The human candidate tumor suppressor gene HIC1 recruits CtBP through a degenerate GLDLSKK motif

Abstract

HIC1 (hypermethylated in cancer) and its close relative HRG22 (HIC1-related gene on chromosome 22) encode transcriptional repressors with five C(2)H(2) zinc fingers and an N-terminal BTB/POZ autonomous transcriptional repression domain that is unable to recruit histone deacetylases (HDACs). Alignment of the HIC1 and HRG22 proteins from various species highlighted a perfectly conserved GLDLSKK/R motif highly related to the consensus CtBP interaction motif (PXDLSXK/R), except for the replacement of the virtually invariant proline by a glycine. HIC1 strongly interacts with mCtBP1 both in vivo and in vitro through this conserved GLDLSKK motif, thus extending the CtBP consensus binding site. The BTB/POZ domain does not interact with mCtBP1, but the dimerization of HIC1 through this domain is required for the interaction with mCtBP1. When tethered to DNA by fusion with the Gal4 DNA-binding domain, the HIC1 central region represses transcription through interactions with CtBP in a trichostatin A-sensitive manner. In conclusion, our results demonstrate that HIC1 mediates transcriptional repression by both HDAC-independent and HDAC-dependent mechanisms and show that CtBP is a HIC1 corepressor that is recruited via a variant binding site.

FIG. 1.

Identification in the HIC1 and HRG22 proteins of an evolutionary conserved GLDLSKK/R motif related to the CID consensus. (Left) Schematic drawing of the HIC1 and HRG22 proteins from various species. Hu, human; Ck, chicken; Zf, zebra fish. Degrees of conservation (percentages of identity and similarity) between the human HRG22 and HIC1 proteins obtained with the BLAST program are indicated. (Right) The evolutionarily conserved GLDLSKK/R motif can be aligned with CtBP-binding motif PXDLSXK/R, originally found in the Ad E1A proteins and later in a still-growing list of proteins from various species. Notably, the proline residue (see also Fig. 4) is replaced by a glycine (underlined) in HIC1 and HRG22. Mu, murine; Xe, Xenopus laevis.

FIG. 2.

mCtBP1 specifically interacts with the central region of HIC1 in yeast two-hybrid assays. The baits contain the Gal4 DNA-binding domain fused to the HIC1 central region (CR) or BTB/POZ domain. The preys contain the activation domain of either VP16 (fused to mCtBP1) or Gal4 (fused to HIC1). The baits and preys were transfected in L40a yeast cells. -, pGAD vector. The transcriptional activation levels attained by the bait-prey interactions were estimated by measuring the β-galactosidase (β-gal) activity in liquid cultures of individual yeast colonies. Mean values and standard deviations from two independent experiments are shown.

FIG. 3.

mCtBP1 specifically interacts with the central region of HIC1 through the conserved GLDLSKK motif in mammalian two-hybrid assays. (A) Schematic structures of the Gal4 DNA-binding domain (construct 1) and of various Gal4-HIC1 chimeras (constructs 2 to 7). Numbering refers to HIC1 residues. Hatched box, GLDLSKK motif; black ovals, five zinc fingers. Gal4-NC4 (construct 8), containing the CID motif of the Net transcription factor, was used as a positive control (9). (B) The GLDLSKK motif is required for the interaction with mCtBP1. Luc and β-galactosidase assays were performed on total extracts from RK13 cells that had been transiently transfected with 750 ng of the pG5Luc vector (Promega), 50 ng of the pSG5 β-galactosidase construct as a control of transfection efficiency, 100 ng of the indicated Gal4 construct, and 100 ng of the VP16 activation domain (white bars) or VP16 activation domain-tagged mCtBP1 (grey bars). After normalization to β-galactosidase activity, the data were expressed as Luc activity relative to the activity of pG5Luc with empty control vectors, which was given an arbitrary value of 1. Results presented are the mean values and standard deviations from two independent transfections in triplicate.

FIG. 4.

Definition of a new CID consensus. The new consensus binding site for CtBP is shown according to the nomenclature first proposed by Postigo and Dean (45), who used outlined, boldface, and plain characters of different sizes based on the frequency of the residue. The leucine (outlined) remains the only invariant residue in the core motif. The previously invariant proline can be replaced by a glycine (underlined) as shown in this study for the naturally occurring HIC1 and HRG22 proteins and as previously suggested by in vitro assays with synthetic peptides (35). V∗, potential variant residue recently identified. Indeed, 6 out of 41 CtBP-binding partners cloned from a mouse embryonic library in a yeast two-hybrid screen contain a VLDLS motif, but the functionality of this motif still has to be demonstrated (54). Similarly, short-range Drosophila repressor Giant interacts with dCtBP but indirectly through an unknown bZIP protein, whereas the Giant VLDLSRR motif could recruit an as yet unidentified corepressor (40).

FIG. 5.

HIC1 interaction with mCtBP1 in eukaryotic cells requires the GLDLSKK motif. (A) mCtBP1 interacts with Flag-HIC1 but not with Flag-HIC1 ΔGLDLSKK in the eukaryotic GST pull-down assay. Cos-1 cells were cotransfected with the indicated combination of expression vectors. Cell extracts were analyzed by the GST pull-down assay. Immunoblot analysis was used to detect the proteins retained on the beads by protein-protein interaction (top) with the GST proteins (bottom; 5 μl of each total cell extract; input). This blot was stripped and reprobed (middle) with the anti-Flag M2 monoclonal antibody to detect epitope-tagged HIC1 proteins. WB, Western blot. (B) In vivo interaction of Flag-HIC1 and Flag-HIC1 mutants with mCtBP1. Cos-7 cells were mock transfected (lane 1) or transfected with the indicated expression vectors. Flag-tagged proteins were immunoprecipitated (IP) from cell lysates with the anti-Flag M2 monoclonal antibody. The resulting immunoprecipitates were then Western blotted and analyzed with the anti-CtBP1 rabbit polyclonal antibody (top). ∗, nonspecific band immunoprecipitated by the M2 antibody in each extract under the conditions used. Five microliters of each total cell extract (input) was resolved by SDS-PAGE and immunoblotted with the anti-Flag antibody to control for HIC1 protein expression (middle). This blot was stripped and probed with the rabbit anti-CtBP1 polyclonal antibody to ascertain the presence of the exogenous mCtBP1.

FIG. 6.

HIC1 and mCtBP1 colocalize at nuclear dots in transfected Cos-1 cells. (a to c) Transfected Cos-1 cells were labeled with the anti-Flag M2 monoclonal antibody (a and b) or a rabbit anti-CtBP1 polyclonal antibody (c). Flag-HIC1 (a) and the Flag-HIC1 ΔGLDLSKK mutant (b) have punctate nuclear localizations, whereas mCtBP1 (c) has a diffuse nuclear and cytoplasmic localization. (d to f) When HIC1 and CtBP1 were cotransfected, HIC1 (d) recruited CtBP (e) onto nuclear dots (as shown in panel f by the merge). Note the presence in this section of a cell not transfected by Flag-HIC1, where the ectopically expressed mCtBP1 exhibits its typical nuclear and cytoplasmic diffuse pattern (e and f, bottom). (g to i) Flag-HIC1 ΔGLDLSKK mutant exhibits a punctate nuclear localization (g) but is unable to recruit mCtBP1 (h) onto these nuclear dots (i, merge).

FIG. 7.

Inducible HIC1 can recruit endogenous CtBP. (A) Characterization of the stable inducible EcRCHO-pINDFlag-HIC1 clone 6 cell line. EcR-CHO pIND-Flag-HIC1 clone 6 cells were untreated (−) or treated with 10 μM ponasterone A for 48 h (+) and analyzed by conventional immunofluorescence microscopy with the anti-HIC1 PAb325 polyclonal antibody (left) or the anti-CtBP E12 monoclonal antibody (middle). Hoechst staining of the same field is shown on the right. Note the weak expression of HIC1 in a few uninduced cells, probably due to the leakiness of the promoter. (B) Endogenous nuclear CtBP proteins can interact with HIC1 in nuclear extracts prepared from the stable inducible EcRCHO-pIND-Flag-HIC1 clone 6 cell line. Nuclear extracts were prepared as described previously (41) from the untreated EcR-CHO pIND-Flag-HIC1 inducible cell line (−; lanes 1, 3, 5, and 7) or from cells treated with 10 μM ponasterone A for 48 h (+; lanes 2, 4, 6, and 8). Aliquots were immunoprecipitated with the indicated rabbit preimmune serum (lanes 3 and 4) or with two distinct anti-HIC1 immune sera directed against the HIC1 C-terminal region (lanes 5 to 8). The immunoprecipitates, 10% of each nuclear extract (input; lanes 1 and 2) and 3 μl of a mCtBP1-programmed reticulocyte lysate as a control (lane 9) were resolved by SDS-PAGE and immunoblotted with the anti-Flag M2 monoclonal antibody (top). Notably, a small amount of HIC1 can be detected in the uninduced cells (lanes 1, 5, and 7), presumably due to the leakiness of the hsp promoter. The blot was stripped and probed with the anti-CtBP monoclonal antibody (bottom) to detect the interaction with endogenous CtBP (lanes 6 and 8). The band detected in lanes 3 and 4 by the anti-CtBP monoclonal antibody is not endogenous CtBP, as clearly shown by its distinct migration in SDS-PAGE (compare with lanes 6, 8, and 9). It rather corresponds to a nonspecific band brought down by the polyclonal anti-HIC1 preimmune rabbit serum under these experimental conditions. Interestingly, the small amount of HIC1 protein present in the uninduced cells is able to coimmunoprecipitate a proportionally smaller amount of endogenous CtBP (lanes 5 and 7).

FIG. 8.

The HIC1 central region contains two TSA-sensitive repression domains. (Left) The Gal4 DNA-binding domain (construct 1) and the various Gal4-HIC1 chimeras (constructs 2 to 4) used in this assay are shown. Numbering refers to HIC1 residues. Hatched box, GLDLSKK motif. (Right) RK13 cells were transiently transfected in triplicate with 200 ng of the indicated constructs and 750 ng of the pG5Luc reporter. The cells were treated 24 h later with 300 nM TSA (dissolved in dimethyl sulfoxide [DMSO]) (hatched bars) or mock treated with an equal volume of DMSO (-, white bars) for a further 24 h before harvesting. The Luc activity was normalized to the β-galactosidase activity of a cotransfected pSG5 β-galactosidase construct (50 ng). After normalization to β-galactosidase, the data were expressed as Luc activity relative to the activity of pG5Luc with empty control vectors, which was given an arbitrary value of 1. The results are the mean values and standard deviations from two independent transfections in triplicate.

FIG. 9.

The full-length HIC1 protein and the ΔGLDLSKK mutant can associate with endogenous HDAC activity. (A) Cos-1 cells were transfected with expression vectors encoding either Flag-HIC1 or Flag-HIC1 ΔGLDLSKK. Flag-BCL-6, which can recruit HDAC-1, and Flag-HDAC-1 itself were used as positive controls, whereas cells transfected with the empty Flag expression vector (−) were used as a negative control to measure background levels. The HDAC activities coimmunoprecipitated with the Flag-HIC1 and Flag-HIC1 ΔGLDLSKK proteins were measured in duplicate from two independently transfected plates, with the standard deviations indicated. (B) To demonstrate the coretention of bona fide deacetylase activity, an experiment similar to that in panel A was performed, except that the immunoprecipitates were divided into two aliquots. One was left untreated (grey bars), and the other was incubated in the presence of 300 nM TSA (black bars) before the enzymatic assay.