The NF-#cB transcription factor complex is sequestered in the cytoplasm by the inhibitory protein I#cB-a (MAD-3). Various cellular stimuli relieve this inhibition by mechanisms largely unknown, leading to NF-KB nuclear localization and transactivation of its target genes. It is demonstrated here with human T lymphocytes and monocytes that different stimuli, including tumor necrosis factor a and phorbol 12-myristate 13-acetate, cause rapid degradation ofIOcB-a, with concomitant activation of NF-cB, followed by a dramatic increase in IKB-a mRNA and protein synthesis. Transfection studies reveal that the IKB-a mRNA and the encoded protein are potently induced by NF-cB and by homodimers of p65 and of c-Rel. We propose a model in which NF-.cB and IKB-a mutually regulate each other in a cycle: saturating amounts of the inhibitory IKB-a protein are destroyed upon stimulation, allowing rapid activation of NF-KB. Subsequently, IcB-a mRNA and protein levels are qulickly induced by the activated NF-cB. This resurgence of IKB-a protein acts to restore an equilibrium in which NF-KB is again inhibited.NF-KB is a dimeric transcription factor that binds and regulates gene expression through decameric cis-acting KB DNA motifs (reviewed in refs. 1 and 2). Although a p50/p65 heterodimer has traditionally been referred to as NF-KB and remains the prototypical and most abundant form, it has been recognized recently that several distinct but closely related homo-and heterodimeric factors are responsible for KB site-dependent DNA binding activity and regulation. The various dimeric factors are composed of members of the family of Rel-related polypeptides. One subclass of this family, distinguished by its proteolytic processing from precursor forms and lack of recognized activation domains, includes p50 (NFKB1) (3-6) and pSOB (NFKB2, pS2) (7-10), whereas the second subclass contains recognized activation domains and includes p65 (RelA) (11-13), RelB (14,15),18) Activation of the NF-KB transcription factor and various related forms can be initiated by a variety of agents, including tumor necrosis factor a (TNF-a) and phorbol 12-myristate 13-acetate (PMA) (1, 2). Activation proceeds through a post-translational event in which preformed cytoplasmic NF-KB is released from a cytoplasmic inhibitory protein, (20)(21)(22)(23). IKB-a inhibits transactivation of the p50/p65 heterodimer, by binding to the p65 component, blocking the dimer's translocation to the nucleus (20,21,23). IKB-a also inhibits complexes containing c-Rel or RelB (24,25). IKB-a blocks binding in vitro of various NF-KB dimers to KB binding sites in DNA (11,12,15,22,26,27). Because the latter effect requires nuclear IKB-a, its relevance, in vivo, is unknown. Although lKB-a is generally a cytoplasmic protein (21, 23), it and its chicken homolog (pp4O) have also been detected in the nucleus (refs. 28-30 and K.B., G.F., and U.S., unpublished results). In addition to the wellcharacterized and cloned IKB-a and its chicken and rat homologs (24, 31), another biochemically d...

Bcl-3 is an I kappa B-related protein with ankyrin repeat motifs. Its gene is located at a site of recurrent translocations in a subset of B cell chronic lymphocytic leukemias. Bcl-3 associates tightly with p50B (NFKB2, p52) homodimers in cells, and together these proteins form a ternary complex with DNA at kappa B sites. Such an association functionally leads to a novel and potent form of transactivation through the kappa B motif: the tethering of Bcl-3 to DNA via the p50B homodimers allows Bcl-3 to transactivate directly, while p50B homodimers alone cannot. Transactivation mediated by Bcl-3 requires two cooperating domains located amino- and carboxy-terminal to the ankyrin domain. Bcl-3 is localized to the nucleus, and a Bcl-3-p50B complex is detected in certain lymphoid cells. Our data reveal a novel role for Bcl-3, distinct from that of the inhibitor I kappa B. The results have implications for tumorigenesis.

Acetylation and other modifications on histones comprise histone codes that govern transcriptional regulatory processes in chromatin. Yet little is known how different histone codes are translated and put into action. Using fluorescence resonance energy transfer, we show that bromodomain-containing proteins recognize different patterns of acetylated histones in intact nuclei of living cells. The bromodomain protein Brd2 selectively interacted with acetylated lysine 12 on histone H4, whereas TAF(II)250 and PCAF recognized H3 and other acetylated histones, indicating fine specificity of histone recognition by different bromodomains. This hierarchy of interactions was also seen in direct peptide binding assays. Interaction with acetylated histone was essential for Brd2 to amplify transcription. Moreover association of Brd2, but not other bromodomain proteins, with acetylated chromatin persisted on chromosomes during mitosis. Thus the recognition of histone acetylation code by bromodomains is selective, is involved in transcription, and potentially conveys transcriptional memory across cell divisions.