Epstein-Barr virus (EBV)-genome-negative human lymphoma lines, Ramos and BJAB, can be converted by EBV in vitro into EBV-genome-positive virus nonproducer lines. These cell lines have been used to identify surface antigens unique to EBV, with the expectation that such determinants might be related to the antigenic target responsible for EBV-specific immunosurveillance. Antisera prepared in rabbits immunized with either whole cells or purified plasma membranes were used in immunoblot experiments to analyze antigenic differences resulting from expression of the resident EBV genome. Unexpectedly, an increase in polypeptides of 32 and 70 kilodaltons was consistently observed in the EBV-converted Ramos lines. In contrast, these antigens were not expressed in BJAB or in its EBV-converted lines. These data suggested that p32 and gp70 might be murine leukemia virus (MuLV)-coded antigens because Ramos, but not BJAB, had been passaged in athymic nude mice during establishment of this cell line. This conclusion was confirmed by using antisera specific for MuLV p30 and gp70. Retroviral antigens were expressed constitutively at low levels in Ramos. Quantitative immunoblotting showed that EBV conversion of Ramos amplified the expression of MuLV proteins 3- to 5-fold. The molecular mechanism responsible for the enhanced expression is unknown. The biological relevance of this phenomenon is also not clear, but the interaction between a DNA and a RNA tumor virus in a Burkitt lymphoma line that carries both viruses may have important biological consequences in relation to retrovirus latency and tumor induction. These results also show that caution must be used when ascribing "uniqueness" to EBV-determined antigens, particularly in the Ramos lines. This warning extends also to the use of Ramos cell lines as immunogens, because immunization of rabbits elicited antibodies that recognized proteins of the same size as the retroviral antigens.