Phosphorylation of the CCAAT displacement protein (CDP)/Cux transcription factor by cyclin A-Cdk1 modulates its DNA binding activity in G(2).

Stable DNA binding by the mammalian CCAAT displacement protein (CDP)/Cux transcription factor was previously found to be up-regulated at the G(1)/S transition as the result of two events, dephosphorylation by the Cdc25A phosphatase and proteolytic processing, to generate an amino-truncated isoform of 110 kDa. In S phase, CDP/Cux was shown to interact with and repress the core promoter of the p21(WAF1) gene. Here we demonstrate that DNA binding by p110 CDP/Cux is down-modulated as cells progress into G(2). Accordingly, cyclin A-Cdk1 was found to bind to CDP/Cux and modulate its DNA binding activity in vitro and in vivo. Interaction with CDP/Cux required the presence of both cyclin A and a cyclin-dependent kinase (Cdk)-activating kinase-activated Cdk1 and involved the Cut homeodomain and a downstream Cy motif. Phosphorylation of serines 1237 and 1270 caused inhibition of DNA binding in vitro. In cotransfection studies, cyclin A-Cdk1 inhibited CDP/Cux stable DNA binding and prevented repression of the p21(WAF1) reporter. In contrast, mutant CDP/Cux proteins in which serines 1237 and 1270 were replaced with alanines were not affected by cyclin A-Cdk1. In summary, our results suggest that the phosphorylation of CDP/Cux by cyclin A-Cdk1 contributes to down-modulate CDP/Cux activity as cells progress into the G(2) phase of the cell cycle.

S phase-specific proteolytic cleavage is required to activate stable DNA binding by the CDP/Cut homeodomain protein.

The CCAAT displacement protein (CDP), the homologue of the Drosophila melanogaster Cut protein, contains four DNA binding domains that function in pairs. Cooperation between Cut repeat 3 and the Cut homeodomain allows stable DNA binding to the ATCGAT motif, an activity previously shown to be upregulated in S phase. Here we showed that the full-length CDP/Cut protein is incapable of stable DNA binding and that the ATCGAT binding activity present in cells involves a 110-kDa carboxy-terminal peptide of CDP/Cut. A vector expressing CDP/Cut with Myc and hemagglutinin epitope tags at either end generated N- and C-terminal products of 90 and 110 kDa, suggesting that proteolytic cleavage was involved. In vivo pulse/chase labeling experiments confirmed that the 110-kDa protein was derived from the full-length CDP/Cut protein. Proteolytic processing was weak or not detectable in G(0) and G(1) but increased in populations of cells enriched in S phase, and the appearance of the 110-kDa protein coincided with the increase in ATCGAT DNA binding. Interestingly, the amino-truncated and the full-length CDP/Cut isoforms exhibited different transcriptional properties in a reporter assay. We conclude that proteolytic processing of CDP/Cut at the G(1)/S transition generates a CDP/Cut isoform with distinct DNA binding and transcriptional activities. These findings, together with the cleavage of the Scc1 protein at mitosis, suggest that site-specific proteolysis may play an important role in the regulation of cell cycle progression.

Role of the multifunctional CDP/Cut/Cux homeodomain transcription factor in regulating differentiation, cell growth and development.

CDP/Cux/Cut proteins are an evolutionarily conserved family of proteins containing several DNA binding domains: one Cut homeodomain and one, two or three Cut repeats. In Drosophila melanogaster, genetic studies indicated that Cut functions as a determinant of cell-type specification in several tissues, notably in the peripheral nervous system, the wing margin and the Malpighian tubule. Moreover, Cut was found to be a target and an effector of the Notch signaling pathway. In vertebrates, the same functions appear to be fulfilled by two cut-related genes with distinct patterns of expression. Cloning of the cDNA for the CCAAT-displacement protein (CDP) revealed that it was the human homologue of Drosophila Cut. CDP was later found be the DNA binding protein of the previously characterized histone nuclear factor D (HiNF-D). CDP and its mouse counterpart, Cux, were also reported to interact with regulatory elements from a large number of genes, including matrix attachment regions (MARs). CDP/Cut proteins were found generally to function as transcriptional repressors, although a participation in transcriptional activation is suggested by some data. Repression by CDP/Cut involves competition for binding site occupancy and active repression via the recruitment of a histone deacetylase activity. Various combinations of Cut repeats and the Cut homeodomains can generate distinct DNA binding activities. These activities are elevated in proliferating cells and decrease during terminal differentiation. One activity, involving the Cut homeodomain, is upregulated in S phase. CDP/Cut function is regulated by several post-translational modification events including phosphorylation, dephosphorylation, and acetylation. The CUTL1 gene in human was mapped to 7q22, a chromosomal region that is frequently rearranged in various cancers.

The role of CDP in the negative regulation of CXCL1 gene expression.

The CXC chemokine, melanoma growth stimulatory activity/growth-regulated protein, CXCL1 is an important modulator of inflammation, wound healing, angiogenesis, and tumorigenesis. Transcription of CXCL1 is regulated through several cis-acting elements including Sp1, NF-kappa B, and an element that lies immediately upstream of the NF-kappa B element, the immediate upstream region (IUR). A transcription element data base search indicated that the IUR element contains a binding site for the transcriptional repressor, human CUT homeodomain protein/CCAAT displacement protein (CDP). It is shown here that in electrophoretic mobility shift assays, complexes obtained with the IUR oligonucleotide probe are supershifted by anti-CDP antibodies and that a CDP polypeptide containing a high affinity DNA binding domain binds to the sequence GGGATCGATC in the IUR element. In Southwestern blot analyses, oligonucleotides containing the wild-type IUR sequence, but not a mutant oligonucleotide with substitutions in the GGGATCGATC sequence, bind a 170--180-kDa protein. Furthermore, overexpression of the CDP protein blocks CXCL1 promoter activity in reporter gene assays, whereas overexpression of an antisense CDP construct leads to a significant increase in CXCL1 promoter activity. Mutations in the IUR element, which map in the putative CDP-binding site, inhibit the binding of CDP to the IUR element and favor increased transcription from the CXCL1 promoter. Based on these results, we propose that transcriptional regulation of the CXCL1 gene is mediated in part by CDP, which could play an important role in inflammatory processes and tumorigenesis.

CCAAT displacement activity involves CUT repeats 1 and 2, not the CUT homeodomain.

The CCAAT displacement protein, the homolog of the Drosophila melanogaster CUT protein, contains four DNA-binding domains: three CUT repeats (CR1, CR2, and CR3) and the CUT homeodomain (HD). Using a panel of fusion proteins, we found that a CUT repeat cannot bind to DNA as a monomer, but that certain combinations of domains exhibit high DNA-binding affinity: CR1+2, CR3HD, CR1HD, and CR2HD. One combination (CR1+2) exhibited strikingly different DNA-binding kinetics and specificities. CR1+2 displayed rapid on and off rates and bound preferably to two C(A/G)AT sites, organized as direct or inverted repeats. Accordingly, only CR1+2 was able to bind to the CCAAT sequence, and its affinity was increased by the presence of a C(A/G)AT site at close proximity. A purified CCAAT displacement protein/CUT protein exhibited DNA-binding properties similar to those of CR1+2; and in nuclear extracts, the CCAAT displacement activity also required the simultaneous presence of a C(A/G)AT site. Moreover, CR1+2, but not CR3HD, was able to displace nuclear factor Y. Thus, the CCAAT displacement activity requires the presence of an additional sequence (CAAT or CGAT) and involves CR1 and CR2, but not the CUT homeodomain.

CDP/Cut DNA binding activity is down-modulated in granulocytes, macrophages and erythrocytes but remains elevated in differentiating megakaryocytes.

DNA binding by the CCAAT-displacement protein, the mammalian homologue of the Drosophila melanogaster Cut protein, was previously found to increase sharply in S phase, suggesting a role for CDP/Cut in cell cycle progression. Genetic studies in Drosophila indicated that cut plays an important role in cell-type specification in several tissues. In the present study, we have investigated CDP/Cut expression and activity in a panel of multipotent hematopoietic cell lines that can be induced to differentiate in vitro into distinct cell types. While CDP/Cut DNA binding activity declined in the pathways leading to macrophages, granulocytes and erythrocytes, it remained elevated in megakaryocytes. CDP/Cut was also highly expressed in primary megakaryocytes isolated from mouse, and some DNA binding activity could be detected. Altogether, these results raise the possibility that CDP/Cut may be a determinant of cell type identity downstream of the myelo-erythroid precursor cell. Another possibility, which does not exclude a role in lineage identity, is that CDP/Cut activity in megakaryocytes is linked to endomitosis. Indeed, elevated CDP/Cut activity in differentiating megakaryocytes and during the S phase of the cell cycle suggests that it may be required for DNA replication.

Exon/intron structure and alternative transcripts of the CUTL1 gene.

The human CUTL1 gene (Cut-like 1) is a candidate tumor suppressor gene located on chromosome 7 at band 22, a region that is frequently deleted in several human cancers. The gene spans at least 340kb and contains 33 exons. Synthesis of five different transcripts involves two promoter regions, two polyadenylation sites and seven alternative splicing events. The two polyadenylation sites are located at the ends of exons 24 and 33 and are separated by approximately 40kb. Transcription is initiated in two genomic regions, giving rise to alternate first exons which are spliced to a common exon 2. All transcripts contain exons 2 to 14, but differ in their 3' regions. Exon 14 can be spliced alternatively to the beginning or the middle of exon 15, or to exon 25, generating transcripts with exons 15 to 24 or exons 25 to 33. Moreover, exon 16 can be spliced out from the mature transcripts that contain exons 15 to 24. Overall, five distinct transcripts are generated as a result of alternative transcription initiation, splicing and polyadenylation. We discuss potential mechanisms by which alternate polyadenylation site usage may affect alternative splicing events and vice versa.

Human cut-like repressor protein binds TGFbeta type II receptor gene promoter.

Resistance to the growth inhibitory effects of transforming growth factor beta (TGFbeta) has been associated with decreased levels of the TGFbeta type II receptor (TbetaR-II) and has been correlated with tumorigenicity. Previously, we reported an A --> G mutation at position -364 in the TbetaR-II promoter in A431 tumor cells which results in reduced TbetaR-II promoter activity. In this study, we show that the CDP/Cut (CCAAT displacement protein) transcription factor, a transcriptional repressor, binds both the wild type and the mutant TbetaR-II promoter. We also demonstrate that the A --> G mutation increases CDP/Cut binding affinity, and that overexpression of CDP/Cut reduces transcription from TbetaR-II promoter reporter constructs. Increased binding of the CDP/Cut repressor protein, as a result of a mutation at position -364, represents a novel mechanism of regulation in a neoplastic cell of the promoter of a tumor suppressor gene, TbetaR-II.

Promoter analysis of the murine T-cell protein tyrosine phosphatase gene.

The T-cell protein tyrosine phosphatase (TC PTP) is expressed ubiquitously at all stages of mammalian development. However, mRNA levels fluctuate in a cell-cycle-dependent manner, reaching peak levels in late G1, and rapidly decreasing in S phase. Furthermore, TC PTP being present in higher amounts in lymphoid tissues, we have recently shown that it is essential for proper maintenance of both the bone marrow micro-environment and B- and T-cell functions. In order to better understand the elements controlling the expression pattern of this gene, we have isolated and characterized approx. 4kb of the murine TC PTP promoter. DNA sequencing of the proximal 5' region revealed the absence of both TATAA and CAAT boxes. Primer extension analysis and S1 nuclease mapping techniques identified multiple transcription initiation sites. Functional promoter activity was determined using transfection experiments of promoter deletion constructs fused to a CAT reporter construct. Our results indicate that the minimal promoter sequence required for functional expression is contained within the first 147bp of the TC PTP promoter. In addition, consistent with the cell-cycle-dependent expression of TC PTP, we localized a domain between 492 and 1976bp from the transcription initiation site through which repression occurs. In conclusion, although initiator-driven transcription allows for ubiquitous expression of TC PTP, we define general transcription motifs present within the promoter that may mediate specific modulations of the TC PTP gene.

Refined mapping of the region of loss of heterozygosity on the long arm of chromosome 7 in human breast cancer defines the location of a second tumor suppressor gene at 7q22 in the region of the CUTL1 gene.

In breast cancer, loss of heterozygosity (LOH) has been described on the long arm of chromosome 7, at band q31, suggesting the presence of a tumor suppressor gene in this region. In this study, we have identified a second region of LOH on 7q, at band 7q22. Deletion of genetic material at 7q22 was found in all tumor types and grades and was associated with increased tumor size. The region of LOH at 7q22 in every case included one or more of three polymorphic markers that are located within the CUTL1 gene. LOH of 7q22 has also been documented in the case of human uterine leiomyomas (Zeng et al., 1997; Ishwad et al., 1997). Interestingly, in both leiomyomas and mammary tumors induced in transgenic mice expressing the Polyomavirus (PyV) large T (LT) antigen, immunocomplexes of CUTL1 and PyV LT antigen were detected (Webster et al., 1998). Altogether, genetic data in human breast cancer and biochemical analyses in breast tumors from transgenic mice suggest that CUTL1 is a candidate tumor suppressor gene.

The mammalian Cut homeodomain protein functions as a cell-cycle-dependent transcriptional repressor which downmodulates p21WAF1/CIP1/SDI1 in S phase.

Cut is a homeodomain transcription factor which has the unusual property of containing several DNA-binding domains: three regions called Cut repeats and the Cut homeodomain. Genetic studies in Drosophila melanogaster indicate that cut plays important roles in the determination and maintenance of cell-type specificity. In the present study, we show that mammalian Cut proteins may yet play another biological role, specifically in proliferating cells. We found that the binding of Cut to a consensus binding site varies during the cell cycle. Binding was virtually undetectable in G0 and early G1, but became very strong as cells reached S phase. This was shown to result both from an increase in Cut expression and dephosphorylation of the Cut homeodomain by the Cdc25A phosphatase. We also show that the increase in Cut activity coincides with a decrease in p21WAF1/CIP1/SDI1 mRNAs. In co-transfection experiments, Cut proteins repressed p21WAF1/CIP1/SDI1 gene expression through binding to a sequence that overlaps the TATA box. Moreover, p21WAF1/CIP1/SDI1 expression was repressed equally well by either Cdc25A or Cut. Altogether, these results suggest a model by which Cdc25A activates the Cut repressor which then downregulates transcription of p21WAF1/CIP1/SDI1 in S phase. Thus, in addition to their role during cellular differentiation, Cut proteins also serve as cell-cycle-dependent transcriptional factors in proliferating cells.

The induction of uterine leiomyomas and mammary tumors in transgenic mice expressing polyomavirus (PyV) large T (LT) antigen is associated with the ability of PyV LT antigen to form specific complexes with retinoblastoma and CUTL1 family members.

The inactivation of certain tumor suppressor genes is thought to play an important role in the genesis of a number of tumor types. For example, inactivation of the Retinoblastoma (Rb) tumor suppressor is frequently observed in a proportion of sporadic human breast cancers. While these studies suggest that inactivation of key tumor suppressor genes may play an important role in the induction of mammary cancers, direct evidence supporting this contention is lacking. Because polyomavirus (PyV) Large T (LT) antigen is known to associate with and inactivate certain members of the Rb family (p105Rb, p107, p130), we have derived transgenic mice which express PyV LT antigen in the mammary epithelium. As expected mammary epithelial-specific expression of PyV LT antigen resulted in the induction of mammary tumors which correlated with their capacity to associate with Rb family members. In addition to mammary carcinomas, female transgenic mice expressing the PyV LT transgene frequently develop uterine leiomyomas. Because loss of heterozygosity involving the human CUTL1 (Cut like 1) gene located at chromosomal position 7q22 has been recently implicated in sporadic human uterine leiomyomas, we tested the hypothesis that PyV LT antigen may also form specific complexes with CUTL1. The results of these analyses revealed that specific complexes of CUTL1 and PyV LT antigen could be detected in both leiomyomas and mammary tumors. Taken together, these observations suggest that PyV LT antigen may be involved in inducing these tumors by sequestering both CUTL1 and Rb growth regulatory proteins.

DNA binding by cut homeodomain proteins is down-modulated by casein kinase II.

The Drosophila and mammalian Cut homeodomain proteins contain, in addition to the homeodomain, three other DNA binding regions called Cut repeats. Cut-related proteins thus belong to a distinct class of homeodomain proteins with multiple DNA binding domains. Using nuclear extracts from mammalian cells, Cut-specific DNA binding was increased following phosphatase treatment, suggesting that endogenous Cut proteins are phosphorylated in vivo. Sequence analysis of Cut repeats revealed the presence of sequences that match the consensus phosphorylation site for casein kinase II (CKII). Therefore, we investigated whether CKII can modulate the activity of mammalian Cut proteins. In vitro, a purified preparation of CKII efficiently phosphorylated Cut repeats causing an inhibition of DNA binding. In vivo, overexpression of the CKII alpha and beta caused a decrease in DNA binding by Cut. The CKII phosphorylation sites within the murine Cut (mCut) protein were identified by in vitro mutagenesis as residues Ser400, Ser789, and Ser972 within Cut repeat 1, 2, and 3, respectively. Cut homeodomain proteins were previously shown to function as transcriptional repressors. Overexpression of CKII reduced transcriptional repression by mCut, whereas a mutant mCut protein containing alanine substitutions at these sites was not affected. Altogether our results indicate that the transcriptional activity of Cut proteins is modulated by CKII.

Loss of heterozygosity and reduced expression of the CUTL1 gene in uterine leiomyomas.

Cytogenetic analyses has revealed deletions and/or rearrangments at several chromosomal positions in approximately half of uterine leiomyomas. The most frequent genetic alteration, deletion of 7q22, was found in approximately 35% of studied cases with cytogenetic abnormalities (128/366=35%). The same chromosomal band was also found to be deleted in a fraction of acute myeloid leukemias and myelodysplastic syndromes. The frequent deletion of 7q22 in some tumors suggest that a tumor suppressor gene may be located in this region. The human Cut-like homeobox gene, CUTL1, is one of the genes localized to 7q22 and it was shown previously to encode a transcriptional repressor that down-modulates the expression of c-Myc. Activation of the c-Myc oncogenic potential has been shown in many cancers to result from alterations in one or the other of its several mechanisms of regulation. These observations led us to hypothesize that CUTL1 could act as a tumor suppressor gene. In the present study, we have identified polymorphic markers within and directly adjacent to CUTL1 at 7q22 and demonstrated that these markers are present in a commonly deleted region in seven out of 50 uterine leiomyomas samples examined. Furthermore, Northern blot analysis revealed that CUTL1 mRNA levels were reduced in eight tumors out of 13. These results suggest that CUTL1 may act as a tumor suppressor gene whose inactivation could be of pathological importance in the etiology of uterine leiomyomas.

Cux/CDP homeodomain protein binds to an enhancer in the rat c-mos locus and represses its activity.

The c-mos gene is transcribed in male and female germ cells, in differentiating myoblasts and in 3T3 cells from cell-specific promoters. We characterized the rat testis promoter, which contains a TATA-box and one binding site for a testis-specific transcription factor TTF-D, as well as a region which can act as enhancer, which is located approx. 2 kb upstream of the c-mos AUG start codon. It binds three factors at sites I, II and III as determined in DNAse I footprint assays. We demonstrated that a member of the NF-1/CTF family of transcription factors binds site II. Here we report the cloning of the protein that binds to enhancer site III. This protein is the rat homolog of human hCut/CDP, mouse Cux/CDP and canine Clox. hCut/Cux/CDP/Clox (hereafter called Cux/CDP), a 160 kDa protein containing multiple repeats and a homeodomain, negatively regulates the mammalian c-myc, gp91-phox and N-Cam genes. Using bacterially produced murine GST-Cux fusion proteins and GST-Cux deletion mutants, we find that Cux repeat CR3 and the homeodomain are both required for efficient binding to enhancer site III. Mouse lung and testis nuclear Cux/CDP bind to site III as determined in electrophoretic gel mobility supershift assays using two different anti-hCut specific monoclonal antibodies. Transfections of CAT constructs containing the enhancer fragment linked to a minimal promoter demonstrated that Cux/CDP represses c-mos enhancer activity.

DNA binding by cut homeodomain proteins is down-modulated by protein kinase C.

The Drosophila and mammalian Cut homeodomain proteins contain, in addition to the homeodomain, three other DNA binding regions called Cut repeats. Cut-related proteins thus belong to a distinct class of homeodomain proteins with multiple DNA binding domains. Using nuclear extracts from mammalian cells, Cut-specific DNA binding was increased following phosphatase treatment, suggesting that endogenous Cut proteins are phosphorylated in vivo. Sequence analysis of Cut repeats revealed the presence of sequences that match the consensus phosphorylation site for protein kinase C (PKC). Therefore, we investigated whether PKC can modulate the activity of mammalian Cut proteins. In vitro, a purified preparation of PKC efficiently phosphorylated Cut repeats, which inhibited DNA binding. In vivo, a brief treatment of cells with calphostin C, a specific inhibitor of PKC, led to an increase in Cut-specific DNA binding, whereas phorbol 12-myristate 13-acetate, a specific activator of PKC, caused a decrease in DNA binding. The PKC phosphorylation sites within the murine Cut (mCut) protein were identified by in vitro mutagenesis as residues Thr415, Thr804, and Ser987 within Cut repeats 1-3, respectively. Cut homeodomain proteins were previously shown to function as transcriptional repressors. Activation of PKC by phorbol 12-myristate 13-acetate reduced transcriptional repression by mCut, whereas a mutant mCut protein containing alanine substitutions at these sites was not affected. Altogether, our results indicate that the transcriptional activity of Cut proteins is modulated by PKC.

The human cut homeodomain protein can repress gene expression by two distinct mechanisms: active repression and competition for binding site occupancy.

By analogy with other homeodomain proteins conserved in evolution, mammalian Cut proteins are believed, as in Drosophila melanogaster, to play an important role in determining cell type specificity in several tissues. At the molecular level, Cut proteins appear to serve as transcriptional repressors. In this study, we have examined the mechanism by which the human Cut (hCut) protein down-regulates gene expression. The homeodomain and the three regions called Cut repeats are evolutionarily conserved and were previously shown to function as DNA binding domains. The carboxy-terminal region, although it does not show amino acid sequence homology per se, in all cases is enriched in alanine and proline residues, a distinctive feature of some transcriptional repression domains. Our results reveal two distinct modes of repression: competition for binding site occupancy and active repression. On one hand, the composite DNA binding domain formed by Cut repeat 3 and the Cut homeodomain was shown to bind to CCAAT and Sp1 sites within the tk gene promoter and to reduce gene expression, presumably by preventing activation by the corresponding transcription factors. On the other hand, the carboxy-terminal region of mammalian Cut proteins was found to function as an active repression domain in a distance-independent manner. We have further narrowed this activity to two subdomains that can independently repress activated transcription. Finally, we present a model to illustrate the two mechanisms by which Cut proteins repress gene expression.

Plasminogen activator inhibitor 1 messenger RNA expression and molecular evidence for del(7)(q22) in uterine leiomyomas.

We analyzed the expression of plasminogen activator inhibitor 1 (PAI-1) in 16 leiomyomas and adjacent myometrium of women who underwent a hysterectomy while in the proliferative (n = 8) and secretory phases (n = 8) of the menstrual cycle. We localized the PAI-1 and its mRNA expression in smooth muscle and vessel endothelial cells of uterine tissues using immunocytochemistry and in situ hybridization. The expression of PAI-1 mRNA was higher in 11 (68.75%) of 16 leiomyomas compared with the adjacent myometrium (leiomyoma/myometrium ratio, 1.4-3.0; mean, 2.045). The leiomyoma:myometrium ratio of PAI-1 mRNA expression did not change during the proliferative (Phase I) and secretory (Phase II) phases of the menstrual cycle. In the remaining five samples, the leiomyoma:myometrium ratio of PAI-1 mRNA expression was close to 1 (0.8-1.2; mean, 0.92). Because the locus of the PAI-1 gene is on chromosome 7q22, we screened for loss of heterozygosity (LOH) in these samples using the PAI-1 marker and D7S471, an anonymous marker 12 cM telomeric to PAI-1. Four of five samples with low leiomyoma:myometrium ratio had LOH for the PAI-1 and/or D7S471 markers. The fifth sample demonstrated a noninformative analysis for these markers but had LOH for the D7S515, D7S666, and D7S518 markers, all centromeric to PAI-1. Because del(7)(q22), associated with a relatively low PAI-1 mRNA expression, can deregulate matrix proteinases and growth factors' activity in leiomyomas, it is conceivable that del(7)(q22) results in heterogeneous leiomyoma biology.

DNA-binding specificity of the cut repeats from the human cut-like protein.

The Drosophila Cut and mammalian Cut-like proteins contain, in addition to the homeodomain, three other DNA-binding regions called Cut repeats. Cut-like proteins, therefore, belong to a distinct class of homeodomain proteins with multiple DNA-binding domains. In this study, we assessed the DNA-binding specificity of the human Cut repeats by performing PCR-mediated random oligonucleotide selection with glutathione S-transferase fusion proteins. Cut repeat 1, Cut repeat 3, and Cut repeat 3 plus the homeodomain selected related yet distinct sequences. Therefore, sequences selected by one of the fusion proteins were often, but not always, recognized by the other proteins. Consensus binding sites were derived for each fusion protein. In each case, however, some selected sequences diverged from the consensus but were confirmed to be high-affinity recognition sites by electrophoretic mobility shift assay. We conclude that Cut DNA-binding domains have broad, overlapping DNA-binding specificities. Determination of dissociation constants indicated that in addition to the core consensus, flanking sequences have a moderate but significant effect on sequence recognition. Evidence from electrophoretic mobility shift assay, DNase footprinting, and dissociation constant analyses strongly suggested that glutathione S-transferase/Cut fusion proteins bind to DNA as dimers. The implications of these findings are discussed in relation to the DNA-binding capabilities of Cut repeats. In contrast to other studies, we found that the human Cut-like protein does not preferably bind to a site that includes an ATTA homeodomain-binding motif. Here we demonstrate that the native human Cut-like protein recognizes more efficiently a site containing an ATCGAT core consensus flanked with G/C-rich sequences.