- Email: [email protected]
The Caenorhabditis elegans Protein CTBP-1 Defines a New Group of THAP Domain-Containing CtBP Corepressors
J. Mol. Biol. (2008) 375, 1–11
doi:10.1016/j.jmb.2007.10.041
Available online at www.sciencedirect.com
COMMUNICATION
The Caenorhabditis elegans Protein CTBP-1 Defines a New Group of THAP Domain-Containing CtBP Corepressors Hannah R. Nicholas⁎, Jason A. Lowry, Tina Wu and Merlin Crossley School of Molecular and Microbial Biosciences, University of Sydney, Sydney, New South Wales 2006, Australia Received 2 August 2007; received in revised form 9 October 2007; accepted 16 October 2007 Available online 22 October 2007
The C-terminal binding proteins (CtBPs) play roles in diverse cellular processes including transcriptional regulation, Golgi membrane fission, and synaptic ribbon formation. In the context of transcriptional regulation, they function as corepressors, interacting with promoter-bound transcription factors and recruiting a large protein complex that contains chromatinmodifying enzymes. We recently described the structure of a Thanatosassociated protein (THAP) domain that is found in a new member of the CtBP family, the Caenorhabditis elegans CTBP-1 protein. We have identified additional THAP domain-containing CtBPs in the nematode, echinoderm, and cephalochordate lineages. The distribution of these lineages within the animal kingdom suggests that the ancestral form of the animal CtBPs may have contained a THAP domain that was subsequently lost in the vertebrate and arthropod lineages. We also provide functional data indicating that CTBP-1 represses gene expression and homodimerizes and interacts with PXDLS-containing partner proteins, three key features of the previously characterized animal CtBPs. CTBP-1 is therefore the founding member of a new subgroup within the CtBP corepressor family, the THAP domaincontaining CtBPs. © 2007 Elsevier Ltd. All rights reserved.
Edited by J. Karn
Keywords: CtBP; THAP domain; gene regulation; corepressor; Caenorhabditis elegans
Members of the C-terminal binding protein (CtBP) family are multifunctional proteins that have been ascribed roles in transcriptional corepression, Golgi maintenance, and synaptic ribbon formation.1 As transcriptional corepressors, CtBPs associate with DNA-bound transcription factors, many of which contain a short sequence with the consensus PXDLS.2 CtBPs function as part of a large corepressor complex that contains chromatin-modifying enzymes including histone deacetylases (HDACs), methyltransferases, and a demethylase.3,4 CtBPs have been reported to interact with many transcription factors,
*Corresponding author. E-mail address: [email protected]. Abbreviations used: CtBP, C-terminal binding protein; THAP, Thanatos-associated protein; HDAC, histone deacetylase; AN, ANGUSTIFOLIA; EST, expressed sequence tag; DBD, DNA-binding domain.
suggesting that CtBP-mediated gene repression is a widely utilized mechanism. The CtBPs share primary sequence and structural similarity with the D-isomer-specific 2-hydroxyacid dehydrogenases, which are composed of two functional domains: a substrate-binding domain and a nucleotide-binding domain.5,6 In addition to an analogous N-terminal substrate-binding domain and central nucleotide-binding domain, the CtBPs possess an intrinsically unstructured C-terminal region.6,7 The PXDLS-binding cleft through which CtBP contacts most of its partner proteins resides primarily in the substrate-binding domain. The central nucleotide-binding domain contains the nucleotide-binding consensus G/AXGXXGX17D (where X is any amino acid) and mediates interaction with NAD(H) as well as dimerization.6 Also conserved within the nucleotide-binding domain of the CtBPs is a histidine residue that is required for the catalytic activity of the D-isomer-specific 2-hydroxyacid dehydrogenase enzymes.5,8 Although CtBPs
0022-2836/$ - see front matter © 2007 Elsevier Ltd. All rights reserved.
2 have been shown to reduce pyruvate to lactic acid, this enzymatic activity has not been shown to be required for the repression function of the CtBPs.9–13 CtBP-related genes have been identified in numerous animal species. Invertebrates including Drosophila melanogaster, Anopheles gambiae, Apis mellifera, and Tribolium castaneum contain a single CtBP gene.14–16 In contrast, the studied vertebrates including human, mouse, rat, chicken, quail, zebrafish, and Xenopus contain two CtBPs (named CtBP1 and CtBP2). CtBP-like proteins have also been identified in a number of plant species. The best characterized of these is the Arabidopsis thaliana protein called ANGUSTIFOLIA (AN).17–19 Despite its similarity to the animal CtBPs, AN appears to be unable to associate with the PXDLS motif.20 Furthermore, the nucleotide-binding motif and dehydrogenase catalytic histidine are not conserved in AN. 17,18,20 Consistent with this lack of conservation of key functional elements, phylogenetic analysis indicates that the plant members of the CtBP family form a unique subfamily, suggesting that they may serve functions that are specific to plants.18 We recently described the structure of a protein domain, termed Thanatos-associated protein (THAP) domain, contained within a new member of the CtBP family, the CTBP-1 protein from the nematode Caenorhabditis elegans.21 Here, we present evidence that CTBP-1 defines a new subgroup of THAP–CtBP corepressors that extends across animal lineages. Our results suggest that the THAP–CtBPs represent the ancestral form of CtBP, with THAP domains having subsequently been lost in vertebrate and various other lineages. The C. elegans ctbp-1 gene encodes a CtBP Homologue with an N-Terminal THAP domain Early analysis of the genome of C. elegans predicted an open reading frame named F49E10.5 encoding a
CTBP-1, a THAP Domain-Containing CtBP Corepressor
612-amino-acid CtBP homologue (T34290), which has been cited in sequence comparisons of members of the CtBP family.11,18 Subsequently, expressed sequence tag (EST) data prompted the merging of F49E10.5 with its neighbor, F49E10.6. We have named this revised predicted gene ctbp-1.21 To verify this prediction, we sequenced products arising from PCR amplification of cDNA derived from RNA extracted from a mixed-stage population of C. elegans hermaphrodites. Five clones were sequenced, two of which corresponded to the predicted cDNA (NM_076582). The remaining clones prompt two amendments to the predicted transcript structure (Fig. 1a). First, there is an alternative 3′ splice site within the 3rd intron, which, when used, results in the inclusion of an additional 9 nucleotides in exon 4. Second, within the predicted 13th exon, there is an intron of 48 nucleotides, retained in some transcripts, which necessitates the definition of a 14th exon. We have not investigated the functional significance of these alternative splice isoforms. Conceptual translation of the ctbp-1 cDNA (NM_076582) yields a protein of 727 amino acids in which a search of the National Center for Biotechnology Information Conserved Domain Database identified two domains: amino acids 1–89 comprise a THAP domain, which we have described previously,21 while amino acids 178–505 show similarity to the D-isomer-specific 2-hydroxyacid dehydrogenases (Fig. 1b). Amino acids 90–177, linking the THAP domain to the dehydrogenase domain, and the C-terminal residues 506–727 do not contain any detectable structural motifs. Comparison of the dehydrogenase domain of CTBP-1 (residues 178–505) with that from human CtBP1 (residues 27–353) reveals 55% identity and 74% similarity of the amino acid sequences. Furthermore, several features of potential functional importance are conserved in the C. elegans protein (Fig. 1c).
Fig. 1. The C. elegans ctbp-1 gene encodes a protein that contains a THAP domain and is homologous to animal CtBPs. (a) Structure of the ctbp-1 transcript. Exons are depicted as boxes and introns are depicted as lines. Amendments to the predicted gene structure are shown below the schematic. Splice donor and acceptor sites are shown in bold, intronic sequences are shown in lower case, and exonic sequences are shown in upper case. Underlined in sequence (i) is an alternative 3′ splice site in the 3rd intron. Sequence (ii) shows the 13th intron, which is retained in some transcripts. (b) Schematic representations of the human CtBP1 and C. elegans CTBP-1 proteins showing domains identified in the National Center for Biotechnology Information Conserved Domain Database. Dark gray shading indicates the D-isomerspecific 2-hydroxyacid dehydrogenase domain. Light gray shading indicates the THAP domain. (c) ClustalW alignment of representative vertebrate (hCtBP1, NP_001319; hCtBP2, AAH52276) and insect (dCtBP, NP_001014617) CtBPs with C. elegans CtBP (CTBP-1, NP_508983). Symbols below the alignment denote the degree of conservation observed in each column. Asterisks indicate identical residues, “:” indicates conservative substitutions, and “.” indicates semiconservative substitutions. Features of potential functional importance are indicated above the alignment. “^” indicates residues lining the PXDLS-binding cleft, a line indicates the nucleotide-binding motif GXGXXGX17D, and “‡” indicates the His residue that is critical for the catalytic activity of the D-isomer-specific 2-hydroxyacid dehydrogenases. Dark gray shading indicates the D-isomer-specific 2-hydroxyacid dehydrogenase domain. Light gray shading indicates the THAP domain. Methods: Reverse transcriptase DNA amplification reactions were performed to obtain the ctbp-1 cDNA as follows. Total RNA was extracted from a mixed-stage population of C. elegans N2 (wild type) hermaphrodites using TRI Reagent® (Sigma-Aldrich) as described by the manufacturer. First-strand cDNA was then synthesized using the SuperScript™ II First-Strand Synthesis System for RT-PCR (Invitrogen) and oligo(dT)20 primers. The ctbp-1 cDNA was amplified by PCR (5BamcTC, CGGGATCCATGCCGACGACTTGTGGATTTCC; 3CeCtBPcDNA, AGTTGTCACAACCTCGAGAAC), and the product was then used as a template for a nested PCR (5BamcTC; 3EcoflC, CGGAATTCTTATGTGGCCAATGGTTGCTC). The product was cloned into the BamHI-EcoRI pBSSK (Stratagene) vector, and five clones were sequenced. One clone that corresponded to the predicted ctbp-1 cDNA sequence (NM_076582) was used for all subsequent clonings and is referred to as CTBP-1.
CTBP-1, a THAP Domain-Containing CtBP Corepressor
Fig. 1 (legend on previous page)
3
4 First, all the residues that line the PXDLS-binding cleft are present in the CTBP-1 sequence (Leu180, Ala182, Met193, Thr201, Ala203, Phe204, Cys205, His214, and Val217). Second, the CTBP-1 protein contains the nucleotide-binding consensus GXGXXGX17D (residues 332–355). Third, the catalytic histidine that is essential for dehydrogenase activity is conserved in the CTBP-1 protein (His443). This high degree of sequence homology argues for the inclusion of CTBP-1 in the CtBP family of proteins. As noted, unlike previously described members of this protein family, CTBP-1 contains an N-terminal THAP domain. Defined by a zinc-coordinating consensus sequence Cys-X2–4-Cys-X35–53-Cys-X2His, our structural data have shown that the THAP domain belongs to the zinc finger superfamily and contains a treble clef motif as its structural core.21 Although sequence searches have
CTBP-1, a THAP Domain-Containing CtBP Corepressor
identified over 100 animal proteins that contain one or more THAP domains,22 the vast majority of these have not been characterized. Those that have been studied appear to be involved in transcriptional regulation; for example, human THAP1 acts as a transcription factor for cell-cycle target genes23 and human THAP7 acts as a transcriptional repressor.24,25 The THAP domains of these proteins appear to be integral to their transcriptional activities, with the THAP domain in THAP1 serving as a DNAbinding domain (DBD)22 and the corresponding domain in THAP7 recruiting HDAC3.25 We have shown that, like the THAP domain of the human protein THAP1, the C. elegans CtBP THAP domain can bind DNA.21 Our structural data revealed the presence of a large, positively charged face on the C. elegans CtBP THAP domain, which may constitute the DNA contact surface.21
Fig. 2. The CtBPs from diverse organisms contain highly conserved THAP domains. (a) Phylogenetic tree based on the tree of life [tolweb.org/tree/phylogeny.html] depicting representative species in which a CtBP homologue has been either previously identified or identified in the current study. “X” indicates branches in which the CtBPs do not contain a THAP domain. Species in which a THAP domain-containing CtBP was identified are shaded gray. “?” denotes branches in which ESTs or genomic contigs show the presence of at least part of the THAP domain although the complete genomic sequence was unavailable. (b) Sequence alignment of five THAP domains from CtBPs: C. elegans (NP_508983), C. briggsae (CAE68813), C. remanei (GenBank whole genome shotgun sequence AAGD01000137), S. purpuratus (XP_001178649), and B. floridae (GenBank EST: BW870057 and Genome v1.0 scaffold_232). Conserved residues are indicated by asterisks. Seven charged residues that cluster on one face of the C. elegans CtBP THAP domain are shaded in gray. Methods: Available CtBP sequences were collected from GenBank and European Molecular Biology Laboratory databases. The C. elegans and S. purpuratus protein sequences were used to search (tblastn) the EST and genomes within GenBank and the following specialized genome databases: C. intestinalis, B. floridae, and D. pulex [genome.jgi-psf.org/euk_home.html].
CTBP-1, a THAP Domain-Containing CtBP Corepressor
5
Fig. 3. CTBP-1 represses gene transcription. (a) A Western blot was performed to examine expression of Gal4DBD fused to mCtBP2, CTBP-1, and CTBP-1(90–727) in transiently transfected COS-1 cells. (b) Gal4DBD–CtBP fusion constructs were tested for their ability to repress LexA–VP16-activated expression of a luciferase reporter in COS-1 cells (n = 4, ±SD). Methods: The full-length ctbp-1 cDNA was used as a template for the PCR amplification of a truncated cDNA lacking the first 267 nucleotides (5BamcnonTC, CGGGATCCGAAAGTAAGAACAGTGATCG; 3EcoflC). The PCR product was cloned into the BamHI-EcoRI pBSSK (Stratagene) vector and sequenced. The truncated cDNA is referred to as CTBP-1(90–727). For expression of proteins fused to the Gal4DBD in mammalian cells, the pcDNA3-Gal4 plasmid was made by ligating a HindIII-BamHI Gal4DBD fragment that had been excised from pGBT9(new) (derived from pGBT9 from Clontech) into the HindIII-BamHI sites of the pcDNA3 (Invitrogen) vector. CTBP-1 and CTBP-1(90–727) were excised from pBSSK-CTBP-1 and pBSSK-CTBP-1(90–727), respectively, and cloned into BamHI-EcoRI pcDNA3-Gal4 to create pcDNA3-Gal4–CTBP-1 and pcDNA3-Gal4–CTBP-1(90–727), respectively. The pcDNA3-Gal4–mCtBP2 plasmid has been described previously.11 The firefly luciferase reporter vector pGL2-(Gal4)5-(LexA)2-E1B-Luc and LexA–VP16 mammalian expression plasmid pCMV-LexA(1–202)–VP16(410–490) were gifts from Luke Gadreau and Mark Ptashne (The Sloan–Kettering Institute, New York, NY, USA). COS-1 cells were cultured as described previously28 and transfected with FuGENE6 (Roche Diagnostics) transfection reagent according to the manufacturer's instructions. Western blot assays were performed to assess the relative expression levels of the Gal4 fusion proteins used in the mammalian repression assays. Nuclear extracts were made from 10-cm dishes of COS-1 transfected with 4 μg of the relevant expression plasmid as has been described previously.28 The proteins were detected on Western blots using an anti-Gal4DBD antibody. Relative expression of fusion proteins was quantified using ImageJ image-processing software.29 To examine CtBP repression of reporter gene expression, we transfected six-well plates of COS-1 cells. The following plasmids were used: 3 μg of the pGL2-(Gal4)5-(LexA)2-E1B-Luc reporter, 1 μg of the pCMV-LexA(1–202)–VP16(410–490) expression vector, and either 1 μg of pcDNA3-Gal4 or 0.25, 0.5, or 1 μg of pcDNA3-Gal4–mCtBP2, pcDNA3-Gal4–CTBP-1, or pcDNA3-Gal4– CTBP-1(90–727). Renilla luciferase vector pRL-Luc (10 ng; Promega) was cotransfected to enable adjustment of firefly luciferase measurements to control for transfection efficiency. Luciferase activity was measured 48 h posttransfection in a Turner Designs model TD-20/20 Luminometer with the dual-luciferase reporter assay system (Promega), enabling calculation of the firefly:Renilla luciferase ratio (FF/R luciferase ratio).
CTBP-1, a THAP Domain-Containing CtBP Corepressor
6 Additional CtBPs from diverse lineages, nematodes, echinoderms and cephalochordates contain a THAP domain Given that the C. elegans protein CTBP-1 is the only member of the CtBP family identified to date that contains a THAP domain, we asked whether THAP domain-containing CtBPs might exist in other organisms. We used BLAST searches of sequenced genomes and EST databases to identify any such proteins. Within the nematode phylum, we have identified THAP domain-containing CtBPs in two additional Caenorhabditis species, C. remanei and C. briggsae, and have found genomic sequences indicating the pre-
sence of a similar THAP domain in the more distant nematode Haemonchus contortus (Fig. 2a, Supplementary Fig. 1). We have also found evidence for THAP domain-containing CtBPs beyond the nematodes. Specifically, sequence data support the existence of four THAP domain-containing CtBPs among the deuterostomes: three in the echinoderm lineage and one in the cephalochordate lineage. In the echinoderm lineage, genomic sequences supported by ESTs show that the CtBP of the sea urchin Strongylocentrotus purpuratus contains a THAP domain. Also within this lineage, ESTs indicate the presence of part of a THAP domain in the CtBP of both the common urchin Paracentrotus lividus and the sea star Asterina pectinifera. As complete genomic
Fig. 4 (legend on next page)
CTBP-1, a THAP Domain-Containing CtBP Corepressor
sequence data are unavailable for these latter two species, we have been unable to confirm that the entire THAP domain is present. In the cephalochordate lineage, genomic sequences supported by ESTs show that the lancelet Branchiostoma floridae also has a THAP domain-containing CtBP. The relative positions of these species on the tree of life are depicted in Fig. 2a. Two other surveyed species in which CtBPs were identified that do not contain a THAP domain are also shown: a urochordate, the sea squirt Ciona intestinalis (JGI genome v2.0 scaffold chr_05q), and a crustacean, the water flea Daphnia pulex (dpulex_jgi0609051 scaffold_3). Although CtBP homologues from at least 17 species had previously been described, these species represent only two animal clades: vertebrates and insects. Our sequence searches have identified 10 additional homologues from species representing diverse lineages within the animal kingdom. Importantly, among these new CtBP homologues, we have identified THAP domain-containing CtBPs in three groups: nematodes, echinoderms, and cephalochordates. The positions of these animals in the tree of life suggest that the THAP–CtBP fusion might in fact represent the ancestral form of the animal CtBP. Our identification of a crustacean CtBP that does not contain a THAP domain, together with previous reports of insect CtBPs that similarly do not contain this domain,14–16 suggests that the THAP domain has been lost in the arthropod phylum. Similarly, in the chordates, our data are consistent with loss of
7 the THAP domain in both the urochordate and the vertebrate lineages. Given that the THAP–CtBP fusion is widespread within the animal kingdom and that it may represent the ancestral form of the animal CtBP, we conclude that the THAP domaincontaining CtBPs represent an important new group within the CtBP family of proteins. The THAP domains of the CtBPs are highly conserved The amino acid sequences of the THAP domains from the identified THAP domain-containing CtBPs were compared (Fig. 2b). In addition to the zinccoordinating consensus (Cys-X2–4-Cys-X35–53-CysX2-His), 4 additional residues are highly conserved among all THAP domains and form part of the hydrophobic core of the domain (Pro29, Trp39, Phe61, and Pro72, numbered according to the C. elegans CTBP-1 THAP domain).21 Both elements are present in all the CtBP-associated THAP domains that we have identified, suggesting that they are true THAP domains. Notably, the similarity between these CtBP-associated THAP domains extends beyond these 8 consensus residues. Indeed, among the five identified CtBP THAP domains, 60 of the 89 residues (67%) are identical in all sequences (Fig. 2b). This identity suggests that the CtBP THAP domains are likely to serve similar functions. Supportive of this notion, the residues that constitute the positively charged surface patch on the THAP
Fig. 4. CTBP-1 homodimerizes and interacts with a PXDLS-containing partner protein through the PXDLS-binding cleft. This partner protein can repress transcription in a CtBP-dependent manner. (a–c) Yeast two-hybrid assays were performed to examine the interactions of CTBP-1 with partner proteins. The assays were performed by fusing the test proteins to the C-terminus of either the Gal4DBD or the Gal4 activation domain (Gal4AD). Growth on -His-Leu-Trp plates (pictured) indicates that the two proteins interact. Homodimerization of wild-type CTBP-1 was assessed (a), as was the interaction of CTBP-1 with a putative PXDLS-containing transcription factor from C. elegans, PAG-3 (b). The PXDLS-like motif in PAG-3 was mutated and the mutant protein (PAG-3 ΔNL) was tested for its capacity to interact with CTBP-1 to examine the PXDLS dependence of the interaction between CTBP-1 and PAG-3 (b). The dependence of the interaction between CTBP-1 and PAG-3 on the PXDLS-binding cleft of CTBP-1 was assessed by mutating one residue within the putative PXDLS-binding cleft and testing the capacity of the mutant protein (CTBP-1 A203E) to interact with PAG-3 (c). To confirm the integrity of the mutant protein CTBP-1 A203E, we also tested the capacity of this protein to dimerize with CTBP-1 (a). (d) Gal4DBD–PAG-3 fusion constructs were tested for their ability to repress LexA–VP16-activated expression of a luciferase reporter in COS-1 cells (n = 4, ±SD). Methods: The CTBP-1 insert was subcloned into the BamHI-EcoRI sites of the pGAD10 (Clontech) vector and pGBT9(new) vector to allow expression in the yeast two-hybrid system as both Gal4AD and Gal4DBD fusions. The A203E mutation was introduced into CTBP-1 by overlap PCR mutagenesis (5CeCtBPA203E, GTTGCAACAGTTGAGTTTTGCGATGCCCAG; 3CeCtBPA203E, CTGGGCATCGCAAAACTCAACTGTTGCAAC). The BamHI-EcoRI-digested mutant PCR product was ligated into the BamHI-EcoRI sites of pGAD10 and pGBT9(new) vectors to generate pGAD10-CTBP-1-A203E and pGBT9(new)-CTBP-1-A203E. The pag-3 cDNA was amplified by PCR (5PAG3Bam, CGGGATCCATGAGCACTGAGCAGGTGTC; 3PAG3Eco, CGGAATTCTTATGACACGCTCAAATTCAG) from cDNA synthesized as described for the ctbp-1 cDNA. The product was cloned into the BamHI-EcoRI sites of the pGAD10 vector to generate pGAD10-PAG-3 and sequenced. The ΔNL mutation was introduced into PAG-3 by PCR mutagenesis (5PAG3Bam; 3PAGEcoMut, CGGAATTCTTATGACACGCTCGAAGCCAG), and the product was cloned into the BamHI-EcoRI sites of the pGAD10 vector to generate pGAD10-PAG-3 ΔNL and sequenced. The Clontech yeast two-hybrid system was used according to the manufacturer's protocols, with test proteins expressed in yeast strain HF7c as either Gal4DBD or Gal4AD fusions. Transformant colonies were selected on -Leu-Trp plates and patched onto -His-Leu-Trp plates, and growth was scored after 72 h of incubation. PAG-3 and PAG-3 ΔNL were excised from pGAD10-PAG-3 and pGAD10-PAG-3 ΔNL, respectively, and cloned into BamHI-EcoRI pcDNA3-Gal4 to create pcDNA3-Gal4–PAG-3 and pcDNA3-Gal4–PAG-3 ΔNL, respectively. To examine PAG-3 repression of reporter gene expression, we transfected six-well plates of COS-1 cells. The following plasmids were used: 3 μg of the pGL2(Gal4)5-(LexA)2-E1B-Luc reporter, 1 μg of the pCMV-LexA(1–202)–VP16(410–490) expression vector, and either 0.5 or 1 μg of pcDNA3-Gal4–PAG-3 or pcDNA3-Gal4–PAG-3 ΔNL. Assays were performed and normalized as described in the legend to Fig. 3.
8 domain of CTBP-1 (Arg27, Lys30, Arg31, Arg36, Arg55, Lys66, and Lys67) are identical in all but one of the newly identified CtBP THAP domains. In the remaining THAP domain, that of B. floridae, although 2 of the residues are not identical, the charge of the residues is maintained; Arg31 is replaced by a Lys residue and Lys67 is replaced by an Arg residue. The conservation of the charge of this face, which we have suggested may represent the DNA-binding surface of the THAP domain,21 in each of the newly identified CtBP THAP domains suggests that they are also likely to possess DNAbinding activity. C. elegans CTBP-1 represses gene expression The above data indicate that the C. elegans protein CTBP-1 represents the founding member of a new subgroup of the CtBP family proteins within the animal kingdom, the THAP domain-containing CtBPs. Given their novelty, we sought to determine whether members of this subgroup are functionally orthologous to previously characterized animal CtBPs. To this end, we first investigated whether CTBP-1 represses gene expression. The established assay for testing the repression capacity of CtBPs involves fusing the protein to a heterologous DBD (Gal4DBD, from the yeast transcription factor Gal4) and testing its effect on Gal4responsive reporter constructs.11,26,27 A luciferase reporter gene under the control of a promoter containing two LexA sites and five Gal4-binding sites upstream of the adenoviral E1B promoter was used. This promoter was activated by expression of LexA–VP16 fusion proteins, and the capacity of Gal4DBD–CtBP fusions to repress this expression was assayed. The Gal4–CTBP-1 fusion protein repressed expression of the luciferase reporter, albeit to a lesser extent than a Gal4–mCtBP2 fusion protein (Fig. 3b). Taking account of the low level of expression of the Gal4–CTBP-1 fusion (6.5-fold lower expression) relative to the expression of the Gal4– mCtBP2 fusion (Fig. 3a), we found that these two proteins possess comparable repression capacity. The corepressor function of the animal CtBPs is thus conserved within the THAP-containing subgroup. Given that the THAP domain is not present in any of the previously characterized CtBP corepressors, we hypothesized that this domain might be dispensable for the repression activity of CTBP-1. Indeed, when a truncated version of CTBP-1 that lacks the THAP domain is fused to the Gal4DBD [Gal4DBD–CTBP-1(90–727)], the repression activity is equivalent to that of the full-length protein, indicating that the THAP domain is not required for repression (Fig. 3b). If, as is predicted by our previous findings,21 the CTBP-1 THAP domain serves primarily as a DBD, removal of the THAP domain in the assay system employed here would not be expected to alter repression capacity; the DNA-binding activity of the THAP domain is rendered redundant by the presence of the Gal4DBD.
CTBP-1, a THAP Domain-Containing CtBP Corepressor
CTBP-1 homodimerization and PXDLS binding Previous analyses have demonstrated that homodimerization and PXDLS-binding capacity represent two important functional capacities of CtBPs that act as corepressors.9,26 Testing the dimerization of CtBP family members and their interactions with PXDLSmotif-bearing partner proteins has previously been carried out successfully using the yeast two-hybrid system.11,15,18,27 We therefore used this system to determine whether CTBP-1 possesses these activities. The data shown in Fig. 4a demonstrate that CTBP-1 can homodimerize in yeast. We next asked whether CTBP-1 could interact with a PXDLScontaining transcription factor. For this assay, we selected the PAG-3 protein from C. elegans as a candidate CTBP-1 interactor. PAG-3 is a zinc finger protein with homology to Drosophila Senseless and mammalian Gfi that has been proposed to function as a transcriptional repressor.30 A PXDLS-like motif had previously been noted in PAG-3, making it a candidate partner for CTBP-1.31 As shown in Fig. 4b, CTBP-1 and PAG-3 can indeed interact in this assay. To ensure that the interaction was occurring through the expected PXDLS contact motif, we disrupted this motif in PAG-3 by mutagenesis. It is known that mutation of the core Asp-Leu portion of PXDLS motifs to Ala-Ser disrupts interactions with CtBPs (these mutations are referred to as ΔDL mutations).5,11 Accordingly, we tested for an interaction between CTBP-1 and a PAG-3 protein in which the PXDLS-like motif VLNLS had been mutated to VLASS (PAG-3 ΔNL). This mutation abrogated the interaction, suggesting that CTBP-1 directly targets the PXDLS-like motif in PAG-3 (Fig. 4b). Next, we tested whether the PAG-3 contact surface of CTBP-1 coincided with the predicted PXDLSbinding cleft within CTBP-1. Previously, mutation of an Ala residue within this cleft to Glu has been shown to disrupt interaction of mammalian CtBPs with PXDLS-containing partners. 6 We therefore tested the effect of the analogous mutation (A203E) on the capacity of CTBP-1 to interact with PAG-3. This mutation abrogated the interaction, suggesting that a functional PXDLS-binding cleft is conserved in the C. elegans CTBP-1 protein (Fig. 4c). To confirm that the inability of the CTBP-1 A203E mutant to interact with PAG-3 is not due to a disruption of folding or expression of the CTBP-1 protein in yeast, we also assessed the capacity of this mutant protein to dimerize with wild-type CTBP-1 (Fig. 4a). Interaction of CTBP-1 A203E with CTBP-1 suggests that the integrity of CTBP-1 is not compromised by the A203E mutation. PAG-3 can function as a CTBP-1-dependent repressor We next sought to determine whether PAG-3 can, as would be predicted by its interaction with CTBP1 and as has been suggested by previous in vivo studies, function as a transcriptional repressor.30 To this end, we fused PAG-3 to the Gal4DBD and
CTBP-1, a THAP Domain-Containing CtBP Corepressor
Fig. 5. Possible modes of THAP–CtBP recruitment to promoters. (a) A THAP domain-containing CtBP may be recruited by a PXDLS-containing transcription factor (TF) to a promoter that does not contain a THAP-binding site. (b) ATHAP domain-containing CtBP may be recruited to a promoter through interactions with a PXDLS-containing TF, and the THAP domain may stabilize CtBP at the promoter. (c) A THAP domain-containing CtBP may be targeted to other promoters directly through the DNAbinding activity of the THAP domain. In this model, the PXDLS-binding pocket of the CtBP might mediate contacts with PXDLS-containing repression effectors such as HDACs.
assessed the capacity of this fusion protein to repress expression of a Gal4-dependent promoter. The Gal4DBD–PAG-3 fusion protein repressed expression of the reporter (Fig. 4d). To determine the role played by CtBP in this repression activity, we tested whether the ΔNL mutation, which we had shown to disrupt the interaction of PAG-3 with CTBP-1, affected the repression capacity of PAG-3. On the same promoter, the Gal4DBD–PAG-3 ΔNL mutant protein failed to repress (Fig. 4d). Western blotting confirmed that Gal4DBD–PAG-3 ΔNL was expressed at levels equivalent to the wild-type protein (data not shown). Together, these assays suggest that PAG-3 can act as a CtBP-dependent repressor protein. The finding that CTBP-1 has the capacity to interact with PXDLS-containing partner proteins is in-
9 triguing. While interactions with PXDLS-containing transcription factors are required for the recruitment of the vertebrate and insect CtBPs to promoters, the intrinsic DNA-binding capacity of THAP-containing CtBPs provides an alternative mechanism for recruitment to DNA and might be expected to abrogate the requirement. We have, however, shown that CTBP-1 interacts with a PXDLS-containing transcription factor, PAG-3. Importantly, we have also demonstrated that this transcription factor requires contact with CtBP for its repression activity, suggesting that PAG-3 is a CtBP-dependent repressor in vivo. A PXDLS-like motif has been noted in another C. elegans transcription factor, ZAG-1, and such a motif is also found in EGL-43.32–34 The mammalian counterparts of these proteins, ZEB1/2 and Evi-1, respectively, are CtBP-dependent repressors. 35, 36 Therefore, ZAG-1 and EGL-43 are likely to be CtBP-dependent repressors in vivo. Thus, at least in some promoter contexts, the THAP domain-containing CTBP-1 is likely to be recruited by PXDLS-containing transcription factors (Fig. 5a). On some promoters, the DNA binding of the THAP domain may not be required for recruitment but could contribute to the stabilization of CTBP-1 at the promoter following recruitment by a PXDLScontaining transcription factor (Fig. 5b). In other promoter contexts, recruitment of CTBP-1 may be mediated by its THAP domain (Fig. 5c). In such cases, it is possible that the PXDLS cleft in CTBP-1 may mediate contact with PXDLS-containing effectors of repression, such as HDAC4.37 In each of the described scenarios, in addition to roles in DNA binding, like previously described zinc-binding domains,38 the THAP domain of CtBP may also serve as a protein–protein interaction module, mediating interactions with other transcriptional regulators. Given our finding of the widespread presence of THAP domain-containing CtBPs within the animal kingdom, clarifying the role of the THAP domain in CtBP-mediated repression will be an important goal.
Acknowledgements We would like to thank L. Gadreau and M. Ptashne for the plasmids pGL2-(Gal4)5-(LexA)2E1B-Luc and pCMV-LexA(1–202)–VP16(410–490) and A. Verger and J.P. Mackay for critical reading of the manuscript. This work was supported by a National Health and Medical Research Council Program Grant awarded to M.C., J.P. Mackay, and J.M. Matthews. H.R.N. is a University of Sydney Postdoctoral Research Fellow.
Supplementary Data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/ j.jmb.2007.10.041
10
References 1. Chinnadurai, G. (2003). CtBP family proteins: more than transcriptional corepressors. BioEssays, 25, 9–12. 2. Turner, J. & Crossley, M. (2001). The CtBP family: enigmatic and enzymatic transcriptional co-repressors. BioEssays, 23, 683–690. 3. Shi, Y., Sawada, J., Sui, G., Affar el, B., Whetstine, J. R., Lan, F. et al. (2003). Coordinated histone modifications mediated by a CtBP co-repressor complex. Nature, 422, 735–738. 4. Shi, Y., Lan, F., Matson, C., Mulligan, P., Whetstine, J. R., Cole, P. A. et al. (2004). Histone demethylation mediated by the nuclear amine oxidase homolog LSD1. Cell, 119, 941–953. 5. Schaeper, U., Boyd, J. M., Verma, S., Uhlmann, E., Subramanian, T. & Chinnadurai, G. (1995). Molecular cloning and characterization of a cellular phosphoprotein that interacts with a conserved C-terminal domain of adenovirus E1A involved in negative modulation of oncogenic transformation. Proc. Natl. Acad. Sci. USA, 92, 10467–10471. 6. Nardini, M., Spano, S., Cericola, C., Pesce, A., Massaro, A., Millo, E. et al. (2003). CtBP/BARS: a dual-function protein involved in transcription corepression and Golgi membrane fission. EMBO J. 22, 3122–3130. 7. Nardini, M., Svergun, D., Konarev, P. V., Spano, S., Fasano, M., Bracco, C. et al. (2006). The C-terminal domain of the transcriptional corepressor CtBP is intrinsically unstructured. Protein Sci. 15, 1042–1050. 8. Bernard, N., Johnsen, K., Gelpi, J. L., Alvarez, J. A., Ferain, T., Garmyn, D. et al. (1997). D-2-Hydroxy-4methylvalerate dehydrogenase from Lactobacillus delbrueckii subsp. bulgaricus. II. Mutagenic analysis of catalytically important residues. Eur. J. Biochem. 244, 213–219. 9. Kumar, V., Carlson, J. E., Ohgi, K. A., Edwards, T. A., Rose, D. W., Escalante, C. R. et al. (2002). Transcription corepressor CtBP is an NAD(+)-regulated dehydrogenase. Mol. Cell, 10, 857–869. 10. Balasubramanian, P., Zhao, L. J. & Chinnadurai, G. (2003). Nicotinamide adenine dinucleotide stimulates oligomerization, interaction with adenovirus E1A and an intrinsic dehydrogenase activity of CtBP. FEBS Lett. 537, 157–160. 11. Turner, J. & Crossley, M. (1998). Cloning and characterization of mCtBP2, a co-repressor that associates with basic Kruppel-like factor and other mammalian transcriptional regulators. EMBO J. 17, 5129–5140. 12. Phippen, T. M., Sweigart, A. L., Moniwa, M., Krumm, A., Davie, J. R. & Parkhurst, S. M. (2000). Drosophila C-terminal binding protein functions as a contextdependent transcriptional co-factor and interferes with both mad and groucho transcriptional repression. J. Biol. Chem. 275, 37628–37637. 13. Sutrias-Grau, M. & Arnosti, D. N. (2004). CtBP contributes quantitatively to Knirps repression activity in an NAD binding-dependent manner. Mol. Cell. Biol. 24, 5953–5966. 14. Nibu, Y., Zhang, H., Bajor, E., Barolo, S., Small, S. & Levine, M. (1998). dCtBP mediates transcriptional repression by Knirps, Kruppel and Snail in the Drosophila embryo. EMBO J. 17, 7009–7020. 15. Poortinga, G., Watanabe, M. & Parkhurst, S. M. (1998). Drosophila CtBP: a Hairy-interacting protein required for embryonic segmentation and hairy-mediated transcriptional repression. EMBO J. 17, 2067–2078.
CTBP-1, a THAP Domain-Containing CtBP Corepressor
16. Mani-Telang, P. & Arnosti, D. N. (2007). Developmental expression and phylogenetic conservation of alternatively spliced forms of the C-terminal binding protein corepressor. Dev. Genes Evol. 217, 127–135. 17. Folkers, U., Kirik, V., Schobinger, U., Falk, S., Krishnakumar, S., Pollock, M. A. et al. (2002). The cell morphogenesis gene ANGUSTIFOLIA encodes a CtBP/ BARS-like protein and is involved in the control of the microtubule cytoskeleton. EMBO J. 21, 1280–1288. 18. Kim, G. T., Shoda, K., Tsuge, T., Cho, K. H., Uchimiya, H., Yokoyama, R. et al. (2002). The ANGUSTIFOLIA gene of Arabidopsis, a plant CtBP gene, regulates leafcell expansion, the arrangement of cortical microtubules in leaf cells and expression of a gene involved in cell-wall formation. EMBO J. 21, 1267–1279. 19. Cho, K. H., Shindo, T., Kim, G. T., Nitasaka, E. & Tsukaya, H. (2005). Characterization of a member of the AN subfamily, IAN, from Ipomoea nil. Plant Cell Physiol. 46, 250–255. 20. Tsukaya, H. (2007). A new member of the CtBP/BARS family from plants: Angustifolia. In CtBP Family Proteins (Chinnadurai, G., ed), pp. 112–118, Landes Bioscience, Georgetown, TX. 21. Liew, C. K., Crossley, M., Mackay, J. P. & Nicholas, H. R. (2007). Solution structure of the THAP domain from Caenorhabditis elegans C-terminal binding protein (CtBP). J. Mol. Biol. 366, 382–390. 22. Clouaire, T., Roussigne, M., Ecochard, V., Mathe, C., Amalric, F. & Girard, J. P. (2005). The THAP domain of THAP1 is a large C2CH module with zinc-dependent sequence-specific DNA-binding activity. Proc. Natl Acad. Sci. USA, 102, 6907–6912. 23. Cayrol, C., Lacroix, C., Mathe, C., Ecochard, V., Ceribelli, M., Loreau, E. et al. (2007). The THAP-zinc finger protein THAP1 regulates endothelial cell proliferation through modulation of pRB/E2F cell-cycle target genes. Blood, 109, 584–594. 24. Macfarlan, T., Parker, J. B., Nagata, K. & Chakravarti, D. (2006). Thanatos-associated protein 7 associates with template activating factor-Ibeta and inhibits histone acetylation to repress transcription. Mol. Endocrinol. 20, 335–347. 25. Macfarlan, T., Kutney, S., Altman, B., Montross, R., Yu, J. & Chakravarti, D. (2005). Human THAP7 is a chromatin-associated, histone tail-binding protein that represses transcription via recruitment of HDAC3 and nuclear hormone receptor corepressor. J. Biol. Chem. 280, 7346–7358. 26. Thio, S. S., Bonventre, J. V. & Hsu, S. I. (2004). The CtBP2 co-repressor is regulated by NADH-dependent dimerization and possesses a novel N-terminal repression domain. Nucleic Acids Res. 32, 1836–1847. 27. Quinlan, K. G., Verger, A., Kwok, A., Lee, S. H., Perdomo, J., Nardini, M. et al. (2006). Role of the C-terminal binding protein PXDLS motif binding cleft in protein interactions and transcriptional repression. Mol. Cell. Biol. 26, 8202–8213. 28. Perdomo, J., Verger, A., Turner, J. & Crossley, M. (2005). Role for SUMO modification in facilitating transcriptional repression by BKLF. Mol. Cell. Biol. 25, 1549–1559. 29. Abramoff, M. D., Magelhaes, P. J. & Ram, S. J. (2004). Image processing with ImageJ. Biophoton. Int. 11, 36–42. 30. Jia, Y., Xie, G., McDermott, J. B. & Aamodt, E. (1997). The C. elegans gene pag-3 is homologous to the zinc finger proto-oncogene gfi-1. Development, 124, 2063–2073. 31. Jafar-Nejad, H. & Bellen, H. J. (2004). Gfi/Pag-3/ senseless zinc finger proteins: a unifying theme? Mol. Cell. Biol. 24, 8803–8812.
CTBP-1, a THAP Domain-Containing CtBP Corepressor
32. Clark, S. G. & Chiu, C. (2003). C. elegans ZAG-1, a Znfinger-homeodomain protein, regulates axonal development and neuronal differentiation. Development, 130, 3781–3794. 33. Wacker, I., Schwarz, V., Hedgecock, E. M. & Hutter, H. (2003). zag-1, a Zn-finger homeodomain transcription factor controlling neuronal differentiation and axon outgrowth in C. elegans. Development, 130, 3795–3805. 34. Rimann, I. & Hajnal, A. (2007). Regulation of anchor cell invasion and uterine cell fates by the egl-43 Evi-1 proto-oncogene in Caenorhabditis elegans. Dev. Biol. 308, 187–195. 35. Postigo, A. A. & Dean, D. C. (1999). ZEB represses transcription through interaction with the corepressor CtBP. Proc. Natl Acad. Sci. USA, 96, 6683–6688.
11 36. Nitta, E., Izutsu, K., Yamaguchi, Y., Imai, Y., Ogawa, S., Chiba, S. et al. (2005). Oligomerization of Evi-1 regulated by the PR domain contributes to recruitment of corepressor CtBP. Oncogene, 24, 6165–6173. 37. Zhang, C. L., McKinsey, T. A., Lu, J. R. & Olson, E. N. (2001). Association of COOH-terminal-binding protein (CtBP) and MEF2-interacting transcription repressor (MITR) contributes to transcriptional repression of the MEF2 transcription factor. J. Biol. Chem. 276, 35–39. 38. Liew, C. K., Simpson, R. J., Kwan, A. H., Crofts, L. A., Loughlin, F. E., Matthews, J. M. et al. (2005). Zinc fingers as protein recognition motifs: structural basis for the GATA-1/friend of GATA interaction. Proc. Natl Acad. Sci. USA, 102, 583–588.
doi:10.1016/j.jmb.2007.10.041
Available online at www.sciencedirect.com
COMMUNICATION
The Caenorhabditis elegans Protein CTBP-1 Defines a New Group of THAP Domain-Containing CtBP Corepressors Hannah R. Nicholas⁎, Jason A. Lowry, Tina Wu and Merlin Crossley School of Molecular and Microbial Biosciences, University of Sydney, Sydney, New South Wales 2006, Australia Received 2 August 2007; received in revised form 9 October 2007; accepted 16 October 2007 Available online 22 October 2007
The C-terminal binding proteins (CtBPs) play roles in diverse cellular processes including transcriptional regulation, Golgi membrane fission, and synaptic ribbon formation. In the context of transcriptional regulation, they function as corepressors, interacting with promoter-bound transcription factors and recruiting a large protein complex that contains chromatinmodifying enzymes. We recently described the structure of a Thanatosassociated protein (THAP) domain that is found in a new member of the CtBP family, the Caenorhabditis elegans CTBP-1 protein. We have identified additional THAP domain-containing CtBPs in the nematode, echinoderm, and cephalochordate lineages. The distribution of these lineages within the animal kingdom suggests that the ancestral form of the animal CtBPs may have contained a THAP domain that was subsequently lost in the vertebrate and arthropod lineages. We also provide functional data indicating that CTBP-1 represses gene expression and homodimerizes and interacts with PXDLS-containing partner proteins, three key features of the previously characterized animal CtBPs. CTBP-1 is therefore the founding member of a new subgroup within the CtBP corepressor family, the THAP domaincontaining CtBPs. © 2007 Elsevier Ltd. All rights reserved.
Edited by J. Karn
Keywords: CtBP; THAP domain; gene regulation; corepressor; Caenorhabditis elegans
Members of the C-terminal binding protein (CtBP) family are multifunctional proteins that have been ascribed roles in transcriptional corepression, Golgi maintenance, and synaptic ribbon formation.1 As transcriptional corepressors, CtBPs associate with DNA-bound transcription factors, many of which contain a short sequence with the consensus PXDLS.2 CtBPs function as part of a large corepressor complex that contains chromatin-modifying enzymes including histone deacetylases (HDACs), methyltransferases, and a demethylase.3,4 CtBPs have been reported to interact with many transcription factors,
*Corresponding author. E-mail address: [email protected]. Abbreviations used: CtBP, C-terminal binding protein; THAP, Thanatos-associated protein; HDAC, histone deacetylase; AN, ANGUSTIFOLIA; EST, expressed sequence tag; DBD, DNA-binding domain.
suggesting that CtBP-mediated gene repression is a widely utilized mechanism. The CtBPs share primary sequence and structural similarity with the D-isomer-specific 2-hydroxyacid dehydrogenases, which are composed of two functional domains: a substrate-binding domain and a nucleotide-binding domain.5,6 In addition to an analogous N-terminal substrate-binding domain and central nucleotide-binding domain, the CtBPs possess an intrinsically unstructured C-terminal region.6,7 The PXDLS-binding cleft through which CtBP contacts most of its partner proteins resides primarily in the substrate-binding domain. The central nucleotide-binding domain contains the nucleotide-binding consensus G/AXGXXGX17D (where X is any amino acid) and mediates interaction with NAD(H) as well as dimerization.6 Also conserved within the nucleotide-binding domain of the CtBPs is a histidine residue that is required for the catalytic activity of the D-isomer-specific 2-hydroxyacid dehydrogenase enzymes.5,8 Although CtBPs
0022-2836/$ - see front matter © 2007 Elsevier Ltd. All rights reserved.
2 have been shown to reduce pyruvate to lactic acid, this enzymatic activity has not been shown to be required for the repression function of the CtBPs.9–13 CtBP-related genes have been identified in numerous animal species. Invertebrates including Drosophila melanogaster, Anopheles gambiae, Apis mellifera, and Tribolium castaneum contain a single CtBP gene.14–16 In contrast, the studied vertebrates including human, mouse, rat, chicken, quail, zebrafish, and Xenopus contain two CtBPs (named CtBP1 and CtBP2). CtBP-like proteins have also been identified in a number of plant species. The best characterized of these is the Arabidopsis thaliana protein called ANGUSTIFOLIA (AN).17–19 Despite its similarity to the animal CtBPs, AN appears to be unable to associate with the PXDLS motif.20 Furthermore, the nucleotide-binding motif and dehydrogenase catalytic histidine are not conserved in AN. 17,18,20 Consistent with this lack of conservation of key functional elements, phylogenetic analysis indicates that the plant members of the CtBP family form a unique subfamily, suggesting that they may serve functions that are specific to plants.18 We recently described the structure of a protein domain, termed Thanatos-associated protein (THAP) domain, contained within a new member of the CtBP family, the CTBP-1 protein from the nematode Caenorhabditis elegans.21 Here, we present evidence that CTBP-1 defines a new subgroup of THAP–CtBP corepressors that extends across animal lineages. Our results suggest that the THAP–CtBPs represent the ancestral form of CtBP, with THAP domains having subsequently been lost in vertebrate and various other lineages. The C. elegans ctbp-1 gene encodes a CtBP Homologue with an N-Terminal THAP domain Early analysis of the genome of C. elegans predicted an open reading frame named F49E10.5 encoding a
CTBP-1, a THAP Domain-Containing CtBP Corepressor
612-amino-acid CtBP homologue (T34290), which has been cited in sequence comparisons of members of the CtBP family.11,18 Subsequently, expressed sequence tag (EST) data prompted the merging of F49E10.5 with its neighbor, F49E10.6. We have named this revised predicted gene ctbp-1.21 To verify this prediction, we sequenced products arising from PCR amplification of cDNA derived from RNA extracted from a mixed-stage population of C. elegans hermaphrodites. Five clones were sequenced, two of which corresponded to the predicted cDNA (NM_076582). The remaining clones prompt two amendments to the predicted transcript structure (Fig. 1a). First, there is an alternative 3′ splice site within the 3rd intron, which, when used, results in the inclusion of an additional 9 nucleotides in exon 4. Second, within the predicted 13th exon, there is an intron of 48 nucleotides, retained in some transcripts, which necessitates the definition of a 14th exon. We have not investigated the functional significance of these alternative splice isoforms. Conceptual translation of the ctbp-1 cDNA (NM_076582) yields a protein of 727 amino acids in which a search of the National Center for Biotechnology Information Conserved Domain Database identified two domains: amino acids 1–89 comprise a THAP domain, which we have described previously,21 while amino acids 178–505 show similarity to the D-isomer-specific 2-hydroxyacid dehydrogenases (Fig. 1b). Amino acids 90–177, linking the THAP domain to the dehydrogenase domain, and the C-terminal residues 506–727 do not contain any detectable structural motifs. Comparison of the dehydrogenase domain of CTBP-1 (residues 178–505) with that from human CtBP1 (residues 27–353) reveals 55% identity and 74% similarity of the amino acid sequences. Furthermore, several features of potential functional importance are conserved in the C. elegans protein (Fig. 1c).
Fig. 1. The C. elegans ctbp-1 gene encodes a protein that contains a THAP domain and is homologous to animal CtBPs. (a) Structure of the ctbp-1 transcript. Exons are depicted as boxes and introns are depicted as lines. Amendments to the predicted gene structure are shown below the schematic. Splice donor and acceptor sites are shown in bold, intronic sequences are shown in lower case, and exonic sequences are shown in upper case. Underlined in sequence (i) is an alternative 3′ splice site in the 3rd intron. Sequence (ii) shows the 13th intron, which is retained in some transcripts. (b) Schematic representations of the human CtBP1 and C. elegans CTBP-1 proteins showing domains identified in the National Center for Biotechnology Information Conserved Domain Database. Dark gray shading indicates the D-isomerspecific 2-hydroxyacid dehydrogenase domain. Light gray shading indicates the THAP domain. (c) ClustalW alignment of representative vertebrate (hCtBP1, NP_001319; hCtBP2, AAH52276) and insect (dCtBP, NP_001014617) CtBPs with C. elegans CtBP (CTBP-1, NP_508983). Symbols below the alignment denote the degree of conservation observed in each column. Asterisks indicate identical residues, “:” indicates conservative substitutions, and “.” indicates semiconservative substitutions. Features of potential functional importance are indicated above the alignment. “^” indicates residues lining the PXDLS-binding cleft, a line indicates the nucleotide-binding motif GXGXXGX17D, and “‡” indicates the His residue that is critical for the catalytic activity of the D-isomer-specific 2-hydroxyacid dehydrogenases. Dark gray shading indicates the D-isomer-specific 2-hydroxyacid dehydrogenase domain. Light gray shading indicates the THAP domain. Methods: Reverse transcriptase DNA amplification reactions were performed to obtain the ctbp-1 cDNA as follows. Total RNA was extracted from a mixed-stage population of C. elegans N2 (wild type) hermaphrodites using TRI Reagent® (Sigma-Aldrich) as described by the manufacturer. First-strand cDNA was then synthesized using the SuperScript™ II First-Strand Synthesis System for RT-PCR (Invitrogen) and oligo(dT)20 primers. The ctbp-1 cDNA was amplified by PCR (5BamcTC, CGGGATCCATGCCGACGACTTGTGGATTTCC; 3CeCtBPcDNA, AGTTGTCACAACCTCGAGAAC), and the product was then used as a template for a nested PCR (5BamcTC; 3EcoflC, CGGAATTCTTATGTGGCCAATGGTTGCTC). The product was cloned into the BamHI-EcoRI pBSSK (Stratagene) vector, and five clones were sequenced. One clone that corresponded to the predicted ctbp-1 cDNA sequence (NM_076582) was used for all subsequent clonings and is referred to as CTBP-1.
CTBP-1, a THAP Domain-Containing CtBP Corepressor
Fig. 1 (legend on previous page)
3
4 First, all the residues that line the PXDLS-binding cleft are present in the CTBP-1 sequence (Leu180, Ala182, Met193, Thr201, Ala203, Phe204, Cys205, His214, and Val217). Second, the CTBP-1 protein contains the nucleotide-binding consensus GXGXXGX17D (residues 332–355). Third, the catalytic histidine that is essential for dehydrogenase activity is conserved in the CTBP-1 protein (His443). This high degree of sequence homology argues for the inclusion of CTBP-1 in the CtBP family of proteins. As noted, unlike previously described members of this protein family, CTBP-1 contains an N-terminal THAP domain. Defined by a zinc-coordinating consensus sequence Cys-X2–4-Cys-X35–53-Cys-X2His, our structural data have shown that the THAP domain belongs to the zinc finger superfamily and contains a treble clef motif as its structural core.21 Although sequence searches have
CTBP-1, a THAP Domain-Containing CtBP Corepressor
identified over 100 animal proteins that contain one or more THAP domains,22 the vast majority of these have not been characterized. Those that have been studied appear to be involved in transcriptional regulation; for example, human THAP1 acts as a transcription factor for cell-cycle target genes23 and human THAP7 acts as a transcriptional repressor.24,25 The THAP domains of these proteins appear to be integral to their transcriptional activities, with the THAP domain in THAP1 serving as a DNAbinding domain (DBD)22 and the corresponding domain in THAP7 recruiting HDAC3.25 We have shown that, like the THAP domain of the human protein THAP1, the C. elegans CtBP THAP domain can bind DNA.21 Our structural data revealed the presence of a large, positively charged face on the C. elegans CtBP THAP domain, which may constitute the DNA contact surface.21
Fig. 2. The CtBPs from diverse organisms contain highly conserved THAP domains. (a) Phylogenetic tree based on the tree of life [tolweb.org/tree/phylogeny.html] depicting representative species in which a CtBP homologue has been either previously identified or identified in the current study. “X” indicates branches in which the CtBPs do not contain a THAP domain. Species in which a THAP domain-containing CtBP was identified are shaded gray. “?” denotes branches in which ESTs or genomic contigs show the presence of at least part of the THAP domain although the complete genomic sequence was unavailable. (b) Sequence alignment of five THAP domains from CtBPs: C. elegans (NP_508983), C. briggsae (CAE68813), C. remanei (GenBank whole genome shotgun sequence AAGD01000137), S. purpuratus (XP_001178649), and B. floridae (GenBank EST: BW870057 and Genome v1.0 scaffold_232). Conserved residues are indicated by asterisks. Seven charged residues that cluster on one face of the C. elegans CtBP THAP domain are shaded in gray. Methods: Available CtBP sequences were collected from GenBank and European Molecular Biology Laboratory databases. The C. elegans and S. purpuratus protein sequences were used to search (tblastn) the EST and genomes within GenBank and the following specialized genome databases: C. intestinalis, B. floridae, and D. pulex [genome.jgi-psf.org/euk_home.html].
CTBP-1, a THAP Domain-Containing CtBP Corepressor
5
Fig. 3. CTBP-1 represses gene transcription. (a) A Western blot was performed to examine expression of Gal4DBD fused to mCtBP2, CTBP-1, and CTBP-1(90–727) in transiently transfected COS-1 cells. (b) Gal4DBD–CtBP fusion constructs were tested for their ability to repress LexA–VP16-activated expression of a luciferase reporter in COS-1 cells (n = 4, ±SD). Methods: The full-length ctbp-1 cDNA was used as a template for the PCR amplification of a truncated cDNA lacking the first 267 nucleotides (5BamcnonTC, CGGGATCCGAAAGTAAGAACAGTGATCG; 3EcoflC). The PCR product was cloned into the BamHI-EcoRI pBSSK (Stratagene) vector and sequenced. The truncated cDNA is referred to as CTBP-1(90–727). For expression of proteins fused to the Gal4DBD in mammalian cells, the pcDNA3-Gal4 plasmid was made by ligating a HindIII-BamHI Gal4DBD fragment that had been excised from pGBT9(new) (derived from pGBT9 from Clontech) into the HindIII-BamHI sites of the pcDNA3 (Invitrogen) vector. CTBP-1 and CTBP-1(90–727) were excised from pBSSK-CTBP-1 and pBSSK-CTBP-1(90–727), respectively, and cloned into BamHI-EcoRI pcDNA3-Gal4 to create pcDNA3-Gal4–CTBP-1 and pcDNA3-Gal4–CTBP-1(90–727), respectively. The pcDNA3-Gal4–mCtBP2 plasmid has been described previously.11 The firefly luciferase reporter vector pGL2-(Gal4)5-(LexA)2-E1B-Luc and LexA–VP16 mammalian expression plasmid pCMV-LexA(1–202)–VP16(410–490) were gifts from Luke Gadreau and Mark Ptashne (The Sloan–Kettering Institute, New York, NY, USA). COS-1 cells were cultured as described previously28 and transfected with FuGENE6 (Roche Diagnostics) transfection reagent according to the manufacturer's instructions. Western blot assays were performed to assess the relative expression levels of the Gal4 fusion proteins used in the mammalian repression assays. Nuclear extracts were made from 10-cm dishes of COS-1 transfected with 4 μg of the relevant expression plasmid as has been described previously.28 The proteins were detected on Western blots using an anti-Gal4DBD antibody. Relative expression of fusion proteins was quantified using ImageJ image-processing software.29 To examine CtBP repression of reporter gene expression, we transfected six-well plates of COS-1 cells. The following plasmids were used: 3 μg of the pGL2-(Gal4)5-(LexA)2-E1B-Luc reporter, 1 μg of the pCMV-LexA(1–202)–VP16(410–490) expression vector, and either 1 μg of pcDNA3-Gal4 or 0.25, 0.5, or 1 μg of pcDNA3-Gal4–mCtBP2, pcDNA3-Gal4–CTBP-1, or pcDNA3-Gal4– CTBP-1(90–727). Renilla luciferase vector pRL-Luc (10 ng; Promega) was cotransfected to enable adjustment of firefly luciferase measurements to control for transfection efficiency. Luciferase activity was measured 48 h posttransfection in a Turner Designs model TD-20/20 Luminometer with the dual-luciferase reporter assay system (Promega), enabling calculation of the firefly:Renilla luciferase ratio (FF/R luciferase ratio).
CTBP-1, a THAP Domain-Containing CtBP Corepressor
6 Additional CtBPs from diverse lineages, nematodes, echinoderms and cephalochordates contain a THAP domain Given that the C. elegans protein CTBP-1 is the only member of the CtBP family identified to date that contains a THAP domain, we asked whether THAP domain-containing CtBPs might exist in other organisms. We used BLAST searches of sequenced genomes and EST databases to identify any such proteins. Within the nematode phylum, we have identified THAP domain-containing CtBPs in two additional Caenorhabditis species, C. remanei and C. briggsae, and have found genomic sequences indicating the pre-
sence of a similar THAP domain in the more distant nematode Haemonchus contortus (Fig. 2a, Supplementary Fig. 1). We have also found evidence for THAP domain-containing CtBPs beyond the nematodes. Specifically, sequence data support the existence of four THAP domain-containing CtBPs among the deuterostomes: three in the echinoderm lineage and one in the cephalochordate lineage. In the echinoderm lineage, genomic sequences supported by ESTs show that the CtBP of the sea urchin Strongylocentrotus purpuratus contains a THAP domain. Also within this lineage, ESTs indicate the presence of part of a THAP domain in the CtBP of both the common urchin Paracentrotus lividus and the sea star Asterina pectinifera. As complete genomic
Fig. 4 (legend on next page)
CTBP-1, a THAP Domain-Containing CtBP Corepressor
sequence data are unavailable for these latter two species, we have been unable to confirm that the entire THAP domain is present. In the cephalochordate lineage, genomic sequences supported by ESTs show that the lancelet Branchiostoma floridae also has a THAP domain-containing CtBP. The relative positions of these species on the tree of life are depicted in Fig. 2a. Two other surveyed species in which CtBPs were identified that do not contain a THAP domain are also shown: a urochordate, the sea squirt Ciona intestinalis (JGI genome v2.0 scaffold chr_05q), and a crustacean, the water flea Daphnia pulex (dpulex_jgi0609051 scaffold_3). Although CtBP homologues from at least 17 species had previously been described, these species represent only two animal clades: vertebrates and insects. Our sequence searches have identified 10 additional homologues from species representing diverse lineages within the animal kingdom. Importantly, among these new CtBP homologues, we have identified THAP domain-containing CtBPs in three groups: nematodes, echinoderms, and cephalochordates. The positions of these animals in the tree of life suggest that the THAP–CtBP fusion might in fact represent the ancestral form of the animal CtBP. Our identification of a crustacean CtBP that does not contain a THAP domain, together with previous reports of insect CtBPs that similarly do not contain this domain,14–16 suggests that the THAP domain has been lost in the arthropod phylum. Similarly, in the chordates, our data are consistent with loss of
7 the THAP domain in both the urochordate and the vertebrate lineages. Given that the THAP–CtBP fusion is widespread within the animal kingdom and that it may represent the ancestral form of the animal CtBP, we conclude that the THAP domaincontaining CtBPs represent an important new group within the CtBP family of proteins. The THAP domains of the CtBPs are highly conserved The amino acid sequences of the THAP domains from the identified THAP domain-containing CtBPs were compared (Fig. 2b). In addition to the zinccoordinating consensus (Cys-X2–4-Cys-X35–53-CysX2-His), 4 additional residues are highly conserved among all THAP domains and form part of the hydrophobic core of the domain (Pro29, Trp39, Phe61, and Pro72, numbered according to the C. elegans CTBP-1 THAP domain).21 Both elements are present in all the CtBP-associated THAP domains that we have identified, suggesting that they are true THAP domains. Notably, the similarity between these CtBP-associated THAP domains extends beyond these 8 consensus residues. Indeed, among the five identified CtBP THAP domains, 60 of the 89 residues (67%) are identical in all sequences (Fig. 2b). This identity suggests that the CtBP THAP domains are likely to serve similar functions. Supportive of this notion, the residues that constitute the positively charged surface patch on the THAP
Fig. 4. CTBP-1 homodimerizes and interacts with a PXDLS-containing partner protein through the PXDLS-binding cleft. This partner protein can repress transcription in a CtBP-dependent manner. (a–c) Yeast two-hybrid assays were performed to examine the interactions of CTBP-1 with partner proteins. The assays were performed by fusing the test proteins to the C-terminus of either the Gal4DBD or the Gal4 activation domain (Gal4AD). Growth on -His-Leu-Trp plates (pictured) indicates that the two proteins interact. Homodimerization of wild-type CTBP-1 was assessed (a), as was the interaction of CTBP-1 with a putative PXDLS-containing transcription factor from C. elegans, PAG-3 (b). The PXDLS-like motif in PAG-3 was mutated and the mutant protein (PAG-3 ΔNL) was tested for its capacity to interact with CTBP-1 to examine the PXDLS dependence of the interaction between CTBP-1 and PAG-3 (b). The dependence of the interaction between CTBP-1 and PAG-3 on the PXDLS-binding cleft of CTBP-1 was assessed by mutating one residue within the putative PXDLS-binding cleft and testing the capacity of the mutant protein (CTBP-1 A203E) to interact with PAG-3 (c). To confirm the integrity of the mutant protein CTBP-1 A203E, we also tested the capacity of this protein to dimerize with CTBP-1 (a). (d) Gal4DBD–PAG-3 fusion constructs were tested for their ability to repress LexA–VP16-activated expression of a luciferase reporter in COS-1 cells (n = 4, ±SD). Methods: The CTBP-1 insert was subcloned into the BamHI-EcoRI sites of the pGAD10 (Clontech) vector and pGBT9(new) vector to allow expression in the yeast two-hybrid system as both Gal4AD and Gal4DBD fusions. The A203E mutation was introduced into CTBP-1 by overlap PCR mutagenesis (5CeCtBPA203E, GTTGCAACAGTTGAGTTTTGCGATGCCCAG; 3CeCtBPA203E, CTGGGCATCGCAAAACTCAACTGTTGCAAC). The BamHI-EcoRI-digested mutant PCR product was ligated into the BamHI-EcoRI sites of pGAD10 and pGBT9(new) vectors to generate pGAD10-CTBP-1-A203E and pGBT9(new)-CTBP-1-A203E. The pag-3 cDNA was amplified by PCR (5PAG3Bam, CGGGATCCATGAGCACTGAGCAGGTGTC; 3PAG3Eco, CGGAATTCTTATGACACGCTCAAATTCAG) from cDNA synthesized as described for the ctbp-1 cDNA. The product was cloned into the BamHI-EcoRI sites of the pGAD10 vector to generate pGAD10-PAG-3 and sequenced. The ΔNL mutation was introduced into PAG-3 by PCR mutagenesis (5PAG3Bam; 3PAGEcoMut, CGGAATTCTTATGACACGCTCGAAGCCAG), and the product was cloned into the BamHI-EcoRI sites of the pGAD10 vector to generate pGAD10-PAG-3 ΔNL and sequenced. The Clontech yeast two-hybrid system was used according to the manufacturer's protocols, with test proteins expressed in yeast strain HF7c as either Gal4DBD or Gal4AD fusions. Transformant colonies were selected on -Leu-Trp plates and patched onto -His-Leu-Trp plates, and growth was scored after 72 h of incubation. PAG-3 and PAG-3 ΔNL were excised from pGAD10-PAG-3 and pGAD10-PAG-3 ΔNL, respectively, and cloned into BamHI-EcoRI pcDNA3-Gal4 to create pcDNA3-Gal4–PAG-3 and pcDNA3-Gal4–PAG-3 ΔNL, respectively. To examine PAG-3 repression of reporter gene expression, we transfected six-well plates of COS-1 cells. The following plasmids were used: 3 μg of the pGL2(Gal4)5-(LexA)2-E1B-Luc reporter, 1 μg of the pCMV-LexA(1–202)–VP16(410–490) expression vector, and either 0.5 or 1 μg of pcDNA3-Gal4–PAG-3 or pcDNA3-Gal4–PAG-3 ΔNL. Assays were performed and normalized as described in the legend to Fig. 3.
8 domain of CTBP-1 (Arg27, Lys30, Arg31, Arg36, Arg55, Lys66, and Lys67) are identical in all but one of the newly identified CtBP THAP domains. In the remaining THAP domain, that of B. floridae, although 2 of the residues are not identical, the charge of the residues is maintained; Arg31 is replaced by a Lys residue and Lys67 is replaced by an Arg residue. The conservation of the charge of this face, which we have suggested may represent the DNA-binding surface of the THAP domain,21 in each of the newly identified CtBP THAP domains suggests that they are also likely to possess DNAbinding activity. C. elegans CTBP-1 represses gene expression The above data indicate that the C. elegans protein CTBP-1 represents the founding member of a new subgroup of the CtBP family proteins within the animal kingdom, the THAP domain-containing CtBPs. Given their novelty, we sought to determine whether members of this subgroup are functionally orthologous to previously characterized animal CtBPs. To this end, we first investigated whether CTBP-1 represses gene expression. The established assay for testing the repression capacity of CtBPs involves fusing the protein to a heterologous DBD (Gal4DBD, from the yeast transcription factor Gal4) and testing its effect on Gal4responsive reporter constructs.11,26,27 A luciferase reporter gene under the control of a promoter containing two LexA sites and five Gal4-binding sites upstream of the adenoviral E1B promoter was used. This promoter was activated by expression of LexA–VP16 fusion proteins, and the capacity of Gal4DBD–CtBP fusions to repress this expression was assayed. The Gal4–CTBP-1 fusion protein repressed expression of the luciferase reporter, albeit to a lesser extent than a Gal4–mCtBP2 fusion protein (Fig. 3b). Taking account of the low level of expression of the Gal4–CTBP-1 fusion (6.5-fold lower expression) relative to the expression of the Gal4– mCtBP2 fusion (Fig. 3a), we found that these two proteins possess comparable repression capacity. The corepressor function of the animal CtBPs is thus conserved within the THAP-containing subgroup. Given that the THAP domain is not present in any of the previously characterized CtBP corepressors, we hypothesized that this domain might be dispensable for the repression activity of CTBP-1. Indeed, when a truncated version of CTBP-1 that lacks the THAP domain is fused to the Gal4DBD [Gal4DBD–CTBP-1(90–727)], the repression activity is equivalent to that of the full-length protein, indicating that the THAP domain is not required for repression (Fig. 3b). If, as is predicted by our previous findings,21 the CTBP-1 THAP domain serves primarily as a DBD, removal of the THAP domain in the assay system employed here would not be expected to alter repression capacity; the DNA-binding activity of the THAP domain is rendered redundant by the presence of the Gal4DBD.
CTBP-1, a THAP Domain-Containing CtBP Corepressor
CTBP-1 homodimerization and PXDLS binding Previous analyses have demonstrated that homodimerization and PXDLS-binding capacity represent two important functional capacities of CtBPs that act as corepressors.9,26 Testing the dimerization of CtBP family members and their interactions with PXDLSmotif-bearing partner proteins has previously been carried out successfully using the yeast two-hybrid system.11,15,18,27 We therefore used this system to determine whether CTBP-1 possesses these activities. The data shown in Fig. 4a demonstrate that CTBP-1 can homodimerize in yeast. We next asked whether CTBP-1 could interact with a PXDLScontaining transcription factor. For this assay, we selected the PAG-3 protein from C. elegans as a candidate CTBP-1 interactor. PAG-3 is a zinc finger protein with homology to Drosophila Senseless and mammalian Gfi that has been proposed to function as a transcriptional repressor.30 A PXDLS-like motif had previously been noted in PAG-3, making it a candidate partner for CTBP-1.31 As shown in Fig. 4b, CTBP-1 and PAG-3 can indeed interact in this assay. To ensure that the interaction was occurring through the expected PXDLS contact motif, we disrupted this motif in PAG-3 by mutagenesis. It is known that mutation of the core Asp-Leu portion of PXDLS motifs to Ala-Ser disrupts interactions with CtBPs (these mutations are referred to as ΔDL mutations).5,11 Accordingly, we tested for an interaction between CTBP-1 and a PAG-3 protein in which the PXDLS-like motif VLNLS had been mutated to VLASS (PAG-3 ΔNL). This mutation abrogated the interaction, suggesting that CTBP-1 directly targets the PXDLS-like motif in PAG-3 (Fig. 4b). Next, we tested whether the PAG-3 contact surface of CTBP-1 coincided with the predicted PXDLSbinding cleft within CTBP-1. Previously, mutation of an Ala residue within this cleft to Glu has been shown to disrupt interaction of mammalian CtBPs with PXDLS-containing partners. 6 We therefore tested the effect of the analogous mutation (A203E) on the capacity of CTBP-1 to interact with PAG-3. This mutation abrogated the interaction, suggesting that a functional PXDLS-binding cleft is conserved in the C. elegans CTBP-1 protein (Fig. 4c). To confirm that the inability of the CTBP-1 A203E mutant to interact with PAG-3 is not due to a disruption of folding or expression of the CTBP-1 protein in yeast, we also assessed the capacity of this mutant protein to dimerize with wild-type CTBP-1 (Fig. 4a). Interaction of CTBP-1 A203E with CTBP-1 suggests that the integrity of CTBP-1 is not compromised by the A203E mutation. PAG-3 can function as a CTBP-1-dependent repressor We next sought to determine whether PAG-3 can, as would be predicted by its interaction with CTBP1 and as has been suggested by previous in vivo studies, function as a transcriptional repressor.30 To this end, we fused PAG-3 to the Gal4DBD and
CTBP-1, a THAP Domain-Containing CtBP Corepressor
Fig. 5. Possible modes of THAP–CtBP recruitment to promoters. (a) A THAP domain-containing CtBP may be recruited by a PXDLS-containing transcription factor (TF) to a promoter that does not contain a THAP-binding site. (b) ATHAP domain-containing CtBP may be recruited to a promoter through interactions with a PXDLS-containing TF, and the THAP domain may stabilize CtBP at the promoter. (c) A THAP domain-containing CtBP may be targeted to other promoters directly through the DNAbinding activity of the THAP domain. In this model, the PXDLS-binding pocket of the CtBP might mediate contacts with PXDLS-containing repression effectors such as HDACs.
assessed the capacity of this fusion protein to repress expression of a Gal4-dependent promoter. The Gal4DBD–PAG-3 fusion protein repressed expression of the reporter (Fig. 4d). To determine the role played by CtBP in this repression activity, we tested whether the ΔNL mutation, which we had shown to disrupt the interaction of PAG-3 with CTBP-1, affected the repression capacity of PAG-3. On the same promoter, the Gal4DBD–PAG-3 ΔNL mutant protein failed to repress (Fig. 4d). Western blotting confirmed that Gal4DBD–PAG-3 ΔNL was expressed at levels equivalent to the wild-type protein (data not shown). Together, these assays suggest that PAG-3 can act as a CtBP-dependent repressor protein. The finding that CTBP-1 has the capacity to interact with PXDLS-containing partner proteins is in-
9 triguing. While interactions with PXDLS-containing transcription factors are required for the recruitment of the vertebrate and insect CtBPs to promoters, the intrinsic DNA-binding capacity of THAP-containing CtBPs provides an alternative mechanism for recruitment to DNA and might be expected to abrogate the requirement. We have, however, shown that CTBP-1 interacts with a PXDLS-containing transcription factor, PAG-3. Importantly, we have also demonstrated that this transcription factor requires contact with CtBP for its repression activity, suggesting that PAG-3 is a CtBP-dependent repressor in vivo. A PXDLS-like motif has been noted in another C. elegans transcription factor, ZAG-1, and such a motif is also found in EGL-43.32–34 The mammalian counterparts of these proteins, ZEB1/2 and Evi-1, respectively, are CtBP-dependent repressors. 35, 36 Therefore, ZAG-1 and EGL-43 are likely to be CtBP-dependent repressors in vivo. Thus, at least in some promoter contexts, the THAP domain-containing CTBP-1 is likely to be recruited by PXDLS-containing transcription factors (Fig. 5a). On some promoters, the DNA binding of the THAP domain may not be required for recruitment but could contribute to the stabilization of CTBP-1 at the promoter following recruitment by a PXDLScontaining transcription factor (Fig. 5b). In other promoter contexts, recruitment of CTBP-1 may be mediated by its THAP domain (Fig. 5c). In such cases, it is possible that the PXDLS cleft in CTBP-1 may mediate contact with PXDLS-containing effectors of repression, such as HDAC4.37 In each of the described scenarios, in addition to roles in DNA binding, like previously described zinc-binding domains,38 the THAP domain of CtBP may also serve as a protein–protein interaction module, mediating interactions with other transcriptional regulators. Given our finding of the widespread presence of THAP domain-containing CtBPs within the animal kingdom, clarifying the role of the THAP domain in CtBP-mediated repression will be an important goal.
Acknowledgements We would like to thank L. Gadreau and M. Ptashne for the plasmids pGL2-(Gal4)5-(LexA)2E1B-Luc and pCMV-LexA(1–202)–VP16(410–490) and A. Verger and J.P. Mackay for critical reading of the manuscript. This work was supported by a National Health and Medical Research Council Program Grant awarded to M.C., J.P. Mackay, and J.M. Matthews. H.R.N. is a University of Sydney Postdoctoral Research Fellow.
Supplementary Data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/ j.jmb.2007.10.041
10
References 1. Chinnadurai, G. (2003). CtBP family proteins: more than transcriptional corepressors. BioEssays, 25, 9–12. 2. Turner, J. & Crossley, M. (2001). The CtBP family: enigmatic and enzymatic transcriptional co-repressors. BioEssays, 23, 683–690. 3. Shi, Y., Sawada, J., Sui, G., Affar el, B., Whetstine, J. R., Lan, F. et al. (2003). Coordinated histone modifications mediated by a CtBP co-repressor complex. Nature, 422, 735–738. 4. Shi, Y., Lan, F., Matson, C., Mulligan, P., Whetstine, J. R., Cole, P. A. et al. (2004). Histone demethylation mediated by the nuclear amine oxidase homolog LSD1. Cell, 119, 941–953. 5. Schaeper, U., Boyd, J. M., Verma, S., Uhlmann, E., Subramanian, T. & Chinnadurai, G. (1995). Molecular cloning and characterization of a cellular phosphoprotein that interacts with a conserved C-terminal domain of adenovirus E1A involved in negative modulation of oncogenic transformation. Proc. Natl. Acad. Sci. USA, 92, 10467–10471. 6. Nardini, M., Spano, S., Cericola, C., Pesce, A., Massaro, A., Millo, E. et al. (2003). CtBP/BARS: a dual-function protein involved in transcription corepression and Golgi membrane fission. EMBO J. 22, 3122–3130. 7. Nardini, M., Svergun, D., Konarev, P. V., Spano, S., Fasano, M., Bracco, C. et al. (2006). The C-terminal domain of the transcriptional corepressor CtBP is intrinsically unstructured. Protein Sci. 15, 1042–1050. 8. Bernard, N., Johnsen, K., Gelpi, J. L., Alvarez, J. A., Ferain, T., Garmyn, D. et al. (1997). D-2-Hydroxy-4methylvalerate dehydrogenase from Lactobacillus delbrueckii subsp. bulgaricus. II. Mutagenic analysis of catalytically important residues. Eur. J. Biochem. 244, 213–219. 9. Kumar, V., Carlson, J. E., Ohgi, K. A., Edwards, T. A., Rose, D. W., Escalante, C. R. et al. (2002). Transcription corepressor CtBP is an NAD(+)-regulated dehydrogenase. Mol. Cell, 10, 857–869. 10. Balasubramanian, P., Zhao, L. J. & Chinnadurai, G. (2003). Nicotinamide adenine dinucleotide stimulates oligomerization, interaction with adenovirus E1A and an intrinsic dehydrogenase activity of CtBP. FEBS Lett. 537, 157–160. 11. Turner, J. & Crossley, M. (1998). Cloning and characterization of mCtBP2, a co-repressor that associates with basic Kruppel-like factor and other mammalian transcriptional regulators. EMBO J. 17, 5129–5140. 12. Phippen, T. M., Sweigart, A. L., Moniwa, M., Krumm, A., Davie, J. R. & Parkhurst, S. M. (2000). Drosophila C-terminal binding protein functions as a contextdependent transcriptional co-factor and interferes with both mad and groucho transcriptional repression. J. Biol. Chem. 275, 37628–37637. 13. Sutrias-Grau, M. & Arnosti, D. N. (2004). CtBP contributes quantitatively to Knirps repression activity in an NAD binding-dependent manner. Mol. Cell. Biol. 24, 5953–5966. 14. Nibu, Y., Zhang, H., Bajor, E., Barolo, S., Small, S. & Levine, M. (1998). dCtBP mediates transcriptional repression by Knirps, Kruppel and Snail in the Drosophila embryo. EMBO J. 17, 7009–7020. 15. Poortinga, G., Watanabe, M. & Parkhurst, S. M. (1998). Drosophila CtBP: a Hairy-interacting protein required for embryonic segmentation and hairy-mediated transcriptional repression. EMBO J. 17, 2067–2078.
CTBP-1, a THAP Domain-Containing CtBP Corepressor
16. Mani-Telang, P. & Arnosti, D. N. (2007). Developmental expression and phylogenetic conservation of alternatively spliced forms of the C-terminal binding protein corepressor. Dev. Genes Evol. 217, 127–135. 17. Folkers, U., Kirik, V., Schobinger, U., Falk, S., Krishnakumar, S., Pollock, M. A. et al. (2002). The cell morphogenesis gene ANGUSTIFOLIA encodes a CtBP/ BARS-like protein and is involved in the control of the microtubule cytoskeleton. EMBO J. 21, 1280–1288. 18. Kim, G. T., Shoda, K., Tsuge, T., Cho, K. H., Uchimiya, H., Yokoyama, R. et al. (2002). The ANGUSTIFOLIA gene of Arabidopsis, a plant CtBP gene, regulates leafcell expansion, the arrangement of cortical microtubules in leaf cells and expression of a gene involved in cell-wall formation. EMBO J. 21, 1267–1279. 19. Cho, K. H., Shindo, T., Kim, G. T., Nitasaka, E. & Tsukaya, H. (2005). Characterization of a member of the AN subfamily, IAN, from Ipomoea nil. Plant Cell Physiol. 46, 250–255. 20. Tsukaya, H. (2007). A new member of the CtBP/BARS family from plants: Angustifolia. In CtBP Family Proteins (Chinnadurai, G., ed), pp. 112–118, Landes Bioscience, Georgetown, TX. 21. Liew, C. K., Crossley, M., Mackay, J. P. & Nicholas, H. R. (2007). Solution structure of the THAP domain from Caenorhabditis elegans C-terminal binding protein (CtBP). J. Mol. Biol. 366, 382–390. 22. Clouaire, T., Roussigne, M., Ecochard, V., Mathe, C., Amalric, F. & Girard, J. P. (2005). The THAP domain of THAP1 is a large C2CH module with zinc-dependent sequence-specific DNA-binding activity. Proc. Natl Acad. Sci. USA, 102, 6907–6912. 23. Cayrol, C., Lacroix, C., Mathe, C., Ecochard, V., Ceribelli, M., Loreau, E. et al. (2007). The THAP-zinc finger protein THAP1 regulates endothelial cell proliferation through modulation of pRB/E2F cell-cycle target genes. Blood, 109, 584–594. 24. Macfarlan, T., Parker, J. B., Nagata, K. & Chakravarti, D. (2006). Thanatos-associated protein 7 associates with template activating factor-Ibeta and inhibits histone acetylation to repress transcription. Mol. Endocrinol. 20, 335–347. 25. Macfarlan, T., Kutney, S., Altman, B., Montross, R., Yu, J. & Chakravarti, D. (2005). Human THAP7 is a chromatin-associated, histone tail-binding protein that represses transcription via recruitment of HDAC3 and nuclear hormone receptor corepressor. J. Biol. Chem. 280, 7346–7358. 26. Thio, S. S., Bonventre, J. V. & Hsu, S. I. (2004). The CtBP2 co-repressor is regulated by NADH-dependent dimerization and possesses a novel N-terminal repression domain. Nucleic Acids Res. 32, 1836–1847. 27. Quinlan, K. G., Verger, A., Kwok, A., Lee, S. H., Perdomo, J., Nardini, M. et al. (2006). Role of the C-terminal binding protein PXDLS motif binding cleft in protein interactions and transcriptional repression. Mol. Cell. Biol. 26, 8202–8213. 28. Perdomo, J., Verger, A., Turner, J. & Crossley, M. (2005). Role for SUMO modification in facilitating transcriptional repression by BKLF. Mol. Cell. Biol. 25, 1549–1559. 29. Abramoff, M. D., Magelhaes, P. J. & Ram, S. J. (2004). Image processing with ImageJ. Biophoton. Int. 11, 36–42. 30. Jia, Y., Xie, G., McDermott, J. B. & Aamodt, E. (1997). The C. elegans gene pag-3 is homologous to the zinc finger proto-oncogene gfi-1. Development, 124, 2063–2073. 31. Jafar-Nejad, H. & Bellen, H. J. (2004). Gfi/Pag-3/ senseless zinc finger proteins: a unifying theme? Mol. Cell. Biol. 24, 8803–8812.
CTBP-1, a THAP Domain-Containing CtBP Corepressor
32. Clark, S. G. & Chiu, C. (2003). C. elegans ZAG-1, a Znfinger-homeodomain protein, regulates axonal development and neuronal differentiation. Development, 130, 3781–3794. 33. Wacker, I., Schwarz, V., Hedgecock, E. M. & Hutter, H. (2003). zag-1, a Zn-finger homeodomain transcription factor controlling neuronal differentiation and axon outgrowth in C. elegans. Development, 130, 3795–3805. 34. Rimann, I. & Hajnal, A. (2007). Regulation of anchor cell invasion and uterine cell fates by the egl-43 Evi-1 proto-oncogene in Caenorhabditis elegans. Dev. Biol. 308, 187–195. 35. Postigo, A. A. & Dean, D. C. (1999). ZEB represses transcription through interaction with the corepressor CtBP. Proc. Natl Acad. Sci. USA, 96, 6683–6688.
11 36. Nitta, E., Izutsu, K., Yamaguchi, Y., Imai, Y., Ogawa, S., Chiba, S. et al. (2005). Oligomerization of Evi-1 regulated by the PR domain contributes to recruitment of corepressor CtBP. Oncogene, 24, 6165–6173. 37. Zhang, C. L., McKinsey, T. A., Lu, J. R. & Olson, E. N. (2001). Association of COOH-terminal-binding protein (CtBP) and MEF2-interacting transcription repressor (MITR) contributes to transcriptional repression of the MEF2 transcription factor. J. Biol. Chem. 276, 35–39. 38. Liew, C. K., Simpson, R. J., Kwan, A. H., Crofts, L. A., Loughlin, F. E., Matthews, J. M. et al. (2005). Zinc fingers as protein recognition motifs: structural basis for the GATA-1/friend of GATA interaction. Proc. Natl Acad. Sci. USA, 102, 583–588.