Tetrameric Oligomerization Mediates Transcriptional Repression by the BRCA1-dependent Kruppel-associated Box-Zinc Finger Protein ZBRK1*

The Kruppel-associated box (KRAB)-zinc finger protein ZBRK1 has been implicated in the transcriptional regulation of DNA damage-response genes that function in cell growth control and survival. Recently, we described a novel BRCA1-dependent C-terminal transcriptional repression domain (CTRD) within ZBRK1, the mode of repression of which is functionally distinguishable from that of the N-terminal KRAB repression domain within ZBRK1. The identification of BRCA1 binding-competent but repression-defective CTRD mutants further revealed that BRCA1 binding is necessary, but not sufficient, for ZBRK1 CTRD function. During an unbiased search for possible co-regulators of the CTRD, we identified ZBRK1 itself, suggesting that ZBRK1 can oligomerize through its CTRD. Herein we explore the physical and functional requirements for ZBRK1 oligomerization in ZBRK1-directed transcriptional repression. Protein interaction analyses confirmed that ZBRK1 can homo-oligomerize both in vitro and in vivo and further mapped the ZBRK1 oligomerization domain to the CTRD C terminus. Biochemical analyses, including protein cross-linking and gel filtration chromatography, revealed that ZBRK1 homo-oligomers exist as tetramers in solution. Functionally, ZBRK1 oligomerization facilitates ZBRK1-directed transcriptional repression through ZBRK1 response elements; requirements for oligomerization-dependent repression include the ZBRK1 CTRD and KRAB repression domains but not the DNA binding activity of ZBRK1. These observations suggest that higher order oligomers of ZBRK1 may assemble on target ZBRK1 response elements through both protein-DNA and CTRD-dependent protein-protein interactions. These findings thus reveal an unanticipated dual function for ZBRK1 in both DNA binding-dependent and -independent modes of transcriptional repression and further establish the CTRD as a novel protein interaction surface responsible for directing homotypic and heterotypic interactions necessary for ZBRK1-directed transcriptional repression.

ZBRK1 1 (zinc finger and BRCA1-interacting protein with a KRAB domain-1) is a member of the Kruppel-associated boxzinc finger protein (KRAB-ZFP) family of transcriptional repressors (1,2). The ϳ300 members of this family comprise a significant proportion of the transcription factor complement of the human proteome and are believed to occupy important regulatory roles in development, differentiation, and transformation (3)(4)(5)(6)(7)(8)(9). In addition to its identification as the product of a gene up-regulated in senescent fibroblasts, ZBRK1 was independently identified as a BRCA1-dependent transcriptional repressor of the gene encoding GADD45a, a functionally important DNA damage-response effector that functions in G 2 /M cell cycle checkpoint control and the maintenance of genomic stability (2,10,11). Previous functional analyses revealed that ZBRK1 represses GADD45a gene transcription through an intron 3 DNA-binding site in a manner dependent upon direct interaction with BRCA1 and further suggested a model in which ZBRK1 and BRCA1 function coordinately to repress GADD45a gene transcription in the absence of genotoxic stress (2). Evidence to support this model derives from the recent observation that DNA damage-induced ubiquitin-mediated proteolysis of ZBRK1 leads to de-repression of GADD45a (12). Although the DNA damage-induced degradation of ZBRK1 is a BRCA1-independent phenomenon, familial breast cancer-derived mutants of BRCA1 that disrupt its interaction with ZBRK1 nonetheless abrogate its co-repressor activity, suggesting that its ZBRK1 co-repressor function may be important for the tumor suppressor properties of BRCA1 (2,12).
In addition to GADD45a, potential ZBRK1 binding sites have been identified in other DNA damage-response genes that are also regulated by BRCA1, including p21, Bax, and GADD153 (2). This observation suggests a potentially broader role for BRCA1 and ZBRK1 in the coordinate transcriptional control of functionally diverse DNA damage response genes. As part of our initial effort to explore the mechanism by which BRCA1 mediates sequence-specific transcriptional repression through ZBRK1, we recently identified and functionally characterized a novel C-terminal transcriptional repression domain (CTRD) within ZBRK1 (13). Structurally, the CTRD comprises the last four zinc fingers and an atypical C-terminal extension within the eight-fingered ZBRK1 protein (13). Functionally, the CTRD represses transcription in a BRCA1-dependent, histone deacetylase-dependent, and promoter-specific manner and is thus distinguishable from the N-terminal KRAB repres-sion domain in ZBRK1, which exhibits no BRCA1 dependence and broad promoter specificity (13).
During the course of our functional dissection of ZBRK1, we identified CTRD mutants that are competent for BRCA1 binding but nonetheless defective for transcriptional repression (13). This observation suggests that BRCA1 binding is necessary but not sufficient for the ZBRK1 CTRD repression function and implies a role for additional co-regulators of the CTRD. During an unbiased search for CTRD-interacting proteins and, therefore, potential CTRD co-regulators, we unexpectedly identified ZBRK1 itself, suggesting that ZBRK1 can homo-oligomerize through its CTRD. Herein, we explore the physical and functional requirements for ZBRK1 oligomerization in ZBRK1-directed repression. Our results reveal the ZBRK1 CTRD to be a novel interaction surface responsible for directing both homotypic and the heterotypic interactions essential for ZBRK1 repression and further suggest an unanticipated dual role for ZBRK1 in both DNA binding-dependent and -independent modes of transcriptional repression. We discuss the implications of these findings for BRCA1-dependent ZBRK1 repression.
Cell Culture, Transfections, and Reporter Assays-U2OS human osteosarcoma cells were cultured as described (13). For transient reporter assays, 2 ϫ 10 5 cells were seeded per well of a 6-well plate 24 h before transfection. When 70% confluent, the cells were transfected with Fu-GENE 6 (Roche Diagnostics) following the manufacturer's instructions. The amounts of expression and reporter plasmids utilized in each experiment are indicated in the legends to Figs. 4 -6. A total of 1or 2 g, respectively, of DNA per well was transfected when pG 5 TK-Luc or pZRE 4 -Luc were used as reporter templates. Each transfection also included an internal control plasmid, pCH110 (15), expressing ␤-galactosidase under control of the SV40 promoter. Forty-eight hours posttransfection, cells were harvested and lysed in 250 l/well reporter lysis buffer (Promega, Madison, WI). Transfected cell lysates (20 l) were analyzed for luciferase activity using the luciferase assay system (Promega) and for ␤-galactosidase activity using the Galacto-Light Plus chemiluminescent reporter assay (BD Biosciences). Each transfection was performed a minimum of three times in duplicate.
Protein Expression and Purification-MBP-ZBRK1 ⌬K, MBP-ZBRK1 CTRD, and MBP proteins were each expressed in and purified from E. coli strain BL21 Star (DE3) pLysS (Invitrogen). The procedures for protein induction, extraction, and purification of MBP-ZBRK1 ⌬K have been described previously (13). For MBP-ZBRK1 CTRD and MBP, cells were grown at 37°C to an A 600 of 0.6. Isopropyl-1-thio-␤-D-galactopyranoside was added to a final concentration 0.3 mM, and the cells were transferred to 25°C for another 3.5 h. Cells were pelleted, washed once with phosphate-buffered saline, and then resuspended in MBP binding buffer (20 mM HEPES, pH 7.5, 1 mM EDTA, 500 mM NaCl, and 10 mM ␤-mercaptoethanol) supplemented with protease inhibitors (0.4 g/ml aprotinin, 0.5 g/ml chymostatin, 0.5 g/ml leupeptin, 0.5 g/ml pepstatin, 0.5 mM phenylmethylsulfonyl fluoride, and 0.5 mM benzamidine-HCl). Resuspended cells were frozen and thawed once followed by sonication (three times for 1 min each on ice) and clarification by centrifugation at 30,000 ϫ g for 30 min. MBP-ZBRK1 CTRD and MBP proteins were purified from clarified lysates by affinity chromatography on amylose resin (New England Biolabs, Beverly, MA). Briefly, clarified lysates were incubated with amylose resin in batch at for 1 h at 4°C, washed with MBP binding buffer, and eluted with MBP binding buffer containing 0.5% maltose in column format.
MBP Binding Assays-MBP-ZBRK1 ⌬K was purified from clarified bacterial cell extract (150 l) by incubation with amylose resin (15 l) in batch for 30 min at 25°C, followed by washing two times for 5 min each at 25°C in MBP binding buffer supplemented with protease inhibitors and three times for 5 min each at 25°C in lysis 300 buffer (50 mM Tris-HCl, 300 mM NaCl, 5 mM EDTA, 0.1% Nonidet P-40, and 10 mM ␤-mercaptoethanol) supplemented with protease inhibitors. FLAG-ZBRK1 or its mutant derivatives were labeled with [ 35 S]methionine (TNT SP6 quick coupled transcription/translation system; Promega) and incubated with amylose-immobilized MBP-ZBRK1 ⌬K in lysis 300 buffer for 1 h at 4°C. Binding reactions were washed with Lysis 300 buffer four times for 10 min each at 4°C and subsequently boiled in 30 l of 1.5ϫ Laemmli sample buffer. Eluates were resolved by SDS-PAGE (10% gel) and visualized by PhosphorImager analysis (Amersham Biosciences).
Chemical Cross-linking Assays-U2OS cells at 70% confluency on 10-cm plates were transfected with either 12 g of pCS2ϩ-FLAG-ZBRK1 or 6 g of pCS2ϩ-FLAG-ZBRK1 ⌬C1 in combination with 6 g of pCS2ϩ. Thirty-six hours post-transfection, cells were harvested and lysed in 300 l of lysis 150H buffer (20 mM HEPES, pH 7.5, 150 mM NaCl, 5 mM EDTA, and 0.1% Nonidet P-40) by sonication at four intervals of 7 s each using a Branson 450 sonifier (Branson Ultrasonics) at the lowest setting. The lysate was clarified by centrifugation at 20,000 ϫ g for 20 min at 4°C. Clarified lysates (36 l) were subjected to chemical cross-linking by the addition of disuccinimidyl suberate (DSS) (25 mM) to a final concentration of 2.25 mM. Cross-linking reactions were performed at 25°C and terminated by the addition of 2ϫ Laemmli sample buffer followed by boiling for 3 min (16). Cross-linked products were resolved by SDS-PAGE (10% gel) and processed for immunoblot analysis using anti-FLAG M2 monoclonal antibody (Sigma-Aldrich).
Purified MBP-ZBRK1 CTRD and MBP proteins were each subjected to cross-linking by the addition of DSS (25 mM) to a final concentration of 0.75 mM. Cross-linking reactions were performed at 25°C and terminated by the addition of 2ϫ Laemmli sample buffer followed by boiling for 3 min. Cross-linked products were resolved by SDS-PAGE (10% gel) followed by silver staining.
Gel Filtration Chromatography-U2OS cells at 70% confluency on 10-cm plates were transfected with either 12 g of pCS2ϩ-FLAG-ZBRK1 or 6 g of pCS2ϩ-FLAG-ZBRK1 ⌬C1 in combination with 6 g of pCS2ϩ. Thirty-six hours post-transfection, cells were harvested and lysed in 150 l of lysis 150T buffer (50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 5 mM EDTA, 0.5% Nonidet P-40, and 25 mM pentaethylene glycol monooctyl ether (Sigma-Aldrich)) by sonication at four intervals of 7 s each using a Branson 450 sonifier (Branson Ultrasonics) at the lowest setting. The lysate was clarified by centrifugation at 20,000 ϫ g for 20 min at 4°C. Clarified lysates were concentrated 5-fold by centrifugation at 10,000 rpm using a Nanosep 30K omega filtration unit (Pall Corporation, East Hills, NY). Gel filtration chromatography was performed on an Amersham Biosciences SMART System using a Superdex 200 PC 3.2/30 column (Amersham Biosciences) and a flow rate of 40 l/min. Column fractions (45 l) were resolved by SDS-PAGE (10% gel) and visualized by immunoblot analysis using an anti-FLAG M2 monoclonal antibody (Sigma-Aldrich).

ZBRK1 Homo-oligomerizes in Vitro-
To identify novel interaction partners and potential co-regulators of the ZBRK1 CTRD, we employed two independent experimental approaches. First, we used the CTRD (amino acids 319 -532 of ZBRK1) as a bait protein to screen a human fetal brain cDNA library by a yeast-two hybrid assay. From among ϳ1 ϫ 10 6 independent clones screened, we identified ZBRK1 itself as a CTRD interaction partner (data not shown). Second, we expressed a ZBRK1 KRAB domain deletion derivative (ZBRK1 ⌬K; amino acids 144 -532 of ZBRK1) as an MBP chimera in E. coli, immobilized the fusion protein on amylose resin, and probed HeLa cell nuclear extract for interacting proteins using an MBP pull-down assay. Identification of ZBRK1 ⌬K-interacting proteins by matrix-assisted laser desorption ionization time-of-flight-based peptide mass fingerprinting revealed ZBRK1 itself to be among the interaction partners thus captured (data not shown). Coupled with the identification of ZBRK1 as a CTRD-interacting protein by a yeast two-hybrid assay, this observation suggested the possibility that ZBRK1 might homo-oligomerize through a direct interaction involving its CTRD.
To confirm that ZBRK1 can bind directly to itself and to map the relevant interaction surface required for this interaction, we tested the ability of in vitro translated wild-type ZBRK1 and ZBRK1 CTRD mutant derivatives for their respective abilities to bind to MBP-ZBRK1 ⌬K in an MBP pull-down assay (Fig. 1). As reported previously, deletion of the KRAB domain permits the expression and purification of an otherwise insoluble fulllength recombinant ZBRK1 protein (13). MBP-ZBRK1 ⌬K bound full-length ZBRK1 efficiently, thus confirming that ZBRK1 can directly self-associate in vitro (Fig. 1, A and B). Next, we examined the structural determinants within the CTRD required for ZBRK1 self-association. Analysis of C-terminal ZBRK1 truncation derivatives revealed that the deletion of only nine amino acids from the CTRD C terminus was sufficient to eliminate ZBRK1 oligomerization (Fig. 1A). By contrast, individual disruption of CTRD zinc fingers 5-8 by substitution mutagenesis revealed that each "broken finger" mutant (13) bound MBP-ZBRK1 ⌬K comparably to wild-type ZBRK1, thus revealing that the CTRD zinc fingers are individually dispensable for ZBRK1 oligomerization (Fig. 1B). Taken together, these results demonstrate that ZBRK1 can homooligomerize in vitro through a direct interaction involving its CTRD C terminus.
ZBRK1 Homo-oligomerizes in Vivo-To determine whether ZBRK1 can homo-oligomerize in vivo, we first examined the ability of ZBRK1 derivatives independently tagged with either T7 or FLAG epitopes to reciprocally co-immunoprecipitate one another following their transient over-expression in U2OS human osteosarcoma cells. T7-tagged ZBRK1 could be efficiently precipitated by FLAG-specific antibodies only in the presence of FLAG-tagged ZBRK1 ( Fig. 2A). Reciprocally, FLAG-tagged ZBRK1 could be efficiently precipitated by T7-specific antibodies only when T7-tagged ZBRK1 was co-expressed (Fig. 2B). These results confirm that ZBRK1 can homo-oligomerize in vivo. Next, to map the relevant interaction surface(s) required for ZBRK1 oligomerization in vivo, we examined a set of FLAGtagged ZBRK1 CTRD mutant derivatives for their respective abilities to be co-immunoprecipitated along with T7-tagged wild-type ZBRK1 following their transient co-expression in U2OS cells. Consistent with the results of in vitro binding analyses, these in vivo binding studies revealed the CTRD C terminus to be critical and the CTRD zinc fingers to be dispensable for ZBRK1 self-association. Thus, deletion of only nine amino acids from the ZBRK1 C terminus largely abolished ZBRK1 self-association (Fig. 2C), whereas individual ZBRK1 derivatives bearing targeted disruptions of CTRD zinc fingers 5-8 bound to T7-tagged ZBRK1 as efficiently as wild-type ZBRK1 (Fig. 2D). Taken together, these results demonstrate that ZBRK1 can homo-oligomerize both in vitro and in vivo and furthermore delimit the oligomerization interface to the CTRD C terminus.
The ZBRK1 CTRD Mediates Tetrameric Oligomerization-To characterize the oligomeric state of ZBRK1 in vivo, we first employed the chemical cross-linker DSS to covalently capture ZBRK1 protein complexes present in extracts of U2OS cells transfected with FLAG-tagged ZBRK1 (16). Cross-linked samples were analyzed by SDS-PAGE and immunoblot analysis using FLAG epitope tag-specific antibodies. In the presence of DSS, a significant proportion of FLAG-tagged ZBRK1 migrated with an apparent molecular mass of ϳ240 kDa, consistent with the size of a ZBRK1 tetramer (Fig. 3A). Importantly, no cross-linked products were observed when an oligomerization-defective ZBRK1 mutant (⌬C1) was over-expressed in U2OS cells, demonstrating that DSS-induced formation of high molecular mass ZBRK1 complexes is strictly dependent upon the intact CTRD C terminus (Fig. 3A).
Although the electrophoretic properties of FLAG-tagged ZBRK1 in DSS-cross-linked U2OS cell extracts suggests the formation of tetramers, we could not exclude the possibility that additional cellular proteins were also incorporated into the ZBRK1 complex, particularly because we know that ZBRK1 interacts with other proteins in the cell, including BRCA1 (2,13). The concentration of ectopically expressed ZBRK1 in these transfection experiments considerably exceeds that of its endogenous counterpart, and these conditions are likely to drive ZBRK1 self-association and mask the influence of endogenous interacting proteins on the size of a cross-linked ZBRK1 complex. Nonetheless, we sought independent confirmation of the stoichiometry of ZBRK1 oligomerization through its CTRD. To this end, we therefore subjected highly purified recombinant MBP-ZBRK1 CTRD to DSS cross-linking. Cross-linked samples were analyzed by SDS-PAGE followed by silver staining to visualize MBP-ZBRK1 CTRD oligomers. In the presence of DSS, a significant proportion of MBP-ZBRK1 CTRD migrated with an apparent molecular mass of ϳ270 kDa, consistent with the size of an MBP-ZBRK1 CTRD tetramer (Fig. 3B). Importantly, no cross-linked products were observed when MBP alone was treated with DSS, indicating that DSS-induced formation of a high molecular mass MBP-ZBRK1 CTRD complex is strictly dependent upon the ZBRK1 CTRD (Fig. 3B). Furthermore, immunoblot analysis using ZBRK1-specific antibodies confirmed that the high molecular mass complex observed upon DSS cross-linking of highly purified MBP-ZBRK1 CTRD is derived from MBP-CTRD oligomerization (data not shown).
As an independent approach to determine the stoichiometry of ZBRK1 oligomerization through its CTRD, we examined the respective gel filtration profiles of ectopically expressed wildtype ZBRK1 and its oligomerization-defective mutant (⌬ C1) present in extracts of transfected U2OS cells. Immunoblot analysis of column fractions using a FLAG epitope-specific antibody revealed a peak of wild-type ZBRK1 protein in fractions corresponding to an apparent molecular mass of ϳ240 kDa, most consistent with the size of a ZBRK1 tetramer (Fig.  3C). We also noted a minor peak of wild-type ZBRK1 protein in a fraction corresponding to an apparent ZBRK1 monomer of 58 kDa. By contrast, an oligomerization-defective ZBRK1 derivative (⌬ C1) peaked exclusively in column fractions corresponding to an apparent monomer of ϳ57 kDa (Fig. 3C). Taken together, these results indicate that ZBRK1 tetramerizes both in vitro and in vivo in a manner that is strictly dependent upon the integrity of its CTRD C terminus.
ZBRK1 Homo-oligomerization Potentiates ZBRK1-directed Transcriptional Repression-To explore the functional consequence of ZBRK1 homo-oligomerization, we initially exploited our previous observation that the ZBRK1 CTRD can function as an autonomous transcriptional repression domain when tethered to a heterologous DNA-binding domain from the yeast transcription factor GAL4 (13). First, we examined the ability of full-length ZBRK1 to potentiate transcriptional repression FIG. 2. ZBRK1 homo-oligomerizes in vivo through its CTRD C terminus. A and B, T7 and FLAG-tagged wild-type ZBRK1 proteins were ectopically expressed either alone or together in U20S cells, and whole cell extracts from transfected cells were subjected to immunoprecipitation (IP) using antibodies specific for either the FLAG (A) or T7 (B) epitopes. Immunoprecipitates were resolved by SDS-PAGE (10% gel) and processed by immunoblot analysis using both T7 and FLAG epitope-specific antibodies as indicated. Input represents 10% of the total whole cell extract subjected to immunoprecipitation. C and D, T7-tagged wild-type (WT) ZBRK1 was ectopically expressed in U20S cells along with FLAG-tagged WT ZBRK1 or its indicated ⌬C (C) or BF (D) derivatives. Whole cell extracts from transfected cells were subjected to immunoprecipitation (IP) using antibodies specific for the T7 epitope. Immunoprecipitates were resolved by SDS-PAGE (10% gel) and processed by immunoblot analysis using both T7-and FLAG epitope-specific antibodies as indicated. Input represents 10% of the total whole cell extract subjected to immunoprecipitation. directed by GAL4-ZBRK1 CTRD in a transient repression assay using a reporter template bearing multimerized GAL4 DNA-binding sites located upstream of the herpes simplex virus TK promoter in U2OS cells (Fig. 4A) (13). Consistent with our previous studies (13), we observed that GAL4-ZBRK1 CTRD repressed transcription from this reporter template at least 10-fold in a dose-dependent manner (data not shown). To examine the influence of wild-type ZBRK1 on repression directed by GAL4-ZBRK1 CTRD, we transfected a sub-optimal nanogram quantity of the GAL4-ZBRK1 CTRD expression vector that supports only ϳ2.5-fold repression of reporter gene activity. Under these conditions, ectopic expression of wild-type ZBRK1 enhanced this level of repression up to 4-fold in a dose-dependent manner (Fig. 4B). Importantly, ZBRK1 had minimal influence on reporter activity in the presence of the GAL4 DNA-binding domain alone, confirming that the repressive effect of ZBRK1 in this assay derives from its direct recruitment to the CTRD and not to DNA (Fig. 4B). This result demonstrates for the first time that, in addition to its well documented role as a sequence-specific DNA-binding transcriptional repressor (2,13), ZBRK1 can also function as a corepressor independent of its inherent DNA binding activity.
Next, we examined the structural features within ZBRK1 required for its co-repressor function. Relative to wild-type ZBRK1, oligomerization-defective CTRD C-terminal truncation mutants (⌬C1, ⌬C2, and ⌬C3) were completely defective for ZBRK1 co-repressor activity, implying a strict reliance on oligomerization for ZBRK1 to function as a co-repressor (Fig.  5A). Analysis of ZBRK1 derivatives bearing targeted disruptions of CTRD zinc fingers 5-8 (BF 5-8) revealed these mutants to be partially defective for the ZBRK1 co-repressor function (Fig. 5B). Because these CTRD zinc finger mutations do not disrupt ZBRK1 oligomerization (see Figs. 1 and 2) but do disrupt CTRD interactions with additional co-repressors including BRCA1 (13), this result suggests that CTRD-mediated recruitment of additional co-repressors also likely contributes to the ability of ZBRK1 to function as a DNA binding-independent co-repressor. Finally, we examined the contribution of the FIG. 3. ZBRK1 tetramerization requires its CTRD C terminus. A, schematic representation of FLAG-tagged wild-type ZBRK1 and its oligomerization-defective ⌬C1 deletion derivative. U2OS cells were transfected with expression vectors for either FLAG-tagged WT ZBRK1 or its ⌬C1 deletion derivative as indicated. Whole cell extracts from transfected cells were subjected to cross-linking with 2.25 mM DSS for 0, 1.5, or 3 min as indicated, and cross-linking reactions were terminated with Laemmli sample buffer. Cross-linked proteins were resolved by SDS-PAGE (10% gel) and visualized by immunoblot analysis using FLAG epitope-specific antibodies. Arrowheads indicate un-cross-linked monomers of FLAG-tagged ZBRK1 and its ⌬C1 deletion derivative as well as cross-linked ZBRK1 complexes. Molecular weight markers (M r ) are indicated. Asterisk denotes a cross-reacting cellular protein recognized by the FLAG antibody. B, schematic representation of recombinant MBP-ZBRK1 CTRD and MBP. Purified recombinant MBP-ZBRK1 CTRD and MBP were subjected to cross-linking with 0.75 mM DSS for 0, 1.5, or 3 min as indicated, and cross-linking reactions were terminated with Laemmli sample buffer. Cross-linked proteins were resolved by SDS-PAGE (10% gel) and visualized by silver stain analysis. Arrowheads indicate un-cross-linked monomers of MBP-ZBRK1 CTRD and MBP as well as cross-linked MBP-ZBRK1 CTRD complexes. Molecular weight markers (M r ) are indicated. C, whole cell extracts from U2OS cells transfected with FLAG-tagged wild-type (WT) ZBRK1 or an oligomerization-defective mutant derivative (⌬C1) were independently subjected to Superdex 200 gel filtration chromatography. Individual column fractions were resolved by SDS-PAGE (10% gel) and analyzed for the presence of FLAG-tagged ZBRK1 by immunoblot analysis with FLAG epitope-specific antibodies. Arrowheads indicate the relative positions of marker protein peaks. Input represents 5% of the total whole cell extract subject to gel filtration chromatography. N-terminal KRAB repression domain within ZBRK1 to its corepressor activity. Relative to wild-type ZBRK1, a ZBRK1 mutant derivative bearing two substitution mutations (DV 3 AA) within the KRAB domain previously shown to disrupt KRAB repression function (17) was significantly defective for co-repressor activity (Fig. 5C). As this mutation does not affect ZBRK1 oligomerization, this result suggests that KRAB domain-associated co-repressors also contribute to the ability of ZBRK1 to function as a DNA binding-independent co-repressor of the CTRD. Taken together these results reveal that ZBRK1, in a DNA-binding-independent manner, can potentiate the repression function of DNA-bound CTRD through oligomerization and the consequent recruitment of co-repressors through both its KRAB and CTRD repression domains.
We next asked whether ZBRK1 CTRD-mediated oligomerization contributes to ZBRK1-directed transcriptional repression from ZBRK1 response elements (ZREs). To this end, we employed a reciprocal transient repression assay in which we examined the ability of the CTRD to potentiate ZBRK1-directed repression from a reporter template bearing multimerized ZREs (Fig. 6A). The fact that the CTRD by itself has no ZRE binding activity permitted us to assess the influence of the CTRD on ZBRK1-directed repression solely by virtue of its ability to oligomerize with DNA-bound ZBRK1. Consistent with our previous studies (2), we observed that wild-type ZBRK1 repressed transcription from a reporter template bearing multimerized ZREs located upstream of the HSV TK pro-moter in U2OS cells at least 10-fold in a dose-dependent manner (data not shown). To examine the influence of the ZBRK1 CTRD on repression directed by wild-type ZBRK1, we transfected a sub-optimal nanogram quantity of the ZBRK1 expression vector that supports only ϳ2.5-fold repression of transcription from this reporter template (Fig. 6B). Under these conditions, ectopic expression of GAL4-CTRD enhanced this level of repression up to 4-fold in a dose-dependent manner (Fig. 6B). Importantly, GAL4-CTRD had minimal influence on reporter activity in the absence of ZBRK1, confirming that the CTRD potentiates ZBRK1 repression in this assay by virtue of its direct recruitment to ZBRK1 and not to DNA (Fig. 6B). Furthermore, an oligomerization-defective CTRD mutant (GAL4-CTRD ⌬C1) bearing a nine amino acid deletion from the CTRD C terminus was completely defective in its ability to potentiate ZBRK1 repression in this assay, confirming that CTRD-mediated oligomerization is a requirement for its ZBRK1 co-repressor activity (Fig. 6C). Collectively, these experiments reveal an unanticipated dual function for ZBRK1 in both DNA binding-dependent and -independent modes of repression, the latter of which derives from the ability of ZBRK1 to homo-oligomerize through its CTRD.

DISCUSSION
The identification of ZBRK1 itself during an unbiased search for ZBRK1 CTRD co-regulators prompted us to characterize the structural requirements for and the functional consequences of ZBRK1 homo-oligomerization. Herein we demonstrate that ZBRK1 undergoes tetrameric oligomerization both in vitro and in vivo, and we further identify the ZBRK1 CTRD C terminus to be both necessary and sufficient for tetrameric oligomerization. Functionally, we demonstrate that ZBRK1 oligomerization facilitates ZBRK1-directed transcriptional repression and concomitantly identify several unique and biologically significant features of ZBRK1 relevant to its role as a KRAB-ZFP transcriptional repressor. First, by virtue of its ability to homooligomerize, we identify an unanticipated role for ZBRK1 as a DNA binding-independent co-repressor in addition to its well characterized role as a sequence-specific DNA-binding transcriptional repressor (2). In this regard, we identify both the KRAB and CTRD repression domains within ZBRK1 to be necessary for oligomerization-dependent repression. Whether or not these domains are sufficient for oligomerization-dependent repression remains to be established. Second, we reveal the ZBRK1 CTRD to be a novel protein interaction surface responsible for directing both homotypic and the heterotypic interactions necessary for BRCA1-dependent ZBRK1 transcriptional repression.
In general, oligomerization enhances regulatory diversity among sequence-specific DNA-binding transcription factors by expanding the number of potential response elements to which oligomers can effectively bind and/or by increasing the number and type of transcriptional regulatory domains that can effectively function at a core promoter. Presently, we do not know whether and how ZBRK1 oligomerization might influence the choice of DNA sequences recognized by ZBRK1. The 15-base pair consensus ZRE was originally identified by sampling a random pool of double-stranded oligonucleotides with a recombinant ZBRK1 derivative comprised solely of the eight ZBRK1 zinc fingers but lacking the C-terminal extension that we show here to be critical for oligomerization (2). Future studies will be required to determine whether and how oligomerization-competent forms of ZBRK1 preferentially recognize an expanded ZRE. Interestingly, our most recent studies revealed that the inherent DNA binding activity of ZBRK1 zinc fingers 1-7 on a consensus ZRE is constrained by the ZBRK1 C terminus, suggesting a potential role for the ZBRK1 oligomerization domain in regulating the FIG. 4. ZBRK1 has a DNA binding-independent co-repressor activity. A, schematic representation of FLAG-tagged ZBRK1, GAL4-ZBRK1-CTRD, GAL4 DNA-binding domain alone, and the pG 5 TK-lucerferase (5XGAL4 TK Luc) reporter template used in transfection assays. B, U2OS cells were transfected with 100 ng of the pG 5 TKluciferase reporter template and 600 ng of either pM (expressing the GAL4 DNA-binding domain alone) or pM-ZBRK1-CTRD (expressing GAL4-ZBRK1-CTRD) without or with the indicated nanogram amounts of pCS2ϩ-FLAG-ZBRK1 expressing FLAG-tagged wild-type ZBRK1. In this and all subsequent transient reporter assays involving effector plasmid titrations, the total amount of DNA in each transfection was fixed by reciprocal titration of the corresponding backbone expression plasmid. Also, in this and all subsequent transient reporter assays the relative luciferase activity represents the ratio of the luciferase activity obtained in a particular transfection to that obtained in cells transfected with only the reporter and pM expression vectors alone. Luciferase activities were first normalized to ␤-galactosidase activity obtained by co-transfection of the SV40-␤-galactosidase vector (25 ng) as described previously (13). Error bars represent the S.D. from the average of at least three independent transfections performed in duplicate.
sequence-specific association of ZBRK1 with a consensus ZRE (13). Although our previous studies clearly revealed that ZBRK1 can bind to and repress transcription from a single ZRE (2,13), it nonetheless remains to be established whether and how ZBRK1 oligomerization might facilitate synergistic binding and/or repression from multimerized ZREs.
Our work reveals a unique and unanticipated dual function for ZBRK1 in both DNA binding-dependent and -independent modes of transcriptional repression. With respect to the latter, we demonstrate that ZBRK1 can function as a co-repressor in a manner dependent upon the integrity of its CTRD and KRAB repression domains but not upon its inherent DNA binding activity. This observation raises the possibility that the regulatory potential of ZBRK1 on certain target genes may derive from its recruitment via protein-protein rather than protein-DNA interactions. To our knowledge, this represents the first Relative luciferase activities were calculated as described in the legend to Fig. 4. Bottom, to confirm that wild-type FLAG-tagged ZBRK1 and its mutant derivatives were expressed at roughly equivalent levels in functional assays, equivalent proportional aliquots of transfected U20S cell lysates used in functional assays (middle) were also resolved by SDS-10% PAGE and subjected to immunoblot analysis using FLAG epitope-specific antibodies.
FIG. 6. CTRD-mediated oligomerization is essential for ZBRK1 co-repressor activity. A, schematic representation of GAL4-ZBRK1-CTRD, GAL4-ZBRK1-CTRD ⌬C1, FLAG-tagged ZBRK1, and a pZRE 4 TK-luciferase reporter (4XZRE TK Luc) template used in transfection assays. B, U20S cells were transfected with 100 ng of the pZRE 4 TK-Luc reporter template without or with 800 ng of pCS2ϩ-FLAG-ZBRK1 expressing FLAG-tagged ZBRK1 and the indicated nanogram amounts of pM-ZBRK1 CTRD expressing GAL4-ZBRK1 CTRD. C, U2OS cells were transfected with 100 ng of the pZRE 4 TK-Luc reporter template without or with 800 ng of pCS2ϩ-FLAG-ZBRK1 expressing FLAG-tagged ZBRK1 and normalized amounts of pM-based plasmids expressing GAL4-ZBRK1 CTRD (1000 ng) or GAL4-ZBRK1-CTRD ⌬C1 (500 ng). To confirm that wild-type (WT) GAL4-ZBRK1 CTRD and GAL4-ZBRK1-CTRD ⌬C1 were expressed at equivalent levels in functional assays, equivalent proportional aliquots of transfected U20S cell lysates used in the functional assays were also resolved by SDS-PAGE (10% gel) and subjected to immunoblot analysis using FLAG epitope-specific antibodies. example of a KRAB-ZFP that can function dually as both a DNA-binding repressor as well as a DNA binding-independent co-repressor. Thus, we propose that higher order oligomers of ZBRK1 may assemble on a target ZRE through protein-DNA and protein-protein interactions, the latter involving oligomerization through the CTRD C terminus.
Our work further provides unique insight into the structural and functional organization of ZBRK1 as a member of the KRAB-ZFP family. Like other KRAB-ZFPs, ZBRK1 harbors an N-terminal KRAB repression domain followed by a C 2 H 2 zinc finger DNA-binding domain. However, ZBRK1 also harbors an atypical C-terminal extension that is generally absent from the larger family of KRAB-ZFPs. Our previous work revealed this C-terminal extension to comprise part of a novel BRCA1-dependent CTRD that also includes zinc fingers 5-8 within ZBRK1 (13). Herein, we describe an additional and unique role for the CTRD as an oligomerization interface sufficient for directing homotypic interactions critical for ZBRK1 repressor function. Within the CTRD we demonstrate that zinc fingers 5-8 are dispensable, whereas the C-terminal extension is essential for ZBRK1 oligomerization. This C-terminal extension shows no obvious sequence similarity to the SCAN domain that mediates selective oligomerization between certain KRAB-ZFPs (1,18). However, a BLAST search of available protein databases reveals several additional KRAB-ZFPs (ZNF577, ZNF613, and FLJ12644), each of which carries an extended C terminus of variable length with sequence homology to the ZBRK1 C-terminal extension. Interestingly, these proteins exhibit the same rank order with respect to their degree of homology and physical proximity to ZBRK1 on chromosome 19q, suggesting that they may have arisen from gene duplication (19 -21). It will be of future interest to determine whether and how these KRAB-ZFPs functionally oligomerize with ZBRK1 and/or BRCA1. Although zinc fingers 5-8 within the CTRD are dispensable for oligomerization, our previous studies have nonetheless shown them to be essential for mediating interactions with additional co-repressors, including BRCA1 (2, 13). Thus, our studies collectively reveal the CTRD to be a novel protein-interaction surface responsible for directing both homotypic and heterotypic interactions necessary for BRCA1dependent ZBRK1 transcriptional repression.
Finally, our identification of a novel oligomerization domain within ZBRK1 suggests that the regulatory potential of ZBRK1 could extend to genes that do not contain ZREs. For example, ZBRK1 could repress transcription through its hetero-oligomerization with other sequence-specific DNA binding transcription factors through direct protein-protein interactions involving its C terminus. This, in turn, could expand the repertoire of target genes subject to coordinate transcriptional control by both ZBRK1 and BRCA1, because the repressive potential of ZBRK1 is dependent upon BRCA1 (2,13). Consistent with this possibility, our yeast two-hybrid screen for CTRDinteracting proteins identified, in addition to ZBRK1 itself, the oligomeric transcription factors SRF and ATF-1 (22)(23)(24)(25)(26). Interestingly, these transcription factors control the expression of genes with diverse functions in cell growth control and survival, and future studies will seek to establish whether and how ZBRK1 physically and functionally interacts with these transcriptional regulators. Nonetheless, our work here sheds new light on the structural and functional organization of ZBRK1 as a KRAB-ZFP and further defines the multifunctional nature of its unique CTRD.