JOURNAL OF VIROLOGY, Aug. 1992, P. 4737-4747

Vol. 66, No. 8

0022-538X/92/084737-11$02.00/0 Copyright © 1992, American Society for Microbiology

Spike Protein-Nucleocapsid Interactions Drive the Budding of Alphaviruses MAARIT SUOMALAINEN, PETER LILJESTROM, AND HENRIK GAROFF*

Department of Molecular Biology, Karolinska Institute, Novum, S-141 57 Huddinge, Sweden Received 12 February 1992/Accepted 14 April 1992

Semliki Forest virus (SFV) particles are released from infected cells by budding of nucleocapsids through plasma membrane regions that are modified by virus spike proteins. The budding process was studied with recombinant SFV genomes which lacked the nucleocapsid protein gene or, alternatively, the spike genes. No subviral particles were released from cells which expressed only the nucleocapsid protein or the spike proteins. Virus release was found to be strictly dependent on the coexpression of the nucleocapsid and the spike proteins. These results provide direct proof for the hypothesis that the alphavirus budding is driven by nucleocapsidspike interactions. The importance of the viral 42S RNA for virus assembly and budding was investigated by using the heterologous vaccinia virus-T7 expression system for the synthesis of the SFV structural proteins. The results demonstrate that the viral genome is not absolutely required for formation of budding competent nucleocapsids, since small amounts of viruslike particles were assembled in the absence of 42S RNA.

Enveloped viruses package their genomes into protein shells, and this capsid (or core) structure is in turn surrounded by a lipid bilayer envelope during the final step of virus maturation, the budding process. The envelope contains viral spike proteins, which mediate the delivery of the capsid and the associated viral genome into the cytosol of a new host cell (29, 61). Thus, the assembly of infectious particles is critically dependent on two steps: first, capsid proteins must selectively encapsidate the viral genome from a pool of other viral and cellular nucleic acids, and second, the capsid must be enveloped by a membrane containing the viral spike proteins. To meet these two criteria, virus assembly is expected to proceed via specific interactions between the capsid proteins and the viral genome on the one hand and the capsid and spike proteins on the other hand. Surprisingly, recent studies on retroviruses have shown that the putative capsid-spike interaction is not essential for retrovirus budding, since the Gag precursor alone can direct the budding process (10, 20; see also references 28, 50, and 54). Furthermore, these and other studies have demonstrated that the viral genome is not needed for the formation of a budding-competent retrovirus core (20, 34). Our studies are focused on the assembly of the alphavirus Semliki Forest virus (SFV), which replicates in the host cell cytoplasm (18, 52, 57). The SFV genome is a single-stranded 42S RNA molecule of positive polarity. The 42S RNA serves as an mRNA for four nonstructural proteins (nsPl to nsP4) which form the viral replicase and transcriptase complex (reviewed in reference 65). The structural proteins of SFV, the nucleocapsid protein C and the two transmembrane envelope proteins p62 and El, are synthesized as a polyprotein precursor from a subgenomic 26S RNA. The aminoterminal C protein has serine protease activity and autoproteolytically cleaves itself from the nascent polyprotein chain (1, 8, 23, 40). The newly synthesized C proteins associate with the genomic RNA in the cytoplasm to form the nucleocapsid structures (for a review, see references 51 and 77). The envelope proteins are cotranslationally inserted into the membrane of the endoplasmic reticulum via alternating *

signal and stop transfer sequences (17, 19, 24, 30, 39). Cleavage events that generate p62 and El are mediated by the host enzyme signal peptidase. In the endoplasmic reticulum, p62 and El form heterodimeric complexes which are transported via the Golgi to the plasma membrane (37, 74, 82). At a late stage during transport (at the post-Golgi stage), the p62 subunit is proteolytically processed at its external domain, and this cleavage event generates the E2 form that is found in mature virions (9). Cleavage of p62 is not essential for virus assembly, but it is required for the activation of viral entry functions (27, 35, 36, 43, 47, 75). Budding of alphaviruses occurs at the plasma membrane, and in contrast to the process in retroviruses, the former process is believed to be driven by nucleocapsid-spike interactions (reviewed in references 51 and 56). Studies with ts mutants, which encode defective spike proteins at the restrictive temperature, have provided some support for such a budding mechanism. Electron microscopy (EM) analysis of the ts-virus-infected cells have shown that the cells are devoid of budding structures at the restrictive temperature, whereas such structures are readily observed in cells infected with the wild-type virus (5, 48). This suggests that the envelopment of the alphavirus nucleocapsid is dependent on nucleocapsid-spike interactions. However, the possible release of spikeless, enveloped nucleocapsids from the ts-virus-infected cells was never thoroughly analyzed by quantitative biochemical assays. Furthermore, it is unclear whether the C proteins of these ts mutants are actually able to form assembly-competent nucleocapsids at the restrictive temperature or whether these structures also contain mutations that render them incompetent for assembly, as is the case with the spike proteins. In this study, we have used genetically defined deletion variants of SFV to study the molecular interactions that drive the budding of alphaviruses. Our results conclusively show that nucleocapsid-spike interactions are needed for alphavirus budding. Furthermore, we found that the viral genome is not absolutely essential for formation of buddingcompetent nucleocapsids, since small amounts of viruslike particles were assembled in the absence of the genomic 42S RNA.

Corresponding author. 4737

4738

SUOMALAINEN ET AL.

MATERIALS AND METHODS Plasmid constructs. (i) pSP6-SFV4. Plasmid pSP6-SFV4 contains a full-length cDNA clone of the SFV genome under the control of the SP6 polymerase promoter (32). The RNA transcribed from this plasmid, SFV-wt RNA, codes for both the SFV nonstructural proteins and the structural proteins and is thus infectious. (ii) pSFV-c. Plasmid pSEV-c was constructed by changing the first codon of p62 (TCC) on pSP6-SFV4 to a translational stop codon (TAG) by site-directed mutagenesis. Furthermore, the 2,104-bp NdeI fragment from the p62-El coding region was deleted to ensure that polypeptides resulting from possible readthrough at the stop codon would not interfere with the interpretation of the results. The RNA transcribed from this construct, SFV-c RNA, directs only the synthesis of the SFV nonstructural proteins and the structural C protein. (ifi) pSFV-spike. Plasmid pSFV-spike was constructed by replacing the 947-bp EcoRI-XhoI fragment of pSP6-SFV4 by the 135-bp EcoRI-XhoI fragment from p62dhfr (17). Thus, pSFV-spike contains a precise deletion of the C gene, and the RNA transcribed from this plasmid, SFV-spike RNA, directs only the synthesis of the SFV nonstructural proteins and the structural p62, 6K, and El proteins. (iv) pSFV-c+spike. Plasmid pSFV-c+spike contains two promoters for subgenomic RNA transcription: a 3' proximal promoter on the negative strand directs the synthesis of the mRNA for the C protein, and the genes for the envelope proteins are under a 3' distal promoter. pSFV-c+spike was made by replacing the 75-bp XbaI-BglII fragment of pSFVspike with the 2,630-bp XbaI-HindIII fragment from pSFV-c. The BglII and HindIII sites were filled in with Klenow fragment prior to ligation. The RNA transcribed from this plasmid, the SFV-c+spike RNA, directs synthesis of the SFV nonstructural and structural proteins. (v) pSFV-spike+c. Plasmid pSFV-spike+c is identical to pSEV-c+spike except that the 3' proximal promoter on the negative strand directs the synthesis of the mRNA for the envelope proteins, while the C gene is under the 3' distal promoter. pSFV-spike+c was constructed by replacing the 75-bp XbaI-BglII fragment of pSFV-c with the 3,917-bp XbaI-HindIII fragment from pSFV-spike. The BglII and HindIII sites were filled in with Klenow fragment prior to ligation. (vi) pGEMl-T726S. Plasmid pGEMl-T726S contains a cDNA copy of the 26S mRNA under the control of the T7 promoter in pGEM1 (Promega). pGEMl-T726S was constructed by joining the 2,820-bp HindIII-EcoRI fragment of pGEM1, the 3,036-bpXAhol-HindIII fragment of pSP6-SFV4, and the 947-bp EcoRI-XhoI fragment of pSP6-SFV4. RNA transcription and metabolic labeling of transfected cells. In vitro transcription and electroporation of in vitromade RNA into BHK-21 cells was done as previously described (32). Electroporated cell suspensions were diluted 1:20 in complete BHK-21 medium (BHK-21 medium [GIBCO] supplemented with 5% fetal calf serum, 20 mM HEPES [N-2-hydroxyethylpiperazine-N'-2-ethanesulfonic acid, pH 7.3], and 20% tryptose phosphate broth). At 14 to 16 h postelectroporation, the medium was replaced with a methionine-free minimum essential medium (GIBCO) supplemented with 0.2% bovine serum albumin and 10 mM HEPES. After 30 min at 37°C, the medium was replaced with the same methionine-free medium containing 100 ,Ci of [35S]methionine (Amersham) per ml, and the cells were incubated at 37°C for 30 min (pulse). After the pulse, cells

J. VIROL.

were washed once with complete BHK-21 medium containing a 10-fold excess of cold methionine and then incubated in the same medium for 1, 3, or 5 h (chase). After the chase, the culture media were collected and clarified by centrifugation in an Eppendorf centrifuge (5 min at 5,000 rpm at 4°C). The cell monolayers were placed on ice, washed twice with ice-cold phosphate-buffered saline (PBS), and solubilized with Nonidet P-40 lysis buffer containing 10 mM iodoacetamide (68). The nuclei were removed by centrifugation in an Eppendorf centrifuge (5,000 rpm for 5 min at 4°C). Assays. (i) Assembly of nucleocapsids. Nucleocapsid assembly was analyzed as described in reference 69. Briefly, lysates from the 1-h chase point (see above) were incubated with EDTA (final concentration, 25 mM) on ice for 15 min before application onto 15 to 30% (wt/wt) sucrose gradients in TNE buffer (50 mM Tris-HCl [pH 7.4], 100 mM NaCl, 1 mM EDTA) containing 0.1% Nonidet P-40. To study the kinetics and efficiency of nucleocapsid assembly in SFV-ctransfected cells, the cell cultures were pulsed for 2 min and chased for 5 or 30 min. Centrifugation was performed in a Beckman SW41 rotor at 40,000 rpm for 2 h at 4°C. The gradients were fractionated from the bottom, and radioactivity was analyzed by liquid scintillation counting. Alternatively, the fractions were trichloroacetic acid (TCA) precipitated by adding an equal volume of 20% TCA. The pellets were resuspended in sodium dodecyl sulfate (SDS)-sample buffer (74) and run on 10% minigels as previously described (68) except that the gels were processed for autoradiography without methanol-acetic acid fixing. The radioactivity in C-protein bands was quantitated as described in reference 74. (ii) Transport of envelope proteins to plasma membrane. Envelope protein transport was assayed by monitoring the cleavage of p62 to E2 by SDS-polyacrylamide gel electrophoresis (PAGE) (p62 cleavage occurs at the post-Golgi stage; 9). (iii) Virus maturation. Viral structural proteins were immunoprecipitated from cell lysates with monoclonal anti-C (36-1-9 [21]), monoclonal anti-El (UM 8.139 [4]), and monoclonal anti-E2 (UM 5.1 [4]) antibodies. Virus particles were immunoprecipitated from clarified culture supernatants with the monoclonal anti-E2 antibody. In the case of SFV-c transfection, culture supernatants were analyzed by anti-C immunoprecipitations. Both the anti-E2 and the anti-C monoclonal antibodies are broadly reacting antibodies which recognize different forms of their respective antigens (21, 66a, 74). Equal volumes of cell lysates and corresponding culture supernatants were used for immunoprecipitations. The immunocomplexes were brought down with Pansorbin (10% [wt/vol] in PBS; Calbiochem, La Jolla, Calif.), with rabbit anti-mouse immunoglobulins (Dakopatts a/s, Glostrup, Denmark) as a linking antibody when necessary. Cell lysate immunoprecipitates were washed as described previously (74), and culture supernatant immunoprecipitates were washed twice with 10 mM Tris-HCl (pH 7.5). Immunoprecipitates were solubilized in SDS-sample buffer by heating at 70°C for 2 min and were analyzed by SDS-PAGE. The radioactivity in protein bands was quantitated (see above), and the percentage of El (or C) in the culture medium {[Elmed/(Ellys + Elmed)] x 100%, where Elmed is the amount of El in the medium and Ellys is the amount of El in the lysate} at the 5-h chase point was determined. (iv) Determination of virus titers. Serial dilutions of the 5-h chase media were used to infect BHK-21 cells grown on coverslips. At 8 to 10 h postinfection, cells were processed for indirect immunofluorescence as described in reference 71

BUDDING OF ALPHAVIRUSES

VOL. 66, 1992

and double stained with monoclonal F13 antibody (71; F13 was kindly provided by David Vaux, European Molecular Biology Laboratory, Heidelberg, Germany) and a polyclonal anti-E2 antiserum. Fluorescein-conjugated sheep anti-mouse and rhodamine-conjugated goat anti-rabbit antibodies (BioSys, Compiegne, France) were used as second antibodies. Virus titers of 5-h chase media (infectious units per milliliter) were determined by counting the cells positive for F13 staining from dilutions which gave 100 to 300 positive cells per coverslip. When using low multiplicities of infection, we found that first rounds of infections could mostly be detected at 8 to 10 h postinfection. However, if groups of positive cells were observed, these were counted as one infectious center. (v) EM. Virus particles were harvested from cleared culture supernatants by pelleting them through a glycerol cushion (50% glycerol in TNE buffer) in an SW28 rotor at 25,000 rpm for 90 min at 4°C. Virus was resuspended in TNE buffer, collected on carbon-coated grids, and stained for 60 s with 1% sodium silicotungstate (pH 7.4). (vi) Isopycnic centrifugation. At 12 h postelectroporation (or in the case of vaccinia virus-T7 expression, at 5 h posttransfection), media were replaced with methionine-free medium. After 30 min at 37°C, media were replaced with methionine-free medium containing 100 ,Ci of [35S]methionine per ml and 1% fetal calf serum, and cells were incubated at 37°C for 12 h. Cell culture media were collected, and cell debris was removed by centrifugation at 3,000 rpm for 10 min (4°C) in a Heraeus Minifuge T table centrifuge. Virus particles were pelleted from the clarified culture supernatants as described above. The virus pellet was resuspended in TNE, and the suspension was loaded onto a 25 to 50% (wt/wt) sucrose gradient in TNE and subjected to centrifugation in a Beckman SW28.1 rotor for 16 h at 25,000 rpm and 4°C (14). Gradients were fractionated from the bottom, and the radioactivity in each fraction was determined by liquid scintillation counting. Peak fractions were TCA precipitated, and the resultant pellets were taken up in SDS-sample buffer and analyzed by SDS-PAGE. Vaccinia virus-T7 expression. Recombinant vaccinia virus (vTF7-3; 12) containing the T7 RNA polymerase gene was kindly provided by Bernard Moss, National Institutes of Health, Bethesda, Md. Stocks of vTF7-3 were prepared in BHK-21 cells as described previously (38). Near-confluent monolayers of BHK-21 cells grown on 35-mm-diameter dishes were infected with vTF7-3 diluted in Dulbecco's PBS (with magnesium and calcium; GIBCO) at a multiplicity of infection of -15 to 20 for 30 min at 37°C. Lipofectin reagent (15 ,ul; Bethesda Research Laboratories, Life Technologies, Inc., Gaithersburg, Md.) and 5 p,g of pGEM1-T726S plasmid DNA were each mixed with a 750-pl aliquot of Optimem-1 (GIBCO). After the 30-min infection, the virus inoculum was removed, and the two Optimem-1 mixtures were combined and added to the cells. After 5 h at 37°C, the cells were pulsed and chased as described above for electroporated cells. For SFV-tr and vTF7-3 mixed infections, near-confluent monolayers of BHK-21 cells were simultaneously infected with vTF7-3 (-15 to 20 PFU per cell) and SFV-tr (-6 infectious units per cell) for 30 min at 37°C and subsequently transfected with pGEM1-T726S as described above (in preliminary experiments, these conditions gave the best yield of infectious virus particles). At 5 h posttransfection, cells were pulsed and chased as described above for electroporated cells. Preparation of SFV-tr virus stocks has been described

previously (31).

1

p62

2

,> X

--

El

--0

C

-

3 4 5

MD -

-

-

d

-

4739

6 7

am_ .

-

-

do

FIG. 1. Synthesis of SFV structural proteins in cells transfected with recombinant SFV genomes. In vitro-made RNAs were electroporated into BHK-21 cells, and at 14 to 16 h postelectroporation, cells were pulse-labeled with [35S]methionine for 10 min, chased for 10 min, and solubilized. Equal volumes of cell lysates were subjected to immunoprecipitation with a combination of anti-El, antiE2, and anti-C monoclonal antibodies. Immunoprecipitates were analyzed by SDS-PAGE and fluorography. Lane 1, SFV-wt RNA; lane 2, SFV-c RNA; lane 3, SFV-spike RNA; lane 4, SFV-c and SFV-spike RNAs; lane 5, SFV-c+spike RNA; lane 6, SFV-spike+c RNA; lane 7, vaccinia virus-T7-driven expression of wild-type 26S RNA.

RESULTS

Expression of a recombinant SFV genome lacking the genes for envelope proteins. pSP6-SFV4 (32) contains a full-length cDNA clone of SFV under the SP6 promoter. The cDNA clone can be used as a template for the in vitro production of infectious 42S RNA. To elucidate the roles of the C and envelope proteins in the assembly of SFV, we analyzed particle formation when the C protein or, alternatively, the envelope proteins were expressed in the absence of other structural proteins. Plasmid pSFV-c coding only for the nonstructural proteins and the C protein was used as a template for in vitro production of SFV-c RNA, which was transfected into BHK-21 cells by electroporation. As shown in Fig. 1, lane 2, the SFV-c genome directed the synthesis of C proteins, which were recognized by a monoclonal anti-C antibody and which had the same molecular mass as the C protein made from the wild-type SFV genome (SFV-wt; Fig. 1, lane 1). The minor upper band in lane 2 corresponds to a C dimer. In order to determine whether the C proteins were assembled into intracellular nucleocapsids, pulse-labeled SFV-c cell lysates were fractionated on 15 to 30% sucrose gradients. As shown in Fig. 2B, lysates contained a major fast-sedimenting peak with labeled C protein (data not shown). This peak migrated at the same position as the intracellular nucleocapsids from SFV-wt-transfected cells (Fig. 2A). By quantitating the amount of C protein in this nucleocapsid peak after both a short and a long chase, it was concluded that C proteins made from the SFV-c genome were incorporated into intracellular nucleocapsids with kinetics and efficiency essentially similar to those of C proteins synthesized in infection by wild-type SFV (data not shown). To test whether these nucleocapsids can bud from the cell, SFV-c-transfected cells were pulse-labeled for 30 min and after 1, 3, or 5 h of chase, and C proteins in the cell lysates and the culture media were immunoprecipitated with the monoclonal anti-C antibody. Culture medium was used for immunoprecipitation both with and without addition of Nonidet P-40 prior to addition of the antibodies in order to detect enveloped and naked nucleocapsids, respectively. Although high amounts of C protein could be detected in cell lysates (Fig. 3B, lanes 1 to 3), no naked or enveloped C proteins were present in the culture medium (Fig. 3B, lanes 4 to 9). In contrast, when cells were transfected with SFV-wt, 17% of total C protein was found in extracellular virions at the 5-h chase point (Fig. 3A and Table 1). These

J. VIROL.

SUOMALAINEN ET AL.

4740

E

8-

E

E 0

E 0

0

10 Fraction No

20

FIG. 2. Fractionation of cell lysates on sucrose gradients. Transfected cells were pulse-labeled for 30 min and chased for 1 h. Cell lysates were layered on 15 to 30% (wt/wt) linear sucrose gradients containing 0.1% Nonidet P.40. Gradients were centrifuged for 2 h at 4°C in an SW41 rotor at 40,000 rpm and then fractionated. Radioactivity in fractions was measured by direct liquid scintillation counting (A) or, alternatively, aliquots of fractions were first TCA precipitated, pellets were analyzed by SDS-PAGE, and radioactivity in C protein bands was determined (B to D). Lysates were from SFV-wt-transfected cells (A), SFV-c-transfected cells (B), cells expressing wild-type 26S RNA (C), or cells coexpressing the SFV-tr genome and the wild-type 26S RNA (D).

results strongly suggest that the release and envelopment of SFV nucleocapsids require the viral envelope proteins. Expression of a recombinant SFV genome lacking the C gene. It has been previously shown that the envelope proteins of Sindbis virus are capable of maintaining extensive

lateral interactions in virion-derived membrane structures from which the underlying nucleocapsid has been removed (73). Furthermore, the formation of lateral interactions between the envelope proteins has been postulated to be an important facilitating factor in the budding process (13, 56, 73). To test whether SFV envelopelike structures can be released from the host cell in the absence of intracellular nucleocapsids, a recombinant SFV genome containing a precise deletion of the C protein-encoding region was constructed. The corresponding SFV-spike RNA, when transfected into BHK-21 cells, directed the synthesis of p62 and El of correct sizes (Fig. 1, lane 3), and these were recognized by monoclonal anti-E2 and anti-El antibodies. Comparison of spike expression levels in SFV-wt- and SFVspike-transfected cells (Fig. 1, lanes 1 and 3, respectively) revealed that significantly fewer viral membrane proteins were made from the SFV-spike genome. At present, we have no clear explanation for this phenomenon, but it could be due to a change in RNA secondary structure. A detailed phenotypic analysis of p62/E2 and El made from the SFVspike genome has shown that the heterodimeric complex of p62 and El is formed and that this complex is transported to the cell surface with normal kinetics (81). The possible release of envelope structures from pulse-labeled SFVspike-transfected cells was analyzed by immunoprecipitation from the culture supernatants with the monoclonal anti-E2 antibody. As shown in Fig. 3C, no E2 or El could be immunoprecipitated from the culture supernatants after 1- to 5-h chases, although the corresponding cell lysates contained high levels of these proteins. In contrast, the SFVwt-transfected cells had incorporated 64% of the El into extracellular virions at the 5-h chase point (Fig. 3A and Table 1; see also reference 47). This suggests that SFV envelope structures cannot be released from infected cells in the absence of C-protein expression. Coexpression of the SFV-c and SFV-spike genomes results in formation of virus particles. The inability of the nucleocapsid and the spikes to form any extracellular subviral particles on their own implies that the budding of SFV requires both the C protein and the envelope proteins. To confirm this, we performed a complementation experiment and cotransfected the SFV-c and SFV-spike genomes into BHK-21 cells. As shown in Fig. 1, lane 4, C, El, and p62 of correct sizes were expressed in the SFV-c- and SFV-spikecotransfected cells, but, similar to the SFV-spike transfection, the expression level of envelope proteins was again reduced compared with that of the wild-type-transfected cells (Fig. 1, lane 1). Fractionation of pulse-labeled cell lysates on 15 to 30% sucrose gradients demonstrated that C proteins in cotransfected cells were assembled into intracellular nucleocapsids (data not shown). Envelope proteins were transported to the cell surface, as judged by the efficient cleavage of p62 to E2 (Fig. 3D). To analyze the export of viral structural proteins from the SFV-c- and SFV-spike-transfected cells, media from these cultures were subjected to immunoprecipitation with the monoclonal anti-E2 antibody after all chase time points (Fig. 3D). Beginning from the 1-h chase point, the anti-E2 antibody coimmunoprecipitated C, El, and E2 from the culture media. Coimmunoprecipitation of these proteins suggested that they were released in virionlike structures. To confirm that the exported C, El, and E2 had been incorporated into virus particles, we tested 5-h chase media for the presence of infectious virus. Since virus particles were expected to carry only the SFV-c or the SFV-spike genome, we had to use indirect immunofluorescence rather than plaque assay to

BUDDING OF ALPHAVIRUSES

VOL. 66, 1992

A

lysate

B

medium

lysate

C

_ _

_po -_

1 h 3h 5h 1 h 3h 5h 1 h 3h 5h

2

3

5

E22 E1 -_

mm

-00-

6

medium

lysate

D

4

C

1

-MS=

El

.

2

3

4

5

6

7

8

9

mo-

1

2-

1

23

-A-

4

lysate

F

medium

_ __ -bo -1

1

6

1 h 3h 5h 1 h 3h E2

-

5

4

lysate

E

1 h 3h 5h 1 h 3h 5h _

3

2

medium

lh 3h 5h 1h 3h 5h

-m

4___M4M-__

lysate

C

"_4

1

E2

medium

medium

det.

1 h 3h 5h 1 h 3h 5h p62 E2 E1

4741

56

medium

1 h 3h 5h 1 h 3h 5h

5h

p62 E2 El

4-

..-O 3 4 _ _6

C

_W,

il5

_ _

'.

'I

.,

.1.

-_

2 3 4 5 6 FIG. 3. Virus release from cells transfected with recombinant SFV genomes. Transfected cells were pulse-labeled for 30 min and chased for the times indicated (1 to 5 h). Viral structural proteins were immunoprecipitated from cell lysates with a combination of anti-El, anti-E2, and anti-C monoclonal antibodies. Virus particles in culture supernatants were immunoprecipitated with monoclonal anti-E2 antibody. In the case of SFV-c, culture media were subjected to immunoprecipitation with monoclonal anti-C antibody in the presence (+) or absence (-) of 1% Nonidet P-40 (det). Equal volumes of cell lysates and corresponding culture supernatants were used for immunoprecipitations. (A) SFV-wt transfection. The presence of excess uncleaved p62 in the cell extracts is due to slight cytopathicity of the prolonged wild-type infection. (B) SFV-c transfection. The upper band in the cell lysates is a host contaminant protein. (C) SFV-spike transfection. (D) SFV-c and SFV-spike cotransfection. The protein migrating between E2 and El in the cell lysates is a host contaminant protein. (E) SFV-c+spike transfection. (F) SFV-spike+c transfection. 1

23

determine the number of infectious particles. BHK-21 cells were incubated with serial dilutions of 5-h chase media, and after 8 to 10 h, cells were processed for immunofluorescence and double stained with a polyclonal anti-E2 antiserum and the monoclonal F13 antibody (71). The F13 antibody specifically stains cells infected with SFV, but instead of being nucleocapsid specific as originally claimed, it most likely recognizes some component of the viral replication machinery (67). Thus, all infected cells were expected to be positive for the F13 staining, and a proportion of these cells, i.e., those infected with the SFV-spike genome, were expected also to be positive for anti-E2 staining. Infected cells were clearly found, and as expected, not all of the F13-positive cells were positive for anti-E2 (data not shown). By counting

45

6

1

cells positive for F13 staining, the titer of virus produced was estimated to be 7 x 107 infectious units per ml (Table 1). To further validate virus production, we analyzed the particulate material in culture media by negative-stain EM and by isopycnic centrifugation on 25 to 50% sucrose gradients. As shown by EM (Fig. 4B) the particles produced from the SFV-c- and SFV-spike-cotransfected cells had sizes and shapes similar to those of wild-type SFV virions (Fig. 4A).

i,A

t

BC

TABLE 1. Quantitation of virus production RNA

SFV-wt

SFV-c SFV-spike SFV-c-SFV-spike SFV-c+spike SFV-spike+c Vaccinia virus-T7-26S Vaccinia virus-T7-26S+SFV-tr

(%Infectivity' of wild type)

2.6 x 108 0 0 7.2 x 107 (28) 2.0 x 107 (8) 0.7 x 107 (3) 0 0.2 x 107 (1)

Exportedb El

C

64

17 0

0 51 10 21 11 16

2 7 3 7 8

a Titer of the 5-h chase medium given as infectious units per ml. Values are means from two or three experiments. b Data represent the means of two to four experiments and are given as percentage of C or El released into the culture medium compared with total C or El (in cell hysate plus medium) at 5-h chase points.

FIG. 4. Electron micrographs of negatively stained virus particles produced from cells transfected with SFV-wt RNA (A), SFV-c and SFV-spike RNAs (B), or SFV-c+spike RNA (C) or from cells expressing wild-type 26S RNA only (D). Note that many of the viruses exhibit extrusions from the membrane layer (blebs), which are considered experimental artifacts (42). Bar = 25 nm.

4742

J. VIROL.

SUOMALAINEN ET AL.

E 36000

particles at 28% of the efficiency of the wild-type infection. To estimate the efficiency of incorporation of pulse-labeled structural proteins into virus particles, we determined the percentage of extracellular El at the 5-h chase point. A similar quantitation was done for the C protein. The results are shown in Table 1. Despite the lower expression level of the p62 and El proteins, the SFV-c- and SFV-spike-cotransfected cells released envelope proteins with an efficiency essentially similar to that of the SFV-wt-transfected cells; i.e., 51% of El was extracellular after the 5-h chase, whereas the corresponding export efficiency was 64% in the SFV-wt transfection. In contrast, the export of labeled C proteins from SFV-c- and SFV-spike-cotransfected cells (2%) was inefficient compared with that in the wild-type transfection (17% exported). Expression of recombinant SFV genomes containing two promoters for subgenomic RNA transcription. Since particle formation in the complementation system described above required that both the SFV-c and SFV-spike genomes be delivered into the same cell, the lower level of virus production compared with wild-type infection could have resulted from poor cotransfection efficiencies. To ensure that every

-

24000

-

12000

-

'

10000 8000

EL 6000 4000

2000

',

-

5000 -

C

.

,

.

, s

4000

E.

3000

0

10

20

30

fraction No of virus 5. FIG. Density analysis particles released from SFVwt-transfected cells (A) or SFV-c- and SFV-spike-cotransfected cells (B) or from cells expressing wild-type 26S RNA (C). Cells were metabolically labeled with [35S]methionine for 12 h, and the particulate material released into the medium was analyzed on linear 25 to 50% sucrose gradients. Gradients were subjected to 16 h of centrifugation in an SW28.1 rotor at 25,000 rpm and 4°C and then fractionated. Radioactivity in fractions was measured by liquid scintillation counting. The measured densities of peak fractions 9 and 14 were 1.198 and 1.173 g/ml, respectively.

Density analysis demonstrated that the majority of released particles had the same density as wild-type SFV virions (Fig. SB). A proportion of the particles released from SFV-c- and SFV-spike-cotransfected cells had a density greater than that of wild-type virions (fraction 9 in Fig. SB). SDS-PAGE analysis revealed that this population differed from the virionlike particles by having a vast excess of labeled C proteins compared with the amount of labeled envelope proteins. These denser structures remain to be further characterized. The results described above demonstrated that the cotransfection of SFV-c and SFV-spike genomes into host cells restored virus production. The efficiency of particle formation in this complementation system was estimated by comparing the titer of infectious virus in 5-h chase media with that produced in the wild-type infection. Table 1 shows that SFV-c and SFV-spike complementation produced virus

transfected cell expresses mRNAs for both the C and the envelope proteins, recombinant SFV genomes containing two promoters for subgenomic RNA transcription were constructed, and the coding regions for the C and envelope proteins were placed separately under these two promoters. In the SFV-c+spike variant, the mRNA for the C protein is read from the 3' proximal promoter on the negative strand, and in SFV-spike+c, the order is reversed: the 3' proximal promoter on the negative strand directs the synthesis of the mRNA for the envelope proteins. SFV-c+spike and SFVspike+c RNAs were transfected into BHK-21 cells by electroporation, and as shown in Fig. 1, lane 5 (SFVc+spike) and lane 6 (SFV-spike+c), C, El, and p62 of correct sizes were synthesized in both cases. The expression levels of envelope proteins in the SFV-c+spike transfection were now almost the same as in the SFV-wt transfection, whereas the SFV-spike+c-transfected cells again contained reduced amounts of envelope proteins. In both transfections, envelope proteins were transported to the cell surface, as judged by the cleavage of p62 to E2 (Fig. 3E and F for SFV-c+spike and SFV-spike+c, respectively). Fractionation of pulse-labeled cell lysates on 15 to 30% sucrose gradients revealed that C proteins made from the SFVc+spike genome were incorporated into wild-type-like intracellular nucleocapsids (data not shown). C proteins in SFVspike+c-transfected cells were also found in nucleocapsid structures, but these sedimented slightly faster than wildtype nucleocapsids on sucrose gradients (data not shown). Figures 3E and F show that structural proteins from SFVc+spike- and SFV-spike+c-transfected cells were incorporated into particles. These represented infectious viruses, since 5-h chase media from both the SFV-c+spike-transfected cells and the SFV-spike+c-transfected cells contained 2.0 x 107 and 0.7 x 107 infectious units per ml, respectively, when assayed by immunofluorescence. Furthermore, the particles produced from the SFV-c+spike transfection were similar to SFV virions in negatively stained EM samples (Fig. 4C; EM analysis was not done for the SFV-spike+c particles). Comparison of the titers of infectious virus in 5-h chase media revealed that the use of recombinant genomes containing two promoters for subgenomic RNA transcription did not improve the yield of virus particles: the efficiency of virus production in the SFV-c+spike transfection was about

VOL. 66, 1992

BUDDING OF ALPHAVIRUSES

medium

lysate

A

lysate

B

1 h 3h 5h o/n 1 h 3h 5h o/n p62 E2 E1

_ _ -_

C

-_

4743

medium

1 h 3h 5h o/n 1 h 3h 5h o/n TR

_

E2 EE1

-_-

-711V~ ~~~~~~~~~ 4m

i

C

1

2

3

4

5 6 7 8

_-_W___ _W

4-

A4'

5 6 7 8 9 10 1 2 3 4 FIG. 6. (A) Vaccinia virus-T7-driven expression of SFV 26S RNA. BHK-21 cells were infected with vaccinia vTF7-3 virus and subsequently transfected with pGEM1-T726S plasmid DNA. At 5 h postinfection, cells were pulse-labeled for 30 min and chased for the indicated times. SFV structural proteins were immunoprecipitated from cell lysates and culture media as described in the legend to Fig. 3. (B) Coexpression of SFV-tr and wild-type 26S subgenomes. BHK-21 cells were double infected with vaccinia vTF7-3 virus and SFV-tr virus and then transfected with the pGEM1-T726S plasmid DNA. At 5 h postinfection, cells were pulse-labeled for 30 min and chased for the indicated times. SFV structural proteins were immunoprecipitated from cell lysates and culture media as described in the legend to Fig. 3. Transferrin receptor (TR) proteins in the 1-h chase lysate were immunoprecipitated with a monoclonal anti-transferrin receptor antibody (lane 10). Lane 9, virus protein markers.

8% of that in the wild-type infection, and the corresponding value for the SFV-spike+c transfection was 3% (Table 1). The poor virus production was also reflected in the export efficiencies of labeled C, El, and E2: at the 5-h chase point, only 10 and 7% of the El and C proteins, respectively, were released from the SFV-c+spike-transfected cells. Export from the SFV-spike+c-transfected cells was similarly inefficient: 21% of El proteins and 3% of C proteins were extracellular at the 5-h chase point. Structural proteins are assembled into viruslike particles in the absence of the genomic 42S RNA. Newly synthesized C proteins are very rapidly and very efficiently incorporated into intracellular nucleocapsids in alphavirus-infected cells, and the viral genome is the only RNA species found in these structures (59, 70). In vitro, however, a variety of singlestranded nucleic acids can promote the assembly of Sindbis virus C proteins into nucleocapsidlike structures (78, 79). We wanted to test whether such structures are formed in vivo in the absence of the SFV genomic 42S RNA and whether these nucleocapsids can interact with envelope proteins to form virus particles. The vaccinia virus-T7 system was chosen for the expression of the SFV structural proteins. BHK-21 cells were infected with the recombinant vTF7-3 vaccinia virus encoding the bacteriophage T7 RNA polymerase and subsequently transfected with the pGEMlT726S plasmid by using Lipofectin. pGEMl-T726S contains a cDNA copy of 26S mRNA under the T7 promoter. As shown in Fig. 1, lane 7, the vaccinia virus-T7 expression system directed the synthesis of p62, El, and C, which comigrated on SDS-PAGE with the authentic counterparts synthesized in the SFV-wt-transfected cells. These SFV proteins were not present in negative control cells, which had been infected with vTF7-3 and then mock transfected (data not shown). During a longer chase, the p62 protein was efficiently cleaved to E2 (Fig. 6A), indicating that the envelope proteins were transported to the cell surface. Assembly of intracellular nucleocapsids was examined by fractionating pulse-labeled cell lysates on 15 to 30% sucrose gradients. As shown in Fig. 2C, virtually all C proteins were found in structures that sedimented considerably slower than the wild-type nucleocapsids. This suggests that the 42S RNA is needed for efficient assembly of intracellular nucleocapsids. Given the inability of the 26S expression system to pro-

mote assembly of intracellular nucleocapsids, we expected this system to be unable to produce virus particles. Surpris-

ingly, labeled E2, El, and C could be immunoprecipitated from culture media with the monoclonal anti-E2 antibody (Fig. 6A). C proteins in the media were accessible to anti-C antibodies only after detergent solubilization, indicating that the C proteins were enveloped (data not shown). EM analysis of the particulate material in the culture medium revealed the presence of particles which had an SFV virionlike morphology (Fig. 4D). These particles were not seen in EM analysis of the culture medium of mock-transfected cells (data not shown). The labeled particulate material separated into two peaks in isopycnic gradient analysis (Fig. SC). The peak of higher density (fraction 9) was also found in media from mock-transfected cells, and proteins other than the SFV structural proteins were also present in this peak (data not shown). In contrast, the particles banding at the position corresponding to the density of the SFV virions (fraction 14) were composed of E2, El, and C proteins only. Since the viruslike particles were noninfectious in the infectivity assay (as expected), the efficiency of particle formation could not be directly compared with those of the other systems. The release of labeled structural proteins (Table 1; 11% for El and 7% for C) suggests that the particle production was comparable to that in SFV-c+spike transfection and thus considerably less than that in wild-type infection. Taken together, the results described above demonstrate that the 26S expression system is capable of driving the formation of SFV particles. This indicates that 42S RNA is not absolutely required for virus assembly. However, since labeled C, El, and E2 proteins were incorporated into extracellular particles less efficiently than in wild-type infection, this suggested that 42S RNA might enhance budding, possibly indirectly via promoting assembly of intracellular nucleocapsids. Therefore, it was of interest to see whether the introduction of a replicating SFV genome into the 26S expression system would facilitate virus production. BHK-21 cells were double infected with vTF7-3 and a recombinant SFV-tr virus (31) and then transfected with the pGEMl-T726S plasmid. SFV-tr virions carry a recombinant SFV genome which contains an intact nonstructural region, but the genes for the structural proteins have been replaced with a gene encoding the human transferrin receptor. This

4744

SUOMALAINEN ET AL.

recombinant genome becomes efficiently packaged into infectious virus particles in a helper genome-based in vivo packaging system (31). Figure 6B shows that the doubleinfected-transfected cells synthesized large amounts of SFV structural proteins and of the transferrin receptor, indicating that both the wild-type 26S mRNA and the SFV-tr genome were expressed. However, extracellular particle production was not significantly enhanced under these conditions. The virus particles produced were now at least partially infectious (Table 1), since 5-h chase media contained 2 x 106 infectious units per ml. However, this amount was 100-fold less than that produced in wild-type infection. Sucrose gradient analysis of pulse-labeled cell extracts revealed that the majority of C proteins were again found in structures that were distinct from those of wild-type nucleocapsids (Fig. 2D). Thus, we conclude that no significant enhancement in virus production was obtained by the inclusion of the SFV-tr genome into the 26S expression system. Although the double-infected-transfected cells produced some infectious virus particles, a 100% coexpression efficiency for wild-type 26S mRNA and the SFV-tr genome could not be confirmed by an indirect immunofluorescence assay. Therefore, the failure of the SFV-tr "complementation" perhaps simply resulted from the inability of SFV and vaccinia virus to efficiently coreplicate in the same cell.

DISCUSSION Several ts mutants of SFV and Sindbis virus which have maturation defects on spike proteins at the restrictive temperature and are thus deficient in the release of infectious virions have been described. These ts mutants were considered useful tools for testing the importance of nucleocapsidspike interactions in virus budding. In SFV tsl- and ts7- and Sindbis virus ts23-infected cells, the mutated spike proteins are not transported to the plasma membrane at the restrictive temperature, and EM analysis of these cells has shown no budding structures at the plasma membrane under these conditions (5, 48, 49, 58). In Sindbis virus ts20-infected cells, the altered spike proteins are able to reach the cell surface, and in these cells, the plasma membrane is lined with nucleocapsids which are not engaged in the budding process (5, 49, 58). A simple conclusion from these studies is that the nucleocapsid cannot bud from the cell without the correct spike-protein interaction. However, the possible release of enveloped, spikeless nucleocapsids from the above-described ts-mutant-infected cells was never tested by quantitative pulse-chase analysis. Moreover, in the case of Sindbis virus ts23 and SFV tsl and ts7, it is completely unknown whether the synthesized nucleocapsid structures are assembly competent at the restrictive temperature. Complementation studies with SFV ts mutants have been unsuccessful, and those with Sindbis virus ts mutants have been very inefficient (7, 63, 64, 66). In the case of Sindbis virus ts2O, the C gene has been found to be correct (33). However, in this case, most nucleocapsids might be bound to defective spikes on the cell surface, as suggested by EM analysis, and therefore not be capable of budding on their own. Considering these problems, the ts-virus studies do not offer any conclusive evidence for the possible importance of nucleocapsid-spike interactions during the budding process. In the present work, we used genetically defined variants of SFV and quantitative biochemical assays to conclusively show that the budding of alphaviruses is mediated by nucleocapsid-spike interactions. When C proteins were expressed from the recombinant SFV-c genome that lacked the

J. VIROL.

genes for the spike proteins, intracellular nucleocapsids were assembled in an apparently normal way, but these nucleocapsids were not released from the cell. Formation of extracellular particles was restored by expression of SFV-c together with the SFV-spike genome, which directed the synthesis of p62 and El. It is noteworthy that virus production by this SFV-c and SFV-spike complementation was significantly less efficient than that obtained during wild-type infections. The most likely explanation for this is that poor expression of spike proteins in the SFV-c- and SFV-spikecotransfected cells affected overall virus production. However, improved virus yields were not obtained from cells which were transfected with the double promoter RNA SFV-c+spike, although this genome directed high expression of all the structural proteins. As the SFV-c+spike and the SFV-spike+c RNAs are longer than the wild-type genome, it is possible that the size of the encapsidated RNA affected the structure and therefore the budding competence of the SFV-c+spike and SFV-spike+c nucleocapsids. Indeed, nucleocapsids from SFV-spike+c-transfected cells sedimented faster on sucrose gradients than nucleocapsids from wild-type-transfected cells. No differences were observed between wild-type and SFV-c+spike nucleocapsids, but minor alterations in nucleocapsid structures might not have been detected by this method. However, it cannot be ruled out that the impaired virus production in all complementation systems was related to some as-yet-unknown mechanism by which the polyprotein expression strategy facilitates virus assembly. From simple topological considerations, it appears most likely that alphavirus nucleocapsids interact with spike heterodimers via the cytoplasmic domain of E2. This domain is 31 residues long, and it is the only part of the spike heterodimer that protrudes into the cytoplasm. However, direct proof for this nucleocapsid-E2 tail interaction is still lacking. For Sindbis virus, it has been shown that amino acid substitutions in the cytoplasmic tail of E2 can lead to defects in virion assembly (15). One explanation for this is that the mutations had altered the nucleocapsid binding site. Recently, Vaux et al. (71) tried to use the anti-idiotype approach to demonstrate structural complementarity between the cytoplasmic tail of the SFV E2 protein and the nucleocapsid surface. However, recent results from our laboratory indicate that the idiotype network antibodies did not reconstruct the nucleocapsid-E2 tail interaction: the anti-idiotypic F13 antibody, which was raised against the anti-E2 tail antibodies, is not nucleocapsid specific as originally claimed but instead most likely recognizes some component of the replication machinery (67). In another recent study, the SFV nucleocapsid has been shown to interact in vitro with synthetic peptides corresponding to the cytoplasmic tail of E2 (41). However, the main control used to determine the specificity of this interaction was the ability of the synthetic peptides to suppress F13 staining of infected cells. One major conclusion from our present data is that the budding mechanism of alphaviruses is different from that of retroviruses: alphavirus budding requires interactions between a cytoplasmic component (the nucleocapsid) and a membrane component (the spike proteins), whereas retrovirus budding can be driven by a cytoplasmic component (the Gag precursor) alone (10, 20, 45). Comparison of alphavirus and retrovirus capsid proteins reveals one distinct difference between these two. The alphavirus C protein lacks any obvious hydrophobic regions (16, 25), and it is thus possible that alphavirus nucleocapsids can bind to membranes only via transmembrane spike proteins. Most retroviral Gag

VOL. 66, 1992

precursors, on the other hand, contain a myristic acid at their amino termini, and this fatty acid modification can perhaps facilitate a direct interaction of the Gag precursor with membrane lipids (10, 20, 44, 46, 53, 76). However, the myristate most likely is not the sole determinant that mediates binding of Gag to membranes, since not all Gag precursors have this modification (11, 26; see also reference 80). Similar to alphaviruses, the envelopment of the hepatitis B virus core has been shown to be dependent on interaction of the core with viral envelope proteins (6). A unique feature of hepatitis B virus is the ability of envelope proteins to independently form subviral lipoprotein particles (55). This suggests that there are strong lateral interactions between the hepatitis B virus spike proteins. The expression of the SFV-spike genome did not result in formation of extracellular envelope particles. However, this does not rule out the possibility that lateral spike interactions have an important role in the budding process. Such lateral contacts are clearly present in the alphavirus envelope, as shown by recent cryo-EM studies (13, 72; see also reference 2). These studies demonstrated that in virion envelopes, the E2-E1 heterodimers are organized as trimeric units arranged in a T=4 icosahedral lattice. Lateral spike-spike interactions seem to persist in Sindbis virus-derived membrane structures from which the nucleocapsids have been removed (73). The phenotypes of the Sindbis virus ts2O and ts103 mutants might reflect the importance of such lateral spike interactions in the virus budding process. The ts103 mutation results in the formation of multicored virus particles, and the ts2O mutation arrests virus budding at the level of the plasma membrane (5, 62). In both cases, the mutations have been mapped to the external domain of E2 (22, 33). One possible interpretation of these results is that the mutations hinder proper lateral contacts between the spike heterodimers and thus cause the aberrant budding phenotypes. However, another possibility is that the mutations affect the E2-E1 heterodimeric association and that this alteration interferes with the binding of spikes to the nucleocapsid. In the present work, we also analyzed the role of the 42S RNA genome in virus assembly and budding. When the vaccinia virus-T7 system was used for the expression of the 26S subgenome, we found that small amounts of particles were released from the cells, although we were unable to detect nucleocapsid structures in the cell lysates. The released particles were very similar in appearance to SFV wild-type virions, as judged by EM analysis. The particles also had the same density as SFV virions. Upon solubilization of the particles with neutral detergent, a structure sedimenting at the same position as viral nucleocapsids was found (data not shown). Whether these "nucleocapsids" contain any RNA is as yet unknown, since a proper analysis of the nucleic acid content is complicated by the low yield of extracellular particles. Taken together, the results indicate that alphavirus structural proteins can assemble into enveloped viruslike particles without the presence of the viral genome, though less efficiently than in wild-type infection. The abilities of nucleocapsid proteins to form capsid structures in the absence of the viral genome have also been documented in some other virus groups. The retrovirus Gag protein can assemble into budding-competent core structures without the viral genome (20, 34). Similarly, the hepatitis B virus and measles virus core proteins can form capsidlike structures in the absence of viral genomes, but it is not known whether these are budding competent (3, 60).

BUDDING OF ALPHAVIRUSES

4745

ACKNOWLEDGMENTS We are very grateful to Bernard Moss for providing vaccinia virus vTF7-3, to Harm Snippe for providing the monoclonal anti-El and anti-E2 antibodies, and to Irene Greiser-Wilke for providing the monoclonal anti-C antibody. EM analyses were performed by Kjell Hultenby, and his assistance is gratefully acknowledged. We also thank Johanna Wahlberg for providing the pSFV-spike plasmid, Antti Salminen for his contributions to the construction of the pSFV-c plasmid, Kerstin Forsell for technical assistance, Jimena Parga Rios for cell culture, and Joan Smyth for critically reading the manuscript. This work was supported by the Swedish Medical Research Council (B90-12X-08272-03A), the Swedish National Board for Technical Development (90-00347P), and the Swedish Natural Science Research Council (B-BU-9353-303). REFERENCES 1. Aliperti, G., and M. J. Schlesinger. 1978. Evidence for an autoprotease activity of Sindbis virus capsid protein. Virology 90:366-369. 2. Anthony, R. P., and D. T. Brown. 1991. Protein-protein interactions in an alphavirus membrane. J. Virol. 65:1187-1194. 3. Birnbaum, F., and M. Nassal. 1990. Hepatitis B virus nucleocapsid assembly: primary structure requirements in the core protein. J. Virol. 64:3319-3330. 4. Boere, W. A. M., T. Harmsen, J. Vinje, B. J. Benaissa-Trouw, C. A. Kraaieveld, and H. Snippe. 1984. Identification of distinct antigenic determinants on Semliki Forest virus by using monoclonal antibodies with different antiviral activities. J. Virol. 52:575-582. 5. Brown, D. T., and J. F. Smith. 1975. Morphology of BHK-21 cells infected with Sindbis virus temperature-sensitive mutants in complementation groups D and E. J. Virol. 15:1262-1266. 6. Bruss, V., and D. Ganem. 1991. The role of envelope proteins in hepatitis B virus assembly. Proc. Natl. Acad. Sci. USA 88: 1059-1063. 7. Burge, B. W., and E. R. Pfefferkorn. 1966. Complementation between temperature-sensitive mutants of Sindbis virus. Virology 30:214-223. 8. Choi, H.-K, L. Tong, W. Minor, P. Dumas, U. Boege, M. G. Rossmann, and G. Wengler. 1991. The crystal structure of Sindbis virus core protein and a proposed structure of the capsid. Nature (London) 354:37-43. 9. de Curtis, I., and K. Simons. 1988. Dissection of Semliki Forest virus glycoprotein delivery from the trans-Golgi network to the cell surface in permeabilized BHK cells. Proc. Natl. Acad. Sci. USA 85:8052-8056. 10. Delchambre, M., D. Gheysen, D. Thines, C. Thiriart, E. Jacobs, E. Verdin, M. Horth, A. Burny, and F. Bex. 1989. The GAG precursor of simian immunodeficiency virus assembles into virus-like particles. EMBO J. 8:2653-2660. 11. Erdie, C. R., and J. W. Wills. 1990. Myristylation of Rous sarcoma virus Gag protein does not prevent replication in avian cells. J. Virol. 64:5204-5208. 12. Fuerst, T. R., E. G. Niles, F. W. Studier, and B. Moss. 1986. Eukaryotic transient-expression system based on recombinant vaccinia virus that synthesize bacteriophage T7 RNA polymerase. Proc. Natl. Acad. Sci. USA 83:8122-8126. 13. Fuller, S. D. 1987. The T=4 envelope of Sindbis virus is organized by interactions with a complementary T=3 capsid. Cell 48:923-934. 14. Gaedigk-Nitschko, K, and M. J. Schlesinger. 1990. The Sindbis virus 6K protein can be detected in virions and is acylated with fatty acids. Virology 175:274-281. 15. Gaedigk-Nitschko, K., and M. J. Schlesinger. 1991. Site-directed mutations in Sindbis virus E2 glycoprotein's cytoplasmic domain and the 6K protein lead to similar defects in virus assembly and budding. Virology 183:206-214. 16. Garoff, H., A.-M. Frischauf, K. Simons, H. Lehrach, and H. Delius. 1980. The capsid protein of Semliki Forest virus has clusters of basic amino acids and prolines in its amino-terminal region. Proc. Natl. Acad. Sci. USA 77:6376-6380. 17. Garoff, H., D. Huylebroeck, A. Robinson, U. Tiliman, and P.

4746

18. 19. 20.

21. 22.

23. 24.

25. 26.

27.

28. 29. 30.

31.

32.

33.

34.

35. 36. 37.

38.

SUOMALAINEN ET AL.

Liljestrom. 1990. The signal sequence of the p62 protein of Semliki Forest virus is involved in initiation but not in completing chain translocation. J. Cell Biol. 111:867-876. Garoff, H., C. Kondor-Koch, and H. Riedel. 1982. Structure and assembly of alphaviruses. Curr. Top. Microbiol. Immunol. 99:1-50. Garoff, H., K. Simons, and B. Dobberstein. 1978. Assembly of Semliki Forest virus membrane glycoproteins in the membrane of the endoplasmic reticulum in vitro. J. Mol. Biol. 124:587-600. Gheysen, D., E. Jacobs, F. de Foresta, C. Thiriart, M. Francotte, D. Thines, and M. De Wilde. 1989. Assembly and release of HIV-1 precursor Pr55SaS virus-like particles from recombinant baculovirus-infected cells. Cell 59:103-112. Greiser-Wilke, I., V. Moennig, O.-R. Kaaden, and L. T. M. Figueiredo. 1989. Most alphaviruses share a conserved epitopic region on their nucleocapsid protein. J. Gen. Virol. 70:743-748. Hahn, C. S., C. M. Rice, E. G. Strauss, E. M. Lenches, and J. H. Strauss. 1989. Sindbis virus ts103 has a mutation in glycoprotein E2 that leads to defective assembly of virions. J. Virol. 63:34593465. Hahn, C. S., and J. H. Strauss. 1990. Site-directed mutagenesis of the proposed catalytic amino acids of the Sindbis virus capsid protein autoprotease. J. Virol. 64:3069-3073. Hashimoto, K., S. Erdei, S. Keranen, J. Saraste, and L. Kaariainen. 1981. Evidence for a separate signal sequence for the carboxy-terminal envelope glycoprotein El of Semliki Forest virus. J. Virol. 38:34-40. Helenius, A., and H. S6derlund. 1973. Stepwise dissociation of the Semliki Forest virus membrane with Triton X-100. Biochim. Biophys. Acta 307:287-300. Henderson, L. E., R. C. Sowder, G. W. Smythers, and S. Oroszlan. 1987. Chemical and immunological characterizations of equine infectious anemia virus gag-encoded proteins. J. Virol. 61:1116-1124. Jain, S. K., S. DeCandido, and M. Kielian. 1991. Processing of the p62 envelope precursor protein of Semliki Forest virus. J. Biol. Chem. 266:5756-5761. Kawai, S., and H. Hanafusa. 1973. Isolation of defective mutant of avian sarcoma virus. Proc. Natl. Acad. Sci. USA 70:34933497. Kielian, M., and S. Jungerwirth. 1990. Mechanism of enveloped virus entry into cells. Mol. Biol. Med. 7:17-31. Liljestrom, P., and H. Garoff. 1991. Internally located cleavable signal sequences direct the formation of Semliki Forest virus membrane proteins from a polyprotein precursor. J. Virol. 65:147-154. Liljestr6m, P., and H. Garoff. 1991. A new generation of animal cell expression vectors based on the Semliki Forest virus replicon. Bio/Technology 9:1356-1361. Liljestrom, P., S. Lusa, D. Huylebroeck, and H. Garoff. 1991. In vitro mutagenesis of a full-length cDNA clone of Semliki Forest virus: the small 6,000-molecular-weight membrane protein modulates virus release. J. Virol. 65:4107-4113. Lindqvist, B. H., J. DiSalvo, C. M. Rice, J. H. Strauss, and E. G. Strauss. 1986. Sindbis virus mutant ts2O of complementation group E contains a lesion in glycoprotein E2. Virology 151:1020. Linial, M. L., and A. D. Miller. 1990. Retroviral RNA packaging: sequence requirements and implications, p. 125-152. In R. Swanstrom and P. K. Vogt (ed.), Retroviruses, strategies of replication. Springer-Verlag, Berlin. Lobigs, M., and H. Garoff. 1990. Fusion function of the Semliki Forest virus spike is activated by proteolytic cleavage of the envelope glycoprotein p62. J. Virol. 64:1233-1240. Lobigs, M., J. M. Wahlberg, and H. Garoff. 1990. Spike protein oligomerization control of Semliki Forest virus fusion. J. Virol. 64:5214-5218. Lobigs, M., H. Zhao, and H. Garoff. 1990. Function of Semliki Forest virus E3 peptide in virus assembly: replacement of E3 with an artificial signal peptide abolishes spike heterodimerization and surface expression of El. J. Virol. 64:4346-4355. Mackett, M., G. L. Smith, and B. Moss. 1985. The construction and characterisation of vaccinia virus recombinants expressing

J. VIROL.

foreign genes, p. 191-211. In D. M. Glover (ed.), DNA cloning. IRL Press, Oxford. 39. Melancon, P., and H. Garoff. 1986. Reinitiation of translocation in the Semliki Forest virus structural polyprotein: identification of the signal for the El glycoprotein. EMBO J. 5:1551-1560. 40. Melancon, P., and H. Garoff. 1987. Processing of the Semliki Forest virus structural polyprotein: role of the capsid protease. J. Virol. 61:1301-1309. 41. Metsikko, K., and H. Garoff. 1990. Oligomers of the cytoplasmic domain of the p62/E2 membrane protein of Semliki Forest virus bind to the nucleocapsid in vitro. J. Virol. 64:4678-4683. 42. Murphy, F. A. 1980. Togavirus morphology and morphogenesis, p. 241-316. In R. W. Schlesinger (ed.), The togaviruses, biology, structure, replication. Academic Press, Inc., New York. 43. Presley, J. F., and D. T. Brown. 1989. The proteolytic cleavage of PE2 to envelope glycoprotein E2 is not strictly required for the maturation of Sindbis virus. J. Virol. 63:1975-1980. 44. Rein, A., M. R. McClure, N. Rice, R. B. Luftig, and A. M. Schultz. 1986. Myristylation site in Pr65-gag is essential for virus particle formation by Moloney murine leukemia virus. Proc. Natl. Acad. Sci. USA 83:7246-7250. 45. Rhee, S. S., H. Hui, and E. Hunter. 1990. Preassembled capsids of type D retroviruses contain a signal sufficient for targeting specifically to the plasma membrane. J. Virol. 64:3844-3852. 46. Rhee, S. S., and E. Hunter. 1987. Myristylation is required for intracellular transport but not for assembly of D-type retrovirus capsids. J. Virol. 61:1045-1053. 47. Salminen, A., J. M. Wahlberg, M. Lobigs, P. Liljestrom, and H. Garoff. 1992. Membrane fusion process of Semliki Forest virus II: cleavage dependent reorganization of the spike protein complex controls virus entry. J. Cell Biol. 116:349-357. 48. Saraste, J., C.-H. von Bonsdorff, K. Hashimoto, L. Kiiariainen, and S. Keranen. 1980. Semliki Forest virus mutants with temperature-sensitive transport defect of envelope proteins. Virology 100:229-245. 49. Saraste, J., C.-H. von Bonsdorff, K. Hashimoto, S. Keranen, and L. Kiariainen. 1980. Reversible transport defect of virus membrane glycoproteins in Sindbis virus mutant infected cells. Cell Biol. Int. Rep. 4:279-286. 50. Scheele, C. M., and H. Hanafusa. 1971. Proteins of helperdependent RSV. Virology 45:401-410. 51. Schlesinger, M. J., and S. Schlesinger. 1986. Formation and assembly of alphavirus glycoproteins, p. 121-148. In S. Schlesinger and M. J. Schlesinger (ed.), The Togaviridae and Flaviviridae. Plenum Press, New York. 52. Schlesinger, S., and M. J. Schlesinger. 1986. The Togaviridae and Flaviviridae. Plenum Press, New York. 53. Schultz, A. M., and A. Rein. 1989. Unmyristylated Moloney murine leukemia virus Pr65rag is excluded from virus assembly and maturation events. J. Virol. 63:2370-2373. 54. Shields, A., 0. N. Witte, E. Rothenberg, and D. Baltimore. 1978. High frequency of aberrant expression of Moloney murine leukemia virus in clonal infections. Cell 14:601-609. 55. Simon, K., V. R. Lingappa, and D. Ganem. 1988. Secreted hepatitis B surface antigen polypeptides are derived from a transmembrane precursor. J. Cell Biol. 107:2163-2168. 56. Simons, K., and H. Garoff. 1980. The budding mechanism of enveloped animal viruses. J. Gen. Virol. 50:1-21. 57. Simons, K., and G. Warren. 1984. Semliki Forest virus: a probe for membrane traffic in the animal cell. Adv. Protein Chem. 36:79-132. 58. Smith, J. F., and D. T. Brown. 1977. Envelopment of Sindbis virus: synthesis and organization of proteins in cells infected with wild-type and maturation-defective mutants. J. Virol. 22: 662-678. 59. Soderlund, H. 1973. Kinetics of formation of the Semliki Forest virus nucleocapsid. Intervirology 1:354-361. 60. Spehner, D., A. Kirn, and R. Drillien. 1991. Assembly of nucleocapsidlike structures in animal cells infected with a vaccinia virus recombinant encoding the measles virus nucleoprotein. J. Virol. 65:6296-6300. 61. Stegmann, T., R. W. Doms, and A. Helenius. 1989. Proteinmediated membrane fusion. Annu. Rev. Biophys. Biophys.

VOL. 66, 1992

Chem. 18:187-211. 62. Strauss, E. G., C. R. Birdwell, E. M. Lenches, S. E. Staples, and J. H. Strauss. 1977. Mutants of Sindbis virus. II. Characterization of a maturation-defective mutant, ts103. Virology 82:122149. 63. Strauss, E. G., E. M. Lenches, and J. H. Strauss. 1976. Mutants of Sindbis virus. I. Isolation and partial characterization of 89 new temperature-sensitive mutants. Virology 74:154-168. 64. Strauss, E. G., and J. H. Strauss. 1980. Mutants of alphaviruses: genetics and physiology, p. 393-426. In R. W. Schlesinger (ed.), The Togaviruses, biology, structure, replication. Academic Press, Inc., New York. 65. Strauss, E. G., and J. H. Strauss. 1986. Structure and replication of the alphavirus genome, p. 35-90. In S. Schlesinger and M. J. Schlesinger (ed.), The Togaviridae and Flaviviridae. Plenum Press, New York. 66. Strauss, E. G., H. Tsukeda, and B. Simizu. 1983. Mutants of Sindbis virus. IV. Heterotypic complementation and phenotypic mixing between temperature-sensitive mutants and wildtype Sindbis and western equine encephalitis viruses. J. Gen. Virol. 64:1581-1590. 66a.Suomalainen, M. Unpublished data. 67. Suomalainen, M., and H. Garoff. 1992. Alphavirus spike-nucleocapsid interaction and network antibodies. J. Virol. 66:51065109. 68. Suomalainen, M., H. Garoff, and M. Baron. 1990. The E2 signal sequence of rubella virus remains part of the capsid protein and confers membrane association in vitro. J. Virol. 64:5500-5509. 69. Ulmanen, I., H. Soderlund, and L. Kaariiinen. 1976. Semliki Forest virus capsid protein associates with the 60S ribosomal subunit in infected cells. J. Virol. 20:203-210. 70. Ulmanen, I., H. Soderlund, and L. Kiiriainen. 1979. Role of protein synthesis in the assembly of Semliki Forest virus nucleocapsid. Virology 99:265-276. 71. Vaux, D. J. T., A. Helenius, and I. Mellman. 1988. Spikenucleocapsid interaction in Semliki Forest virus reconstructed using network antibodies. Nature (London) 336:36-42. 72. Vogel, R. H., S. W. Provencher, C.-H. von Bonsdorff, M. Adrian, and J. Dubochet. 1986. Envelope structure of Semliki

BUDDING OF ALPHAVIRUSES

73. 74.

75.

76.

77.

78.

79.

80.

81. 82.

4747

Forest virus reconstructed from cryo-electron micrographs. Nature (London) 320:533-535. von Bonsdorff, C.-H., and S. C. Harrison. 1978. Hexagonal glycoprotein arrays from Sindbis virus membranes. J. Virol. 28:578-583. Wahlberg, J. M., W. A. Boere, and H. Garoff. 1989. The heterodimeric association between the membrane proteins of Semliki Forest virus changes its sensitivity to mildly acidic pH during virus maturation. J. Virol. 63:4991-4997. Watson, D. G., J. M. Moehring, and T. J. Moehring. 1991. A mutant CHO-Kl strain with resistance to Pseudomonas exotoxin A and alphaviruses fails to cleave Sindbis virus glycoprotein PE2. J. Virol. 65:2332-2339. Weaver, T. A., and A. T. Panganiban. 1990. N myristylation of the spleen necrosis virus matrix protein is required for correct association of the Gag polyprotein with intracellular membranes and for particle formation. J. Virol. 64:3995-4001. Wengler, G. 1990. Structure and synthesis of the core protein: role in regulation of assembly and disassembly of alphavirus and flavivirus cores, p. 227-236. In M. A. Brinton and F. X. Heinz (ed.), New aspects of positive-strand RNA viruses. American Society for Microbiology, Washington. Wengler, G., U. Boege, G. Wengler, H. Bischoff, and K. Wahn. 1982. The core protein of the alphavirus Sindbis virus assembles into core-like nucleoproteins with the viral genome RNA and with other single-stranded nucleic acids in vitro. Virology 118:401-410. Wengler, G., G. Wengler, U. Boege, and K. Wahn. 1984. Establishment and analysis of a system which allows assembly and disassembly of alphavirus core-like particles under physiological conditions in vitro. Virology 132:401-412. Wills, J. W., R. C. Craven, R. A. Weldon, Jr., T. D. Nelle, and C. R. Erdie. 1991. Suppression of retroviral MA deletions by the amino-terminal membrane-binding domain of p60". J. Virol. 65:3804-3812. Zhao, H., and H. Garoff. Unpublished data. Ziemiecki, A., H. Garoff, and K. Simons. 1980. Formation of the Semliki Forest virus membrane glycoprotein complexes in the infected cell. J. Gen. Virol. 50:111-123.