Human T-cell lymphotropic virus type 1 (HTLV-1) is the causative agent of adult T-cell leukemia (40, 55) and a neurodegenerative disorder called tropical spastic paraparesis/HTLV-1-associated myelopathy (19, 37). HTLV-1 replication is dependent on the viral transactivator, Tax, and three 21-bp repeat elements, collectively referred to as the Tax response element (TxRE), localized in the U3 region of the proviral long terminal repeat (LTR). Each 21-bp repeat contains a core cyclic AMP (cAMP) response element (CRE) flanked by 5′ G-rich and 3′ C-rich sequences that confer Tax-dependent transactivation and dictate the assembly of Tax, CREB (CRE-binding protein), and the 21-bp repeat into a ternary (3°) complex (38, 46, 53, 56, 57). Tax interacts with CREB via the latter’s basic domain and amino acid residues in its immediate vicinity (2, 39, 54). A cellular homolog of CREB, ATF-1, can also interact with Tax in the form of a CREB/ATF-1 heterodimer or as an ATF-1 homodimer (11, 44, 57). Due to subtle amino acid sequence differences in the ATF-1 basic domain, Tax binds ATF-1 with reduced affinity (2, 57). Like CREB, Tax functions as a dimer (47). Analyses of Tax mutants have suggested that the NH2-terminal domain of Tax is important for CREB bZip (basic domain-leucine zipper) binding, while the midsection and the COOH terminus of Tax are important for subunit dimerization and possibly interaction with basal transcription factors, respectively (1, 23, 42, 47). Several groups have also reported that Tax enhances the dimerization and DNA binding of bZip proteins in vitro (3, 7, 12, 39, 50, 52). Transcriptional activation via the TxRE, CREB (and/or ATF-1), and Tax is independent of cellular signaling processes. Irrespective of the cell types used, cotransfection of a Tax expression plasmid and a reporter containing the TxRE results in 30- to 100-fold transcriptional induction. This contrasts with cellular somatostatin, proenkephalin, and phosphoenolpyruvate carboxyl kinase genes, whose expression is under CRE control and becomes induced only when the cAMP signaling pathway is activated. Work from many laboratories has indicated that an elevation in the intracellular levels of cAMP, as a consequence of cell surface signaling events, results in activation of protein kinase A, which can phosphorylate a sequence-specific DNA-binding protein, CREB, at amino acid residue Ser-133 (21, 22). Phospho-CREB subsequently recruits a transcriptional coactivator, CBP (CREB-binding protein) (16, 31, 35), or its homolog, p300 (4, 5), to CRE-containing enhancers to augment mRNA transcription (16, 31, 35). CBP and p300 appear to serve as integrators of numerous cellular signaling processes (26). The massive sizes of CBP and p300 allow them to interact with many cellular transcription factors in a signal-dependent and sometimes mutually exclusive fashion. To date, steroid and retinoid hormone receptors, CREB, c-Jun, c-Myb, Sap-1a, c-Fos, MyoD, p53, Stat-1/2, NF-κB, pp90rsk, TATA-binding protein, and TFIIB have been found to interact with CBP and p300 (see reference 26 for a review). This list will probably continue to grow. The oncoproteins of two DNA tumor viruses, adenovirus E1A and simian virus 40 large T antigen, also target and affect CBP/p300 functions (4, 5, 8). P/CAF (p300/CBP-associated factor), a histone acetyltransferase (HAT) that acetylates histones H3 and H4 but not H2A or H2B, has also been shown to bind to the CBP/p300 domain where E1A binds (51). Finally, in addition to their proposed roles as adaptor or coactivator molecules for transcription, CBP and p300 possess intrinsic HAT activity for all four core histones (9, 36). For this reason, they are suggested to play an important role in changing the nucleosomal structure at or around certain promoters (9, 36). The HAT activities of CBP/p300 and P/CAF presumably release promoter sequences from the repressive effect of the nucleosome and render them accessible to basal transcription factors and RNA polymerase II for transcriptional initiation. Using fluorescence anisotropy techniques, Kowk et al. have recently shown that HTLV-1 Tax directly binds CBP and p300 (30), results which have been confirmed as well as extended (20). These observations, combined with previous reports that Tax, CREB bZip, and TxRE specifically form a 3° complex, suggests that both CBP and p300 can be recruited to CREB/TxRE in a manner that bypasses signal-induced CREB phosphorylation by interacting with Tax. Thus, via the action of Tax, viral transcription can proceed constitutively without extracellular stimuli. The protein domain in CBP that is important for binding Tax has been localized to amino acid residues 451 to 682 (30). Interestingly, this domain has also been found to be important for binding phosphorylated CREB, c-Jun, c-Myb, and Sap-1a (6, 10, 17, 27, 30). In this study, we show by electrophoretic mobility shift assay (EMSA) that both the full-length p300 and residues 451 to 682 of CBP (CBP451-682) form stable quaternary (4°) complexes with Tax, CREB bZip, and the HTLV-1 21-bp repeat in vitro. The presence of CBP451-682 or p300 greatly facilitated the formation of a stable, high-order protein-DNA complex on the HTLV-1 21-bp repeat. The 4° complex readily forms on the HTLV-1 21-bp repeat but not on a CRE with nonspecific flanking sequences. While ATF-1 bZip could participate in 4° complex assembly, ATF-2 and ATF-3 failed to do so. Analyses of a series of Tax deletion mutants by glutathione S-transferase (GST) pull-down assays and EMSA, coupled with peptide competition, further revealed Tax amino acid residues 81 to 95 (81QRTSKTLKVLTPPIT95) to be important for CBP/p300 binding. This region lies between the domains for binding CREB bZip and for Tax subunit dimerization (1, 47). Three amino acid residues (88KVL90) in this region constitute sites that are rapidly cleaved by trypsin and chymotrypsin, suggesting that they are highly exposed and accessible to protein-protein interactions. We further report a series of point mutations targeted to this domain in Tax, which result in impairments in LTR transactivation and generate proteins that are defective in the ability to interact with CBP/p300. These data provide direct in vivo evidence in support of the role of CBP/p300 in Tax-mediated transactivation. Importantly, mutations that inactivate the COOH-terminal transactivation domain of Tax did not significantly alter the incorporation of CBP and p300 in vitro, suggesting that 4° complex formation alone might not be sufficient for transcriptional activation. Interactions with additional cellular factors via the COOH-terminal transactivation domain of Tax are likely to be important for efficient LTR-driven gene expression.