The NF-kappa B and Sp1 motifs of the human immunodeficiency virus type 1 long terminal repeat function as novel thyroid hormone response elements.

We report that thyroid hormone (T3) receptor (T3R) can activate the human immunodeficiency virus type 1 (HIV-1) long terminal repeat (LTR). Purified chick T3R-alpha 1 (cT3R-alpha 1) binds as monomers and homodimers to a region in the LTR (nucleotides -104 to -75 [-104/-75]) which contains two tandem NF-kappa B binding sites and to a region (-80/-45) which contains three Sp1 binding sites. In contrast, human retinoic acid receptor alpha (RAR-alpha) and mouse retinoid X receptor beta (RXR-beta) do not bind to these elements. However, RXR-beta binds to these elements as heterodimers with cT3R-alpha 1 and to a lesser extent with RAR-alpha. Gel mobility shift assays also revealed that purified NF-kappa B p50/65 or p50/50 can bind to one but not both NF-kappa B sites simultaneously. Although the binding sites for p50/65, p50/50, and T3R, or Sp1 and T3R, overlap, their binding is mutually exclusive, and with the inclusion of RXR-beta, the major complex is the RXR-beta-cT3R-alpha 1 heterodimer. The NF-kappa B region of the LTR and the NF-kappa B elements from the kappa light chain enhancer both function as T3 response elements (TREs) when linked to a heterologous promoter. The TREs in the HIV-1 NF-kappa B sites appear to be organized as a direct repeat with an 8- or 10-bp gap between the half-sites. Mutations within the NF-kappa B motifs which eliminate binding of cT3R-alpha 1 also abolish stimulation by T3, indicating that cT3R-alpha 1 binding to the Sp1 region does not independently mediate activation by T3. The Sp1 region, however, is converted to a functionally strong TRE by the viral tat factor. These studies indicate that the HIV-1 LTR contains both tat-dependent and tat-independent TREs and reveal the potential for T3R to modulate other genes containing NF-kappa B- and Sp1-like elements. Furthermore, they indicate the importance of other transcription factors in determining whether certain T3R DNA binding sequences can function as an active TRE.