key: cord-0724676-qbmcw4ck
authors: Chen, Si-Yi; Matsuoka, Yumiko; Compans, Richard W.
title: Golgi complex localization of the Punta Toro virus G2 protein requires its association with the G1 protein
date: 1991-07-31
journal: Virology
DOI: 10.1016/0042-6822(91)90148-5
sha: 40b271768b2444fab4c42ce4866dc88215ac43c9
doc_id: 724676
cord_uid: qbmcw4ck

Abstract The glycoproteins of bunyaviruses accumulate in membranes of the Golgi complex, where virus maturation occurs by budding. In this study we have constructed a series of full length or truncated mutants of the G2 glycoprotein of Punta Toro virus (PTV), a member of the Phlebovirus genus of the Bunyaviridae, and investigated their transport properties. The results indicate that the hydrophobic domain preceding the G2 glycoprotein can function as a translocational signal peptide, and that the hydrophobic domain near the C-terminus serves as a membrane anchor. A G2 glycoprotein construct with an extra hydrophobic sequence derived from the N-terminal NSM region was stably retained in the ER, and was unable to be transported to the Golgi complex. The full-length G2 glycoprotein, when expressed on its own, was transported out of the ER and expressed on the cell surface, whereas the G1 and G2 proteins when expressed together are retained in the Golgi complex. A truncated anchor-minus form of the G2 glycoprotein was found to be secreted into the culture medium, but was retained in the Golgi complex when coexpressed with the G1 glycoprotein. These results indicate that the G2 membrane glycoprotein is a class I membrane protein which does not contain a signal sufficient for Golgi retention, and suggest that its Golgi localization is a result of association with the G1 glycoprotein.

INTRODUCTION sequence of a trans Golgi resident membrane protein, P-galactoside a-2,6-sialyltransferase, was found to be involved in Golgi localization . Portions of the sequences of several ER resident membrane proteins were also recently shown to be required for their retention (Jackson et a/., 1990; Kuroki et a/., 1989; Nilsson et al,, 1989; Poruchynsky et al., 1985; Stirzaker and Both, 1989) . However, the precise structure features of the retention signals and the mechanism for the intracellular retention of resident membrane proteins of the ER and Golgi complex are still unknown.

A central problem in cell biology is to determine how membrane proteins are sorted and targeted to various subcellular compartments. Recent evidence supports the view that selective transport of proteins from the endoplasmic reticulum (ER) to the cell surface in the central vacuolar system could simply occur by bulk flow: according to this model, after folding and assembly of secretory and membrane proteins into their native 3-dimensional structure, they are transported out of the ER to the plasma membrane, and only proteins which have specific retention signals are retained in intracellular compartments (Klausner, 1989; Pelham, 1988; Rose and Doms, 1988; Rothman, 1987; Wieland et a/., 1987) . Assembly of enveloped viruses of several families, e.g., coronaviruses, bunyaviruses, and flaviviruses, takes place by budding at intracellular membranes, and the corresponding viral glycoproteins accumulate intracellularly rather than being transported to the cell surface (Leary and Blair, 1980; Matsuoka et a/., 1988; Murphy et a/., 1973; Tooze and Tooze, 1985) . Machamer and Rose (1987) found that the first of three hydrophobic membrane-spanning domains of the coronavirus El glycoprotein, a Golgi membrane protein, is required for its retention. The NH,-terminal ' Present address: Secretech, Inc., 1025 18th Street South, Birmingham, AL 35205.

* To whom reprint requests should be addressed.

The glycoproteins of bunyaviruses are model Golgi membrane proteins, since the viruses are formed by budding at the smooth membranes of the Golgi complex (Murphy et a/., 1973; Smith and Pifat, 1982) and the viral glycoproteins accumulate in the Golgi complex during virus infection (Chen et al., 1991; Gahmberg et a/., 1986; Kuismanen et a/., 1982 Kuismanen et a/., , 1984 Madoff and Lenard, 1982) as well as when expressed by recombinant vaccinia viruses (Matsuoka et al., 1988; Pensiero et a/., 1988; Pettersson et al,, 1988; Wasmoen et a/., 1988) . Punta Toro virus (PTV) is a member of the phlebovirus genus of the family Bunyaviridae (Bishop et al., 1980) . Its structural components include three single-stranded RNA genomic segments, L, M, and S, and three major structural proteins, a nucleoprotein, N, and two major membrane glycoproteins, Gl and G2. These glycoproteins are encoded by the M genomic segment and are translated from a single mRNA as a precursor glycoprotein and cleaved cotranslationally into individual proteins (Eshita et al., 1985; Schmaljohn and Patterson, 1990; Suzich and Collett, 1988) .

The gene product order, NH,-NS, (nonstructural protein)-Gl-G2-COOH, of the PTV M RNA segment has been determined by N-terminal amino acid sequence analysis (Ihara et al., 1985) . One potential Nlinked glycosylation site (Asn-X-Sernhr) in the Gl protein and four in the G2 protein were found in the sequence (Ihara et a/,, 1985) . The predicted amino acid sequence of the polyprotein precursor of PTV reveals that distinct hydrophobic regions precede both the Gl and G2 proteins. There are also hydrophobic domains near the carboxy termini of the Gl and G2 proteins, followed by charged amino acids, similar to the "stop transfer" sequences seen in the transmembrane domains of other viral envelope proteins (Garoff et al., 1980; lhara et al., 1985; Rose et al., 1980) . This suggests that each of the two glycoproteins encoded in the polyprotein precursor may have its own signal sequence and membrane anchor sequence, analogous to the internal signal sequence and membrane anchor sequences described for the El protein of the alphaviruses, Semliki forest virus, and Sindbis virus (Garoff et al., 1980; Hashimoto et a/., 1981; Rice and Strauss, 1981) . However, virtually no direct evidence is available about the sequences responsible for targeting of the nascent PTV proteins to the ER membrane or the domains responsible for membrane anchorage. The PTV G 1 and G2 glycoproteins, when expressed from a recombinant vaccinia virus, are specifically retained in the Golgi complex (Matsuoka et al., 1988; Chen et al., 1991) . The NS, sequence was ruled out as a retention signal for the glycoproteins, indicating that the structural features of the Gl and/or G2 glycoproteins are responsible for their Golgi retention.

To investigate the molecular basis for ER translocation, membrane anchorage and Golgi retention of the PTV G2 glycoprotein, we constructed a series of fulllength and truncated G2 glycoprotein mutants. In the present paper, we have investigated the transport properties of the resulting molecules using vaccinia virusbased expression systems.

BHK, CV-1, HeLa, HeLa T4+, TK-143, and Vero cells were maintained in Dulbecco's modified Eagle's medium (DMEM) supplemented with lOq/o newborn bovine serum (in the presence of BUdR for TK-143 cells). A recombinant pGEM-G containing the PTV Gl and G2 coding sequence under control of a T7 promoter was described previously (Matsuoka et al., 1988) . PTV was obtained from USAMRIID (Frederick, MD) and passaged in Vero cells. Polyclonal and monoclonal antibodies against PTV Gl and G2 glycoproteins were generously supplied by Drs. J. F. Smith and D. Pifat (USAMRIID, Frederick, MD) . The monoclonal antibody to immunoglobulin heavy chain binding protein (BiP) (Bole et al., 1986) 

All DNA manipulation was carried out as described by Maniatis et a/. (1986) . To construct the G2 recombinant designated as pGEM-G2(SH) (Fig. l) , the pGEM-M plasmid (Matsuoka et al., 1988) was linearized by digestion with Stul and partially digested with Hind Il. A 5.8-kb fragment was recovered by gel-purification, treated with Klenow DNA polymerase, and then religated. The G2 insert of the resulting recombinant was then cut by Pstl and EcoRI, recovered by gel purification, and finally subcloned into Pstl and EcoRI-cut pGEM-4Z under control of the T7 promoter. To construct the G2 recombinant designated as pGEM-G2(HH), pGEM-M was partially digested with Hincll, and a 5.6-kb fragment was gel-purified and religated. To construct the Gl -truncated G2 recombinant designated as pGEM-G(A-), the plasmid pGEM-G (Matsuoka et a/., 1988) was cut at an Xbal site located just before the 19 amino acid hydrophobic sequence which is the presumed membrane anchor, in addition to two Xbal sites in the Gl sequence and polylinker of the vector. Two DNA fragments were recovered from a low-melting agarose gel, purified with a Geneclean kit, and then religated. The recombinant which lacked the coding region (45 amino acids), for the G2 hydrophobic 19 amino acids and C-terminal sequences was then cut at a S&l site next to an Xbal site in the polylinker. insert DNA was located downstream of the T7 promoter sequence in plasmid pGEM-3.

To construct the G2 recombinant designated as pGEM-G2, an upstream primer (P-l ,5'-GGGAATTCA-TGAGAAGATTCAAAACAACT-3') containing an EcoRl site followed by the sequence of nucleotides 2387 to 2419 of the PTV M genomic RNA, and a downstream primer (5'-CACTGCAGTCAGTlIXlTClTGATAlT-3'), containing the sequence of nucleotides 3941 to 3961 followed by a Pstl site, were used to amplify the G2 DNA fragment by the polymerase chain reaction (PCR). The amplified DNA fragment containing an internal ATG, a putative signal peptide, and G2 coding sequence was then cut by EcoRl and Pstl, purified by a Geneclean kit, and ligated into /FcoRI and Pstl-cut pGEM-3Z. To construct the truncated G2 recombinant designated as pGEM-G2(A-), the upstream primer P-l, and a downstream primer (5'-CAAAGCTTACTTTAGT-ATGGCCTTCATAGG-3') containing the sequence of nucleotides 3839 to 3860 of the PTV M RNA followed by a stop Codon TAA, were used to amplify the DNA fragment by PCR. The amplified DNA fragment was cut by EcoRI, and then ligated into EcoRI and Smal-cut pGEM-3Z. The nucleotide sequence of the PCR-amplified DNA products in pGEM-3Z was analyzed by restriction endonucleases and confirmed by Sanger's di-deoxynucleotide chain termination method (Sanger et al., 1977) using a sequenase DNA sequence kit (United States Biochemical Corp., Cleveland, OH). The SP6 promoter and T7 promoter primers (Promega Biotech., Madison, WI) were used for sequencing by the dideoxy method.

Selected DNAfragments of the PTV M genomic segment were amplified by PCR with the plasmid pGEM-G as template DNA (Saiki et a/., 1988) in a DNA thermal cycler(Perkin-Elmer-Cetus, Norwalk, CT) using thefollowing conditions: upstream and downstream primers were 30 pM each in final concentration, dNTP were 100 &I each, template plasmid DNA was 100 ng, Taq DNA polymerase was 2.5 units, other reaction conditions were used according to the GeneAmp DNA amplification kit instructions in a total volume of 100 ~1. The reaction mixtures were overlayed with 75 ~1 of mineral oil, and the reactions were carried out for 25 cycles. Each cycle included a heat denaturation step at 92" for 1 min, followed by annealing of primers to the DNA template at 50" for 1 min, and DNA chain extension with Taq polymerase at 72" for 2 min. After the reaction was completed, the PCR products were purified with a Geneclean kit. The recombinant pGEM-G2(HH) plasmid DNA was linearized and RNA was synthesized using SP6 polymerase. The SP6-G2 transcripts or control samples without RNA transcripts were translated in rabbit reticulocyte lysates in the presence of [35S]methionine. The protein products were immunoprecipitated with polyclonal antibody to PTV and analyzed by SDS-PAGE. The molecular weight markers were pyruvate kinase, 58 kDa; fumarase, 48.5 kDa; lactic dehydrogenase, 36.5 kDa; and triosephosphate isomerase, 26.6 kDa.

To construct the vaccinia recombinant viruses designated as W-G2(SH) and VV-G2(HH) (Fig. l) , the G2 glycoprotein coding sequence was excised from pGEM-G2 (SH) by Smal digestion or from pGEM-G2(HH) by Smal-Hincll digestion, ends filled in with the Klenow DNA polymerase, inserted into the Smal site of pSC1 1, and then checked for orientation (Chakrabarti et al., 1985) . For isolation of vaccinia recombinants, CV-1 cells were infected with vaccinia virus (strain IHD-J) at an m.o.i. of 0.05. At 2 hr p.i., cells were transfected with a calcium phosphate precipitate of 10 pg of a pSCl1 recombinant plasmid and 15 pg of salmon sperm DNA/ml in HEPES-buffered saline (Graham and Van der Eb, 1973) . To select recombinant% TK-143 cells were infected with 50-100 PFU of virus in the presence of BUdR (25 pg/ml). At 48 p.i. the monolayers were overlayered with lo/o low-melting agarose containing 300 pg/ml of 5-bromo-4-chloro-3-indolyl-p-ogalactopyranosidase (X-Gal), and blue plaques were picked and purified by several rounds of plaque purification. Stocks of recombinant viruses were prepared in CV-1 cells, and titrated in TK-143 cells. For vaccinia recombinant virus expression, HeLa T4+ cells, which are relatively resistant to cytopathic effects of vaccinia virus (Matsuoka eta/., 1988) , were infected with recombinant vaccinia virus with an m.o.i. of 10 for 30 min and incubated for 16 hr before radiolabeling and immunoprecipitation.

T7 polymerase transient expression A recombinant vaccinia virus containing the T7 RNA polymerase gene (VV-T7) was obtained from Dr. B. Moss (Fuerst er al., 1986) . HeLa T4+ cells were infected with W-T7 at an m.o.i. of 20 for 30 min and washed with PBS three times, and then 10 pg of recombinant plasmid DNA mixed with 50 ~1 of water and 50 ~1 of lipofectin was added into the serum-free DMEM medium of the infected cells and incubation continued for 12 hr before radiolabeling and immunoprecipitation.

Recombinant plasmid DNA under control of the SP6 promoter (4.0 pg) was linearized by Kpnl digestion, and was then incubated at 37" for 2 hr in a reaction mixture containing 100 mM dithiothreitol, 2.5 mM ATP, 2.5 mM GTP, 2.5 mNI UTP, 2.5 ml\/l CTP, 1 unit RNasin (Promega Biotech., Madison, WI), and 10 units SP6 polymerase (Promega Biotech., Madison, WI) in a final volume of 50 ~1. The reaction was terminated by the addition of 1 unit of RQ 1 DNase (Promega Biotech., Madison, WI) at 37" for 10 min. The mixture was extracted with phenol/chloroform, before being precipitated with ethanol. RNA transcripts (1 pg) were incubated at 30" for 90 min with 17.5 ~1 of rabbit reticulocyte lysate (Promega Biotech., Madison, WI), 0.5 rnM of each amino acid except for methionine, and 25 &i [35S]methionine/cysteine in final volume of 25 ~1. The translation product was immunoprecipitated and analyzed by SDS-PAGE.

Transfected or infected cells were washed once with phosphate-buffered saline (PBS) at indicated time points postinfection or transfection and incubated in methionine-free medium for 3 hr. Cells were then labeled with [35S]methionine/cysteine (100 &i/ml) in methionine-free medium for 15 min and chased in Eagle's medium containing 10 mM methionine for indicated periods. For surface immunoprecipitation, these cells were incubated with polyclonal antibody against PTV diluted in PBS for 60 min in 4", washed three times in ice-cold PBS, and lysed with 0.3 ml of cell lysis buffer (50 mM Tris-HCI (pH 7.5), 0.15 M NaCI, 1 YO Triton X-l 00, 0.1 Oh SDS, 20 mNI EDTA). Nuclei were removed by centrifugation at 13,000 g for 5 min at 4". For immunoprecipitation of intracellular proteins, cells were lysed, and monoclonal or polyclonal antibodies were then added to supernatants of the cell lysates and incubated at 37" for 90 min or at 4' overnight. All samples were incubated with Protein A-Sepharose CL-4B at 37" for 90 min or at 4" overnight (Pharmacia Inc. Piscataway, NJ). The precipitates were pelleted, washed three times with cold lysis buffer, resuspended in Laemmli sample buffer with or without @-mercaptoethanol and analyzed by SDS-PAGE (Laemmli, 1970) .

Endo H digestion lmmunoprecipitated samples were resuspended in 200 ~1 of 0.1 M sodium acetate, pH 5.5, and then divided into two equal aliquots. Samples were incubated with 8 mU of endo H/ml or without endo H for 16 hr at 37", and then centrifuged at 15,OOOg for 5 min (Kuismanen, 1984) . The precipitates were resuspended in Laemmli sample buffer, boiled for 5 min, and then analyzed by SDS-PAGE.

HeLa T4+ cells grown on glass coverslips were infected with vaccinia virus at an m.o.i. of 10, or VV-T7 at an m.o.i. of 20 following transfection with recombinant plasmid DNA. At indicated times post-infection, cells were washed with PBS and fixed with ethanol containing 5% acetic acid for 10 min at -20" for intracellular immunofluorescence or with 1% formaldehyde for 5 min at room temperature for surface immunofluorescence. Cells were then washed with PBS and reacted with monoclonal or polyclonal antibodies at 37" for 30 min followed by a fluorescein-conjugated goat antimouse IgG incubation. After a final washing, coverslips were mounted and observed by a Nikon Optiphot microscope equipped with a modified B2 tube.

Localization of Pn/ membrane glycoproteins expressed using a T7 transient expression system

We used a vaccinia T7 transient expression system to compare PTV glycoprotein localization with that observed with a recombinant vaccinia virus in several cell lines. This transient expression system is based on infection of the cells with a recombinant vaccinia virus expressing bacteriophage T7 RNA polymerase and subsequent transfection with the plasmid DNA containing the gene to be expressed under control of the T7 promoter. Transcription of the gene is mediated by the T7 polymerase produced in the cytoplasm by the recombinant vaccinia virus (Fuerst et a/., 1986) . We used a modification of the system by cloning the genes of interest into pGEM-3, pGEM-3Z, or pGEM-4Z plasmids under the control of a T7 promoter, but lacking a transcriptional termination sequence (Pattnaik and Wertz, 1990) . The pGEM-G plasmid, which contains the PTV Gl and G2 coding sequence under control of the T7 promoter, was initially used to investigate the expression and localization of the glycoproteins (Fig.  1) . It was found that the Gl and G2 glycoproteins were expressed, cleaved, and glycosylated in the transfected cells as detected by immunoprecipitation and SDS-PAGE (not shown), indicating that expression and processing of the Gl and G2 proteins occurred efficiently using the T7 expression system. HeLa T4+ cells retained relatively normal morphology for over 24 hr post-transfection while expressing high levels of the Gl and G2 glycoproteins. The localization of the Gl and G2 proteins was found to be in a perinuclear Golgi pattern, as observed previously in vaccinia recombinant W-G-infected cells (Fig. 2) . Furthermore, a similar pattern of Golgi localization of the proteins was also detected in the Vero, HeLa, or BHK cell lines when the proteins were expressed by the T7 transient system at early time points post-transfection. Thus, the Golgi targeting of PTV glycoproteins occurs using various cell types and expression systems, and either the T7 transient expression system or the vaccinia recombinant system can be used to study the targeting of glycoprotein constructs.

To determine whether the hydrophobic domain preceding the G2 protein is required as a translocational signal peptide, the recombinant pGEM-G2(HH), which lacks the coding sequence for six hydrophobic amino acids in the hydrophobic sequence preceding the G2 protein, was constructed (Fig. 1) . To identify the products produced in vitro, pGEM-G2(HH) plasmid DNA was analyzed by in vitro transcription and translation followed by immune precipitation. A translation product of molecular weight around 52 kDa was found in the pGEM-G2(HH) plasmid translation samples (Fig. 3) . To characterize the expressed protein product in viva, a recombinant vaccinia virus designated as VV-G2(HH) was constructed (Fig. 1) . HeLa T4+ and CV-1 cells were infected with the VV-G2(HH) recombinant and newly synthesized proteins were immunoprecipitated with antibody against PTV. However, no detectable G2 specific protein was found in the infected cell lysates (data not shown). The lack of a detectable protein product may reflect lack of translocation into the ER. In contrast, when the hydrophobic domain preceding G2 was preserved, the G2 protein was detected in viva (see below), indicating an essential role of this hydrophobic sequence for protein expression.

In order to express the individual G2 protein in vivo, a recombinant pGEM-G2(SH), which contains an extra sequence from the NS, region including a possible signal peptide sequence, was initially constructed us-ing convenient restriction sites, and a corresponding recombinant vaccinia virus VV-G2(SH) was obtained (Fig. 1) . To identify the expressed proteins, HeLa T4+ cells were infected with the recombinant W-G2(SH), and radiolabeled proteins were immunoprecipitated with antibody against PTV. A distinct polypeptide band, migrating slightly slower than native G2, was observed in the cell lysates of W-G2 (SH)-infected cells. Transport of the G2 protein was determined by surface and intracellular immunoprecipitation and immunofluorescence staining. In VV-G2(SH)-infected cells, the G2 protein was exclusively detected intracellularly, not on the cell surface, and it appeared to accumulate without degradation when analyzed by immunoprecipitation and SDS-PAGE (Fig. 4A) . Consistent with this, the G2 protein was also only detected in the cytoplasm by indirect immunofluorescence staining (Figs. 5c and 5d). A perinuclear Golgi pattern as seen in VV-G-infected cells was not observed in the infected cells even after a 4-hr cycloheximide treatment, and the distribution pattern of the protein was similar to that of BiP, a resident ER protein (Figs. 5a-5c), indicating that the protein product is accumulating in the ER. Similar results were obtained using a recombinant pGEM-G2(SH) and the VV-T7 transient expression system (not shown).

Endo H treatment, which cleaves high-mannose Nlinked oligosaccharides, was used to monitor whether the proteins are transported out of the ER and reach the medial Golgi complex, where processing reactions convert the high mannose forms to the endo H-resistant complex form of oligosaccharides (Kornfeld and Kornfeld, 1985) . The G2 proteins expressed by W-G2(SH) remained sensitive to endo H treatment after a 4-hr chase. In the cells treated with Brefeldin A (BFA), which causes the Golgi enzymes to redistribute to the ER (Lippincott-Schwartz et al., 1989) , a small fraction of the protein became endo H resistant (Fig. 6A) , indicating that glycosylation sites of the protein are accessible to glycosyltransferases under this condition. Thus, the G2 protein with an extra NS, sequence may have a structural defect which blocks its transport out of the ER, possibly due to the presence of extra amino acids.

Transport to cell surface of the full-length G2 protein

To investigate the possibility that the transport defect of the G2 protein with an extra N-terminal sequence is due to a structural defect, another construct which only contains the thirteen hydrophobic amino acids preceding the G2 protein and the complete G2 protein sequence was constructed using PCR (Fig. 1) . To detect protein expression, HeLa T4+ cells infected with VV-T7 were transfected with this construct desig- FIG. 9 . Localization of the glycoproteins expressed by pGEM-G(A-). HeLa T4+ cells grown on coverslips were infected with VV-T7 for 30 min, and then transfected with the plasmid DNA. After incubating for 12 hr, the cells were incubated with cycloheximide (50 fig/ml) at 37" for4 hr prior nated pGEM-G2. A polypeptide band around 55 kD, indistinguishable from native G2 in PTV-infected cells, was found in the cell lysates of pGEM-G2 transfected cells (Fig. 48) . To determine the transport properties of the G2 protein, surface and intracellular immunoprecipitation and immunofluorescence staining were used. In contrast to W-G2(SH)-infected cells, a fraction of the G2 protein was detected on the surface of pGEM-G2transfected cells by immunoprecipitation and SDS-PAGE after a 120-min chase, and the fraction of the G2 proteins expressed on the cell surface gradually increased as the chase continued (Fig. 4B) . The proportion of the surface proteins expressed by pGEM-G2 was quantitated by densitometric scanning of the autoradiograms, and the results indicate that the T,,2 for transport of the molecules to the cell surface was about 6 hr (Fig. 4B) . Furthermore, the surface G2 protein was detected by staining nonpermeabilized cells with MAbs to the G2 protein (Figs. 5e and 5f). The G2 proteins sometime appeared as a doublet, as seen previously when expressed by VV-G or a Hantaan virus M gene recombinant, which was believed to be due to heterogeneous glycosylation of the proteins (Matsuoka eta/., 1988; Pensiero et al., 1988) . Furthermore, the G2 protein expressed by pGEM-G2 slowly became partially endo H resistant during longer chases (Fig.  6B) . Thus, our results indicate that the full-length G2 protein expressed by the pGEM-G2 recombinant is slowly transported out of the ER and expressed on the cell surface.

To determine the requirement of the hydrophobic domain near the G2 C-terminus for membrane anchorage, and its effect on localization of the glycoprotein, a truncated G2 protein was constructed in which the hydrophobic and C-terminal domains were deleted (Fig.  1) . The truncated G2 protein contains 25 cysteine residues from a total of 26 residues in the G2 protein (lacking one in the hydrophobic domain) and all four potential glycosylation sites, and was thus expected to maintain a conformation similar to the native N-external domain. HeLa T4+ cells infected with W-T7 were transfected with the recombinant pGEM-G2(A-), pulselabeled, and chased for indicated times. A protein band with an estimated mol. wt. of 53 kDa, consistent with the truncated G2 protein (32 amino acids were deleted), was found in the cell lysates (Fig. 7) . The truncated protein was detected in the culture medium after a 120-min chase and the fraction of the protein in the medium increased with time of chase (Fig. 7) . The mobility of the protein in the medium was slightly slower than that in the cell lysates, suggesting that the protein was terminally glycosylated before secretion into the medium. The T,,2 for secretion of the molecules was about 4 hr as analyzed by densitometry scanning of the autoradiograms. Thus, the secretion of the truncated G2 protein into the medium indicates that the hydrophobic domain near the C-terminus serves as a membrane anchor, and the glycosylated N-external domain does not contain signals for intracellular retention.

Expression and localization of a truncated Gl-G2 glycoprotein

To determine whether interaction between the Gl and G2 glycoprotein occurs and affects G2 protein localization, we constructed a recombinant pGEM-G(A-), containing the intact Gl sequence and the truncated G2 sequence in which the hydrophobic and C-terminal domains of the G2 protein (45 amino acids) were deleted (Fig. 1) . To identify the protein products, HeLa T4+ cells infected with VV-T7 were transfected with the recombinant plasmid DNA, labeled for 15 min, and then chased for the indicated times. As shown in Fig.  8 , a 63 kDa molecular weight band corresponding to native Gl, and a 52-kDa molecular weight band smaller than the native G2 protein, were found in the cell lysates by SDS-PAGE. No PTV-specific protein was found on the cell surface or in the medium for up to a 12-hr chase. Endo H treatment showed that a fraction of the Gl-truncated G2 protein was incompletely processed to endo H-resistant forms as the chase continued, suggesting that these molecules reached the Golgi complex (Fig. 6C ). The localization of the truncated Gl -G2 proteins was further examined by immunofluorescence staining. No specific fluorescence was detected on the cell surface (Figs. 9e and 9f) . In permeabilized cells, the intracellular localization of both the Gl and truncated G2 proteins appeared predominantly in a perinuclear Golgi pattern after the cell culture was treated with cycloheximide for 4 hr (Figs. 9a and 9b). The Golgi localization of the glycoproteins was further indicated by BFA treatment, which caused the perinuclear distribution pattern of the Gl and truncated G2 proteins to change into a dispersed ER-like pattern (Figs. 9c and 9d), as seen for resident Golgi proteins in response to BFA treatment (Lippincott-Schwartz et a/., 1989) . Thus, in contrast to the secreto fixation, then BFA (10 pg/ml) was added to the culture medium for 20 min. After fixation, cells were stained with Mabs against Gl or G2, followed with fluorescein-labeled goat anti-mouse IgG. a. Gl glycoprotein after cycloheximide treatment. b. Truncated G2 glycoprotein after cycloheximide treatment. c. Gl glycoprotein in cells after BFA treatment. tion into the medium of the truncated G2 protein when expressed on it own, the Gl and truncated G2 glycoproteins are transported out of the ER and retained in the Golgi complex when expressed together using the pGEM-G(A-) recombinant. This result indicates that the Golgi retention of the truncated G2 glycoprotein occurs because of its association with the Gl glycoprotein.

Topogenic sequences for the G2 glycoprotein

We investigated the signal sequence for the second glycoprotein encoded in a bunyavirus polyprotein precursor (G2 in PTV) by analysis of two different G2 constructs. One construct (G2(HH)), containing a partial deletion of the presumptive signal peptide, was transcribed and translated into a G2-specific protein in vitro, but no protein product was detected in cells infected with the corresponding vaccinia recombinant virus. In contrast, using a construct containing the full hydrophobic sequence preceding G2, which has all the characteristics of a signal peptide as described by von Heijne (1984) the G2 protein was found to be stably produced and glycosylated in viva. Newly synthesized membrane proteins that fail to undergo translocation into the ER have been reported to be extremely unstable in the cytoplasm (Gething and Sambrook, 1982; Sekikwa and Lai, 1983) . Thus, a likely reason for not detecting a protein product in W-G2(HH)-transfected cells is rapid proteolysis of the untranslocated protein. These results provide evidence that translocation of the G2 protein into the ER can be mediated by a thirteen amino acid hydrophobic sequence, preceding the G2 protein, as a signal peptide. The internally located signal sequences of the PTV polyproteins are similar to those of the Semliki Forest virus structural polyprotein, although the translocational signal peptide of the El protein is located in the C-terminus of a 6-kDa protein rather than the structural protein p62 (Liljestrom and Garoff, 1991; Melancon and Garoff, 1986) .

The results obtained by comparison of truncated vs full-length G2 proteins indicate that a hydrophobic domain near the C-terminus of the G2 protein serves as a membrane anchor. A truncated G2 protein in which the hydrophobic and C-terminal domains were deleted was found to be secreted into the medium, whereas the intact G2 protein remained membrane-associated. This conclusion is consistent with the membrane topology analysis predicted from the deduced amino acid sequence. Since all potential glycosylation sites in the G2 are located in the N-terminal side of the membrane anchor domain, and the G2 protein is glycosylated, we conclude that the G2 glycoprotein is a class I membrane protein with the topology depicted in Fig. 10 .

The G2 glycoprotein does not contain a signal for Golgi retention

The cDNA sequence information for the M genomic segment from members of the bunyavirus family reveals that the predicted C-terminal amino acid sequence of the second glycoprotein in the polypeptide in several genera (G2 in PTV) shows a common feature which includes multiple positively charged residues: -KHKKS in Hantaan virus (Schmaljohn eta/., 1987) and in SR-1 1 viruses (Arikawa et al., 1990 ) -KYKKS in Uukuniemi virus (Ronnholm and Petterson, 1987) and -IKKKN in PTV (Ihara et al., 1985) . We investigated the possibility that this C-terminal sequence of the short cytoplasmic tail may serve as a retention signal, since the C-terminal tetrapeptides WHDEL of the luminal ER proteins have been well defined as an ER retention signal (Munro and Pelham, 1987; Pelham, 1988 ) and a tiULtiI C;UMt'LtX LUC;HLILHI IUN Ut C-terminal consensus sequence in ER membrane proteins was also identified as a signal for their ER retention (Jackson eta/., 1990; Nilsson eta/., 1989) . Interestingly, the G2 C-terminal sequence corresponds to this consensus ER-retention signal, with lysines positioned three and four or five residues from the C-terminus (Jackson et al., 1990) . In contrast to our expectation, a full-length G2 protein was found to be transported to the cell surface. Further, a truncated Gl -G2 protein, in which the G2 C-terminal sequence was deleted, was transported out of the ER, and still targeted to the Golgi complex. Thus, the C-terminal sequence of the G2 glycoprotein does not serve as an ER retention signal or function to maintain a conformation needed for Golgi retention of the Gl and G2 protein. These results indicate that the consensus motif of basic residues at the C-terminus of transmembrane proteins (Jackson et a/., 1990) does not always represent an ER-retention signal, since the PTV glycoproteins with or without this consensus motif are transported out of the ER.

The intracellular retention of the G2 glycoprotein occurs by its association with the Gl glycoprotein

The G2 glycoprotein expressed in the absence of the Gl protein was found to be transported out of the ER and expressed on the cell surface, although its transport was less efficient than transport of the vesicular stomatitis virus (VSV)-G or influenza hemagglutinin (HA) proteins . Furthermore, the anchor-minus G2 protein was found to be secreted into the medium. Thus, we conclude that the G2 glycoprotein does not contain a specific signal for Golgi retention. In contrast to misfolded or unassembled proteins which fail to exit the ER (Copeland et a/., 1988; Doms et al., 1988; Gething et al., 1986; Rose and Doms, 1988) , the PTV membrane glycoproteins are transported out of the ER, but specifically retained in the Golgi complex. Our results indicate that Golgi localization of the G2 glycoprotein requires its interaction with the Gl glycoprotein. This conclusion is supported by the following evidence: (1) The anchor minus G2 protein is secreted into medium, whereas it is retained in the Golgi complex when coexpressed with the Gl protein; (2) The G2 protein when expressed individually is transported to the cell surface, whereas it is retained in the Golgi complex when expressed together with the Gl protein. Thus, it seems likely that a Golgi retention signal is present on the Gl protein. Studies to identify a possible retention signal in the Gl protein are in progress. Alternatively, a functional retention signal might require formation of a complex between the Gl and G2 proteins. We recently observed that the majority of the Gl and G2 glycoprotein are assembled into noncova-IHt i-'UNIH IUKU VIKUY b;IL YKUItlN 505 lently linked Gl -G2 heterodimers, and that the G2 Cterminal and transmembrane domains are not required for dimerization (S.Y. Chen and R. W. Compans, manuscript in preparation). The dimerization of the Gl and G2 glycoproteins was also recently reported in Uukuniemi virus-infected cells (Persson and Pettersson, 1991) .

Except for the G2 C-terminal sequence, the bunyavirus Gl and G2 proteins have no significant sequence homology among different genera, but the hydropathy profiles of the glycoproteins of bunyaviruses reveal a striking similarity, suggesting a common transmembrane topology (Elliott, 1990; Schmaljohn and Patterson, 1990) . Similarly, glycosyltransferases, which are resident membrane proteins of the Golgi complex and the ER, show virtually no sequence homology in their deduced amino acid sequences, but they have a common transmembrane topology with a short NH,-terminal cytoplasmic tail, a 16-to 20-amino acid signal-anchor domain, and an extended stem region which is followed by the large COOH-terminal catalytic domain Shaper et al., 1988; Weinstein et al., 1987) . When various Golgi resident membrane proteins, the coronavirus El glycoprotein (a class III membrane protein), bunyavirus Gl and G2 glycoproteins (class I membrane proteins), and glycosyltransferases (class II membrane proteins) are compared, no sequence homology or similarity of hydropathy profiles among these proteins was found. We believe that these membrane proteins do not employ the same strategy as luminal ER proteins which use a short linear C-terminal sequence as a retention signal. Alternatively, all of these proteins may have a similar signal patch as proposed for lysosomal enzymes specifically recognized by Iv-acetylglucosaminylphosphotransferase (Dahms et a/., 1989; Kornfeld, 1990) , which cannot be observed by analysis of the linear sequence.

Coding properties of the S and M genome segments of Sapporo rat virus: Comparison to other causative agents of hemorrhagic fever with renal syndrome

Bunyaviridae. ln "Virology

Poattranslational association of immunoglobulin heavy chain binding protein with nascent heavy chains in nonsecreting and secreting hybridomas

Vaccinia virus expression of ,&galactosidase provides visual screening of recombinant virus plaques

Assembly and polarized release of Punta Toro virus and effects of Brefeldin A

Conversion of a Golgi apparatus sialyltransferase to a aecretory protein by replacement of the NH,-terminal signal anchor with a signal peptide

Mannose 6-phosphate receptors and lysosomal enzyme targeting

Differential effects of mutation in three domains on folding, quaternary structure, and intracellular transport of vesicular stomatitis virus G protein

Molecular biology of the Bunyaviridae

Analysis of the mRNA transcription processed of snowshoe hare bunyavirus S and M RNA species

Eukaryotic transient expression system based on recombinant vaccinia virus that synthesizes T7 RNA polymerase

Uukuniemi virus glycoproteins accumulate in and cause morphological changes of the Golgi complex in the absence of virus maturation

Nucleotide sequence of cDNA coding for Semliki Forest virus membrane glycoproteins

Expression of wild-type and mutant forms of influenza hemagglutinin: the role of folding in intracellular transport

Construction of influenza haemagglutinin genes that code for intracellular and secreted forms of the protein

A new technique for the assay of infectivity of human adenovirus 5 DNA. Viro/ogy52

Evidence for a separate signal sequence for the carboxy terminal envelope glycoprotein E 1 of Semliki Forest virus

Complete sequences of the glycoproteins and M RNA of Punta Toro phlebovirus compared to those of Rift Valley fever virus

Identification of a consensus motif for retention of transmembrane proteins In the endoplasmlc reticulum

Novel N-terminal amino acid sequence required for retention of a hepatitis B virus glycoprotein in the endoplasmic reticulum

Sorting and traffic in the central vacuolar system

Lysosomal enzyme targeting

Assembly of asparaginelinked oligosaccharides

Uukuniemi virus maturation: Accumulation of virus particles and viral antigens in the Golgi complex

Posttranslational processing of Uukuniemi virus glycoprotein Gl and G2

Cleavage of the structural proteins during the assembly of the head of bacteriophage T4

Sequential events in the morphogenesis of Japanese encephalitis virus

Internally located cleavable signal sequences direct the formation of Semliki Forest virus membrane proteins from a polyprotein precursor

Rapid redistribution of Golgi proteins into the ER in cells treated with Brefeldin A: Evidence for membrane cycling from Golgi to ER

A specific transmembrane domain of a coronavirus El glycoprotein is required for its retention in the Golgi region

A membrane glycoprotein that accumulates intracellularly: cellular processing of the large glycoprotein of Lacrosse virus

Molecular Cloning: A Laboratory Manual

Intracellular accumulation of Punta Toro virus glycoproteina expressed from cloned cDNA

Reinitiation of translocation in the Semliki Forest virus structural polyprotein: identification of the signal for the El glycoprotein

A C-terminal signal prevents secretion of luminal ER proteins

Bunyaviridae: Morphologic and morphogenetic similarities of Bunyamwera serologic supergroup viruses and several other arthropodborne viruses

Short cytoplasmic sequences serve as retention signals for tranamembrane proteins in the endoplasmic reticulum

Replication and amplification of defective interfering particle RNAs of vesicular atomatitis virus in cells expressing viral proteins from vectors containing cloned cDNAs

Evidence that luminal ER proteins are sorted from secreted proteins in a post-ER compartment

Expression of the Hantaan virus M genome segment by using a vaccinia virus recombinant. 1. Viral

Formation and intracellular transport of a heterodimeric viral spike protein complex

Bunyavirus membrane glycoproteins as models for Golgi-specific proteins

Deletion into an NH,-terminal hydrophobic domain result in secretion of rotavirus VP7, a resident endoplasmic reticulum membrane glycoprotein

Nucleotide sequence of the 26s mRNA of sindbis virus and deduced sequences of the encoded virus structural proteins. froc

Complete nucleotide sequence of the M RNA segment of Uukuniemi virus coding the membrane glycoproteins Gl and G2

Regulation of protein export from the endoplasmic reticulum

Vesicular stomatitis virus glycoprotein is anchored in viral membrane by a hydrophobic domain near the COOH terminus

Protein sorting by selective retention in the endoplasmic reticulum and Golgi stack

Primer-directed enzymatic amplification of DNA with a thermostable DNA polymerase

DNA sequencing with chain terminating inhibitors

Replication of Bunyaviridae

Hantaan virus M RNA: Coding strategy, nucleotide sequence, and gene order

Defects in functional expression of an influenza virus hemagglutinin lacking the signal peptide sequences

Characterization of the full length cDNA for murine

Morphogenesis of sandfly fever viruses (Bunyaviridae family)

The signal peptide of the rotavirus glycoprotein VP7 is essential for its retention in the ER as an integral membrane protein

Rift Valley fever virus M segment: Cell-free transcription and translation of virus-complementary RNA. L/iro/ogy 164

Infection of AtT20 murine pituitary tumor cells by mouse hepatitis virus strain A59: virus budding is restricted to the Golgi region. 1

How signal sequences maintain cleavage specificity. 1

Rift Valley fever M segment: Cellular localization of M segment-encoded proteins

Primary structural of @-galactoside a2,6-sialyltransferase

The rate of bulk flow from the endoplasmic reticulum to the cell surface