinity NCM complexes. The recent finding that the VSV M protein interacts with host proteins containing WW domains, which may be involved in the budding of virions, provides examples of potential host factors that may be involved in assembling NCM complexes (Craven et al., 1999; Harty et al., 1999). In our experiments, such a host factor may have been lost or inactivated during preparation of the subcellular fractions or may require additional cofactors for its activity, such as divalent cations or nucleotides. Future experiments will be directed toward identifying such a factor that commits nucleocapsids to assemble into virions. MATERIALS AND METHODS Cells and viruses Wild-type VSV (New Jersey serotype or Indiana serotype, San Juan strain) was grown in BHK cells. Infections were carried out in Dulbecco’s modified Eagle’s medium, DMEM 1 2% FCS (Life Technologies). Virus was purified by infecting cells at 0.01 PFU/cell and incubating for 18–24 h. Radiolabeled virus was prepared by adding 10 mCi/ml of [ 35S]methionine (Amersham Pharmacia Biotech) to the medium. Virus was purified from culture supernatants clarified by low-speed centrifugation (3300 g, 20 min, 4°C). The culture supernatants were centrifuged in a Ti45 rotor (Beckman) at 35,000 rpm for 1 h at

306

FLOOD AND LYLES

4°C to pellet the virus. The virus pellet was resuspended in 10 mM Tris, pH 8.1, and banded in 10–60% (w/w) sucrose gradients containing 10 mM Tris, pH 8.1, by centrifugation in an SW41 rotor (Beckman) at 35,000 rpm for 2.5 h at 4°C. The viral band was collected with a Pasteur pipette, diluted with 10 mM Tris, pH 8.1, and pelleted as before in a Ti45 rotor. The viral pellet was resuspended in 10 mM Tris containing aprotinin (2 mg/ ml) and pepstatin A (1 mg/ml) (Boehringer Mannheim). Virus was stored on ice and used in the binding assays within 48 h. Influenza virus X47 and Sendai virus (Z strain) were grown in embryonated eggs and purified as described in (Lyles et al., 1985). Protein concentrations were determined by the method of Lowry et al. (1951). Isolation of subcellular fractions BHK cells were infected with VSV at 20 PFU/cell and incubated for 4.5 h at 37°C in DMEM 1 2% FCS. For the preparation of radiolabeled subcellular fractions, infected cells were pulse labeled with 100 mCi/ml of [ 35 S]methionine for 30 min at 4 h postinfection. All subsequent steps were carried out at 4°C using prechilled buffers. Cells were scraped off the dishes, pelleted (1000 g, 5 min, 4°C), and washed once with PBS-A. Approximately 2 3 10 7 cells were resuspended in 10 ml of homogenization buffer (10 mM Tris, pH 8.1, or 7.5, 10 mM NaCl, 0.25 mM MgCl 2 ) and placed on ice for 15 min. Cells were pelleted and resuspended in 0.5 ml homogenization buffer and disrupted by 30–40 strokes with a Dounce homogenizer. Lysis was verified by microscopy. Concentrated NaCl was added immediately after lysis to raise the NaCl concentration to 150 mM to prevent aggregation of M protein. Aprotinin (2 mg/ml) and pepstatin A (1 mg/ml) were added after lysis. Nuclei were removed by centrifugation (2000 g, 5 min). The postnuclear supernatant was centrifuged in a TLA-45 rotor (Beckman) at 40,000 rpm for 1 h at 4°C. The resulting supernatant was the cytosolic fraction. This material was placed on ice and used within 24 h. The pelleted material was resuspended in 0.5 ml homogenization buffer by Dounce homogenization and mixed with 2.5 ml 66% (w/w) sucrose. The sample was overlaid with 1.2 ml of 40% (w/w) sucrose and 0.8 ml of 10% (w/w) sucrose containing 10 mM Tris, pH 8.1, or 7.5, 10 mM NaCl and 1 mM EDTA. The sample was centrifuged in an SW50.1 rotor (Beckman) at 35,000 rpm for 16 h at 4°C. Membranes were collected from the 40/10% interface with a pipette and concentrated by pelleting the diluted membrane material in a TLA-45 rotor at 40,000 rpm for 1 h at 4°C. Membranes were resuspended with a Dounce homogenizer in 10 mM Tris with aprotinin and pepstatin A. The membrane material was used immediately.

Isolation and quantitation of intracellular nucleocapsids and nucleocapsids stripped of M protein BHK cells were infected at 10 PFU/cell and intracellular nucleocapsids were isolated at 5 h postinfection. 35 S-radiolabeled intracellular nucleocapsids were prepared from infected cells pulse labeled from 3 to 5 h postinfection with methionine-free DMEM to which [ 35S]methionine was added at 50 mCi/ml. Cells were scraped off the dishes, pelleted (1000 g, 5 min, 4°C) and washed twice with PBS-A. Approximately 3 3 10 7 cells were resuspended in 0.9 ml of 10 mM Tris, 10 mM NaCl, pH 8.1, followed by addition of 0.1 ml of 10% Triton X-100 in 10 mM Tris, 1 M NaCl, pH 8.1, to lyse the cells. Aprotinin (2 mg/ml) and pepstatin A (1 mg/ml) were added, and the cells were incubated on ice for 15 min. Lysis was verified by microscopy. Nuclei were removed by centrifugation (2000 g, 5 min, 4°C). The postnuclear supernatant (1 ml) was layered on top of a 10–60% (w/w) sucrose gradient containing 10 mM Tris, pH 8.1, and 100 mM NaCl. The lysate was centrifuged in an SW41 rotor at 35,000 rpm for 5 h at 4°C. Fractions (0.5 ml) were collected by puncturing the bottom of the tube and the fractions were analyzed by scintillation counting or SDS– PAGE and staining with Coomassie blue to identify the fractions containing nucleocapsids. Appropriate fractions were dialyzed overnight against 10 mM Tris, 10 mM NaCl, pH 8.1, and 0.1 mM dithiothreitol. Intracellular nucleocapsids were quantitated by comparing the amount of N protein in each preparation to the known amount of N protein in purified VSV by SDS–PAGE followed by Coomassie blue staining and densitometry. In purified virions, N protein is 33% of the total protein (Thomas et al., 1985). Intracellular nucleocapsids were used within 48 h of lysing the cells. In some cases it was necessary to concentrate the intracellular nucleocapsids. This was done by layering the dialyzed nucleocapsids over a cushion of 0.3 ml 90% glycerol in 10 mM Tris, 10 mM NaCl, placed at the bottom of an SW41 tube. The sample was centrifuged at 40,000 rpm for 90 min at 4°C. The glycerol cushion was collected and used without further dialysis. Radiolabeled virion nucleocapsids were isolated from virions labeled with 20 mCi/ml [ 35S]methionine. Approximately 0.5 mg/ml of purified virions were incubated on ice for 30 min with 1% Triton X-100, 0.25 M NaCl, and 0.2 mg/ml dithiothreitol. The sample was centrifuged for 10 min at 1000 g at 4°C to remove large aggregates. The supernatant was layered over a 10–66% (w/w) sucrose gradient containing 10 mM Tris, pH 8.1, and 250 mM NaCl and centrifuged in an SW50.1 rotor at 40,000 rpm for 5 h at 4°C. Fractions were collected and the appropriate fractions were dialyzed overnight, and the amount of N protein was quantitated as was done for

ASSEMBLY OF VSV NCM COMPLEXES

intracellular nucleocapsids. Virion nucleocapsids were used within 24 h of solubilizing freshly purified virus. Binding of intracellular M protein to virion NCM complexes, intracellular nucleocapsids, and virion nucleocapsids stripped of M protein. An 35S-labeled cytosolic or membrane fraction prepared from infected cells (approximately 50 mg/ml total protein) was mixed with purified VSV-IND, VSV-NJ, influenza virus, or Sendai virus in buffer containing 10 mM Tris, pH 8.1, and 120 or 60 mM NaCl in a final volume of 120 ml. Triton X-100 (0.5% final concentration) was added to solubilize the virus envelopes and cellular membranes. Samples were incubated at room temperature for $5 min to allow for complete equilibration of M protein in NCM complexes and then were centrifuged for 25 min at 28 lb/in. 2 (approx. 100,000 g) in a Beckman Airfuge. Supernatant and pellet fractions were analyzed by SDS–PAGE in gels made with 10% acrylamide with either 0.27 or 0.5% bis-acrylamide to provide better resolution of the viral N and P proteins. The amount of radiolabeled M protein was quantitated by phosphorescence imaging (Molecular Dynamics). The percentage of radiolabeled M protein in the pellet was calculated as the amount of M protein in the pellet divided by the sum of M protein in the supernatant and pellet fractions. Binding of cytosolic and membrane-derived M protein to intracellular or stripped nucleocapsids was measured in a similar manner except that non-radiolabeled nucleocapsids were substituted for purified virions. Apparent dissociation constants were determined by nonlinear least squares fit of the data to Eq. 1 in Lyles and McKenzie (1998) (Curvefit Software, Kevin Raner Software, Australia). Sedimentation of nucleocapsids in gradients containing M protein Linear sucrose gradients containing M protein throughout were generated from 10 and 60% w/v sucrose solutions containing equal amounts of a soluble extract from virions containing M protein. Purified virions (approximately 1 mg/ml) were solubilized in 10 mM Tris, pH 8.1, 0.25 M NaCl, and 0.5% Triton X-100, incubated on ice for 30 min, and centrifuged in a Ti70.1 rotor (Beckman) at 60,000 rpm for 20 min at 4°C to pellet the viral nucleocapsids. The supernatant containing M protein and G protein was used to make 10 and 60% (w/v) sucrose solutions by dissolving 0.6 or 3.6 g of sucrose, respectively, in 3.0 ml of the supernatant from solubilized virions and the final volume was adjusted to 6.0 ml with 10 mM Tris buffer. The final concentrations in the gradients were 10 mM Tris, 0.25% Triton X-100, 125 mM NaCl, and 4–7 mM M protein. Gradients lacking M protein were made similarly except that the buffer used for solubilizing viri-

307

ons was used in place of the soluble extract. Either intracellular or virion nucleocapsids stripped of M protein were layered on the gradients and centrifuged in a TLS-55 rotor (Beckman) at 55,000 rpm for 30 min at 4°C. Fractions were collected from the bottom of the tube, precipitated with 10% trichloroacetic acid, and analyzed by SDS–PAGE. The amount of N protein in each fraction was quantitated by phosphorescence imaging. Flotation analysis of nucleocapsids binding to membranes Binding of stripped nucleocapsids to membranes was performed essentially as described by Chong and Rose (1993). Membranes isolated from VSV or mock-infected cells were mixed with stripped nucleocapsids in 0.2 ml of 10 mM Tris, pH 7.5, 120 mM NaCl, and incubated at 37°C for 1 h, with intermittent mixing. The samples were placed at the bottom of an SW50.1 tube and mixed with 1.0 ml of 66% sucrose by pipetting. The samples were overlaid with 3.0 ml of 40% and 0.8 ml of 10% (w/w) sucrose solutions containing 10 mM Tris, pH 7.5, and 120 mM NaCl. The samples were centrifuged at 35,000 rpm for 16 h at 4°C. Fractions were collected, precipitated with 10% trichloroacetic acid, and analyzed by SDS– PAGE. The amount of N protein in each fraction was quantitated by phosphorescence imaging. Negative stain electron microscopy Intracellular or stripped nucleocapsids isolated as described previously were stained with 2% phosphotungstate (PTA) on 300-mesh, Formvar/carbon-coated copper grids (Electron Microscopy Sciences). Bacitracin at 0.2 mg/ml (Sigma) was added to dialyzed nucleocapsids (approx. 0.05 mg/ml) and to PTA to facilitate spread of sample and stain on the grid. Bacitracin and PTA were filtered twice through 0.2-mm cellulose acetate syringe filters (Corning). A drop of sample was floated on a grid for 5 min, the excess sample was blotted away with Whatman filter paper, and then the grid was then floated on a drop of 2% PTA, pH 5.5, for 1 min. The excess stain was blotted away and the dried grids were examined in a Phillips TEM-400 electron microscope at 80 keV. ACKNOWLEDGMENTS We thank Drs. Griffith Parks, David Ornelles, and Roy Hantgan for helpful advice and comments on the manuscript. This work was supported by Public Health Service Grant AI 15892 from the National Institute of Allergy and Infectious Diseases. Electron microscopy was performed in the Electron Microscopy Core Laboratory of the Comprehensive Cancer Center of Wake Forest University, supported in part by Public Health Service Grant CA 12197 from the National Cancer Institute.

308

FLOOD AND LYLES

REFERENCES Barge, A., Gaudin, Y., Coulon, P., and Ruigrok, R. W. (1993). Vesicular stomatitis virus M protein may be inside the ribonucleocapsid coil. J. Virol. 67, 7246–7253. Butler, P. J. (1984). The current picture of the structure and assembly of tobacco mosaic virus. J. Gen. Virol. 65, 253–279. Carroll, A. R., and Wagner, R. R. (1979). Role of the membrane (M) protein in endogenous inhibition of in vitro transcription by vesicular stomatitis virus. J. Virol. 29, 134–142. Chong, L. D., and Rose, J. K. (1993). Membrane association of functional vesicular stomatitis virus matrix protein in vivo. J. Virol. 67, 407–414. Clinton, G. M., Little, S. P., Hagen, F. S., and Huang, A. S. (1978). The matrix (M) protein of vesicular stomatitis virus regulates transcription. Cell 15, 1455–1462. Craven, R. C., Harty, R. N., Paragas, J., Palese, P., and Wills, J. W. (1999). Late domain function identified in the vesicular stomatitis virus M protein by use of rhabdovirus-retrovirus chimeras. J. Virol. 73, 3359– 3365. Garoff, H., Hewson, R., and Opstelten, D. J. E. (1998). Virus maturation by budding. Microbiol. Mol. Biol. Rev. 62, 1171–1190. Gaudin, Y., Barge, A., Ebel, C., and Ruigrok, R. W. (1995). Aggregation of VSV M protein is reversible and mediated by nucleation sites: Implications for viral assembly. Virology 206, 28–37. Harty, R. N., Paragas, J., Sudol, M., and Palese, P. (1999). A proline-rich motif within the matrix protein of vesicular stomatitis virus and rabies virus interacts with WW domains of cellular proteins: implications for viral budding. J. Virol. 73, 2921–2929. Justice, P. A., Sun, W., Li, Y., Ye, Z., Grigera, P. R., and Wagner, R. R. (1995). Membrane vesiculation function and exocytosis of wild-type and mutant matrix proteins of vesicular stomatitis virus. J. Virol. 69, 3156–3160. Knipe, D. M., Baltimore, D., and Lodish, H. F. (1977). Separate pathways of maturation of the major structural proteins of vesicular stomatitis virus. J. Virol. 21, 1128–1139. Li, Y., Luo, L., Schubert, M., Wagner, R. R., and Kang, C. Y. (1993). Viral liposomes released from insect cells infected with recombinant baculovirus expressing the matrix protein of vesicular stomatitis virus. J. Virol. 67, 4415–4420. Lowry, O. H., Rosebrough, N. J., Farr, A. L., and Randall, R. J. (1951). Protein measurement with the Folin phenol reagent. J. Biol. Chem. 193, 265–275. Luan, P., and Glaser, M. (1994). Formation of membrane domains by the envelope proteins of vesicular stomatitis virus. Biochemistry 33, 4483–4489. Luan, P., Yang, L., and Glaser, M. (1995). Formation of membrane domains created during the budding of vesicular stomatitis virus. A model for selective lipid and protein sorting in biological membranes. Biochemistry 34, 9874–9883. Lyles, D. S., McKenzie, M., and Parce, J. W. (1992). Subunit interactions of vesicular stomatitis virus envelope glycoprotein stabilized by binding to viral matrix protein. J. Virol. 66, 349–358. Lyles, D. S., and McKenzie, M. O. (1998). Reversible and irreversible steps in assembly and disassembly of vesicular stomatitis virus: Equilibria and kinetics of dissociation of nucleocapsid-M protein complexes assembled in vivo. Biochemistry 37, 439–450. Lyles, D. S., McKenzie, M. O., and Hantgan, R. R. (1996a). Stopped-flow, classical, and dynamic light scattering analysis of matrix protein binding to nucleocapsids of vesicular stomatitis virus. Biochemistry 35, 6508–6518. Lyles, D. S., McKenzie, M. O., Kaptur, P. E., Grant, K. W., and Jerome, W. G. (1996b). Complementation of M gene mutants of vesicular stomatitis virus by plasmid-derived M protein converts spherical extracellular particles into native bullet shapes. Virology 217, 76–87. Lyles, D. S., McKinnon, K. P., and Parce, J. W. (1985). Labeling of the cytoplasmic domain of the influenza virus hemagglutinin with fluo-

rescein reveals sites of interaction with membrane lipid bilayers. Biochemistry 24, 8121–8128. Lyles, D. S., Puddington, L., and McCreedy, B. J., Jr. (1988). Vesicular stomatitis virus M protein in the nuclei of infected cells. J. Virol. 62, 4387–4392. McCreedy, B. J., Jr., and Lyles, D. S. (1989). Distribution of M protein and nucleocapsid protein of vesicular stomatitis virus in infected cell plasma membranes. Virus Res. 14, 189–205. McCreedy, B. J., Jr., McKinnon, K. P., and Lyles, D. S. (1990). Solubility of vesicular stomatitis virus M protein in the cytosol of infected cells or isolated from virions. J. Virol. 64, 902–906. Morita, K., Vanderoef, R., and Lenard, J. (1987). Phenotypic revertants of temperature-sensitive M protein mutants of vesicular stomatitis virus: Sequence analysis and functional characterization. J. Virol. 61, 256–263. Naeve, C. W., Kolakofsky, C. M., and Summers, D. F. (1980). Comparison of vesicular stomatitis virus intracellular and virion ribonucleoproteins. J. Virol. 33, 856–865. Newcomb, W. W., and Brown, J. C. (1981). Role of the vesicular stomatitis virus matrix protein in maintaining the viral nucleocapsid in the condensed form found in native virions. J. Virol. 39, 295–299. Newcomb, W. W., Tobin, G. J., McGowan, J. J., and Brown, J. C. (1982). In vitro reassembly of vesicular stomatitis virus skeletons. J. Virol. 41, 1055–1062. Odenwald, W. F., Arnheiter, H., Dubois-Dalcq, M., and Lazzarini, R. A. (1986). Stereo images of vesicular stomatitis virus assembly. J. Virol. 57, 922–932. Ogden, J. R., Pal, R., and Wagner, R. R. (1986). Mapping regions of the matrix protein of vesicular stomatitis virus which bind to ribonucleocapsids, liposomes, and monoclonal antibodies. J. Virol. 58, 860–868. Ohno, S., and Ohtake, N. (1987). Immunocytochemical study of the intracellular localization of M protein of vesicular stomatitis virus. Histochem. J. 19, 297–306. Ono, K., Dubois-Dalcq, M. E., Schubert, M., and Lazzarini, R. A. (1987). A mutated membrane protein of vesicular stomatitis virus has an abnormal distribution within the infected cell and causes defective budding. J. Virol. 61, 1332–1341. Reidler, J. A., Keller, P. M., Elson, E. L., and Lenard, J. (1981). A fluorescence photobleaching study of vesicular stomatitis virus infected BHK cells. Modulation of G protein mobility by M protein. Biochemistry 20, 1345–1349. Simonsen, C. C., Batt-Humphries, S., and Summers, D. F. (1979). RNA synthesis of vesicular stomatitis virus-infected cells: In vivo regulation of replication. J. Virol. 31, 124–132. Soria, M., Little, S. P., and Huang, A. S. (1974). Characterization of vesicular stomatitis virus nucleocapsids. I. Complementary 40 S RNA molecules in nucleocapsids. Virology 61, 270–280. Sun, W., Huang, L., and Wagner, R. R. (1994). Common distribution of antigenic determinants and complementation activity on matrix proteins of two vesicular stomatitis virus serotypes. J. Gen. Virol. 75(Pt 4), 937–943. Thomas, D., Newcomb, W. W., Brown, J. C., Wall, J. S., Hainfeld, J. F., Trus, B. L., and Steven, A. C. (1985). Mass and molecular composition of vesicular stomatitis virus: A scanning transmission electron microscopy analysis. J. Virol. 54, 598–607. Wertz, G. W., Davis, N. L., and Patton, J. (1987). The role of proteins in vesicular stomatitis virus RNA replication. In “The Rhabdoviruses” (R. R. Wagner, Ed.), pp. 271–293. Plenum, New York. Whelan, S. P., and Wertz, G. W. (1999). The 59 terminal trailer region of vesicular stomatitis virus contains a position-dependent cis-acting signal for assembly of RNA into infectious particles. J. Virol. 73, 307–315. Wilson, T., and Lenard, J. (1981). Interaction of wild-type and mutant M protein vesicular stomatitis virus with nucleocapsids in vitro. Biochemistry 20, 1349–1354.