The Crystal Structure of CCG1/TAFII250-interacting Factor B (CIB)*

The general transcription initiation factor TFIID and its interactors play critical roles in regulating the transcription from both naked and chromatin DNA. We have isolated a novel TFIID interactor that we denoted as CCG1/TAFII250-interacting factor B (CIB). We show here that CIB activates transcription. To further understand the function of this protein, we determined its crystal structure at 2.2-Å resolution. The tertiary structure of CIB reveals an α/β-hydrolase fold that resembles structures in the prokaryotic α/β-hydrolase family proteins. It is not similar in structure or primary sequence to any eukaryotic transcription or chromatin factors that have been reported to date. CIB possesses a conserved catalytic triad that is found in other α/β-hydrolases, and our in vitro studies confirmed that it bears hydrolase activity. However, CIB differs from other α/β-hydrolases in that it lacks a binding site excursion, which facilitates the substrate selectivity of the other α/β-hydrolases. Further functional characterization of CIB based on its tertiary structure and through biochemical studies may provide novel insights into the mechanisms that regulate eukaryotic transcription.

Eukaryotic gene expression is mainly regulated at the level of transcription initiation from chromatin templates (1)(2)(3). This regulation involves the orchestrated interplay of chromatin factors, gene-specific DNA-binding factors, general transcription factors, and transcription machinery elements and their interactors. Eukaryotic transcription is essentially regulated in three distinct steps. The first step is the alteration of chromatin structure through the modification and remodeling of chromatin components. A major target of the structural changes in chromatin is the nucleosome, which is the fundamental repeating unit of chromatin that consists of histones and DNA in a precise stoichiometric balance. Histone modification enzymes, ATP-dependent remodeling factors, and histone chaperones are involved in regulating the histone-DNA and histone-histone interactions that lead to particular chromatin structures. Of these chromatin regulators, the tertiary structures of histone acetyltransferase (HAT), 1 histone deacetylase homologues, histone methyltransferases, and histone chaperones have been solved (4 -7). The second step of eukaryotic transcription regulation is the binding of regulatory transcription factors to promoter/enhancer elements. The tertiary structures of numerous types of DNA-binding domains that are present in these factors have been determined, and their DNA binding interactions have also been well characterized and classified. However, structures of most transcriptional activation/regulatory domains are not yet available. This may be because such unstable activation/regulatory domains are induced to fit to the basal transcription machinery (8). The third step is the regulation of RNA polymerization by RNA polymerase II and several general transcription initiation and elongation factors (9,10). One of these general transcription initiation factors is TFIID, which is known to interact with various kinds of transcription/chromatin factors and to activate transcription initiation. The tertiary structures of RNA polymerase II and general transcription initiation and elongation factors, including domains of TATA box-binding protein (TBP), TBP-associated factor (TAF), TFIIA, TFIIB, TFIIE, TFIIF, and S-II (TFIIS), have recently been determined (11). Nevertheless, the mechanisms by which the activated general transcription initiation and elongation factors regulate RNA polymerization are not yet fully understood.
TFIID is a multiprotein complex that is composed of TBP (12) and TAFs (13,14), and it plays a central role in both basal and regulatory transcription from the naked DNA template. TFIID binds to the TATA box of the core promoter and initiates transcription by forming the preinitiation complex with other general transcription initiation factors and RNA polymerase II (9,10). It also interacts with various kinds of regulatory transcription factors as well as cofactors and activates transcription in cooperation with these elements (15)(16)(17). Some studies suggest that TFIID also plays a key role in regulating transcription from a chromatin DNA template. In one of these studies, it was found that the binding of TFIID to the TATA box of the core promoter during nucleosome assembly potentiates the subsequent initiation by RNA polymerase II (18). Moreover, * This work was supported by a grant-in-aid for scientific research from the Ministry of Education, Science, Sports, and Culture of Japan, by the New Energy and Industrial Technology Development Organization, and by Exploratory Research for Advanced Technology of the Japan Science and Technology Corporation. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The crystallographic analyses have revealed that there are multiple histone-fold pairs in TFIID (19,20). In addition, a biochemical study has suggested that a subset of four yeast histone-fold TAFs forms a histone-fold octamer (21).
The largest subunit of TFIID (22,23) is CCG1/TAF II 250. CCG1 and TAF II 250 are acronyms. CCG1 stands for cell cycle arrest in G 1 , and TAF II 250 stands for TATA box-binding protein-associated factor with a molecular weight of 250 kDa. It has been reported that CCG1 has three enzymatic activities, namely, HAT activity (24), histone H1-specific ubiquitin-activating/conjugating activity (25), and protein kinase activity (26). CCG1 has two bromodomains that are also found in various chromatin factors and that bind to the acetylated aminoterminal tails of histones (27). These structural and functional observations indicate that the function of CCG1 is closely related to the alteration of chromatin structure. In addition, cell lines that express CCG1 bearing an amino acid substitution suffer cell cycle arrest that leads to apoptosis, which suggests that CCG1 plays a role in growth and apoptosis (28).
Considering the characteristics of TFIID described above, we speculated that the molecules that interact with TFIID are likely to be chromatin factors and/or transcriptional regulators/ cofactors. Indeed, when we isolated and characterized various TFIID interactors, we found that one of the TAF interactors is a MOZ-Ybf2/Sas3-Sas2-Tip60 (MYST)-type HAT denoted as Tip60 (Tat-interactive protein 60) (29 -32) and that one of the CCG1 interactors is a histone chaperone denoted as CIA (CCG1-interacting factor A) (33)(34)(35)(36)(37). These proteins are highly conserved among various species, and they regulate several nuclear events, including DNA replication, DNA repair, transcription, silencing, and apoptosis, which indicates that they play important roles in regulating chromatin.
When we screened for proteins that interact with the conserved HAT domain of CCG1 by using the yeast two-hybrid system, we also identified the novel factor CCG1-interacting factor B (CIB). The primary sequence of CIB does not show any similarities to those of any of the transcription/chromatin factors known to date. To understand the functional role of CIB, we have determined the crystal structure of CIB refined at 2.2-Å resolution. These data are reported here. The presented structure shows a fold that is similar to one found in prokaryotic ␣/␤-hydrolase family proteins, even though these proteins show very little sequence homology with CIB.

EXPERIMENTAL PROCEDURES
Yeast Two-hybrid Screen-We isolated the cDNA of CIB (70 -210 aa) during a yeast two-hybrid screen (38) that aimed to search for factors that interact with CCG1 (GenBank TM /EBI accession number Q96IU4). A full-length clone (1-210 aa) of CIB was isolated by plaque hybridization using a partial CIB clone as a probe. The coding sequence of its cDNA was also identified in the mouse expressed sequence tag data bases (accession number Q8VCR7). DNA encoding a highly conserved region (HCR) of the human CCG1 protein (447-1112 aa) was cloned in-frame with the Gal4 DNA-binding domain in pAS1-CYH2 using NdeI and BamHI sites to make the pAS1-CYH2-CCG1-HCR construct. A human peripheral lymphocyte cDNA library fused to the Gal4 activation domain in pACT at the XhoI site was co-transformed with the pAS1-CYH2-CCG1-HCR plasmid into Y190 yeast cells in which the His3 and ␤-galactosidase genes are integrated under the GAL4 element as the reporter. The yeast transformants were selected on minimal medium lacking tryptophan, leucine, and histidine and containing 25 mM 3-aminotriazole. The resultant colonies were tested for ␤-galactosidase activity to detect the specific interactions. The ␤-galactosidase assay was performed as described previously (39).
Direct Interaction between CIB and CCG1 Proteins-The CIB and CCG1 (447-788 or 777-1111 aa) proteins were expressed in Escherichia coli and purified to near homogeneity. The full-length CIB protein or buffer was immobilized with CNBr-activated Sepharose and used in the pull-down assay. Thus, 0.2 g of CCG1 (447-788 or 777-1111 aa) protein was mixed with either 1 g of CIB protein that had been immobilized with Sepharose or with Sepharose alone in the binding buffer (25 mM Hepes-KOH, pH 7.5, 2 mM MgCl 2 , 2 mM CaCl 2 , 10% glycerol, 150 mM KCl, 20 mM ZnCl 2 ) and incubated for 1 h at 4°C. This was followed by three washes with the binding buffer. The pellet was then subjected to SDS-PAGE and Western blotting using an anti-His tag antibody.
Genomic Southern Blot Hybridization-Hybridization was performed for 18 h at 65°C in 5ϫ saline/sodium phosphate/EDTA, 2% SDS, 10ϫ Denhardt's solution, followed by three washes for 10 min in 2ϫ SSC, 0.05% SDS at room temperature and two washes for 20 min in 0.1ϫ SSC, 0.1% SDS at 60°C. Autoradiography was performed for 48 h at Ϫ80°C. Human genomic DNA (Clontech, Inc.) was digested with EcoRI, BamHI, HindIII, or PstI. The digested DNA was then subjected to 0.7% agarose electrophoresis, transferred to a nylon membrane, and fixed with ultraviolet light. The DNA probe used was the same as that used in the Northern blot hybridization assay.
Detection of CIB mRNA Transcripts in Human Tissues-A multiple human tissue mRNA blot (Clontech, Inc.) was hybridized with DNA probes that had been radiolabeled by [ 32 P]dCTP using a random priming kit (Amersham Biosciences). The DNA probe was generated by XhoI digestion of CIB cDNA. Hybridization was performed for 18 h at 42°C in 50% formamide, 5ϫ saline/sodium phosphate/EDTA, 2% SDS, 10ϫ Denhardt's solution, followed by three washes for 10 min in 2ϫ SSC, 0.05% SDS at room temperature and two washes for 20 min in 0.1ϫ Cellular Localization of CIB-CIB was cloned in-frame with green fluorescent protein in pEGFP. pEGFP-CIB was transfected into COS cells by Lipofectin (Invitrogen) according to the manufacturer's instructions. The COS cells were cultured in Dulbecco's modified Eagle's medium supplemented with 5% fetal bovine serum at 37°C in a humidified atmosphere of 5% CO 2 and 95% air. Thirty-six h after transfection, the cells were examined under a fluorescent microscope (Zeiss) equipped with fluorescein isothiocyanate and ultraviolet filters.
Transcriptional Activation Assay-CIB (70 -210 aa) was cloned inframe with the Gal4 DNA-binding domain in pAS1-CYH2 using NdeI and XhoI sites to make the pAS1-CYH2-CIB construct. pAS1-CYH2-CIB was transformed in Y190 yeast cells in which the ␤-galactosidase gene is integrated under the GAL4 element as the reporter. The trans-formants were tested by the ␤-galactosidase assay to detect transcriptional activation activity.
Expression, Purification, and Crystallization of Human CIB-The expression, purification, and crystallization of human CIB have been described elsewhere (40,41).
Structure Determination-All diffraction data were collected at the beamline BL18B of the Photon Factory, Tsukuba, Japan. The structure was solved by the multiple isomorphous replacement method at 2.8-Å resolution and refined to 2.2 Å with data from a native crystal (Table I).
The native data were indexed and reduced with DENZO, and intensities were scaled using SCALEPACK (42). The derivative data were indexed and processed with DPS (43). Two uranium sites and one gold site were found and refined using SOLVE (44). The experimental map produced by SOLVE was improved by solvent flattening with SOLO-MON incorporated in CCP4 (45). After solvent flattening, the map was  reasonably good enough to trace the secondary structure elements using O (46). The map was further improved by combining experimental phases with calculated phases obtained from the initial fitting. After the first refinement round, the R-factor was 29.4%, and R free was 34.3%. The model was refined with CNS (47) using a maximum likelihood target including amplitude and phase probability distributions. Phasecombined A -weighted electron density maps were used throughout to guide further model building and to place water molecules. The current refined model at a 2.2-Å resolution consists of 208 aa, one sulfate molecule, and 191 water molecules with a final R-factor of 18.9% and an R free of 23.5%. The electron densities for one residue at each of the amino and carboxyl termini were absent. PROCHECK (48) showed Ser-111, in an unfavorable (, ) region that is unique in this protein family, combined with excellent stereochemistry.
Hydrolase Assay-p-Nitrophenyl butyrate was used as the substrate for the hydrolase reaction. Thus, 6.25-50 ng of CIB wild type and 0.5 mM p-nitrophenyl butyrate were mixed in phosphate-buffered saline and incubated at 20°C for 10 min. At the end of the reaction, the optical density at 400 nm was measured to detect p-nitrophenol, the product of the hydrolysis reaction.
Coordinates-Coordinates have been deposited in the Protein Data Bank (code 1IMJ).

RESULTS
Interaction of CIB with the CCG1-HAT Domain and Transcriptional Activation Activity of CIB-We used a yeast twohybrid system employing the HAT domain of CCG1 as bait to screen for proteins that interact with human CCG1 and identified two novel proteins named CIA (33)(34)(35)(36)(37) and CIB (40,41). To confirm that CIB interacts specifically and directly with CCG1-HAT, we performed two-hybrid and pull-down assays.
The two-hybrid assay shows that CIB does interact specifically with CCG1-HAT (Fig. 1A), whereas the pull-down assay (Fig.  1B) indicates that CIB protein binds directly and specifically to the 777-1111-aa domain of CCG1 but not to the 447-788-aa domain. To confirm that the CIB protein is encoded in the human genome, we performed genomic Southern blot hybridization using CIB cDNA as a probe. This indicated that CIB is indeed a human factor (Fig. 2A). Northern blot hybridization analysis also shows that CIB mRNA is expressed in almost all human tissues (Fig. 2B), which indicates that CIB is ubiquitously expressed. The CIB protein is found both in the nucleus and the cytosol (Fig. 3A), which suggests that it is involved in both nuclear and cytosolic reactions. As CIB interacts with the CCG1-HAT domain of the TFIID subunit, it is likely that CIB is involved in transcriptional regulation. Moreover, Fig. 3B indicates that CIB has transcriptional activation activity. All these results indicate that the CIB protein is indeed a general factor in humans that interacts with CCG1-HAT.
The Overall Structure of CIB-Because the primary structure of CIB differs from that of all other known eukaryotic transcription factors, we chose to resolve its tertiary structure in the hope that this might reveal its function. The structure of CIB protein was solved by multiple isomorphous replacement and refined to a crystallographic R-factor of 18.9% at 2.2-Å resolution (Figs. 4 and 5 and Table I). The tertiary structure has an ␣/␤-fold (Fig. 6). The ␤-sheet consists of seven parallel ␤-strands (␤1, ␤3, ␤4, ␤5, ␤6, ␤7, and ␤8) and one antiparallel amino-terminal ␤-strand (␤2), and it is flanked by helices ␣1 and ␣6 on one side and by helices ␣2, ␣3, ␣4, and ␣5 on the other side. The ␤-strands form a twisted ␤-sheet. A long loop comprising 69 -90 aa connects ␤-strand ␤4 and helix ␣2. The outer two strands, ␤1 and ␤8, are oriented nearly perpendicular to each other, which results in a left-handed superhelical twist in the ␤-sheet. The first two strands, ␤1 and ␤2, run antiparallel to each other and are connected by a hairpin turn. The remaining ␤-strands in the ␤-sheet run parallel to strand ␤1.
Structure-based Functional Analyses of CIB-To elucidate the function of CIB, its structure was compared with the protein structures in the Protein Data Bank using the program DALI (49). The search revealed good structural similarity with prokaryotic bromoperoxidase (Fig. 7A)  The amino and carboxyl termini of the protein are labeled, and the helices and strands are numbered. ␣1␣ and ␣1␤ are the former and the latter parts of the ␣1 helix, respectively. The catalytic triad Ser-111, His-188, and Asp-162 is shown (see Fig. 9). The sulfate arising from the crystallization process is denoted as SO 4 . The secondary structure assignment is based on the program PROCHECK (48). This figure was generated by Molscript (72). atoms of CIB with these proteins gave root mean square deviations of 2.2 Å for 179 aa of bromoperoxidase (Fig. 7B), 2.3 Å for 183 aa of 2-hydroxy-6-oxo-6-phenylhexa-2,4-dienoate hydrolase, 2.7 Å for 165 aa of carboxylesterase, and 2.4 Å for 177 aa of soluble epoxide hydrolase. Lipase (55), which also belongs to the ␣/␤-hydrolase superfamily, has a z-score of 11.7 and a root mean square deviation of 3.1 Å for 168 C␣-atoms with CIB.
Although CIB shows little sequence similarity with these proteins, the tertiary structure of CIB distinctly resembles the ␣/␤-domain of these proteins. Maximal sequence similarity (23%) occurred with the ␣/␤-domain of bromoperoxidase (1-128 aa and 214 -277 aa of bromoperoxidase) (Fig. 8). The sequence alignment of these proteins reveals that they share some conserved structural motifs. The sequence alignment results for human CIB indicate that it is more similar to prokaryotic hydrolases than to any known eukaryotic factors. However, although the CIB structure is similar to that of the ␣/␤-hydro-lases, CIB does not possess the additional ␣-helical domain part that is important in determining the substrate selectivity of the other ␣/␤-hydrolases. Hence, it is possible that factors that interact with CIB may act to facilitate the binding of the appropriate native substrate to CIB.
Active Site Region of CIB-In addition to the overall similarity in the folding topology found between the ␣/␤-hydrolase proteins and CIB, a "nucleophile elbow" motif, i.e. a strandnucleophile-helix motif, that is highly conserved in ␣/␤-hydrolases (56) was found in the 105-122-aa region of CIB (Fig. 9A). The , angles for the potential nucleophile Ser-111 of CIB fall into an unfavorable region in the Ramachandran plot, which is a characteristic of the nucleophile in a nucleophile elbow. The Gly-X-Ser-Y-Gly consensus sequence motif around the nucleophile residue that is common to many hydrolase enzymes (57) is absent in CIB. Instead, both glycine positions are replaced by Ser.
As with other serine esterases, CIB possesses a catalytic triad that is formed by Ser-111, Asp-162, and His-188. The nucleophile of a catalytic triad in an ␣/␤-hydrolase is either a serine or a cysteine, depending on its hydrolase function. The nucleophile in CIB, Ser-111, is positioned in a sharp turn that connects the strand (␤5) and the helix (␣3). Regarding the other two residues of the catalytic triad of CIB, His-188 and Asp-162 (Fig. 9A), His-188 is located on a loop between strand ␤8 and helix ␣6. This is perfectly situated to shuttle protons between the nucleophile Ser-111 and the general base Asp-162, which is located in a loop between ␤7 and ␣5. The distance between NE His-188 and OG Ser-111 and the distance between ND1 His-188 and OD2 Asp-162 are 2.85 and 2.75 Å, respectively.
To assess the catalytic triad of CIB, we compared it to that of the serine protease (Fig. 9B). The serine-type proteases that belong to the families S9, S10, S15, S28, and S33 also fall into the clan of ␣/␤-hydrolases (ESTHER data base, http://bioweb. ensam.inra.fr/ESTHER/). When we compared the CIB structure with these proteases, the CIB structure was closely related to the ␣/␤-domain of prolyl aminopeptidase from Serratia marcescens with a z-score of 11.0. Aminopeptidase catalyzes the removal of amino-terminal proline residue from peptides. This protease has been found mainly in bacteria and in some plants. The tertiary structure of this protein contains two domains: ␣/␤-hydrolase fold domain and a small ␣-helical domain that covers the substrate-binding region from solvent. The superposition of ␣/␤-domain of this protein with the CIB structure gave the root mean square deviation value of 2.7 Å for 163 residues (Fig. 9B). The catalytic residues Ser-113, Asp-268, and His-296 of aminopeptidase are positioned in the same way as the catalytic residues in CIB (Fig. 9B). Although the CIB does not share its sequence with those of ␣/␤-hydrolases, its surprising and intriguing structural similarity, combined with the preservation of the arrangement of the catalytic triads, suggests that they possibly evolved from a common ancestor.
The active site is located in a deep cleft running across the carboxyl ends of the central ␤-sheet (Fig. 9A). A long flexible loop comprising 69 -90 aa that runs across one side of the cleft contains mainly hydrophobic residues. This inverted "V" shapelike loop may move away to make room for a substrate that is being bound. At one side of the cleft, the hydrophobic residues Ile-41, Leu-112, Ile-84, Gly-85, and Leu-87 line up, probably so that they can form hydrophobic interactions with a substrate moiety. Arg-42 and Lys-141 lie at the edge of the cleft region and may potentially interact electrostatically with the substrate moiety. Because electrostatic surface potential analysis shows that the active site region is predominantly composed of a hydrophobic pocket (Fig. 10), the candidate substrate is likely to be hydrophobic. The general protein surface of CIB is also mainly hydrophobic (Fig. 10). The total number of negatively charged aa (Asp and Glu) and positively charged aa (Arg and Lys) is 17 and 13, respectively. Such low numbers of charged aa in CIB indicate that the surface is rather hydrophobic. The hydrophobicity of CIB is high when compared with bromoperoxidase (51), 2-hydroxy-6-oxo-6-phenylhexa-2,4-dienoate hydrolase (52), soluble epoxide hydrolase (53), and carboxylesterase (54) (Fig. 10). This appreciable difference of CIB relative to the other hydrolases may indicate a unique feature of CIB and suggests that this hydrophobic region might be responsible for protein-protein interactions in transcriptional regulation.
Biochemical Activities of CIB-The comparative studies of the primary and tertiary structures of CIB suggest it may have hydrolase activity. To test this, we used p-nitrophenyl butyrate, a typical substrate for hydrolases, and purified the recombinant CIB protein to near 95% homogeneity (Fig. 11A). CIB induced the hydrolysis of p-nitrophenyl butyrate into p-nitrophenol in a dose-dependent manner (Fig. 11B). Thus, as predicted, CIB possesses hydrolase activity. The specific activity of CIB was about 50 units/mg, which is comparable with E. coli esterase (58). To our knowledge, CIB is the first eukaryotic transcriptional factor/cofactor that has been found to be a hydrolase that bears a bromoperoxidase-related ␣/␤-hydrolase fold.

DISCUSSION
Significant Features of the Structure of CIB-The structure of CIB differs from the structures of all other eukaryotic transcription factors. Instead, it bears a structure that resembles the ␣/␤-hydrolase fold found in several prokaryotic enzymes. Interestingly, unlike most other hydrolases, CIB lacks a binding site excursion in its ␣/␤-hydrolase fold structure. The binding site excursion, which acts as a lid, is believed to support the substrate binding activity of a hydrolase as it is positioned above the substrate-binding site (54). How CIB recognizes its substrate is thus unclear. One possibility is that a binding partner complements the lack of a binding site excursion in CIB. This notion is supported by the presence of a long flexible loop near the active site region of CIB (Fig. 5). When the binding partner of CIB binds, it may induce the loop to undergo a conformational change that then facilitates substrate binding. We speculate that CCG1, more specifically, its HAT domain, could be a candidate binding partner. Supporting this notion is that CIB and CCG1-HAT may share the same substrate because they interact with each other. That substrate could be acetyl-CoA because when we incubated CIB with acetyl-CoA and subjected the mix to time-of-flight mass spectrometry analysis, we detected a peak that corresponds to an acetyl-CoA-CIB complex (data not shown). Thus, CCG1-HAT may play an important role in the substrate binding of CIB (discussed further in the next section).
Prediction of the CIB Substrate That Is Involved in Transcriptional Regulation-The identification of the native substrate of CIB and analysis of complex crystal structures of CIB with its substrate and/or with CCG1 will most likely reveal a novel transcription-regulatory mechanism, as CIB is a nuclear hydrolase enzyme and is the first such enzyme shown to be involved in transcription. As suggested above, acetyl-CoA may be the substrate that is bound by CIB. This is because the likely binding partner of CIB is CCG1, the largest subunit of TFIID, which consists of many functional domains including a HAT domain (517-976 aa) (24 -27). As it is known that HAT activity (i.e. GCN5) can transfer an acetyl group from acetyl-CoA to the amino terminus of histone in vitro (59), it may be that the acetylation of histones is regulated by CIB. That CIB may bind to acetyl-CoA is supported by the tertiary structure of the active site region of CIB, which suggests that this cleft may provide sufficient space to bind acetyl-CoA. A pull-down assay also indicates that CIB selectively binds to the 777-1111-aa domain of CCG1, which encompasses the acetyl-CoA-binding site of CCG1-HAT. This suggests that CIB participates in the acetyltransferase reaction of CCG1. However, additional biochemical studies have failed to demonstrate that CIB has acetyltransferase, deacetylase, or other acetyl-CoA-related activities (data not shown). Therefore, it will be necessary to search for the native substrate for the hydrolase activity of CIB or its regulators before the biological role of this protein can be fully understood.
Functions and Biological Roles of CIB-The acetylation reactions of CCG1 may regulate transcription as it has been established that the acetylation and deacetylation of highly conserved lysine residues in the amino-terminal tails of histones correlate with transcriptional activation and inactivation, respectively (1-3). Moreover, it is known that the acetylation of DNA-binding proteins regulates their activities (60,61), and CCG1 has been shown to acetylate histone and TFIIE (24,62). In addition, the HAT domain of CCG1 maps to the central and most conserved portion of CCG1. We showed in this report that CIB interacts with CCG1-HAT directly (Fig. 1B). That CIB binds to this conserved domain suggests that it is involved in regulating histone acetylation and chromatin organization along with CCG1. Supporting this notion is that a single amino acid mutation of CCG1 in hamster cell lines results in their arrest in G 1 phase of the cell cycle followed by apoptosis. As CIB is the binding partner of CCG1, this observation suggests that CCG1 and CIB both play a role in cell proliferation and apoptosis (28).
It is likely that further investigation of the relationship between TFIID and CIB will help to reveal the full extent of the biological activities of CIB. That these two molecules interact at a physiological level is suggested by the fact that TFIID and CIB both exist in the nucleus and that they interact directly with each other in vitro (Figs. 1B and 3A). This is supported by the fact that CIB is expressed in almost all tissues as shown by our Northern blot hybridization analysis (Fig. 2B). TFIID appears to participate in a variety of fundamental biological phenomena. First, mutations of TFIID subunits cause cell cycle arrest (28). Second, a variant of TFIID plays a specific role in embryonic development (63,64). In addition, cells bearing TFIID subunit mutants undergo drastic morphological changes due to a defect in cytokinesis (65). As CIB directly interacts with TFIID and exists in all major tissues, the analysis of the relationship between CIB and TFIID will shed a light to understand the biological relevance of CIB.
The Relationship between CIB and Other CCG1 Interactors-We previously isolated CIA as a protein that interacts with the CCG1 bromodomain (35). This domain is also found among chromatin factors. CIA binds to the carboxyl terminus of FIG. 10. The CIB protein has a highly hydrophobic surface. The electrostatic surface potentials of CIB and several ␣/␤-hydrolases (Protein Data Bank codes 1BRT, 1C4X, and 1AUO) were calculated using GRASP (75). The left panel shows the front view of the image that has been rotated 180°with respect to the right panel. All four structures are drawn in the same orientation. The putative active site of CIB is indicated in the left panel. histone H3 and probably contributes to transcription through its histone chaperone activity (33). In addition to transcription, CIA functions in DNA replication, DNA repair, cell cycle, and apoptosis (34,66). CIA has a genetic link with and interacts physically with MYST-HAT and blocks the acetylation activity of its complex (67). CIA also interacts with the SWI/SNF chromatin-remodeling complex (68). Acetylation involves the hydrolysis of acetyl-CoA, whereas chromatin remodeling requires the hydrolysis of ATP. Thus, through its putative acetyl-CoAand ATP-hydrolyzing activity, CIB may cooperate with or inhibit CIA and CIA interactors, thus contributing to the regulation of chromatin organization.
Another example of a CCG1 interactor is the retinoblastoma tumor suppressor protein Rb, which regulates the cell cycle, tumor formation, cell differentiation, and senescence (69). Rb interacts directly with CCG1 through multiple domains and inhibits the intrinsic kinase activity of CCG1 (70). Rb is regulated by its phosphorylation and dephosphorylation, which involves the hydrolysis of a phosphoryl bond. It would be of interest to investigate whether CIB could regulate the phosphorylation state of Rb through its hydrolase activity.
Finally, as CIB directly interacts with CCG1, which is involved in a wide range of biological phenomena, including gene expression, chromatin organization, and cell cycle, through its biochemical activities, which include HAT, kinase, and ubiquitinase activities, further analyses of the function of CIB in relation to the activities of CCG1 and CCG1 interactors are likely to yield novel insights into the physiological roles that are played by each of these elements.