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, 206-2 14 (199 1)

Site-Directed Mutations in Sindbis Virus E2 Glycoprotein’s Cytoplasmic Domain and the 6K Protein Lead to Similar Defects in Virus Assembly and Budding KERSTIN GAEDIGK-NITSCHKO’ Department

of Molecular

Microbiology,

Washington

AND University

MILTON J. SCHLESINGER* School of Medicine,

St. Louis, Missouri

63 110

Received February 22, 199 1; accepted April 1, 199 1 Site-directed mutagenesis was used to obtain four mutants with amino acid replacements in the cytoplasmic domain of the E2 glycoprotein and three with replacements in the 6K protein of Sindbis virus. All but one of these mutants yielded progeny virus after transfection of chicken embryo fibroblasts with RNA prepared by in vitro transcription of the virus cDNA; however, even this nonproducer mutant made virus structural proteins in the transfected cells. The other six mutants divided into two groups based on growth in chicken embryo fibroblasts. One group of four mutants (two in E2 and two in 6K) was indistinguishable from wild-type in formation of infectious virus in avian cells while the other group, consisting of two mutants, grew significantly slower. All six mutants grew slower than the parental wild-type virus in mosquito cells. In avian cells, all mutants produced extracellular particles at a slower rate than the wild-type and many of the particles contained multiple nucleocapsids, based on electron microscopy and kinetics of thermal inactivation, One of the E2 mutants with a cysteine changed to alanine and the 6K mutant with four cysteines replaced were deficient in covalent-bound palmitic acid. Two mutants with changes near the signalase cleavage sites between E2 and 6K and between 6K and El appeared to be defective in proteolytic processing. Despite individual differences, all of these mutants and the two previously described produced similar phenotypes in which multicored infectious virus o 1991 Academic PESS, Inc. particles were released more slowly from mosquito cells than from avian cells.

activities of network antibodies (Vaux et al., 1988). Results of the latter showed that antibodies raised against peptides with sequences derived from the cytoplasmic domain of the E2 glycoprotein could be used to obtain anti-idiotypes that reacted with epitopes on the virus nucleocapsid. Thus, there must be complementarity in structure between these virus components. The actual envelopment of the nucleocapsid by the lipid bilayer is presumed to be driven by the binding of the membrane-embedded glycoproteins to nucleocapsids (Ziemiecki and Garoff, 1978; Fuller, 1987; Metsikko and Garoff, 1990). Based on this model, we predicted that mutations altering amino acids in the cytoplasmic domain of E2 should affect virus assembly. The ability to make such site-specific mutations became possible when Sindbis virus RNA, prepared by in vitro transcription of a cDNA encoding the entire sequence of the Sindbis virus genome, was found to be infectious (Rice et a/., 1987). This system has allowed us to prepare site-directed mutations in the E2 cytoplasmic domain and examine effects on virus budding. In this paper we describe four such mutants as well as another three mutants with alterations in the sequence of the small hydrophobic protein noted as 6K. We tested the latter because we found that the 6K protein was important for efficient virus assembly and release (Gaedigk-Nitschko et al., 1990). Although these seven mutants have unique defects, we were surprised to discover that six of them

INTRODUCTION The release of newly replicated alphaviruses from host cells requires specific interactions between nucleocapsids and transmembranal virus-specific glycoproteins. A model based on this kind of protein-protein interaction for the assembly and budding of Semliki Forest and Sindbis viruses, the prototypic alphaviruses, was first proposed more than 10 years ago (Garoff and Simons, 1974; Smith and Brown, 1977; Simons and Garoff, 1980; Strauss et al., 1980). In the model, the virus nucleocapsid was postulated to bind to cytoplasmic domains of the virus transmembranal glycoproteins that had been transported to the plasma membrane of the infected cell. Once positioned at the cell’s plasma membrane, the nucleocapsid acted as a “trap” for binding additional glycoprotein spikes and these interactions led to the envelopement and budding of new virions. This model was initially based on electron microscopic analyses of virus-infected cells (Acheson and Tamm, 1967; Brown and Waite, 1972; Birdwell et al., 1973; Torrisi and Bonatti, 1985) but it has been strongly supported by recent data describing

’ Current address: GSF-Forschungszentrum fur Umwelt und Gesundheit, GmbH, Department of Molecular Cell Pathology, Ingolstaedter Landstrasse 1, D-8042 Neuherberg, Germany. ’ To whom correspondence and reprint requests should be addressed. 0042-6822191

$3.00

Copyright 0 199 1 by Academic Press. Inc All rights of reproduction in any form resewed.

206

SITE-DIRECTED

MUTATIONS

AFFECTING

produced very similar phenotypes and we discuss these data in terms of the model described above.

SINDBIS VIRUS ASSEMBLY

(A) Sindbis Virus Genome NSPI 5’

MATERIALS AND METHODS The preparations of cells and virus, reagents for recombinant DNA experiments and methods for preparing site-directed mutations were carried out as previously described (Gaedigk-Nitschko et al., 1990). The reconstructed cDNA of the Sindbis virus was sequenced in the region surrounding the newly inserted mutationally altered fragment (Sanger eta/., 1977). The specific sites altered and the substituted amino acids in the carboxy terminus of the E2 glycoprotein and the 6K protein are presented in Fig. 1. In vitro transcription of linearized cDNA and transfection of cells has been described (Levis et al., 1986). Three separate transfections were performed and the specific infectivities of the RNAs (plaques/pg RNA) were similar in each. One infectious RNA that carried three site mutations in E2 gave only one plaque in three transfections. We suspected this plaque arose by a revertant virus of some kind and although it was phenotypically distinct from wild-type, we have not included it in the data here. All mutants except for the one nonproducer were plaque purified and grown from low m.o.i. infection on chicken embryo fibroblasts to prepare stocks of virus. The methods for plaque assays, immunofluorescence, isolation of virus from the cell culture fluid, nucleocapsids from cell lysates, extraction and analysis of intracellular virus proteins, and quantitation of virus proteins and RNA were previously described (Gaedigk-Nitschko et a/., 1990; Gaedigk-Nitschko and Schlesinger, 1990). Virus was also purified by affinity chromatography with Cellufine (Amicon). The resin as supplied by the manufacturer was washed thoroughly with phosphate-buffered saline solution to adjust the salt concentration to 0.15 M NaCI. Approximately 200 PI of packed resin was suspended with the culture medium from 1O6 infected cells and rocked for 45 min. The resin was centrifuged and washed three times with saline solution and the protein removed by boiling the resin with 100 ~1 of Laemmli SDS/PAGE loading buffer. Analysis of surface glycoproteins with anti-Sindbis virus antibodies was carried out on cells (3 X 106) that had been pulse labeled for 15 min with [35S]cysteine and chased for 60 min. The pulse labeled cells served as a “background” since it requires about 20 min for the nascent glycoprotein to appear on the cell surface. The pulse- and chase-labeled cells were washed several times with cold saline solution followed by incubation with an IgG fraction of anti-Sindbis virus antibodies at 4” for 30 min. Unreacted antibodies were removed and cell monolayers washed with saline several times.

207

NSPP

NSPB

NSP4

c

P62

EK

El

1w

j

3’

(B) E2 carboxyl-terminus 395 399

415416

419

Ii S. A.

I C

--E.C.L.T.P.Y.A.L.A.P.N.A.V.I.P.T.S.L.A.L.L.C.C.V.R.S.A.N.A.--(6K) I F

1 S

(C) 6K carboxyl-terminus 343536 3839 ---L.M.R.C.C.S.C.C.L.P.F.L.V.V.A.G.A.Y.A.L.A.K.V.D.A.--(El) III A S. A.

52 I A

Ii S. A.

FIG. 1. The Sindbis vrrus genome (A) and partial sequences of E2 cytoplasmic domain (B) and 6K protein (C) with indicated mutated sites and amino acrds replacements. The open squares in (A) indicate the portions of the genome shown in (B) and (C). The numbering of the residues is based on the sequence of Strauss and Strauss (1986) starting with position 1 of mature E2 and 6K.

Unlabeled Sindbis virus (10 pg) was added and the cells lysed by addition of 0.4 ml RIPA buffer (Kelley and Schlesinger, 1982). Two additional 0.4-ml samples of RIPA were used to collect cell lysates and the antibody-antigen complex was precipitated by addition of lyophilized and washed preparation of S. aureus, Cowan strain as previously described (Gaedigk-Nitschko and Schlesinger, 1990). The immunoprecipitates were washed with RIPA buffer and solubilized by boiling with 50 ~1 of Laemmli SDS/PAGE loading buffer. Quantitative analyses were carried out on bands separated in 10% SDS/PAGE (Gaedigk-Nitschko et a/., 1990).

RESULTS The seven mutants analyzed in this study segregated into three groups when tested for formation of infectious virus in cultured chicken embryo fibroblasts. Group I showed no progeny virus from the initial transfection although we could detect by immunofluorescence the formation of virus glycoproteins in about 1O/O of the cells. This value is about the optimum that we could attain for transfection with in-vitro transcribed viral RNA. Only one of the mutants, which had two cysteines at positions 4 15 and 416 of E2 replaced with serine and alanine (Fig. l), was in this “nonproducer” group. There were not enough transfected cells to determine how this mutation affected virus replication although the production of virus glycoprotein antigens indicated that the block was not in the early events in virus replication. Group II mutants produced infectious virus at a significantly slower rate than wild-type virus in both avian and insect cells (Table 1) and released particles at 8 to 27% that of wild-type from avian cells (Table 2). Two

GAEDIGK-NITSCHKO

208 TABLE 1 FORMATIONOF INFECTIOUSVW? E2 mutants

Cells CEF 4 hr 5 6 7.5 c7-10 16hr 24

Wildtype

0.08 1 8 20 20 200

3958 (Ill)

0.1 10 20 30 1 100

6K mutants

399F 419C (111) (11)

0.1 10 20 30 0.2 30

0.001 0.003 0.04 8 0.5 10

35S,36A; 38S,39A (III)

34A (111)

52Ab (11)

0.08 1.0 8 10

0.08 1.0 8 20

0.004 0.007 0.05 0.40

0.3 15

0.1 20

0.01 0.3

a Measured as plaque-forming units (X10*) on CEF at 37”. lo6 cells were infected with an m.o.i. = 20. CEF were grown at 37”; C7-10 were grown at 28”. b Small plaques. c Mutant group-refer to text.

mutants were in this class-one from E2 at position 419 (replacement of serine by cysteine) and one from the 6K protein at position 52 (replacement of lysine by alanine). In contrast, the four mutants in group III produced infectious virus as well as the wild-type in avian cells but grew slower than the wild-type in mosquito cells (Table 1). Despite their similarity to wild-type virus in growth of infectious virus in avian cells, these four mutants secreted virus particles at a rate that was 8 to 38% that of the wild-type virus, as measured either by radiolabeling of protein or RNA (Table 2). One mutant in group III had a change in E2 at position 395 (replacement of cysteine by serine) and another had the tyrosine at position 399 in E2 replaced by phenylalanine. For the two other group III mutants, one had a change at position 34 in the 6K protein (replacement of arginine by alanine) and one had changes at positions 35.36.38.39 (replacement of cysteines by serines and alanines). Three events in the replication of Sindbis virus were analyzed in avian cells infected with plaque-purified mutant viruses: (1) formation of nucleocapsids, (2) transport of the glycoproteins to the cell surface; and (3) proteolytic conversion leading to E2 maturation. During a 2-hr labeling period, the ratio of intracellular nucleocapsid protein to total viral capsid protein for wild-type virus was 0.4 compared with values of 0.6 to 0.9 for the mutants (Table 2). About 9% of the glycoproteins made in a 15-min pulse of wild-type virus-infected cells could be detected on the cell surface after a 60-min chase. The equivalent values for the mutant-

AND SCHLESINGER

infected cells ranged from 9 to 19% (Table 2). These relative increases in amounts of cell-associated viral products (nucleocapsids and surface glycoproteins) in the mutant-infected cells were consistent with the accompanying decrease in the rate of virus-particle secretion noted above and they indicated also that the mutations had not affected synthesis of virus structural gene products or their normal cellular locations. The proteolytic processing of the E2 glycoprotein from p62 (Schlesinger and Schlesinger, 1972) was measured in pulse-chase studies of wild-type and mutant virus-infected cells. Changes in the rate of p62 processing in these mutants were not considered significantly different from the wild-type (Table 2), but the two mutants in group II showed a qualitative change in the pattern of glycoproteins as detected by SDS/PAGE of pulse-labeled infected cells. Mutant 41 SC-infected cells contained a protein band that moved slower whereas mutant 52A-infected cells had a band that moved faster than the normal virus glycoproteins (Fig. 2, arrowheads). The slower band was immunoprecipitated by anti E2 antibodies and the faster one was immunoprecipitated by both anti 6K and anti El antibodies (data not presented). We have reported previously that the 6K protein appears in infected cells in two forms with mobilities in gels of 6K and 4K, and 6K was more palmitoylated than 4K (Gaedigk-Nitschko and Schlesinger, 1990). The ratio of 6K to 4K protein in cells infected with wild type virus was 0.2, and the values (0.14-0.2) for all the mutants except the 6K mutant deficient in four cysteines were not considered significantly different from the wild-type. Two of the mutants, 52A and 35S,36A, 38S,39A showed changes in the mobility of the 6K protein (Fig. 2, bottom). In mutant 52A, the 6K protein and also its 4K isoform migrated slightly slower than wildtype after SDS/PAGE. The mobility of the 6K protein in mutant 35S,36A,38S,39A was anomalous, presumably because the four cysteines had been removed. A single cysteine in the sequence of 6K at position 23 accounted for the labeled bands at the bottom of Fig. 2. We speculate that the changes in virus glycoproteins in 419C and 52A and perhaps in the 6K and 4K proteins of 52A occurred because the amino acid substitutions, which were near the carboxy terminus of E2 and 6k, led to new signal peptidase cleavage sites in the E2-6K and 6K-El junctions (Fig. 3). The carboxyterminal hydrophobic sequences of E2 and 6K can function as signal sequences for membrane translocation with cleavage between E2 and 6K and 6K and El carried out by signalase (Melancon and Garoff, 1986; Liljestrom and Garoff, 1991). Two of the mutants, 395s of E2 and 35S,36A,

SITE-DIRECTED

MUTATIONS

AFFECTING

SINDBIS VIRUS ASSEMBLY

209

TABLE 2 FORMATIONOF EXTRACELLULARVIRUS PARTICLES,INTRACELLULARNUCLEOCAPSIDS,AND VIRUS GLYCOPROTEINS E2 mutants

Particle secretionb C V V/(V + C)% RNAd Nucleocapsid formatior? V+NC NC/(V + NC) Maturation of E2’ P62 to E2 Glycoprotein transportg % at surface

6K mutants

Wild-type

395s (Ill)

399F (111)

419c (11)

43.8 5.8 13 15

60.0 3.1 5 (38) 5 (33)

57.0 1.3 2 (15) 5 (33)

60.4 1.4 2 (15) 4 (27)

45 0.4

21 0.6

34 0.8

3

4

9

6

35S,36A,37S,38A (111)

36A (111)

52A (11)

53.8 0.6 1 (8) 4 (27)

50.1 1.1 2 (15) 2 (13)

55.7 0.4 1 (8) 1 (8)

48 0.7

27 0.9

31 0.7

25 0.9

3

2

4

3

2

19

19

9

16

14

a Refers to mutation group (see text). ’ Measured as virus-specific 35S-labeled proteins, cpm x 1Oe3. C-cell extracts; V, virions isolated from the media. Labeling period was a 60-min chase after a 15-min pulse at 4 hr postinfection of 1O6 cells with m.o.i. = 20. ’ Numbers in parentheses are the percentage of the wild-type value. d Numbers are the fold increase in counts per minute of [3H] uridine incorporated into virus particles isolated from media by adsorption to Cellufine (see Materials and Methods). Infected cells (5 X 106; m.o.i. = 30) were labeled at 6 hr postinfection with 100 pCi [3H]uridine (37 Ci/mm). Samples of media were collected every 30 min for 2 hr and precipitated with trichloroacetic acid. Values are based on differences between 7 and 8 hr postinfection. a Numbers are the cpm X 1 Om3in the capsid protein measured after SDS/PAGE. V, virus particles isolated from the media by Cellufine adsorption. NC, nucleocapsids extracted from infected cells followed by rate zonal centrifugation In a sucrose gradient. In the experiment, 10’ cells were infected with an m.o.i. of 50 and cells labeled from 4 to 6 hr postinfection with 30 &i of [35S]methionine. ’ Numbers are the fold increase in the ratio of cpms in E2:p62 as measured from bands excised after SDS/PAGE in 10% polyacrylamide gels. The samples measured were from chase periods of 10 and 20 min after a 15-min pulse of [35S]methionine given 4 hr postinfection. g Numbers are the ratio of cpms in surface glycoproteins El and E2 to total labeled glycoproteins in lo6 cells infected with m.o.i. = 40 and labeled at 5 h post infection for 45 min following a 15.min pulse of 100 &i of [35S]cysteine. The values are corrected for a background that is obtained from the 15.min pulse-labeled cells. Refer to Materials and Methods for details.

38S,39A of 6K were altered in cysteines that have been postulated to be sites of palmitoylation (Magee eta/., 1984; Gaedigk-Nitschko and Schlesinger, 1990). Our analysis of infected cells labeled with [3H]palmitic acid showed that the 6K mutant lacking four cysteines had low levels of lipid label in the 6K protein (Fig. 2, middle, and Table 3). The values for fatty acylation of the glycoproteins showed that there was less lipid in E2 of mutant 3958, based on normalizing values to the El lipid label. For 6K mutants, both the 34A and 52A mutants had low levels of acylated El and E2. The basis for the results with the latter two mutants was unclear, although we noted above that the El glycoprotein formation was altered in the 52A mutant (Fig. 2). Our initial studies with two site-directed mutations in the 6K Sindbis virus protein revealed that the mutated virions were more thermostable than the wild-type. We tested the six new mutants for thermal stability at 56“ and found that all were more stable than wild-typevirus (Table 4). The viruses examined in Table 4 were grown

in avian cells but similar studies with viruses grown in insect cells showed that the mutants were much more thermostable than wild-type, and mutants from both avian and insect cells showed multihit thermal inactivation kinetics (data not presented). These results were indicative of multicored virus particles which we detected previously by electron microscopy of infected cells and purified virions (Gaedigk-Nitschko et al., 1990). Similar kinds of multicored particles were found near surfaces of cells infected with these mutants (Fig. 4) but other differences were also observed. For example, the E2 399F mutant (Fig. 4C) showed an unusually large number of nucleocapsids aligned underneath the cell surface membrane and the 6K 52A mutant had patches of dense aggregates scattered along membranes of infected cells (Fig. 4D). DISCUSSION With the seven mutants described here and two previous ones, we have examined thus far nine muta-

210

GAEDIGK-NITSCHKO wt

395s

P c

419C

399s

P c

PC

35S, 36A 38S, 39A

34A

52A

PC

PC

PC

PC

AND SCHLESINGER TABLE 3 EFFECT

OF MUTATIONS

E’i= E2

6K

OF

FATTY ACYLATION

PROTEINS

virus proteins formed in mutant and FIG. 2. Pattern of intracellular wild type virus-infected chicken embryo fibroblasts. The mutants are listed at the top; Wt is wild-type. P refers to a 15-min pulse of [35S]methionine (25 &i to 1 O6cells infected at an m.o.i. of 20) added 4.5 hr postinfection. C refers to a 60-min chase after the pulse. Top: the glycoproteins (~62, El and E2) and capsid were separated by SDS/ PAGE in a 10% polyacrylamide gel. Arrowheads refer to the slowermoving protein in 419C and the faster-moving protein in 52A (refer to the text). Middle: A portion of a SDS/PAGE analysis that shows the 6K protein from 5 X 10’ infected cells labeled for 15 min with 200 &i of [3H]palmitic acid (60 Ci/mm) at 6 hr postinfection (m.o.i. = 50). Bottom: A portion of a SDS/PAGE analysis that shows immunoprecipitates obtained with anti 6K antibodies (Gaedigk-Nitschko and Schlesinger, 1990) and extracts of 1 O7 cells labeled for 2 hr at 6 hr postinfection (m.o.i. = 20) with 150 &i [35S]cysteine. In the two lower panels 5-2096 gradient polyacrylamide gels were used to separate proteins.

tions in the short region of the Sindbis virus genome that encodes the cytoplasmic carboxyl tail of the E2 glycoprotein and the 6K protein. A summary of the defects found in these mutants is in Table 5. All mutations affected virus particle release and virion structure. Only one mutation so severely impaired the structure of the

NORMAL

SITE:

SITE:

E2 -S.L.A.L.L.C.C.V.R.S.A.N.A.-/-E.F.T.E.T.M.S-

3958

399F

419C

PE2+E2’ El E2/El

2.0 1.0 2.0

1.3 1.1 1.2

1.3 0.8 1.6

E-(GP’$ 6K 6K/E

10.1 1.0 0.1

11.3 1.6 0.1

12.4 1.2 0.1

52A

1.6 0.8 2.0

1.8 0.9 2.0

0.5 0.4 1.3

0.9 0.3 3.0

13.0 1.2 0.1

13.7 0.3 /-my

7.4 0.7 0.1

7.3 1.1 0.2

after a 15.min label of after a 60.min

label of

6K

6K

SITE:

34A

virus protein that progeny virions could not be detected. Of the eight remaining, all showed surprisingly similar phenotypes even though they have specific individual defects. Three of the 6K mutants were altered in the cluster of cysteines near the middle of this very hydrophobic polypeptide. All of these were defective in palmitoylation, thus allowing us to definitively assign these cysteines as sites for fatty acylation. One of the E2 mutants had a cysteine replaced and this mutant showed deficient fatty acylation. This result supports our earlier prediction, based on the release of fatty acid by neutral hydroxylamine, that the sites of fatty acylation in E2 were cysteines (Magee et al., 1984). Two of the mutants, E2 419C and 6K 52A, contained amino acid substitutions close to sites of signal peptidase cleavage and both produced an altered pattern of glycoproteins. Based on immunoprecipitation with specific antibodies, the altered glycoprotein band in the 6K mutant 52A contained both 6K and El sequences. The

TABLE 4

-S.L.A.L.L.C.C.V.R.&A.N.A.E.F.T.E.T.M.S-

SITE:

35S,36A,38S,39A

a Figures are cpm X 10m3 measured [3H]palmitate. * Figures are cpm X 10m3 measured [3H]palmitate.

THERMAL

NORMAL

6K mutants

Wildtype

C-+

NEW

AND

E2 mutants

P62U

NEW

ON LEVELS

E2,

IN El,

INACTIVATION

El

E2 mutants

-C.C.L.P.F.L.V.V.A.G.A.Y.L.A.K.V.D.A.-/-E.H.A.T.T.-

-C.C.L.P.F.L.V.V.A.G.A.Y.L.A.-,-~.V.D.A.E.H.A.T.T-

FIG. 3. Predicted changes in the signalase cleavage site between E2 and 6K and between 6K and El as a result of the 419C and 52A mutations. The new sites are based on a computer program predicting signal peptidase processing sites. (Folz and Gordon, 1987; Von Heijne, 1983). The amino acid substitutions are noted by *, The probability of cleavage at the new sites is estimated to be about one-half that of the normal sites.

OF MUTANT

Time at 56” 0 10 min 20 min 30 min

VIRUSES’

6K mutants

Wild-type

3958

399F

419C

35S,36A,38S,39A

34A

52A

9.4 5.7 2.7

9.2 8.5 6.8 5.8

8.8 8.3 7.3 6.4

8.6 7.6 7.2 6.3

8.8 8.6 7.8 6.8

8.5 7.8 7.8 7.0

8.0 7.5 7.0 6.2

’ Values are plaque-forming embryo fibroblasts.

units: [log,,]. Virus was from chicken

SITE-DIRECTED

MUTATIONS

AFFECTING

SINDBIS VIRUS ASSEMBLY

211

FIG. 4. Electron micrographs of virus-infected chicken embryo fibroblasts. (A) E2 mutant 3958; (B) E2 mutant 419C; (C) E2 mutant 399F; (D) 6K mutant 52A. Refer to text and Gaedigk-Nitchko ef al. (1990) for experimental details. Arrows show multicored particles. The electron microscopy was performed by Marilyn Aach-Levy of the Department of Cell Biology and Physiology, Washington University School of Medicine.

altered band in the E2 mutant 419C reacted with E2 antibodies but not with 6K antibodies; however, the latter negative result could be due to a masking of the antigenic determinants in 6K as a result of the putative E2-6K polyprotein. The change in levels of virus glycoproteins and 6K protein that would result from these new cleavage sites might account for the slow release and altered form of virus. Aberrant interactions between heterodimers are postulated to explain the phenotype of the Sindbis virus temperature-sensitive mutant, ts 103. Like the mutants described here, this ts mutant produced multicored particles at the nonpermissive temperature and showed multihit thermal inactivation kinetics (Strauss et al., 1977); however, the amino acid substitution in this mutant occurs in the outer structure of E2 (Hahn eta/., 1989). Other temperature sensitive mutants that are defective in virus assembly have been described. One of these, noted ts20 (Burge and Pfefferkorn, 1970) contains a substitution in the extracytoplasmic domain of the E2 glycoprotein (Lindqvist et a/., 1986) and is defective in processing p62 at the nonpermissive temperature (Bracha and

Schlesinger, 1976). Cells infected with this mutant accumulate nucleocapsids at the plasma cell membrane (Brown and Smith, 1975); thus, they resemble to some extent the E2 399F mutant described here although it should be noted that the latter is not temperature sensitive in growth nor significantly altered in p62 processing. Six of the mutants (four here and the two previously reported) produced infectious virus like wild-type in avian cells but released virus particles at a slower rate than wild-type and the particles exhibited multihit thermal inactivation kinetics (not shown) indicative of multicored virions. Growth of these mutants was slower than that for wild type in insect cells. To account for these changes we consider the following scheme, noted earlier, for alphavirus assembly. Assembly and budding to form normal sing/e-cored lipid enveloped particles from the cell’s plasma membrane require severe bending of the lipid bilayer which is “driven” in part by protein-protein interactions between trimers of E2-El heterodimers and the nucleocapsid. We should note, however, that both Semliki

GAEDIGK-NITSCHKO

212

AND SCHLESINGER

TABLE 5 SUMMARY OF PROPERTIESOF E2 AND 6K MUTANTS Growth Mutant E2 395s 399F 415s 416A 419c 6K 34A 35s 36A 38s 39A I 52A 39s” 35S8 36A

CEF

c7-10

Normal Normal

Slow Slow

None

Particle

Kinetics of thermal inactivation

Others

Multicored Multicored

Multihit Multihit

Less Palmitate Excess membrane-bound

Not detected

Not determined

No progeny virus Altered E2-6K cleavage

Slow

Slow

Multicored

Multihit

Normal

Slow

Multicored

Multihit

Normal

Slow

Multicored

Multihit

Almost no palmitate

Slow Normal

Slow Slow

Not detected Multicored

Multihit Multihit

Altered 6K-El cleavage; Less palmitate

Normal

Slow

Multicored

Multihit

Less palmitate

a Data in Gaedigk-Nitschko

nucleocapsids

surface aggregate

et al. (1990).

Forest virus and Sindbis virus can be assembled with uncleaved E2 glycoproteins (Presley and Brown, 1989; Russell et a/., 1989; Garoff, 199 1). A specific configuration of the polypeptides in the cytoplasmic tails of the E2 trimer complex is recognized by a complementary structure on the surface of the nucleocapsid (Vaux et a/., 1988). Three of the E2 mutations described here are considered to affect this binding between nucleocapsid and glycoprotein. The group 1 mutant which lacked two cysteines may be so defective in the structure of the cytoplasmic domain that essentially no binding of nucleocapsid occurs. This could result if the putatively acylated cysteines near the carboxy terminus of E2 anchor the cytoplasmic domain in the membrane in a manner that constrains the structure of the polypeptide. Substitution of the cysteine at position 395 appeared to have removed a palmitoylation site that might anchor the amino terminus portion of the E2 cytoplasmic domain. The fatty acids might also affect lipid asymmetry by selective insertion into one leaflet of the bilayer. The tyrosine at 399 is in a hydrophobic string of seven to eight amino acids that are very highly conserved among alphaviruses (Rice and Strauss, 1981; Garoff et a/., 1980; Dalgarno et a/., 1983; Kinney et al,, 1986; Levinson et al., 1990; Hahn et a/., 1988). Replacement of this tyrosine by phenylalanine led to a mutant phenotype; thus, we suggest that the tyrosine in this sequence is critical for capsid-glycoprotein binding. The role of the 6K protein in assembly and budding is not yet clear. Earlier we showed that 6K protein is part

of the virion structure but there is only about one 6K protein per trimer (Gaeidgk-Nitschko and Schlesinger, 1990). Based on the model above, we postulate that 6K plays a role in “packing” of trimers and also contributes to the binding between E2 tails and nucleocapsid. The single lysine near the carboxy terminus of 6K can be cross-linked to the E2 protein of the virion with a membrane impermeable reagent (N. C. Collier and M. J. Schlesinger, unpublished data) suggesting that some of the trimers are noncovalently complexed with 6K protein. Previously (Gaedigk-Nitschko et al., 1990) we speculated that this very lipophilic protein, which has only 5 of its 52 amino acids as charged residues and contains 4 covalent fatty acids, facilitated lipid “flipping” that should accompany bending of the lipid bilayer during budding. Perhaps, lipid rearrangement is critical to trimer packing. One intriguing property of many of the mutants described here is their differential growth in mosquito cells as contrasted with avian cells, with more severe defects in insect cells. In this regard, their phenotype is exactly opposite to a temperature-sensitive mutant described by Durbin and Stollar (1984). This mutant grows normally in mosquito cells but forms small plaques in mammalian cells and is temperature sensitive for growth in the latter. At the nonpermissive temperature, nucleocapsids accumulate at the plasma membrane as in the ts20 mutant noted above, thus indicating a defect in viral assembly in mammalian cells. The mutation results from three changes in the extracytoplasmic domain of the E2 protein one of

SITE-DIRECTED

MUTATIONS

AFFECTING

which leads to hyperglycosylation (Durbin and Stellar, 1986). The differences between insect and vertebrate cells are considerable and even the growth of wild-typevirus is significantly slower in insect cells than in vertebrate cells incubated at the same temperature (i.e., 30”). One property that distinguishes these two cell types and is possibly important in the mutant phenotypes noted here is the variation in lipid composition (Luukkonen et al., 1977). The insect cell membrane is more “fluid” than the vertebrate cell membrane and this difference has been noted also for the lipid bilayer of viruses grown on the particular cell (Moore et al., 1976). We speculate that the probability of a successful nucleation event involving nucleocapsids, trimers of glycoprotein heterodimers and membrane associated 6K protein might be lower in a more fluid membrane than in a more viscous lipid bilayer. Thus, any mutation that affected the complex of trimers and 6K proteins and nucleocapsid binding would be exacerbated in insect cells. We hope to test this hypothesis by comparing virus growth in cells that can be altered in lipid and cholesterol content. In addition, complementation tests of the host range defects exhibited by these mutants may reveal additional kinds of protein-protein interactions important in virus assembly and budding.

ACKNOWLEDGMENTS We thank Cynthia Nettrourfor technical assistance. We are greatly appreciative of the assistance from Marilyn Aach-Levy for electron microscopy and C. Rice for infectious Sindbis virus cDNA. We thank Steve Nothwehr for computer analysis of the signalase cleavage sites. Dr. Gaedlgk-Nitschko was supported by a fellowship from the Boehnnger lngelheim Fonds. This study was supported by a grant from the U.S. Public Health Service (Al 19494).

REFERENCES ACHESON, N. H., and TAMM, I, (1967). Replication of Semliki Forest virus: An electron microscopic study. virology 32, 128-143. BIRDWELL,C. R., STRAUSS,E. G., and STRAUSS,J. H. (1973). Replication of Sindbis virus. Ill. An electron microscopic study of virus maturation using the surface replica technique. Virology 56, 429438. BRACHA, M., and SCHLESINGER,M. 1. (1976). Defects in RNA+ temperature-sensitive mutants of Sindbis virus and evidence for a complex of PE2-El viral glycoproteins. Virology 74, 441-449. BROWN, D. T., and SMITH, J. F. (1975). Morphology of BHK-21 cells infected with Sindbis virus temperature sensitive mutants in complementation groups D and E. /. Viral. 15, 1262-l 266. BROWN, D. T., WAITE, M. R. F., and PFEFFERKORN,E. R. (1972). Morphology and morphogenesis of Sindbis virus as seen with freezeetching techniques. /. Viral. 10, 524-536. DALGARNO, L., RICE, C. M., and STRAUSS,J. H. (1983). Ross rivervirus 26s RNA: Complete nucleotide sequence and deduced sequence of the encoded structural proteins. Virology 129, 170-l 87. DURBIN, R. K., and STOLLAR,V. (1984). A mutant of Sindbis virus with

SINDBIS VIRUS ASSEMBLY

213

a host-dependent defect in maturation associated with hyperglycosylation of E2. Virology 135, 331-344. DURBIN, R. K., and STOLLAR, V. (1986). Sequence analysis of the E2 gene of a hyperglycosylated, host restricted mutant of Sindbis virus and estimation of mutation rate from frequency of revertants. Virology 154, 135-143. FOLZ, R. J., and GORDON, J. I. (1987). Computer-assisted predictions of signal peptidase processing sites. Biochem. Biophys. Res. Commun. 146, 870-877. FULLER,S. D. (1987). The T=4 envelope of Sindbis virus is organized by interactions with a complementary T=3 capsid. Cell 48, 923934. GAEDIGK-NITSCHKO,K., DING, M., and SCHLESINGER,M. J. (1990). Sitedirected mutations in the Sindbis virus 6K protein reveal sites for fatty acylation and the underacylated protein affects virus release and virion structure. Virology 175, 282-29 1. GAEDIGK-NITSCHKO,K., and SCHLESINGER,M. 1. (1990). The Sindbis virus 6K protein can be detected in virions and is acylated with fatty acid. Virology 175, 274-281. GAROFF, H., and SIMONS, K. (1974). Location of the spike glycoprotein in the Semliki Forest virus membrane. Proc. Nat/. Acad. Sci. USA 71, 3988-3992. GAROFF. H., FRISCHAUF,A.-M., SIMON% K., LEHRACH, H., and DELIUS, H. (1980). Nucleotide sequence of cDNA coding for Semliki Forest virus membrane glycoproteins. Nature 288, 236-241. HAHN, C. S., RICE, C. M., STRAUSS, E. G., LENCHES, E. M., and STRAUSS,J. H. (1989). Sindbis virus ts 103 has a mutation in glycoprotein E2 that leads to defective assembly of virions. 1. Viral. 63, 3459-3465. HAHN, C. S.. LUSTIG, S., STRAUSS, E. G., and STRAUSS.J. H. (1988). Western equine encephalitis virus IS a recombinant virus. Proc. Natl. Acad. Sci. USA 85, 5997-6001. HELENIUS, A., and KARTENBECK,1. (1980). The effects of octylglucoside on the Semliki Forest virus membrane: evidence for a spikeprotein-nucleocapsid interaction. Eur. J. Biochem. 106, 613-618. KELLEY, P. M.. and SCHLESINGER,M. J. (1982). Antibodies to two major chicken heat shock proteins crossreact with similar proteins in widely divergent species. Mol. Cell. Biol. 2, 267-274. KINNEY, R. M., JOHNSON,B. J., BROWN, D. W., and TRENT. D. W. (1986). Nucleotide sequence of the 26s RNA of the virulent Trinidad donkey strain of Venezuelan equine encephalitis virus and deduced sequences of the encoded structural proteins. Virology 152,400413. LEVINSON,R. S., STRAUSS,J. H., and STRAUSS,E. G. (1990). Complete sequence of the genomic RNA of O’nyong-nyong virus and its use in the construction of alphavirus phylogenetic trees. Virology 175, 110-123. LEVIS, R.. WEISS, B. G., TSIANG, M., HUANG, H., and SCHLESINGER,S. (1986). Deletion mapping of Sindbis virus DI RNAs derived from cDNAs defines the sequences essential for replication and packaging. Cell 44, 137-l 45. LIUESTROM. P., and GAROFF, H. (1991). Internally located cleavable signal sequences direct the formation of Semliki Forest virus membrane proteins from a polyprotein precursor. /. Viral. 65, 147154. LINDQVIST, B. H., DISALVO, J., RICE, C. M., STRAUSS, J. H., and STRAUSS,E. G. (1986). Sindbis virus mutant ts20 of complementation group E contains a lesion in glycoprotein E2. Virology 151, 1O-20. LUUKKONEN,A., VON BONSDORFF, C.-H., and RENKONEN, 0. (1977). Characterization of Semliki Forest virus grown in mosquito Cells. Comparison with the virus from hamster cells. virology 78, 331335. MAGEE, A. I., KOYAMA, A. H., MALFER, C., WEN, D., and SCHLESINGER,

214

GAEDIGK-NITSCHKO

M. J. (1984). Release of fatty acids from virus glycoproteins by hydroxylamine. Biochim. Biophys. Acta 798, 156-l 66. MELANCON. P., and GAROFF, H. (1986). Reinitiation of translocation in the Semliki Forest virus structural polyprotein: Identification of the signal for the El glycoproteln. EMBO J. 5, 1551-l 560. METSIKKO, K., and GAROFF, H. (1990). Oligomers of the cytoplasmic domain of the p62/E2 membrane protein of Semliki Forest virus bind to the nucleocapsid in vitro. J. Viral. 64, 4678-4683. MOORE, N. F., BARENHOLZ,Y., and WAGNER, R. R. (1976). MicroviscosIty of togavirus membranes studied by fluorescence depolarization: influence of envelope proteins and the host cell. J. Viral. 19, 126-135. PRESLEY,J. F., and BROWN, D. T. (1989). The proteolytic cleavage of PE2 to envelope glycoprotein E2 is not strictly required for the maturation of Sindbis virus. J. Viral. 63, 1975-l 980. RICE, C. M., LEVIS, R., STRAUSS,J. H., and HUANG, H. V. (1987). Production of infectious RNA transcripts from Sindbis virus cDNA clones: Mapplng of lethal mutations, rescue of a temperature-sensitive marker, and in vitro mutagenesis to generate defined mutants. 1. Viral. 61, 3809-3819. RUSSELL,D. L., DALRYMPLE,J. M., and JOHNSON, R. E. (1989). Sindbis virus mutations which coordinately affect glycoprotein processing, penetration and virulence in mice. J. Vkol. 63, 1619-l 629. SANGER. F., NICKLEN, S., and COULSON, A. R. (1977). DNA sequencing with chain-terminating inhibitors. froc. Nat/. Acad. Sci. USA 81,2050-2054. SCHLESINGER,S., and SCHLESINGER,M. J. (1972). FormatIon of Sindbls virus proteins: Identification of precursor of one of the envelope proteins. J. Viral. 10, 925-932. SIMONS, K.. and GAROFF,H. (1980). The budding mechanism of enveloped animal viruses. J. Gen. Viral. 50, l-21.

AND SCHLESINGER SMITH, J. F., and BROWN, D. T. (1977). Envelopment of Sindbis virus: Synthesis and organization of proteins in cells infected with wild type and maturation-defective mutants. J. Viral. 22, 662-678. STRAUSS,E. G.. BIRDWELL,C. R., LENCHES, E. M. STAPLES,S. E., and STRAUSS.J. H. (1977). Mutants of Sindbis virus. Il. Characterization of a maturation-defective mutant, ts 103. Virology 82, 122-l 49. STRAUSS. E. G., LENCHES, E. M., and STAMREICH-MARTIN, M. A. (1980). Growth and release of several alphaviruses in chick and BHK cells. J. Gen. Viral. 49, 297-307. STRAUSS,E. G., and STRAUSS,1. H. (1986). Structure and replication of the alphavirus genome. In “The Togaviridae and Flaviviridae” (S. Schlesinger and M. J. Schlesinger, Eds.), pp. 35-90. Plenum, New York. TORRISI,M. R., and BONAITI, S. (1985). lmmunocytochemical study of the partition and distribution of Sindbis virus glycoproteins in freeze-fractured membranes of infected baby hamster kidney cells. J. CeIBiol. 101, 1300-1306. VAUX, D. J. T., HELENIUS,A., and MELLMAN I. (1988). Spike-nucleocapsid interaction in Semliki Forest virus reconstructed using network antibodies. Nature 336, 36-42. VON HEIJNE.G. (1986). A new method for predicting signal cleavage sites. Nucleic Acids Res. 14, 4683-4690. WAHLBERG,J., BOERE,W. A. M., and GAROFF, H. (1989). The heterodimerit association between the membrane proteins of Semliki Forest virus changes its sensitivity to low pH during virus maturation. J. Viral. 63,4991-4997. WEISS, B., NITSCHKO, H., GHA~AS, I., WRIGHT, R.. and SCHLESINGER,S. (1989). Evidence for specificity in the encapsidation of Sindbis virus RNA. /. Viral. 63, 53 1O-53 18. ZIEMIECKI. A., and GAROFF, H. (1978). Subunit composition of the membrane glycoprotein complex of Semliki Forest virus. J. Mol. Biol. 122, 259-269.