To facilitate the investigation of hepatitis B virus (HBV) sequence variation, we recently established a method for functional analysis of PCR-amplified full-length HBV genomes. This study aimed at estimating the number of mutations introduced during amplification of genomes from samples from patients with low levels of viremia and their influence on replication and antigen expression. Wild-type HBV DNA template molecules in concentrations like those present in samples from patients with very low levels of viremia were amplified, sequenced (30 kb total), and functionally tested. We found that Taq polymerase and a Taq-Pwo polymerase mixture introduced an average of 5.7 and 3.1 mutations per genome, respectively, corresponding to polymerase error rates of 12.1 x 10(-5) and 6.0 x 1(0-5). One of 8 genomes (12%) amplified with Taq polymerase, but 7 of 17 genomes amplified with Taq-Pwo polymerases (41%), remained replication competent. All replication-competent genomes expressed HBs and HBe antigens and had an average of only 0.9 mutations per genome. In contrast, replication-defective genomes had an average of 5.4 mutations, which frequently also disturbed viral antigen expression. From these data we conclude that many of the replication-competent HBV genomes from a clinical specimen will retain their replication and antigen expression phenotypes even after extensive amplification with Taq-Pwo polymerases. Because replication competence is highly sensitive to random mutations, it is the best marker for the identification of HBV genomes with few or no PCR-introduced mutations.