VIROLOGY
177,578-587
(1990)
Antigenic,
Sequence,
GILLIAN
and Crystal Variation
in Influenza
6 Neuraminidase’
M. AIR,*,‘W. GRAEME LAVER,t MING LUO,* STEPHEN J. STRAY,* GAYLA LEGRONE,* AND ROBERT G. WEBSTER+
*Department of Microbiology, University ofAlabama at Birmingham, Birmingham, Alabama 35294; John Curtin School of Medical Research, Australian National University, P. 0. Box 334, Canberra, and *St. Jude Children’s Research Hospital, P. 0. Box 3 18, Memphis, Tennessee Received
December
13, 1989; accepted
March
tlnfluenza Research Unit, ACT 260 1, Australia; 38 10 1
19, 1990
The neuraminidase (NA) genes of influenza B viruses B/Maryland/59, B/Hong Kong/8/73, B/Singapore/222/79, B/ Oregon/5/80, B/USSR/100183, BNictoria13185, BILeningradll79l86, B/Memphis/6/86, and B/Memphis/3/89 have been sequenced. The deduced amino acid sequences show high variability in the stalk domain of the NA, but a surprising degree of sequence conservation in the head region which carries all the antigenic and enzyme activity. The variable region coding for the neuraminidase stalk also translates into a variable section in the overlapping NB polypeptide, which is coded from a second reading frame that overlaps the first 100 amino acids of NA. The influenza B NAs are antigenically distinguishable with monoclonal antibodies in neuraminidase-inhibition tests, even when there is only one amino acid sequence difference. However, seven of nine escape mutants selected with monoclonal antibodies were distinguished only by the antibody used for selection. When NA heads of influenza B viruses are crystallized, there are remarkable differences in crystal morphology between neuraminidases which have very few sequence changes. 0 1990 Academic
Press.
Inc.
INTRODUCTION
in antigenic properties and sequence is similar to that of the HA (Webster et a/., 1982; Martinez et al., 1983; Colman and Ward, 1985). NA heads, which retain the tetrameric configuration of the intact molecule, can be released from several influenza viruses by proteases. These heads carry all the enzyme activity and the antigenic determinants. In some cases, the heads have been crystallized in a form suitable for structure analysis by X-ray diffraction methods. The three-dimensional structure of NA heads from two subtypes of influenza A (N2 and N9) has been determined (Varghese et al., 1983; Baker et al., 1987). Crystals of NA heads have also been obtained from several strains of influenza B (Bossart et al., 1988; Lin eta/., 1990) and from one of these complexed with antibody Fab (Laver et al., 1988), and the crystal structure determinations by X ray diffraction are in progress. In order to interpret the structure in terms of function and antigenicity, and to determine if the erratic antigenie and sequence variation found in influenza B HAS are duplicated in the NA, we have studied the antigenic properties and determined the NA gene sequence of 10 epidemic strains of influenza B, and have further investigated antigenic variation by selecting and characterizing escape mutants with monoclonal antibodies raised against the NA of influenza virus B/Lee/40.
There are fundamental differences in the way antigenie variation is manifested in influenza 6 viruses compared to the sequential pattern of antigenic drift seen in influenza A (Schild et a/., 1973; Lu et a/., 1983; Oxford et al., 1984; Hovanec and Air, 1984). Influenza B viruses change antigenic properties erratically, and the differences seem to involve fewer amino acid sequence changes than those found in influenza A antigens (Krystal et al., 1983; Yamashita et al., 1988). These analyses have so far been restricted to the hemagglutinin (HA). This surface glycoprotein enables the virus to bind to sialic acid receptors on the surface of host cells and then facilitates entry by a membrane fusion process. The second major surface glycoprotein of both influenza A and 6 viruses is the neuraminidase (NA), whose enzyme activity removes terminal sialic acid and thus destroys receptors for the virus. Although the role of this activity in viral infection has never been completely defined, it has been shown to assist in viral release and spread (Palese et al., 1974). Antibodies to the NA indirectly neutralize virus infectivity and protect animals against infection (Webster et al., 1982, 1988). In influenza A neuraminidases, the extent of variation
MATERIALS ’ Sequence EMBUGenBank M30639. ’ To whom 0042.6822/90
data from this article have been deposited Data Libraries under Accession Nos. reprint
requests
should
$3.00
CopyrIght 0 1990 by Academic Press, Inc. All rights of reproduction in any form reserved.
with the M30631-
AND
METHODS
Viruses The viruses used in this work MarylandI59, B/Hong Kong/8/73,
be addressed. 578
were B/Lee/40, B/ B/Hong Kong18173
579
VARIATION IN INFLUENZA B NEURAMINIDASE
(HG), B/Singapore/222/79, B/Singapore/222/79 (HG), B/Oregon/5/80, B/Oregon/5/80 (HG), B/USSR/l 00/83, B/USSR/l 00/83 (HG), B/Victoria/3/85, B/Leningrad/ 179/86, B/Memphis/6/86, B/Memphis/6/86 (HG), and B/Memphis/3/89 (HG). High-growth (HG) reassortant viruses were prepared by reasserting with B/HK (HG), which is a reassortant virus, B/Hong Kong/8/73 (HA) Lee/40 (NA) [RI, prepared and characterized as previously described (Webster, 1970; Wei et a/., 1987). The NAs of reassot-tant high growth (HG) viruses were identified using monoclonal antibodies. Except for B/Leningrad/l79/86, the influenza B viruses for this study were obtained from the repository at St. Jude’s Hospital, and grown either at St. Jude’s or at the Australian National University. B/Leningrad/86 viral cores were obtained from Dr. lgor Kharitenikov of the lvanovsky Institute of Virology, Moscow, USSR. Viruses were grown in embryonated eggs and purified by adsorption to and release from red blood cells followed by sucrose density centrifugation (Laver, 1969).
1988). The purified virus particles were suspended in 0.15 NI NaCI, Tris buffer pH 8.0, and 1 m&I CaCI, and incubated with TPCK-trypsin (Worthington Biochemicals) for 2 hr at 37”. Virus cores were removed by centrifugation (30,000 rpm for 1 hr) and the supernatant, which contained 100% of the NA activity, was centrifuged at 50,000 rpm (Ti 60 rotor) for 20 hr. The NA heads formed a pellet which, with most type B neuraminidases, had a crystalline appearance. The pellet was dissolved in saline (0.5 ml) and the NA heads purified by sucrose density gradient centrifugation (5-20% sucrose in 0.15 M NaCI, SW65 rotor, 58,000 rpm at 5” for 8 hr. Fractions containing NA were pooled, dialyzed against saline, and then against 70% saturated (NH&SO,. The precipitated NA heads were spun down, dissolved in the minimum amount of saline, and dialyzed against saline. The NA heads were crystallized from hanging drops by vapor diffusion using a variety of solvent systems. Preparation of viral RNA and sequence analysis
Serological tests Neuraminidase and neuraminidase-inhibition assays were done with fetuin substrate as described (AymardHenry et a/., 1973). Enzyme-linked immunosorbent assays (ELISA) were carried out as described (Kida et a/., 1982). Monoclonal antibodies Monoclonal antibodies were made as previously described (Webster and Berton, 1981; Laver eta/., 1988). The hybridomas were derived from spleen cells of a Balb/c mouse immunized with B/HK (HG) virus, or B/ Ore/80 virus (10 pg protein), and boosted 1 month later with 40 pg purified trypsin-released NA. The cells were fused 3 days postboost with myeloma P3X63-Ag8.653 cells. The hybridomas were screened by ELISA assay and by neuraminidase-inhibition tests. Hybridomas producing antibodies to NA were cloned in soft agar and grown as ascites in Balb/c mice. Ascitic fluid was collected from pristane-primed Balb/c mice. Selection of escape mutants Escape mutants were obtained by selection with neutralizing monoclonal antibodies as previously described (Webster and Berton, 1981) except that the selection and cloning were done in Madin Darby Canine Kidney (MDCK) cells. Preparation and crystallization
of NA heads
NA heads were prepared for crystallization similarly to the methods previously described (Bossart et a/.,
This was done as previously described (Air et a/., 1985). The starting material was either intact virus, or viral “cores” from which the NA had been removed by trypsin for crystallization. RESULTS 1. Sequences of NA genes of influenza B viruses isolated from 1940 to 1989 RNA templates were isolated from influenza viruses B/Mar-$and/59, B/Hong Kong18173 and (HG), BlSingapare/222/79 and (HG), B/Oregon/5/80 (HG), B/USSR/ 100/83 and (HG), B/Victoria/3/85, B/Leningrad/l 79/ 86, B/Memphis/6/86 and (HG), and B/Memphis/3/ 89 (HG). Sequences of cDNA transcribed from viral RNA were determined using the dideoxy method with the oligonucleotide primers specific for RNA segment 6 previously described (Wei et a/., 1987); they were designed from the B/Lee/40 NA sequence of Shaw et al. (1982) or the 5’ 56 nucleotide sequence of B/Maryland/59 NA gene of Stoeckle et al. (1987). In cases where these failed to prime or gave ambiguities, alternative oligonucleotides (usually 15-mers) were synthesized. Most of the NA cDNA sequences begin at nucleotide 31 (from a primer spanning nucleotides 14-29). In B/Len/86 there is apparently some change in the 3’ sequence of the vRNA, and no oligonucleotides made from this region would prime synthesis. The sequence for this strain begins at nucleotide 5 1, from a primer spanning 37-49. All the sequencing reactions were run more than once, the gels were read by at least two people, and the differ-
580
AIR ET AL.
FIG. 1. Nucleotide sequence of the NA gene of BIMem/3/89 and differences in the sequences of B/Mem/6/86, B/Len/179/86. B/Vie/3/85, B/ USSR/100/83, B/Ore/5/80, BISing1222179, B/Hong Kong/8/73, BIMatylandI59. and BILeei40. The B/Lee/40 sequence is that of Shaw et a/., (1982). This differs at 3 nucleotides from the B/Lee NA sequence in the B/HK (HG) reassortant virus (Wei et a/., 1987). The sequences are given
VARIATION
B/Lee,40 B/Mary/59 B/XK/8/J3 B/Sing/79 B,ore,ao B,“SSR/B3 B,“iC,3,85 B,Le”,i79,86 B,Mem,6,86 B,f+m/3/89
IN INFLUENZA
B NEURAMINIDASE
250
201 v
LIKIKYGERY
TDTYHSYRNN
iLRTQESACN
CIGGDCYLMI
TDGSASGISE
301
L A
CRFLKIREGR
IIKEIFPTGR
“EHTEECTCG
FASU?Y.IECA
B B I CRDNSUMR
400
350 A
K
N N N
*/sing,79 B/Ore/m
N N
B/USSR,83 B,“iC,3,85 *,Le”,179,86 B/&m/6/86 *,Mem,3,89
N N N N PPVKLNVETD
TAEIRLMCTE
TYLDTPRPDD
GSITGPCESD
WL L I.
N
E
L D
GDKGRGGIKG
GF”HQRMASK
450
I : ! “SMKEPGWYS
IGRWYSRTMS
FGFEIKDKKC
D”PCIGIEn”
HDGGKKTWHS
KTERMGMEL’I
“KYDGDPWTD
SDALAPSGVM
466
D E E E E
IS I
TL
L
401
FIG. of the strains could
300 K K K K K K K
H H
B/Lee,40 B/Mary/59 B/W/8/73
B/Lee,40 B/Mary/59 B/HK/8/73 B/Sing/79 8,0re,80 B,“SSR,t?3 B,“iC,3,85 B,Len,179,86 B,Mem,6,86 B/~em,3,89
581
f
AATAIYCLMG
SGQLLWtn”T
G”E!mL*
2. Amino acrd sequence of the NA deduced from the nucleotide sequence of B/Mem/3/89 and differences in the amino acid sequences other strains as in the legend to Fig. 1. The signal-anchor and stalk domains are indicated. NA heads are released by trypsin from all the shown; the site of cleavage was determined by protein sequencing for B/Lee/40 and BIHW73 (Bossart et a/., 1988). Sequences which be sites of N-glycosyiation are underlined.
ences between strains were checked by direct comparison of the sequencing gel films. Forfourviruses (B/HKf8/73, B/Sing/222/79, B/USSR/ 100/83, and B/Mem/6/86), the NA gene was completely sequenced from both the parental virus and its HG reassortant. The NA nucleotide sequences of parental virus and HG reassortant were identical. The nucleotide sequence of the cDNA transcribed from RNA segment 6 of B/Mem/3/89 and differences in the NA genes of the earlier strains and in B/Lee/40 NA gene (Shaw et a/., 1982) are shown in Fig. 1. It is
clear from Fig. 1 that B/Lee/40 is rather different to the others, and this is also seen when predicted amino acid sequences are compared (Fig. 2). The quantitative differences in nucleotides, amino acids, and amino acids of the NA head domain starting from the trypsin cleavage site at amino acid 70 are shown in Table 1. These were derived by the program “Distances” of the University of Wisconsin packageversion 6.1 (Devereux eta/., 1984) but the results were modified so that insertions and deletions are counted as differences. The nucleotide sequence of B/Lee/40 is most similar to B/
as cDNA transcribed from the viral RNA using NA-specific primers. The sequences shown start at nucleotide 41. Data is available for most strains starting at nucleotide 30, and is included in the data submitted to Genbank although not shown here. The remaining 5’terminal cDNA sequences are available for B/Lee (Shaw eta/., 1982; Wei era/., 1987) and B/Maryland/59 (Stoeckle eta/., 1987).
582
AIR
ET AL.
TABLE DIFFERENCES BETWEEN INFLUENZA BlMdl59 B/Lee/40 BlMdl59 BIHW73 B/Sing/79 B/Ore/80 B/USSR/83 BlVicl85 B/Len/86 B/Mem/86 B/Mem/89 Note.
B/HW73
93, 25, 17 0
Lengths
B/Sing/79
96, 30. 20 32, 10, 5 0
of sequences
compared:
1517
B NAs (NUCLEOTIDES, B/Ore/80
98, 33, 22 43, 12, 8 29, 6, 5 0
1
BIUSSW83
108,32,21 45, 14,7 30,7,4 17,4, 1 0
nucleotides
(starting
at nucleotide
466 amino
116,36.24 53,19,11 41, 15,lO 30, 11) 7 30, 10,6 27. 8, 3 9.3,2 0
acids,
397 amino
BlMemI86
B/Mem/89
115.36,24 52, 19. 11 39, 15, 10 28, 11,7 28, 10,6 25, 8, 3 7,3,2 4, 0, 0 0
115,37,25 57, 22, 12 43, 14,9 33, 12.6 33, 10,5 31,g.z 2528, 5 22, 53 18, 5, 3 0
acids
in NA heads.
Trypsin-released NA heads from B/Sing/222/79 (HG), B/Oregon/80 (HG), B/USSR/83 (HG), B/Mem/86 (HG), and B/Mem/89 (HG) were purified as described and crystallized in hanging drops by vapor diffusion against a variety of precipitants. Single large crystals of each of these strains with the exception of B/USSR/83 and B/Mem/86 were grown (Fig. 4). B/USSR/83 only formed long, thin needles under the conditions used for the other B NAs, and B/Mem/86 yielded only amorphous material. Crystals of B/Sing/222/79 never grew large enough for diffraction studies. Preliminary data for B/Lee/40 and B/HW8/73 NA crystals have been previously reported (Bossart eta/., 1988), but both of these crystals show disorder making structure determination from these crystals difficult and maybe impossible. A different, well-ordered crystal form of B/Lee/40 has now been grown (Lin et al., 1990). Crystals of B/Ore/51 80 were also disordered, but B/Mem/89 NA when grown from PEG 3350 produced crystals, some of which showed disorder but a few were suitable for Xray diffraction analysis. Experiments to improve the crystal quality are still in progress. Preliminary crystallographic data for B/Mem/89 NA shows that these crystals diffract X-rays to 3.0-A resolution, and diffraction data have been collected for the three-dimensional 50
c
I
T
I I A
H
MNNATFNYTN * 1
B/Len/86
113,34,23 55. 19, 11 41, 16, 10 31, 12,7 31,11,6 28, 9, 3 0
41)
ACIDS IN HEADS)
2. Crystals of B neuraminidases
1
B/ORE/SO B/USSR/86 B/VIC/85 B/LEN/86 B/NE”,86 B,MEM,89
BNicl85
112,33,23 46, 15, 10 34, 11,7 23, 7,4 25,8, 3 0
Sing/79 while the amino acid sequence is most similar to BlMaryl59. Between B/Maryland/59 and B/Mem/89 there is less than 4% variation in nucleotide sequences, and less than 5% variation in amino acid sequences. The amino acid sequence differences are not distributed randomly in the polypeptide. There is a high proportion of changes in the stalk region of the NA, which was previously found to be more variable than the rest of the NA in influenza A subtypes (Blok and Air, 1982a,b; Els et a/., 1985). In influenza B this region is of higher significance, since it also codes for the NB polypeptide in a second reading frame (Shaw eta/., 1982). NB is made in large quantities during the infectious cycle, and its function is still unknown. It is a transmembrane protein with the N-terminal 18 amino acids exposed on the exterior of the cell membrane, 22 amino acids presumably spanning the membrane, and the C-terminal 60 amino acids forming an intracellular domain. Two carbohydrate addition sites have been identified in the external domain (Williams and Lamb, 1986, 1988). We therefore examined the amino acid sequences of NB which would be coded by the cDNA sequences shown in Fig. 1. The results are shown in Fig. 3. Surprisingly, the deduced NB sequences show considerable variation in the same region as the greatest variation in NA.
B/LEE/40 B/MARY/59 B/HK/8/73 B/SING/79
AMINO ACIDS, AMINO
L T T T T
I I
c H
N TN K NK TNK NN TN K
T T T T vNPISHIRCS
“IITICVSFI t
TTTKIFI-KNNCT
“1LT”FGYIA membrane
100 ” ” ” ” ” ”
PO
RVHK R”
T
A AR
G
A KH
K
H ” NNDIGLRERI
P KCSGCEPFCN
KRDDISSPRT
GVOIPSPILP
GLNLSESTPN
-3
FIG. 3. Amino acid sequences of the NB polypeptide deduced from the nucleotide sequence. as are the two sites of glycosylation (asterisks) determined by Williams and Lamb (1988).
The membrane-spanning
domain
is indicated,
VARIATION
IN INFLUENZA
6 NEURAMINIDASE
583
FIG. 4. Crystals of influenza B NA heads of four strains. The crystals were grown by vapor diffusion in hanging drops. The precipitant for B/ Stng/79, B/USSR/83, and B/Ore/80 NAs was 1.4 M KH,PO, + 3 AI K,HPO, mixed in the volume ratio of 2: 1, and for B/Mem/89 it was 15% PEG MW 3350 in 0.15 M NaCI. The BIMem/89 and B/Ore/80 crystals could be grown to about 0.4 mm on an edge, but B/Sing/79 has so far not yielded any crystals larger than 0.1 mm. B/USSR/83 has not yielded any crystals suitable for diffraction studies from any precipitant so far tested.
structure determination. graphic data are reported 3. Antigenic
properties
The preliminary cr-ystalloelsewhere (Lin eta/., 1990). of influenza
B NA
Table 2 shows the results of neuraminidase inhibition (NI) tests of viruses used in the sequence studies with monoclonal antibodies made to the NA of B/Lee/
40 [from B/HK (HG) reassortant virus; 5 hybridomas], B/HW8/73 (3 hybridomas), and B/Oregon/5/80 (1 hybridoma). In general, there was considerable cross-reactivity between the NAs of influenza B viruses. With the exception of B/Lee, there was cross-reactivity between the NAs of each of the field isolates. Monoclonal antibodies to B/Lee NA varied in cross-reactivity; one was specific only for B/Lee, one also cross-reacted
584
AIR ET AL. TABLE 2 CROSS-REACTIVITY
BETWEEN THE NEURAMINIDASES
Monoclonal antibodies to NA of
OF INFLUENZA B VIRUSES
NI titers to the following influenza viruses
Virus
No.
Lee140
Md/1/59
HWa/73
Sing1222179
Orel5/80
ussR1100783
Vic13/85
Mem/6/86
Mem/3/89
Lee/40 Lee/40 Lee140 Lee140 Lee/40 HW73 HW73 HW73 Ore/80
Bl 473/l 319 l/6 83 B19 B14 B12 i 48/i
3.85 3.33 3.08 2.80 3.92
< < 1.99 2.92 3.22 3.25 3.21 2.14 2.03
< < < 2.41 1.93 3.53 3.40 3.00 2.79
< 0.92 1.41 2.38 3.20 2.94 3.34 3.12 2.92
< < < 2.59 2.62 3.60 4.02 2.96 2.60
< < < < < < 2.24 3.50 2.19
< < < < < < 2.05 3.33 1.79
< < < < < < < 3.62 1.45
< < < < < < 2.02 2.78 2.02
3;co6 <
177,578-587
(1990)
Antigenic,
Sequence,
GILLIAN
and Crystal Variation
in Influenza
6 Neuraminidase’
M. AIR,*,‘W. GRAEME LAVER,t MING LUO,* STEPHEN J. STRAY,* GAYLA LEGRONE,* AND ROBERT G. WEBSTER+
*Department of Microbiology, University ofAlabama at Birmingham, Birmingham, Alabama 35294; John Curtin School of Medical Research, Australian National University, P. 0. Box 334, Canberra, and *St. Jude Children’s Research Hospital, P. 0. Box 3 18, Memphis, Tennessee Received
December
13, 1989; accepted
March
tlnfluenza Research Unit, ACT 260 1, Australia; 38 10 1
19, 1990
The neuraminidase (NA) genes of influenza B viruses B/Maryland/59, B/Hong Kong/8/73, B/Singapore/222/79, B/ Oregon/5/80, B/USSR/100183, BNictoria13185, BILeningradll79l86, B/Memphis/6/86, and B/Memphis/3/89 have been sequenced. The deduced amino acid sequences show high variability in the stalk domain of the NA, but a surprising degree of sequence conservation in the head region which carries all the antigenic and enzyme activity. The variable region coding for the neuraminidase stalk also translates into a variable section in the overlapping NB polypeptide, which is coded from a second reading frame that overlaps the first 100 amino acids of NA. The influenza B NAs are antigenically distinguishable with monoclonal antibodies in neuraminidase-inhibition tests, even when there is only one amino acid sequence difference. However, seven of nine escape mutants selected with monoclonal antibodies were distinguished only by the antibody used for selection. When NA heads of influenza B viruses are crystallized, there are remarkable differences in crystal morphology between neuraminidases which have very few sequence changes. 0 1990 Academic
Press.
Inc.
INTRODUCTION
in antigenic properties and sequence is similar to that of the HA (Webster et a/., 1982; Martinez et al., 1983; Colman and Ward, 1985). NA heads, which retain the tetrameric configuration of the intact molecule, can be released from several influenza viruses by proteases. These heads carry all the enzyme activity and the antigenic determinants. In some cases, the heads have been crystallized in a form suitable for structure analysis by X-ray diffraction methods. The three-dimensional structure of NA heads from two subtypes of influenza A (N2 and N9) has been determined (Varghese et al., 1983; Baker et al., 1987). Crystals of NA heads have also been obtained from several strains of influenza B (Bossart et al., 1988; Lin eta/., 1990) and from one of these complexed with antibody Fab (Laver et al., 1988), and the crystal structure determinations by X ray diffraction are in progress. In order to interpret the structure in terms of function and antigenicity, and to determine if the erratic antigenie and sequence variation found in influenza B HAS are duplicated in the NA, we have studied the antigenic properties and determined the NA gene sequence of 10 epidemic strains of influenza B, and have further investigated antigenic variation by selecting and characterizing escape mutants with monoclonal antibodies raised against the NA of influenza virus B/Lee/40.
There are fundamental differences in the way antigenie variation is manifested in influenza 6 viruses compared to the sequential pattern of antigenic drift seen in influenza A (Schild et a/., 1973; Lu et a/., 1983; Oxford et al., 1984; Hovanec and Air, 1984). Influenza B viruses change antigenic properties erratically, and the differences seem to involve fewer amino acid sequence changes than those found in influenza A antigens (Krystal et al., 1983; Yamashita et al., 1988). These analyses have so far been restricted to the hemagglutinin (HA). This surface glycoprotein enables the virus to bind to sialic acid receptors on the surface of host cells and then facilitates entry by a membrane fusion process. The second major surface glycoprotein of both influenza A and 6 viruses is the neuraminidase (NA), whose enzyme activity removes terminal sialic acid and thus destroys receptors for the virus. Although the role of this activity in viral infection has never been completely defined, it has been shown to assist in viral release and spread (Palese et al., 1974). Antibodies to the NA indirectly neutralize virus infectivity and protect animals against infection (Webster et al., 1982, 1988). In influenza A neuraminidases, the extent of variation
MATERIALS ’ Sequence EMBUGenBank M30639. ’ To whom 0042.6822/90
data from this article have been deposited Data Libraries under Accession Nos. reprint
requests
should
$3.00
CopyrIght 0 1990 by Academic Press, Inc. All rights of reproduction in any form reserved.
with the M30631-
AND
METHODS
Viruses The viruses used in this work MarylandI59, B/Hong Kong/8/73,
be addressed. 578
were B/Lee/40, B/ B/Hong Kong18173
579
VARIATION IN INFLUENZA B NEURAMINIDASE
(HG), B/Singapore/222/79, B/Singapore/222/79 (HG), B/Oregon/5/80, B/Oregon/5/80 (HG), B/USSR/l 00/83, B/USSR/l 00/83 (HG), B/Victoria/3/85, B/Leningrad/ 179/86, B/Memphis/6/86, B/Memphis/6/86 (HG), and B/Memphis/3/89 (HG). High-growth (HG) reassortant viruses were prepared by reasserting with B/HK (HG), which is a reassortant virus, B/Hong Kong/8/73 (HA) Lee/40 (NA) [RI, prepared and characterized as previously described (Webster, 1970; Wei et a/., 1987). The NAs of reassot-tant high growth (HG) viruses were identified using monoclonal antibodies. Except for B/Leningrad/l79/86, the influenza B viruses for this study were obtained from the repository at St. Jude’s Hospital, and grown either at St. Jude’s or at the Australian National University. B/Leningrad/86 viral cores were obtained from Dr. lgor Kharitenikov of the lvanovsky Institute of Virology, Moscow, USSR. Viruses were grown in embryonated eggs and purified by adsorption to and release from red blood cells followed by sucrose density centrifugation (Laver, 1969).
1988). The purified virus particles were suspended in 0.15 NI NaCI, Tris buffer pH 8.0, and 1 m&I CaCI, and incubated with TPCK-trypsin (Worthington Biochemicals) for 2 hr at 37”. Virus cores were removed by centrifugation (30,000 rpm for 1 hr) and the supernatant, which contained 100% of the NA activity, was centrifuged at 50,000 rpm (Ti 60 rotor) for 20 hr. The NA heads formed a pellet which, with most type B neuraminidases, had a crystalline appearance. The pellet was dissolved in saline (0.5 ml) and the NA heads purified by sucrose density gradient centrifugation (5-20% sucrose in 0.15 M NaCI, SW65 rotor, 58,000 rpm at 5” for 8 hr. Fractions containing NA were pooled, dialyzed against saline, and then against 70% saturated (NH&SO,. The precipitated NA heads were spun down, dissolved in the minimum amount of saline, and dialyzed against saline. The NA heads were crystallized from hanging drops by vapor diffusion using a variety of solvent systems. Preparation of viral RNA and sequence analysis
Serological tests Neuraminidase and neuraminidase-inhibition assays were done with fetuin substrate as described (AymardHenry et a/., 1973). Enzyme-linked immunosorbent assays (ELISA) were carried out as described (Kida et a/., 1982). Monoclonal antibodies Monoclonal antibodies were made as previously described (Webster and Berton, 1981; Laver eta/., 1988). The hybridomas were derived from spleen cells of a Balb/c mouse immunized with B/HK (HG) virus, or B/ Ore/80 virus (10 pg protein), and boosted 1 month later with 40 pg purified trypsin-released NA. The cells were fused 3 days postboost with myeloma P3X63-Ag8.653 cells. The hybridomas were screened by ELISA assay and by neuraminidase-inhibition tests. Hybridomas producing antibodies to NA were cloned in soft agar and grown as ascites in Balb/c mice. Ascitic fluid was collected from pristane-primed Balb/c mice. Selection of escape mutants Escape mutants were obtained by selection with neutralizing monoclonal antibodies as previously described (Webster and Berton, 1981) except that the selection and cloning were done in Madin Darby Canine Kidney (MDCK) cells. Preparation and crystallization
of NA heads
NA heads were prepared for crystallization similarly to the methods previously described (Bossart et a/.,
This was done as previously described (Air et a/., 1985). The starting material was either intact virus, or viral “cores” from which the NA had been removed by trypsin for crystallization. RESULTS 1. Sequences of NA genes of influenza B viruses isolated from 1940 to 1989 RNA templates were isolated from influenza viruses B/Mar-$and/59, B/Hong Kong18173 and (HG), BlSingapare/222/79 and (HG), B/Oregon/5/80 (HG), B/USSR/ 100/83 and (HG), B/Victoria/3/85, B/Leningrad/l 79/ 86, B/Memphis/6/86 and (HG), and B/Memphis/3/ 89 (HG). Sequences of cDNA transcribed from viral RNA were determined using the dideoxy method with the oligonucleotide primers specific for RNA segment 6 previously described (Wei et a/., 1987); they were designed from the B/Lee/40 NA sequence of Shaw et al. (1982) or the 5’ 56 nucleotide sequence of B/Maryland/59 NA gene of Stoeckle et al. (1987). In cases where these failed to prime or gave ambiguities, alternative oligonucleotides (usually 15-mers) were synthesized. Most of the NA cDNA sequences begin at nucleotide 31 (from a primer spanning nucleotides 14-29). In B/Len/86 there is apparently some change in the 3’ sequence of the vRNA, and no oligonucleotides made from this region would prime synthesis. The sequence for this strain begins at nucleotide 5 1, from a primer spanning 37-49. All the sequencing reactions were run more than once, the gels were read by at least two people, and the differ-
580
AIR ET AL.
FIG. 1. Nucleotide sequence of the NA gene of BIMem/3/89 and differences in the sequences of B/Mem/6/86, B/Len/179/86. B/Vie/3/85, B/ USSR/100/83, B/Ore/5/80, BISing1222179, B/Hong Kong/8/73, BIMatylandI59. and BILeei40. The B/Lee/40 sequence is that of Shaw et a/., (1982). This differs at 3 nucleotides from the B/Lee NA sequence in the B/HK (HG) reassortant virus (Wei et a/., 1987). The sequences are given
VARIATION
B/Lee,40 B/Mary/59 B/XK/8/J3 B/Sing/79 B,ore,ao B,“SSR/B3 B,“iC,3,85 B,Le”,i79,86 B,Mem,6,86 B,f+m/3/89
IN INFLUENZA
B NEURAMINIDASE
250
201 v
LIKIKYGERY
TDTYHSYRNN
iLRTQESACN
CIGGDCYLMI
TDGSASGISE
301
L A
CRFLKIREGR
IIKEIFPTGR
“EHTEECTCG
FASU?Y.IECA
B B I CRDNSUMR
400
350 A
K
N N N
*/sing,79 B/Ore/m
N N
B/USSR,83 B,“iC,3,85 *,Le”,179,86 B/&m/6/86 *,Mem,3,89
N N N N PPVKLNVETD
TAEIRLMCTE
TYLDTPRPDD
GSITGPCESD
WL L I.
N
E
L D
GDKGRGGIKG
GF”HQRMASK
450
I : ! “SMKEPGWYS
IGRWYSRTMS
FGFEIKDKKC
D”PCIGIEn”
HDGGKKTWHS
KTERMGMEL’I
“KYDGDPWTD
SDALAPSGVM
466
D E E E E
IS I
TL
L
401
FIG. of the strains could
300 K K K K K K K
H H
B/Lee,40 B/Mary/59 B/W/8/73
B/Lee,40 B/Mary/59 B/HK/8/73 B/Sing/79 8,0re,80 B,“SSR,t?3 B,“iC,3,85 B,Len,179,86 B,Mem,6,86 B/~em,3,89
581
f
AATAIYCLMG
SGQLLWtn”T
G”E!mL*
2. Amino acrd sequence of the NA deduced from the nucleotide sequence of B/Mem/3/89 and differences in the amino acid sequences other strains as in the legend to Fig. 1. The signal-anchor and stalk domains are indicated. NA heads are released by trypsin from all the shown; the site of cleavage was determined by protein sequencing for B/Lee/40 and BIHW73 (Bossart et a/., 1988). Sequences which be sites of N-glycosyiation are underlined.
ences between strains were checked by direct comparison of the sequencing gel films. Forfourviruses (B/HKf8/73, B/Sing/222/79, B/USSR/ 100/83, and B/Mem/6/86), the NA gene was completely sequenced from both the parental virus and its HG reassortant. The NA nucleotide sequences of parental virus and HG reassortant were identical. The nucleotide sequence of the cDNA transcribed from RNA segment 6 of B/Mem/3/89 and differences in the NA genes of the earlier strains and in B/Lee/40 NA gene (Shaw et a/., 1982) are shown in Fig. 1. It is
clear from Fig. 1 that B/Lee/40 is rather different to the others, and this is also seen when predicted amino acid sequences are compared (Fig. 2). The quantitative differences in nucleotides, amino acids, and amino acids of the NA head domain starting from the trypsin cleavage site at amino acid 70 are shown in Table 1. These were derived by the program “Distances” of the University of Wisconsin packageversion 6.1 (Devereux eta/., 1984) but the results were modified so that insertions and deletions are counted as differences. The nucleotide sequence of B/Lee/40 is most similar to B/
as cDNA transcribed from the viral RNA using NA-specific primers. The sequences shown start at nucleotide 41. Data is available for most strains starting at nucleotide 30, and is included in the data submitted to Genbank although not shown here. The remaining 5’terminal cDNA sequences are available for B/Lee (Shaw eta/., 1982; Wei era/., 1987) and B/Maryland/59 (Stoeckle eta/., 1987).
582
AIR
ET AL.
TABLE DIFFERENCES BETWEEN INFLUENZA BlMdl59 B/Lee/40 BlMdl59 BIHW73 B/Sing/79 B/Ore/80 B/USSR/83 BlVicl85 B/Len/86 B/Mem/86 B/Mem/89 Note.
B/HW73
93, 25, 17 0
Lengths
B/Sing/79
96, 30. 20 32, 10, 5 0
of sequences
compared:
1517
B NAs (NUCLEOTIDES, B/Ore/80
98, 33, 22 43, 12, 8 29, 6, 5 0
1
BIUSSW83
108,32,21 45, 14,7 30,7,4 17,4, 1 0
nucleotides
(starting
at nucleotide
466 amino
116,36.24 53,19,11 41, 15,lO 30, 11) 7 30, 10,6 27. 8, 3 9.3,2 0
acids,
397 amino
BlMemI86
B/Mem/89
115.36,24 52, 19. 11 39, 15, 10 28, 11,7 28, 10,6 25, 8, 3 7,3,2 4, 0, 0 0
115,37,25 57, 22, 12 43, 14,9 33, 12.6 33, 10,5 31,g.z 2528, 5 22, 53 18, 5, 3 0
acids
in NA heads.
Trypsin-released NA heads from B/Sing/222/79 (HG), B/Oregon/80 (HG), B/USSR/83 (HG), B/Mem/86 (HG), and B/Mem/89 (HG) were purified as described and crystallized in hanging drops by vapor diffusion against a variety of precipitants. Single large crystals of each of these strains with the exception of B/USSR/83 and B/Mem/86 were grown (Fig. 4). B/USSR/83 only formed long, thin needles under the conditions used for the other B NAs, and B/Mem/86 yielded only amorphous material. Crystals of B/Sing/222/79 never grew large enough for diffraction studies. Preliminary data for B/Lee/40 and B/HW8/73 NA crystals have been previously reported (Bossart eta/., 1988), but both of these crystals show disorder making structure determination from these crystals difficult and maybe impossible. A different, well-ordered crystal form of B/Lee/40 has now been grown (Lin et al., 1990). Crystals of B/Ore/51 80 were also disordered, but B/Mem/89 NA when grown from PEG 3350 produced crystals, some of which showed disorder but a few were suitable for Xray diffraction analysis. Experiments to improve the crystal quality are still in progress. Preliminary crystallographic data for B/Mem/89 NA shows that these crystals diffract X-rays to 3.0-A resolution, and diffraction data have been collected for the three-dimensional 50
c
I
T
I I A
H
MNNATFNYTN * 1
B/Len/86
113,34,23 55. 19, 11 41, 16, 10 31, 12,7 31,11,6 28, 9, 3 0
41)
ACIDS IN HEADS)
2. Crystals of B neuraminidases
1
B/ORE/SO B/USSR/86 B/VIC/85 B/LEN/86 B/NE”,86 B,MEM,89
BNicl85
112,33,23 46, 15, 10 34, 11,7 23, 7,4 25,8, 3 0
Sing/79 while the amino acid sequence is most similar to BlMaryl59. Between B/Maryland/59 and B/Mem/89 there is less than 4% variation in nucleotide sequences, and less than 5% variation in amino acid sequences. The amino acid sequence differences are not distributed randomly in the polypeptide. There is a high proportion of changes in the stalk region of the NA, which was previously found to be more variable than the rest of the NA in influenza A subtypes (Blok and Air, 1982a,b; Els et a/., 1985). In influenza B this region is of higher significance, since it also codes for the NB polypeptide in a second reading frame (Shaw eta/., 1982). NB is made in large quantities during the infectious cycle, and its function is still unknown. It is a transmembrane protein with the N-terminal 18 amino acids exposed on the exterior of the cell membrane, 22 amino acids presumably spanning the membrane, and the C-terminal 60 amino acids forming an intracellular domain. Two carbohydrate addition sites have been identified in the external domain (Williams and Lamb, 1986, 1988). We therefore examined the amino acid sequences of NB which would be coded by the cDNA sequences shown in Fig. 1. The results are shown in Fig. 3. Surprisingly, the deduced NB sequences show considerable variation in the same region as the greatest variation in NA.
B/LEE/40 B/MARY/59 B/HK/8/73 B/SING/79
AMINO ACIDS, AMINO
L T T T T
I I
c H
N TN K NK TNK NN TN K
T T T T vNPISHIRCS
“IITICVSFI t
TTTKIFI-KNNCT
“1LT”FGYIA membrane
100 ” ” ” ” ” ”
PO
RVHK R”
T
A AR
G
A KH
K
H ” NNDIGLRERI
P KCSGCEPFCN
KRDDISSPRT
GVOIPSPILP
GLNLSESTPN
-3
FIG. 3. Amino acid sequences of the NB polypeptide deduced from the nucleotide sequence. as are the two sites of glycosylation (asterisks) determined by Williams and Lamb (1988).
The membrane-spanning
domain
is indicated,
VARIATION
IN INFLUENZA
6 NEURAMINIDASE
583
FIG. 4. Crystals of influenza B NA heads of four strains. The crystals were grown by vapor diffusion in hanging drops. The precipitant for B/ Stng/79, B/USSR/83, and B/Ore/80 NAs was 1.4 M KH,PO, + 3 AI K,HPO, mixed in the volume ratio of 2: 1, and for B/Mem/89 it was 15% PEG MW 3350 in 0.15 M NaCI. The BIMem/89 and B/Ore/80 crystals could be grown to about 0.4 mm on an edge, but B/Sing/79 has so far not yielded any crystals larger than 0.1 mm. B/USSR/83 has not yielded any crystals suitable for diffraction studies from any precipitant so far tested.
structure determination. graphic data are reported 3. Antigenic
properties
The preliminary cr-ystalloelsewhere (Lin eta/., 1990). of influenza
B NA
Table 2 shows the results of neuraminidase inhibition (NI) tests of viruses used in the sequence studies with monoclonal antibodies made to the NA of B/Lee/
40 [from B/HK (HG) reassortant virus; 5 hybridomas], B/HW8/73 (3 hybridomas), and B/Oregon/5/80 (1 hybridoma). In general, there was considerable cross-reactivity between the NAs of influenza B viruses. With the exception of B/Lee, there was cross-reactivity between the NAs of each of the field isolates. Monoclonal antibodies to B/Lee NA varied in cross-reactivity; one was specific only for B/Lee, one also cross-reacted
584
AIR ET AL. TABLE 2 CROSS-REACTIVITY
BETWEEN THE NEURAMINIDASES
Monoclonal antibodies to NA of
OF INFLUENZA B VIRUSES
NI titers to the following influenza viruses
Virus
No.
Lee140
Md/1/59
HWa/73
Sing1222179
Orel5/80
ussR1100783
Vic13/85
Mem/6/86
Mem/3/89
Lee/40 Lee/40 Lee140 Lee140 Lee/40 HW73 HW73 HW73 Ore/80
Bl 473/l 319 l/6 83 B19 B14 B12 i 48/i
3.85 3.33 3.08 2.80 3.92
< < 1.99 2.92 3.22 3.25 3.21 2.14 2.03
< < < 2.41 1.93 3.53 3.40 3.00 2.79
< 0.92 1.41 2.38 3.20 2.94 3.34 3.12 2.92
< < < 2.59 2.62 3.60 4.02 2.96 2.60
< < < < < < 2.24 3.50 2.19
< < < < < < 2.05 3.33 1.79
< < < < < < < 3.62 1.45
< < < < < < 2.02 2.78 2.02
3;co6 <
