Human bocavirus (HBoV) was discovered 16 years ago by Allander et.al. using a new molecular method for screening respiratory samples collected from children who had respiratory tract infection (RTI) of unknown etiology [1
]. Additional serological and quantitative PCR research have provided compelling evidence that the newly discovered virus (later classified as HBoV1) acts as an etiologic agent for respiratory tract infections [2
]. In the following years, three other genotypes (HBoV2-4) were discovered [4
], primarily from the gastrointestinal tract; therefore, it seems to be involved in pathogenesis of gastroenteritis. Although HBoV2-4 has also been detected in samples from the respiratory tract but with a much lower frequency, their role in the pathogenesis of respiratory infections is unclear.
HBoVs are small viruses up to 26 nm in diameter, nonenveloped, DNA viruses that belong to the Parvoviridae
genus. Members of two species within Bocaparvovirus
genus infect humans: HBoV1 and HBoV3 belonging to the species Primate bocaparvovirus 1
, and HBoV 2 and HBoV 4 belonging to the species Primate bocaparvovirus 2
]. HBoV possess a linear, single-stranded DNA genome of 5543 nucleotides (nt) long with non-identical terminal hairpins of 140 and 122 nt that play a key role in virus replication [7
]. Recent advances in molecular biology research of HBoV1 revealed that during its replication in the polarized/nondividing airway epithelial cells HboV1 expresses six nonstructural proteins: NP1, NS1, NS1-70, NS2, NS3, and NS4, depending on splicing mRNAs within ORF 1 [9
]. The mRNA spliced at the D2-A2 sites results in a shift of the NS1 ORF at the C-terminus and encodes NS1. Unspliced mRNA that reads the D2-A2 intron encodes the NS1-70 protein, and alternative splicing within the NS1-coding region generates mRNAs that encode NS2, NS3 and NS4 [9
]. Moreover, HBoV1 also expresses viral non-coding RNA (BocaSR) and three structural proteins VP1, VP2, and VP3. The BocaSR is the first identified RNA polymerase III (Pol III) transcribed viral non-coding RNA in small DNA viruses [9
HBoVs are considered to be highly diverse pathogens characterized by a rapid evolution [5
]; therefore, worldwide surveillance of HBoVs’ genetic evolution is necessary. Although both mutation and recombination are responsible for bocavirus evolution, in a HBoV genome prone to rapid evolution with low level of polymorphisms, recombination seems to play a dominant role [5
In this study, we focused on an investigation of the prevalence, phylogenetics, and evolution of HBoVs from pediatric patients with RTI, which will help to reveal the molecular epidemiology and phylogeny of the circulating HBoV in Croatia.
This study was performed on 957 Croatian children hospitalized with RTI during period of four years revealed prevalence of HBoV of 7.6%. Furthermore, HBoV was the fifth most frequently detected virus in respiratory tract samples after RV, RSV, AdV and PIV. Previously published retrospective study from Croatia revealed a high detection rate of HBoV among infants and small children with LRTI that required hospitalization (i.e., 23.1% of those with proven viral etiology) [25
]. The higher frequency detection of HBoV in comparison to this study may be explained by different designs of the studies, specifically the subject’s age. Namely, the mentioned study included small children up to three years, while this study was performed on patients up to 18 years of age. High HBoV detection in small children is also corroborated by the results of this study; 86.2% of HBoV positive children were younger than three years of age. Moreover, the most recent meta-analysis indicated that being <5 years old is a risk factor for HBoV infection [26
]. The same meta-analysis that included 35 studies involving 32,656 subjects from 16 European countries showed that HBoV prevalence varied from 2.0% to 45.69% with a pooled estimate rate of 9.57% [26
]. Another review that included 311 studies from 50 countries all over the world performed between 2005 and 2016, showed the average prevalence of HBoV in respiratory tract samples ranged from 1.0% to 56.8%, depending on the country, with the worldwide HBoV total prevalence estimates of 6.3% [27
], which is consistent with our results. Furthermore, the same study reported the rate of co-infections in subjects with respiratory infections, and HBoV-positivity ranged from 8.3% to 100%, with total co-infection estimates in the 193 studies covered of 52.4% [27
]. In our study, HBoV was co-detected with another respiratory virus in 82.2% cases. High co-detection is a well-recognized characteristic of bocavirus infection, which is probably result of prolonged shedding of HBoV1 to the nasopharynx, including for weeks and months [28
]. Furthermore, prolonged shedding of HBoV complicates the diagnosis of acute HBoV infection, thus for accurate diagnosis quantitative PCR, serology, or HBoV1, mRNA detection is recommended diagnostic approaches [29
]. This also affects the seasonal distribution studies since the actual HBoV1 infection may have occurred months before the current sampling during a later RTI episode.
Most HBoV cases in this study were detected during winter months, from November to February. The northwest part of Croatia, where the study was conducted, has a temperate climate region. Observed seasonal distribution is similar to the prevalence previously reported for temperate regions, where HBoV1 infection mostly occurs in winter and spring, but different from the HBoV1 epidemics in in subtropical regions [31
Sequencing and phylogenetic analysis shows all samples detected during this study belong to HBoV1, which is in agreement with other studies showing that HBoV1 is primarily a respiratory pathogen, while other HBoVs are prevalent in gastrointestinal infections [32
]. Studies have shown that HBoV2-4 can be detected in respiratory samples at rates of 0.4–4.3%; albeit, their role in respiratory illness remains inconclusive [35
]. Nevertheless, we have considered this in the sequence assembly step: all samples best assembled to HBoV-1 reference. There was no phylogenetic grouping based on year of isolation.
Out of 29 sequenced samples, amplification of nearly complete genome in one reaction was successful for five samples, unrelated to whether HBoV was detected as single infecting virus or infection with multiple respiratory viruses was detected. It was difficult to obtain long PCR products because of complexity and quality of clinical specimens (nasopharyngeal and pharyngeal flocked swabs combined) although our primers targeted conserved regions at positions which allowed us to divide the genome in three parts. This was most evident in four samples (HR1223-20, HR142-17, HR387-18, HR421-18), which had much lower coverage. BLAST search of these samples revealed that these reads contain primarily human DNA. Nevertheless, with the approach to divide the genome in three smaller segments, we were able to sequence 24 samples and additional five samples contained two out of three genome segments.
Analysis of nucleotide differences show limited heterogeneity between samples and only small differences were observed between different genes. In agreement with previous studies [37
], genes encoding non-structural proteins are more conserved than VP proteins, which bind to the surface cell receptors and are responsible for transporting the genome to the nucleus [27
]. Moreover, VP3 protein (previously called VP2) is the major antigenic determinant [39
]; thus, it is expected to be more variable given that mutation and recombination are major drivers of viral evolution [11
Of the five substitutions in NS1 protein found in our samples, two are located in the middle helicase domain [9
] but fall outside of four conserved Walker motifs which execute 3′–5′ helicase function [9
]. Three substitutions are located in the C-terminal part of NS1 protein, which is predicted to have transcription transactivation capability, but has not been studied [9
VP1 unique region (VP1u) contains phospholipase A2 (PLA2) domain between aa residues 11–66 [42
]. Within our samples, a substitution at amino acid residue 17 between Arg and Lys was observed. Most of our samples have arginine at this position, while two samples (HR199-17 and HR1289-20) have lysine. Phospholipase activity of this domain is required for endosomal escape [43
], but since both Lys and Arg have positively charged side chains, we do not believe it could have an impact on these processes.
Among our samples, we also observed a change at residue 461 within variable region VR-VIIIB, which is reported to be important for contact with Fab [44
]. The impact of this substitution was not examined. However, VR-VIIIB was previously identified as a potential target for the development of a peptide vaccine that would be broadly neutralizing against multiple HBoV strains [44
]. Nevertheless, further research that includes available HBoV1 cell culture systems are required to find whether the substitutions and variable sites detected in the sequenced strains had an effect on virus production, infectivity or reactivity with neutralizing anti-HBoV1 antibodies.
Calculated nucleotide based evolutionary rates are consistent with previous studies all showing that HBoV evolves at rates of about 10−4
substitutions per site and year [37
]. Slowest evolving was gene encoding NS1-70 protein, which was also the most conserved based on differences analyzed at the nucleotide level. The NP1 protein had a faster rate of evolution, although amino acid analysis shows this protein is completely conserved among our samples, indicating that synonymous substitutions play main role in the evolution of this gene. A study by Lu et al. also shows a slightly higher substitution rate of NP1
Intra-genotypes recombination appears to play a significant role in the evolution of bocaparvovirus [12
]. Using comparison of genome organization and phylogenetic analysis, it was shown that HBoV3 NS1
sequences cluster with the homologous sequences of the HBoV1 strain, but conversely, the VP1/VP2
sequences of HBoV3 are similar to HBoV2, providing evidence that HBoV3 may have resulted from recombination between the HBoV1 and HBoV2 viruses [27
]. A study from Thailand, using full-length sequence analysis, revealed that an unusual strain of HBoV4 was the result of recombination between HBoV2 and HBoV4 strains [11
]. Furthermore, whole genome sequencing of Novosibirsk HBoV isolates detected an isolate that emerged via recombination between HBoV3 and HBoV4 [45
]. Although HBoVs are considered to be diverse and frequently recombinant pathogens, especially HBoV2-4 that are primarily replicate in the gastrointestinal tract, recombination was not detected in the samples investigated in this study.
In conclusion, the prevalence of HBoV found in this study are consistent with previously published data confirming HBoV1 as the dominant human bocavirus that causes respiratory infections. Nevertheless, further molecular studies are needed to continuously monitor the evolution of human bocaviruses.