key: cord-0697230-f9otvby6
authors: Boursnell, M.E.G.; Brown, T.D.K.; Binns, M. M.
title: Sequence of the membrane protein gene from avian coronavirus IBV
date: 1984-12-31
journal: Virus Research
DOI: 10.1016/0168-1702(84)90019-4
sha: 946ec9b4ff8f3a1cd7efc576ba15e9fe3742c073
doc_id: 697230
cord_uid: f9otvby6

Abstract cDNA clones prepared from genomic RNA of coronavirus IBV have been sequenced. The nucleotide sequence for the complete 5' region of mRNA C, which is not present in mRNAs A and B, has been determined. A sequence of 1224 bases is presented which contains a long open reading frame predicting a polypeptide of molecular weight 25 443. This is in agreement with the molecular weight of 23 000 reported for the unglycosylated form of the membrane polypeptide.

Avian infectious bronchitis virus (IBV) is a member of the family Coronaviridae. The coronaviruses are large enveloped viruses with positive-stranded RNA genomes (Siddell et al., 1983 ). The pleomorphic virus particle is generally spherical in shape and contains three major protein structures: the membrane protein, the nucleocapsid protein and the surface projections which form the distinctive 'corona' (Cavanagh. 1981) . The membrane or M protein comprises a polypeptide of molecular weight 23 000 (23k) which is glycosylated to different extents to form glycopolypeptides of molecular weights ranging from 26k to 34k (Stern et al., 1982; Stern and Sefton, 1982b; Cavanagh, 1983) . The oligosaccharides of the M polypeptide are of the high mannose type and are linked to the polypeptide by N-glycosidic linkages (Stern and Sefton, 1982b; Cavanagh, 1983) . The major portion of the membrane protein appears to be embedded in the virion membrane (Cavanagh, 1981) with about 20-40% projecting outside the lipid envelope. Work on the membrane protein of the murine coronavirus MHV-A59 suggests that part of the molecule may also project from the inner surface of the membrane (Sturman et al., 1980) . Six major species of RNA have been observed in IBV-infected cells (Stern and Kennedy, 1980a) . These include RNA F. which is the same length as the viral genome, and five smaller species, RNAs A-E, RNA A being the smallest. These mRNAs consist of a 3'-coterminal 'nested' set with each RNA species containing all the sequences present in the smaller RNAs in addition to some ' unique' sequences at its 5' end (Stern and Kennedy, 1980b) . In vitro translation studies of fractionated and gel-purified mRNAs from IBV have shown that mRNAs A and C code for the nucleocapsid and membrane polypeptides respectively and that RNA E codes for a precursor polypeptide of the surface projection or spike (Stern et al., 1982: Stern and Sefton, 1984) . The sizes of these primary translation products correspond well to the lengths of the 'unique' sequences at the 5' end of each mRNA. Thus it is probable that only the 5' portion of each messenger species is translated.

In this paper we report the DNA sequence of a cloned cDNA copy of IBV genomic RNA in the region corresponding to the 5' end of messenger RNA C. Translation of the sequence predicts a polypeptide of molecular weight 25.4k which is in agreement with the molecular weight of 23k reported for the unglycosylated form of the membrane polypeptide (Stern et al., 1982) .

The preparation of oligo(dT)-primed cDNA clones has been previously described (Brown and Boursnell, 1984) . Briefly, virion RNA was isolated from IBV strain Beaudette grown in embryonated eggs. cDNA was produced by oligo(dT)-primed reverse transcription of the RNA, followed by self-primed reverse transcription to generate the second strand. Sl-treated cDNA was dC-tailed using terminal transferase, annealed to dG-tailed pAT153 (Twigg and Sherratt, 1980) and transformed into E. coli HBlOl. Ampicillin-sensitive colonies were selected for further characterisation.

Viral clones were identified as described by hybridisation with a probe prepared by polynucleotide kinase labelling of alkali-treated, full-length IBV genomic RNA. Restriction sites were mapped on a series of clones and this enabled construction of a continuous map, 3.3 kb in length. Hybridisation with a kinaselabelled poly(U) probe has confirmed that these clones include the poly(A) sequences at the 3' terminus of the viral genome.

cDNA clones A specific oligonucleotide primer, 13 bases long, complementary to a suitable sequence approximately 300 bases from the 5' end of the existing oligo(dT)-primed clones, was synthesised using the phosphotriester method (Gait et al., 1982) . This was used to prime reverse transcription on IBV strain Beaudette genomic RNA. It was passed over a column of Sephadex G-100 equilibrated in 10 mM Tris-HCl pH 7.5, 1 mM EDTA. The excluded fractions were pooled and ethanol precipitated.

The cDNA/RNA hybrids were further fractionated on a Sepharose CL4B column equiiibrated in 0.3 M NaCl, 10 mM Tris-HCl pH 8.0, 1 mM EDTA. Excluded fractions were pooled and ethanol precipitated.

This material was dC-tailed (Roychoudhury and Wu, 1980) . The tailed cDNA was annealed with dG-tailed pBR322 and transformed (Hanahan, 1983) into E. cofi strain LE392. Viral clones were again identified by hybridisation with a kinase-labelled viral probe, and the overlap with the existing clones confirmed by DNA sequencing after reckoning into pUC9 (Messing and Vieria, 1982) . These additional clones enabled construction of a continuous map extending 3520 bases from the 3' end of the viral genome.

Formaldehyde-agarose gel analysis of IB V mRNAs 1.5% formaldehyde-agarose gels were run essentially as described by Maniatis et al. (1982) . Total RNA samples from IBV-infected chick kidney cell cultures were run overnight at 60 V on 16 cm vertical gels. IBV mRNAs were detected by blotting onto nitrocellulose and probing with nick-translated cloned IBV sequences (Maniatis et al., 1982) . Molecular weights were calculated by comparison with the mobilities of DNA restriction fragments and E. coli and chicken ribosomal RNAs.

Sequencing was carried out essentially as described by Maxam and Gilbert (1980) . Plasmid DNA was prepared by a modification of the method of Holmes and Quigley (1981) . For sequencing some regions of the DNA, restriction digests of the clones, or of the viral insert, were recloned into the plasmid pUC9 allowing sequencing from adjacent vector restriction sites (Messing and Vieria, 1982) . Fragments recloned included Alu I digests of the C5.136 insert, and the complete inserts of the two clones C5.136 and 142. This latter procedure allowed sequencing in from the ends of both clones. DNA restriction fragments were 3' end-labelled with ]cu-32P]dNTPs using Klenow polymerase, or 5' end-labelled with [y-32P]ATP using T4 polynucleotide kinase. The two labelled ends were separated by digestion with a second restriction enzyme, electrophoresis on 3.5% or 5% polyacrylamide gels and extraction of the fragments as described (Maxam and Gilbert, 1980) . The depurination reaction was carried out in 66% formic acid for 10 min at 2O*C, after which the samples were treated in the same way as the pyrimidine reaction. The products of the sequencing reactions were analysed on 0.3 mm, 8% polyacrylamide sequencing gels. Additional sequencing was carried out by the dideoxy method of Sanger et al. (1977) , after recloning of Barn Hl digests of (X.136 into the filamentous phage M13mp9, using the Amersham Ml3 cloning and sequencing kit (Amersham International). [ff-] 32P dATP was used in the sequencing reactions and the products were analysed on 0.3 mm, 6% polyacrylamide sequencing gels. All sequencing gels were dried down onto silane-treated glass plates as described by Garoff and Ansorge (1981) . Sequence data were stored and analysed on an Apple He microcomputer using the programs of Larson and Messing (1983) . The hydrophilicity plot from these programs was used with the hydrophilicity values of Kyte and Doolittle (1982) .

Results 1224 base pairs of DNA sequence have been determined from two clones: C5.136 and 142. The positions of these two clones and of the region of sequence presented here are shown in Fig. 1 . In Fig. 2a the arrows show the direction and amount of sequence information determined from individual restriction sites. 97% of the sequence has been determined on both strands, all restriction sites used for the sequencing have been 'sequenced through' from other sites, and a large section has been sequenced by both enzymatic and chemical methods. Fig. 2a also shows restriction sites used in the sequencing.

Sites marked above the line were used for Maxam and Gilbert sequencing directly from the clones. Sequencing from the Alu I sites marked below the line was carried out on fragments recloned into pUC9 (see Materials and Methods). Also shown below the line are the ends of the clones from which sequencing was carried out after recloning into pUC9.

A search was made for initiation and termination codons in the three possible reading frames. These initiation and termination signals are shown in Fig. 2b with the main open reading frames marked. Also shown is the 5' terminus of mRNA B as determined by Sl mapping (Brown and Boursnell, 1984) . Sl mapping of the 5' terminus of mRNA C, using clone 142, proved very difficult, probably due to the small size of the protected fragment. The lack of suitable restriction sites precluded the use of a procedure involving 5' end labelling. The approximate position of the 5' terminus of mRNA C was therefore determined by mRNA length measurements on formaldehyde-agarose gels. These measurements give an estimated iength of 3400 bases. Coronavirus mRNAs have a 5' leader sequence which is fused to the 'body' of the message during transcription (Lai et al., 1983; Baric et al., 1983; Spaan et al_ 1983) . In the case of mRNA C, therefore, the length of the IBV leader sequence, approximately 60 bases (Brown and Boursneli, manuscript in preparation), has been subtracted from the measured length of the mRNA to give the length of the 'body' of the message which is present on genomic RNA.

The 1224 base pairs of sequence are shown in Fig. 3 with a translation of the largest open reading frame. This open reading frame (ORF) of 225 amino acids predicts a polypeptide of molecular weight 25 443. Two smaller ORFs, 4 and 5 (see Fig. 2b ), which follow on directly from the large ORF in the same reading frame, predict polypeptides of 5896 and 3055 daltons respectively. There are also two small ORFs, 2 and 3, present within the sequences of the largest ORF. The ORF starting with the ATG codon at position 1147 is the start of the putative 7500 dalton polypeptide encoded by mRNA B (Brown and Boursnell, manuscript submitted). 

The organisation of the messenger RNAs of coronaviruses and in vitro translation studies have led to the hypothesis that it is only the 'unique' 5' sequences of each mRNA which are translated (Stern and Kennedy, 1980b; Lai et al., 1981) . The sequence shown in Fig. 3 predicts a large open reading frame (ORF 1) of 225 amino acids. If translated this would code for a protein of molecular weight 25.4k. Gel-purified mRNA C, translated in vitro, gives a 23k polypeptide identical in size to the unglycosylated form of the membrane polypeptide (Stern et al.. 1982; Stern and Sefton, 1984) . Hence, the data is entirely consistent with this open reading frame being the coding'sequence for the membrane polypeptide. The sequence also predicts four other open reading frames (ORFs) longer than 25 amino acids. Two of these, ORFs 2 and 3, of 32 and 45 amino acids, respectively, lie within the coding sequences for the membrane polypeptide. Examination of the sequences flanking the initiation codon for the membrane polypeptide gene shows that they conform to one of the sequences preferred for functional eucaryotic initiation codons (Kozak, 1983) . Because of this, and because initiation at anything other than the first AUG codon is known to be very rare (Kozak, 1983) , it seems probable that the 25.4k product of the 225 amino acid open reading frame, ORF 1, is the only polypeptide produced from mRNA C. It is possible, however, that there are minor mRNA species, intermediate in size between mRNAs B and C, which would enable a product to be translated from one or both of the ORFs which are present between the end of the membrane gene and mRNA B. 85 bases before the initiation codon for the membrane polypeptide gene is the sequence CTTAACAA which also occurs 101 bases before the first open reading frame of mRNA A (probably the nucleocapsid gene) and 28 bases before 7.5k open reading frame encoded by mRNA B (Boursnell and Brown, manuscript submitted). The point at which this sequence occurs is also known to mark the 5' termini of the 'bodies' of mRNAs A and B (Brown and Boursnell, 1984) . Thus it seems likely that this sequence also represents the 5' terminus of mRNA C, and its position relative to the estimated position of the end of the body of mRNA C helps to confirm this. It is interesting to note that there appears to be greater homology between the sequences present at the 5' termini of mRNAs A and B than to that which we are postulating occurs at the end of mRNA C (see Fig. 4 ). There is, however, a 12-base homology, AAAACTTAACAA, between the sequences at the 5' termini of mRNAs B and C. The amino acid sequence of the membrane polypeptide reveals several interesting features. It is known that the carbohydrate side chains of the glycosylated forms of Cl TATAC~TATG~TAGAA~ACTT~ACAIATCCGQ~ATTAOAAGC~GTT~ - Fig. 4 . Nucleotide sequences at the 5' termini of mRNAs A, B and C. Some sequence homologies are underlined. the 23k polypeptide are N-linked (Stern and Sefton. 1982b; Cavanagh, 1983) . There are only two sites, Asn-X-Thr, in the amino acid sequence, at which N-linked glycosylation can occur (Hubbard and Ivatt, 1981) . These are at the NH,-terminus at amino acids 3 and 6. This agrees well with the data of Stern et al. (1982) who have shown that glycosylation of the 23k polypeptide occurs at the NH,-terminus. A hydropathicity profile (Fig. 5a) , which shows the local hydrophobic tendencies (positive values) or hydrophilic tendencies (negative values) of a polypeptide chain (Kyte and Doolittle, 1982) shows that there are several hydrophobic regions in the membrane polypeptide.

In particular there are three stretches where the mean hydrophobicities taken over 19 amino acids greatly exceed the lowest value generally associated with membrane-spanning polypeptides (Kyte and Doolittle, 1982) . These regions, which are underlined in Fig. 3 , each consist of a contiguous length of at least 20 uncharged amino acids which are highly enriched in hydrophobic residues. It is possible that each of these regions spans the membrane once, with the short hydrophilic sections separating them being at the surfaces of the membrane. However, since these hydrophilic sections are only very short, it may be that all the hydrophobic NH,-terminal region is buried in the membrane with only the hydrophilic COOH-terminal exposed. It is known that the membrane proteins of the 23k family do not undergo any major post-translational proteolytic processing (Stern and Sefton, 1982a) , and indeed there is no apparent hydrophobic signal sequence at the NH,-terminus which might be cleaved off. However, it is possible that one of the internal hydrophobic Amino acid number regions acts as a signal sequence as may be the case for chicken ovalbumin (Lingappa et al., 1979) . Treatment of the viral particle with the protease bromelain reduces the molecular weight of all the glycosylated forms of the membrane polypeptide to a single polypeptide whose molecular weight is approximately lk less than the size of the unglycosylated membrane polypeptide (D. Cavanagh, personal communication). Thus, the protease appears to cleave at least the first six amino acids with their associated variable carbohydrate side chains from the NH,-terminus of the membrane protein. The fact that this is achieved by protease treatment of largely intact virus particles suggests that the carbohydrate side chains, and hence the NH,terminus, are located on the outside of the viral membrane.

This would have been expected by analogy with the membrane (El) polypeptide of MHV and with other viral glycoproteins (Sturman and Holmes, 1977; Rose and Gallione, 1981) . The 75 or so amino acids of the COOH-terminal region of the polypeptide form a hydrophilic region which may protrude from the inner surface of the viral envelope. For MHV-A59 Sturman et al. (1980) have shown that there is some interaction between the membrane (El) protein and the viral RNA. If this is also the case for IBV it is interesting to note that of the 18 basic residues in the membrane polypeptide, 14 of them are in the COOH-terminal half of the molecule. It is possible that these basic residues are involved in binding to the RNA.

Comparison of the sequence of the IBV membrane protein and the El protein of MHV-A59 (Peter Rottier, personal communication) at the RNA sequence level reveals essentially no homology. However, comparison of the predicted amino acid sequences (Fig. 6) shows a remarkable degree of homology especially in the amino terminal half of the molecule. In particular, the sequence within the first two potential membrane-spanning regions seem to be highly conserved, with 24 out of 44 ammo acids the same and most of the differences involving pairs of amino acids with similar properties. A region of even greater homology occurs in the hydrophilic stretch of residues at position 104 (Fig. 6 ) in which 15 out of 19 residues are the same. These similarities at the amino acid levtsl are strikingly reflected in the hydropathicity profiles (Fig. 5) . This protein sequence conservation in the hydrophobic NH,-terminal half of the polypeptide suggests a high degree of selective pressure for maintenance of the structure of the putative trans-membrane regions of coronavirus membrane proteins.

In conciusion then, the sequence presented here shows good agreement with the observed and expected properties of the membrane polypeptide of infectious bronchitis virus,

We are grateful to Penny Gatter, Anne Foulds, Ian Foulds and Bridgette Britton for excellent technical assistance. This research was carried out under Research Contract No. GBI-2-Oil-UK of the Biomolecular Engineering Programme of the Commission of the European Communities. We wish to thank Dr. Greg Winter, MRC Laboratory of Molecular Biology, Cambridge, for help in the preparation of the synthetic oiigonuc~eotjde primer, and David Cavanagh for valuable discussion during the writing of this paper. We would also like to thank Dr. Peter Rottier for allowing us to use the unpublished sequence of the El protein of MHV-A59. Some of this data was presented in a preliminary form at the EMBO workshop C~~o~~vj~~e~: ~ofec~l~r Si&gy and ~ar~~ge~es~~, 1983, Zeist, The Netherlands.

Characterisation of replicativc intermediate RNA of mouse hepatitis virus: presence of leader RNA sequences on nascent chains

Poliovirus type 3 molecular cIonin8 of the genome and nudeotide sequence of the region encoding the protease and polymerase proteins

Coronavirus IBV glycopoiypeptides: size of their polypeptide moieties and nature of their oiigosacch~des

Rapid synthesis of oIig~eox~~bonucIeotides VII. Solid phase synthesis of oiig~eoxyribonucieotides by a continuous flow phosphotriester method on a kiesel8uhr-polyamide support

Improvements in DNA sequencing gels

Studies on transformation of &cc&+&in coli with plasmids

A rapid boiling method for the preparation of bacterial plasmids

SC. and Ivatt, R.J. (1981) Synthesis and processing of asparagine-linked oligosaccharides. Annu. Rev. B&hem. 50, 555-584. Kozak, M. (1983) Comparison of initiation of protein synthesis in procaryotes, eucaryotes and organelles. Microbial. Rev. 47, l-45. Kyte, J. and Doolittle, R.F. (1982)