We have screened a human DNA library using as probe a chimpanzee sequence that contains homology to the polymerase gene ofthe endogenous baboon virus. One set ofoverlapping clones spans about 20 kilobases and contains regions of DNA sequence homology to the gag p30, gag p15, and polymerase genes of Moloney murine leukemia virus. Furthermore, the spacings are the same as in Moloney virus between these sequences and a 480-nucleotide region that has the structural characteristics of a 3' copy of the long terminal repeat sequence. Hybridization of the cloned DNA to restriction digests of human DNA indicates that the human genome contains only two copies closely related to the sequence and 10 less closely related copies. This retroviral sequence appears to have been in its present chromosomal location prior to the divergence ofman and chimpanzee because the human and chimpanzee clones have 3-4 kilobases of identical 3' flanking sequence.Retroviruses have been shown to be capable ofinducing tumors by two basic mechanisms. The first involves the incorporation of a cellular transforming gene, or oncogene, into a retrovirus so that it is expressed whenever the virus infects cells (1). The number of oncogenes in a mammalian genome is uncertain but is clearly more than one (2). The second mechanism, demonstrated by the induction of bursal lymphomas in chickens by avian leukosis virus, involves the ability of the retroviruses to integrate a DNA copy ofthe viral genome into the chromosomal DNA of the host. In the rare case in which the viral DNA is integrated adjacent to an oncogene, the promoter sequences present in the termini of the integrated virus can activate the oncogene (3).As a result of the ability of the retroviruses to integrate into host DNA and, thus, become part of the host genome, many, if not most, mammals have accumulated tens or hundreds of integrated retroviral sequences. These endogenous viral sequences thus provide a reservoir of viral genes which might be activated by developmental or environmental factors including mutagens. These newly activated viruses then might induce tumors either by the mechanisms described above or through other mechanisms, such as inactivation of a critical gene by integration within the gene (4). Evidence that the activation of endogenous viral genes may be significant is provided by the observation that formation of a recombinant virus derived from two endogenous viruses is correlated with spontaneous leukemias in the AKR mouse (5), a strain with a high incidence of leukemia.The critical questions concerning retroviruses are whether they are involved in human carcinogenesis and, if so, to what degree. Attempts to answer these questions have been hampered by a dearth of human retroviruses. There is currently one human retrovirus (6). This virus is not an endogenous human retrovirus because related sequences are not found in normal human DNA (7). There is also a potential endogenous human retroviral sequence that was cloned from human DNA (8). To date this clone has b...