20
Arch Virol (1996) 141:2057-2076 _ Archives ..... Vi rology © Springer-Verlag 1996 Printed in Austria Divergence of VP7 genes of G1 rotaviruses isolated from infants vaccinated with reassortant rhesus rotaviruses Q. Jin 1'2, R. L. Ward3, D. R. Knowlton3, Y. B. Gabbay 4, A. C. Linhares4, R. Rappaport 5, P. A. Woods 1, R. I. Glass 1, and J. R. Gentsch 1 1Viral Gastroenteritis Section,Division of Viral and Rickettsial Diseases, National Center for InfectiousDiseases, Centers for Disease Control and Prevention, Atlanta, Georgia, U.S.A.; 2National Laboratory of Molecular Virology and Genetic Engineering, Institute of Virology, Chinese Academyof PreventiveMedicine,Beijing, People's Republic of China; 3Division of Infectious Diseases,Children's Hospital Medical Center, Cincinnati, Ohio, U.S.A.; 4VirologySection, Instituto Evandro Chagas, Para, Brazil; 5Wyeth-Ayerst Research, Biotechnology and MicrobiologyDivision, Radnor, Pennsylvania,U.S.A. AcceptedJuly 4, 1996 Summary. A large placebo-controlled efficacy trial of the rhesus tetravalent (RRV-TV) and serotype G1 monovalent (RRV-S1) rotavirus vaccines was conducted in 1991-1992 at 24 sites across the United States. Protection was 49% and 54% against all diarrhea but 80% and 69% against very severe gastroenteri- tis for the two vaccines, respectively. Post-vaccination neutralizing antibody titers to the G1 Wa strain, whose VP7 protein is nearly identical to that of the D strain of rotavirus contained in both vaccines, did not correlate with protec- tion against subsequent illness with G1 strains. This result raised the possibility that in infants who developed post-vaccination neutralizing antibody to Wa, breakthrough (i.e., vaccine failure-the occurrence of rotavirus diarrhea after immunization) may have been due to infection by G1 strains that were sufficient- ly antigenically distinct from the vaccine strain to evade the neutralizing antibodies elicited by vaccination. To test this hypothesis, we initially compared post-vaccination neutralizing antibody titers of vaccinees against Wa and G1 breakthrough strains using sera flom subjects who experienced breakthrough. Post-immunization neutralizing antibody titers to Wa elicited by vaccination were significantly (P < 0.001) greater than to the breakthrough strains subse- quently obtained from these subjects. This difference did not, however, correlate with lack of protection since similar differences in titer to Wa and breakthrough strains were found using post-vaccination sera from vaccinees who either experienced asymptomatic rotavirus infections or no infections. To determine

Divergence of VP7 genes of G1 rotaviruses isolated from infants vaccinated with reassortant rhesus rotaviruses

  • Upload
    q-jin

  • View
    215

  • Download
    2

Embed Size (px)

Citation preview

Arch Virol (1996) 141:2057-2076

_ A r c h i v e s . . . . .

Vi rology © Springer-Verlag 1996 Printed in Austria

D i v e r g e n c e o f V P 7 g e n e s o f G 1 rotav iruses i so la ted f r o m in fant s

v a c c i n a t e d with r e a s s o r t a n t rhesus ro tav iruses

Q. Jin 1'2, R. L. Ward 3, D. R. Knowlton 3, Y. B. Gabbay 4, A. C. Linhares 4, R. Rappaport 5, P. A. Woods 1, R. I. Glass 1, and J. R. Gentsch 1

1 Viral Gastroenteritis Section, Division of Viral and Rickettsial Diseases, National Center for Infectious Diseases, Centers for Disease Control and Prevention, Atlanta, Georgia, U.S.A.; 2 National Laboratory of Molecular Virology and Genetic Engineering, Institute of Virology,

Chinese Academy of Preventive Medicine, Beijing, People's Republic of China; 3 Division of Infectious Diseases, Children's Hospital Medical Center, Cincinnati, Ohio, U.S.A.; 4 Virology Section,

Instituto Evandro Chagas, Para, Brazil; 5 Wyeth-Ayerst Research, Biotechnology and Microbiology Division,

Radnor, Pennsylvania, U.S.A.

Accepted July 4, 1996

Summary. A large placebo-controlled efficacy trial of the rhesus tetravalent (RRV-TV) and serotype G1 monovalent (RRV-S1) rotavirus vaccines was conducted in 1991-1992 at 24 sites across the United States. Protection was 49% and 54% against all diarrhea but 80% and 69% against very severe gastroenteri- tis for the two vaccines, respectively. Post-vaccination neutralizing antibody titers to the G1 Wa strain, whose VP7 protein is nearly identical to that of the D strain of rotavirus contained in both vaccines, did not correlate with protec- tion against subsequent illness with G1 strains. This result raised the possibility that in infants who developed post-vaccination neutralizing antibody to Wa, breakthrough (i.e., vaccine failure-the occurrence of rotavirus diarrhea after immunization) may have been due to infection by G1 strains that were sufficient- ly antigenically distinct from the vaccine strain to evade the neutralizing antibodies elicited by vaccination. To test this hypothesis, we initially compared post-vaccination neutralizing antibody titers of vaccinees against Wa and G1 breakthrough strains using sera flom subjects who experienced breakthrough. Post-immunization neutralizing antibody titers to Wa elicited by vaccination were significantly (P < 0.001) greater than to the breakthrough strains subse- quently obtained from these subjects. This difference did not, however, correlate with lack of protection since similar differences in titer to Wa and breakthrough strains were found using post-vaccination sera from vaccinees who either experienced asymptomatic rotavirus infections or no infections. To determine

2058 Q. Jin et al.

the genetic basis for these differences, we compared the VP7 gene sequences of Wa with vaccine strain D, 12 G1 breakthrough strains, and 3 G1 control strains isolated during the same trial from placebo recipients. All breakthrough strains were distinct from Wa and D in antigenically important regions throughout the VP7 protein, but these differences were conserved between breakthrough and placebo strains. Furthermore, a comparative analysis of the deduced amino sequences from VP7 genes of G1 rotaviruses from 12 countries indicated that four distinct lineages have evolved. All breakthrough and control strains from the U.S. vaccine trial were in a lineage different from strain D, the serotype G1 vaccine strain. Although the overall results do not support our original hypo- thesis that immune selection of antigenically distinct escape mutants led to vaccine breakthrough in subjects with a neutralization response to Wa, it cannot be excluded that breakthrough could be partially due to antigenic differences in the VP7 proteins of currently circulating G1 strains.

Introduction

Group A rotaviruses are the most important etiologic agents of severe gastroen- teritis in children, and as a result, the development of an efl'ective vaccine against severe rotavirus diarrhea is a top public health priority [18]. Numerous rotavirus serotypes have been characterized on the basis of neutralizing anti- bodies to the outer capsid proteins VP7 (G-serotypes) and VP4 (P-serotypes), of which at least four G-serotypes and one P-serotype (composed of two subtypes, 1A and 1B) have been shown to be of major importance in human disease [10, 17, 32, 34, 37]. Antibodies to the VP7 and VP4 proteins protect against disease in animal models, suggesting that development of type-specific neutralizing anti- bodies may be important in protection against human illness [25]. Early observations that high levels of neutralizing antibodies to the infecting serotype correlated with protection against natural rotavirus disease in infants were consistent with this hypothesis and suggested that an effective vaccine would have to produce immunity to common human G types [2]. However, studies showing that type-specific protection is elicited by natural rotavirus infection are contradicted by other studies showing that symptomatic reinfection of individ- ual infants with the same G-serotype is possible [26]. Whether reinfections occur because of antigenic variation of these strains due to short-lived immunity in infants who seroconverted to the primary infecting strain, or because titers of serum neutralizing antibody are not directly related to protection, is unknown.

The first rotavirus vaccine candidates were heterologous simian (RRV) and bovine (WC3 and RIT4237) rotavirus strains whose G-serotypes were different from 3 or all 4 of the most common rotaviruses isolated from children with diarrhea. These attenuated, live, oral ("Jennerian") vaccines failed to induce consistently heterotypic protection to common human rotavirus serotypes in some settings, reinforcing the hypothesis that serotype-specific immunity may be necessary [19]. In an effort to elicit neutralizing antibody to some of the most important human rotaviruses, reassortant strains with G1, G2, G3, and G4 VP7

Divergence of G1 rotavirus VP7 genes 2059

genes, or a P1A VP4 gene, and most or all other genes from heterotypic RRV (serotype G3) or bovine WC3 (serotype G6) rotaviruses have been constructed [3, 4, 19, 24].

A large efficacy trial comparing the RRV-TV vaccine, containing G1 to G4 VP7 genes, to the RRV-S1 vaccine containing G1 VP7, and placebo con- trol was recently carried out at 24 sites across the U.S. The overall efficacies for RRV-S1 and RRV-TV were found to be 54% and 49%, respectively, for all diarrhea and 69% and 80% protection, respectively, for very severe gastroenteritis [30]. Between 84% (RRV-S1) and 92% (RRV-TV) of the vac- cinees seroconverted following vaccination based on rises in rotavirus IgA and neutralizing antibody to RRV, but < 35 % had neutralizing antibody rises to any human G1 to G4 prototype strain [30]. Furthermore, there was no significant association between seroconversion to the G1 strain and protection against G1 rotaviruses, the predominant circulating strains during the trial (M. E. Mack and R. L. Ward, unpubl, data). Finally, it was also found that some infants who seroconverted to G1 rotavirus following vaccination experienced rota- virus diarrhea (i.e. vaccine failure or breakthrough) with the same serotype, including several subjects who developed relatively high titers of neutraliz- ing antibody to the G1 strain Wa (D. R. Knowlton and R. L. Ward, unpubl. data).

Although there are a number of possible explanations for these observations, one possible reason is that there were sufficient antigenic differences between the VP7 proteins of the G1 vaccine strain and the circulating G1 strains to permit the latter to escape neutralization and cause illness. To test this hypothesis, we first analyzed the neutralizing antibody titers of post-vaccination sera from infants who had vaccine-induced serotype G1 neutralizing antibodies to strain Wa but still experienced serotype G1 illnesses. For test rotaviruses, we used both the subjects illness strains and that Wa strain whose VP7 gene is nearly identical to the G1 strain in the vaccine (i.e., D. strain). Representative illness strains, whose neutralization titers to the corresponding post-vaccination infant serum ranged from values comparable to those found for the Wa strain to titers that were much lower, were selected for sequencing. Sequences of the entire VP7 gene open reading frames and the hypervariable regions of VP4 from these strains were compared to those of the vaccine strain D and prototype rotavirus Wa, control strains isolated at the same time and in the same locale, and diverse community diarrhea strains from multiple countries.

Materials and methods

Subjects and study design for the rotavirus vaccine trial

Enrollment of subjects for the rotavirus vaccine trial (312US) conducted at 24 sites across the U.S. has been described [30]. In brief, 1213 infants aged 5-25 weeks were administered three separate doses of RRV-S1 (412 subjects), RRV-TV (402 subjects), or placebo (399 subjects) and monitored for rotavirus disease during the subsequent rotavirus season (1991 1992). The RRV-S1 vaccine consisted of 4 × l0 s PFU of the serotype G1 strain D x RRV

2060 Q. Jin et al.

reassortant, while the RRV-TV vaccine consisted of 1 x 105 PFU of the serotype G3 RRV, plus 1 x 105 PFU each of the G1, G2 and G4 reassortants containing 10 genes from RRV and the VP7 genes of human rotavirus strains D (serotype G1), DS-1 (serotype G2) and ST-3 (serotype G4) [3, 19, 23, 24]. Serologic evaluation was performed on children who developed rotavirus-associated diarrhea following the first dose of vaccine until the end of the rotavirus season, and on 450 randomly selected control children.

Rotavirus strains

Breakthrough strains

Vaccine failure (designated breakthrough in this study) among children participating in the 312US vaccine trial was defined as experiencing an episode of rotavirus diarrhea 2 weeks or more after receiving the third dose of vaccine regardless of the severity of the gastroenteritis [30]. Fecal suspensions were prepared from rotavirus positive specimens and the strains were isolated by cultivation in MA 104 cells as described previously [36]. Atternpts to culture-adapt virus were made only with G1 isolates from subjects who had measurable (i.e., a titer of >~ 10) post-vaccination neutralizing antibody titers to the Gt prototype strain Wa. In all, 30 out of 34 serotype G1 specimens were successfully cultured from U.S. breakthrough strains. Because of their limited number, non-G1 breakthrough strains were not examined in this study. Three control strains collected during the same year from children who were placebo recipients (controls) in the vaccine trial were also sequenced in this study.

Community diarrhea strains

Other serotype G1 rotaviruses were obtained between t983 to 1988 from infants with diarrhea in Bangladesh (Ban-48/-59, 1988), China (Chi-45/-46, 1987), Costa Rica (Cos-69/-70, 1987), Egypt (Egy-7/-8, 1988), Israel (Isr-55/-56, 1987), Korea (Kor-54/-64, 1988), Zambia (Zam-52, 1988), and Brazil (Brz-2, 1985) ([21, 37]° J. Mathewson, pers. comm.). In addition, 4 strains from Brazil (strain pairs Brz-3 and Brz-4, and Brz-5 and Brz-6 isolated in t991-1992, respectively) were isolated from 2 infants who had sequential infections [22]. These 4 strains were isolated from placebo controls participating in a RRV-TV vaccine trial in Belem, Brazil. Strains Wa and D were originally isolated in the U.S. in 1974, isolates from Japan (Jpn-417/-418/-421) were obtained in Tokyo in 1991 and strains fi'om Australia (Aus-90/-91) were collected in Melbourne in 1990 and 1991 [1, t6, 28, 38].

Serum neutralizin 9 antibody determinations

Serum specimens collected 3 to 5 weeks after the third dose of vaccine from subjects who subsequently developed serotype G1 rotavirus gastroenteritis were tested for neutralizing antibody to the reference human serotype G1 strain Wa and to the corresponding culture- adapted infecting strains by an enzyme immunoassay neutralization procedure previously described [20].

RNA extraction and cDNA synthesis

Viral double-stranded (ds)-RNA was prepared by a modification of a previously described glass powder method [I1]. Briefly, 10% suspensions of fecal specimens or virus infected MA-104 cell tysates were extracted with phenol-chloroform (i:I), and ds-RNA was isolated

Divergence of G1 rotavirus VP7 genes 2061

by adsorption to RNAID (Bio 101, Inc., La Jolla, CA) and etution with water. Synthesis of cDNA was carried out by reverse transcription-polymerase chain reaction (RT-PCR) from each VP7 and VP4 encoding gene or gene segment. The oligonucleotide primer pairs Beg9 (nucleotides 1-28 of the VP7 gene of strain Wa), Rvg9 (nucleotides 1062- 1044 of the VP7 gene of strain Wa) and con3 (nucleotides 11-32 of strain Wa gene 4), Con2 (nucleotides 868-887 of strain Wa gene 4) were used for amplification of full length VP7 genes and VP4 fragments, respectively [11, 13]. The RT-PCR was performed by one cycle of reverse transcription (60min at 42°C), 35 cycles of PCR (lmin at 94°C, 2 min at 42 °C, and 3 rain at 72 °C) and a final extension cycle (8 rain at 72 °C). The ampli- fied PCR products were separated by electrophoresis on 1.2% SeaKem GTG agarose gels (FMC Bioproducts, Rockland, ME) and then purified using a QIAEX Gel extraction Kit (QIAGEN Inc., Chatsworth, CA) according to the protocol provided by the manu- facturer.

Dideo x ynucleotide sequencing

The nucleotide sequences of RT-PCR derived cDNA fragments were determined by the dideoxy chain termination method with fluorescently labeled dideoxynucleotides and Amplitaq DNA polymerase from Applied Biosystems (Foster City, CA) [31]. PCR pro- ducts were sequenced following the manufacturers instructions (30 second at 96°C, 15 second at 50°C, and 4min at 60°C, 25 cycles) using primers aBT1 (nucleotides 3t4-335 of the VP7 gene of strain Wa) and 4Con6 (5'-TATATTCATTACACTTAGATTCTTG-3', nucleotides 640-664 of gene 4). The sequences of primers Con2 and Con3 which amplify nucleotides 11 887 of the VP4 gene and primers Beg9 and End9 which amplify the entire VP7 gene have been described [-11, 13]. Cycle-sequencing products were electrophoresed in 6% polyacrylamide gels containing 8 M urea on the ABI model 373A automatic sequencing apparatus.

Data analysis

Sequences were analyzed using the University of Wisconsin Genetics Computer Group (GCG) and the Phylip35 computer programs [6, 9]. Distance matrices for deduced VP7 amino acid sequences used to generate phylogenetic trees were calculated with the Protdist program using the Kimura protein method. The evolutionary tree presented in this work was prepared with the Kitsch module of the Phylip3.5 software programs using a neighbor- joining method [9]. Trees generated with other methods, including the Fitch module of the Phylip3.5 programs, and the Growtree program of the GCG package gave comparable results differing in some details, but leading to the same overall conclusions as presented below, i.e. that several very distinct strain lineages exist, and that strain D, the current G1 vaccine strain, resides in a lineage distinct from serotype G1 clinical isolates of the U.S. Examples of these differences were that the Growtree program of the GCG package classified lineage IV as an intermediate cluster between lineages I and II and that the Kitch program clustered strain Egy-7 as a highly distinct sub-lineage of lineage II, rather than as an equally distinct sub-lineage of lineage IV. Previously published VP7 gene sequences used in the construction of dendrograms or amino acid alignments (Wa, D, Jpn-417/-418/-421, Aus-90/- 91, Rus-91) were obtained from GenBank or the publications [16, 28, 38].

Accession numbers

The VP7 and VP4 sequences in this paper have been assigned GenBank accession numbers U26362 to U26395 and U26754 to U26766, respectively.

2062 Q. Jin et al.: Divergence of G1 rotavirus VP7 genes

Results

Neu~ralizin 9 antibody titers of post-vaccination sera amon 9 vaccinees who experienced serotype G1 rotavirus illnesses

We first examined the hypothesis that G1 rotavirus illnesses following vaccina- tion in subjects who produced neutralizing antibody to the G1 prototype strain Wa could be due to low neutralizing antibody titers to circulating G1 strains. Thirty serotype G1 strains from vaccinees who had post-vaccination neutraliz- ing antibodies to strain Wa and later became ill with rotavirus diarrhea were culture-adapted. Each child's post-vaccination serum neutralizing antibody titer was determined against the child's individual strain and reference isolate Wa whose VP7 protein differs by a single residue, i.e. amino acid (aa) 123, from that of the D strain used in the RRV-S1 vaccine and included as one component of the RRV-TV vaccine [16]. The overall geometric mean titers (GMT's) for the 30 subjects were significantly (P < 0.001) higher against strain Wa (65.1) than to the illness strains (30.4) (Table 1). Thus, subjects who developed neutralizing anti- body to Wa following vaccination generally had lower titers to the circulating G1 strains that caused subsequent illnesses. For some subjects, the difference was quite large (e.g., a > 50-fold difference in titer for the strain isolated from subject Oh-11), while in others it was negligible (e.g., subject Va-12). This result indicated that either the strains causing illnesses varied from one another in thief VP7 neutralization epitopes, some being very similar to Wa and others being quite different, or the relative amounts of neutralizing antibody elicited against different VP7 epitopes of the G1 vaccine strain varied extensively between subjects, especially those given RRV-S 1.

To differentiate between these first two possibilities, neutralizing antibody titers of post-vaccination sera from subjects who experienced G1 rotavirus illnesses were retested against two illness strains (Mo-12 and Va-12) along with Wa. The sera were chosen from the ten subjects with G1 rotavirus illnesses who developed the highest post-vaccination neutralizing antibody to Wa in order to permit the maximum possibility to observe variations in triers. Also included were sera from vaccinees with high post-vaccination neutralizing antibody titers to Wa who either experienced an asymptomatic G1 rotavirus infection or who had no rotavirus infection based on neutralizing antibody titers in sera collected after the rotavirus season. The latter groups were included to determine whether the anti-Wa and anti-illness strain titers in post-vaccination sera were more nearly equal in vaccinees who did not experience G1 rotavirus illnesses (vaccine successes) than in vaccinees who did experience G1 illnesses.

The range of serum neutralizing antibody titers of the ten vaccinees included who had experienced G1 illnesses (Table 2) were generally very similar to the range of titers found against the rotaviruses that caused the illnesses (Table 1). Thus, it appeared that the relative responses against different VP7 neutralization epitopes on the G1 vaccine strain varied between subjects. However, the titers against Mo-12 and Va-12 were sometimes quite different (see subjects Ky-206 and RI-49), and all but 2 of the 26 subjects had higher titers to Va-12 than to

Table 1. Post-vaccination neutralizing antibody titers in vaccinees who experienced G1 rotavirus illnesses measured against the Wa and culture-adapted illness strains

Subject no. Vaccine Neutralizing antibody titer

Wa illness Wa illness strain strain

O h - l l RRV-S1 252 5 50.4 b Ar-7 185 17 10.9 Oh-25 115 10 10.5 ~ DC-6 182 22 8.3 b Oh-31 454 102 4.5 Ar- 11 22 5 4.4 RI-14 46 11 4.2 b Oh-40 20 5 4.0 b Az-910 66 18 3.7 Pa-4 27 8 3.4 Mo-12 255 96 2.7 b Md-233 102 44 2.3 Va-12 520 403 1.3 b Ga-125 38 28 1.3 Ga-192 93 100 0.9 Ca-15 89 113 0.8 Oh-8 11 14 0.8 Nb-5 70 98 0.7

G M T subtotal:

RRV-TV

83.6 26.9" 3.11

Nb-49 121 49 2.5 b

Ky-274 15 6 2.5 Ga-122 164 76 2.2 Tn-16 43 20 2.2 b RI-43 57 55 1.0 b

Mo-38 21 20 1.0 Md-216 53 56 1.0 b RI-17 29 31 0.9 RI-39 24 29 0.8 b Tx-75 30 36 0.8 NY-7 78 98 0.8 Oh-7 38 63 0.6

G M T subtotal: 44.5 36.4 1.22 G M T total: 65.1 30.4 a 2.14

a p < .00t relative to anti-Wa G M T (Student t test) b Chosen for sequence analysis study. Of the 12 strains chosen, the corresponding severity

scores of the subjects were: ~< 8 (3 subjects), > 8 to ~< 14 (6 subjects), > 114 (2 subjects), and not determined (1 subject) [30]. These scores were not significantly different from those of other vaccinees or placebo recipients who experienced rotavirus disease (R. L. Ward, unpubl, data). Overall, for this vaccine trial, no correlation was found between severity of illness caused by G 1 strains and the subjects' neutralizing antibody titers to strain Wa (R. L. Ward, unpubl. data)

2064

Table 2.

Q. Jin et al.

Post-vaccination neutralizing antibody titers of vaccinees against Wa and illness strains Mo-12 and Va-12

Subjects Subject no. Neutralizing Ab titers a

Wa Mo-12 Va-12

Symptomatic Mo-12 373 117 125 infections: Va-12 338 159 352

RRV-S1 Ar-7 272 18 35 Vaccinees Md-233 92 5 26

Oh-3t 212 68 99 Oh-l l 198 5 5 DC-6 159 5 5 Oh-25 137 5 7

RRV-TV Ga-122 152 29 63 Vaccinees Nb-49 132 43 90

Asymptomatic Ky-206 392 5 93 infections: Oh- 17 208 34 86

RRV-S1 RI-49 207 6 97 Vaccinees Ga- 173 377 63 164

Ga-212 195 19 38 Nb-3 172 13 45 SC-43 110 18 42

'Md-15 107 10 28

No infection: Ky-277 363 41 82 RRV-S1 RI-60 840 80 121 Vaccinees Ga-170 1320 85 115

Md-131 327 248 300 Mo-20 536 79 71 Md-ll7 654 244 196 NY-2 466 76 192 Tn-41 448 11 33

Neutralizing antibody titers were significantly higher to Wa than to Mo-12 and Va-12 for all groups (P < .001, Students t test)

Mo-12 (Table 2). Thus, there were also antigenic differences between the illness strains based on these results.

The neutral izing an t ibody titers of all three groups of subjects (symptomat i - cally infected, asymptomat ica l ly infected, and not infected) were significantly (P < 0.001) higher against Wa than to Mo-12 and Va-12 illness strains using post-vaccinat ion sera (Table 2). Fur the rmore , a l though the 16 "protec ted" vaccinees had higher neutral izing an t ibody titers to Wa, Mo-12 and Va-12 than the o ther two groups, Mo-12 and Va-12 did not have propor t iona te ly higher neutral izing an t ibody titers compared to Wa than vaccinees who experienced G1 rotavirus illnesses (Table 2).

Divergence of G1 rotavirus VP7 genes 2065

Sequence analysis of the VP7 gene of U.S. breakthrough and control strains

Because of the overall significant differences in anti-Wa neutralizing antibody titers found in vaccinees relative to the titers in the circulating strains examined, it seemed possible that the circulating G1 strains may have significant differences from Wa in their VP7 gene sequences. To test this hypothesis, we selected 12 breakthrough strains that had neutralizing antibody titer ratios (titer against Wa/titer against subject's strain determined with each subject's post-vaccination serum) between approximately 50:1 and 1:1, and compared their VP7 sequences to 3 control strains isolated during the same rotavirus season from placebo recipients at 3 of the study sites (see Table 1).

The entire 978 nucleotide open reading frame for the VP7 genes of 12 breakthrough and 3 control strains from placebo recipients was sequenced and an alignment of their deduced amino acid sequences was prepared using strain Wa as the reference virus. A comparison of selected regions of this alignment is shown (Table 3). It was striking that most illness strains, both from vaccinees and placebo recipients, shared one or more amino acid substitutions relative to the Wa strain in previously characterized antigenic regions A, B, and C, and other changes in or near a hypervariable region designated VR4, and at amino acids 170 and 307 [8, 12, 14]. Most of these substitutions were found not to be associated with culture adaptation, since strains whose sequence was determined using fecal rotaviruses as the starting material showed similar patterns of changes (Table 3).

The analysis of the antigenic regions of VP7 suggested that the breakthrough and control strains may be more closely related to each other than to reference strain Wa. To examine this in more detail, a distance matrix table was prepared from the amino acid sequences deduced both from the VP7 open reading frame and from VP4 amino acids 71-204 (hypervariable region) of these strains (Table 4). Overall divergence of 3.5% to 6.7% was observed between these recent clinical isolates and the reference VP7 sequence of strain Wa, while divergence of 0% to 6.7% was found between the clinical isolates themselves. For the hypervariable region of VP4, the divergences ranged from 2.27% to 8.71% compared to Wa, and 0% to 13.04% between each other. However, for all recently isolated U.S. strains examined except for Mo-15, Tn-16, and Tn-39, the VP7 proteins and VP4 hypervariable regions were more closely related to each other than to Wa (divergence range: 0% to 2.17% for VP7 proteins and 0% to 1.51% for VP4 hypervariable region peptides, respectively). Our results also suggested that for the most part, the VP7 and VP4 genes of breakthrough and control strains exhibited covariation as there was little evidence that any of them were formed by reassortment between strains of different VP4 or VP7 lineages (e.g., all Ohio isolates had identical VP4 peptides and closely related VP7 proteins, and the two Missouri strains 12 and 15, had distantly related VP7 and VP4 proteins). All serotype G1 strains whose VP4 variable regions were se- quenced belonged to genotype P [8] using the recently adopted classification system described by Estes [8a].

2066 Q. Jin et al.

Table 3. Amino acid substitutions compared to the Wa strain in VP7 protein antigenic regions and sit among breakthrough and control strains

Strain Region VR4 Region A Region B Region C Site A Site t 65-76 87 100 i42-150 208-224 170 307

Wa A/65 V/66 G/96 D/97 S/147 Q/208 M/217 V/170 V/30' T/68 E/74 1/218 N/221 V/75

breakthrough ~ Oh-25 ~ T/65 A/68 _b N/147 A/208 I I

6/74 1/75 Mo-12 ~ T/65 A/68 - N/147 D/221 t I

0/74 1/75 Va-12 T/65 A/68 E/97 N/147 V/218 I 1

G/74 1/75 Md-216 T/65 A/68 E/97 N/147 V/218 I I

0/74 1/75 Tn-16 T/65 A/68 E/97 N/147 V/218 I I

0/74 1/75 RI-14 T/65 A/68 E/97 N/147 V/218 I I

0/74 M/75 Oh-40 T/65 A/68 E/97 N/147 V/218 I I

0/74 M/75 RI-43 T/65 A/68 E/97 N/147 V/218 I I

0/74 M/75 Oh-ll T/65 A/68 E/97 N/147 V/218 I l

6/74 M/75 DC-6 T/65 A/66 E/97 N/147 V/218 I I

A/68 0/74 1/75

RI-39 T/65 A/68 E/97 N/147 V/218 I I G/74 1/75

Nb-49 T/65 A/68 E/97 N/147 - I I 0/74 1/75

Placebo Oh-64 T/64 A/68 -- N/147 V/218 I i

G/74 1/75 Tn-39" T/64 A/68 - N/147 V/218 I -

G/74 1/75 Mo-15 a T/64 A/68 D/96 E/97 N/147 T/217 V/218 I

G/74 1/75

a Fecal isolates; all other strains were culture adapted b Indicates identity with strain Wa in the region or site

Tab

le 4

. D

ista

nce

mat

rix

(am

ino

aci

d su

bst

itu

tio

ns/

i00

res

idue

s) b

etw

een

the

VP

7 p

rote

ins

and

VP

4 v

aria

ble

reg

ion

pep

tid

es (

amin

o a

cids

71

-20

4)

of

U,S

. b

reak

thro

ug

h a

nd

co

ntr

ol

stra

ins

Str

ain

VP

7 d

ista

nces

bet

wee

n t

he f

oll

ow

ing

str

ains

RI-

14

R

I-3

9

RI-

43

M

o-1

2

Mo

-15

bV

a-1

2

Oh

-ll

Oh

-25

O

h-4

0

Oh

-64

b N

eb-4

9

Md

-21

6

DC

-6

Tn

-16

T

n-3

9 b

W

a

RI-

14

0,

61

0 1.

86

5.4

0.61

0

1.55

0

0.93

1.

23

0,61

0,

92

0.92

2.

81

3,45

R

I-3

9

0 0.

61

1.86

5.

4 0,

61

0.61

1,

55

0.61

0.

93

1.23

0.

61

0.92

0.

92

2,81

3,

77

RI-

43

0 0

1.86

5.

4 0.

92

0 1,

55

0 0.

93

1.23

0.

61

0.92

0.

92

2.81

3.

45

Mo

-12

0.

75

0.75

0.

75

6.73

1.

86

1.86

0.

92

1.86

1,

55

1.86

1.

86

2.17

2.

17

3.45

3.

77

Mo

-15

7.

88

7,88

7.

88

8.71

5.

4 5.

4 6.

39

5.4

5,74

6.

06

5.4

5.73

5.

73

6.38

6.

73

Va-

12

0.75

0,

75

0.75

0

8.71

0.

61

1.55

0.

61

0,93

1.

23

0 0.

92

0.31

2.

81

3.77

O

h-t

l 1.

51

1,51

1,

5t

0.75

9.

56

0.75

1,

55

0 0,

62

1,23

0.

61

0.92

0.

92

2.81

3,

45

Oh

-25

1,

51

1.51

1.

51

0.75

9.

56

0.75

0

1.55

1.

25

1.55

1,

55

1.86

1.

86

3.13

3.

45

Oh

-40

1.

51

1.51

1.

51

0,75

9.

56

0.75

0

0 1.

25

1.23

0.

61

0.92

0.

92

2.81

3.

45

Oh

-64

_a

-

..

..

..

..

..

1.

55

0.93

1.

24

0.93

2.

50

3.46

N

eb-4

9

1.51

1.

51

1.51

0.

75

9.56

0.

75

1.51

1.

51

1.51

-

1.23

1.

55

1.55

3.

45

3.77

M

d-2

16

1.

51

1.51

1.

51

0.75

9.

56

0.75

1.

51

1.51

1.

51

- 1.

51

0.92

0.

31

2.81

3.

77

WaD

C-6

1.

51

1.51

1.

51

0.75

9.

56

0.75

1.

51

1.51

1.

51

1.51

1.

51

1.23

3.

13

4.09

T

n-1

6

4.62

4.

62

4.62

3.

83

13.0

4 3.

83

4.62

4.

62

4.62

-

4.62

4.

62

4.62

3.

13

4.09

T

n-3

9

..

..

..

..

..

..

..

..

..

.

4.09

W

a 2.

27

2.27

2.

27

1.51

8.

71

1.51

2.

27

2.27

2.

27

2.27

2.

27

2.27

5.

42

<

t~

o~

t~

e.+ .<

0~

t~

RI-

14

R

1-39

R

I-4

3

Mo

-12

M

o-1

5b

Va-

12

O

h-l

l O

h-2

5

Oh

-40

O

h-6

4b

Nb

-49

M

d-2

16

D

C-6

T

n-1

6

Tn

-39

b

Wa

VP

4 d

ista

nce

s b

etw

een

the

ab

ov

e st

rain

s

No

t se

quen

ced

u S

trai

ns i

sola

ted

fro

m p

lace

bo

rec

ipie

nts

C2~

O

', -,

.a

2068 Q. Jin et al.

Comparison of breakthrough and control strains from the U.S. to serotype G1 strains from other countries

Our results indicate that serotype G1 breakthrough and control strains from the U.S. are distinguishable from strain G 1 prototype Wa both serologically, and in VP7 amino acid sequence. Since the VP7 protein of Wa is nearly identical to that of the current vaccine strain D, we wanted to determine if other serotype G1 isolates also showed these sequence differences. To do this, we compared G1 isolates from 11 other countries, including 18 community diarrhea isolates from 8 nations (China, Zambia, Israel, Bangladesh, Egypt, Korea, Brazil, and Costa Rica) characterized in this study, and to 8 previously sequenced isolates from 4 countries (U.S., Japan, Australia, and Russia) [16, 28, 38]. Overall deduced amino acid divergences of 1.23% to 7.74% were observed between global strains and the Wa VP7 protein, while divergences of 0% to 8.77% were found between the isolates themselves (results not shown). Evolutionary analysis of the distance matrix data of 42 strains from 11 countries demonstrated that their VP7 proteins fell into at least three major, and one minor lineages (designated lineages I to IV) based on inter-strain deduced amino acid homologies and patterns of amino acid substitutions (Fig. 1). One strain (Egy-7) showed no close amino acid homology

lineage II

Oh-ll Oh-40 DO-6 RH4 . ^. RI-4 .~,us-u ]

- Tn-16 R~ -39Oh-I Md-216 ~.~/T I l l / )S-70

Va-12 "'-~ ~ ' / . ~ / M ~ . . ~ Aus-90 NOh- ~ 4 2 5 ~ ~ . . . / Rtl s*;lrl.39

Mo-15 ~

Bao-48 ~ ~ / / " / / d p n - 4 1 7 ~ / j / / / Jpn-418 / . . / / / /

Isr-55- / / / / I Zam-52" f / . . . . . . . /

Ch[-45 L.n,-4~ / 1

Kor-64'/// dpm421 Kor-54 1

lineage tV Egy-7

~ lsr-56 Egy-8

grz-3 lineage I I I

0.1

Fig. 1. Phylogram of VP7 proteins of global serotype G1 strains compared to U.S. break- through and control strains showing genetic distance between strains. The phylogenetic tree was generated from a VP7 amino acid multiple sequence file using the Protdist and Kitsch modules of the Phylip3.5 program to calculate a distance matrix and draw the tree [6, 9]. Two and three letter codes were used to distinguish the positions of U.S. strains from those of

isolates from other countries, respectively

Divergence of G1 rotavirus VP7 genes 2069

with any other strain, but was arbitrarily designated a lineage IV isolate because its closest clustering pattern was with this group.

The pattern of amino acid substitutions between strains was analyzed in more detail using an amino acid alignment with Wa as the reference VP7 sequence (Fig. 2). The VP7 proteins within each lineage possessed distinctive amino acid residues that were found in most members of the same lineage at most positions, but were not present in most other strains, e.g., lineage I (IIe-57, Ser-68, Thr-217 and Iie-281), lineage II (Ser-38, Ala-68, Glu-97), lineage III (Ala-65, Thr-68, Glu-75, Val-76 and Ser-147), and lineage IV (Val-16, Thr-43, Thr-101, Iie-193, Tyr-235, Asp-278, Iie-326). Most of the amino acid substitutions relative to Wa for global strains, as well as for U.S. isolates, occurred in and around previously identified variable regions A, B, C, VR3, VR4, and VR6 [8, 14].

In contrast to the majority of the isolates, only 7 strains were in lineage III with Wa and D. These isolates have no substitutions in regions A, B, and C, but most are distinct from Wa and D at sites VR3, VR6, and at amino acid 291. A neutralization site has recently been defined at amino acid 291 [5, 7].

Discussion

Rotavirus vaccines have not been completely effective in protecting vaccinated infants against rotavirus diarrhea, even when challenge is due to a serotype included in the vaccine, and when the infants have mounted a neutralizing antibody response to the challenge strain. In this report, we examined the hypothesis that antigenic variation in the rotavirus VP7 protein of circulating serotype G1 strains could be one factor contributing to lack of protection (i.e., breakthrough) against these strains in infants who had post-vaccination neu- tralizing antibodies to the prototype G1 Wa strain following vaccination with RRV-S1 or RRV-TV. To do this, we compared the neutralizing antibody responses of these subjects against the breakthrough strains that caused their illnesses to those of prototype Wa. In addition, we sequenced the VP7 genes of these strains to see if changes in deduced VP7 amino acid sequences could be associated with breakthrough.

Consistent with our hypothesis, we found serum neutralizing antibody titers stimulated by these vaccines were overall significantly greater to Wa than to the G1 rotavirus strains subsequently isolated from vaccinated infants with diar- rheal disease, suggesting that these strains differed antigenically from Wa, and by inference, from serotype G1 vaccine strain D. However, lower responses to circulating G1 rotavirus srains relative to Wa were also found in subjects considered to be vaccine successes, i.e. 16 vaccinees who either became asym- ptomatically infected or experienced no natural rotavirus infection following vaccination (see Table 2). Thus, we could not correlate neutralizing titers to the clinical isolates with protection against diarrhea in infants who became infected after vaccination. Furthermore, we have not excluded the possibility that some of the titer differences could be due to variation in replication potential of the clinical isolates compared to the test G1 strain, Wa [29].

2070 Q. Jin et al.

O ~ ~ J

o

! ~ z

ta.a z

C

>-

_ J

; ~ . . J

V- b -

N

~ . . . . . . . . . . . . . . . . . ~ ~

Wa

Jpn

-41

7

Jpn

-41

8

Ch

i-4

5

N

Ch

i-4

6

N

Zam

-52

N

Isr-

55

I

N

Ban

-48

N

Ban

-59

N

Mo-

15

N

Mo-

12

N

0h-2

5 N

M

D-2

16

N

Va-

12

N

Tn-

16

N

0h

-li

N

Oh-

40

N

RI-

14

N

R

I-4

3

N

Aus

-91

N

RI-

39

N

0h

-64

F N

N

b-49

N

D

C-6

N

T

n-39

S

A

N

C

os-6

9 S

N

C

os-7

0 S

N

A

us-9

0 S

N

R

us-9

1 S

E

N

D

S

E

gy-8

S

Is

r-5

6

S

Brz

-2

S

I B

rz-4

S

S

I

Brz

-5

S

S

I B

rz-6

S

[

Brz

3

S

I Jp

n-42

1 N

K

or-

64

N

K

or-

54

N

E

gy-7

S

N

110

VR

6 B

(V

RT)

G

WPT

GSV

YFK

EY

SN

IVD

FSV

DP

QLY

CD

YN

L V

LMK

YD

QS

LE L

DM

SE

LAD

LI L

NEW

LCN

PMD

N

C

N

T

C (V

RS

) 21

8 V

TLY

YY

QQ

SG

ESN

KWIS

MG

S S

CTV

KV

CP

LN T

QTL

GIG

CQ

T TN

VD

SFE

MI

I I I I i I E

I I I I I I f I I I I I I I I I I I I [ I I

D

A

L TV

TV

TV

TV

TV

TV

TV

TV

A

V

V

V

V

V

V

V

V

V

V

V

V

V

V

V

V

V

V

A

Y

I I I TV

LIN

EA

GE

I I I I 1 II

II

II

II

II

II

II

II

II

II

II

II

II

II

II

II

II

II

II

III

III

III

III

III

III

iII

III

IV

IV

IV

IV

< <

qQ

bo

O

-.--.1

Fig

. 2

(con

tinue

d)

219

VR9

326

Wa

AE

NE

, gLA

IVD

VV

DG

INH

,K,IN

LTT

TTC

TIR

N

CKK

LGPR

ENV

AVIQ

VGG

SNV

LDIT

ADPT

TN

PQTE

RM

MR

VN W

KKW

WQ

VFYT

IV

DY

INQ

IVQ

VM

SKR

SRSL

N S

AAFY

YRV

LIN

EAG

E

dpn-

417

L I

I I

R

I dp

n-41

8 L

I I

I R

I

Chi

-45

I I

R

i C

hi-4

6 I

I R

I

Zam

-52

Y I

I A

I R

I

[sr-

65

N

I

i R

I

Ban

-48

T I

I R

I

Ban

-59

[ I

I R

I

Mo-

15

T R

I

N

R

V D

[

Mo-

12

D

S I

LI

0h-2

5 I

II

MD

~216

A

I II

V

a-12

A

I [I

In

-16

A

I II

0

h-l

l I

[[

Oh-

40

I II

R

I-14

I

II

RI-

43

I II

A

us-9

1 F

II

RI-

39

V i

II

0h

-64

I

II

Nb-4

9 I

II

DC

-6

I II

Cos

-69

II

Cos

-70

II

Tn-

39

II

Aus

-90

F tI

R

us-9

1 II

D

Il

l E

gy-8

II

I Is

r-5

6

Ill

Brz

-2

M

Ill

Brz

-4

P R

R

II

I B

rz-5

R

Il

l B

rz-6

R

Il

l B

rz-3

D

G

R

Il

l Jp

n-42

1 Y

I D

L

R

I IV

K

or-6

4 Y

I D

R

I

IV

Kor

-54

Y E

I D

R

I

IV

Egy

-7

Q

I [

R

IV

Fig

. 2.

Com

pari

son

of d

educ

ed a

min

o ac

id s

eque

nces

of

the

enti

re V

P7

gene

of

36 h

uman

ser

otyp

e G

1 co

mm

unit

y st

rain

s is

olat

ed f

rom

diff

eren

t ge

ogra

phic

loca

tion

s ar

ound

the

wor

ld. T

he p

revi

ousl

y re

port

ed s

eque

nces

of s

trai

ns J

pn-4

17,

-418

and

-42

1 (a

cces

sion

num

bers

Dt6

326,

D16

327,

an

d D

1632

8),

Aus

-90

(acc

essi

on n

umbe

r M

9300

6),

Aus

-91,

and

Rus

-91

(acc

essi

on n

umbe

r $8

3903

) w

ere

obta

ined

fro

m G

enB

ank

or th

e pu

blic

atio

ns

them

selv

es [

16,

28,

38].

Seq

uenc

e al

ignm

ents

wer

e pr

epar

ed b

y th

e G

CG

pro

gram

PIL

EU

P u

sing

a p

rogr

essi

ve a

lignm

ent

met

hod

[6].

The

VP

7 pr

otei

n se

quen

ce o

f st

rain

Wa

is s

how

n on

the

top

, an

d th

e am

ino

acid

s th

at d

iffer

fro

m W

a ar

e sh

own

for

the

rem

aini

ng s

trai

ns

b-~

b,~

Divergence of G1 rotavirus VP7 genes 2073

Our serologic results are consistent with the vaccine trial as a whole in that the RRV-TV and RRV-S1 vaccines both provided good protection overall against rotavirus disease [30], but neither showed a significant correlation between seroresponses in neutralizing antibody to the G1 strain Wa and protection against any diarrhea or severe disease associated with G1 rotaviruses (Mack and Ward, unpubl, data). A similar lack of association between protection against rotavirus disease and serotype-specific serum neutralizing antibody titers was previously reported in children following natural rotavirus infection [35]. These results suggest that better correlates of immunity to rotavirus disease remain to be identified. Possibly, intestinal immune responses which were not examined in this study would correlate better with protection against rotavirus disease as well as severity of rotavirus illnesses.

Sequence analysis of breakthrough strains confirmed our serological results and demonstrated that circulating G1 rotaviruses are antigenically distinct from Wa and from vaccine strain D. Although neutralizing antibody titers in post- vaccination sera against the G1 rotavirus strains that caused subsequent illnesses were generally lower than against Wa, in some cases these titers were nearly equal. Variations in the ratios of the titers against Wa compared to those against strains causing illnesses appeared to be primarily associated with differences in the relative antibody responses from one subject to another, rather than to major antigenic differences in the G1 rotaviruses circulating during this study. This suggestion was supported by the finding that no apparent VP7 sequence changes could be identified that were responsible for variations in the relative neutraliz- ation titers to the Wa and breakthrough strains. Furthermore, most substitu- tions in control and breakthrough strains relative to Wa were the same, indicating that differences in sequences of VP7 proteins were not restricted to the breakthrough strains.

Although these results do not support our original hypothesis, they demon- strate that the VP7 protein of vaccine strain D is antigenically distinct from the U.S. strains and most circulating G1 isolates worldwide examined in this study, especially in antigenically important regions A, B, and C where recognized neutralization epitopes involved in serotype determination and protection are found. Some of the substitutions are in amino acids that are critical components of neutralization epitopes identified in previous studies (e.g., the S to N change at aa 147, or the D to E at aa 97) [5, 331. In addition, the presence of human serum antibodies to epitopes within sites A and C correlate with protection from rotavirus diarrhea in adult volunteers challenged with virulent rotavirus, or in naturally infected children [15, 27]. Our sequencing results suggest that these epitopes may be absent on the VP7 proteins of circulating G1 strains in the U.S. which could theoretically lead to reduced protection. Thus, it cannot be excluded that the observed antigenic changes play some role in vaccine failure.

In contrast to U.S. strains, several vaccine placebo strains isolated during a vaccine trial of RRV-TV in Brazil and characterized in this study were much more closely related to Wa overall, and in antigenic regions A to C. Thus, it may be interesting to serologically and molecularly characterize G1 breakthrough

2074 Q. Jin et al.

strains isolated in the same Brazilian trial, to determine if subjects who had serum neutral izing antibodies to Wa can subsequent ly experience illness caused by strains closely related to Wa. The results of such a s tudy m a y help clarify whether amino acid substi tut ions in antigenic regions of VP7 are impor tan t de te rminants of b reak through .

Acknowledgements

We thank Anne Mather, Brian Holloway, Robert Holman, and Steve Monroe for help with this investigation. We also thank the members of the U.S. Rotavirus Vaccine Efficacy Group who performed the vaccine trial from which clinical specimens used for this study were collected.

References

1. Bessarab IN, Epifanova NV, Novikova NA, Borodin AM (1991) An analysis of the gene coding for the basic neutralizing antigen VP7 of human rotavirus isolate 1407. Vopr Virusol 36:480 483

2. Chiba S, Nakata S, Urasawa T, Urasawa S, Yokoyama T, Morita Y, Taniguchi K,Nakao T (1986) Protective effect of naturally acquired homotypic and heterotypic rotavirus antibodies. Lancet 2:417-421

3. Clark HF, Borian FE, Plotkin SA (1990) Immune protection of infants against rotavirus gastroenteritis by a serotype 1 reassortant of bovine rotavirus WC-3. J Infect Dis 161: 1 099-1 104

4. Clark HF, Welsko D, Offit PA (1992) Infant responses to rotavirus WC3 reassortants containing human rotavirus VP7, VP4 or VP7 + VP4. Antimicrob Agents Chemother, Abstract 1394

5. Coulson B, Kirkwood C (1991) Relation of VP7 amino acid sequence to monoclonal antibody neutralization of rotavirus and rotavirus monotype. J Virol 65:5 968-5 974

6. Devereux J, Haeberli P, Smithies O (t984) A comprehensive set of sequence analysis programs for the VAX. Nucleic Acids Res 12:387 395

7. Dunn SJ, Ward RL, McNeal MM, Cross TL, Greenberg HB (1993) Identification of a new neutralization epitope on VP7 of human serotype 2 rotavirus and evidence for electropherotype differences caused by single nucleotide substitutions. Virology 197: 397-404

8. Dyall-Smith ML, Lazdins I, Tregear GW, Holmes IH (1986) Location of the major antigenic sites involved in rotavirus serotype-specific neutralization. Proc Natl Acad Sci USA 83:3 465-3 468

8a. Estes MK (1996) Rotaviruses and their replication. In: Fields BN, Knipe DM, Howley PM, Chanock RM, Melnick JL, Morath TP, Roeyman B, Strauss SE (eds) Fields virology, 3rd ed, vol 2. Lippincott-Raven, Philadelphia, pp 16251655

9. Felsenstein J (1989) PHYLIP-phylogeny inference package (version 3.2). Cladistics 5: 164 166

10. Flores J, Taniguchi K, Green K, Perez-Schael I, Garcia D, Sears J, Urasawa S, Kapikian AZ (1988) Relative frequencies of rotavirus serotypes 1, 2, 3, and 4 in Venezuelan infants with gastroenteritis. J Clin Microbiol 26:2 092-2 095

11. Gentsch JR, Glass RI, Woods P, Gouvea V, Gorziglia M, Flores J, Das BK, Bhan MK (1992) Identification of group A rotavirus gene 4 types by potymerase chain reaction. J Clin MicrobioI 30:1 365-1 373

Divergence of G1 rotavirus VP7 genes 2075

12. Glass RI, Keith J, Nakagomi O, Nakagomi T, Askaa J, Kapikian AZ, Chanock RM, Flores J (1985) Nucleotide sequence of the structural glycoprotein VP7 gene of Nebraska calf diarrhea virus rotavirus: comparison with homologous genes from four strains of human and animal rotaviruses. Virology 141:292-298

13. Gouvea V, Glass RI, Woods P, Taniguichi K, Clark HF, Forrester B, Fang ZY (11990) Polymerase chain reaction amplification and typing of rotavirus nucleic acids from stool specimens. J Clin Microbiol 28:276-282

14. Green KY, Hoshino Y, Ikegami N (1989) Sequence analysis of the gene encoding the serotype-specific glycoprotein (VP7) of two new human rotavirus serotypes. Virology 168:429-433

15. Green KY, Kapikian AZ (1992) Identification of VP7 epitopes associated with protec- tion against human rotavirus illness or shedding in volunteers. J Virol 66:548-553

16. Green K, Midthun K, Gorziglia M, Hoshino Y, Kapikian A, Chanock R, Flores J (1987) Comparison of the amino acid sequences of the major neutralization protein of four human rotavirus serotypes. Virology 161:153 159

17. Gunasena S, Nakagomi O, Isegawa Y, Kaga E, Nakagomi T, Steele AD, Flores J, Ueda S (1993) Relative frequency of VP4 gene alleles among human rotaviruses recovered over a 10 year period (1982-1991) from Japanese children with diarrhea. J Clin Microbio131: 2 195-2 197

18. Institute of Medicine (1986) The prospects of immunizing against rotavirus. In: New vaccine development: diseases of importance in developing countries, vol 2. National Academy Press, Washington, pp D13-1-D13-12

19. Kapikian AZ (1994) Jennerian and modified Jennerian approach to vaccination against rotavirus diarrhea in infants and young children: An introduction. In: Kapikian AZ (ed) Viral infections of the gastrointestinal tract. Marcel Dekker, New York, pp 409- 417

20. Knowlton DR, Spector DM, Ward RL (1991) Development of an improved method for measuring neutralizing antibody to rotavirus. J Virol Methods 33:127-134

21. Linhares AC, Gabbay YB, Freitas RB, da Rosa ES, Mascarenhas JD, Loureiro EC (1989) Longitudinal study of rotavirus infections among children from Belem, Brazil. Epi- demiol Infect 102:129-145

22. Linhares AC, Gabbay YB, Mascarenhas JDP, Freitas RB, Oliveira CS, Bellesi N, Monteiro TF, Lins-Lainson Z, Ramos FLP, Valente SA (1994) Estudo prospectivo das infeccdes por rotavirus em Belem, Para, Brazil: ume abordagem clinico- epidemiologica. J Pediatr 70:220-225

23. Midthum K, Greenberg HB, Hoshino Y (1985) Reassortant rotaviruses as potential live rotavirus vaccine candidates. J Virol 53:949-954

24. Midthun K, Hoshino Y, Kapikian AZ, Chanock RM (1986) Single gene substitution rotavirus reassortants containing the major neutralization protein (VPT) of human rotavirus serotype 4. J Clin Microbiol 24:822 826

25. Offit PA, Clark HF, Blavat G, Greenberg HB (1986) Reassortant rotaviruses containing structural proteins VP3 and VP7 from different parents induce antibodies protective against each parental serotype. J Virol 60:491-496

26. O'Ryan M, Matson D, Estes M, Bartlett A, Pickering L (1990) Molecular epidemiology of rotavirus in children attending day care centers in Houston. J Infect Dis 162:810-816

27. O'Ryan ML, Matson DO, Estes MK, Picketing LK (1994) Anti-rotavirus G type- specific antibodies in children with natural rotavirus infections. J Infect Dis 169:504-511

28. Palombo Ea, Bishop RF, Cotton RGH (1993) Intra- and inter-season genetic variability in the VP7 gene of serotype 1 (monotype ta) rotavirus clinical isolates. Arch Virol 130: 57 69

2076 Q. Jin et al.: Divergence of G1 rotavirus VP7 genes

29. Parkar RA, Pallansch MA (1992) Using the virus challenge dose in the analysis of virus neutralization assays. Stat Med 11:1 253-1 262

30. Rennels MB, Glass RI, Dennehy PH, Bernstein DI, Pichichero ME, Zito ET (1996) Safety and efficacy of high-dose rhesus-human reassortant rotavirus vaccines: report of the national muIticenter trial. Pediatrics 97:7-13

31. Sanger F, Nicklen S, Coulson AR (1977) DNA sequencing with chain-terminating inhibitors. Proc Natl Acad Sci USA 74:5 463-5 467

32. Steele AD, Garcia D, Sears J, Gerna G, Nakagomi O, Flores J (1993) Distribution of VP4 gene alleles in human rotaviruses by using probes to the hyperdivergent region of the VP4 gene. J Clin Microbiol 31:1 735-t 740

33. Taniguchi K, Hoshino Y, Nishikawa K, Green KY, Maloy WL, Morita Y, Urasawa S, Kapikian AZ, Chanock RM, Gorziglia M (1988) Cross-reactive and serotype-specific neutralization epitopes on VP7 of human rotavirus: nucleotide sequence analysis of antigenic mutants selected with monoclonal antibodies. J Virol 62:1 870-1 874

34. Urasawa S, Urasawa T, Taniguchi K, Wakasugi F, Kobayashi N, Chiba S, Sakurada N, Morita M, Morita O, Tokieda M, Kawamoto H, Minekawa Y, Obseto M (1989) Survey of human rotavirus serotypes in different locales in Japan by enzyme-linked immunosor- bent assay with monoclonal antibodies. J Infect Dis 160:44-51

35. Ward RL, Clemens JD, Knowlton DR, Rao MR, van Loon FP, Huda N, Ahmed F, Schiff GM, Sack DA (1992) Evidence that protection against rotavirus diarrhea after natural infection is not dependent on serotype-specific neutralizing antibody. J Infect Dis 166:1 251-1 257

36. Ward RL, Knowlton DR, Pierce MJ (1984) Efficiency of human rotavirus propagation in cell culture. J Clin Microbiol 19:748-753

37. Woods PA, Gentsch J, Gouvea V, Mata L, Simhon A, Santosham M, Bai Z-S, Urasawa S, Glass RI (1992) Distribution of serotypes of human rotavirus in different populations. J Clin Microbiol 30:781-785

38. Xin K-Q, Morikawa S, Fang Z-Y, Mukoyama A, Okuda K, Ushijima H (1993) Genetic variation in VP7 gene of human rotavirus serotype 1 (G1 type) isolated in Japan and China. Virology 197:813-816

Authors' address: Dr. J. R. Gentsch, Viral Gastroenteritis Section, MS-G04, Center for Disease Control and Prevention, 1600 Clifton Road N. E., Atlanta, Georgia 30333, U.S.A.

Received May 24, 1996