Characterization of a Novel Influenza Hemagglutinin, H15: Criteria for Determination of Influenza A Subtypes

VIROLOGY

217, 508–516 (1996) 0145

ARTICLE NO.

Characterization of a Novel Influenza Hemagglutinin, H15: Criteria for Determination of Influenza A Subtypes ¨ SS,‡ JOHN MACKENZIE,§ and ROBERT G. WEBSTER*,Ø,1 ¨ HM,* NANNAN ZHOU,*,† JOCHEN SU CAROLIN RO *Department of Virology and Molecular Biology, St. Jude Children’s Research Hospital, 332 North Lauderdale, Memphis, Tennessee 38105; †Department of Microbiology, Jiangxi Medical College, 161 Bayi Road, Nanchang 330008, Jiangxi, China; ‡Bundesinstitut fu¨r gesundheitlichen Verbraucherschutz und Veterina¨rmedizin, Diedersdorfer Weg 1, 12277 Berlin, Germany; §Department of Microbiology, University of Queensland, St. Lucia, Brisbane, Queensland 4072, Australia; and ØDepartment of Pathology, University of Tennessee, Memphis, 800 North Madison, Memphis, Tennessee 38163 Received October 2, 1995; accepted January 12, 1996 Two viruses with a novel hemagglutinin (HA), A/duck/Australia/341/83 and A/shearwater/West Australia/2576/79, have been isolated from a duck and a shorebird in Australia. Hemagglutination inhibition and double immunodiffusion assays failed to reveal cross-reactivity with any of the known subtypes (H1 to H14). We therefore propose that these viruses constitute a new HA subtype, H15. Sequence analysis of the HA genes confirmed the serologic findings. When compared at the amino acid level, the HA1 region of the H15 subtype differs from those of the other subtypes by 30% and more. This degree of heterogeneity is also found among HA genes of other subtypes. Thus we propose that amino acid sequence data should be evaluated when determining the HA subtypes of influenza A viruses. Sequence comparison and phylogenetic analysis suggested that the HA subtype H15 is most closely related to the H7. Compared to the H7 HA, the H15 acquired a 30-nucleotide insertion within HA1 at position 253 which is located in the globular head of the molecule. This finding suggests that RNA recombination, although a rare event in nature, may play an important role in the evolution of influenza viruses. q 1996 Academic Press, Inc.

INTRODUCTION

coded by the HA1 subunit, which are much more variable than those in the HA2 region which forms large parts of the HA stalk. Thus Kawaoka et al. (1990) suggested that nucleotide and amino acid sequence information of the HA1 should also be evaluated when determining HA subtypes of influenza A viruses. During surveillance studies in Australia, influenza A viruses were isolated from various healthy waterbirds and shorebirds (Mackenzie et al., 1984). Two isolates could not successfully be assigned to any of the 14 known HA subtypes. We compared the antigenic and genetic characteristics of the HA from A/duck/Australia/ 341/83 (H?N8) and A/shearwater/West Australia/2576/79 (H?N9) to the known HA subtypes. Our results show that wild birds in Australia harbor influenza A viruses with HA molecules that are antigenically and genetically distinct from the 14 known HA subtypes. We therefore propose that the HA molecules of A/duck/Australia/341/83 and A/shearwater/West Australia/2576/79 belong to a new subtype, H15.

Influenza A viruses have been isolated from many avian species worldwide (Slemons et al., 1974; Hinshaw et al., 1980; Su¨ss et al., 1994). Wild aquatic birds are considered the primordial reservoir of influenza A viruses, because this population harbors the entire spectrum of subtypes. According to the current classification system, these viruses are assigned to specific subtypes based on the antigenicity of the major surface glycoproteins, hemagglutinin (HA) and neuraminidase (NA) (WHO Memorandum, 1980). Fourteen HA and nine NA subtypes have previously been identified based on reactivity in double immunodiffusion assays (WHO Memorandum, 1980; Hinshaw et al., 1982; Kawaoka et al., 1990). This technique distinguishes between viruses of different HA subtypes, but is less sensitive in differentiating viruses within the same subtype (Schild et al., 1980). The molecular basis for different antigenic reactivities of the HAs are differences in the amino acid sequences. Wiley et al. (1981) showed that the antigenic sites for H3 HA molecules are located on the globular head of the HA. This portion of the HA comprises amino acids en-

MATERIALS AND METHODS Viruses A/duck/Australia/341/83 (H?N8; DkAU83) and A/shearwater/West Australia/2576/79 (H?N9; ShWA79) were isolated from cloacal swabs of healthy birds in Australia

1 To whom reprint requests should be addressed at Department of Virology and Molecular Biology, St. Jude Children’s Research Hospital, 332 North Lauderdale, Memphis, TN 38105.

0042-6822/96 $18.00

508

Copyright q 1996 by Academic Press, Inc. All rights of reproduction in any form reserved.

AID

VY 7786

/

6a12$$$341

02-12-96 13:35:11

vira

AP: Virology

NEW INFLUENZA HEMAGGLUTININ IN AUSTRALIA

and were provided by the Australian Animal Health Laboratory, Commonwealth Scientific and Industrial Research Organization, Victoria, Australia. The viruses were cultivated in 11-day-old embryonated chicken eggs. The NA antigens of these viruses were identified as N8 and N9 for DkAU83 and ShWA79, respectively. Preparation of reassortant virus A reassortant virus containing the H? from DkAU83 and the N1 from A/Bellamy/42 (H1N1) (Bel) was prepared as described by Webster (1970). Serological studies Hemagglutination inhibition tests (HI) were performed in microtiter plates with receptor-destroying enzyme-treated antisera (WHO Collaborating Center for Reference and Research on Influenza, 1982). To prepare monospecific antiserum against DkAU83 HA, the DkAU83 (H?)-Bel(N1) reassortant was grown in 11-day-old embryonated chicken eggs. The reassortant present in the allantoic fluid was concentrated with a hollow fiber concentrator (Amicon, Beverly, MA) and purified by differential sedimentation through a 25–70% sucrose gradient. Disruption of the reassortant with sodium dodecyl sulfate (SDS) and electrophoretic separation of the proteins on cellulose acetate (Laver, 1964; Laver and Webster, 1972) were not successful, because the HA of DkAU83 was not stable in 1% SDS (w/v). Thus we used an alternative method. The purified reassortant virus was disrupted by octylglucoside (7.5%) (Johansson et al., 1989). HA and NA were separated from the virus particles by centrifugation at 21,000 g and adsorbed to chicken red blood cells (CRBC). These were washed thoroughly with PBS and tested for residual NA or NP antigen by ELISA with negative results. The CRBC containing 10,000 HAUs of DkAU83 were reinjected intravenously into the chickens from which they were harvested. Intravenous injections were repeated four times (the first booster was given 10 days after the initial injection, and the second through fourth boosts were given 4–5 days after the previous injections), and antiserum was collected 3 days after the last injection. A neuraminidase inhibition assay verified that the antiserum was negative for NA antibodies. The viruses used for preparation of reference antisera are listed in Table 1. Monospecific hyperimmune antisera to each of the HA subtypes were produced in goats as previously described (Webster et al., 1974; Schild et al., 1980). The double immunodiffusion assay was performed as reported by Schild et al. (1980). Viruses were grown, concentrated, and purified as described above. Purified virus was disrupted by 1% sodium lauroyl sarcosinate. The same antisera were used as for the HI assays.

509 TABLE 1

Viruses Used for Preparation of Antisera and Sequence Analysis

Subtype

Virus

GenBank Accession No.

Viruses used for preparation of antisera H1 H1 H1 H1 H2 H2 H3 H3 H3 H4 H5 H6 H6 H7 H7 H8 H9 H10 H11 H11 H12 H13 H14

A/Puerto Rico/8/34 (H1N1) A/Fort Monmouth/1/47 (H1N1) A/swine/Iowa/15/30 (H1N1) A/duck/Alberta/35/76 (H1N1) A/Singapore/1/57 (H2N2) A/mallard/NY/6750/78 (H2N2) A/Hong Kong/1/68 (H3N2) A/equine/Miami/1/63 (H3N8) A/duck/Ukraine/1/63 (H3N8) A/duck/Czechoslovakia/56 (H4N6) A/tern/South Africa/61 (H5N3) A/turkey/Massachusetts/3740/65 (H6N2) A/shearwater/Australia/1/72 (H6N5) A/equine/Prague/1/56 (H7N7) A/FPV/Rostock/34 (H7N1) A/turkey/Ontario/6118/68 (H8N4) A/turkey/Wisconsin/1/66 (H9N2) A/chicken/Germany/N/49 (H10N7) A/duck/England/56 (H11N6) A/duck/Memphis/546/74 (H11N9) A/duck/Alberta/60/76 (H12N6) A/gull/Maryland/704/77 (H13N6) A/mallard/Gurjev/263/82 (H14N5)

Viruses used for sequence homologies and phylogenetic analysis H1 H2 H3 H4 H5 H6 H7 H7 H7 H7 H7 H7 H7 H7 H8 H9 H10 H11 H12 H13 H14

A/duck/Alberta/35/76 (H1N1) A/pintail/Praimoric/625/76 (H2N2) A/duck/Ukraine/1/63 (H3N8) A/duck/Czechoslovakia/56 (H4N6) A/chicken/Pennsylvania/1/83 (H5N2) A/shearwater/Australia/1/72 (H6N5) A/FPV/Rostock/34 (H7N1) A/turkey/Oregon/71 (H7N3) A/duck/Hong Kong/293/78 (H7N2) A/chicken/Leipzig/79 (H7N7) A/seal/Massachusetts/1/80 (H7N7) A/chicken/Victoria/1/85 (H7N7) A/equine/Prague/1/56 (H7N7) A/equine/London/1416/73 (H7N7) A/turkey/Ontario/6118/68 (H8N4) A/turkey/Wisconsin/1/66 (H9N2) A/chicken/Germany/N/49 (H10N7) A/duck/England/56 (H11N6) A/duck/Alberta/60/76 (H12N6) A/gull/Maryland/704/77 (H13N6) A/mallard/Gurjev/263/82 (H14N5)

D10477 L11141 J02109 D90302 J04325 D90303 M24457 M31689 U20461 U20459 K00429 M17735 X62552 M58657 D90304 D90305 M21646 D90306 D90307 D90308 M35997

RNA was extracted from virus-containing allantoic fluid by a previously described phenol–chloroform procedure

(Bean et al., 1980). Single-strand cDNA was synthesized with reverse transcriptase (Life Sciences, Inc., St. Petersburg, FL) and amplified by PCR with Taq DNA polymerase (Promega, Madison, WI) after the procedure of Katz et al. (1990). Oligonucleotide primers were synthesized by the Center of BioTechnology of St. Jude Children’s Research Hospital (Memphis, TN; primer sequences available upon request). The reaction was completed in a

AID

vira

Polymerase chain reaction (PCR) and sequencing

VY 7786

/

6a12$$$342

02-12-96 13:35:11

AP: Virology

¨ HM ET AL. RO

510

programmable thermal controller (MJ Research, Inc., Watertown, MA). Amplified DNA was purified with the Wizard PCR Prep kit (Promega, Madison, WI). The nucleotide sequences of the HA genes were determined by sequencing at least two PCR products of each RNA segment using the dideoxynucleotide chaintermination method (Sanger et al., 1977). Oligonucleotide primers complementary to the HA segment were endlabeled with [g-32P]ATP (ICN Pharmaceuticals Inc., Costa Mesa, CA), annealed to template DNA, and extended by Taq DNA polymerase with the fmol DNA sequencing system (Promega, Madison, WI) in a programmable thermal controller. Reaction products were separated on a 6% polyacrylamide/7.8 M urea gel containing a 1.0 to 5.01 TBE (90 mM Trisborate, pH 8.0, 2 mM EDTA) gradient. The first 24 and the last 20 nucleotides of both HA sequences were obtained from the PCR primer sequences. The sequence between nucleotides 800 and 880 of DkAU83 was also determined by direct RNA sequencing with the dideoxynucleotide chain-termination method. RNA extracted from purified virus was used as a template. Oligonucleotide primers were end-labeled with [g32 P]ATP, annealed to the RNA template, and extended by reverse transcriptase. Sequence analysis and phylogenetic analysis The IntelliGenetics suite, release 5.4 (IntelliGenetics, Inc., Mountain View, CA), was used to translate and align the nucleotide sequence data. The alignment of nucleotide sequences was done using the following parameters: amino acid residue length, 2; deletion weight, 5; length factor, 0; matching weight, 1; nucleic acid residue length, 4; spread factor, 50. Phylogenetic analysis by the maximum parsimony method determined the minimum number of mutations needed to account for the sequence differences (Fitch, 1971). Searches for the most parsimonious trees of the nucleotide sequences were done with PAUP software, version 2.4 (David L. Swofford, Illinois Natural History Survey, Champaign, IL), using the MULPARS and global swap options. In this system, tree length is measured in steps that are equivalent to nucleotide changes. The total tree length is the sum of all branch lengths.

(H7N1), which were replaced by A/Hong Kong/1/68 (H3N2), A/chicken/Mexico/31381-Avilab/94 (H5N2), and A/ruddy turnstone/New Jersey/65/85 (H7N3), respectively. DkAU83 and ShWA79 did not react with any of the reference antisera (HI titers of less than 40 were considered negative, because of nonspecific inhibition of hemagglutination by the hyperimmune goat antisera against subtypes H1 to H14 of up to a titer of 20), nor did the monospecific antisera against the HA of DkAU83 react with the HA of any reference virus (Table 2). Hemagglutination by ShWA79 was inhibited with a-DkAU83 antiserum at an HI titer of 20. In the case of the chicken antiserum against DkAU83, which did not produce nonspecific inhibition of hemagglutination at dilutions as low as 1:10, this value was considered positive, but it was much lower than that for DkAU83 (HI titer: 320). This observation suggests that although both viruses from Australia did react with the DkAU83 antiserum, they are antigenically quite different. Since the HI test does not always detect cross-reactions between different antigens which belong to the same HA subtype, we examined cross-reactivity in double immunodiffusion assays. Based on sequence analysis (see below) detailed comparisons in the double immunodiffusion assay (Schild et al., 1980) were only made between DkAU83 and its closest related subtype, H7, using avian as well as equine H7 viruses. The monospecific goat antisera to the reference H7 viruses produced strong lines of precipitation with their homologous viruses but no precipitation with DkAU83. The chicken antiserum against DkAU83 HA produced a single weak line of precipitation only with the homologous virus that could not be photographed, indicating that the antibodies that reacted well in HI tests were nonprecipitating. Thus this assay confirmed that DkAU83 is antigenically distinct from all known HA subtypes. Sequencing analysis

The antigenic relationship between the HAs of DkAU83, ShWA79, and other HA subtypes was determined through HI tests with monospecific antisera against DkAU83 and the HAs of reference viruses (Table 1). The viruses used in the HI were the same as for preparation of antisera except for A/Aichi/2/68 (H3N2), A/tern/South Africa/61 (H5N3), and A/FPV/Rostock/34

The sequences for the HA genes of DkAU83 and ShWA79 have been entered in the GenBank database and have been assigned Accession Nos. L43916 and L43917, respectively. Both genes are 1762 nucleotides in length, with a single open reading frame that spans position 22 to 1732. The deduced amino acid sequences of the mature HAs (after removal of signal sequences) were aligned with those of other HA subtypes (Table 1) using the Genalign program of the IntelliGenetics Suite, release 5.4. Amino acid insertions or deletions were counted as mismatches. The HA of DkAU83 has the highest homology with the H7 HA gene [77.5% overall homology with A/FPV/Rostock/34 (H7N1); Fig. 1A]. We also compared the homologies between DkAU83 and nine H7 HAs (H7 viruses used are listed in Table 1) with those from other closely related subtypes such as H4 and H14, or H2 and H5 (Fig. 1B). The overall

AID

vira

RESULTS Antigenic analysis

VY 7786

/

6a12$$$342

02-12-96 13:35:11

AP: Virology

NEW INFLUENZA HEMAGGLUTININ IN AUSTRALIA

511

TABLE 2 Comparison of the Putative H15 Subtype with Other HA Subtypes Hemagglutination inhibition assay Viruses Antiserum

DkAU83 (H?N8)

ShWA79 (H?N9)

H1–H14

H1–H14 DkAU83

õ40a 320

õ40 20

80–5120b õ10

Double immunodiffusion assay Virusesc

Antiserum

DkAU83 (H?N8)

EqPr56 (H7N7)

EqLo73 (H7N7)

TyOR71 (H7N3)

TyMN80 (H7N3)

SnPo81 (H7N7)

0 /

/e 0

/ 0

/ 0

/ 0

/ 0

H7 d DkAU83 a

HI titers are the reciprocal of the dilution inhibiting 4 HA of a virus. Each reference serum was tested with the homologous virus. c Abbreviations used: EqPr56, A/equine/Prague/1/56 (H7N7); EqLo73, A/equine/London/1416/73 (H7N7); TyOR71, A/turkey/Oregon/71 (H7N3); TyMN80, A/turkey/Minnesota/1237/80 (H7N3); SnPo81, A/swan/Potsdam/63/81 (H7N7). d H7 antiserum was prepared against the hemagglutinin of A/equine/Prague/1/56 (H7N7) and A/FPV/Rostock/34 (H7N1). Both sera were tested against DkAU83. In the case of the H7 antigens only equine or avian H7 antiserum was used. e /, a single line of precipitation was detected in double immunodiffusion assays. b

amino acid homology between DkAU83 and the H7 HAs ranged from 73.5 to 78.6% compared to 77.1% for H4 and H14, and 76.0% for H2 and H5 HA genes (Fig. 1B). In addition, we assessed the amino acid homologies of the HA subunits. Homologies between HA1s of DkAU83 and H7 viruses range between 67.5 and 71.3%, compared to 66.7% between H4 and H14, and 69.6% for the H2 and H5 subtypes (Fig. 1B). The homology between the HA1 regions of viruses within an established subtype exceeds 80% (Kawaoka et al., 1990; Nobusawa et al., 1991). The HA2 subunit, which forms most of the HA stalk, is very conserved among subtypes. In the case of DkAU83 and the H7 viruses, HA2 amino acid homology is between 82.8 and 90.0% and is lower than that found between the H4 and the H14 subtypes (Fig. 1B). This finding suggests that the DkAU83-like and H7 HAs might have diverged earlier than did the H4 and H14 subtypes, whose HA2 regions are 96.7% homologous. The HA genes from the Australian viruses encode an unprocessed precursor polypeptide of 570 amino acids (Fig. 2A). Assuming DkAU83 HA is processed by a signal peptidase as H7 HAs, we predict that the HA molecules of DkAU83 and ShWA79 consist of an 18-amino-acid signal peptide and the 552-amino-acid mature peptide. In contrast, the mature peptide for the HA of nonpathogenic H7 viruses is only 542 amino acids in length (Fig. 2A). This difference results from a 10-amino-acid insertion at position 253 in the HA1 of DkAU83 and ShWA79. Based on the threedimensional structure of the H3 HA (Wilson et al., 1981; Wiley et al., 1981), we conclude that this inserted sequence

is found within the globular head of the molecule, as are important antigenic epitopes in the HA of other subtypes. Because Khatchikian et al. (1989) and Orlich et al. (1994) found insertions in the influenza HA which originated from cellular and viral genes, respectively, we searched the databases with the 30 nucleotides encoding the additional amino acids. We failed to identify any sequence whose homology to the search string exceeded 72%. The HA genes of DkAU83 and ShWA79 have seven potential glycosylation sites (Asn-X-Ser/Thr; Struck et al., 1978; Fig. 2A). The four sites common to H7 and our Australian viruses (positions 12 and 28 in HA1 and positions 82 and 154 in HA2) are in the stem region of the HA. Two of them (position 12 in the HA1 subunit and position 154 in the HA2 subunit) are highly conserved among all influenza A subtypes. The locations of the other glycosylation sites within the HA1 subunit differ between the two Australian viruses (positions 82 and 156) and H7 viruses (positions 123 and 231). The connecting peptides between HA1 and HA2 of DkAU83 and ShWA79 consist of the same number of amino acids as those of nonpathogenic H7 viruses. They also feature two basic amino acids at the HA cleavage site, histidine at position 03 of the HA1 (substituted by an arginine in the HA of ShWA79) and arginine at position 01. However, both Australian viruses have isoleucine at position 04 (Fig. 2A) instead of proline, which is highly conserved among H7 viruses. There are many differences between the HA genes of DkAU83 and ShWA79, which account for the different

AID

vira

VY 7786

/

6a12$$$343

02-12-96 13:35:11

AP: Virology

¨ HM ET AL. RO

512

FIG. 1. (A) Homology at the amino acid level between the mature HA (signal peptide excluded) of DkAU83 and that of HA subtypes 1 through 14. The viruses are listed in Table 1. A/FPV/Rostock/34 (H7N1) represented the H7 viruses in this comparison. Insertions and deletions were interpreted as mismatches. (B) Homology between the amino acid sequences of HA0, HA1, and HA2 of DkAU83 and nine H7 viruses, compared to homologies between the closely related HA subtypes, H2 and H5, and H4 and H14. The viruses used in the analysis are listed in Table 1. Insertions and deletions were interpreted as mismatches.

antigenic reactivities in HI assays. A total of 50 nucleotide changes led to 20 amino acid substitutions (Fig. 2B)— 1 in the HA signal peptide, 17 in the HA1 subunit, and 2 in the HA2 region. Thus the HA genes are 96.56% homologous in their amino acid sequences. Three nucleotide substitutions are located in the sequence corresponding to the 10-amino-acid insertion in HA1, leading to a single amino acid change at position 259. The potential glycosylation sites are identical for both HA genes. Phylogenetic analysis To examine the phylogenetic relationship between DkAU83, ShWA79, and 26 H7 viruses, we aligned the sequences of the entire HA1 and the first 51 nucleotides of the HA2, because this region was available for all the sequences obtained from GenBank. The signal peptide, the insertions of the Australian viruses, and insertions at the cleavage site of H7 viruses were excluded. The tree was rooted to an H10 HA sequence because of the high homology between H7 and H10 subtypes (Air, 1981; Nobusawa et al., 1991). The shortest tree required 1772 steps to connect all 29 HA genes. This phylogenetic tree (Fig. 3) shows all H7 viruses clustered in one major branch that is separated into a North American avian, a Eurasian avian, and an equine lineage. DkAU83 and ShWA79 are located on a separate branch, suggesting that these two viruses shared only one common ancestor

AID

VY 7786

/

6a12$$$343

02-12-96 13:35:11

HA with the H7 subtype, before H7 HAs formed different lineages. Phylogenetic relationship to all the other subtypes was also determined. The HA nucleotide sequences of DkAU83 and the other 14 HA subtypes were aligned using conserved residues in the amino acid sequences as reference points (Nobusawa et al., 1991). Nucleotide sequences encoding the mature HA were used, and all insertions were excluded from the analysis. The shortest evolutionary path connecting all the subtypes required 5934 steps (Fig. 4). This phylogenetic tree confirms that the HA of DkAU83 is most closely related to H7 HA genes. However, there is a significant distance between DkAU83 and H7 viruses (458 steps), which is very similar to the distances between H4 and H14 (477 steps) and H2 and H5 (475 steps) viruses. DISCUSSION In this report, we describe antigenic and genetic characterization of a new HA from avian influenza A viruses from Australia. Based on results from serological assays and sequence analysis, we propose that these viruses constitute a new subtype, H15. HI assays and double immunodiffusion tests failed to demonstrate any crossreactivity with the 14 known HA subtypes. Molecular analysis of the HA genes supported the findings of the serological assays. The degree of amino acid homology

vira

AP: Virology

NEW INFLUENZA HEMAGGLUTININ IN AUSTRALIA

513

FIG. 2. (A) The amino acid sequence of the HA from DkAU83 compared to an H7 HA gene from a nonpathogenic avian virus (A/swan/Potsdam/ 63/81, GenBank Accession No. U20467). The areas corresponding to the signal peptide (SP), HA1, and HA2 are shown. Amino acid positions are indicated by the numbers below the genes; the symbols and amino acid position numbers above the diagram correspond to potential glycosylation sites. The black insert in the DkAU83 HA represents a 10-amino-acid insertion at position 253. The amino acid sequences of the HA cleavage sites of both genes are shown for comparison. (B) Amino acid differences between the HA genes of DkAU83 and ShWA79. Amino acid positions are indicated for the signal peptide (SP), HA1, and HA2.

for the HA1 of DkAU83 and different H7 HA1 genes (67.5– 71.3%) is similar to that usually seen between closely related but different subtypes (such as H2 and H5; Fig. 1B). In contrast, HA1 genes belonging to the same subtype have similarities higher than 80% (Kawaoka et al., 1990; Nobusawa et al., 1991). In addition, potential glycosylation sites in the globular head region of the H15 HA are located in different positions compared to H7 HA genes. Although isolated at the same time and in the same region, viruses with the new HA were found in different hosts (ducks and shorebirds; isolates with an HA identical to that of ShWA79 were also found in ducks, lesser noddies, and terns) and in combination with different NA subtypes (N8 and N9). The antigenicities of the HAs from DkAU83 and ShWA79 differ; ShWA79 had a lower HI titer with a-DkAU83 antiserum than did the homologous virus. In addition, the HAs of DkAU83 and ShWA79 are significantly different at the amino acid level, with a homology of only 96.56%. Although both viruses contained the 10amino-acid insertion, they had accumulated three nucleotide substitutions leading to one amino acid change within the insertion. These findings show that antigenically and genetically different H15 strains are circulating

in the aquatic bird population in Australia. With the presence of various H15 viruses, it is fair to assume that the H15 subtype emerged a long time ago and since then gained a considerable genetic diversity. Our phylogenetic analysis supported this idea. A common ancestral HA with the H7 subtype existed before H7 HAs were separated into host-specific and geographically distinct lineages. Thus H15 probably emerged at the same point in time as did the H7 subtype which later also established a geographically distinct sublineage in Australia. It is interesting to note that earlier studies with the H15 viruses had suggested that they represented aberrant H7 HAs (Mackenzie et al., 1984). So far, viruses with the H15 HA have only been isolated in Australia. There is some precedent for the special location of Australia in terms of influenza virus evolution. Gorman et al. (1990) showed that the evolution of avian NP genes follows a geographic pattern, with a separate avian lineage limited to Australia. Phylogenetic analysis has shown a similar geographic confinement for H7 HA genes (Ro¨hm et al., 1995). Thus, the isolated evolution of H15 in Australia is not unlikely and may account for the absence of this subtype among viruses from the rest of the world.

AID

vira

VY 7786

/

6a12$$$343

02-12-96 13:35:11

AP: Virology

¨ HM ET AL. RO

514

obtained a mutant of A/seal/Mass./1/80 (H7N7) with a 60-nucleotide insertion at the HA cleavage site which was due to intersegmental RNA recombination between the HA and the nucleoprotein gene of the same virus. Although we were unable to trace the 30-nucleotide insert of our Australian viruses to any viral or cellular gene, nonhomologous recombination between viral and cellular RNA may have led to the insertion. Nonhomologous recombination in influenza viruses has been described several times. Fields and Winter (1982) proposed that defective–interfering RNAs are derived by intersegment recombination. Bergmann et al. (1992) described a generation of influenza A viruses that resulted from RNA recombination events in a ribonucleoprotein transfection system. Since the HA in which the insertion occurred originally is unknown, speculation about the mechanism responsible for the initial recombination event, which might include a copy choice or a breakage-joining process (Lai, 1992), is difficult. That the 30-nucleotide insertion lies within the H15 HA gene, which is translated into a stable surface glycoprotein, emphasizes the possible importance of RNA recombination in creating genetic diversity among influenza viruses. In vitro RNA recombination has been reported for several RNA viruses. Members of the picornavirus and coronavirus families demonstrated RNA recombination between viral RNAs (Cooper, 1977; Makino et al., 1986). Bujarski and Kaesberg (1986) showed that RNA FIG. 3. Phylogenetic tree for the HA genes of DkAU83, ShWA79, and H7 influenza A viruses. Nucleotide positions 76 to 1038 were used for H7 HA genes, and 76 to 1048 for DkAU83 and ShWA79. All insertions were excluded prior to the analysis. The tree was rooted to an H10 HA sequence [A/chicken/Germany/N/49 (H10N7), GenBank Accession No. M21646]. The nucleotide tree required 1772 steps to join the 29 HA sequences. Horizontal distances are proportional to the number of nucleotide changes required to join nodes and HA sequences. Vertical lines are for spacing branches and labels. The arrow to the left indicates the direction of the H10 HA sequence from the root node. The HA sequences used in this analysis were obtained from GenBank or literature (Accession Nos.: K00429, M17735, M24457, M31689, M58657, U20458, U20459, U20461–U20471, X61627, X62552–X62554, X62556, X62557, and X62559; reference for A/chicken/Brescia/1902 (H7N7) and A/FPV/Weybridge/27 (H7N7) is Klimov et al. (1992).

The additional 10 amino acids in the HA1 region are an interesting feature of the H15 HA molecule. Although insertions of this length occur uncommonly among influenza A HAs, in vivo and in vitro examples have been reported. All equine H7 viruses have acquired a 10amino-acid insert at the HA cleavage site, compared to their avian counterparts (Gibson et al., 1992). Even longer insertions at the HA cleavage site were obtained by adapting H7 viruses to replication in chicken embryo cells. Khatchikian et al. (1989) described a mutant of A/turkey/Oregon/71 (H7N3) containing a 54-nucleotide insertion at the HA cleavage site that originated from a region of cellular 28 S ribosomal RNA. Orlich et al. (1994)

FIG. 4. Phylogenetic tree of HA genes of DkAU83 and viruses of different HA subtypes. The nucleotide sequences encoding the mature peptide were used for this analysis. All insertions were excluded prior to the analysis. The tree was rooted by its midpoint. The nucleotide tree required 5934 steps to join the 15 HA sequences. Horizontal distances are proportional to the number of nucleotide changes required to join nodes and HA sequences. Vertical lines are for spacing branches and labels. The HA sequences used in this analysis were obtained from GenBank and are listed in Table 1. A/turkey/Oregon/71 (H7N3) represented the H7 viruses.

AID

vira

VY 7786

/

6a12$$$343

02-12-96 13:35:11

AP: Virology

NEW INFLUENZA HEMAGGLUTININ IN AUSTRALIA

plant viruses (e.g., brome mosaic virus) may undergo RNA recombination when under selective pressure. RNA recombination has only infrequently been observed in vivo. One example is the western equine encephalitis virus, which arose by recombination between eastern equine encephalitis virus and a Sindbis-like virus (Hahn et al., 1988). Meyers et al. (1989) identified that recombination between viral and cellular RNAs led to a 228nucleotide insertion within the genome of a bovine viral diarrhea virus. The insertion in the influenza H15 HA once more demonstrates that RNA recombination events might be more common in virus evolution than was shown in the past. Influenza A viruses are assigned to different HA subtypes based on their antigenic reactivity in HI and double immunodiffusion assays (Schild et al., 1980; WHO Memorandum, 1980). The antigenicity of a molecule is determined by its amino acid sequence. The five antigenic sites of influenza H3 HA molecules are located at the globular head of the HA protein (Wiley et al., 1981). Antigenic sites of other HA subtypes probably are also located at the globular head, in light of the significant conservation of structurally important residues (e.g., cystine, glycine, and proline; Nobusawa et al., 1991). Therefore, we recommend comparison of the amino acid homologies of the HA1 genes as an additional criterion in the assignment of HA subtypes as did Kawaoka et al. (1990) when characterizing the H14 subtype. These HA1 amino acid sequence homologies are usually 70% or less for viruses of different subtypes and exceed 80% for those within a subtype. Since carbohydrate side chains may significantly influence antigenicity by masking potentially antigenic amino acid residues (Caton et al., 1982), analyzing the potential glycosylation sites within the globular head region may also be helpful when designating the HA subtype. However, different glycosylation patterns can be found between viruses belonging to the same subtype. What drives the generation of influenza A subtypes? All influenza A subtypes are found in wild aquatic birds. This significant genetic diversity seems to contrast with the high genetic conservation among avian viruses within a subtype. Several studies have shown that nucleotide mutations continuously occur in avian influenza genes, but with no net accumulation of amino acid changes in avian virus proteins (Gorman et al., 1990, 1991; Bean et al., 1992). These findings suggest that avian influenza viruses have reached an adaptive optimum with their hosts, where genetically conserved subtypes coexist in the same animal. The cause for divergent evolution and the reasons for continued coexistence of all influenza A subtypes in birds are unresolved. Does selective pressure like immunological pressure or adaptation to a new host species drive the generation of influenza A subtypes? Or, alternatively, does mere separation of genetic pools facilitate the emergence of different subtypes?

AID

VY 7786

/

6a12$$$343

02-12-96 13:35:11

515

Given the present static host–virus relationship of aquatic birds and avian influenza viruses, antibody selection is probably irrelevant. But emergence of the H15 HA in Australia and possible limitation to this continent make geographical separation of an Australian virus reservoir a very likely reason for the evolution of a new influenza HA subtype. It also provides the unique opportunity to test this hypothesis. If geographical separation facilitated the emergence of H15, these viruses in fact should be limited to Australia. Further surveillance studies of avian species worldwide could provide evidence for such a limitation of H15 viruses to the Australian continent. However, we have to consider that although viruses may originally be reservoir specific, they could spill over to another geographically separated reservoir by changes in avian migration patterns. ACKNOWLEDGMENTS We thank Dan Wehr and Scott Krauss for excellent technical assistance, Drs. P. K. Murray and A. J. Della-Porta for providing the viruses, Dr. C. Naeve and the St. Jude Children’s Research Hospital Center for BioTechnology for preparing the oligonucleotides, Krishna Sankhavaram and the St. Jude Children’s Research Hospital Molecular Biology Computer Facility for computer support, Amy Frazier for editorial support, and Dayna Baker for typing the manuscript. This work was supported in part by Grants RO1 AI-08831 and RO1 AI-29680 from the National Institutes of Allergy and Infectious Diseases, by Grant P30 CA-21765 from the National Cancer Institute, and by the American Lebanese Syrian Associated Charities (ALSAC).

REFERENCES Air, G. M. (1981). Sequence relationships among the hemagglutinin genes of 12 subtypes of influenza A virus. Proc. Natl. Acad. Sci. USA 78, 7639–7643. Bean, W. J., Jr., Sriram, G., and Webster, R. G. (1980). Electrophoretic analysis of iodine-labeled influenza virus RNA segments. Anal. Biochem. 102, 228–232. Bean, W. J., Schell, M., Katz, J., Kawaoka, Y., Naeve, C., Gorman, O., and Webster, R. G. (1992). Evolution of the H3 influenza virus hemagglutinin from human and nonhuman hosts. J. Virol. 66, 1129–1138. Bergmann, M., Garcia-Sastre, A., and Palese, P. (1992). Transfectionmediated recombination of influenza A virus. J. Virol. 66, 7576–7580. Bujarski, J. J., and Kaesberg, P. (1986). Genetic recombination between RNA components of a multipartite plant virus. Nature 321, 528–531. Caton, A. J., Brownlee, G. G., Yewdell, J. W., and Gerhard, W. (1982). The antigenic structure of the influenza virus A/PR/8/34 hemagglutinin (H1 subtype). Cell 31, 417–427. Cooper, P. D. (1977). Genetics of picornaviruses. In ‘‘Comprehensive Virology 9’’ (H. Fraenkel-Conrat and R. R. Wagner, Eds.), pp. 133– 207. Plenum, New York/London. Fields, S., and Winter, G. (1982). Nucleotide sequences of influenza virus segments 1 and 3 reveal mosaic structure of a small viral RNA segment. Cell 28, 303–313. Fitch, W. M. (1971). Toward defining the course of evolution: Minimum change for a specific tree topology. Syst. Zool. 20, 406–416. Gibson, C. A., Daniels, R. S., Oxford, J. S., and McCauley, J. W. (1992). Sequence analysis of the equine H7 influenza virus hemagglutinin gene. Virus Res. 22, 93–106. Gorman, O. T., Donis, R. O., Kawaoka, Y., and Webster, R. G. (1990). Evolution of influenza A virus PB2 genes: Implications for evolution of the ribonucleoprotein complex and origin of human influenza A virus. J. Virol. 64, 4893–4902.

vira

AP: Virology

¨ HM ET AL. RO

516

Gorman, O. T., Bean, W. J., Kawaoka, Y., Donatelli, I., Guo, Y., and Webster, R. G. (1991). Evolution of influenza A virus nucleoprotein genes: Implications for the origins of H1N1 human and classical swine viruses. J. Virol. 65, 3704–3714. Hahn, C. S., Lustig, S., Strauss, E. G., and Strauss, J. H. (1988). Western equine encephalitis virus is a recombinant virus. Proc. Natl. Acad. Sci. USA 85, 5997–6001. Hinshaw, V. S., Webster, R. G., and Turner, B. (1980). The perpetuation of orthomyxoviruses and paramyxoviruses in Canadian waterfowl. Can. J. Microbiol. 26, 622–629. Hinshaw, V. S., Air, G. M., Gibbs, A. J., Graves, L., Prescott, B., and Karunakaran, D. (1982). Antigenic and genetic characterization of a novel hemagglutinin subtype of influenza A viruses from gulls. J. Virol. 42, 865–872. Johansson, B. E., Bucher, D. J., and Kilbourne, E. D. (1989). Purified influenza virus hemagglutinin and neuraminidase are equivalent in stimulation of antibody response but induce contrasting types of immunity to infection. J. Virol. 63, 1239–1246. Katz, J. M., Wang, M., and Webster, R. G. (1990). Direct sequencing of the HA gene of influenza (H3N2) virus in original clinical samples reveals sequence identity with mammalian cell-grown virus. J. Virol. 64, 1808–1811. Kawaoka, Y., Yamnikova, S., Chambers, T. M., Lvov, D. K., and Webster, R. G. (1990). Molecular characterization of a new hemagglutinin, subtype H14, of influenza A virus. Virology 179, 759–767. Khatchikian, D., Orlich, M., and Rott, R. (1989). Increased viral pathogenicity after insertion of a 28 S ribosomal RNA sequence into the haemagglutinin gene of an influenza virus. Nature 340, 156–157. Klimov, A., Pro¨sch, S., Scha¨fer, J., and Bucher, D. (1992). Subtype H7 influenza viruses: comparative antigenic and molecular analysis of the HA-, M-, and NS-genes. Arch. Virol. 122, 143–161. Lai, M. M. C. (1992). RNA recombination in animal and plant viruses. Microbiol. Rev. 56, 61–79. Laver, W. G. (1964). Structural studies on the protein subunits from three strains of influenza virus. J. Mol. Biol. 9, 109–124. Laver, W. G., and Webster, R. G. (1972). Studies on the origin of pandemic influenza. II. Peptide maps of the light and heavy polypeptide chains from the hemagglutinin subunits of A2 influenza viruses isolated before and after the appearance of Hong Kong influenza. Virology 48, 445–455. Mackenzie, J. S., Edwards, E. C., Holmes, R. M., and Hinshaw, V. S. (1984). Isolation of ortho- and paramyxoviruses from wild birds in Western Australia, and the characterization of novel influenza A viruses. Aust. J. Exp. Biol. Med. Sci. 62, 89–99. Makino, S., Keck, J. G., Stohlman, S. A., and Lai, M. M. C. (1986). High-

frequency RNA recombination of murine coronaviruses. J. Virol. 57, 729–737. Meyers, G., Ru¨menapf, T., and Thiel, H.-J. (1989). Ubiquitin in a togavirus. Nature 341, 491. Nobusawa, E., Aoyama, T., Kato, H., Suzuki, Y., Tateno, Y., and Nakajima, K. (1991). Comparison of complete amino acid sequences and receptor-binding properties among 13 serotypes of hemagglutinins of influenza A viruses. Virology 182, 475–485. Orlich, M., Gottwald, H., and Rott, R. (1994). Nonhomologous recombination between the hemagglutinin gene and the nucleoprotein gene of an influenza virus. Virology 204, 462–465. Ro¨hm, C., Horimoto, T., Kawaoka, Y., Su¨ss, J., and Webster, R. G. (1995). Do hemagglutinin genes of highly pathogenic avian influenza viruses constitute unique phylogenetic lineages? Virology 209, 664–670. Sanger, F., Nicklen, S., and Coulson, A. R. (1977). DNA sequencing with chain-terminating inhibitors. Proc. Natl. Acad. Sci. USA 74, 5463– 5467. Schild, G. C., Newman, R. W., Webster, R. G., Major, D., and Hinshaw, V. S. (1980). Antigenic analysis of influenza A virus surface antigens: Considerations for the nomenclature of influenza virus. Arch. Virol. 63, 171–184. Slemons, R. D., Johnson, D. C., Osborn, J. S., and Hayes, F. (1974). Type A influenza viruses isolated from wild free-flying ducks in California. Avian Dis. 18, 119–125. Struck, D. K., Lennarz, W. J., and Brew, K. (1978). Primary structural requirements for the enzymatic formation of the N-glycosidic bond in glycoproteins. J. Biol. Chem. 253, 5786–5794. Su¨ss, J., Scha¨fer, J., Sinnecker, H., and Webster, R. G. (1994). Influenza virus subtypes in aquatic birds of Eastern Germany. Arch. Virol. 135, 101–114. Webster, R. G. (1970). Antigenic hybrids of influenza A viruses with surface antigens to order. Virology 42, 633–642. Webster, R. G., Isachenko, V. A., and Carter, M. (1974). A new avian influenza virus from feral birds in the USSR: Recombination in nature? Bull. WHO 51, 325–332. WHO Collaborating Center for Reference and Research on Influenza (1982). ‘‘Concepts and Procedures for Laboratory-Based Influenza Surveillance.’’ WHO Collaborating Center for Reference and Research on Influenza, Centers for Disease Control, Atlanta, GA, B17– B35. WHO Memorandum (1980). A revision of the system of nomenclature for influenza viruses. Bull. WHO 58, 585–591. Wiley, D. C., Wilson, I. A., and Skehel, J. J. (1981). Structural identification of the antibody-binding sites of Hong Kong influenza haemagglutinin and their involvement in antigenic variation. Nature 289, 373–378. Wilson, I. A., Skehel, J. J., and Wiley, D. C. (1981). Structure of the haemagglutinin membrane glycoprotein of influenza virus at 3A˚ resolution. Nature 289, 366–373.

AID

vira

VY 7786

/

6a12$$$344

02-12-96 13:35:11

AP: Virology