Previous Article | Next Article ![]()
Journal of Clinical Microbiology, September 2004, p. 4338-4343, Vol. 42, No. 9
0095-1137/04/$08.00+0 DOI: 10.1128/JCM.42.9.4338-4343.2004
Copyright © 2004, American Society for Microbiology. All Rights Reserved.
Regional Laboratory of Virology, Baranya County Institute of State Public Health Service,1 Department of Medical Microbiology and Immunology,3 Department of Medical Genetics and Child Development, Faculty of Medicine, University of Pécs, Pécs, Hungary,4 Department of Animal Health and Well-Being, University of Bari, Bari, Italy2
Received 18 February 2004/ Returned for modification 10 May 2004/ Accepted 27 May 2004
|
|
|---|
|
|
|---|
Thus far, at least 10 G types (G1 to G6, G8 to G10, and G12) and 11 P types (P[1], P[3] to P[6], P[8] to P[11], P[14], and P[19]) have been identified in humans. Most of these specificities can also be found in animals, suggesting that certain human G and P types, especially those that are rarely detected, might have originated from animal strains either by direct interspecies transmission or by reassortment of cognate genes between heterologous and homologous strains. Serotypes G6 and G8, identified frequently in ruminants, or serotype G5, common in pigs but rare in humans, exemplify this hypothesis (2, 4, 6, 11). Although direct epidemiological evidence is still lacking, studies based on full genome hybridization and, more recently, on sequence and phylogenetic analysis seem to support the hypothesis that interspecies transmission may occur (4, 13, 29), even though such analyses usually cannot demonstrate direct evolutionary ancestry between recent animal and human strains. Indeed, gene sequences of animal strains segregate in most instances apart from those of human strains, suggesting either a strong diversification after the host switching or unequal distribution of distinct phylogenetic lineages in different hosts.
Early molecular epidemiological data together with recent surveillance studies identified five predominant G types (G1 to G4 and G9) and three common P types (P[4], P[6], and P[8]) mainly in six combinations (P[8],G1; P[4],G2; P[8],G3; P[8],G4; P[6],G9; and P[8],G9) to have clinical and epidemiological significance, particularly in temperate climates. By extrapolation of the data collected from international reports, these six strains are responsible for more than 90 to 95% of all hospitalization cases, while the remaining 5 to 10% of cases are associated with unusual combinations of the common P and G types (e.g., P[4],G1; P[6],G3; or P[8],G2) and with rare strains, some of which seem to have primarily local relevance (e.g., P[9],G6 in Hungary or P[8],G5 in South America). Although epidemiological observations suggest the existence of genetic linkages between the most common human VP4 and VP7 specificities (genotype P[8] is in linkage with the G1, G3, G4, and G9 specificities, whereas genotype P[4] is associated mainly with the G2 specificity), in this regard the relatively prevalent human P type, P[6], may represent an exception based on data that it has been found to be in association with most of the human VP7 specificities (5, 6, 13, 29).
During the past two decades, the epidemiological role of genotype P[6] has been substantially reconsidered. In early studies P[6] strains were detected exclusively from children with asymptomatic infection, suggesting that the P[6] specificity is associated with virus attenuation and, therefore, raising hopes that such strains may serve as components of future vaccine candidates (8, 32). However, subsequent epidemiological investigations identified P[6] strains as important human pathogens. To date, P[6] strains have been detected from both asymptomatic neonatal infections and symptomatic infantile infections in association with a wide variety of G types (5, 6, 13, 18, 29). At present there is no epidemiological data on neonatal asymptomatic infections in Hungary; however, genotype P[6] strains, in linkage with G4 specificity, have been sporadically identified from episodes of gastroenteritis of infants and young children (2).
Although P[6] is one of the three most prevalent VP4 genotypes in humans (6), the relatively few sequence data available in the public databases suggest that the heterogeneity within this genotype might be presently unexplored. Sequencing studies have demonstrated that almost all the human P[6] strains detected worldwide belong to one lineage, M37-like. However, substantial differences were observed in the VP8* sequences of two additional P[6] strains, namely, AU19, a unique human G1 rotavirus displaying a supershort RNA electropherotype that was isolated from a child affected with severe diarrhea and dehydration in Japan (26), and the porcine G4 strain, Gottfried, that was originally isolated from the intestinal content of a suckling pig with diarrhea (3). These findings indicate that strains AU19 and Gottfried may be considered prototypes of two additional P[6] lineages (26) and also suggest that the P[6] VP4 gene has undergone genetic and antigenic drift, similar to what has been described for the P[8] VP4 gene and for the G1, G3, G4, G6, and G9 VP7 genes (4, 12, 16, 17, 22, 24, 25, 34, 36). Interestingly, all three lineages of genotype P[6] represent serologically distinct subtypes, designated P2A[6] (M37-like) and P2C[6] (AU19-like) for the human strains and P2B[6] (Gottfried-like) for the porcine strain (20, 26).
In order to gain further insights into the genetic variability within genotype P[6], in this study we sequenced and analyzed partial sequences of the VP4 gene (VP8*) of P[6],G4 strains identified during an ongoing surveillance in Hungary.
Rotavirus-positive stool samples collected between 1992 and 2002 from children with diarrhea were serotyped and genotyped as was previously reported (2). Based on the genotype profile we randomly selected a total of six P[6],G4 strains for analysis. The laboratory work and subsequent computer analysis consisted of the following steps. Briefly, viral RNA was extracted using the guanidine thiocyanate/Glass Milk method as previously described (9). The RNA was heat denatured in the presence of the VP4-specific consensus primer (Con3), chilled immediately on ice, and subsequently reverse transcribed using the avian myeloblastosis virus reverse transcriptase (Promega). The resulting cDNA strand was amplified with Taq DNA polymerase (Promega) with primers Con3 and Con2 (9). The amplicons were electrophoresed in a low-melting-point agarose gel, subsequently excised, and extracted using the QIAquick gel extraction kit (QIAGEN). Cycle sequencing was carried out using the BigDye 1.1 kit (Applied Biosystems) with primers utilized in the PCR. Dye-labeled products were precipitated with ethanol in the presence of sodium acetate and then were dried. The products were subsequently resuspended in formamide and then run on an automated sequence analyzer (type 3100; ABI Prism). Raw sequence data were edited with the GeneDoc 1.1 program (27). Related sequences were searched in the public databases using the BLAST algorithm (1), and an alignment was made from our sequences and the downloaded sequences with the ClustalW algorithm in DAMBE software (35). The phylogenetic analysis was performed with the MEGA2 program using the p-distance algorithm (obtained by dividing the number of nucleotide differences by the total number of nucleotides compared) and the neighbor-joining method supported by bootstrap analysis (19).
The primer pair Con3/Con2 amplifies a product of about 880 bp in length (9), including the hypervariable region and a small fragment of the conserved region of the VP4 gene, called VP8* and VP5*, respectively. Our human P[6] strains gave amplicons of the expected sizes (data not shown). BLAST search identified the closest relatives among P[6] strains, demonstrating that our strains belong to genotype P[6]. This result confirmed the genotyping results obtained by multiplex reverse transcription-PCR (2 and unpublished data). Sequence comparison and phylogenetic analyses (Table 1 and Fig. 1) indicated, however, that each of these six strains was distinct from the human P[6] strains characterized previously. Indeed, phylogenetic analysis identified at least five separate lineages. In agreement with previous studies, the vast majority of human strains clustered in a single large lineage, M37-like, while two minor lineages, represented each by a single strain, were the human AU19-like and the porcine Gottfried-like lineages. The Hungarian human P[6],G4 strains clustered in two novel lineages, designated lineage IV (designated for the BP1198/98-like strains) and lineage V (designated for the BP720/93-like strains), a phylogenetic finding that was statistically supported by bootstrap analysis (Fig. 1).
|
View this table: [in a new window] |
TABLE 1. Distance matrix of sequence divergence within the VP8* gene (spanning nt 200 to 700 and aa 65 to 230) of P[6] strains
|
![]() View larger version (29K): [in a new window] |
FIG. 1. Phylogenetic tree for the VP8* fragment of VP4 gene (nt 200 to 700) of strains with P[6] specificity by using the neighbor-joining algorithm. Bootstrap values greater than 60% are indicated at the branch nodes. The scale bar is proportional to the genetic distance. An outgroup sequence (P[19]) was included to better understand the phylogenetic relationships among P[6] strains. Data about the associated G serotype, the country of isolation, and whether the strain was identified from symptomatic (S) or asymptomatic (AS) infections are displayed where information was available. Unknown data are indicated with a question mark.
|
2.2%; aa,
3.6%) and lineage V (nt,
6.0%; aa,
4.8%) fell virtually into the same range (Table 1). In contrast, divergence between strains of lineage IV and strains belonging to the four other lineages were considerably higher (nt, 14.7 to 20.1%; aa, 8.5 to 21.2%). Comparable divergences were found between lineage V and the other lineages (nt, 14.7 to 22.1%; aa, 8.5 to 21.8%; Table 1). In this comparison the lowest interlineage variation was found between strains of lineage IV and lineage V (nt, 14.7 to 16.1%; aa, 8.5 to 11.5%; Table 1). When a longer sequence (795 bp; nt 56 to 850) was analyzed, we found essentially the same ranges of divergence with both the M37-like human strains (lineage IV, 14.7 to 17.1% for nt and 9.8 and 16.3% for aa; lineage V, 15.5 to 17.9% for nt and 12.9 to 17.0% for aa) and the Gottfried strain (lineage IV, 19.4 to 19.9% for nt and 15.2 to 15.9% for aa; lineage V, 18.4 to 19.5% for nt and 15.2 to 16.3% for aa; data not shown). Overall, the sequence variation for the partial VP4 gene fell approximately into the same ranges in any interlineage comparison, except for lineage IV and lineage V strains, which were somewhat more closely related to each other. Direct inspection of the sequence alignment (Fig. 2) revealed that all Hungarian P[6] strains and the atypical Japanese human P[6] strain share several amino acid substitutions with the porcine strain Gottfried (aa 98, R instead of K; aa 101, V instead of I; aa 105, I instead of L; aa 108, I instead of V; aa 134, S instead of T; aa 147, A instead of V; aa 170, G instead of Y; aa 171, G instead of N; aa 172, R instead of S). Some residues (aa 91, V instead of I; aa 196, V instead of I) were conserved between the M37-like strains and the porcine strain Gottfried, but they were absent in the other human lineages. All these changes may represent a reversion to or, presumably, retention of the ancestral sequence, suggesting that the three minor human lineages are in an intermediate position between the major human M37-like lineage and the porcine virus prototype Gottfried. Thus, assuming that the conserved sites shared by several lineages represent the ancestral sequence, the presence of conserved residues observed exclusively in the human lineages very likely represent amino acid sites that were exposed to more intensive genetic drift in pigs. Alternatively, these conserved amino acids might have importance in the process of adaptation to the human host by the human P[6] lineages.
![]() View larger version (59K): [in a new window] |
FIG. 2. An alignment for the partial amino acid sequences of genotype P[6] specificity. The amino acid sites addressed to be involved in attenuation are highlighted (white letters on black background) (5, 28). The residues conserved across the minor lineages of human P[6] strains and the porcine strain Gottfried are also shaded (gray). The line separates the asymptomatic strains (above) and symptomatic strains (below). Only substitutions distinct from the consensus (Gottfried) are shown.
|
In addition to the lack of consistent amino acid change patterns, neither phylogenetic studies nor antigenic characterization identified significant differences in the VP4 gene among symptomatic and asymptomatic P2A[6] rotavirus strains (15; see also Fig. 1). In contrast, the VP4 genes of the unique human P[6] strain, AU19, and the porcine Gottfried strain could be distinguished both phylogenetically and antigenically, and therefore they were classified as P2C[6] and P2B[6], respectively (20, 26). Because there is no absolute correlation between P genotypes and P serotypes (7), the lack of antigenic data on the novel genotype P[6] strains does not allow prediction of their serotype specificities. Serologic and molecular characterization of group A rotavirus isolates indicate that strains belonging to a particular P serotype usually share >89% amino acid identity in their full-length VP4 genes (10). However, exceptions to these data have been observed, e.g., for strains 69 M (P4[10]) and H-2 (P4[12]), which carry the same serotype specificity despite the relatively low amino acid sequence identity (85.3%; data not shown). Thus, although in this study we did not determine full-length VP4 genes, the finding that all three previously identified subtypes of P2[6] belong to separate genetic lineages warrants further studies to investigate the effect of amino acid substitutions on the antigenic properties of the novel P?[6] strains after their adaptation to cell culture.
In summary, in the present study we described two formerly undetected lineages of genotype P[6] in humans. Representatives of the globally distributed M37-like P[6] lineage were not identified among the few isolates characterized. Thus, we were not able to assess the relative epidemiological importance of these novel strains from our study. Nonetheless, our results suggest that the variability within P[6] specificity is larger than previously thought. Collection of comprehensive sequence data about genotype P[6] strains will be necessary to reveal the significance of these strains in Hungary and other parts of the world. It would also be useful to understand whether these two novel lineages include genuine human strains or whether they originated from animal strains by interspecies transmission. The presence of several amino acid changes in the Hungarian and the Japanese strains' VP8* gene compared to the sequence of the porcine P[6] prototype is compatible with the hypothesis that the P[6] VP4 gene has been introduced into the human pool of rotavirus genes on a number of occasions by multiple, independent reassortment events. Alternatively, a hypothesis could be that a single heterologous reassortment event (perhaps with porcine strains) occurred that was followed by different rates of diversification for each human lineage. According to this hypothesis different lineages may undergo various selective pressures and evolve less or more quickly, reaching a sort of evolutionary stasis when completely adapted to their host. Such a model has been formulated for influenza viruses, where avian strains, unlike mammalian strains, seem to possess low evolutionary rates, having reached an optimal adaptation in their primary host, the birds (31, 33).
Full-genome analysis by RNA-RNA hybridization or sequencing of selected genes may provide insights into the possible origin of the novel genotype P[6] rotaviruses.
Nucleotide sequence accession numbers. Strains and GenBank accession of their partial VP4 gene are the following: strain BP720/93, accession no. AJ621503; BP1198/98, AJ621504; BP1338/99, AJ621507; BP271/00, AJ621502; BP1227/02, AJ621505; and BP1231/02, AJ621506.
This study was supported by the Hungarian Research Fund (OTKA; T032933).
|
|
|---|
This article has been cited by other articles:
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Copyright © 2009 by the American Society for Microbiology. For an alternate route to Journals.ASM.org, visit: http://intl-journals.asm.org | More Info»