Several observations suggest that bacteria induce autoimmunity in primary biliary cirrhosis (PBC). Since no PBC-specific bacterial species could be identified, it can be speculated that the triggers are non-species-specific bacterial proteins. This hypothesis would imply that several or even all bacterial species can trigger PBC. Therefore, we investigated whether PBC exhibits immune reactions to non-species-specific bacterial antigens. Yersinia enterocolitica O3 was screened for the presence of proteins that were labeled by immunoblotting using PBC sera. We focused our investigations on a 160-kDa protein, which was further enriched and characterized by partial N-terminal amino acid sequencing. The prevalence of antibodies to this protein was determined by immunoblotting in a variety of diseases. The 160-kDa protein was identified as the beta-subunit of bacterial RNA-polymerase, a highly conserved bacterial protein with a very high degree of sequence identity among all bacterial species. Antibodies to the beta-subunit of bacterial RNA polymerase were specific for this protein. Until now no mammalian protein could be found that cross-reacts with these antibodies. The prevalence of antibodies to the beta-subunit of bacterial RNA polymerase (ARPA) using the protein from Yersinia enterocolitica O3 (serum dilution 1:1000) was: healthy controls (HC, N = 101) 7.9%, primary biliary cirrhosis (PBC, N = 61) 32.8%, autoimmune hepatitis type 1 (AIH, N = 46) 26.1%, alcoholic liver cirrhosis (ALC, N = 44) 9.1%, Crohn's disease (CD, N = 38) 7.9%, ulcerative colitis (UC, N = 24) 8.3%, primary sclerosing cholangitis + UC (PSC/UC, N = 11) 0%, acute yersiniosis (Yers, N = 36) 19.4%, acute infection with Campylobacter jejuni (Camp, N = 10) 0%, acute Q-fever (QF, N = 16) 6.25%, chronic hepatitis C (HCV, N = 39) 7.7%, c-ANCA-positive vasculitis (Vasc, N = 40) 15%, systemic lupus erythematosus (SLE, N = 28) 10.7%, and malaria tropica (MT, N = 24) 16.7%. There was no significant difference between PBC and AIH. The group of autoimmune liver diseases (PBC + AIH, N = 107, 29.9%) differed highly significantly from HC, chronic inflammatory bowel diseases (CD + UC + PSC/UC, N = 73, 6.8%), ALC, and HCV and also differed significantly (P = 0.01) from the group with bacterial and parasitic diseases (Yers + Camp + QF + MT, N = 86,13.95%) and from the group with Vasc + SLE (N = 68,13.2%). Testing of ARPA using the protein from E. coli yielded nearly identical results. In conclusion, an increased prevalence of antibodies to the beta-subunit of bacterial RNA polymerase, a highly conserved non-species-specific bacterial protein, can be found in primary biliary cirrhosis, but also in autoimmune hepatitis type I. These findings do not add an argument for a bacterial trigger of PBC. Rather, they suggest that ARPA belong to the pool of natural antibodies that are up-regulated in autoimmune liver diseases.