Patients with primary biliary cirrhosis (PBC) have both serologic and tissue evidence of infection. A recently identified human betaretrovirus was originally cloned from the biliary epithelium cDNA library of a patient with PBC. By conducting a BLASTN search, the initial partial pol gene fragment was found to have 95% to 97% nucleotide homology with mouse mammary tumor virus (MMTV) and with retrovirus sequences derived from human breast cancer samples. Using an anti-p27(CA) MMTV antibody, viral proteins were detected in the perihepatic lymph nodes but not in liver tissue samples from patients with PBC, suggesting a higher viral burden in lymphoid tissue. Therefore, in the current study, we used lymph node DNA to clone the proviral genome of the human betaretrovirus from two patients with PBC using a polymerase chain reaction (PCR) walking methodology with conserved primers complementary to MMTV. The human betaretrovirus genome contains five potential open reading frames (ORF) for Gag, protease (Pro), polymerase (Pol), envelope (Env), and superantigen (Sag) proteins that are collinear with their counterparts in MMTV. Alignment studies performed with characterized MMTV and human breast cancer betaretrovirus amino acid sequences revealed a 93% to 99% identity with the p27 capsid proteins, a 93% to 97% identity with the betaretrovirus envelope proteins, and a 76% to 85% identity with the more variable superantigen proteins. Phylogenetic analysis of known betaretrovirus superantigen proteins showed that the human and murine sequences did not cluster as two distinct species. In conclusion, human betaretrovirus nucleic acid sequences have been cloned from patients with PBC. They share marked homology with MMTV and human breast cancer-derived retrovirus sequences.