Ó Springer-Verlag 2000
Med Microbiol Immunol (2000) 189: 1±6
ORIGINAL INVESTIGATION
Hiroshi Komada á Morihiro Ito á Machiko Nishio Mitsuo Kawano á Hisataka Ohta á Masato Tsurudome Shigeru Kusagawa á Myles O'Brien á Hisanori Bando Yasuhiko Ito
N-Glycosylation contributes to the limited cross-reactivity between hemagglutinin neuraminidase proteins of human parain¯uenza virus type 4A and 4B Received: 21 January 2000
Abstract cDNAs encoding human parain¯uenza virus type 4B (hPIV-4B) hemagglutinin neuraminidase (HN) protein were cloned and the nucleotide sequences were determined. A high degree of identity (81.4%) was observed between the nucleotide sequences of hPIV-4A and -4B HN proteins, and an 87.3% identity was found between the deduced amino acid sequences. This degree of identity is considered to be greater than immunological similarity between hPIV-4A and -4B HN proteins determined using monoclonal antibodies. To elucidate the causes of the antigenic dierence between HN proteins of hPIV-4A and -4B, we constructed three cDNAs of hPIV-4B HN whose potential N-glycosylation sites were partially or completely the same as in hPIV-4A HN cDNA. We compared the antigenicity of the expressed wild-type and mutant proteins, and found that the antigenicities of the mutant hPIV-4B HN proteins were more similar to the hPIV-4A HN protein than to the
H. Komada á M. Ito á M. Nishio á M. Kawano M. Tsurudome á S. Kusagawa á Y. Ito (&) Department of Microbiology, Mie University School of Medicine, Tsu, Mie 514-8507, Japan e-mail:
[email protected]; Fax: +81-59-2315008 H. Komada Department of Microbiology, Suzuka University of Medical Science, Suzuka, Mie 510-0293, Japan H. Ohta Department of Biology, Faculty of Science, Science University of Tokyo, Noda, Chiba, 278-8510, Japan M. O'Brien Mie Prefectural College of Nursing, Tsu, Mie 514-0116, Japan H. Bando Department of Applied Bioscience, Faculty of Agriculture, Hokkaido University, Sapporo, Hokkaido 060-8589, Japan
non-mutant hPIV-4B HN protein. This study indicated that the antigenic diversity between hPIV-4A and -4B was partly caused by deletion or creation of glycosylation sites, showing that the point mutations resulting in deletion or creation of glycosylation sites is one of the initial steps leading to the division of virus into subtypes. Key words Human parain¯uenza virus type 4 hemagglutinin neuraminidase protein á N-Glycosylation site á Antigenicity
Introduction Human parain¯uenza viruses which belong to the Paramyxovirus genus of the family Paramyxoviridae are known to be important pathogens in respiratory tract disease of infants and children. The viruses are divided into four types (hPIV-1, 2, 3 and 4), and type 4 is further subdivided into 4A and 4B. hPIV-4 was ®rst isolated and characterized by Johnson et al. [9], and antigenic heterogeneity was observed using polyclonal antibodies [2]. A large number of monoclonal antibodies (mAbs) directed against hPIV-4 have been produced [11], the structural polypeptides have been determined, and the antigenicity between the two subtypes 4A and 4B has been compared. In addition, we have cloned the NP [14], P [15], M [16] and F [13] genes of both hPIV-4A and -4B, and the hemagglutinin neuraminidase (HN) gene [1] of hPIV-4A. However, the gene structures of the HN of hPIV-4B and of the L proteins of hPIV-4A and -4B have not yet been determined. Two envelope glycoproteins of hPIV, F and HN, have important roles in virus replication. F mediates the fusion between the viral envelope and the cell membrane (fusion without) and between cell membranes (fusion within). HN is bi-functional, i.e., it has both hemagglutinin and neuraminidase activities, and is involved in the ®rst and the last steps of viral infection. In addition, envelope glycoproteins are the only antigens that have been shown to induce antibodies which neutralize
2
infectivity. Sequence analyses of HN proteins of hPIV-1 [5], Sendai virus (SV) [19], hPIV-3 [4], bovine PIV-3 (bPIV-3) [26], and Newcastle disease virus (NDV) [18] have shown that there are signi®cant amino acid similarities between the HN proteins. However, antigenic comparison of hPIV-1 and SV HN proteins using mAbs against hPIV-1 [12] and mAbs against SV [5, 12] indicated limited conservation of epitopes. Signi®cant antigenic diversity between HN proteins of hPIV-3 and bPIV-3 was also reported [22, 23]. Furthermore, we previously found that there was some antigenic diversity between HN proteins of hPIV-4A and -4B [11]. In the present study, the nucleotide sequence of hPIV-4B HN was determined and the structural features were compared with the HN proteins of hPIV-4A. NGlycosylation sites which were speci®c for hPIV-4B HN were changed by point mutations into those which were speci®c for hPIV-4A HN, the cDNAs were constitutively expressed in L929 cells, and the antigenic characteristics were determined by ¯ow cytometry with the use of mAbs directed against HN proteins of hPIV-4A and -4B.
Materials and methods Cells Primary monkey kidney cells and L929 cells were cultivated in Eagle's minimum essential medium (MEM) containing 5% fetal calf serum. The Toshiba strain (M-25) of hPIV-4A and the 68-333 strain of hPIV-4B were used in this study. Cloning and sequencing of hPIV-4B HN gene The cDNA libraries against nucleocapsid RNA and the mRNA of hPIV-4B obtained from the virus-infected primary monkey kidney cells were constructed according to Okayama-Berg's [21] or Gubler-Homann's [7] method. Using a DNA fragment derived from the recombinant plasmid pG4HF1 containing the sequence upstream from the hPIV-4A HN gene [1] as a probe, the libraries were screened and the clones containing a fragment of HN gene were sequenced by dideoxy chain termination method [25] using [a-35S]dATP (Dupont, NEN research Product, Boston, Mass.). Sequence analyses were aided by SDC-GENETYX programs. Construction of expression vector plasmids carrying HN genes The hPIV-4B HN expression plasmid pcDS-4BHN1 was constructed by polymerase chain reaction (PCR). Two cDNA clones which were from the ®rst and the last half of HN gene were ampli®ed using two pairs of primers which contained PstI and DraI, and DraI and EcoRI sites, respectively. Each cDNA was ®rst tailed with dG, and subcloned into the dC tailed pBluescript. The cDNA inserts were excised by digestion with PstI and DraI, and DraI and EcoRI, and then inserted between PstI and EcoRI sites of pcDLSRa296. The orientation of the inserted sequence was also analyzed using the restriction enzymes. pcDS-4BHN1 was digested with SalI, and inserted between SalI sites of pKan-2 where the gene of resistance to G418 is located. pcDL-SRa296 and pKan-2 were donated kindly by Dr. Takebe, NIH, Japan [27]. Three mutant cDNAs of hPIV-4B HN were made by using a point mutation kit according to the manufacturer's instructions (Clontech, Calif.).
Transfection of L929 cells with the plasmids The plasmid DNA transfection was performed using lipofectin reagent according to the method described previously [20]. Establishment of L929 cells that constitutively express wild-type or mutant hPIV-4B HN protein L929 cells were transfected with expression vector plasmid carrying wild-type or mutant hPIV-4B HN gene. For establishment of L929 cell lines stably expressing HN protein, the transfected cells were cultured in soft agar containing G418 (Geneticin; Life Technologies). To con®rm the expression of the proteins, ELISA and FACS analysis were performed. Flow cytometry The cells (5 ´ 105) were incubated with various mAbs for 1 h and were then reacted with FITC-conjugated anti-mouse Igs goat serum. Immuno¯uorescence-stained cells were analyzed on an FACScan (Becton Dickinson) using Consort 30 software (Becton Dickinson) [20]. The nucleotide sequence data reported here will appear in the DDBJ, EMBL and GenBank nucleotide sequence databases with the following accession number: AB006958.
Results Gene structure and predicted amino acid sequence of hPIV-4B HN protein The nucleotide sequence of hPIV-4B HN gene is presented in the positive sense DNA. The overall sequence of HN gene consisted of 2,538 nucleotides excluding the poly A tract. The sequence contained a large open reading frame (ORF) of 1,725 nucleotides with 575 codons. The ®rst initiation codon ATG found at nucleotides 288± 290 was in the most favored codons [17] and the termination codon (TAG) exists at nucleotides 2,010±2,012. If this ORF is transcribed from the ®rst ATG to TAG, a protein of molecular weight 66,109 will be produced. The deduced amino acid sequence encoded is presented in Fig. 1. As shown in Fig. 1, the predicted sequence of the HN protein contains only one major hydrophobic anchor region at the N terminus and seven potential asparaginelinked glycosylation sites with two additional sites of less favorable potential caused by the presence of proline. Bando et al. [1] found extraordinarily long noncoding regions in the hPIV-4A HN gene, which were about 300 and 500 nucleotides at its 5¢ and 3¢ ends, respectively. We also found similar long noncoding regions in the hPIV4B HN gene. There was an ORF in the 3¢ end noncoding region of hPIV-4B HN gene, corresponding to nucleotide 2,367±2,535 in phase with HN, which has the potential to code 56 amino acids. Comparison of the nucleotide and amino acid sequences between hPIV-4A and -4B HN proteins There is a high degree of identity (81.4%) in the nucleotide sequences of hPIV-4A and -4B HN proteins, and a
3 Fig. 1 Comparison of the sequence between the deduced HN proteins of hPIV-4A and -4B. Potential N-glycosylation sites are boxed. Cysteine (#) and proline (+) residues are shown, and single strong hydrophobic domain is underlined. Arrow indicates the position of residues which are conserved in all paramyxovirus HNs (D and Y) other than hPIV-4 HNs (N and F). The numbers (300) and (301) are for the alignment of paramyxovirus HNs [1] (HN hemagglutinin neuraminidase, hPIV human parain¯uenza virus)
high degree of identity (87.3%) is observed in the deduced amino acid sequences (Fig. 1). This degree of identity is considered to be greater than the immunological similarity between hPIV-4A and -4B HN proteins determined using the mAbs [11]. There are two major dierences in N-glycosylation sites between hPIV-4A HN and -4B HN proteins, at
residues 126±128, NRS (4A) and SHR (4B), and at residues 411±413, NIF (4A) and NIS (4B). Among amino acids which are important determinants of the secondary structure of protein, all 16 cysteines are conserved between hPIV-4A and -4B HN proteins (Fig. 1). However, there are some dierences in the position of proline residues (Fig. 1). Hydropathy
4
pro®les of both proteins are almost identical (data not shown). Subsequently, a phylogenetic tree was constructed for hPIV-1 [5], hPIV-3 [4], Sendai virus (SV) [19], Newcastle disease virus (NDV) [18], hPIV-2 [10], simian virus 5 (SV5 ) [8], mumps virus (MuV) [29], hPIV-4A [1] and hPIV-4B HN proteins using the neighbor joining method [24] (Fig. 2), showing that the division of hPIV-4 into subtypes occurred more recently as compared with the division into SV and hPIV-1 or into hPIV-3 and bPIV-3. Eect of N-glycosylation on the antigenicity of expressed hPIV-4B HN proteins To elucidate the causes of the antigenic dierence between hPIV-4A and -4B HN proteins, point mutations were introduced into the hPIV-4B HN cDNA, and three cDNAs whose N-glycosylation sites were similar to hPIV-4A HN cDNA were obtained. The ®rst mutant cDNA has NRS instead of SHR at amino acid residues 126±128 (Fig. 1 ), the second has NIF instead of NIS at residues 411±413 (Fig. 1), and the third has both NRS at residues 126±128 and NIF at residues 411±413 (Fig. 1). Thereafter, each of the three mutant hPIV-4B HN cDNA was transfected into L929 cells, and L929 cell clones which expressed the mutant hPIV-4B HN proteins were selected by ¯ow cytometry. The antigenicities of the three HN proteins were analyzed by ¯ow cytometry with the use of L929 cells stably expressing each mutant HN protein and mAbs against HN proteins of hPIV-4A and/or -4B. Table 1 shows the reactivities of various mAbs to the expressed mutant proteins. The antigenicities of expressed HN proteins were the same as those of virus HN proteins of hPIV-4A and -4B. Out of 11 mAbs, 2 (A028 and A158) against hPIV-4A HN reacted with the expressed hPIV-4B HN protein; the same 2 mAbs reacted with the ®rst mutant hPIV-4B HN protein (originally at that point there was no N-glycosylation site, but there is in the mutant protein); 3 mAbs (A140, A028 and A158) reacted with the second mutant hPIV-4B HN protein (originally at that point there was an N-glycosylation site, but the mutant
Fig. 2 Phylogenetic tree for the HN proteins of paramyxoviruses (SV Sendai virus, NDV Newcastle disease virus, MuV mumps virus)
protein does not have it), indicating that the site at 411± 413 is antigenically more important than the site at 126± 128. Furthermore, the third mutant of hPIV-4B HN protein whose N-glycosylation sites were the same as hPIV-4A HN protein showed reactivities to 5 mAbs (A138, A139, A140, A028 and A158) against hPIV-4A HN protein. The 3 mutant proteins lacked reactivity to mAb B23, which reacted with virus HN proteins of hPIV-4A. These results indicate that N-glycosylation of HN protein is closely related with the immunological relationship between hPIV-4A HN and -4B HN proteins and that factor(s) other than N-glycosylation determined the antigenicities of hPIV-4 HN proteins.
Discussion The nucleotide sequence of HN glycoprotein gene of hPIV-4B was determined. The neuraminidase activity of hPIV-4B is as low as that of hPIV-4A (data not shown). The important part of the catalytic site of neuraminidase consists of two amino acids, aspartic acid (at residue 300) and tyrosine (at residue 301) [1]. Aspartic acid and tyrosine are conserved among all the paramyxoviruses other than hPIV-4, and the two are located within the active site of neuraminidase. The ®rst amino acid of the two, aspartic acid, is also conserved in in¯uenza virus NA proteins and it is one of the 18 residues which de®ne the pocket and the rim of the active site of neuraminidase [3]. When hPIV-3-infected cells were incubated with anti-hPIV-3 mAbs inhibiting neuraminidase activity, variants with substitutions at the neighboring residues 299 and 302 were selected [28]. Both hPIV-4A and -4B HN proteins, however, have asparagine (at 300) and phenylalanine (at 301). These results suggested that the two amino acids, asparagine and phenylalanine, are located within the active site of neuraminidase of hPIV-4 A and -4B HN proteins, and that this substitution is related to a low neuraminidase activity. The results of the sequence analysis showed that the HN proteins of hPIV-4A and -4B are quite similar in structure and function. However, they can be distinguished immunologically, and why only hPIV-4 of the
5 Table 1 Antigenic analysis of the three mutant hPIV-4B HN proteins. Values are given as: ), 0±25% of the maximum of the dierence in mean intensity ratio between sample and control; +, 25±50% of the maximum of the dierence in mean intensity ratio between sample and control; ++, 50% or more of the maximum of the dierence in mean intensity ratio between sample and control. mAb
A 138 A 141 A 145 A 146 A 152 A 156 A 139 A 140 A 147 A 028 A 158 B 23 B 24 B 28 B 32 B 30 B 31 a
hPIV-4A HN
++ ++ ++ ++ ++ + ++ ++ ++ ++ + + ) ) ) + )
hPIV-4B HN
) ) ) ) ) ) ) ) ) ++ + + + + + ++ ++
First mutant
) ) ) ) ) ) ) ) ) ++ + ) ) ) ) ++ )
Values for hPIV-4A: average of two dierent clones. For the other columns: average of two dierent experiments on one clone, and of one experiment on another clone. Control cells were stained only with FITC-conjugated secondary mAb (hPIV human parain¯uenza virus, HN hemagglutinin neuraminidase, mAb monoclonal antibody)
Second mutant
) ) ) ) ) ) ) + ) ++ + ) ) ) ) ++ )
Third mutant
+ ) ) ) ) ) + + ) ++ + ) ) ) ) ++ )
Previously studied strainsa hIV-4A (M-25)
hPIV-4B (68-333)
+ + + + + + + + + + + + ) ) ) + )
) ) ) ) ) ) ) ) ) + + + + + + + +
Reactivities determined by ELISA in the previous study [11]
four types of hPIV can be divided into two subtypes has been an interesting problem. Immunological analysis using mAbs revealed that antigenic dierence of HN proteins is greater than that of the other proteins [11]. The present investigation show that the potential N-glycosylation sites of hPIV-4B HN protein were somewhat dierent from those of hPIV-4A HN protein. Consequently, to analyze the eect of glycosylation on the antigenic characteristics, we constructed three cDNAs whose N-glycosylation sites were similar to 4A HN cDNA, and compared the antigenicity of the expressed proteins. The mutant hPIV-4B HN protein, which has the same glycosylation sites as hPIV-4A HN, had quite similar antigenicity to hPIV-4A HN protein, showing that the N-glycosylation sites are very important for the antigenicity of hPIV-4 HN proteins. The HN protein of hPIV-1 is closely related antigenically to that of SV. Antigenic analysis, however, with anti-SV mAbs [5] and with anti-hPIV-1 mAbs [12], demonstrated that these HN proteins can be distinguished. Gorman et al. [6] reported that removal of carbohydrate moieties did not change epitope recognition of anti-hPIV-1 HN mAbs or cross-reactive anti-SV mAbs, indicating that carbohydrate moieties do not contribute to the antigenic properties of SV HN and hPIV-1 HN proteins. On the other hand, our present results show that N-glycosylation is responsible for the antigenic dierences between hPIV-4A and -4B HN proteins, and that the reason why hPIV-4A and -4B were identi®ed as dierent subtypes may come partly from the dierence of N-glycosylation sites on the HN proteins. However, all the antigenic dierence in hPIV-4 HN
protein could not be explained by N-glycosylation only, because some mAbs lacked responsiveness to the third mutant hPIV-4B HN protein whose N-glycosylation site are the same as in hPIV-4A HN protein and one hPIV4B HN-speci®c mAb, which had reactivity with hPIV4A HN protein, did not recognize the mutant hPIV-4B HN proteins whose N-glycosylation sites were similar to hPIV-4A HN protein. This indicates that factor(s) other than carbohydrate moieties is involved in the antigenic dierence. Although the positions of cysteines are the same in hPIV-4A and -4B HN proteins, those of proline are dierent; one possibility is that the dierences of proline sites may in¯uence antigenicities of hPIV-4B. The ®ndings that newly created antigenicity of mutant HN proteins is not necessarily strong showed that factor(s) other than carbohydrate moieties is related to the degree of immunoreactivity. In parain¯uenza viruses, subtypes can be found only in hPIV-4. Since division of hPIV-4 into subtypes occurred very recently as compared with division into SV and hPIV-1 or into hPIV-3 and bPIV-3, information on the molecular basis of subdivision into hPIV-4 subtypes is of use for clari®cation of the initial step of molecular evolution of the virus. This study shows that the antigenic diversity between hPIV-4A and -4B was partly caused by deletion or creation of glycosylation sites, indicating that point mutations resulting in deletion or creation of glycosylation sites is one of the initial steps leading to the division of virus into subtypes. Acknowledgement This work was supported in part by a grant from the Ministry of Education, Science, Culture and Sports of Japan.
6
References 1. Bando H, Kondo K, Kawano M, Komada H, Tsurudome M, Nishio M, Ito Y (1990) Molecular cloning and sequence analysis of human parain¯uenza type 4A virus HN gene: its irregularities in structure and activities. Virology 175: 307±312 2. Canchola J, Vargosco AJ, Kim HW, Parrot RH, Chritmas E, Jeries B, Chanock RM (1964) Antigenic variation among newly isolated strains of parain¯uenza type 4 virus. Am J Hyg 79: 357±364 3. Colman PM, Varghese JN, Laver WG (1983) Structure of the catalytic and antigenic sites in in¯uenza virus neuraminidase. Nature 303: 41±44 4. Elango N, Coligan JE, Jambou RC, Venkatsen S (1986) Human parain¯uenza type 3 virus hemagglutinin-neuraminidase glycoprotein: nucleotide sequence of mRNA and limited amino acid sequence of the puri®ed protein. J Virol 57: 481±489 5. Gorman WL, Gill DS, Scroggs RA, Portner A (1990) The hemagglutinin-neuraminidase glycoproteins of human parain¯uenza virus type 1 and Sendai virus have high structurefunction similarities with limited antigenic cross-reactivities. Virology 176: 211±221 6. Gorman WL, Pridgen C, Portner A (1991) Glycosylation of hemaggulutinin-neuraminidase glycoprotein of human parain¯uenza virus type 1 aects its functional but not its antigenic properties. Virology 183: 83±90 7. Gubler U, Homann BJ (1983) A simple and very ecient method for generating cDNA libraries. Gene 25: 263±269 8. Hiebelt SW, Paterson RC, Lamb RA (1985) Identi®cation and predicted sequence of a previously unrecognized small hydrophobic protein, SH, of the paramyxovirus simian virus 5. J Virol 55: 744±751 9. Johnson KM, Chanock RM, Cook MK, Huebner RJ (1960) Studies of a new human hemadsorption virus I. Isolation, properties and characterization. Am J Hyg 71: 81±92 10. Kawano M, Bando H, Yuasa T, Kondo K, Tsurudome M, Komada H, Nishio M, Ito Y (1990) Sequence determination of the hemaggulutinin-neuraminidase (HN) gene of human parain¯uenza type 2 virus and the construction of a phylogenetic tree for HN proteins of all the paramyxoviruses that are infectious to humans. Virology 174: 308±313 11. Komada H, Tsurudome, Ueda M, Nishio M, Bando H, Ito Y (1989) Isolation and characterization of monoclonal antibodies to human parain¯uenza virus type 4 and their use in revealing the antigenic relationship between subtype 4A and 4B. Virology 171: 27±37 12. Komada H, Kusagawa S, OÈrvell C, Tsurudome M, Nishio M, Bando H, Kawano M, Matsumura H, Norrby E, Ito Y (1992) Antigenic diversity of human parain¯uenza virus type 1 isolates and their immunological relationship with Sendai virus by using monoclonal antibodies. J Gen Virol 73: 875±884 13. Komada H, Bando H, Ito M, Ohta H, Kawano M, Nishio M, Tsurudome M, Watanabe N, Ikemura N, Kusagawa S, Mao X, O'Brien M, Ito Y (1995) Sequence analysis of human parain¯uenza virus type 4A and 4B fusion proteins. J Gen Virol 76: 3205±3210 14. Kondo K, Bando H, Kawano M, Tsurudome M, Komada H, Nishio M, Ito Y (1990) Sequencing analysis and comparison of parain¯uenza virus type 4A and 4B NP protein genes. Virology 174: 27±37
15. Kondo K, Bando H, Tsurudome M, Kawano M, Nishio M, Ito Y (1990) Sequence analysis of the phosphoprotein (p) genes of human parain¯uenza type 4A and 4B viruses and RNA editing at transcript of p genes: the number of G residues added is imprecise. Virology 178: 321±326 16. Kondo K, Fujii M, Nakamura T, Bando H, Kawano M, Tsurudome M, Komada H, Kusagawa S, Nishio M, Ito Y (1991) Sequence characterization of matrix protein genes of parain¯uenza virus types 4A and 4B. J Gen Virol 72: 2283± 2287 17. Kozak M (1986) Point mutations de®ne a sequence ¯anking the ATG initiator codon that modulates transcription by eukaryotic ribosome. Cell 44: 283±292 18. McGinnes LW, Wilde A, Morrison TG (1987) Nucleotide sequence of the gene encoding the Newcastle disease virus hemagglutinin-neuraminidase protein and comparisons of paramyxovirus hemagglutinin-neuraminidase protein sequences. Virus Res 7: 187±202 19. Miura N, Nakatani Y, Ishiura M, Uchida T, Okada Y (1985) Molecular cloning of a full-length cDNA encoding the hemagglutinin-neuraminidase glycoprotein of Sendai Virus. FEBS Lett 188: 112±116 20. Nishio M, Tsurudome M, Komada H, Kawano M, Tabata N, Matsumura H, Ikemura N, Watanabe N, Ito Y (1994) Fusion properties of cells constitutively expressing human parain¯uenza virus type 4A haemaggulutinin-neuraminidase and fusion glycoproteins. J Gen Virol 75: 3517±3523 21. Okayama H, Berg P (1982) High eciency cloning of full length cDNAs. Mol Cell Biol 2: 161±170 22. Ray R, Compans RW (1986) Monoclonal antibodies reveal extensive antigenic dierence between the hemaggulutininneuraminidase glycoprotein of human and bovine parain¯uenza 3 viruses. Virology 148: 232±236 23. Rydbeck R, Love A, Orvell C, Norrby E (1987) Antigenic analysis of human and bovine parain¯uenza virus type 3 strains with monoclonal antibodies. J Gen Virol 68: 2135±2160 24. Saitou N, Nei M (1987) The neighbour-joining method: a new model for reconstituting phylogenetic trees. Mol Biol Evol 4: 406±425 25. Sanger F, Nicklen S, Coulson AR (1977) DNA sequencing with chain-terminating inhibitor. Proc Natl Acad Sci USA 74: 5463±5467 26. Suzu S, Sakai Y, Shioda T, Shibuta H (1987) Nucleotide sequence of the bovine parain¯uenza 3 virus genome: the genes of the F and HN glycoproteins. Nucleic Acids Res 15: 2945± 2958 27. Takebe Y, Seiki M, Fujisawa J, Hoy F, Yokota K, Arai K, Yoshida M, Arai N (1988) SRa promoter: an ecient and versatile mammalian cDNA expression system composed of the simian virus 40 early promoter and R-U5 segment of human T-cell leukemia virus type 1 long terminal repeat. Mol Cell Biol 8: 466±472 28. Van Wyke Colelingh KL, Winter CC, Jorgensen ED, Murphy BJ (1987) Antigenic and structural properties of hemagglutinin-neuraminidase glycoprotein of human parain¯uenza virus type 3: antibodies which inhibit infectivity, hemagglutination, and neuraminidase activities. J Virol 61: 1473±1477 29. Waxham MN, Aronowski J, Server AC, Wolinsky JS, Smith JA, Goodman HM (1988) Sequence determination of the mumps virus HN gene. Virology 164: 318±325
