Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
VIROLOGY 251, 227±233 (1998) VY989302 ARTICLE NO. The Attachment Protein of Hendra Virus Has High Structural Similarity but Limited Primary Sequence Homology Compared with Viruses in the Genus Paramyxovirus Meng Yu,* Eric Hansson,* Johannes P. M. Langedijk,² Bryan T. Eaton,* and Lin-Fa Wang*,1 *CSIRO Division of Animal Health, Australian Animal Health Laboratory, P.O. Bag 24, Geelong, Victoria 3220, Australia; and ²Department of Molecular Biology, The Scripps Research Institute, 10550 N. Torrey Pines Road, San Diego, California 92037 Received March 24, 1998; returned to author for revision May 15, 1998; accepted June 12, 1998 The complete nucleotide sequence of the attachment protein gene of Hendra virus, a new member of the subfamily Paramyxovirinae, has been determined from cDNA clones derived from viral genomic RNA. The deduced mRNA is 2565 nucleotides long with one open reading frame encoding a protein of 604 amino acids, which is similar in size to the attachment protein of the members of the subfamily. However, the mRNA transcript is .600 nucleotides longer than others in the subfamily due to the presence of long untranslated regions at both the 59 and 39 ends. The protein is designated G because it lacks both hemagglutination and neuraminidase activities. It contains a hydrophobic transmembrane domain close to the N terminus, eight potential N-linked glycosylation sites, and 18 cysteine residues. Although the HeV G protein had low sequence homology with Paramyxovirinae members, the predicted folding pattern of its extracellular globular head was very similar to that of members of the genus Paramyxovirus, with the location of seven potential pairs of sulfide bonds absolutely conserved. On the other hand, among the seven residues known to be critical for neuraminidase activity, only one was conserved in the Hendra virus G protein compared with at least six in HN proteins of paramyxoviruses and rubulaviruses and four in H proteins of morbilliviruses. The biological significance of this finding is discussed. © 1998 Academic Press Paramyxovirus and Rubulavirus genera, the Morbillivirus genus, and the Pneumovirus genus have been designated HN, H, and G, respectively, depending on whether they display haemagglutination (H) and neuraminidase (N) activities. Pneumoviruses lack both activities. Recently, the neuraminidase-negative status of morbilliviruses was challenged by Langedijk et al. (1997), who demonstrated that the attachment proteins of rinderpest virus (RPV) and peste des petits ruminant virus (PPRV), members of the genus Morbillivirus, manifest enzyme activity. The function of the neuraminidase is not well understood in members of the family Paramyxoviridae. In influenza viruses, members of the family Orthomyxoviridae, the H and N activities are encoded by two different genes, and N has been shown to be necessary for the release of progeny virus from infected cells by cleavage of sialic acids (Palese et al., 1974). Another role for the neuraminidase activity in orthomyxoviruses and paramyxoviruses may be for virus transport through the sialic acid-rich mucus secretions that serve to protect the respiratory tract, and therefore the substrate specificity of the enzyme may be related to the host and the site of infection (Langedijk et al., 1997). In September 1994, a new member of the Paramyxoviridae family was identified as the causative agent of an outbreak of respiratory disease that resulted in the death of horses and one person (Murray et al., 1995; O'Sullivan et al., 1996). The virus superficially resembled morbilliviruses and was provisionally named equine morbillivirus (EMV). As information on the sequence of the genome INTRODUCTION The family Paramyxoviridae was reclassified into two subfamilies in 1993: the Paramyxovirinae and the Pneumovirinae (Murphy et al., 1995; Lamb and Kolakofsky, 1996). Members of the Pneumovirinae can be distinguished from the Paramyxovirinae on the basis of their nucleocapsid diameter (13±14 vs 18 nm, respectively), length of glycoprotein spike protruding from the virus surface (10±12 vs 15 nm, respectively), and genome organization. Pneumovirinae and Paramyxovirinae members contain more than seven and fewer than seven transcriptional units, respectively. In the subfamily Paramyxovirinae, there are three genera, Paramyxovirus, Rubulavirus and Morbillivirus, which are differentiated on the basis of (1) antigenic cross-reactivity, (2) the presence or absence of neuramidinase activity, and (3) the coding strategies of their respective P genes (Murphy et al., 1995). Members of the family Paramyxoviridae contain two envelope glycoproteinsÐthe fusion protein (F) and the attachment protein, with the latter responsible for binding the virus to specific receptor molecules on the host cell surface. In some cases, it has been demonstrated that the attachment protein is also required for membrane fusion by acting in concert with the F protein (Bousse et al., 1994; Sergel et al., 1993). Attachment proteins of members of the 1 To whom reprint requests should be addressed. Fax: 161±3-5227± 555. E-mail: [email protected]. 227 0042-6822/98 $25.00 Copyright © 1998 by Academic Press All rights of reproduction in any form reserved. 228 YU ET AL. and its organisation became available (Wang et al., 1998; Yu et al., 1998), it was clear that although the virus is a member of the subfamily Paramyxovirinae, it does not fit comfortably into any of the existing genera. Indeed, it seems likely that this new virus represents the prototype of a new genus in the subfamily Paramyxovirinae (Wang et al., 1998; Yu et al., 1998). As a result, the name of the virus was changed to Hendra virus (HeV), after the location of the first outbreak in Brisbane, Australia (Murray et al., 1998; Wang et al., 1998). HeV naturally infects humans and horses and has been isolated from fruit bats, which are believed to be the natural hosts of the virus (Halpin et al., 1996). The virus is also pathogenic in experimentally infected cats, horses, and guinea pigs (Westbury et al., 1995; 1996). The wide host range distinguishes HeV from other known members of the Paramyxovirinae, which generally have narrow host specificity. The sequence and putative structural features of the HeV attachment protein (G) were determined and compared with those of other members of the Paramyxovirinae to determine whether the HeV attachment protein has any unique feature that may be responsible for its wide host range. RESULTS AND DISCUSSION Sequence features of the HeV G gene and protein The complete nucleotide sequence of the G gene was first determined from overlapping clones isolated from a cDNA library and then confirmed by direct sequencing of polymerase chain reaction (PCR) fragments derived from genomic RNA. The gene sequence with the deduced amino acid sequence is presented in Figure 1. The mRNA is 2564 nucleotides (nt) in length and is flanked by the trinucleotide intercistronic sequence CTT (not shown in the figure) as described for the other HeV genes identified so far (Gould, 1996; Wang et al., 1998; Yu et al., 1998). Transcription initiation (59-AGGACCCAAG-39) and termination sequences (59ATTAAGAAAAA-39) were identified on the basis of their conservation with other members of the subfamily and other HeV genes characterized so far (Lamb and Kolakofsky, 1996; Wang et al., 1998). A search for open reading frames (ORFs) indicated only one of significant size in the G mRNA. The ORF encodes a protein that is 604 amino acid residues in length and has a calculated molecular mass of 67,251 Da. There are three in-frame ATG codons (highlighted in Fig. 1) before the hydrophobic transmembrane domain. In theory, translation initiation from any of these three sites could produce a functional attachment protein. According to Kozak (1991), the optimal context for translation initiation has the consensus sequence of GCCA(G)CCATGG, in which the A (or G) residue at 23 and the G residue at 14 play the most important role in modulating initiation efficiency. For the G gene, the third ATG codon has the most favorable Kozak sequence, followed by the second, located immediately after the first. The G gene may therefore produce several attachment proteins that differ in the length of their cytoplasmic domain. If the first (or second) ATG codon is used in translation, the HeV G protein would contain a cytoplasmic domain of ;45 amino acids, which is much longer than that of known Paramyxovirinae members. Regardless of which ATG codon is used for initiation, the cytoplasmic domain is rich in lysine and is highly charged. The significance of this difference is not clear, but a previous report by Parks and Lamb (1990) demonstrated that the cytoplasmic domain of paramyxovirus HN protein does play a role in the assembly and transport of this type II membrane protein. The overall length of the potential translation products merits comment. Initiation at the first ATG codon produces a protein of 604 amino acids, which resembles those of morbilliviruses (604±617 amino acids), whereas initiation at the third ATG codon produces a protein (573 amino acids) much closer in size to that of paramyxovirus HN proteins (572±575 amino acids). It is also interesting to note that there are two in-frame ATG codons located within the hydrophobic transmembrane domain (see Fig. 1). Their location is reminiscent of the alternative translational initiation site used by the respiratory syncytial virus G gene to produce a soluble form of the attachment protein (Roberts et al., 1994). Although the length of the coding region of HeV G protein is similar, the G gene transcript length is much larger in HeV than that in other Paramyxovirinae members. This is due to the presence of longer untranslated regions at both ends of the mRNA. The HeV G gene has 59 and 39 untranslated regions of 233 and 516 nt, respectively, compared with 20±73 and 85±111 nt, respectively, for other members of the subfamily. Long untranslated regions are also found in the N, P, M, and F genes of HeV (Gould, 1995; Wang et al., 1998; Yu et al., 1998). Similar regions characterize the genes of members of the family Filoviridae, including Ebola and Marburg viruses (Feldmann and Klenk, 1996). The function of long untranslated regions in negative-stranded RNA viruses is not known, but a role in increasing mRNA stability is a possibility. The deduced HeV G protein has eight potential N-linked glycosylation sites. Other Paramyxovirinae attachment proteins have five to seven such sites. The distribution of these residues seems to be evenly spread along the molecule, rather than concentrated at a particular region, as in the case of N-glycosylation sites for measles virus H protein (Lamb and Kolakofsky, 1996). The viral G protein is glycosylated in vivo as shown by its mobility in SDS-PAGE and its binding to lectins (Wang et al., 1998; W. Michalski and B. T. Eaton, unpublished results). The HeV G protein has 18 cysteine residues compared with 12±17 for other subfamily members. The importance of cysteine residue location is discussed later. Sequence and structural alignment of HeV G protein with other Paramyxovirinae HN/H proteins Comparison of primary sequences of HeV G and Paramyxovirinae HN and H proteins indicated that the THE HENDRA VIRUS G PROTEIN 229 FIG. 1. DNA and deduced amino acid sequence of the HeV G gene. Amino acid sequence is given in single letter code beneath the DNA sequence. Conserved nucleotide sequences for transcription initiation and termination are underlined. The three in-frame ATG start codons together with their encoded methionine (M) residues are indicated in bold. Important amino acid sequence features are highlighted as follows: cysteine residues, shaded; N-linked glycosylation sites; underlined by a double straight line; hydrophobic transmembrane domain, underlined with a wavy line. The nucleotide sequence reported here has been submitted to the GenBank/EMBL and assigned the accession number AF017149. overall sequence conservation is low (data not shown). The percentage sequence identities vary from 11% to 23%, with HeV G displaying a closer relationship to paramyxoviruses than to morbilliviruses and rubulaviruses. The highest sequence identity (23%) exists between HeV and human parainfluenza virus 1 (hPIV1). However, this is only half of that observed among known paramyxoviruses (45% between human PIV1 and bovine PIV3). This overall low level of sequence identity is consistent with previous analyses based on the sequences of the N, P, M, and F proteins, suggesting that HeV may represent the prototype virus 230 YU ET AL. FIG. 1ÐContinued of a new genus in the subfamily Paramyxovirinae (Wang et al., 1998). The structure of paramyxovirus attachment proteins can be divided into four regions: N-terminal cytoplasmic tail, hydrophobic transmembrane domain, stem, and extracellular globular head. Recently, Langedijk et al. (1997) conducted a detailed sequence and structure alignment of the globular head domains of paramyxovirus and morbillivirus HN and H proteins and defined a highly conserved structure containing a canonical fold, which is conserved in neuraminidases derived from viruses, bacteria, and eukaryotic organisms. Similar sequence and structural alignments were conducted for HeV G protein using sequences derived from the stem and globular head domains. Several important observations were made from the alignment presented in Figure 2: (1) HeV G protein shares many structural features characteristic of viruses in the genus Paramyxovirus. The locations of all seven proposed disulfide bonds are absolutely conserved between HeV and paramyxoviruses, whereas five are conserved between HeV and rubulaviruses THE HENDRA VIRUS G PROTEIN 231 FIG. 2. Sequence and structure alignment of the globular head region of Paramyxovirinae attachment proteins. (See below for abbreviations and sources of sequences used in this analysis.) The locations of the strands of b-sheets 1±6 are given above the sequence. Residues conserved in all nine viruses are shaded, whereas the seven proposed active-site residues are indicated by numbers 1±7. Residues conserved between HeV and neuraminidase-positive viruses (i.e., paramyxoviruses and rubulaviruses) are shown in bold capital letters, and those conserved between HeV and morbilliviruses are shown in bold small letters. Proposed disulfide bonds conserved between HeV and other members of the subfamily are shown by solid lines, with the long range bond indicated by an open box. Disulfide bonds unique to morbillivirus and rubulaviruses are shown in dashed lines. Residue numbering is indicated after each line. Abbreviations (Databank accession number): CDV, canine distemper virus (P24306); PDV, phocine distemper virus (Z36979); MeV, measles virus (P08362); RPV, rinderpest virus (M21512); DMV, dolphin morbillivirus (Z36978); SeV, Sendai virus (P06863); PIV3, bovine parainfluenza virus 3 (P06167); MuV, mumps virus (M19933); and NDV, Newcastle disease virus (P32884). 232 YU ET AL. and between HeV and morbilliviruses. Rubulaviruses and morbilliviruses also have a set of disulfide bonds unique to their respective genera. (2) In addition to cysteine residues, other structurally important residues (such as glycine, proline, and aromatic amino acid residues) are also conserved to a large extent. (3) Among the seven active site residues known to be important for neuraminidase activity (Langedijk et al., 1997), only one residue (residue 7) is conserved in HeV G protein, whereas six are conserved in HN proteins of paramyxoviruses and rubulaviruses and four are conserved in H proteins of morbilliviruses. However, conserved residues are found in the surrounding positions of most proposed active site residues. It has been proposed that the conserved hexapeptide NRKSCS located at 2L01 and 2S1 in paramyxovirus HN is close to the sialic acid binding site and forms part of the neuraminidase activity site (Jorgensen et al., 1987; Morrison and Portner, 1991; Mirza et al., 1994). Binding to sialic acid-containing molecules on the surface of cells initiates the process of infection by viruses in the paramyxovirus and rubulavirus genera (Lamb and Kolakofsky, 1996). A number of gangliosides that contain one or more sialic acids as N-acetylneuraminic acid attached to an internal or a terminal galactose have been identified as isoreceptors for Sendai virus (Markwell et al., 1986). HeV G protein contains only the last two residues of the putative sialic acid binding site (TIHHCS instead of NRKSCS), suggesting that HeV may not bind sialic acid. This is consistent with our experimental findings. Indeed, not only does the virus fail to demonstrate neuraminidase activity in vitro (Murray et al., 1995), but also the removal of sialic acid from susceptible cells by treatment with neuraminidase did not abrogate infection by HeV (B. T. Eaton, unpublished results). In summary, the results indicate that the HeV G mRNA resembles other HeV mRNAs in having long untranslated regions and that the HeV G protein has a longer cytoplasmic tail than other Paramyxovirinae members. Sequence and structural comparison revealed that although the overall primary sequence homology is low, the HeV G protein shared high structural similarities with the HN proteins of paramyxoviruses. Although the biological significance of this discovery is not clear at the present time, a high structural conservation with disposal of neuraminidase activity site residues for the HeV G protein is an interesting and unusual evolutionary step that merits further investigation. MATERIALS AND METHODS Cloning and sequencing The methods for purification of virus, cDNA library construction, and screening have been described previously (Wang et al., 1998). Briefly, HeV was purified by zonal centrifugation in sucrose gradients, and genomic RNA was isolated using standard methods. The Time- Saver cDNA synthesis kit (Pharmacia) was used to make total cDNA using random priming, followed by the addition of an EcoRI adaptor. cDNA fragments were cloned into pZEr0±1 vector (InVitrogen), resulting in a cDNA library of ;100,000 independent primary clones. A total of eight overlapping clones were isolated covering the entire coding region of the G gene. DNA sequencing was performed using the Sequenase Kit (U.S. Biochemicals), and the nucleotide at each position was sequenced at least twice, either by sequencing two overlapping clones or by sequencing the opposite strand of the same clone. The complete cDNA sequence was then confirmed by direct sequencing of PCR fragments derived from genomic RNA without cloning. Sequence analysis and alignment Primary sequence comparison and calculation of percentage sequence identities were done using the AlignPlus Program from Scientific & Educational Software (Durham, North Carolina). Multiple sequence alignments were performed using the Pileup program of the Genetic Computer group (Devereux et al., 1984). Comparative alignments were conducted between the HeV G protein and several paramyxovirus HN proteins and morbillivirus H proteins. A final alignment was conducted using the same format and covering the same regions as in Langedijk et al. (1997). Phylogenetic analysis was carried out using programs in the PHYLIP package (Felsenstein 1993). Distances were calculated with PROTDIST using the alignment of Figure 2 as input. ACKNOWLEDGMENTS We thank Kaylene Selleck and Nadia Mayfield for technical assistance. The genome sequencing work was supported in part by a grant from the National Health and Medical Research Council of Australia. REFERENCES Bousse, T., Takimoto, T., Gorman, W. L., Takahasi, T., and Portner, A. (1994) Regions on the hemagglutinin-neuraminidase proteins of human parainfluenza virus type-1 and Sendai virus important for membrane fusion. Virology 204, 506±514. Devereux, J. P., Haeberli, P., and Smithies, O. (1984). A comprehensive set of sequence analysis programs for the VAX. Nucleic Acids Res. 12, 387±395. Feldmann, H., and Klenk, H.-D. (1996). Marburg and Ebola viruses. Adv. Virus Res. 47, 1±53. Felsenstein, J. (1993). PHYLIP (Phylogeny Inference Package). Version 3.5c. Department of Genetics, University of Washington, Seattle (distributed by the author). Gould, A. R. (1996). Comparison of the deduced matrix and fusion protein sequences of equine morbillivirus with cognate genes of the Paramyxoviridae. Virus Res. 43, 17±31. Halpin, K., Young, P., and Field, H. (1996). Identification of likely natural hosts for equine morbillivirus. Commun. Dis. Intell. 20, 476. Jorgensen, E. D., Collins, P. L., and Lomedico, P. T. 9. (1987). Cloning and nucleotide sequence of Newcastle disease virus hemagglutininneuraminidase mRNA: Identification of a putative sialic acid binding site. Virology 156, 12±24. THE HENDRA VIRUS G PROTEIN Kozak, M. (1991). Structural features in eukaryotic mRNAs that modulate the initiation of translation. J. Biol. Chem. 266, 19867±19870. Lamb, R. A., and Kolakofsky, D. (1996). Paramyxoviridae: the viruses and their replication. In ªFields Virology,º 3rd ed. (B. N. Fields, et al., Eds.), pp 1177±1204. Philadelphia, Lippincott-Raven Publishers. Langedijk, J. P. M., Daus, F. J., and Van Oirschot, J. T. (1997). Sequence and structure alignment of Paramyxoviridae attachment proteins and discovery of enzymatic activity for a morbillivirus hemagglutinin. J. Virol. 71, 6155±6167. Markwell, M. A. K., Moss, J., Hom, B. E., Fishman, P. H., and Sevennerholm, L. (1986). Expression of gangliosides as receptors at the cell surface controls infection of NCTC 2071 cells by Sendai virus. Virology 155, 356±364. Mirza, A. M., Deng, R., and Iorio, R. M. (1994). Site-directed mutagenesis of a conserved hexapeptide in the paramyxovirus hemagglutininneuraminidase glycoprotein: Effects on antigenic structure and function. J. Virol. 68, 5093±5099. Morrison, T., and Portner, A. (1991). Structure, function, and intracellular processing of the glycoproteins of Paramyxoviridae. In ªThe Paramyxovirusesº (D. W. Kingsbury, Ed.), pp. 347±382. Plenum, New York. Murphy, F. A., Fauques, C. M., Bishop, D. H. L., Ghabrial, S. A., Jarvis, A. W., Martelli, G. P., Mayo, M. A., and Summers, M. D. (Eds.). (1995). ªVirus Taxonomy: Classification and Nomenclature of Virus,º pp. 268±274, Spring-Verlag, New York. Murray, K., Selleck, P., Hooper, P., Hyatt, A., Gould, A. R., Gleeson, L., Westbury, H., Hiley, L., Selvey, L., Rodwell, B., and Ketterer, P. (1995). A morbillivirus that caused fatal disease in horses and humans. Science 268, 94±97. Murray, P. K., Eaton, B. T., Hooper, P., Wang, L.-F., Williamson, M., and Young, P. (1998). Flying foxes, horses and humans: A zoonosis caused by a new member of the Paramyxoviridae. In ªEmerging Infectionsº (W. M. Schield, D., Armstrong, and J. M. Hughes, Eds.), pp. 43±58. ASM Press, Washington, DC. 233 O'Sullivan, J. D., Allworth, A. M., Paterson, D. L., Samaratunga, H., Boots, R., Gleeson, L., Snow, T., and Bradfield, J. (1996). Fatal encephalitis due to a novel morbillivirus transmitted from horses. Lancet 349, 93±95. Palese, P., Tobita, K., Ueda, M., and Compans, R. W. (1974). Characterisation of temperature-sensitive influenza virus mutants defective in neuraminidase. Virology 61, 397±410. Parks, G. D., and Lamb, R. A. (1990). Defective assembly and intracellular transport of mutant paramyxovirus haemagglutinin-neuraminidase proteins containing altered cytoplasmic domains. J. Virol. 64, 3605±3616. Ropberts, S. R., Lichtenstein, D., Ball, L. A., and Wertz, G. W. 9. (1994). The membrane-associated and secreted forms of the respiratory syncytial virus attachment glycoprotein G are synthesized from alternative initiation codons. J. Virol. 68, 4538±4546. Sergel, T., McGinnes, L. W., Peeples, M. E., and Morrison, T. G. (1993). The attachment function of the Newcastle disease virus hemagglutinin-neuraminidase protein can be separated from fusion promotion by mutation. Virology 193, 717±726. Wang, L.-F., Michalski, W., Yu, M., Pritchard, L. I., Crameri, G., Shiell, B., and Eaton, B. T. (1998). A novel P/V/C gene in a new Paramyxoviridae virus which causes lethal infection in humans, horses and other animals. J. Virol. 72, 1482±1490. Westbury, H. A., Hooper, P. T., Brouwer, S. L., and Selleck, P. W. (1996). Susceptibility of cats to equine morbillivirus. Aust. Vet. J. 74, 132±134. Westbury, H. A., Hooper, P. T., Selleck, P. W., and Murray, P. K. (1995). Equine morbillivirus pneumonia: Susceptibility of laboratory animals to the virus. Aust. Vet. J. 72, 278±279. Yu, M., Hansson, E., Shiell, B., Michalski, W., Eaton, B. T., and Wang, L.-F. (1998). Sequence analysis of the Hendra virus nucleoprotein gene: Comparison with other members of the subfamily Paramyxovirinae. J. Gen. Virol. 79, 1775±1780.