Serologic and Genomic Characterization of a G12 Human Rotavirus in Thailand

Abstract

The G and P type specificity of the human rotavirus strain T-152 (G12P[9]) isolated in Thailand was serologically confirmed with G12-specific monoclonal antibodies prepared in this study by using a reference G12 strain, L26, as an immunizing antigen and a P[9]-specific monoclonal antibody, respectively. The genomic relationship of strain T-152 with representative human rotavirus strains was examined by means of Northern blot analysis. The results showed that T152 is closely related to strain AU-1 (G3P[9]). Gene 5 (NSP1 gene) of T152, which did not hybridize with those of any other strains examined, was characterized by sequence determination. The T152 NSP1 gene is 1,652 nucleotides in length, encodes 493 amino acids, and exhibits low identity to those of representative human and animal rotaviruses.

Rotavirus, a member of the Reoviridae family, is the most common agent of severe, dehydrating gastroenteritis in infants and young children and in the young of most mammalian species (12). In developing countries, rotavirus infection results in high mortality, and an annual death rate of 500,000 to 600,000 persons has been estimated. Furthermore, in developed countries, rotavirus infection is a cause of high morbidity. Vaccination is thought to be the best way to reduce this significant mortality and morbidity worldwide.

Two rotavirus outer capsid proteins, a glycoprotein VP7 and a protease-sensitive VP4, have independent serotype specificities, G (VP7) serotype and P (VP4) serotype, respectively, and rotaviruses are classified by a binary system as are used for influenza viruses. For P typing, sequence analysis has been adopted due to the lack of readily available typing sera. A total of 15 G serotypes have been reported. Among them, 10 G serotypes have been detected in humans. G1 to G4 are the major G serotypes, with G5, G6, G8 to G10, and G12 being minor or unusual ones (5, 9, 12). In contrast, 22 P types have been recognized, with at least 10 P genotypes having been detected in humans (5, 8, 9, 12). Recently a number of human rotavirus strains with unusual G or P types and rare combinations of G and P types have been detected worldwide (1-9, 15-17, 20, 21, 27-29). For example, G5 was detected in almost half of the rotavirus-positive samples in Brazil (7). G9 is increasing rapidly (3, 17) and has become more common than G4 in many locations. Human G8 strains have been detected in Africa at a high frequency (1, 2, 4). P[8] is the most common, followed by P[4] and P[9]. Recently, P[6], which was first detected as an asymptomatic infection in neonates, has been increasing (1, 4, 9).

Since G12 was first detected in stool specimens collected from diarrheic children under 2 years of age between December 1987 and February 1988 in the Philippines (24, 28), no further report on the detection of G12 in humans or animals has appeared, although extensive surveys on the distribution of the G serotype worldwide have been conducted. In a previous study, however, members of our group detected a human G12P[9] rotavirus, T152, in Thailand and characterized it by means of reverse transcription-PCR and sequence determination (18). Also, in the United States, Griffin et al. (9) found a G12 strain with P[6] specificity. These results imply that the prevalence of G12 strains is expanding. In this study, we analyzed the Thai G12 strain T152 serologically using neutralizing monoclonal antibodies specific to G12 and P[9]. In addition, genomic relatedness between the T152 strain and representative human rotavirus strains was examined by Northern blot hybridization. Furthermore, we found the uniqueness of the NSP1 gene of strain T152.

The following representative human and animal rotavirus strains were employed: KU (human, G1P[8]), K8 (human, G1P[9]), S2 (human, G2P[4]), YO (human, G3P[8]), AU-1 (human, G3P[9]), Hosokawa (human, G4P[8]), OSU (porcine, G5P[7]), NCDV (bovine, G6P[1]), Ty-1 (turkey, G7P[17]), 69 M (G8P[10]), WI-61 (human, G9P[8]), B223 (bovine, G10P[11]), YM (porcine, G11P[7]), L26 (human, G12P[4]), L27 (human, G12P[4]), T152 (human, G12P[9]), L338 (G13P[18]), and FI-23 (equine, G14P[12]). Each rotavirus strain was pretreated with 10 μg of trypsin (type IX, from porcine pancreas and crystallized; Sigma) per ml, inoculated onto MA-104 cells in the presence of trypsin (1 μg/ml), and then harvested 1 to 3 days after infection.

The purified L26 strain was used as an immunizing antigen. P3-X63-Ag8.653 mouse myeloma cells were fused with spleen cells from mice immunized intraperitoneally with L26 as described previously (23). To obtain ascitic fluid, 10⁷ hybridoma cells were inoculated intraperitoneally into Pristane-primed BALB/c mice.

A 1:5 dilution (for hybridoma screening) of culture fluid or twofold serial dilutions (for determination of neutralizing antibody titers) of ascitic fluid was reacted with a virus suspension containing about 500 fluorescent cell-forming units/0.025 ml for 1 h. At 18 to 24 h postinfection, the infected cells were washed with phosphate-buffered saline, fixed with cold (−80°C) ethanol, and then reacted with a 1:30 dilution of anti-human rotavirus rabbit serum for 1 h. After an additional 1-h reaction with a 1:50 dilution of fluorescein isothiocyanate-conjugated anti-rabbit immunoglobulin G goat serum (Seikagaku Kogyo, Tokyo, Japan), the number of fluorescent cells was determined by vertically illuminated fluorescence microscopy. The neutralization titer was expressed as the reciprocal of the highest serum dilution that reduced the fluorescent-cell count by more than 60%.

An enzyme-linked immunosorbent assay with monoclonal antibodies was carried out as described previously (25). The following monoclonal antibodies were used: group A-common YO-156 (directed to VP6), subgroup I-specific S2-37 (VP6), subgroup II-specific YO-5 (VP6), G1-specific KU-4 (VP7), G2-specific S2-2G10 (VP7), G3-specific YO-1E2 (VP7), G4-specific ST-2G7 (VP7), and a group A-common YO-2C2 (VP4).

Rotavirus double-stranded RNA was extracted from stools and culture fluid with a disruption solution comprising 1% sodium dodecyl sulfate, 0.1% 2-mercaptoethanol, and 50 mM EDTA and then with phenol and chloroform. The RNA was electrophoresed in 10% acrylamide gels (2 mm thick) for 16 h at 20 mA at room temperature. RNA segments were visualized by silver staining.

Full-length cDNA of the NSP1 gene of culture-adapted strain T152 was prepared by reverse transcription-PCR. PCR-amplified cDNA was ligated into the pCRII vector with a TA cloning kit (Invitrogen Corp). The PCR products and three cDNA clones were sequenced with the ABI PRISM BigDye terminator cycle sequencing ready reaction kits (PE Biosystems, Chiba, Japan) and an automated sequencer, the ABI PRISM 310 genetic analyzer (PE Applied Biosystems, Foster City, Calif.). Nucleotide sequences were analyzed for construction of a phylogenetic tree using the Neighbor-Joining method.

Northern blot hybridization was carried out as previously described (19). Briefly, after polyacrylamide gel electrophoresis (PAGE) analysis, double-stranded RNA was denatured by soaking the gel in 0.1 N NaOH and 0.25 M NaCl for 20 min and was then neutralized in 4× Tris-acetate-EDTA for 20 min twice and in 1× Tris-acetate-EDTA for 20 min. Electrotransfer of rotavirus RNA to Hybond N+ (Amersham) was conducted at 0.2 mA overnight at 4°C. Hybridization was performed with an enhanced chemiluminescence direct nucleic acid labeling and detection system (Amersham) according to the instructions of the manufacturer. Stringency was regulated by changing the concentration of the SSC solution (1× SSC is 0.15 M NaCl plus 0.015 M sodium citrate) for primary and secondary wash buffers.

For serological characterization of strain T152, we prepared neutralizing monoclonal antibodies specific to G12 rotaviruses. Four hybridoma clones were established in two fusion experiments, using strain L26 (G12P[4]) as an immunizing antigen. In neutralization tests involving various human and animal rotavirus strains, the four monoclonal antibodies all reacted specifically with the Philippine G12 strains L26 and L27 (Table 1). The antibodies also neutralized strain T152 with similar efficiency (Table 1).

TABLE 1.

Reactivity patterns of monoclonal antibodies to human and animal rotaviruses

Strain	G, P type	Neutralizing titer of monoclonal antibody
		1G11	3B5	3G9	4E7
KU	G1P[8]	<100	<100	100	<100
K8	G1P[9]	<100	<100	<100	<100
S2	G2P[4]	<100	<100	<100	<100
YO	G3P[8]	<100	<100	<100	<100
AU-1	G3P[9]	<100	<100	<100	<100
Hosokawa	G4P[8]	100	<100	<100	100
OSU	G5P[7]	<100	<100	<100	100
NCDV	G6P[1]	<100	400	200	200
Ty-1	G7P[17]	<100	<100	200	<100
69M	G8P[10]	<100	400	<100	<100
WI-61	G9P[8]	200	100	<100	200
B223	G10P[11]	<100	200	<100	<100
YM	G11P[7]	<100	100	<100	<100
L26	G12P[4]	3,200	102,400	51,200	102,400
L27	G12P[4]	3,200	102,400	51,200	25,600
T152	G12P[9]	3,200	102,400	12,800	12,800
L338	G13P[18]	<100	100	<100	<100
FI-23	G14P[12]	<100	100	<100	<100

P9-specific K8-2C12 monoclonal antibody prepared previously (13) neutralized strain T152 as well as strains K8 and AU-1 with P[9] specificity (data not shown). Thus, we confirmed serologically that the G and P type specificity of strain T152 was G12P[9].

The overall genomic relatedness determined through RNA-RNA hybridization assays has revealed that there are three distinct human-specific genogroups that are almost not related to one another at all: the Wa, DS-1, and AU-1 genogroups. In this study, we performed Northern blot hybridization to examine the overall genomic relatedness of strain T152 with strains KU, S2, and AU-1, representing the Wa, DS-1, and AU-1 genogroups, respectively, since this assay can reveal the relatedness in a segment-to-segment manner and can permit the reuse of the blot with different probes.

The T152 probe reacted with nine RNA segments of strain AU-1, although the probe showed no reaction or only a faint reaction with a few RNA segments in the genome of strain KU, S2, or L26 (Fig. 1). A reciprocal assay with the AU-1 probe also showed the high relatedness between strains T152 and AU-1. In the assay with the L26 probe, the probe reacted only with RNA segment 7, which may be the VP7 gene, and segment 11 of T152 (Fig. 1). As described previously (14), L26 was found to be partly related to strains S2 and KU, members of the DS-1 genogroup and Wa genogroup, respectively. In contrast, the KU and S2 probes did not exhibit any significant relationship with T152 (Fig. 1). Thus, T152 was shown to be highly related to strain AU-1, except in genes 5 and 7.

FIG. 1.

Northern blot hybridization analysis of strain T152. (A) RNA profiles on PAGE. Lanes: 1, strain KU; 2, strain S2; 3, strain AU-1; 4, strain T152; 5, strain L26. (B) Northern blot analysis using the T152 probe. (C) Northern blot analysis using the KU probe. (D) Northern blot analysis using the AU-1 probe. (E) Northern blot analysis using the L26 probe. (F) RNA profiles on PAGE. Lanes: 1, strain KU; 2, strain S2; 3, strain AU-1; 4, strain L26; 5, strain T152. (G) Northern blot analysis using the S2 probe. Northern blot analyses in panels B to E and G were performed using the blot transferred from the polyacrylamide gel shown in panels A and F, respectively.

With Northern blot hybridization, gene 5 of strain T152 did not react with any of the probes employed and vice versa. This finding prompted us to characterize the T152 gene 5 (the NSP1 gene) by sequencing. The NSP1 gene of strain T152 was found to be 1,652 nucleotides in length and to encode 493 amino acids. Comparison of its sequence with the published NSP1 sequences revealed the uniqueness of the T152 NSP1 gene; the nucleotide sequence identity ranged from 46.8 to 63.5% (Table 2). Phylogenetic analysis of the NSP1 genes showed the species relatedness, and the T152 NSP1 gene was found to be related to those of bovine rotaviruses (Fig. 2).

TABLE 2.

Nucleotide and amino acid sequence homologies of NSP1 gene and protein from strain T152 with NSP1 genes and proteins from representative human and animal rotavirus strains

Strain	G type and P type	Species	Homology (%)
			Nucleotide	Amino acid
KU	G1P[8]	Human	61.5	50.4
DS1	G2P[4]	Human	62.0	51.1
K8	G1P[9]	Human	62.5	51.5
69M	G8P[10]	Human	61.7	49.5
L26	G12P[9]	Human	61.3	51.6
A44	G10P[11]	Cow	63.5	55.8
A5-10	G8P[1]	Cow	60.8	*a
A5-13	G8P[1]	Cow	63.5	53.0
UK	G6P[5]	Cow	63.2	55.0
Gottfried	G4P[6]	Pig	62.1	50.1
OSU	G5P[7]	Pig	62.0	52.8
YM	G11P[7]	Pig	61.5	50.5
FI-14	G3P[12]	Horse	56.9	40.3
H2	G3P[12]	Horse	57.0	40.9
L338	G13P[18]	Horse	56.4	38.0
SA11	G3P[2]	Monkey	57.9	42.5
RRV	G3P[3]	Monkey	55.9	39.8
EW	G3P[16]	Mouse	53.6	39.9
PO-13	G7P[17]	Pigeon	46.8	17.5

^a*, Since NSP1 of strain A5-10 is 40 amino acids in length due to the presence of a nonsense codon at nucleotides 153 to 155, the homology could not be determined.

FIG. 2.

Phylogenetic tree for the nucleotide sequences of the NSP1 genes of strain T152 and other representative human and animal rotaviruses. The bootstrap confidence levels obtained by 1,000 replicates are shown. The bar indicates the variation scale.

Although no reports have appeared on the detection of G12 rotaviruses since the detection of human G12 rotaviruses (prototype strain L26) in the Philippines in 1990 (24, 28), two human G12 strains, T152 and Se585, were very recently detected in Thailand and in the United States, respectively (9, 18). On comparison of the VP7 genes of strains L26, T152, and Se585, higher identity (97.8 and 98.5%, nucleotide and amino acid, respectively) was found between T152 and Se585 than that (90.2 or 91.0% and 92.9 or 94.2%, respectively) between L26 and T152 or Se585. The VP7 gene of strain L26 has accumulated point mutations, and the present G12 strains (strains T152 and Se585) with the VP7 gene may have evolved. However, the genome constellation of the two strains is quite different. Seven and 10 genes, respectively, of Se585 hybridized to those of strains L26 and US1205 of the DS-1 genogroup (9). Few or no hybrids were formed between the genomes of Se585 and Wa or AU-1 (9). Thus, strain Se585 was considered to be a natural single-gene reassortant between strain L26 and strain US1205 or a similar one of the DS1 genogroup. In contrast, the genes of T152 hybridized to nine genes of strain AU-1. The VP7 gene is related to L26, and the origin of the NSP1 gene of T152 is unknown. This is a striking contrast with the finding that RNA segments of Se585 did not react to any of the genes of strain AU-1 (14). These results imply that T152 is also a natural reassortant between L26 and a strain in the AU-1 genogroup. A second reassortment step might have occurred for acquisition of the unusual NSP1 gene from an unknown third strain. To begin with, strain L26 is thought to be a reassortant. In our previous study involving RNA-RNA hybridization assaying, it was found that two or three genes were from the Wa genogroup and five or six were from the DS-1 genogroup (14). Thus, the reassortment of rotaviruses may occur through multiple steps between two reassortants or between a reassortant and a nonreassortant virus, such as strain Wa, DS-1, or AU-1, reference strains of the Wa, DS-1, and AU-1 genogroups, respectively.

In this study, we first prepared G12-specific neutralizing monoclonal antibodies. The G12-specific monoclonal antibodies were potent in neutralization tests and enzyme-linked immunosorbent assays. Although the protein specificity (VP7 or VP4) of the antibodies could not be determined, they may be directed to VP7, since they did not react with any strains of P[4], which is the P type of strain L26 employed as an immunizing antigen for hybridoma production. The use of reassortants or the preparation of mutants resistant to the antibodies is necessary for a final conclusion. These antibodies will be quite useful for large-scale epidemiological surveys for the detection of G12 rotaviruses.

In order to determine the overall genomic relatedness of rotavirus strains, Northern blot assays were performed in this study. Compared to the liquid RNA-RNA hybridization assays used more commonly, Northern blot assays have some advantages: the relatedness in a segment-to-segment manner can be determined, the reuse of blots for a different probe is possible, and there is no necessity for radioisotopes for the preparation of probes. However, depending on the lengths of RNA segments, the cutoff levels might be different; short RNA segments, such as segments 10 and 11, tend to react more readily with each other than longer RNA segments exhibiting the same homology. As a whole, segment-to-segment comparison is very useful. Indeed, we could detect the uniqueness of gene 5 of strain T152.

The gene 5 equivalent to the NSP1 gene (1,652 nucleotides) of T152 is much longer than those of other rotaviruses except for that (1,870 nucleotides) of pigeon strain PO-13 (11). It has been shown that NSP1 genes among rotaviruses exhibit great diversity (10, 26). Among the 11 genes, the NSP1 gene exhibits the least identity. In addition, several truncated NSP1 proteins have been reported for the rearranged NSP1 genes (1). Furthermore, we previously found strains (A5-10 and A5-16) with NSP1 proteins of only 40 or 50 amino acids (22). Some clones of these strains produced very large plaques and induced diarrhea in suckling mice as efficiently as the strains with the normal NSP1 genes (Taniguchi et al., unpublished data). Further characterization of a unique NSP1 gene of strain T152 in vitro and in vivo will provide more useful data.

Although G1 to G4 and P[8] or P[4] rotaviruses are common worldwide, strains with unusual properties appear to prevail more frequently than before. In particular, G9 strains are increasing (3, 17), and we also detected G9 strains at an extremely high frequency in a district in Japan (Taniguchi et al., unpublished data). The detection of G12 strains in Thailand and the United States implies expansion of the distribution of the G12 strains. For the development of an effective rotavirus vaccine and for the examination of rotavirus ecology, it is still necessary to continue to survey the G and P type distributions worldwide.

Nucleotide sequence accession numbers.

The nucleotide sequence data reported in this paper for the NSP1 gene of strain T152 have been deposited with the DDBJ/EMBL/GenBank data libraries under accession no. AB097459.