|Home | About | Journals | Submit | Contact Us | Français|
Transcriptional corepressors play complex roles in developmental gene regulation. These proteins control transcription by recruiting diverse chromatin-modifying enzymes, but it is not known whether corepressor activities are finely regulated in different developmental settings or whether their basic activities are identical in most contexts. The evolutionarily conserved C-terminal binding protein (CtBP) is recruited by a variety of transcription factors that play crucial roles in development and disease. CtBP contains a central NAD(H) binding core domain that is homologous to D2 hydroxy acid dehydrogenase enzymes, as well as an unstructured C-terminal domain. NAD(H) binding is important for CtBP function, but the significance of its intrinsic dehydrogenase activity, as well as that of the unstructured C terminus, is poorly understood. To clarify the biological relevance of these features, we established genetic rescue assays to determine how different forms of CtBP function in the context of Drosophila melanogaster development. The mutant phenotypes and specific gene regulatory effects indicate that both the catalytic site of CtBP and the C-terminal extension play important, if nonessential roles in development. Our results indicate that the structural and enzymatic features of CtBP, previously thought to be dispensable for overall transcriptional control, are critical for modulating this protein's activity in diverse developmental settings.
Developmental gene regulation features an elaborate interplay of transcription activation and repression. A variety of transcriptional cofactors are required in eukaryotes to potentiate the activity of sequence-specific transcription factors that affect these regulatory programs. Cofactors are often crucial for the alteration of chromatin structures at target genes; corepressor complexes involved in gene inactivation have various enzymatic activities, including nucleosome remodeling, histone deacetylation, methylation, and demethylation. Although the importance of corepressors in transcriptional repression has been appreciated at the cellular level, the biological functions of corepressors in the development of multicellular organisms are less well understood.
The C-terminal binding protein (CtBP) is an evolutionarily conserved transcriptional corepressor that plays crucial roles in development and disease (6, 7). Initially identified as a phosphoprotein associated with the C terminus of the adenovirus E1A protein, the CtBP dimer has been shown to interact with a variety of cellular transcription factors containing a PXDLS motif to regulate target genes implicated in multiple cellular processes (35). Drosophila melanogaster CtBP is utilized by transcriptional repressors, such as the Knirps, Krüppel, Giant, and Snail proteins, that are active in patterning the blastoderm embryo (23, 28). Structurally, CtBP shares extensive homology with d-2-hydroxy acid dehydrogenases, including the conserved NAD(H) binding domain and putative catalytic site, and the protein has weak in vitro dehydrogenase activity (15). Although CtBP active site residues are highly conserved, it is not clear whether the dehydrogenase activity participates in CtBP-mediated transcriptional repression, because most cell-based assays have shown that this enzymatic activity is dispensable for repression (12, 15, 19, 34). The NAD(H) binding activity, however, appears to be critical for CtBP repression activity (12, 15, 19, 34). Structural studies revealed that NAD(H) binding induces a conformational change in CtBP, which may explain the importance of this cofactor for the interaction of CtBP with cofactors and transcription factors (15, 40). An NAD(H) binding mutant is still capable of forming dimers in vitro, but there is evidence to suggest that NAD(H) binding can further stimulate CtBP dimerization (2, 19).
Similar to other corepressors that serve as scaffolds to recruit chromatin remodeling enzymes, CtBP diversity is increased by gene duplication, alternative splicing, and posttranslational modifications (25). In invertebrates and vertebrates, multiple CtBP isoforms are expressed in spatially and temporally differentiated patterns, suggesting specialization of function. In Drosophila, the single CtBP gene encodes two major isoforms that are generated through alternative RNA splicing (24, 28, 34). These isoforms, termed CtBPL and CtBPS, are highly conserved in insects (18). The proteins are identical at the N terminus but differ in the C-terminal region: CtBPL has an extended C terminus that is lacking in CtBPS. Tethered directly to DNA, these two isoforms exhibit similar repression activities but have distinct developmental expression profiles, suggesting that they may have overlapping or unique functions (18, 34). The vertebrate CtBP1 and CtBP2 genes also produce various isoforms via either RNA splicing or alternative promoter utilization (36). CtBP1-L and CtBP1-S are encoded by the CtBP1 gene through alternative splicing, while CtBP2 and RIBEYE are produced from the CtBP2 gene by alternative promoter usage. CtBP1-L and CtBP2 exhibit corepressor activities, while CtBP1-S may be involved in Golgi membrane fission and RIBEYE in the function of central nervous system synapses (29, 37).
The unstructured C-terminal domain is present in all vertebrate CtBP proteins and the CtBPL isoform in Drosophila (21). This region contains sites for posttranslational modifications, including phosphorylation and sumoylation, which may play regulatory roles: C-terminal phosphorylation of human CtBP1-L triggers CtBP1-L ubiquitylation and degradation, while sumoylation of this tail regulates CtBP1-L subcellular distribution by stimulating nuclear retention (17, 39, 41). A further level of regulation is mediated by acetylation of a lysine residue in the N terminus of mammalian CtBP2, which can control nuclear localization (44).
Similar to other corepressors, CtBP also forms complexes with multiple chromatin-modifying enzymes. Proteomic studies have identified histone deacetylase, histone methyltransferase, and histone demethylase in CtBP complexes, suggesting that CtBP works as a platform to recruit chromatin modifiers to the template to alter chromatin structure and repress target gene expression (30). CtBP may also play a more complex role than merely serving as a scaffold to bridge transcription factors and histone modifiers. The intrinsic dehydrogenase activity of CtBP may participate in transcriptional regulation, although the significance of this function has remained elusive. In addition, CtBP may also work as a sensor to monitor cellular redox status and regulate transcription accordingly (10, 40). Both NADH and NAD+ stimulate CtBP-repressor interaction, but the binding affinity for the reduced NADH cofactor has been measured to be >100-fold higher than the affinity for NAD+, which may permit CtBP to respond to changes in the balance of these two forms of dinucleotide, providing a coupling between metabolism and transcription (10, 15, 40). Kumar et al. and Balasubramanian et al., however, did not observe the different efficiencies of NADH and NAD+ in stimulating CtBP-E1A interaction in in vitro binding assays; thus, there is some uncertainty regarding relative binding affinities to the reduced and oxidized dinucleotides (2, 15).
Cell-based studies have shed light on the basic repression activities of CtBP, but the functions of CtBP in the context of development are less well understood. CtBP is an essential gene for development in Drosophila, as is the CtBP2 gene in mice (13, 23, 28). Murine CtBP1 mutants are viable but show growth defects, and in Caenorhabditis elegans, CtBP mutants do not die but, in fact, exhibit an extended life span (5, 13). These studies provide only a very general picture of CtBP functions on a global scale; they do not address how specific features of the protein may contribute qualitatively or quantitatively to gene regulation in specific contexts. To better understand the biological relevance of different features of CtBP in a developmental context, we generated genomic rescue constructs that gave us the ability to assay different CtBP forms expressed under the control of native cis regulatory elements in a CtBP-null background, providing the first such detailed dissection of CtBP function in a whole organism. The mutant phenotypes indicate that the putative dehydrogenase function, as well as the C-terminal regulatory domain, are in fact essential for the normal developmental activity of this corepressor.
The 16-kb annotated Drosophila CtBP gene region was divided into three pieces, named CtBP-1st, CtBP-2nd, and CtBP-3rd, and amplified from yw adult genomic DNA using an expand long range dNTPack kit (Roche). BamHI/AvrII, AvrII/BglII, and BglII/KpnI sites were designed into the 5′ and 3′ ends of CtBP-1st, CtBP-2nd, and CtBP-3rd, respectively. A pair of oligonucleotides containing BamHI-AvrII-BglII-KpnI sites were annealed and inserted into the BamHI/KpnI sites of the pBluescript vector for assembly (pBS-adaptor). The CtBP-2nd fragment was first ligated into the T-Easy vector (Promega) and used as a template to generate CtBP-CAT (H315Q) and CtBP-NAD (D204N) mutations (Fig. (Fig.11 D) via QuikChange site-directed mutagenesis (Stratagene). Two tandem Flag tags (2×Flag) were inserted, in frame, into the C termini of the last protein-coding regions of CtBPS and CtBPL using the QuikChange strategy. The modified fragments were assembled in pBS-adaptor in the order CtBP-2nd, CtBP-1st, and CtBP-3rd, and the whole gene region was then subcloned into the pattB vector (a gift from K. Basler, University of Zurich, Zurich, Switzerland) using the BamHI/KpnI sites (BglII/KpnI for the pattB vector). To generate the CtBPS construct, the splicing donor site at the end of exon 5 was mutated from GT to GA and exons encoding the C-terminal extension of CtBPL were deleted (Fig. (Fig.1C).1C). The CtBPL construct was generated by fusing the C-terminal-extension region to the C terminus of CtBPS directly (Fig. (Fig.1C1C).
CtBP03463 (P11590) and CtBP87De-10 (1663) mutant lines and Df(3R)Exel8157 (7973) and Df(3R)BSC615 (25690) deletion mutants were obtained from Bloomington Stock Center and used for rescue assays. CtBP genomic constructs were injected into line attp40 using the phiC31 integrase system at Genetic Services, Inc. The cytogenetic location of the attB site in line attp40 is 25C7.
Drosophila Schneider 2 (S2) cells were grown at 25°C in Schneider Drosophila medium (Gibco) containing 10% heat-inactivated fetal bovine serum (FBS), 50 units penicillin G, and 50 μg/ml streptomycin sulfate. For each transfection, ~1 × 106 cells were seeded per well in a six-well plate and transfected with 200 ng of each DNA using an Effectene transfection kit (Qiagen). Cells were grown for three additional days and lysed in radioimmunoprecipitation buffer (50 mM Tris-Cl, pH 8.0, 150 mM NaCl, 1.0% NP-40, 0.5% sodium deoxycholate, and 0.1% sodium dodecyl sulfate) supplemented with protease inhibitor cocktail (Roche). The protein concentration was measured by Bradford assay. An equivalent amount of the total protein was subjected to SDS-PAGE (NuPAGE Novex 4 to 12% bis-Tris gel) and analyzed by immunoblotting. Images and quantifications were acquired on a LAS-3000 imaging system (Fuji). The antibodies used in this study were mouse anti-Flag M2 antibody (Sigma, 1:10,000), rabbit anti-CtBP serum (1:10,000) (18), mouse anti-β-tubulin (1:5,000, Iowa Hybridoma Bank), and goat anti-mouse and goat anti-rabbit horseradish peroxidase-conjugated secondary antibodies (Pierce, 1:10,000).
Immunofluorescent staining of Drosophila embryos was performed as previously described (1). Briefly, fixed embryos were washed six times in 100% ethanol, once in xylene for 30 min, six times in 100% ethanol, four times in 50% methanol-50% PBT (1× phosphate-buffered saline plus 0.1% Tween 80), and four times in 100% PBT and then blocked once in blocking reagent (1% casein in maleic acid and PBT, 1:1 [vol/vol]) for 1 h at room temperature with rocking. The embryos were then incubated with primary antibody and fluor-conjugated secondary antibody in blocking reagent overnight at 4°C and mounted on slides. Samples were examined using an Olympus FluoView 1000 laser scanning confocal microscope-IX81 system at the Center for Advanced Microscopy at Michigan State University (MSU). The antibodies used were mouse anti-Flag M2 antibody (Sigma, 1:500), rabbit anti-CtBP serum (1:200), and Alexa Fluor 555 goat anti-mouse and Alexa Fluor 488 goat anti-rabbit (1:500; Invitrogen).
The wings were removed from adult flies, mounted onto a metal stub, coated with gold, and imaged with a JEOL 6400V scanning electron microscope (Japan Electron Optics Laboratories) at the Center for Advanced Microscopy at MSU.
To move a white mutant allele into the CtBP03463 stock, male flies (w+/Y; CtBP03463/TM3, Sb1Ser1) were crossed with Ser virgin females (w−/w−; +/TM3, Ser). Offspring carrying white and the CtBP03463 mutant alleles (w−/Y; CtBP03463/TM3, Ser and w−/w+; CtBP03463/TM3, Ser) were crossed inter se, and progeny with white eyes and the Ser balancer (w−/w−; CtBP03463/TM3, Ser and w−/Y; CtBP03463/TM3, Ser) were collected and maintained as new CtBP03463 stock. The same crosses were performed to remove the w+ gene in CtBP87De-10 stock. To conduct the CtBP rescue assay, transgenic flies carrying an individual transgene (Tg) were first crossed with TM3, Sb flies. Male offspring (Tg/+; +/TM3, Sb) were crossed with CtBP03463 virgin females (w−/w−; CtBP03463/TM3, Ser). Offspring containing the CtBP transgene and the CtBP03463 mutant allele (Tg/+; CtBP03463/TM3, Sb) were crossed inter se to assess their ability to rescue the CtBP null phenotype. The CtBP87De-10 mutant was similarly tested, but we could not rescue this stock with any rescue construct. Suspecting that this stock may harbor an additional recessive lethal mutation, we performed additional complementation analysis, placing CtBP87De-10 over CtBP03463 or over two CtBP deficiencies, Df(3R)BSC615 and Df(3R)Exel8157, and all of them could be rescued, suggesting that the CtBP87De-10 chromosome may contain additional recessive lethal mutations not related to CtBP.
Two- to 4-h embryos were collected, fixed, and hybridized with digoxigenin-UTP-labeled RNA antisense probe to eve (31).
Total RNA from rescued adult flies was extracted using an RNeasy mini kit (Qiagen) following the manufacturer's instructions, including the on-column DNase digestion step. Three micrograms of total RNA was in vitro transcribed into cDNA using a high-capacity cDNA reverse transcription kit (Applied Biosystems) with RNase inhibitor in a 40-μl volume. cDNA was then diluted to 25 ng/4 μl, and 25 ng of total cDNA was used in each 16-μl real-time PCR reaction mixture. The final concentration of each primer was fixed to 250 nM. A SYBR green PCR master mix kit (Applied Biosystems) was used in this study, and real-time PCR was performed on an Applied Biosystems 7500 real-time PCR system. The PCR conditions were 1 cycle of 50°C for 2 min, 1 cycle of 95°C for 10 min, and 40 cycles of 95°C for 15 s and 60°C for 1 min. A dissociation step was added to determine the primer specificity. The amplification efficiency of each primer set was calculated by the standard curve method, using 5 series of 5-fold dilutions (25 ng/4 μl, 5 ng/4 μl, 1 ng/4 μl, 0.2 ng/4 μl, and 0.04 ng/4 μl) (26). Putative CtBP direct targets were selected based on physical binding of CtBP to the gene promoter (Drosophila CtBP chromatin immunoprecipitation-microchip [ChIP-chip] assay data, http://intermine.modencode.org/release-18/objectDetails.do?id=211000586). Gene transcripts were measured in three biological replicates, with two technical replicates for each sample, and relative gene expression was determined using the Relative Expression Software Tool (27) with β-tubulin56D (9) as the reference gene.
The single Drosophila CtBP gene produces multiple isoforms through RNA splicing (28, 34). Two major isoforms are expressed throughout development, with CtBPS expressed at higher levels than CtBPL. To faithfully reproduce endogenous CtBP expression levels and patterns, we generated a 16-kb rescue construct containing 8 kb 5′ and 1 kb 3′ of the central coding exons, including a large, 3-kb intron that is spliced out to form the CtBPL isoform. A C-terminal Flag epitope tag was used to track the expression of CtBP transgene products. Five transgene constructs, all inserted at the same locus to eliminate position effects, were assayed to determine the activities of CtBP functional elements (Fig. (Fig.1).1). A wild-type construct, CtBP-WT, preserves the exon and intron structure, as do two mutant forms, CtBP-CAT and CtBP-NAD, with single point mutations eliminating catalytic activity or NAD(H) binding (2, 15, 16). To eliminate catalytic activity, the critical catalytic-active-site histidine was mutated to glutamine in the CtBP-CAT mutant, while NAD(H) binding was disrupted by an aspartic acid-to-asparagine mutation in the CtBP-NAD mutant (Fig. (Fig.1).1). These mutations have been described for previous forms of CtBP (16, 19, 34). A previously described CtBP-NAD mutation affecting conserved glycines present in the NAD(H) binding Rossmann fold (GXGXXG) was not used in this study because of its strong destabilizing effect (19). A fourth gene, CtBPS, produces only the abundant short isoform, while CtBPL joins the alternatively spliced 3′ exon directly to the main coding region to produce this isoform at levels similar to that of the more abundant endogenous CtBPS isoform.
The expression of these genes was first assayed by transient transfection in Drosophila S2 cells followed by Western blotting with anti-Flag M2 antibody to detect the transgenic products. Both the CtBPL and the CtBPS isoform were detected in CtBP-WT, CtBP-CAT, and CtBP-NAD transfections, with higher levels of the short isoform, as expected (Fig. (Fig.22 B, lanes 2, 5, and 6). Similarly, the CtBPL construct produced only the long isoform, while the CtBPS expressed only the short variant (Fig. (Fig.2B,2B, lanes 3 and 4). Consistent with these results, transgenic flies carrying CtBP-WT, CtBP-CAT, and CtBP-NAD transgenes expressed both the CtBPL and the CtBPS isoform, whereas only CtBPL or CtBPS was detected in CtBPL or CtBPS transgenic flies (Fig. (Fig.2C).2C). To compare the expression levels of the exogenous CtBP proteins to those of the endogenous CtBP isoforms, extracts from lines expressing the proteins were analyzed with anti-CtBP antibody that recognizes both endogenous and recombinant proteins, allowing direct comparison of the levels of these proteins (Fig. (Fig.2D).2D). The amounts of exogenous CtBP isoforms were quantified and normalized to the amounts of endogenous CtBP proteins in each sample. As shown in Fig. Fig.2E,2E, the CtBP-WT, CtBPL, CtBPS, and CtBP-CAT transgenic lines produced amounts of proteins similar to the endogenous levels, while the CtBP-NAD expression level was roughly one-half of this level, consistent with a lower stability of this mutant form (19).
Endogenous CtBP is found primarily in the nucleus, consistent with its role as a transcription corepressor. Accordingly, a nuclear localization signal (NLS) was identified in the N terminus of Drosophila CtBP (36). To examine whether NAD(H) binding and the dehydrogenase activity affect protein localization, transgenic Drosophila embryos were stained with anti-CtBP or anti-Flag antibodies to visualize CtBP proteins. CtBPL, CtBPS, and CtBP-CAT (Fig. (Fig.33 B, C, and D) were all localized to the nucleus, as was endogenous CtBP (Fig. (Fig.3A).3A). Protein was detected in both the nucleus and cytoplasm in CtBP-NAD transgenic Drosophila embryos that also have endogenous CtBP (Fig. (Fig.3E)3E) and largely in the cytoplasm in embryos expressing only the NAD(H) binding mutant form of CtBP (Fig. (Fig.3F).3F). We suspect that the partial nuclear localization observed in the presence of endogenous CtBP is directed by heterodimerization between wild-type CtBP and CtBP-NAD mutant proteins; apparently NAD(H) binding is essential for CtBP nuclear localization.
We next carried out genetic rescue assays to test the ability of CtBP isoforms and mutants to rescue a CtBP null mutant, CtBP03463, to assess their functions in a developmental context (23, 28). CtBP03463 is a P element-induced homozygous lethal mutant, and the homozygotes die as pharate adults in the pupal case (24, 28). Different forms of CtBP transgenes were introduced into CtBP homozygous mutants to first test whether they could functionally substitute for zygotic CtBP (Fig. (Fig.4A).4A). In the presence of rescue constructs, a considerable number of adult flies survived despite being homozygous for the CtBP mutation (Fig. (Fig.4B).4B). One or two copies of the CtBPL or CtBPS transgene were sufficient to rescue CtBP lethality. A smaller percentage of flies were rescued by the CtBP-CAT transgene (P < 0.0001; z test), while no complementation at all was observed for the CtBP-NAD transgene (Fig. (Fig.4B4B).
To verify that these rescued flies were indeed homozygous for the CtBP mutation, extracts from adults were analyzed by Western blotting with anti-CtBP antibody. The Flag-tagged forms of CtBP protein, which migrate somewhat more slowly than the endogenous proteins, were observed in all rescued adult flies. A weak band migrating at the size of endogenous CtBPS was also seen in some cases (Fig. (Fig.4C,4C, lanes 2 and 3), which may be a degradation product of the transgene or low levels of the endogenous protein. The absence of this product in preparations from many individual CtBP rescue flies suggests that this is a proteolytic product of the recombinant protein, and low to nonexistent levels of endogenous CtBP are present in these flies (data not shown). To test whether low levels of endogenous CtBP were contributing to rescue, we repeated the rescue with the CtBP87De-10 mutant allele (a null or stronger hypomorphic allele), which was placed over one of two CtBP deletions (28, 32). The same rescue results were obtained, indicating that the rescue activity of these transgenes is not allele specific but can be observed in different CtBP mutant backgrounds (crosses described in Materials and Methods; data not shown).
To eliminate the maternal CtBP contribution and dissect the functions of CtBP variants more stringently, the rescued adult flies were crossed inter se so that the females, and consequently the oocytes, also contained no endogenous CtBP protein, and embryo viabilities were assayed. As shown in Fig. Fig.4D,4D, the hatch rates of embryos rescued with the CtBP-WT and CtBPS transgenes were 69% and 70%, respectively, similar to that of wild-type embryos (yw, 87%), while only 43% and 19% of those rescued with the CtBPL and CtBP-CAT transgenes survived (P < 0.0001; z test), demonstrating a significantly weaker ability to rescue. Introducing one copy of the CtBPL transgene into the CtBPS background did not increase the hatch rate, suggesting that this combination of the two proteins did not provide a more potent activity (data not shown).
CtBP-WT, CtBPL, CtBPS, and CtBP-CAT rescued CtBP03463 lethality, allowing development to proceed to adulthood, but their abilities to provide endogenous CtBP function varied. Apart from the embryonic lethality described above, the main observable phenotype was a marked effect on wing development. A variety of wing defects were observed in zygotic rescued adults, ranging from mildly curly wings to blistered wing phenotypes and, in some cases, severely reduced wings (Fig. (Fig.5A,5A, ii to v). A spectrum of phenotypes was observed in all rescue assays; however, the CtBPL rescued adults exhibited significantly stronger wing phenotypes than did the CtBP-WT and CtBPS lines (Fig. 5B and C). The CtBP-CAT mutant showed an even higher penetrance and consistently exhibited the most severe phenotypes, particularly when its partial lethality was taken into account (Fig. (Fig.4B4B and 5B and C). Scanning electron microscopy images of normal and abnormal wings showed that the organization of wing epithelial cells was disrupted (Fig. (Fig.5A,5A, viii), supporting the idea that CtBP may regulate epithelial gene expression (12). The same wing phenotypes and trends were observed in rescued flies that had been bred for several generations so that there was no maternal contribution of endogenous CtBP and the sole source of CtBP protein was the transgene (data not shown). Interestingly, although the CtBP-NAD mutant failed to rescue and was thus not testable in a CtBP mutant background, when placed in a genetic background containing only one copy of the endogenous CtBP gene, it induced a notched wing phenotype (Fig. (Fig.5A,5A, vi). This phenotype appeared in a dosage-dependent manner; 34% of CtBP heterozygous mutant flies carrying two copies of CtBP-NAD showed notched wings, while no phenotype was observed when these transgenes were present in a wild-type background or if only one copy of CtBP-NAD was present.
Severe depletion of CtBP has been observed to affect bristle development in the adult, through the corepressor's involvement in E(spl)-mediated gene expression in sensory neurons (28, 32, 33). Consistent with these reports, we also observed missing and extra macrochaete bristle phenotypes in CtBP rescued flies, especially with the CtBP-CAT and CtBPL lines (data not shown). Thus, the macroscopic morphological defects associated with the partially functional CtBP alleles are not limited to wing development.
CtBP functions primarily as a transcriptional corepressor; therefore, we speculated that the impaired embryonic viability and wing phenotypes that we observed in CtBP rescued embryos and adult flies may be associated with defects of its repression activity. even-skipped (eve) is a CtBP target whose expression pattern is disrupted in a CtBP mutant (23, 28). We examined the embryonic eve expression pattern in CtBP rescue lines and observed that about 15% of CtBP-WT, CtBPL, and CtBPS rescued embryos showed fusion of eve strips 2/3 and 4 to 6 (Fig. (Fig.6A,6A, ii, and B), similar to aberrant eve expression in a Krüppel (Kr) mutant (23). This phenotype is considerably elevated in CtBP-CAT rescued embryos (35%), and CtBP-NAD embryos, which are not rescued, were even more severely affected (45%), indicating that CtBP activity was affected more by the catalytic and NAD(H) binding mutations. Considering that the CtBP-NAD failed to rescue and, thus, the embryos we examined were a mixture of transgenic and mutant forms, the percentage is likely to be underestimated. In CtBP-CAT rescued embryos specifically, however, we observed a novel pattern of eve disruption, showing reduction of stripes 3 to 6 (Fig. (Fig.6C),6C), which is reminiscent of eve patterns in a knirps (kni) mutant, suggesting that the Kr and kni repression activities may be affected by the catalytic mutant (23).
To test whether genes in other stages are affected by different forms of CtBP, we performed quantitative PCR to measure the transcripts of selected genes that are bound in vivo by CtBP (3, 4). In adult flies, we found that the expression of mmp1, a potential CtBP target that has CtBP bound around the promoter, was increased more than 3-fold in CtBPL lines (Fig. (Fig.6D),6D), with somewhat smaller effects in CtBP-CAT and CtBPS lines. Rel, another potential CtBP target, was induced more than 2-fold specifically in CtBPL mutants. The starvin (stv) gene was upregulated 4-fold in the CtBP-CAT mutant, with a smaller but still significant level of upregulation in the CtBPL line (Fig. (Fig.6D).6D). Expression in the CtBP-WT and CtBPS lines was minimally affected. These effects were reproducibly observed in multiple biological experiments, indicating that the C terminus and dehydrogenase activity are involved in CtBP-mediated repression in this context. For these genes, we observed loss of repression, consistent with loss of corepressor function in CtBP-CAT and CtBPL backgrounds; however, another CtBP target, prd, showed reduced expression in the CtBPL background, suggesting that there may be context-specific effects that allow CtBP to function in a stimulatory manner in some cases.
CtBP is recruited by diverse transcription factors to effect repression of target genes in numerous metazoan regulatory pathways. Reflecting its evolutionary relatedness to dehydrogenases, CtBP has a functional NAD(H) binding cleft that is integral for overall protein structure, as well as for interaction with transcription factors and cofactors. In addition, CtBP contains active-site residues that are both evolutionarily conserved and confer in vitro dehydrogenase activity. In this study, we utilized an in vivo developmental assay to address key questions about the function of CtBP. In a genomic rescue assay that comprehensively tests biological function, we show that the residues required for enzymatic activity, as well as the C-terminal regulatory domain, are essential for normal function in development (Fig. (Fig.55 and and6).6). Interestingly, animals programmed solely with the catalytic mutant protein or the form containing the C-terminal regulatory extension had significantly impaired viability and strong wing phenotypes. Specific transcriptional defects were also evident; embryonic eve expression was severely disrupted in these mutant embryos and the expression levels of particular targets in adults were derepressed (Fig. (Fig.6).6). The overall phenotypes of these mutants were for the most part enhanced manifestations of effects noted with less penetrant alleles, suggesting that loss of enzymatic function or too-extensive provision of the C-terminal extension reduced CtBP activity. The effects noted here are likely to represent a significant loss of activity, because CtBP function exhibits a considerable degree of robustness: the CtBP gene is recessive, showing no phenotype with ~50% of normal dosage, and wild-type flies carrying extra copies of these transgenes (up to ~200% of normal CtBP levels) are unaffected. Despite the lack of overt phenotypes for catalytic or C-terminal variants when expressed in cell-based assays, our analysis clearly indicates that the catalytic activity and C-terminal extension are important features that regulate CtBP function.
The results illuminate the overall functional understanding of CtBP, leading to a model that explains the roles of NAD(H) binding and catalysis (Fig. (Fig.7).7). Previous results demonstrated that NAD(H) binding is important for folding of CtBP and influences dimerization, binding of cofactors, and interaction with transcription factors (2, 15, 16, 22, 36, 40, 43). NAD(H) has been suggested to do more than influence the structure of CtBP complexes, however. The observation that the CtBP-NAD mutant largely localizes in the cytoplasm suggests that NAD(H) binding might also regulate the protein's subcellular localization (Fig. (Fig.3F).3F). In agreement with this hypothesis, mammalian CtBP1-NAD mutants also showed improper nuclear localization in MEF90 (CtBP−/−) cells (16). In addition, because of a reported higher affinity of the corepressor for the reduced NADH dinucleotide over NAD+, CtBP has been suggested to serve as a mediator that links cellular redox status to transcriptional output, and treatment of cells with agents that affect NADH levels can influence CtBP-mediated repression and occupancy of promoters (11, 14, 40, 42).
Building on these insights, an essential additional question is how a putative dehydrogenase activity may function in transcriptional control. Here, we have little in the way of precedent to go on; glyceraldehyde-3-phosphate dehydrogenase (GAPDH) has been implicated in transcriptional regulation as an essential component of the OCA-S/Oct-1 coactivator complex, where it plays a critical role in S-phase activation of histone H2B expression (45). This dehydrogenase is also proposed to function as a redox sensor, monitoring changes during S phase, but how dehydrogenase activity itself affects transcription in this case remains a mystery (8, 38). With respect to CtBP, numerous cell-based studies have shown that the dehydrogenase-defective mutant remains active, suggesting that the enzymatic activity is dispensable for CtBP function in these settings (12, 19, 34). These previous studies were not designed to test whether CtBP dehydrogenase activity is important only under certain physiological conditions or on specific target genes, however. Our genomic rescue data strongly suggest that the dehydrogenase activity is important for proper development and is involved in gene regulation. Two models would explain this function: possibly in response to metabolic signals, the enzymatic activity may be important as a mechanism to interconvert the reduced and oxidized forms of NAD, altering CtBP's structure and, thus, interaction with cofactors, or the CtBP dehydrogenase activity may be important for the reduction or oxidation of specific as-yet-unknown substrates important for gene regulation (Fig. (Fig.7).7). We cannot rule out the possibility that the mutation in the catalytic site also affect CtBP interactions with cofactors, but the results of structural and mutational studies indicate that the catalytic site is located in a buried cleft, separate from the cofactor binding domain (20).
An additional area of CtBP biology concerns possible regulation through the C-terminal region. Diverse metazoan CtBP proteins feature evolutionarily divergent C-terminal extensions that are dispensable for corepressor activity. Indeed, although Drosophila genes and those of other arthropods encode a conserved C-terminal region as an alternatively spliced exon, the major isoform lacks this domain (18). Previous studies have indicated that the C-terminal domain is subject to posttranslational modifications through sumoylation and phosphorylation that may regulate CtBP through alternative cellular localization and degradation (17, 39, 41). Here, we show that mutant animals that only express the CtBPL isoform with the C-terminal extension are viable but exhibit impaired viability, defective wings, and gene expression defects, clearly indicating the functional importance of this domain (Fig. (Fig.4D4D and 5B and C). It appears that equipping all CtBP molecules with this extension reduces CtBP activity, suggesting that this domain serves a negative regulatory function, possibly as a recipient of posttranslational modifications that may reflect signal transduction activity. In total, these results indicate that CtBP may function as a nexus of signal integration, responding to metabolic status through NAD(H) binding and to signaling pathways through the C terminus to affect transcriptional levels. An important question is whether this signaling would affect CtBP function generally or in a gene-specific manner. The differential effects of the CtBP-CAT mutant and CtBPL on eve, mmp1, Rel, prd, and stv (Fig. 6C and D) provide some support for the latter possibility. A genome-wide comparison of expression profiles between CtBP-WT and CtBP-CAT mutants, as well as CtBPL mutants, will provide important insights on this question.
In light of the involvement of CtBP in numerous regulatory events in Drosophila, including gap gene repression function in the blastoderm embryo, development of the peripheral nervous system, and recruitment in Notch, wingless, and transforming growth factor β (TGF-β) signaling pathways, it was initially quite surprising that the prevalent phenotypes we observed in different CtBP rescues were specific for wings. However, further characterization of impaired embryo viability and disrupted eve expression in embryos, as well as bristle phenotypes and altered expression of specific genes in adult flies, clearly shows that the effects are not limited to wings. It is likely that changes in CtBP activity produce various effects in different tissues and that those individuals with the strongest effects in nonwing tissues do not survive.
In summary, we have demonstrated that the dehydrogenase activity of CtBP and the C-terminal domain are important for fine-tuning CtBP function in the context of development. A detailed biochemical characterization of how these activities are integrated remains to be elucidated; important clues will likely come from a comprehensive picture of how possible modifications and metabolic signals are focused through CtBP to affect gene expression at a genomic level.
We thank K. Basler (University of Zurich, Zurich, Switzerland) for providing the pattB vector, W. D. Fakhouri, R. Sayal, and L. Zhang for assistance with confocal microscopy, I. Dworkin for fly wing imaging, the Bloomington Stock Center (Indiana) for fly stocks, anonymous reviewers for comments on the manuscript, and members of the Arnosti lab for useful discussions.
This work was supported by NIH grant no. GM56976 to D.N.A.
Published ahead of print on 15 November 2010.