Characterization of the Surface Proteins of Influenza A (H5N1) Viruses Isolated from Humans in 1997–1998

Virology 254, 115–123 (1999) Article ID viro.1998.9529, available online at http://www.idealibrary.com on

Characterization of the Surface Proteins of Influenza A (H5N1) Viruses Isolated from Humans in 1997–1998 Catherine Bender,*,1 Henrietta Hall,* Jing Huang,* Alexander Klimov,* Nancy Cox,* Alan Hay,† Victoria Gregory,† Keith Cameron,† Wilina Lim,‡ and Kanta Subbarao* *Influenza Branch, Division of Viral and Rickettsial Diseases, Centers for Disease Control and Prevention, Atlanta, Georgia 30333; †Division of Virology, National Institute for Medical Research, Mill Hill, London NW7 1AA, United Kingdom; and ‡Government Virus Unit, Queen Mary Hospital, Hong Kong, China Received September 15, 1998; returned to author for revision October 16, 1998; accepted November 17, 1998 Influenza A (H5N1) viruses infected humans in Hong Kong between May and December, 1997. Sixteen viruses, including 6 from fatal cases, were isolated during this outbreak. Molecular analysis of the surface proteins genes encoding the hemagglutinin (HA) and neuraminidase (NA) of these H5N1 isolates, of a subtype not previously known to infect humans, are presented. The 16 human H5 HA sequences contain multiple basic amino acids adjacent to the cleavage site, a motif associated with highly pathogenic avian influenza A viruses. The phylogenetic relationship among both avian and human H5 hemagglutinins indicates that the human isolates are related directly to isolates that circulated among chickens in the live poultry markets in Hong Kong prior to and during the outbreak in humans. HA sequences from the human isolates and a recent chicken isolate represent a separate clade, within which there are two subgroups that are distinguishable antigenically and by the presence of a potential glycosylation site. Likewise the N1 neuraminidases of the human H5 isolates represent a clade that is evolutionarily distinct from previously characterized N1 neuraminidases. The recent human H5N1 virus NA genes are avian-like, indicating direct introduction from an avian source rather than evolution of a human N1 NA. All of the 16 human NA genes encode a shortened stalk due to a 19-amino acid deletion, also found in the recent avian H5N1 isolates from Hong Kong. Two unique amino acids were identified in the N1 NAs of the recent human isolates; however, it is not known if these residues influence host range. Neither the HA nor the NA genes of the human H5N1 virus isolates show evidence of adaptive changes during the outbreak. Although analyses of the surface protein genes of the H5N1 viruses from this outbreak did not provide immediate answers regarding the molecular basis for virulence, the analyses provided clues to potentially important areas of the genes worth further investigation. © 1999 Academic Press Key Words: human influenza; avian influenza; hemagglutinin; neuraminidase; glycosylation.

previously circulating human influenza viruses and avian influenza A viruses (Webster and Laver, 1972; Kawaoka et al., 1989). The influenza A H1N1 virus that caused the pandemic of 1918, killing .20 million people worldwide, however, does not appear to have originated by genetic reassortment (Gorman et al., 1991; Taubenberger et al., 1997). In May of 1997, a 3-year-old boy in Hong Kong contracted a fatal respiratory illness caused by an influenza A virus, which was subsequently identified as H5N1, a subtype not previously known to infect humans (de Jong et al., 1997). The initial human H5 isolate showed no evidence of genetic reassortment as all eight genes were avian in origin (Claas et al., 1998; Subbarao et al., 1998). Seventeen additional human cases were identified by virus isolation and/or serology between November 1997 and January 1998, raising concern about a potential pandemic. Infections were not restricted by age or sex; six of the cases proved fatal. Avian influenza viruses do not usually replicate efficiently or cause disease in humans (Webster et al., 1981; Bean et al., 1985; Shortridge, 1992; Kurtz et al., 1996), yet in this case, H5N1

INTRODUCTION Influenza A viruses of 15 hemagglutinin (HA) subtypes and 9 neuraminidase (NA) subtypes are known to circulate in birds and other animals, creating a reservoir of potential pandemic strains of influenza A containing new HA or NA genes. Of these 15 HA subtypes, only 3 have been isolated during outbreaks of influenza in humans since 1933, H1, H2 and H3 (Murphy and Webster, 1996). The introduction and subsequent spread in humans of influenza A viruses with HA or HA and NA genes of a novel subtype, called antigenic shift, results in a sudden and major change in virus antigenicity. Lack of protective immunity in the human population against the new HA or HA and NA proteins can result in rapid global spread of the virus, leading to widespread morbidity and mortality. When viruses of the H2N2 and H3N2 subtypes emerged as novel pandemic strains in 1957 and 1968, respectively, there was evidence of genetic reassortment between

1 To whom reprint requests should be addressed. Fax: (404) 6392334. E-mail: [email protected].

115

0042-6822/99 $30.00 Copyright © 1999 by Academic Press All rights of reproduction in any form reserved.

116

BENDER ET AL.

viruses were recovered from respiratory secretions from humans with mild and severe respiratory infections. Migratory waterfowl are thought to play a major role in the maintenance and dissemination of influenza A viruses (Slemons et al., 1974; Hinshaw et al., 1980). A variety of subtypes of avian influenza from the migratory bird population can nonlethally infect domestic poultry, suggesting that migratory waterfowl serve as a reservoir of influenza A viruses (Hinshaw et al., 1980). These observations are consistent with the common occurrence of nonlethal outbreaks in poultry. Highly pathogenic outbreaks of avian influenza have been restricted to two subtypes, H5 and H7 (Wood et al., 1993; Senne et al., 1996a). The expression of the highly pathogenic phenotype has been linked to multiple genes (Rott et al., 1979), but the hemagglutinin is the primary determinant of pathogenicity in birds (Webster and Rott, 1987; Kawaoka and Webster, 1988). Three outbreaks of highly pathogenic avian influenza of the H5 subtype have occurred among poultry in North America in 1966, 1983 and 1995 (Lang et al., 1968; Eckroade and Bachin, 1987; Senne et al., 1996b). Evidence of human infections was sought, but none was documented in association with these outbreaks (Bean et al., 1985). In 1996 there was a report of pathogenic H5 influenza among flocks of geese in Guangdong Province, China (Xiuying et al., 1998). From late March to early May, 1997, three chicken farms in Hong Kong were separately affected by avian influenza outbreaks with mortality rates from 70% to nearly 100%. An influenza A H5N1 virus was obtained from pooled organ material of the dead chickens (Claas et al., 1998) These events temporally preceded the disease outbreak among humans in Hong Kong in 1997. To date four hemagglutinin genes and one neuraminidase gene (Suarez et al., 1998; Subbarao et al., 1998) from this outbreak have been described. We characterized the remaining hemagglutinin and neuraminidase genes of the human H5N1 virus isolates to determine whether there was molecular evidence of adaptation of the viruses to the human population, to analyze the nucleotide sequences of the surface glycoproteins for previously reported molecular markers of virulence and to consider which viruses would be suitable for use in an H5N1 influenza vaccine for humans. RESULTS Antigenic and in vitro growth characteristics The antigenic characteristics of the human H5N1 influenza viruses determined by the hemagglutination-inhibition test are presented in Table 1. Five human H5N1 influenza viruses and corresponding postinfection sera raised in ferrets were used as reference strains. The A/Hong Kong/156/97 had been previously identified as H5 with a hyperimmune goat serum to A/Tern/South

Africa/61 (de Jong et al., 1997). Two antigenic groups, A and B, were clearly differentiated using postinfection ferret sera (Table 2). Group B viruses showed a decrease of fourfold or greater compared with Group A viruses when tested with postinfection sera raised against viruses in Group A. Antisera raised against Group B viruses were more broadly reactive with viruses in Group A, showing differences in titer of less than fourfold compared with titers of homologous (Group B) viruses. The HA sequences of viruses in Group B differed from those in Group A by a single amino acid change that encodes a potential glycosylation site at position 154 (Table 1) (Suarez et al., 1998). The viruses from seven Group A and seven Group B isolates plaqued efficiently in the absence of exogenous trypsin in three cell substrates, chick embryo fibroblasts (CEF), Madin-Darby canine kidney (MDCK) and MadinDarby bovine kidney (MDBK), without significant difference in HA virus titer at 32, 37, and 42°C (data not presented). Viruses that belong to Group A displayed a smaller plaque morphology on CEF than viruses of Group B, and virus yield from MDCK cells and embryonated eggs measured by hemagglutination titer was higher for Group B viruses than for Group A viruses (geometric mean HA titers for Group A viruses grown in MDCK cells and eggs were 12.3 and 8.0, respectively, and for Group B viruses were 36.4 and 58.8, respectively). Genetic characteristics of the hemagglutinin gene The complete nucleotide and deduced amino acid sequence of the HA gene (HA1 and HA2) of the 16 human isolates were compared with H5 isolates previously reported. Comparison of the human H5N1 influenza virus isolates to the chicken H5N1 influenza isolate obtained during the same outbreak (Claas et al., 1998; Suarez et al., 1998) showed that both the human and chicken isolates belonged to a clade that contained isolates found in Europe and Asia since 1978 (Fig. 1). However, these Eurasian virus isolates had the highly pathogenic motif (Arg-Glu-Arg-Arg) at amino acids 323–326. A high degree of homology was present at both the nucleotide and deduced amino acid level among human and avian isolates from this outbreak. The amino acid residue at position 156 further divided the clade (Table 1; Fig. 1). The change from Ala to Ser or Thr at this position created a potential glycosylation site at Asn 154. Seven potential glycosylation sites at amino acids 10, 11, 23, 165, 286, 484, and 543 were maintained among and between the human and chicken isolates. There was good correlation between changes in two residues, Ala156Ser or Ala156Thr and Lys368Gln. In seven of seven of the human isolates and in one chicken isolate with a Ser or Thr at 156, Gln was present at 368. The two exceptions, HK/538/97 and HK/97/98, had Gln at 368 but no potential

SURFACE PROTEINS OF H5N1 VIRUSES FROM HUMANS

117

TABLE 1

HK/97/98

HK/538/97

HK/516/97

HK/482/97

HK/481/97

HK/156/97

CK/HK/728/97

HK/542/97

HK/514/97

HK/503/97

HK/491/97

HK/485/97

HK/532/97

HK/483/97

CK/HK/220/97

CK/HK/915/97

HK/507/97

N I A N I V P R T T K T K M A Q G Q D Q Q K Q L A K K I Q

Group A Isolates HK/488/97

CK/HK/258/97

38 71 83 94 116 131 136 140 156a 188 189 196 218 226 320 322 338 345 349 360 368 388 392 410 426 451 483 519 547

Group B Isolates

HK/486/97

Amino Acid Number

Deduced Amino Acid Differences of Chicken and Human H5N1 Influenza Virus Hemagglutinin Gene Compared to the Hemagglutinin Gene of A/Chicken/258/97

D L T H

V

M L S K S

S

K S

S

S

A

A

A I

A

A

A

A

A

A

A

T

T

T

T

T

T

T H

R S R I T

T

T

T

T

T

T H R

T

T

T

T

H H H

H

K

K

K

K

K

K

H K

K

E P F T R R L L

Note. HA2 domain of the HA gene is shaded. Position 156 represents the third position in a potential glycosylation site defined by N-X-S/T, where X Þ P.

a

glycosylation site at 154. With one exception, all key amino acids believed to be associated with receptor specificity of the H5 HA were conserved among all of the human H5N1 influenza isolates (Garcia et al., 1996) Only HK/514/97 had a conservative amino acid change in one of the amino acids (Val131Leu) identified as part of the putative receptor binding site. Genetic characterization of the neuraminidase gene Sequence analysis of the NA genes revealed all of the human isolates to be of the N1 subtype (Fig. 2). Historically viruses with N1 neuraminidase have circulated in both human and avian hosts and group into two phylogenetically distinct clades based on host of origin. The NAs of all of the H5N1 human isolates from the recent outbreak were related closely to the N1 neuraminidase

from birds, not from humans. (Subbarao et al., 1998; Claas et al., 1998; present data) The stalk region of the human H5N1 influenza neuraminidase, which extends from the viral membrane up to amino acid 85, had a 19-amino acid deletion when aligned with other available N1 sequences (Table 3). When the NA genes of the human viruses were compared to that of the chicken isolate, A/Chicken/Hong Kong/258/97, a number of changes were observed. Residue 412 changed from Met to Leu. There was also a highly conserved substitution at 461 from Gly to Asp, with the single exception of the HK/516/97 isolate. As with the HA gene, there were apparently highly correlated changes at amino acids Ile29Thr, Gln39Lys, and Ile223Thr. The exceptions to the correlation was HK/488/ 97, which did not have the amino acid change at residue

118

BENDER ET AL. TABLE 2 Hemagglutination-Inhibition (HAI) Reactions of Human H5 Influenza Specimens Reference Ferret and Sheep Antisera

HK/483/97

HK/491/97

HK542/97

SHEEP H3

Passageb

H3 WHO Reference Antigen A/Hong Kong/156/97 A/Hong Kong/97/98 A/Hong Kong/483/97 A/Hong Kong/491/97 A/Hong Kong/542/97 Test Antigens Group A A/Hong Kong/481/97 A/Hong Kong/482/97 A/Hong Kong/486/97 A/Hong Kong/488/97 A/Hong Kong/507/97 A/Hong Kong/538/97 Group B A/Hong Kong/485/97 A/Hong Kong/503/97 A/Hong Kong/514/97 A/Hong Kong/532/97

HK/97/98

Reference Antigens

Group B

HK/1561TKYa

Group A

,10 1280 1280 80 320 320

,10 320 320 40 40 80

,10 640 320 640 160 160

,10 1280 640 160 320 320

10 1280 640 160 320 320

1280 ,10 ,10 ,10 ,10 ,10

C2/C3 X/C2 C1/C3 C2/C1 X/C1

1280 640 640 640 1280 1280

160 160 160 160 320 320

320 320 320 640 640 640

640 640 640 1280 1280 1280

640 640 640 640 640 640

,10 ,10 ,10 ,10 ,10 ,10

CX/C3 C1/C2 C2/C2 C1/C2 C1/C2 X/C2

320 320 320 160

80 80 80 40

160 320 320 160

320 640 320 160

320 640 320 320

,10 ,10 ,10 ,10

C2/C1 C1/C1 X/C2 X/C1

Note. Group A antigens are indicated by the shaded area. a To boost titer, sera were concentrated 1:4. The Hong Kong/156/97 infected ferrets were boosted with A/H5/Turkey/Wisconsin/68 virus intranasally to obtain higher antibody titers. b Passage history indicated by: X 5 unknown substrate; C 5 MDCK; number indicates number of passages in substrate.

39, and HK/532/97 and HK/542/97, isolates that had the Ile29Thr change without the amino acid changes at 39 and 223. Replacement of the Thr for Ile at residue 223 creates a potential glycosylation site that divides the human isolates into two phylogenetic groups. The similarity between the NAs of five of the viruses (HK/482/97, HK/486/97, HK/507/97, HK/538/97 and HK/97/98) corresponds to the close relationship between the HAs (Group A), further emphasizing the distinction between Groups A and B. The NAs of the chicken and human isolates shared four new glycosylation sites and lost four glycosylation sites when compared with the NA of the nearest genetic relative, A/Parrot/Ulster/73 (H7N1). There are no other complete avian N1 sequences reported in the literature. The NA of all human and chicken isolates retained the cysteine at position 76 (our numbering is 89), a position thought to be necessary for the formation of infectious virus (Luo et al., 1993). Amino acid residues located in the substrate binding pocket (N2 numbering) (Colman et al., 1983) were completely conserved in all human strains examined.

DISCUSSION Monitoring of variation in the HA and NA genes of influenza A viruses is important in choosing appropriate vaccine strains as well as for identifying new emerging and re-emerging subtypes of influenza in humans. With the appearance of influenza A (H5N1) viruses in the human population for the first time, it was necessary to establish the relationships among the human isolates and between the human and avian isolates. The nucleotide and deduced amino acid sequences of HA confirm that the 16 human isolates were closely related, and the multiple basic amino acid cleavage site identified in the first isolate (Subbarao et al., 1998) was conserved in all the subsequent isolates. Phylogenetic analysis showed that the human H5 HA genes were closely clustered, that the HA sequence of A/Chicken/HK/258/97 was almost indistinguishable from those of the human isolates and that the human and chicken viruses belong to a lineage of viruses that have been isolated in Europe and Asia (Eurasian lineage). A highly pathogenic H5 virus that caused outbreaks among poultry in Mexico in 1995 con-

SURFACE PROTEINS OF H5N1 VIRUSES FROM HUMANS

FIG. 1. Evolutionary relationships of 33 influenza A (H5) viruses based on the nucleotide sequences of the HA gene rooted to the HA of A/Tern/South Africa/61. Horizontal distances are proportional to the number of nucleotide differences between the viruses. *, isolates that have a multibasic cleavage site motif associated with high pathogenecity (R/K-X-R/K-R).

tained an HA with the same multiple basic amino acid cleavage site as these contemporary isolates, but it belonged to a lineage of viruses that have only been isolated in North America (North American lineage). Our analysis confirms previous reports (Claas et al., 1998; Suarez et al., 1998) that the HAs of viruses that infected humans appear to be very similar to one that caused poultry outbreaks in Hong Kong with no cumulative changes to indicate adaptation to the human host. The presence or absence of the glycosylation site at 154 distinguished Group A and Group B H5 influenza viruses both antigenically and genetically. Antibody to Group B viruses showed a greater degree of crossreaction against Group A viruses than vice versa, a feature for consideration in the selection of a vaccine strain for the H5 subtype. Glycosylation of the HA plays a role in antigenic variation by masking and unmasking antigenic sites. The number and distribution of glycosylation sites along the HA1 subunit has been implicated in establishing tissue tropism (Kawaoka et al., 1984), and virulence has been associated with the loss of a glycosylation site at position 11 in field isolates of the H5 subtype (Kawaoka and Webster, 1988). In the chicken and the 16 human isolates

119

from the Hong Kong outbreak, the glycosylation site at residue 11 was conserved. A glycosylation site at amino acid 236, near the receptor-binding pocket, has been suggested to be an adaptive feature of H5 viruses associated with efficient replication and/or transmissibility in poultry (Garcia et al., 1997). Neither the chicken H5N1 influenza nor the 16 human H5N1 influenza viruses acquired this potential glycosylation site. The N1 neuraminidases of the H5 human virus isolates were very closely related to each other and to that of the chicken isolate from the same outbreak but had a 19-amino acid deletion in the stalk not present in the nearest genetic relative analyzed, A/Parrot/Ulster/73 (Claas et al., 1998; Suarez et al., 1998; Subbarao et al., 1998) Glycosylation sites in the stalk of the neuraminidase are believed to play a role in maintaining the tetrameric structure of the protein. Three of the four glycosylation sites present in the NA of Parrot/ Ulster/73 but absent in these viruses are due to the deletion in the stalk region. The number and sequence of amino acid residues in the stalk region is variable, both between and within subtypes. Transfectant viruses with deletions in the stalk region had a reduced ability to grow in MDCK cells as compared to MDBK cells (Luo et al., 1993) and replication of other transfectant viruses in eggs correlated with NA stalk length:

FIG. 2. Evolutionary relationships of 31 influenza A virus neuraminidases based on the nucleotide sequences of the NA gene rooted to the NA of A/Puerto Rico/8/34. Horizontal distances are proportional to the number of nucleotide differences between the viruses.

120

BENDER ET AL. TABLE 3

Amino Acid

CK/HK/258/97

CK/HK/220/97

HK/483/97

HK/532/97

HK/485/97

HK/491/97

HK/503/97

HK/514/97

HK/542/97

HK/156/97

HK/481/97

HK/482/97

HK/486/97

HK/488/97

HK/507/97

HK/516/97

HK/538/97

HK/97/98

Deduced Amino Acid Differences of Chicken and Human H5N1 Influenza Virus Neuraminidase Gene Compared to the Neuraminidase Gene of A/Chicken/258/97

1–15a 21 29 34 36 37 39 40 43 50 53–72b 81 85 88 104 117 118 142 156 166 171 187 201 211 217 221 223 225 227 247 283 334 386 396 412 430 444 450 454 461 466

NS S I V H I Q T P N DEL A L N N I R D R A N G G I K N I R Q S P S S I M R I N V G F

X

X

X

X

X

X

X

X

X

X

X

X N T

X

X

X

X

X

a b

T

T

T

K

K

DEL

DEL

T

T

T

T

T

K

K

K

K

DEL

DEL

DEL

K DEL

DEL

T

T

T

T

T

N

N

L S

L

L

D

D

I P

P T

A

DEL

DEL

DEL

DEL

A

A S

DEL

DEL

D

DEL

DEL T S

DEL

DEL

Y K M K N G T K S E M R

N K T

T

K P S T T L

V L

L

L

L

L

L

V L

L

L

L

L

L

L

D

D

D

D

D

D

V D D

D

D

M D

D

D

D

D L

Sequences obtained for A/Chicken/Hong Kong/258/97 in Genbank did not contain nucleotides for the first 15 residues. Both chicken and human isolates had this 19-amino deletion in the stalk when compared to the nearest genetic relative, A/Parrot/Ulster/73.

the longer the stalk, the higher the virus titer. It has been suggested that higher NA activity might be required for replication in eggs than in tissue culture (Castrucci and Kawaoka, 1993). The human H5N1 influenza viruses achieved lower HA titers in MDCK cells than in eggs (data not presented). The NA with a 19-amino acid deletion may have a different efficiency of enzymatic cleavage in different substrates that

might influence growth characteristics in different host cells and contribute to altered host range. Influenza A H5N1 viruses were not known to infect human hosts before 1997. Gain or loss of potential glycosylation sites in HA clearly plays a role in determining morphological characteristics of the viral isolates, antigenicity and influences selection of potential vaccine strains, but the role these sites play in disease

SURFACE PROTEINS OF H5N1 VIRUSES FROM HUMANS

121

TABLE 4 Human Influenza A (H5N1) Isolates Used in this Study Genbank Accession No. Strain Hong Hong Hong Hong Hong Hong Hong Hong Hong Hong Hong Hong Hong Hong Hong Hong a b

Kong/156/97 Kong/481/97 Kong/483/97 Kong/482/97 Kong/485/97 Kong/486/97 Kong/491/97 Kong/488/97 Kong/542/97 Kong/503/97 Kong/507/97 Kong/516/97 Kong/514/97 Kong/97/98 Kong/532/97 Kong/538/97

Age

Gender

Date of illness

3 2 13 54 24 5 4 2 19 1 3 60 25 34 14 3

M M F M F F M M F M F F F F F M

05/09/97 11/06/97 11/20/97 11/24/97 12/04/97 12/07/97 12/10/97 12/12/97 12/14/97 12/16/97 12/16/97 12/16/97 12/17/97 12/18/97 12/21/97 12/28/97

Outcome

Hemagglutinin

Neuraminidase

Died Discharged Died Died Discharged Discharged Discharged Discharged Discharged Discharged Discharged Died Died Died Discharged Discharged

AF036356a AF046096b AF046097b AF046098b AF102681 AF102671 AF102677 AF102672 AF102678 AF102679 AF102675 AF102673 AF102682 AF102676 AF102680 AF102674

AF036357a AF102663 AF102668 AF102656 AF102664 AF102658 AF102665 AF102657 AF102670 AF102666 AF102659 AF102660 AF102669 AF102661 AF102667 AF102662

Subbarao et al., 1998. Suarez et al., 1998.

is unknown. Previous H5 outbreaks (1983 and 1995) among commercial poultry flocks were of the H5N2 subtype with no associated human disease (Bean et al., 1985). Perhaps the N1 neuraminidase, with its shortened stalk, influences the virus’ host range. This hypothesis could be investigated by restoring the NA stalk length at the level of cDNA by mutagenesis and recovering the modified NA gene into infectious virus by reverse genetics. The human and recent avian H5N1 influenza viruses share amino acid changes from previously circulating H5 and N1 subtypes, with a limited number of unique changes that might contribute to the apparent shift in host range. Comparison of the human H5N1 influenza isolates with additional recent avian isolates from Hong Kong should provide more information about the relationship of the avian to human isolates. If the unique changes are confirmed by further analysis of avian H5N1 influenza isolates, reverse genetic techniques could be used to test the significance of specific residues. MATERIALS AND METHODS Viruses Sixteen of the 18 human cases of H5N1 influenza yielded virus isolates. The viruses used in the present study are listed in Table 4. Isolates were obtained from throat swabs, nasopharyngeal swabs, tracheal aspirates or bronchoalveolar lavage. Viruses were propagated in either embryonated eggs or MDCK cells (Kendal et al., 1982) or obtained by plaque-purifying on chick embryo fibroblasts (CEF) cells.

Antigenic analysis The identity and antigenic characteristics of the 16 human isolates were determined with postinfection ferret or sheep serum (H3 only), using the hemagglutination-inhibition (HAI) assay (Kendal et al., 1982). In all cases, the ferret serum was concentrated 1:4 to increase the working titer. The ferret infected with the A/Hong Kong/156/97 virus also was inoculated intranasally with influenza A/Turkey/Wisconsin/68 virus to boost the initial titer. Nucleotide sequencing Genomic RNA was extracted directly from allantoic fluid or cell culture supernatant with the RNeasy kit (Qiagen, Chatsworth, CA). This viral RNA was used to amplify the HA and NA genes by the reverse-transcriptase–polymerase chain reaction (RT–PCR) followed by sequencing using the ABI Prism dye terminator cycle sequencing kit with an ABI Model 373A DNA Sequencer (Perkin-Elmer, Applied Biosytems Division, Foster City, CA). All reactions were carried out according to the manufacturers’ instructions. Nucleotide sequence of the coding regions is reported. Sequences of amplifying primers and sequencing primers are available upon request. Phylogenetic analysis Sequence data were analyzed using version 8.1 of the sequence analysis software package of the University of Wisconsin Genetics Computer Group (Genetics Com-

122

BENDER ET AL.

puter Group, 1994). Version 3.57 of the Phylogeny Inference Package (PHYLIP) (Felsenstein, 1989) was used to estimate the phylogenies from the nucleotide sequences using the NEIGHBOR program; and the phenogram was drawn using the DRAWGRAM program. The neighborjoining method was selected as the algorithm to describe the relationships because it gives results comparable to those obtained using methods that require longer computational time (Saitou and Imanishi, 1989). Nucleotide sequences for viruses not listed in Table 4 were obtained from GenBank. In vitro growth characteristics Serial 10-fold dilutions of 7 of the 16 human H5N1 influenza viruses were tested for the ability to form plaques on monolayers of CEF, Madin-Darby canine kidney (MDCK) and Madin-Darby bovine kidney (MDBK) cells. Plaque assays were carried out at 32, 37, and 42°C to determine whether the viruses exhibited a temperature-sensitive phenotype. Virus yield from MDCK cultures and embryonated eggs was measured by hemagglutination titers. ACKNOWLEDGMENTS We thank Brian Holloway, Edward George and Melissa Olsen-Rasmussen of the Biotechnology Core Facility at the CDC and Michael Bennett of NIMR for excellent assistance with this project.

REFERENCES Bean, W. J., Kawaoka, Y., Wood, J. M., Pearson, J. E., and Webster, R. G. 1985. Characterization of virulent and avirulent A/Chicken/Pennsylvania/83 influenza A viruses: Potential role of defective interfering RNAs in nature. J. Virol. 54, 151–160. Castrucci, M. R., and Kawaoka, Y. (1993). Biologic importance of neuraminidase stalk length in influenza A virus. J. Virol. 67(2), 759–764. Claas, E., Osterhaus, A., van Beek, R., De Jong, J., Rimmelzwaan, G., Senne, D., Krauss, S., Shortridge, K., and Webster, R. G. (1998). Human influenza A H5N1 virus related to a highly pathogenic avian influenza virus. Lancet 351, 472–477. Coleman, P. M., Varghese, J. N., and Laver, W. G. (1983). Structure of the catalytic and antigenic sites in influenza virus neuraminidase. Nature 303(5), 41–44. de Jong, J. C., Claas, E. C. J., Osterhaus, A. D. M.E., Webster, R. G., and Lim, W. L. (1997). A pandemic warning? [letter] Nature 389(6651), 554. Eckroade, R. J., and Bachin, L. A. (1987). Avian influenza in Pennsylvania—the beginning. In ‘‘Proceedings of the 2nd International Symposium on Avian Influenza,’’ pp. 22–32. US Animal Health Association, Athens, GA. Felsenstein, J. (1989). PHYLIP—Phylogeny inference package (version 3.2). Cladistics 5, 164–166. Garcia, M. C., Crawford, J. M., Latimer, J. L., Riveria-Cruz, E., and Perdue, M. L. (1996). Heterogeneity in the hemagglutinin gene and emergence of the highly pathogenic phenotype among recent H5N2 avian influenza viruses from Mexico. J. Gen. Virol. 77, 1493–1504. Garcia, M. C., Suarez, D. L., Crawford, J. M., Latimer, J. W., Slemons, R. D., Swayne, D. E., and Perdue, M. L. (1997). Evolution of H5 subtype avian influenza A viruses from North America. Virus Res. 51, 115–124. 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. Hinshaw, V. S., Webster, R. G., and Turner, R. G. (1980). The perpetuation of orthomyxoviruses and paramyxoviruses in Canadian waterfowl. Can. J. Microbiol. 26, 622–629. Kawaoka, Y., Krauss, S., and Webster, R. G. (1989). Avian-to-human transmission of the PB1 gene of influenza A viruses in the 1957 and 1968 pandemics. J. Virol. 63 (11), 4603–4608. Kawaoka, Y., Naeve, C. W., and Webster, R. G. (1984). Is virulence of H5N2 influenza viruses in chickens associated with a loss of carbohydrate from the hemagglutinin? Virology 139, 303–316. Kawaoka, Y., and Webster, R. G. (1988). Molecular mechanism of acquisition of virulence in influenza virus in nature. Microb. Pathog. 5, 311–318. Kendal, A. P., Pereira, M. S., and Skehel, J. J. (1982). ‘‘Concepts and procedures for laboratory-based influenza surveillance.’’ U. S. Department of Health and Human Services, Centers for Disease Control and Prevention, Atlanta, GA. Kurtz, J., Manvell, R. J., and Banks, J. (1996). Avian influenza virus isolated from a woman with conjunctivitis [Letter]. Lancet 348, 901– 902. Lang, G., Narayan, O., Rouse, B. T., Ferguson, A. E., and McConnell, M. C. (1968). A new influenza virus A infection in turkeys. II. A highly pathogenic variant, A/Turkey/Ontario/7732/66. Can. Vet. 9, 151–160. Laver, W. G. (1978). Crystallization and peptide mapping of neuraminidase ‘‘heads’’ from H2N2 and H3N2 influenza virus strains. Virology 86, 78–87. Luo, G., Chung, J., and Palese, P. (1993). Alterations of the stalk of the influenza virus neuraminidase: Deletions and insertions. Virus Res. 29, 141–153. Murphy, B. R., and Webster, R. G. (1996). Orthomyxoviruses. In ‘‘Field’s Virology,’’ (B. N. Fields et al., Eds.). Lippincott Raven, Philadelphia. Palese, P., Tobita, K., Ueda, M., and Compans, R. W. (1974). Characterization of temperature-sensitive influenza virus mutants defective in neuraminidase. Virology 61, 397–410. Rott, R., Orlich, M., and Scholtissek, C. (1979). Correlation of pathogenecity and gene constellation of Influenza A viruses. III. Non-pathogenic recombinants derived from highly pathogenic parent strains. J. Gen. Virol. 44, 471–477. Saitou, N., and Imanishi, T. (1989). Relative efficiencies of Fitch-Margolish maximum-parsimony, maximum-likelihood, minimum-evolution and neighbor-joining methods of phylogenetic tree construction in obtaining the correct tree. Mol. Biol. Evol. 6, 514–525. Senne, D. A., Panigraphy, B., Kawaoka, Y., Pearson, J. E., Suss, J., Lipkind, M., Kida, H., and Webster, R. G. (1996a). Survey of the hemagglutinin (HA) cleavage site sequence of H5 and H7 avian influenza viruses: amino acid sequence at the HA cleavage site as a marker for pathogenecity potential. Avian Dis. 40, 425–437. Senne, D. A., Rivera, E., Panigraphy, B., Fraire, M., Kawaoka, Y., and Webster, R. G. (1996b). Characterization of avian influenza H5N2 isolates recovered from chickens in Mexico. In ‘‘Proceedings of the 45th Western Poultry Disease Conference, Cancun, Mexico,’’ (M. M. Jensen and M. A. Ceniceros, Eds.), pp. 5–8. University of California Press, Davis, CA. Shortridge, K. F. (1992). Pandemic influenza: A zoonosis? Semin. Respir. Infect. 7, 11–25. Slemons, R. D., Johnson, D. C., Osborn, J. S., and Hayes, F. (1974). Type-A influenza virus isolated from wild free-flying ducks in California. Avian Dis. 18, 119–124. Suarez, D. L., Perdue, M. L., Cox, N., Rowe, T., Bender, C., Huang, J., and Swayne, D. E. (1998). Comparison of highly virulent H5N1 influenza A viruses isolated from humans and chickens from Hong Kong. J. Virol. 72(8), 6678–6688. Subbarao, K., Klimov, A., Katz, J., Regnery, H., Lim, W., Hall, H., Perdue, M., Swayne, D., Bender, C., Jing, H., Hemphill, M., Rowe, T., Shaw, M., Xu, X., Fukuda, K., and Cox, N. (1998). Characterization of avian

SURFACE PROTEINS OF H5N1 VIRUSES FROM HUMANS influenza A (H5N1) virus isolated from a child with a fatal respiratory illness. Science 279, 393–396. Taubenberger, J. K., Reid, A. H., Kraft, A. E., Bijwaard, K. E., and Fanning, T. G. (1997). Initial characterization of the 1918 ‘‘Spanish’’ influenza virus. Science 275, 1793–1796. Webster, R. G., Geraci, J., Pertusson, G., and Skimson, K. (1981). Conjuctivitis in human beings caused by influenza A virus of seals. N. Eng. J. Med. 304, 911. Webster, R. G., and Laver, W. G. (1972). The origin of pandemic influenza. Bull. WHO 47(4), 449–452. Webster, R. G., Laver, W. G., Air, G. M., and Schild, G. C. (1982).

123

Molecular mechanism of variation in influenza viruses. Nature 296, 115–121. Webster, R. G., and Rott, R. (1987). Influenza virus A pathogenicity: A pivotal role of the hemagglutinin. Cell 50, 665–666. Wood, G. W., McCauley, J. W., Bashruddin, J. B., and Alexander, D. J. (1993). Deduced amino acid sequences at the hemagglutinin cleavage site of avian influenza A viruses of H5 and H7 subtypes. Arch. Virol. 130, 209–217. Xiuying, T., Guobin, T., Chuanshan, Z., Jinfanand, Z., and Kangzhen, Y. (1998). Isolation and characterization of prevalent strains of influenza viruses in China. Chin. J. Anim. Poultry Infect. 20(1), 1–5.