Production of Reassortant Viruses Containing Human Rotavirus VP4 and SA11 VP7 for Measuring Neutralizing Antibody following Natural Infection

CLINICAL AND DIAGNOSTIC LABORATORY IMMUNOLOGY, Sept. 1997, p. 509 514 Vol. 4, No. 5 1071-412X/97/$04.00 0 Copyright 1997, American Society for Microbiology Production of Reassortant Viruses Containing Human Rotavirus VP4 and SA11 VP7 for Measuring Neutralizing Antibody following Natural Infection R. J. GORRELL* AND R. F. BISHOP Department of Gastroenterology and Clinical Nutrition, Royal Children s Hospital, Parkville, Victoria, Australia Received 18 November 1996/Returned for modification 19 March 1997/Accepted 23 May 1997 The outer capsid proteins VP4 and VP7 of group A rotaviruses are both targets of neutralizing antibody produced following natural infection in humans. Of interest is the relative importance and immunodominance of each protein in the generation of a protective immune response. In order to measure neutralizing antibody responses to VP4 and VP7 separately, reassortants bearing VP4 of each of the major human rotavirus P types with VP7 of SA11 origin were successfully produced by neutralizing monoclonal antibody selection. The resulting reassortants, together with reassortants representing each of the major VP7 types, were antigenically characterized with serotype-specific neutralizing monoclonal antibodies and hyperimmune sera. The neutralization proteins of human rotavirus origin were found to be unaffected antigenically by reassortment. The abilities of these reassortants to discriminate between VP4 and VP7 immune responses were evaluated with postinfection sera collected from three patients infected with either a P1A[8],G1, a P1B[4],G2, or a P1A[8], G4 rotavirus strain. The reassortants were shown to be capable of separating the neutralizing antibody responses to VP4 and VP7, with each patient showing a different immune response with respect to VP4 or VP7 immunodominance. These reassortants can now be applied to analyses of individual immune responses to VP4 and VP7 proteins after primary rotavirus infections and reinfections in humans. Rotaviruses (RVs), constituting a genus of the family Reoviridae, are of medical and veterinary importance worldwide, being major etiological agents of acute gastroenteritis in human infants and young animals. The complete virion of group A RV is made up of six structural proteins assembled in three protein layers surrounding a genome of 11 double-stranded RNA (dsrna) segments, each encoding a separate protein. The outer capsid layer consists of the hemagglutinin VP4, encoded by gene segment 4, and VP7, a glycosylated protein encoded in human strains by either gene segment 8 or gene segment 9. Numerous serotypes of the VP4 (P serotype) and VP7 (G serotype) proteins have been determined, resulting in a binary serotyping system (10). Three major P serotypes (serotypes P1A, P1B, and P2) and four major G serotypes (serotypes G1, G2, G3, and G4) are prevalent worldwide in human clinical isolates. VP4 types are also described in terms of gene sequence relatedness, with genotypes [8], [4], and [6] representing serotypes 1A, 1B, and 2, respectively (10). Both VP4 and VP7 are capable of eliciting neutralizing antibody (N-Ab) production during natural infection (13). Little is known, however, about the relative importance of each of these proteins in generating serotype-specific (homotypic) or cross-reactive (heterotypic) immune responses following natural infection in human infants. Currently used assays for detecting the production of serum N-Ab involve measuring the extent to which serum is capable of blocking infection by or neutralizing RV particles. Standard neutralization assays do not discriminate between neutralization mediated via VP4 and VP7 and thus cannot be used to determine the respective roles of VP4 and VP7 in both homotypic and heterotypic responses. * Corresponding author. Mailing address: Department of Gastroenterology and Clinical Nutrition, Royal Children s Hospital, Flemington Rd., Parkville, Victoria, 3052, Australia. Phone: 61 3 9345 5060. Fax: 61 3 9345 6240. E-mail: gastro@cryptic.rch.unimelb.edu.au. Reassortant viruses, or viruses produced by coinfection of two RV strains resulting in mixing of the segmented genomes, can be used to separate the human RV (HRV) VP4 and VP7 proteins. Neutralization of these reassortant viruses can then be used to determine the relative importance of each VP4 and VP7 type in the immune response following natural infection in humans. Production of appropriate reassortants could also be used to examine immune responses to infection with animal RV strains and animal human RV reassortants currently being used as candidate human vaccines. Previous studies examining the relative immunodominance of VP4 and VP7 in human hosts have provided conflicting results. Data from a study of children naturally infected with HRV in which reassortants made from the infecting strain were used suggested that VP4 may be the immunodominant neutralization antigen in the production of a homotypic response (25). Infants given a bovine RV HRV VP7 reassortant vaccine, W179-9, appeared to respond predominantly to the bovine parent, also suggesting that VP4 is important in generating a substantial N-Ab response (3). However, examination of the immune response in infants vaccinated with the bovine RV WC3 vaccine suggested that VP7 is immunodominant (24). A further study in calves infected with the G10 bovine RV B223 strain has suggested that VP7 is the immunodominant neutralizing antigen in the primary response, with VP4 being important in the elicitation of a heterotypic response (27). Results from other animal models suggest that immunodominance may be dependent on the infecting strain (12, 26). The homotypic and heterotypic components of the N-Ab responses against both VP4 and VP7 following natural infection in humans could be examined with a battery of reassortants independently bearing VP4 proteins of each of the three major human P types or VP7 proteins of the four major human G types. These human VP4 and VP7 types could be paired with 509

510 GORRELL AND BISHOP CLIN. DIAGN. LAB. IMMUNOL. VP7 and VP4 proteins, respectively, that are relatively immunologically neutral to human sera, i.e., of animal RV origin. Although selection of reassortants containing a VP4 of animal origin and a VP7 of HRV origin has been achieved successfully (15, 16), reassortants of the reverse combination have proved more elusive. We have derived animal RV HRV reassortants bearing VP4 proteins of each of the three major P types found in human strains. SA11 was used as the source of animal RV genes and was coinfected with the HRV strains Wa (P-type 1A[8]), RV5 (P-type 1B[4]), and ST3 (P-type 2[6]). The desired reassortants were derived with neutralizing monoclonal antibodies (N-MAbs) for selection. In addition, a reassortant incorporating a Wa VP7 and an SA11 VP4 was also selected. This report describes the techniques used in the derivation of these reassortants and confirmation of the antigenic integrity of the neutralization proteins, together with a preliminary study with patient sera to determine the capacities of these reassortants to differentiate the anti-vp4 and -VP7 N-Abs produced in response to natural infection. MATERIALS AND METHODS Viruses. Monolayers of the monkey kidney epithelial cell line MA104, maintained in Dulbecco s modified Eagle s medium (Trace Biosciences, Castle Hill, Australia) supplemented with 10% fetal bovine serum (Trace Biosciences), were used for virus cultivation. Standard and reassortant RVs were grown, following activation with porcine trypsin (10 g/ml; Sigma, St. Louis, Mo.) for 30 min at 37 C, in MA104 cells with maintenance medium containing 1 g of trypsin per ml (DMM-T). Viruses were harvested by freezing-thawing, and the resulting lysates were used for reassortant production and assays. The RV strains used in our laboratory for reassortant production included Wa (P1A[8],G1), RV5(P1B [4],G2), and ST3 (P2[6],G4), representing each of the three major human P types, and the simian RV strain SA11 (P[2],G3). All strains were initially plaque purified three times and were examined for reactivity with the N-MAbs to be used for selection of the reassortants. Reassortants containing the neutralizing antigen combinations RV5 VP7-SA11 VP4 and ST3 VP7-SA11 VP4 were obtained from I. H. Holmes (Department of Microbiology, University of Melbourne). Reassortant virus nomenclature. The nomenclature used for all HRV SA11 reassortant strains in this study was in the form of rhrv-4 or rhrv-7 denoting the human rotavirus parent (Wa, RV5, or ST3) and the neutralization antigen of HRV origin (VP4 or VP7). For example, rwa-4 and rwa-7 bear the VP4 and VP7 proteins of the HRV strain Wa, respectively. Patient samples. Serum samples from patients following primary P1A[8],G1 and P1A[8],G4 RV infections had been collected as part of a previous study in our department (11) or after admission to hospital with a P1B[4],G2 RV strain. Patients were considered to be experiencing a primary RV infection on the basis of an initial absence of immunoglobulin G class RV antibodies in acute-phase sera and demonstration of an immunoglobulin M class RV serum antibody response by enzyme immunoassay (11). MAbs. Ascitic fluids containing the N-MAbs listed in Table 1 were used for selection and characterization of reassortants. N-MAb 2G4 was kindly donated by H. B. Greenberg (Stanford University, Stanford, Calif.). Plaque purification. MA104 monolayers established in six-well plates were inoculated in duplicate with 400 l of RV lysate as 10-fold serial dilutions, following trypsin activation, in DMM-T. Following adsorption at 37 C for 1 h, 3 ml of overlay medium (DMM-T plus 0.75% purified agar) containing DEAEdextran (50 g/ml) was added to each well. A second overlay containing neutral red (80 g/ml) was added after 4 to 5 days to visualize plaques, which were picked at 5 to 7 days and resuspended in 500 l of DMM-T, and 250 l ofthe suspension was passaged in six-well plates following trypsin activation. PAGE separation of dsrna. Polyacrylamide gel electrophoresis (PAGE) was used to determine the RNA composition of virus progeny resulting from each infection. Viral RNA was prepared by phenol-chloroform extraction of infectedcell lysates (9). RNA was separated over a 0.75-mm-thick 12% polyacrylamide gel at 15 ma (20 h, 4 C) with the buffer system of Laemmli (17) and visualized with silver stain (9). Various running conditions and coelectrophoresis were also used in order to distinguish between closely migrating genes for confirmation of the origins of some gene segments in the reassortants. N-Ab assays. Two methods were used to confirm the integrity of antigenic sites on the reassortants. For antigenic characterization of the reassortants with N- MAbs, a fluorescent focus reduction neutralization (FFN) assay was used (5). Further characterization of the reassortants with rabbit or guinea pig anti-rv hyperimmune sera and validation with patient sera was performed by an enzyme immunoassay (EIA)-based method influenced by that of Knowlton et al. (14). All neutralization assays were set up as described previously for the FFN assay (5), TABLE 1. MAbs used for selection and characterization of reassortants MAb Protein specificity Serotype specificity Reference RV4:1 VP7 G1 8 RV4:2 VP7 G1 8 RV4:3 a VP7 G1 8 RV4:4 VP7 G1 6 RV4:5 VP7 G1 6 RV5:1 VP7 G2 8 RV5:3 a VP7 G2 8 RV5:4 VP7 G2 8 RV5:5 VP7 G2 18 RV3:1 VP7 G3 5 RV3:2 VP7 G3 5 RV3:4 VP7 G3 18 RV3:5 a VP7 G3 18 ST3:1 VP7 G4 8 ST3:2 VP7 G4 7 ST3:4 a VP7 G4 7 2G4 a VP4 P[2] 21 F45:3 VP4 P1A[8] 4 F45:4 a VP4 P1A[8] 4 RV5:2 VP4 P1B[4] 8 RV3:3 VP4 P2[6] 5 a N-MAbs used for reassortant selection. and the viral antigen either was detected by indirect immunofluorescence (FFN assay) or was released for EIA detection by freezing-thawing the plates three times. For detection by EIA, microtiter trays were coated with 50 l of preimmune rabbit serum or rabbit anti-sa11 hyperimmune serum diluted in phosphatebuffered saline (PBS) and incubated in a 37 C water bath for 1.5 h. The plates were then washed four times with PBS plus 0.05% (vol/vol) Tween 20 (PBS-T), 45 l of virus-infected cell lysate from each well was added to the microtiter tray wells coated with either anti-rv hyperimmune or preimmune serum, and the trays were left overnight at 4 C. Following washing, 50 l of anti-group A RV MAb, RVA, diluted in PBS-T plus 0.5% (wt/vol) casein (Sigma) was added to each well and the plates were incubated for 2.5 h in a 37 C water bath. After washing, horseradish peroxidase-conjugated sheep anti-mouse immunoglobulin-ab (50 l; Silenus, Hawthorn, Australia) diluted 1/10,000 in PBS-T plus 0.5% casein was added to each well and the plates incubated for 1.5 h in 37 C water bath. Viral antigen was then detected with 50 l of tetramethylbenzidine substrate per well, and the reaction was stopped after 10 min with 25 l of2m H 2 SO 4 per well. The absorbance of each well was determined at 450 nm (optical density at 450 nm), with the difference between the control well (containing preimmune serum) and the anti-rv antibody-coated well being the specific absorbance due to RV antigen. The N-Ab titer was defined as the reciprocal of the antibody dilution giving a 50% reduction in specific absorbance. Production of reassortants containing human rotavirus VP4 and SA11 VP7. MA104 monolayers in six-well plates were coinfected with SA11 paired with each of the HRV strains Wa, RV5, and ST3. A multiplicity of infection of 1 fluorescing cell-forming unit (FCFU) per cell was used for the HRVs (Wa, RV5, and ST3) and 0.1 FCFU/cell was used for SA11. In order to restrict the growth of progeny containing SA11 VP4 after the first cycle of replication, ascitic fluid containing the N-MAb 2G4 (anti-sa11 VP4) was added to the maintenance medium at a concentration 100-fold greater than the titer giving 50% neutralization of SA11. Following incubation at 37 C for 72 h, infected cell lysates were harvested by freezing-thawing, and viral RNA was extracted and examined by PAGE with a 12% polyacrylamide gel. The method of further selection was tailored to each mixture and was guided by the proportion of HRV genes to SA11 genes detected in each coinfection population following PAGE of dsrna with a 12% polyacrylamide gel. For the second passage, two different approaches were taken. Progeny from RV5 SA11 and ST3 SA11 were passaged with no selective pressure, resulting in predominantly HRV genes and faint evidence of SA11 genes. For progeny of Wa SA11, 2G4 was added to the medium after 36 h of selection-free incubation to maintain the dominant presence of Wa genes. After the second passage, the N-MAbs RV4:3, RV5:3, and ST3:1 were used to select against progeny containing VP7 from Wa, RV5, and ST3, respectively. Lysates were passaged in the presence of both the corresponding anti-vp7 MAb and 2G4 until the desired

VOL. 4, 1997 HUMAN ROTAVIRUS VP4 AND SA11 VP7 REASSORTANTS 511 with the lysate resulting from the initial coinfection set up to produce rwa-4 (as described above). Selection-free passaging of the initial coinfection lysate resulted in a mixed population showing equal proportions of all Wa and SA11 genes by PAGE with a 12% polyacrylamide gel of RNA extracted from the lysate. Neutralization, both prior to and during the next passage, with ascitic fluids containing the N-MAbs RV3:5 (anti-g3 VP7) and F45:4 (anti-p1a[8] VP4) resulted in the dominance of Wa gene 9 and SA11 gene 4 in the resulting progeny. A suitable reassortant clone was obtained after two rounds of plaque purification of the population under the selective pressure of RV3:5 alone: the first plaquing with neutralization both prior to and during infection, and the second plaquing with neutralization during infection only. The selected reassortant, containing gene segment 4 of SA11 and all other genes from Wa, was subjected to three further rounds of selection-free plaque purification. Virus stock was then grown in 80-cm 2 flasks, and the infected cell lysate was used for subsequent assays. FIG. 1. Migration patterns of reassortant and parent rotavirus strain gene segments. The photograph was scanned into Photoshop (Adobe) and labeled in Quark Xpress (Adobe). gene profile was achieved, i.e., no or very little detectable Wa or ST3 gene 9 or RV5 gene 8 and no detectable SA11 gene 4. This goal was achieved by neutralization both prior to and during infection or during infection only, after two passages for RV5 SA11, while Wa SA11 and ST3 SA11 required three passages. Plaque purification was then performed with both 2G4 and the appropriate anti-vp7 N-MAb in the overlay. Plaques were passaged and the RNA profile was examined by PAGE with a 12% polyacrylamide gel. For the RV5 SA11 reassortant, a suitable clone was identified and plaque purified three times with no selective pressure. For ST3 SA11, a further round of plaque purification with selection was required before the desired reassortant was obtained, and the reassortant was plaque purified three times. Wa SA11 also required two rounds of plaque purification in the presence of both N-MAbs; however, all plaques obtained following the second plaquing with RV4:3 and 2G4 still contained Wa gene 9. In a subsequent plaque purification rabbit anti-rv4 hyperimmune serum instead of RV4:3 was successfully used in the overlay to neutralize reassortants containing Wa VP7. This was followed by two further rounds of selection-free plaque purification. Stocks of all reassortants were grown in 80-cm 2 flasks, and the infected cell lysates were used for subsequent assays. Production of reassortant containing Wa VP7 and SA11 VP4. Production of a reassortant of Wa and SA11 containing Wa VP7 and SA11 VP4 was achieved Virus RESULTS Production of rhrv-4 reassortants. The RNA migration patterns of the reassortant and parent RV strains are presented in Fig. 1. The gene segment origin for each of the reassortants was determined by examination of the gene profile following PAGE of the genomic dsrna. The gene composition of each of the reassortants is listed in Table 2. Antigenic characterization. Results of neutralization assays with VP4 or VP7 serotype-specific N-MAbs reacted with parent and reassortant viruses confirmed that all reassortants produced carried the VP4 and VP7 antigenic phenotypes predicted by the gene composition (Table 3). In general, the neutralization titers of N-MAbs against parent strains and reassortants bearing the same VP4 or VP7 proteins were identical. Two reassortants gave reactivities different from those of the parent strains. A 10-fold or greater reduction in titer against the reassortant compared with those against the parent viruses was observed with the anti-vp7 N-MAbs RV3:2 and RV3:5 against rst3-4 and SA11 and RV5:4 against rrv5-7 and RV5. The antigenic integrities of the reassortants were also examined by using hyperimmune sera raised in rabbits or guinea pigs to each of the parent viruses (Fig. 2). Five of the six reassortants showed neutralization titers equivalent to that of the parent RV from which the VP7 protein was derived. There was a 100-fold reduction in neutralization titer of hyperimmune sera against rst3-4 compared with that observed with SA11. Because rst3-4 was neutralized by the P2[6] VP4-specific N-MAb to the same extent as the ST3 parent, it was TABLE 2. Gene segment, subgroup, and serotype assignments of reassortant RVs Origin of the Origin of the following gene segment a : following protein: Subgroup b G serotype 1 2 3 4 5 6 7 8 9 10 11 VP7 VP4 P serotype [genotype] SA11 S S S S S S S S S S S SA11 SA11 I G3 P[2] rwa-4 S S W 2 W S W S S S S S SA11 Wa II G3 P1A[8] rwa-7 W W W S W W W W W W W Wa SA11 II G1 P[2] Wa W W W W W W W W W W W Wa Wa II G1 P1A[8] rrv5-4 R R R R R S R S 9 R S R SA11 RV5 I G3 P1B[4] rrv5-7 S S S S S S S R S c S S RV5 SA11 I G2 P[2] RV5 R R R R R R R R R R R RV5 RV5 I G2 P1B[4] rst3-4 St St St St St St St St S S St SA11 ST3 II G3 P2[6] rst3-7 S S S S S S S S St S S ST3 SA11 I G4 P[2] ST3 St St St St St St St St St St St ST3 ST3 II G4 P2[6] a R, strain RV5; S, strain SA11; W, strain Wa; St, strain ST3. Subscript numbers denote migration order in the parent strain. b Subgroup assigned on the basis of gene 6 origin. c Gene 8 from both RV5 and SA11 were present; however, the genes were inseparable in the reassortant by PAGE, and therefore, the migration order was not determinable.

512 GORRELL AND BISHOP CLIN. DIAGN. LAB. IMMUNOL. TABLE 3. Antigenic characterization of reassortant and parent strains using serotype-specific N-MAbs Reciprocal of 50% FFN antibody titer of serotype specific N-Ab to indicated rotavirus strain a Anti-VP4 N-Mabs Anti-VP7 N-MAbs Virus strain 2G4 F45:3 RV5:2 RV3:3 RV3:1 RV3:2 RV3:4 RV3:5 RV4:1 RV4:2 RV4:3 RV4:4 RV4:5 RV5:1 RV5:3 RV5:4 RV5:5 ST3:1 ST3:2 ST3:4 Wa 9.2 10 3 2.7 10 4 7 10 4 1.1 10 6 rwa-7 4.7 10 4 1.3 10 4 1.3 10 4 7.1 10 5 rwa-4 6.8 10 3 6.4 10 5 7.7 10 4 7.5 10 4 7.9 10 6 1.8 10 4 SA11 9.5 10 4 1.2 10 6 2.4 10 4 1.6 10 5 8.9 10 6 1,100 1,450 9.8 10 4 1,130 1,870 RV5 4.6 10 5 1.5 10 5 1.5 10 5 2.2 10 5 4.7 10 5 rrv5-7 1.2 10 5 3.8 10 5 8 10 4 1.2 10 4 3.6 10 5 rrv5-4 2.8 10 5 2.8 10 5 4.5 10 4 3 10 5 3 10 6 SA11 1.2 10 5 4 10 5 7.5 10 4 3.6 10 5 6 10 6 ST3 9.3 10 5 6.8 10 5 4.4 10 5 1.1 10 7 rst3-7 3.8 10 4 9.6 10 5 6.4 10 5 1.6 10 7 rst3-4 9.2 10 5 1.4 10 5 1.2 10 4 8.8 10 4 1 10 6 SA11 1.3 10 4 8.5 10 5 1.3 10 5 1.4 10 5 1 10 7 1,170 a Values in boldface type denote 10-fold or greater differences in titer against the reassortant compared with that against the parent strain., titer of 10 3. assumed that recognition of the rst3-4 VP4 by serum N-Ab would be unaffected by the altered VP7 and, therefore, that the immune response to the isolated P2[6] VP4 would be measurable by neutralization assay. Patient sera. Sera derived from three patients naturally infected with P1A[8],G1, P1B[4],G2, or P1A[8],G4 rotavirus strains were used to determine the abilities of these reassortants to discriminate between neutralizing antibody directed at VP4 and VP7 (Fig. 3). Sera taken at both 4 to 6 weeks (convalescent phase) and 4 months (late convalescent phase) postinfection, in addition to the acute-phase sera, were examined when they were available. Reassortants rwa-4 and rwa-7, representing P1A[8] VP4 and G1 VP7, respectively, and the parent strains Wa and SA11 were reacted against acute-, convalescent-, and late-convalescent-phase sera from a patient hospitalized with a P1A[8],G1 HRV infection (Fig. 3A). Comparison of the titers indicated that the immune response was directed predominantly against the VP4 protein, with the degree of seroconversion to VP4 (23-fold) being more than twice that to VP7 (10-fold). There was no seroconversion to the SA11 parent. Reassortants rrv5-4 and rrv5-7, representing P1B[4] VP4 and G2 VP7, respectively, and the parent strains RV5 and SA11 were reacted against acute- and convalescent-phase sera from a patient hospitalized with a P1B[4],G2 HRV infection (Fig. 3B). The comparative titers indicated a greater response against the VP7 protein than the VP4 protein, with seroconversion to VP7 (12-fold) being almost twice that determined against VP4 (6.5-fold). No seroconversion to the SA11 parent was observed. Reassortants rwa-4 and rst3-7, representing P1A[8] VP4 and G4 VP7, respectively, and the parent strains ST3, Wa, and SA11 were reacted against acute-, convalescent-, and late-convalescent-phase sera from a patient hospitalized with a P1A[8], G4 HRV infection (Fig. 3C). The reverse reassortants rst3-4 and rwa-7 were also reacted with the sera. Comparison of the resulting titers suggested that neutralization was mediated predominantly via the VP7 protein. However, representation of the immune response in this patient as the extent of seroconversion, referred to as the seroconversion index (Fig. 4), showed a greater increase in N-Ab directed against VP4 (27- fold) than against VP7 (6.8-fold). There was no seroconversion to the SA11 parent. DISCUSSION Reassortants representing each of the major HRV VP4 and VP7 types were raised for use in neutralization assays to measure N-Abs to VP4 and VP7 separately. Considerable difficulties were experienced in influencing gene reassortment and in selection of the desired reassortants. The choice of SA11 as the VP7 donor for the production of reassortants containing HRV VP4 was influenced by the prior availability of matching reassortants, together with access to the anti-vp4 N-MAb 2G4. In addition, the migration of SA11 gene segments following PAGE differed sufficiently from those of the HRV parent strains to allow assignment of the genes to either parent, particularly those encoding VP4 and VP7. A consequence of using SA11 as the VP7 donor was the difficulty in eliminating SA11 VP4 from the progeny population. The exceptional growth rate and large plaque size bestowed by the SA11 gene 4 product (22) resulted in a growth advantage to reassortants incorporating this gene, even with N-MAb selection. A single strategy could not be adopted for production of the desired reassortants, contrary to the methods used in the production of reassortants containing an HRV VP7

VOL. 4, 1997 HUMAN ROTAVIRUS VP4 AND SA11 VP7 REASSORTANTS 513 FIG. 2. Antigenic characterization of reassortant and parent rotavirus strains with hyperimmune sera (GP, guinea pig; Rab, rabbit) produced against SA11 and HRVs RV4, RV5, and ST3. paired with an animal strain VP4. Successful N-MAb selection against SA11 VP4 was achieved in the most part by trial and error. The influence of phenotypically mosaic viruses (23) was indeterminable, and the failure of a particular selection regimen could not be predicted for any given parent combination. Success was eventually achieved by varying the concentrations and combinations of N-MAbs and the timing of neutralization. The gene compositions of the reassortants varied depending on the human parent used. Of interest was the cosegregation of the HRV VP6 in rwa-4 and rst3-4 with the SA11 gene 10, the product of which (NSP4) interacts specifically with VP6 and possibly VP4 during transport of the subviral particles across the endoplasmic reticulum (1). This successful heterogeneous interaction has been observed previously with Wa (20), even though SA11 is of a different subgroup, but not with ST3. The antigenic integrity of the HRV VP4 and VP7 proteins in the reassortants was shown to be preserved by using a range of N-MAbs, recognizing a variety of epitopes, and hyperimmune sera. Hyperimmune sera, being a polyclonal antibody population with a range of avidities recognizing many different sites, were more sensitive to overall changes in conformation than specific MAbs. The scarcity of anti-vp4 N-MAbs and the known dominance of anti-vp7 N-Ab in hyperimmune sera (19) may have limited antigenic characterization of VP4 in the reassortants. However, the N-MAbs that were available gave favorable results, suggesting that the antigenic integrity of the HRV VP4s had withstood reassortment. Antigenic characterization of the reassortants revealed some alteration of the VP7 of rst3-4 as a result of reassortment. This is possibly a result of interactions with the heterologous VP4 causing a conformational alteration in order to accommodate the two proteins in the outer capsid. Use of the reassortants to analyze VP4 and VP7 N-Ab responses in serum from three patients after primary infection with different G- and P-type rotaviruses provided preliminary information regarding the immunodominance of the outer capsid proteins. The different responses shown in the three patients, however, suggest no constant immunodominance by either VP4 or VP7 following HRV infection. Examination of the patient sera revealed some of the limitations of using reassortants to measure N-Ab levels. For example, the identical N-Ab titers against the Wa and ST3 parents (Fig. 3C) FIG. 3. N-Ab titers in patient serum against reassortant and parent rotavirus strains following natural rotavirus infection with a P1A[8],G1 strain (A), a P1B[4],G2 strain (B), and a P1A[8],G4 strain (C)., acute phase; 0, convalescent phase;, late convalescent phase.

514 GORRELL AND BISHOP CLIN. DIAGN. LAB. IMMUNOL. FIG. 4. N-Ab levels in response to a P1A[8],G4 rotavirus infection expressed as the fold increase in N-Ab titer with acute-phase levels in serum as the background. The line indicates a fourfold increase; an increase greater than this is defined as a seroconversion., convalescent phase;, late convalescent phase. decreased and increased, respectively, when rwa-4 and rst3-7 were substituted. This suggested differences in the kinetics of neutralization such that antibody interactions with strains containing VP7 and VP4 in their native conformation varied in some way from those with reassortant strains containing the same VP7 or VP4 proteins in isolation. Thus, the titer determined for rst3-7 was higher than the titer determined for ST3, possibly as a result of differences in the efficiency of virus neutralization rather than the presence of more N-Ab recognizing rst3-7. In view of these unpredictable findings, interpretation of the results required a measure other than N-Ab titer in order to compare VP4 and VP7 responses between different patients. The N-Ab titer may be influenced by many factors such as preexisting antibody titers, virion stability and ease of neutralization (28), and changes in epitope recognition due to altered conformation as a result of reassortment (2). As a consequence, the degree of seroconversion to each outer capsid protein, using acute-phase sera titer as a baseline, could be a more accurate parameter for indicating the extent of the immune response against each protein. This approach would remove the influence of neutralization kinetics and virus stability for each strain used and would allow for a comparison of the change in antibody level against each protein as a result of natural infection. Preliminary results obtained with these reassortants were encouraging, and their use for examining neutralizing antibody could allow for the mapping of homotypic and heterotypic immune responses to VP4 and VP7 following HRV infections in humans. Comparison of responses in serum and intestinal contents after primary infections and reinfections could assist in the development of a successful RV vaccine for humans. ACKNOWLEDGMENTS This work was supported by the National Health and Medical Research Council of Australia and the Royal Children s Hospital Research Foundation. REFERENCES 1. Au, K., N. M. Mattion, and M. K. Estes. 1993. A subviral particle binding domain on the rotavirus nonstructural glycoprotein NS28. Virology 194:665 673. 2. Chen, D., M. K. Estes, and R. F. Ramig. 1992. Specific interactions between rotavirus outer capsid proteins VP4 and VP7 determine expression of a cross-reactive, neutralizing VP4-specific epitope. J. Virol. 66:432 439. 3. Christy, C., P. Offit, F. Clark, and J. Treanor. 1993. Evaluation of a bovinehuman rotavirus reassortant vaccine in infants. J. Infect. Dis. 168:1598 1599. 4. Coulson, B. S. 1993. Typing of human rotavirus VP4 by an enzyme immunoassay using monoclonal antibodies. J. Clin Microbiol. 31:1 8. 5. Coulson, B. S., K. J. Fowler, R. F. Bishop, and R. G. H. Cotton. 1985. Neutralizing monoclonal antibodies to human rotavirus and indications of antigenic drift among strains from neonates. J. Virol. 54:14 20. 6. Coulson, B. S., and C. D. Kirkwood. 1991. Relation of VP7 amino acid sequence to monoclonal antibody neutralization of rotavirus and rotavirus monotype. J. Virol. 65:5968 5974. 7. Coulson, B. S., C. D. Kirkwood, P. M. Masendycz, R. F. Bishop, and G. Gerna. 1996. Amino acids involved in distinguishing between monotypes of rotavirus G serotypes 2 and 4. J. Gen. Virol. 77:239 245. 8. Coulson, B. S., J. M. Tursi, W. J. McAdam, and R. F. Bishop. 1986. Derivation of neutralizing monoclonal antibodies to human rotaviruses and evidence that an immunodominant neutralization site is shared between serotypes 1 and 3. Virology 154:301 312. 9. Dyall-Smith, M. L., and I. H. Holmes. 1984. Sequence homology between human and animal rotavirus serotype-specific glycoproteins. Nucleic Acids Res. 12:3973 3982. 10. Estes, M. K. 1996. Rotaviruses and their replication, p. 1625 1655. In B. N. Fields, D. M. Knipe, P. M. Howley, R. M. Chanock, J. L. Melnick, T. P. Monath, B. Roizman, and S. E. Straus (ed.), Fields virology, 3rd ed. Lippincott-Raven Publishers, Philadelphia, Pa. 11. Grimwood, K., J. C. S. Lund, B. S. Coulson, I. L. Hudson, R. F. Bishop, and G. L. Barnes. 1988. Comparison of serum and mucosal antibody responses following severe acute rotavirus gastroenteritis in young children. J. Clin. Microbiol. 26:732 738. 12. Hoshino, Y., L. J. Saif, M. M. Sereno, R. M. Chanock, and A. Z. Kapikian. 1988. Infection immunity of piglets to either VP3 or VP7 outer capsid protein confers resistance to challenge with a virulent rotavirus bearing the corresponding antigen. J. Virol. 62:744 748. 13. Hoshino, Y., M. M. Sereno, K. Midthun, J. Flores, A. Z. Kapikian, and R. M. Chanock. 1985. Independent segregation of two antigenic specificities (VP3 and VP7) involved in neutralization of rotavirus infectivity. Proc. Natl. Acad. Sci. USA 82:8701 8704. 14. Knowlton, D. R., D. M. Spector, and R. L. Ward. 1991. Development of an improved method for measuring neutralizing antibody to rotavirus. J. Virol. Methods 33:127 134. 15. Kobayashi, N., K. Kojima, K. Taniguchi, T. Urasawa, and S. Urasawa. 1994. Genotypic diversity of reassortants between simian rotavirus SA11 and human rotaviruses having different antigenic specificities and RNA patterns. Res. Virol. 145:303 311. 16. Kobayashi, N., K. Taniguchi, T. Urasawa, and S. Urasawa. 1993. Efficient production of antigenic mosaic reassortants of rotavirus with the aid of anti-vp4 and anti-vp7 neutralizing monoclonal antibodies. J. Virol. Methods 44:25 34. 17. Laemmli, U. K. 1970. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227:680 685. 18. Lazdins, I., B. S. Coulson, C. Kirkwood, M. Dyall-Smith, P. J. Masendycz, S. Sonza, and I. H. Holmes. 1995. Rotavirus antigenicity is affected by the genetic context and glycosylation of VP7. Virology 209:80 89. 19. Matsui, S. M., E. R. Mackow, and H. B. Greenberg. 1989. Molecular determinant of rotavirus neutralization and protection. Adv. Virus Res. 36:181 214. 20. Meyer, J. C., C. C. Bergmann, and A. R. Bellamy. 1989. Interaction of rotavirus cores with the nonstructural glycoprotein NS28. Virology 171:98 107. 21. Shaw, R. D., P. T. Vo, P. A. Offit, B. S. Coulson, and H. B. Greenberg. 1986. Antigenic mapping of the surface proteins of rhesus rotavirus. Virology 155:434 451. 22. Taniguchi, K., K. Nishikawa, N. Kobyashi, T. Urasawa, H. Wu, M. Gorziglia, and S. Urasawa. 1994. Differences in plaque size and VP4 sequence found in SA11 virus clones having simian authentic VP4. Virology 198:325 330. 23. Ward, R. L., D. R. Knowlton, and H. B. Greenberg. 1988. Phenotypic mixing during coinfection of cells with two strains of human rotavirus. J. Virol. 62:4358 4361. 24. Ward, R. L., D. R. Knowlton, H. B. Greenberg, G. M. Schiff, and D. I. Bernstein. 1990. Serum-neutralizing antibody to VP4 and VP7 proteins in infants following vaccination with WC3 bovine rotavirus. J. Virol. 64:2687 2691. 25. Ward, R. L., M. M. McNeal, D. S. Sander, H. B. Greenberg, and D. I. Bernstein. 1993. Immunodominance of the VP4 neutralization protein of rotavirus in protective natural infections of young children. J. Virol. 67:464 468. 26. Xu, Z., M. E. Hardy, J. D. Williams, G. N. Woode, and R. F. Ramig. 1993. Immunodominant neutralizing antigens depend on the virus strain during a primary immune response in calves to bovine rotavirus. Vet. Microbiol. 35:33 43. 27. Xu, Z., and G. N. Woode. 1993. Studies on the role of VP4 of G serotype 10 rotavirus (B223) in the induction of heterologous immune response in calves. Virology 196:294 297. 28. Zhou, Y. J. W. Burns, Y. Morita, T. Tanaka, and M. K. Estes. 1994. Localization of rotavirus VP4 neutralization epitopes involved in antibody-induced conformational changes of virus structure. J. Virol. 68:3955 3964.