Role of the C-terminal binding protein PXDLS motif binding cleft in protein interactions and transcriptional repression

Abstract

C-terminal binding proteins (CtBPs) are multifunctional proteins that can mediate gene repression. CtBPs contain a cleft that binds Pro-X-Asp-Leu-Ser (PXDLS) motifs. PXDLS motifs occur in numerous transcription factors and in effectors of gene repression, such as certain histone deacetylases. CtBPs have been depicted as bridging proteins that self-associate and link PXDLS-containing transcription factors to PXDLS-containing chromatin-modifying enzymes. CtBPs also recruit effectors that do not contain recognizable PXDLS motifs. We have investigated the importance of the PXDLS binding cleft to CtBP's interactions with various partner proteins and to its ability to repress transcription. We used CtBP cleft mutant and cleft-filled fusion derivatives to distinguish between partner proteins that bind in the cleft and elsewhere on the CtBP surface. Functional assays demonstrate that CtBP mutants that carry defective clefts retain repression activity when fused to heterologous DNA-binding domains. This result suggests that the cleft is not essential for recruiting effectors. In contrast, when tested in the absence of a fused DNA-binding domain, disruption of the cleft abrogates repression activity. These results demonstrate that the PXDLS binding cleft is functionally important but suggest that it is primarily required for localization of the CtBP complex to promoter-bound transcription factors.

FIG. 1.

Models of CtBP dimers interacting with PXDLS partners. (A) Model of CtBP as a bridging molecule linking a transcription factor (TF) and an effector. An elongated dimer of CtBP as a link between a promoter-bound transcription factor containing a PIDLS motif and the effector molecule HDAC4, which also contains a PIDLS motif. (B) Model where a CtBP dimer is localized to a promoter by binding two PIDLS motif-containing transcription factors and then recruits effectors through a different surface.

FIG. 2.

Generation of CtBP2 derivatives with defective PIDLS binding clefts. CtBP2 derivatives containing either the A58E or the V72R mutation or with a PIDLS extension or a ΔDL mutant PIDLS motif extension were tested for the abilities to dimerize with wild-type CtBP2 and to interact with a recognized PIDLS motif-containing partner protein, BKLF. Yeast growth was monitored, and the results reflecting association of the proteins are shown. (A) Gal4DBD-CtBP2 cleft mutants were tested against Gal4AD-wild-type CtBP2 or BKLF 1-268. (B) Gal4AD-CtBP2 cleft mutant proteins were tested against Gal4DBD-wild-type CtBP2 or BKLF 1-268.

FIG. 3.

Use of CtBP2 cleft derivatives to assess whether FHL3 and Ubc9 bind to CtBP2 in the cleft or elsewhere on the surface of CtBP2. (A) Murine FHL3 fused to the Gal4DBD was tested for the ability to interact with Gal4AD CtBP2 cleft derivatives in the yeast two-hybrid assay. (B) Murine FHL3 and BKLF 1-268 fused to the Gal4DBD were assessed for the ability to bind to the indicated fragments of CtBP2, rat CtBP1-S, and human CtBP1-L in the yeast two-hybrid assay. The two key cleft residues involved in binding to PIDLS motifs (A58 and V72 in CtBP2, A41 and V55 in rCtBP1-S, and A52 and V66 in hCtBP1-L) are indicated. (C) Murine Ubc9 fused to the Gal4AD was tested for the ability to interact with the CtBP2 cleft derivatives fused to the Gal4DBD in the yeast two-hybrid assay. (D) Murine Ubc9 fused to the Gal4AD was assessed for the ability to bind to the indicated fragments of CtBP2 and rat CtBP1-S in the yeast two-hybrid system. The two key cleft residues involved in binding to PIDLS motifs (A58 and V72 in CtBP2 and A41 and V55 in rCtBP1-S) are indicated. (E) Alignment of the N termini of murine CtBP2, rat CtBP1-S, and human CtBP1-L, indicating the regions of CtBP2 implicated in binding to FHL3 and Ubc9. One of the key cleft residues involved in binding to PIDLS motifs (A58 in CtBP2, A41 in rCtBP/BARS, or A52 in hCtBP1-L) is indicated. A summary of the protein-protein interaction results from panels A to D is also displayed on the right of the alignment.

FIG. 4.

Interactions between CtBP2 and its partners. A pull-down assay was performed to detect binding of GST-CtBP2 and GST-CtBP2 A58E to Ubc9, FHL3, and BKLF (1-268). Inputs of whole-cell extracts from transfected COS-1 cells are shown (Input). Interactions with HA-Ubc9, HA-FHL3, and Flag-BKLF (1-268) were monitored by Western blot (WB) analysis with anti-HA (αHA) and anti-Flag (αFlag) antibodies as indicated. For Ubc9, the asterisk indicates an unknown band in the GST lane that does not comigrate with Ubc9.

FIG. 5.

Use of CtBP cleft derivatives to distinguish between CtBP2 partners that bind in the cleft and elsewhere on the surface of CtBP. Yeast two-hybrid assays were performed to examine the interaction between the CtBP2 cleft derivatives and Ikaros (A), TRPS1 (B), Eos (C), and HIPK (D). The results for wild-type CtBP2 and for the CtBP2 fusion with the ΔDL mutant PIDLS extension were identical, and only the wild-type results are presented. A summary of the regions of the test proteins implicated in PIDLS cleft binding to CtBP and in binding to CtBP outside the cleft are indicated. nd, not determined. (E) Amino acid sequence of Eos residues 101 to 532. Residues 364 to 400 are in gray, and the known PIDLS motif, PEDLA, is shaded in black. Our yeast two-hybrid assay results indicate that there is a second PIDLS motif in residues 101 to 363 or 401 to 532.

FIG. 6.

Functional effects on gene repression activity of mutating the PIDLS cleft in CtBP2. (A) Western blot assay performed to examine the expression levels of the Gal4-CtBP2 wild-type and A58E and V72R mutant proteins in COS-1 cells. (B) Testing of Gal4-CtBP2 constructs for the ability to repress firefly luciferase reporter gene expression from the TK promoter in COS-1 cells following transient transfection. (C) Testing of Gal4-CtBP2 constructs for the ability to repress LexA-VP16-activated firefly luciferase reporter gene expression from the E1B promoter in COS-1 cells following transient transfection. (D) Testing of BKLFf-CtBP2 constructs for the ability to repress GATA-1-activated human GH reporter gene expression from the Aγ globin promoter in NIH 3T3 cells following transient transfection. (E) Testing of Gal4-CtBP2 constructs for the ability to repress LexA-VP16-activated firefly luciferase reporter gene expression from the E1B promoter in both CtBP1^+/− CtBP2^+/− and CtBP1^−/− CtBP2^−/− cells following transient transfection. (F) Testing of wild-type CtBP2 and the A58E cleft mutant form for the ability to repress the human E-cadherin (−427/+53)-luciferase promoter in CtBP1^−/− CtBP2^−/− cells. FF/R, firefly/Renilla.