Abstract
We generated a large number of mutations in the hepatitis B virus (HBV) core gene inserted into a bacterial expression vector. The new mutagenesis procedure generated deletions and insertions (as sequence repeats) of various lengths at random positions between M1 and E145 but not substitutions. The R-rich 30-amino-acid C-terminal domain was not analyzed. A total of 50,000 colonies were tested with a polyclonal human serum for the expression of hepatitis B core or e antigen. A total of 110 mutants randomly chosen from 1,500 positive colonies were genotyped. Deletions and insertions were clustered in four regions: D2 to E14, corresponding to the N-terminal loop in a model for the core protein fold (B. Bottcher, S. A. Wynne, and R. A. Crowther, Nature 386:88-91, 1997); V27 to P50 (second loop); L60 to V86 (upper half of the alpha helix forming the N-terminal part of the spike and the tip of the spike); and V124 to L140 (C-terminal part of the C-terminal helix and downstream loop). Deletions or insertions in the remaining parts of the molecule forming the compact center of the fold seemed to destabilize the protein. Of the 110 mutations, 38 allowed capsid formation in Escherichia coli. They mapped exclusively to nonhelical regions of the proposed fold. The mutations form a basis for subsequent analysis of further functions of the HBV core protein in the viral life cycle.