Human Bocavirus in Children with Acute Gastroenteritis, Chile, 1985–2010

We detected human bocavirus in 89 (19.3%) of 462 fecal samples collected during 3 periods from 1985 through 2010 from children <5 years of age in Chile who were hospitalized with acute gastroenteritis. Our findings confirm the long-term circulation of human bocavirus in Chile.

Phylogenetic relationships between isolates from Chile and GenBank reference strains were studied by using MEGA5 software (6). We inferred the evolutionary history using the neighbor-joining method. Bootstrap (1,000 replications) was used to assess the reliability of individual nodes in each phylogenetic tree. Evolutionary distances were computed by using the Kimura 2-parameter method (6). Variance analysis of the bocavirus detection frequency was conducted by using the Kruskal-Wallis test (α = 0.05) using Statdisk 12.0.1 software (Marc Triola and Pearson Education, Inc., New York, NY, USA). The Ethics Committee of the Faculty of Medicine, University of Chile approved the study.
Twenty-two (of 65 HBoV1) nonstructural (NS) 1 partial coding sequences were obtained, and consistent with previous reports, phylogenetic analysis showed that HBoV1 constitutes a genetically homogeneous entity (4,7). The nucleotide divergence average for the HBoV1 Chile isolates was 0.7% (range 0%-1.7%). In the phylogenetic tree analysis, 15 of 22 isolates clustered with prototype strain st1 (GenBank accession no. DQ000495), and the remaining 5 clustered with prototype strain st2 (GenBank accession no. DQ000496) (Figure,  There was no temporal clustering of the HBoV1 isolates. Similarly, the phylogenetic analysis of 391 nt from the nucleocapsid 1 region of 5 HBoV3 isolates from Chile revealed an average divergence of 1.2% (range 0%-1.6%), and the phylogenetic tree analysis showed that the isolates grouped into 2 clusters with no evident temporal clustering ( Figure, panel B).
In contrast, HBoV2 was a genetically heterogeneous group. Analysis of the nucleotide sequence of the NS1 partial region of 16 of 18 isolates yielded an average of 3.0% nt divergence (range 0%-6.0%). CH23-85, the most divergent isolate, showed 5.0%−6.0% nt divergence with other HBoV2 isolates from Chile. Phylogenetic tree analysis demonstrated that this isolate was closely related to the Pakistan strain PK5510, with which it formed a separate cluster (98% nt identity). This same analysis demonstrated 2 additional clusters among HBoV2 isolates from Chile whose intragroup average nucleotide identity reached 99.6% and 99.5%, respectively. Kapoor  represented by prototype strain PK5510 (GenBank accession no. FJ170278); genotype 2, with prototype strain PK2255 (GenBank accession no. FJ170279); and genotype 3, with prototype strain UK648 (GenBank accession no. FJ170280) (4).
We found no other isolates of genotype 1 during the remaining study time (Figure, Table). This prompted speculation that HBoV2 infection is a dynamic phenomenon that can manifest with the emergence of different variants of the agent at different times. This possibility also is rooted in the observation that HBoV2 is highly prone to recombination, and as a consequence presents a high degree of diversity. This diversity is proposed to have given rise to HBoV1, a species with completely different biologic characteristics (7). However, because of the lack of continuous chronologic follow-up of the phenomena found in this study, we cannot ensure such assumptions, and a larger longitudinal study is required to confirm this hypothesis.

Conclusions
Although HBoV1 was originally detected in respiratory secretions of patients with respiratory infection, numerous studies have demonstrated its presence in 1.5%-19% of fecal samples (1)(2)(3)5,8,9). However, after primary respiratory infection this agent can persist with asymptomatic shedding for several months (10,11). Thus, HBoV1 in fecal samples could be due mainly to passive transfer from the respiratory tract (8,12).
Unlike HBoV1, HBoV2-4 have enteric tropism, and their role in gastroenteritis remains unclear (9). HBoV2 has been detected in feces from children with gastroenteritis in a broad range of percentages (1%-21%) alone or in co-infection with other enteropathogens. In contrast to our study, it has been frequently reported as the main HBoV species detected in feces (4,5,9,13). Moreover, in accordance with previous reports that found HBoV3 in low percentages (0%-2%), we detected HBoV3 only in periods A and B (2.3% and 1.5%, respectively) ( Table) and did not detect HBoV4 (5,7,9,14). Although samples were maintained at −80°C until tested, we cannot exclude the possibility that molecular detection of HBoV may be reduced by long-term storage. This fact, along with the co-infection exclusion, may explain the low detection rates of the strains with enteric tropism.
Although the frequency of detection of HBoV species varied among the periods studied, variance analysis indicated no significant differences (p = 0.099) (Table). We cannot confirm circulation or variation in detection in periods during which surveillance was not conducted.
This study confirms long-term circulation of HBoV in Chile and demonstrates the heterogeneity of HBoV2. These findings justify prospective studies to better understand the role of these viruses in childhood gastroenteritis.