A Specific Transmembrane Domain of a Coronavirus E1 Glycoprotein Is Required for Its Retention in the Golgi Region Carolyn E Machamer and John K Rose Molecular Biology and Virology Laboratory The Salk Institute for Biological Studies San Diego California 92138 Abstract The E1 glycoprotein of the avian coronavi ms infectious bronchitis virus contains a short glycosylated amino terminal domain three membrane spanning domains and a long carboxy terminal cyto plasmic domain We show that E1 expressed from cDNA is targeted to the Golgi region as it is in in fected cells E1 proteins with precise deletions of the first and second or the second and third membrane spanning domains were glycosylated thus suggesting that either the first or third transmembrane domain can function as an internal signal sequence The mu tant protein with only the first transmembrane domain accumulated intracellulady like the wild type protein but the mutant protein with only the third transmem brane domain was transported to the cell surface This result suggests that information specifying accumula tion in the Golgi region resides in the first transmem brane domain and provides the first example of an in tracellular membrane protein that is transported to the plasma membrane after deletion of a specific domain T hE intracellular transport and targeting of proteins from their site of synthesis to their correct destination is a process instrumental to malntainence of cellular integrity Membrane bound and secreted proteins are sorted from cytoplasmic proteins and those destined for the nucleus or mitochrondria by virtue of a signal or leader sequence which directs insertion into the lumen of the endoplasmic reticulum Walter et al 1984 Plasma membrane and se creted proteins follow a similar pathway through the cell from the rough endoplasmic reticulum through the Golgi complex en route to the plasma membrane Sabatini et al 1982 Other proteins in this exocytic pathway are retained at points along the way such as cytochrome P 450 in the en doplasmic reticulum Brands et al 1985 and galactosyl transferase in the Golgl complex Roth and Berger 1982 Lysosomal enzymes are sorted to lysosomes after passage through the Golgi complex Some lysosomal enzymes are sorted via a specific marker a mannose 6 phosphate modi fication of asparagine linked N linked 1 oligosaccharides of these enzymes which is recognized by a receptor Sly and Fischer 1982 The process of intracellular transport to the plasma mem brane could be governed by positive sorting signals i e pro C E Machamer s present address is Department of Pathology Yale Univer sity School of Medicine New Haven Connecticut 06510 J K Rose s pres ent address is Departments of Pathology and Cell Biology at the same in stitute 1 Abbreviations used in this paper IBV infectious bronchitis virus MHV mouse hepatitis virus N linked asparagine linked VSV vesicular stomati tis virus teins with these signals would be transported to the cell sur face via a type of receptor mediated process like that involved in targeting of lysosomal enzymes Alternatively as recently suggested by Kelly 1985 and Rothman 1986 pro teins that are destined for the plasma membrane or for con stitutive secretion could move passively with the bulk flow of lipids while proteins secreted in a regulated manner or those retained in intracellular membranes would possess sig nals which selectively retain them from moving with the bulk flow of membranes A third model could invoke both positive signals that enhance the rate of incorporation of proteins into transport vesicles and negative retention signals that ensure intracellular proteins to not be transported beyond a certain point Recombinant DNA technology that uses gene expression and site directed mutagenesis offers a powerful approach to the study of intracellular transport and targeting For exam ple using this approach an involvement of cytoplasmic do mains of proteins has been suggested in facilitating both exocytosis Rose and Bergmann 1983 and endocytosis Lehrman et al 1985 Roth et al 1986 and in targeting of proteins to basolateral membranes of polarized epithelial cells Mostov et al 1986 Puddington et al 1987 Al though specific molecules interacting with these domains have not yet been identified these results suggest that impor tant interactions with cytoplasmic domains do occur Viruses provide simple and useful model systems for studying the signals involved in targeting of membrane pro teins in cells The coronaviruses provide an especially inter esting model because they bud from intracellular membranes rather than from the plasma membrane The best studied 9 The Rockefeller University Press 0021 9525 87 09 1205 10 2 00 The Journal of Cell Biology Volume 105 September 1987 1205 1214 1205 on June 4 2015 jcb rupress org Downloaded from Published September 1 1987 members of the coronavirus group are mouse hepatitis virus MHV and avian infectious bronchitis virus IBV Both viruses specify two glycoproteins called El and E2 reviewed by Sturman and Holmes 1985 E2 forms the virion spike and can be detected in the plasma membrane as well as intra cellular membranes of infected cells In contrast E1 ac cumulates intracellularly and thus appears to play a critical role in intracellular budding of the virus The viral nucleo capsid interacts with the MHV E1 protein presumably the cytoplasmic tail and virion budding occurs at the site of E1 accumulation in infected cells Sturman et al 1980 Dubois Dalque et al 1982 Tooze et al 1984 The sequences of cDNAs encoding E1 proteins from both MHV and IBV have been recently reported Armstrong et al 1984 Boursnell et al 1984 These sequences predict polypeptides with a similar structure a short glycosylated amino terminal domain three hydrophobic domains be lieved to span the membrane three times and a long cytoplas mic domain at the earboxy terminus Studies that use pro tease treatment of intact virus Sturman and Holmes 1977 Cavanagh et al 1986 or of E1 inserted into microsomal membranes Rottier et al 1984 support the model for the structure of El Since there is no cleaved amino terminal sig nal sequence Rottier et al 1984 and the protein is inserted into microsomal membranes in a signal recognition parti cle dependent manner Rottier et al 1985 it has been sug gested that one or more of the putative membrane spanning domains functions as an internal uncleaved signal sequence We report here that the IBV E1 glycoprotein is retained in the Golgi region of the cell in the absence of the other viral proteins when expressed from cDNA In an earlier study using gene expression and mutagenesis techniques we deter mined that the cytoplasmic domain of the IBV El protein did not contain a signal that was capable of retaining the G pro tein of vesicular stomatitis virus VSV in intracellular mem branes Puddington et al 1986 Thus in the present study we chose to search for a retention signal in the amino terminal half of the IBV E1 protein We report that the first of the three membrane spanning domains may constitute such a signal and our results are discussed in terms of cur rent models for protein transport to the plasma membrane Materials and Methods Construction of an Expression Vector Encoding IBV El A cDNA clone encoding the E1 protein of the Beaudette strain M42 of 1BV was derived from viral genomic RNA and kindly provided by D Stern pIBV 5 Stern 1983 A subclone containing the entire coding region was prepared p57 6 and the nucleotide sequence was determined by the proce dure of Maxam and Gilbert 1977 This sequence predicts a polypeptide with 225 amino acids and differs at only two nucleotides from the sequence published by Boursnell et al 1984 for the El protein of the Beaudette strain of IBV These differences were a T instead of a C at nucleotide 167 and a C instead of a T at nucleotide 375 changing the codon for Pro 2 and for Thr 71 to that for Ser and Ile respectively A fragment that contains the en tire coding sequence of IBV El 773 bp was excised from p57 6 with Hpa I and Hha I and incubated with the Klenow fragment of DNA polymerase I and deoxynucleoside triphosphates to remove the 3 overhang After liga tion with Xho I linkers the fragment was cloned into the unique Xho I site of the SV40 based expression vector pJCll9 Sprague et al 1983 and a clone with the insert in the proper orientation for expression from the late SV40 promoter was selected pSV IBVEI This construct includes a 5 un translated region of 52 nucleotides and a 3 untranslated region of 45 nucleotides Oligonucleotide directed Mutagenesis Synthetic oligonucleotides were used to precisely delete coding sequences for putative membrane spanning domains of El The negative strand of the IBV E1 gene was cloned into the Barn HI site of MI3 mp8 as described Bose et al 1984 and single stranded phage DNA was purified for use as the template for mutagenesis Synthetic oligonucleotides 33 mers were de signed to loop out specific regions of the coding sequence by hybridizing with 16 nucleotides on the 5 side of the loop and 17 nucleotides on the 3 side The domain junctions were assigned on the basis of secondary struc ture and hydrophobicity predictions for IBV El Boursnell et al 1984 Rot tier et al 1986 The oligonucleotide used to generate the coding sequence for the mutant protein that lacked the second and third hydrophobic domains Am2 3 was 5 GTA ATC AACAAGACIL TI AAGC 3 and that for the mutant protein that lacked the first and second hydrophobic domains Am1 2 was 5 ICAGCI TTTTAAAGAGGGAGGICr IGI C GCAGC 3 They were synthesized and purified as p viously described Rose et al 1984 Machamer et al 1985 After hybridization of the oligonucleotide with the template DNA primer extension was carried out with the Klenow fragment of DNA polymerase I Boehringer Mannheim Biochemicals Indi anapolis IN by the procedure previously described Adams and Rose 1985 Machamer et al 1985 After transfection of competent Escherichia coil JM103 with the primer extension mixture plaques containing the muta tions were identified by differential hybridization with the 5 32p labeled oligonucleotide DNA containing the mutation was excised and subeloned into the expression vector pJC119 and the mutation confirmed by DNA se quence analysis Maxam and Gilbert 1977 The mutant protein Am2 3 was created by fusion of the codon for Thr 42 to that for Arg 102 deleting 156 nucleotides and Am1 2 by fusion of the codon for Glu 20 to that for Gly 77 deleting 168 nucleotides Cells and Virus Both COS 1 and HeLa cells were maintained in Dulbecco Vogt s modified Eagle s medium with 5 FCS The IBV used in these experiments was adapted to growth in monkey cells by nine consecutive passages of egg grown virus at low multiplicity of infection in Vero cells and was obtained from B Sefton The Salk Institute Tissue culture supernatant from infected Vero cells was used as the inoculum to infect COS I cells for radiolabeling and inununofluorescence The multiplicity of infection was approximately 0 1 and infected cells were assayed for IBV speciflc polypeptides at 29 and 48 h postinfection Preparation of Recombinant Vaccinia Viruses Encoding El and A ml 2 DNA encoding wild type E1 or the mutant protein Aml 2 was excised from the SV40 expression vector with Xho I and the ends were filled in with the Klenow fragment of DNA polymerase I The inserts were then subcloned into the vector pSCll Chakrahatti et al 1985 at the unique Sma I site After selection of clones with inserts in the correct orientation for expres sion from the vaccinia promoter PT SK supercoiled DNA was transfected onto HeLa cells by calcium phosphate coprecipitation The HeLa cells had previously been infected at a multiplicity of 0 05 with wild ty vaccinia vi rus 1 h before addition of the DNA The medium was replaced 18 h later and incubation continued another 2 d Recombinant virus was isolated by including X gal 5 bromo 4 chloro 3 indolyl D alactolrynmoside Boeh ringer Mannheim Biochemicals in the plaquing overlay as described Chakrabarti et al 1985 The pSCll vector contains the E coli lacz gene under control of a late vaccinia promoter thus 5 galactosidase activity results in formation of blue plaques by recombinant virus Recombinant viruses encoding E1 or Aml 2 proteins VVE1 and WArn1 2 were plaque purified three times in HeLa cells and large stocks were prepared HeLa cells 5 x 105 in 35 ram dishes were infected with the recombinant vac cinia viruses at a multiplicity of approximately 4 and analyzed for expres sion 5 h later Antibody Preparation A polyclonal rabbit antiserum was raised to a synthetic peptide correspond ing to the carboxy termina122 amino acids oflBV El The peptide was con jugated to BSA via the penultimate tyrosiue residue with bis diazobenzidlne The Journal of Cell Biology Volume 105 1987 1206 on June 4 2015 jcb rupress org Downloaded from Published September 1 1987 DeCarvalho et al 1964 Two New Zealand white rabbits were immunized with 1 mg conjugate each emulsified with 0 75 ml complete Freund s adju vant in a total volume of 1 5 ml About 30 sites were injected intradermaily with 50 ltl per site Rabbits were boosted with 0 5 mg conjugate 0 75 ml total in the same way every 4 wk ELISA titers to the peptide were mea sured in a solid phase assay that uses a second antibody conjugated with horseradish peroxidase and were approximately 1 10 000 after the second boost This titer did not change significantly after repeated boosts After the second boost the synthetic peptide was discovered to be incorrect in that it contained an extra glycine after the glycine at position 8 of the peptide and subsequent boosts were performed with the correct synthetic peptide Ability of the antiserum to immunoprecipitate the hybrid protein 323 Pnd dington et ai 1986 was detected after the third boost This titer increased until the fifth boost when one of the two rabbits was producing antibody with a slightly higher titer than the other with less background in im munoprecipitates This antiserum was used for all subsequent experiments Anti peptide antibodies were affinity purified on Affigel histamine columns coupled with the peptide Histamine was coupled to Affi gel 10 Bio Rad Laboratories Richmond CA the peptide was then coupled to the conjugated resin via the tyrosine residue with bis diazobenzidine as de scribed above 10 mg of peptide was coupled to 1 mi of resin and remaining active groups were blocked with 1 mg ml ovalbumin Serum 4 mi contain ing antibodies to the peptide was incubated with 1 ml of conjugated resin for 4 h at 4 with end over end inversion After transfer to a column resin was first washed with 20 mi of 10 mM Tris pH 8 0 with 0 2 deoxycbolate and then with 15 mi of 10 rnM Tris pH 8 0 containing 0 5 M NaCI Specific antibodies were eluted with 4 M MgC12 and dialyzed immediately against 10 mM Tris pH 7 4 containing 0 15 M NaC1 and 0 02 NAN3 Pooled fractions with the highest OD2s0 reading were stored at 40C with 100 U mi kallikrein inhibitor Calbiochem Behring Corp La Jolla CA and 0 5 mg mi ovalbumin Approximately 1 5 mg of purified immunoglobulin was recovered from 4 ml of serum Transfection of COS 1 Cells Radiolabeling and Immunoprecipitation COS 1 cells 4 x 105 plated in 35 ram dishes the previous day were trans fected with 10 pg supercoiled DNA using DEAE dextran followed by chio roquine treatment as previously described Adams and Rose 1985 Ap proximately 44 h after transfe tion cells were labeled at 3 C with 50 pCi 35S cysteine 1 300 Ci mmol Amersham Corp Arlington Heights IL in 0 5 ml cysteine free Dulbecco Vogt s modified Eagle s medium containing 4 dialyzed FCS for 1 h or for the time indicated For treatment with tunlcamycin transfected cells were pretreated with 3 pg ml tunlcamycin Sigma Chemical Co St Louis MO for 2 h and then labeled in the pres ence of the same concentration of the drug COS 1 or HeLa cells infected with recombinant vaecinia viruses were labeled 5 h after infection with sS cysteine as described above After labeling cells were lysed at 0 in 0 5 ml of a solution containing 50 mM Tris pH 8 0 1 NP 40 0 4 deoxy cholate 62 5 mM EDTA and 100 U ml kallikrein inhibitor per mi Nuclei were removed by centrifugation at 15 000 g for 1 rain and lysates were ad justed to a final concentration of 0 3 SDS For immunopreeipitation of El proteins 5 Ixl of anti peptide serum not affinity purified was incubated for 2 h with 0 5 ml lysate at 4 Antigen antibody complexes were isolated with protein A bearing Staphylococcus aureus Pansorbin Calbiochem Behring Corp and washed four times with RIPA buffer 10 mM Tris pH 7 4 0 15 M NaC1 1 NP 40 1 dcoxyeholate and 0 1 SDS Pellets were eluted by incubation in Laemmli sample buffer containing 2 2 mer captoethanol at room temperature for 20 rain unless otherwise noted and the S aureus was removed by centrifugation In the experiment shown in Fig 2 B pellets were eluted in 1 M Tris pH 8 8 2 SDS and 2 2 mer captoethanol at 100 C for 1 rain After removal of the S aureus cells super natants were incubated with a final concentration of 0 33 M iodoacetamide for 60 min at room temperature and proteins were precipitated with 9 voi of acetone at 20 C for 2 h After washing in acetone precipitates were dried and resuspended in sample buffer without 2 mereaptoethanol Im munoprecipitates were subjected to electrophoresis in 10 or 15 poly acrylamide gels containing SDS Laemmli 1970 Marker proteins were aC methylated standard molecular weight markers Amersham Corp Labeled proteins were detected by fluorography Bonner and Laskey 1974 Indirect Immunofluorescence Microscopy COS 1 cells grown on coverslips were fixed with paraformaldehyde and per meabilized with NP 40 as described Rose and Bergmann 1982 E1 pro teins were detected by incubation with the affinity purified anti E1 peptide antibody described above 1 30 followed by affinity purified fluorescein conjugated goat anti rabbit IgG 1 50 Southern Bioteehnology Associates Inc Birmingham AL For localization of the Golgi complex coverslips were incubated with rhodamine eonjugated wheat germ agglutinin 1 100 E Y Laboratories Inc San Marco CA Calls were visualized with a Nikon Optiphot microscope equipped with fluorescence epiillumination and a Nikon 40 x oil immersion plan apoehromat objective Treatment of Intact Cells with Proteases Transfected COS 1 cells were radiolabeled as described above with 35S eysteine for 90 min and incubated in the presence of excess unlabeled eysteine for 90 min Cells were then incubated for 15 min at 37 C in 0 5 ml of PBS containing 1 mg ml bromelain Calbiochem Behring Corp and 0 1 mM 2 mercaptoethanol Parallel dishes of transfected cells were in cubated for the same period of time in the absence of bromelain then lysed as usual Bromelain treated cells were collected by centrifugation and washed three times in Tris buffered saline lysed and E1 proteins were im munoprecipitated as described above HeLa ceils infected with VVEI or VVAml 2 were labeled 5 h postinfeetion with 35S eysteine for 60 rain As described above cells were either a mock digested for 15 min at 37 C b treated with 1 mg ml bromelain c treated with I mg ml trypsin TICK Worthington Biochemical Corp Freehold NJ in PBS or d treated in medium with 1 FCS containing 1 mg ml pronase from Streptomyces griseus Boehringer Mannheim Biochemicals which had been previously self digested for 15 rain at 37 Results Construction of an Expression Vector Encoding IBV El A eDNA clone prepared from viral genomic RNA encoding the IBV E1 protein was obtained from D Stem Stern 1983 and subjected to DNA sequence analysis This nucleotide se quence predicts a polypeptide of 225 amino acids and is identical to that for the IBV E1 protein as reported by Bours nell et al 1984 with the exception of two nucleotides Fig 1 see Materials and Methods This sequence includes two potential sites for N linked glycosylation at Asn 3 and Asn 6 both of which are glycosylat