The structural part of the hepatitis C virus (HCV) genome encodes a capsid protein, C and two envelope glycoproteins, E1 and E2, released from the virus polyprotein precursor by signalase(s) cleavage(s). The processing of E1 was investigated by infecting simian cells with recombinant vaccinia viruses expressing parts of the HCV structural proteins. When the predicted E1 sequence was expressed alone (amino acid residues 174-370 of the polyprotein) or with the capsid protein gene (residues 1-370). it showed an apparent molecular mass of 35 kDa as measured by SDS-PAGE analysis. However, when E1 was expressed as part of a truncated C-E1-truncated E2 polypeptide (residues 132-383), the processed E1 product had the expected apparent molecular mass of 31 kDa, suggesting that flanking sequences are necessary for the generation of the mature 31 kDa El form. The N-terminal sequence of the two E1 forms was found to be the same. Analysis of the glycosylation pattern showed that, in both species, only four of the five potential N-linked glycosylation sites were recognized, indicating that glycosylation was not involved in the molecular mass difference. We showed that expression of E1 with or without the hydrophobic stretch of amino acids residues 371-383, defined as the E2 signal sequence, may be responsible for the difference in electrophoretic mobility of the two E1 species. In vitro translation assays and site-directed mutagenesis experiments suggest that this sequence remains part of the 31 kDa E1 mature protein.