Recombinant strains of hepatitis E virus (HEV) with insertions of human genomic fragments or HEV sequence duplications in the sequence encoding the polyproline region (PPR) were previously described in chronically infected patients. Such genomic rearrangements confer a replicative advantage in vitro but little is known about their frequency, location, or origin. As the sequences of only a few virus genomes are available, we analyzed the complete genomes of 114 HEV genotype 3 strains from immunocompromised (n = 85) and immunocompetent (n = 29) patients using the single molecular real-time sequencing method to determine the frequency, location, and origin of inserted genomic fragments, plus the proportions of variants with genomic rearrangements in each virus quasispecies. We also examined the amino acid compositions and post-translational modifications conferred by these rearrangements by comparing them to sequences without human gene insertions or HEV gene duplications. We found genomic rearrangements in 7/114 (6.1%) complete genome sequences (4 HEV-3f, 1 HEV-3e, 1 HEV-3 h, and 1 HEV-3chi-new), all from immunocompromised patients, and 3/7 were found at the acute phase of infection. Six of the seven patients harbored virus-host recombinant variants, including one patient with two different recombinant variants. We also detected three recombinant variants with genome duplications of the PPR or PPR + X domains in a single patient. All the genomic rearrangements (seven human fragment insertions of varying origins and three HEV genome duplications) occurred in the PPR. The sequences with genomic rearrangements had specific characteristics: increased net load ( p < 0.001) and more ubiquitination ( p < 0.001), phosphorylation ( p < 0.001), and acetylation ( p < 0.001) sites. The human fragment insertions and HEV genome duplications had slightly different characteristics. We believe this is the first description of HEV strains with genomic rearrangements in patients at the acute phase of infection; perhaps these strains are directly transmitted. Clearly, genomic rearrangements produce a greater net load with duplications and insertions having different features. Further studies are needed to clarify the mechanisms by which such modifications influence HEV replication.

The enterically transmitted hepatitis E virus (HEV) adopts a unique strategy to exit cells by cloaking its capsid (encoded by the viral ORF2 gene) and circulating in the blood as "quasi-enveloped" particles. However, recent evidence suggests that the majority of the ORF2 protein present in the patient serum and supernatants of HEV-infected cell culture exists in a free form and is not associated with virus particles. The origin and biological functions of this secreted form of ORF2 (ORF2) are unknown. Here we show that production of ORF2 results from translation initiated at the previously presumed AUG start codon for the capsid protein, whereas translation of the actual capsid protein (ORF2) is initiated at a previously unrecognized internal AUG codon (15 codons downstream of the first AUG). The addition of 15 amino acids to the N terminus of the capsid protein creates a signal sequence that drives ORF2 secretion via the secretory pathway. Unlike ORF2, ORF2 is glycosylated and exists as a dimer. Nonetheless, ORF2 exhibits substantial antigenic overlap with the capsid, but the epitopes predicted to bind the putative cell receptor are lost. Consistent with this, ORF2 does not block HEV cell entry but inhibits antibody-mediated neutralization. These results reveal a previously unrecognized aspect in HEV biology and shed new light on the immune evasion mechanisms and pathogenesis of this virus.