Targeting histone deacetylase complexes via KRAB-zinc finger proteins: the PHD and bromodomains of KAP-1 form a cooperative unit that recruits a novel isoform of the Mi-2alpha subunit of NuRD.

Macromolecular complexes containing histone deacetylase and ATPase activities regulate chromatin dynamics and are vitally responsible for transcriptional gene silencing in eukaryotes. The mechanisms that target these assemblies to specific loci are not as well understood. We show that the corepressor KAP-1, via its PHD (plant homeodomain) and bromodomain, links the superfamily of Krüppel associated box (KRAB) zinc finger proteins (ZFP) to the NuRD complex. We demonstrate that the tandem PHD finger and bromodomain of KAP-1, an arrangement often found in cofactor proteins but functionally ill-defined, form a cooperative unit that is required for transcriptional repression. Substitution of highly related PHD fingers or bromodomains failed to restore repression activity, suggesting high specificity in their cooperative function. Moreover, single amino acid substitutions in either the bromodomain or PHD finger, including ones that mimic disease-causing mutations in the hATRX PHD finger, abolish repression. A search for effectors of this repression function yielded a novel isoform of the Mi-2alpha protein, an integral component of the NuRD complex. Endogenous KAP-1 is associated with Mi-2alpha and other components of NuRD, and KAP-1-mediated silencing requires association with NuRD and HDAC activity. These data suggest the KRAB-ZFP superfamily of repressors functions to target the histone deacetylase and chromatin remodeling activities of the NuRD complex to specific gene promoters in vivo.

Solution structure of the PHD domain from the KAP-1 corepressor: structural determinants for PHD, RING and LIM zinc-binding domains.

Plant homeodomain (PHD) domains are found in >400 eukaryotic proteins, many of which are transcriptional regulators. Naturally occurring point mutations or deletions of this domain contribute to a variety of human diseases, including ATRX syndrome, myeloid leukemias and autoimmune dysfunction. Here we report the first structural characterization of a PHD domain. Our studies reveal that the PHD domain from KAP-1 corepressor binds zinc in a cross-brace topology between anti-parallel ss-strands reminiscent of RING (really interesting new gene) domains. Using a mutational analysis, we define the structural features required for transcriptional repression by KAP-1 and explain naturally occurring, disease-causing mutations in PHD domains of other proteins. From a comparison of this PHD structure with previously reported RING and LIM (Lin11/Isl-1/Mec-3) structures, we infer sequence determinants that allow discrimination among PHD, RING and LIM motifs.

Reconstitution of the KRAB-KAP-1 repressor complex: a model system for defining the molecular anatomy of RING-B box-coiled-coil domain-mediated protein-protein interactions.

The KRAB domain is a 75 amino acid residue transcriptional repression module commonly found in eukaryotic zinc-finger proteins. KRAB-mediated gene silencing requires binding to the corepressor KAP-1. The KRAB:KAP-1 interaction requires the RING-B box-coiled coil (RBCC) domain of KAP-1, which is a widely distributed motif, hypothesized to be a protein-protein interface. Little is known about RBCC-mediated ligand binding and the role of the individual sub-domains in recognition and specificity. We have addressed these issues by reconstituting and characterizing the KRAB:KAP-1-RBCC interaction using purified components. Our results show that KRAB binding to KAP-1 is direct and specific, as the related RBCC domains from TIF1alpha and MID1 do not bind the KRAB domain. A combination of gel filtration, analytical ultracentrifugation, chemical cross-linking, non-denaturing gel electrophoresis, and site-directed mutagenesis techniques has revealed that the KAP-1-RBCC must oligomerize likely as a homo-trimer in order to bind the KRAB domain. The RING finger, B2 box, and coiled-coil region are required for oligomerization of KAP-1-RBCC and KRAB binding, as mutations in these domains concomitantly abolished these functions. KRAB domain binding stabilized the homo-oligomeric state of the KAP-1-RBCC as detected by chemical cross-linking and velocity sedimentation studies. Mutant KAP-1-RBCC molecules hetero-oligomerize with the wild-type KAP-1, but these complexes were inactive for KRAB binding, suggesting a potential dominant negative activity. Substitution of the coiled-coil region with heterologous dimerization, trimerization, or tetramerization domains failed to recapitulate KRAB domain binding. Chimeric KAP-1-RBCC proteins containing either the RING, RING-B box, or coiled-coil regions from MID1 also failed to bind the KRAB domain. The KAP-1-RBCC mediates a highly specific, direct interaction with the KRAB domain, and it appears to function as an integrated, possibly cooperative structural unit wherein each sub-domain contributes to oligomerization and/or ligand recognition. These observations provide the first principles for RBCC domain-mediated protein-protein interaction and have implications for identifying new ligands for RBCC domain proteins.

Expression of OVCA1, a candidate tumor suppressor, is reduced in tumors and inhibits growth of ovarian cancer cells.

Loss of all or part of one copy of chromosome 17p is very common in ovarian and breast tumors. OVCA1 is a candidate tumor suppressor gene mapping to a highly conserved region on chromosome 17p13.3 that shows frequent loss of heterozygosity in breast and ovarian carcinomas. Western blot analysis of extracts prepared from breast and ovarian carcinomas revealed reduced expression of OVCA1 compared with extracts from normal epithelial cells from these tissues. Subcellular localization studies indicate that OVCA1 is localized to punctate bodies scattered throughout the cell but is primarily clustered around the nucleus. Attempts to create cell lines that stably expressed OVCA1 from the cytomegalovirus promoter were generally unsuccessful in a variety of different cell lines. This reduction of colony formation was quantified in the ovarian cancer cell line A2780, where it was demonstrated that cells transfected with plasmids expressing OVCA1 had a 50-60% reduction in colony number as compared with appropriate controls, and only a few of these clones expressed OVCA1, albeit at low levels. The clones that expressed exogenous OVCA1 were found to have dramatically reduced rates of proliferation. Reduced growth rates correlated with an increased proportion of the cells in the G1 fraction of the cell cycle compared with the parental cell line and decreased levels of cyclin D1. The low levels of cyclin D1 appeared to be caused by an accelerated rate of cyclin D1 degradation. Overexpression of cyclin D1 was able to override OVCA1's suppression of clonal outgrowth. These results suggest that slight alterations in the level of OVCA1, such as would occur after reduction of chromosome 17p13.13 to hemizygosity, may result in cell cycle deregulation and promote tumorigenesis.

Structure-function studies of the BTB/POZ transcriptional repression domain from the promyelocytic leukemia zinc finger oncoprotein.

The evolutionarily conserved BTB/POZ domain from the promyelocytic leukemia zinc finger (PLZF) oncoprotein mediates transcriptional repression through the recruitment of corepressor proteins containing histone deacetylases in acute promyelocytic leukemia. We have determined the 2.0 A crystal structure of the BTB/POZ domain from PLZF (PLZF-BTB/POZ), and have carried out biochemical analysis of PLZF-BTB/POZ harboring site-directed mutations to probe structure-function relationships. The structure reveals a novel alpha/beta homodimeric fold in which dimer interactions occur along two surfaces of the protein subunits. The conservation of BTB/POZ domain residues at the core of the protomers and at the dimer interface implies an analogous fold and dimerization mode for BTB/POZ domains from otherwise functionally unrelated proteins. Unexpectedly, the BTB/POZ domain forms dimer-dimer interactions in the crystals, suggesting a mode for higher-order protein oligomerization for BTB/POZ-mediated transcriptional repression. Biochemical characterization of PLZF-BTB/POZ harboring mutations in conserved residues involved in protein dimerization reveals that the integrity of the dimer interface is exquisitely sensitive to mutation and that dimer formation is required for wild-type levels of transcriptional repression. Interestingly, similar mutational analysis of residues within a pronounced protein cleft along the dimer interface, which had been implicated previously for interaction with corepressors, has negligible effects on dimerization or transcriptional repression. Together, these studies form a structure-function framework for understanding BTB/POZ-mediated oligomerization and transcriptional repression properties.

KAP-1 corepressor protein interacts and colocalizes with heterochromatic and euchromatic HP1 proteins: a potential role for Krüppel-associated box-zinc finger proteins in heterochromatin-mediated gene silencing.

Krüppel-associated box (KRAB) domains are present in approximately one-third of all human zinc finger proteins (ZFPs) and are potent transcriptional repression modules. We have previously cloned a corepressor for the KRAB domain, KAP-1, which is required for KRAB-mediated repression in vivo. To characterize the repression mechanism utilized by KAP-1, we have analyzed the ability of KAP-1 to interact with murine (M31 and M32) and human (HP1alpha and HP1gamma) homologues of the HP1 protein family, a class of nonhistone heterochromatin-associated proteins with a well-established epigenetic gene silencing function in Drosophila. In vitro studies confirmed that KAP-1 is capable of directly interacting with M31 and hHP1alpha, which are normally found in centromeric heterochromatin, as well as M32 and hHP1gamma, both of which are found in euchromatin. Mapping of the region in KAP-1 required for HP1 interaction showed that amino acid substitutions which abolish HP1 binding in vitro reduce KAP-1 mediated repression in vivo. We observed colocalization of KAP-1 with M31 and M32 in interphase nuclei, lending support to the biochemical evidence that M31 and M32 directly interact with KAP-1. The colocalization of KAP-1 with M31 is sometimes found in subnuclear territories of potential pericentromeric heterochromatin, whereas colocalization of KAP-1 and M32 occurs in punctate euchromatic domains throughout the nucleus. This work suggests a mechanism for the recruitment of HP1-like gene products by the KRAB-ZFP-KAP-1 complex to specific loci within the genome through formation of heterochromatin-like complexes that silence gene activity. We speculate that gene-specific repression may be a consequence of the formation of such complexes, ultimately leading to silenced genes in newly formed heterochromatic chromosomal environments.

Cloning and localization of a human diphthamide biosynthesis-like protein-2 gene, DPH2L2.

Sequence analysis of the candidate tumor suppressor OVCA1 revealed extensive sequence identity and similarity to proteins from a diverse number of species, including the yeast diphthamide biosynthesis protein-2, dph2, which suggested that OVCA1 may be the human homologue to this yeast gene. However, searches of the translated EST database for sequences in common with dph2 and OVCA1 uncovered an EST, h52976, with significant amino acid conservation with dph2. Isolation of a cDNA clone encompassing the EST by RACE methodologies and sequence analysis indicate the identification of a previously unidentified gene that is ubiquitously expressed and maps to chromosome 1p34. Based on amino acid sequence analysis, the 489-amino-acid protein encoded by this novel gene is distinct from OVCA1 and is more closely related to the yeast dph2 gene product. Therefore, we refer to this novel gene as DPH2L2, which constitutes one member of a novel gene family that may be involved in diphthamide biosynthesis in humans.

BAP1: a novel ubiquitin hydrolase which binds to the BRCA1 RING finger and enhances BRCA1-mediated cell growth suppression.

We have identified a novel protein, BAP1, which binds to the RING finger domain of the Breast/Ovarian Cancer Susceptibility Gene product, BRCA1. BAP1 is a nuclear-localized, ubiquitin carboxy-terminal hydrolase, suggesting that deubiquitinating enzymes may play a role in BRCA1 function. BAP1 binds to the wild-type BRCA1-RING finger, but not to germline mutants of the BRCA1-RING finger found in breast cancer kindreds. BAP1 and BRCA1 are temporally and spatially co-expressed during murine breast development and remodeling, and show overlapping patterns of subnuclear distribution. BAP1 resides on human chromosome 3p21.3; intragenic homozygous rearrangements and deletions of BAP1 have been found in lung carcinoma cell lines. BAP1 enhances BRCA1-mediated inhibition of breast cancer cell growth and is the first nuclear-localized ubiquitin carboxy-terminal hydrolase to be identified. BAP1 may be a new tumor suppressor gene which functions in the BRCA1 growth control pathway.

Identification of a gene containing zinc-finger motifs based on lost expression in malignantly transformed rat ovarian surface epithelial cells.

We have used a rat model of epithelial ovarian cancer to identify a gene that shows decreased or lost expression in five of eight independently transformed rat ovarian surface epithelial cell lines compared to the normal progenitor cells. Hence, we refer to this gene as Lot1 (lost on transformation 1; GenBank accession no. U72620). The most abundant transcript of the gene is approximately 6 kb. This sequence contains a 1749-nucleotide open reading frame and, within the 3' untranslated region, 22 near-perfect 60-70-bp repeats and adenine- and uracil-rich areas. The deduced amino acid sequence from the open reading frame contains seven zinc-finger motifs of the C2H2 type, as well as proline-, glutamine-, and glutamic acid-rich areas. The gene maps to the short arm of chromosome one in the rat. Lot1 shows a limited distribution of expression in normal rat tissues, including ovary, which shows abundant expression. Furthermore, examination of DNA derived from multiple species indicates that the gene is widely conserved.

Identification of a zinc-finger gene at 6q25: a chromosomal region implicated in development of many solid tumors.

We have used a rat model of epithelial ovarian cancer to identify a gene that shows decreased or lost expression in independently transformed rat ovarian surface epithelial cell lines compared to the normal progenitor cells. Hence, we refer to this gene as Lot-1 (Lost on transformation 1, GenBank accession no. U72620). Here, we report the cloning of the likely human homologue and its initial characterization. The deduced amino acid sequences of the cDNAs for rat and human LOT-1 (GenBank accession no. U72621) contain seven zinc finger motifs of the C2H2 type as well as proline and glutamine rich areas. The genes share 76.4% identity at the nucleotide level, 67.7% at the amino acid level and 85.5% within the seven zinc finger motifs. LOT-1 is ubiquitously expressed in normal human tissues but was not expressed in four of 11 (36%) human ovarian cancer cell lines or spontaneously transformed human ovarian surface epithelial cells. The human gene maps to chromosome 6 at band q25. We show that there is a 38% incidence of allelic loss at this chromosomal location in human ovarian cancers. This chromosomal region has also been implicated in the genesis of breast, kidney, and pleural mesothelial cancers. We suggest that this newly identified gene is not only of intrinsic interest as a ubiquitously expressed probable transcription factor but is a plausible candidate for the tumor suppressor gene which likely resides in the region of chromosome 6 defined by band q25.

A common mutation in BRCA2 that predisposes to a variety of cancers is found in both Jewish Ashkenazi and non-Jewish individuals.

Recent studies have identified mutations in the breast and (ovarian cancer susceptibility gene 2 (BRCA2), one which has been found in the germline of several males and one female affected with breast cancer. To establish the carrier frequency of this mutation in a large population of individuals affected with cancer, we evaluated constitutional DNA isolated from 83 individuals diagnosed with breast cancer and 93 diagnosed with ovarian cancer at any age, 42 of whom reported a family history of cancer. Using a simple allele-specific PCR-based nonradioactive method, we detected a total of eight individuals (4.5%) carrying a 1-bp deletion at nucleotide 6174 of the BRCA2 gene (6174delT). The age of disease onset in the mutant allele carriers was highly variable and typically late onset (41-72 years for breast cancer and 48-73 years for ovarian cancer). Evaluation of family histories for the eight mutant allele carriers revealed that several individuals had significant cancer histories that included, in addition to breast and/or ovarian cancer, an increased incidence of colon, esophageal, pancreatic, stomach, and hematopoietic cancers. Interestingly, seven of the eight individuals were of Ashkenazi Jewish descent. Haplotype data for the mutant allele carriers using markers spanning the region of the BRCA2 gene on chromosome 13ql2-ql3 suggest that only two of the confirmed Jewish Ashkenazi individuals share a single common ancestry, indicating several independent origins for this mutation. These data provide evidence for the presence of a specific BRCA2 mutation which has its origins in both Jewish Ashkenazi and non-Jewish populations. The observed overrepresentation of specific mutations within a subgroup of the general population may eventually help contribute to the development of inexpensive and routine tests such as the one described in our study.

Two distinct origins of a common BRCA1 mutation in breast-ovarian cancer families: a genetic study of 15 185delAG-mutation kindreds.

We screened 163 women from breast-ovarian cancer-prone families, as well as 178 individuals affected with breast and/or ovarian cancer but unselected for family history, for germ-line mutations in exon 2 of BRCA1, by SSCP analysis and direct sequencing. A total of 25 mutations were detected. Thirteen of 64 Jewish Ashkenazi women and 2 non-Jewish individuals were found to possess the 185delAG mutation. Haplotype data for all 15 individuals, with markers intragenic to BRCA1, suggest that the Jewish Ashkenazi individuals share a common ancestry that is distinct from the lineage shared by the other two women. These data provide the first evidence of two distinct lines of transmission for the 185delAG mutation, only one of which has its origins in the Jewish Ashkenazi population. Our screening also uncovered 10 affected individuals with an 11-bp deletion at nucleotide 188 of BRCA1 (188del11), 4 of whom are Ashkenazi Jews. This is only the third reported mutation detected within the Jewish Ashkenazi population and may represent the second most common alteration in BRCA1 found in Ashkenazi Jews in the United States. The observed overrepresentation of specific mutations within a subgroup of the general population may eventually contribute to the development of inexpensive and routine tests for BRCA1 mutations, as well as to the elucidation of other contributory factors (e.g., diet, environment, and chemical exposures) that may play a key role in cancer initiation and development. The implications of the mutational data, as well as the role that founder effect, demographic history, and penetrance play in the resulting observed phenomena, are discussed.

Identification of two candidate tumor suppressor genes on chromosome 17p13.3.

A second tumor suppressor locus on 17p that is distinct from TP53 has been identified in brain, breast, lung, and ovarian tumors. Using allelic loss mapping and positional cloning methods, we have recently identified two novel genes, which we refer to as OVCA1 and OVCA2, that map to 17p13.3. The two genes are ubiquitously expressed and encode proteins of 443 and 227 amino acids, respectively, with no known functional motifs. Sequence comparison of OVCA1 and OVCA2 revealed extensive sequence identity and similarity to hypothetical proteins from Saccharomyces cerevisiae, Caenorhabditis elegans, and Rattus species. Northern blot analysis reveals that OVCA1 and OVCA2 mRNA were expressed in normal surface epithelial cells of the ovary, but the level of this transcript is significantly reduced or is undetectable in 92% (11/12) of the ovarian tumors and tumor cell lines analyzed. The location, high degree of amino acid conservation, and reduced expression in ovarian tumors and tumor cell lines suggest that decreased expression of these two genes contributes to ovarian tumorigenesis and should be considered candidate tumor suppressor genes.

Characterization of chromosome 9 in human ovarian neoplasia identifies frequent genetic imbalance on 9q and rare alterations involving 9p, including CDKN2.

We have examined 41 forms of ovarian cancer for genetic alterations on chromosome 9 using a combination of five RFLP DNA probes and 15 simple tandem repeat polymorphisms. Genetic imbalance (i.e., loss of heterozygosity, microsatellite instability, amplification) for 1 or more informative markers on chromosome 9 was observed in 66% (27 of 41) of our tumor panel. Genetic imbalance was observed on 9q in 59% (24 of 41) of tumors informative for at least one locus. In contrast, only 13% (5 of 40) of informative tumors demonstrated a genetic alteration involving 9p. Furthermore, allelic loss on 9q was more common in late stage tumors (63%, 17 of 27) and poorly differentiated tumors (75%, 15 of 20) as compared to benign and early stage tumors (30%, 3 of 10). Evaluation of 15 tumors showing limited regions of genetic imbalance has identified 2 candidate tumor suppressor regions on 9q and 1 on 9p. Interestingly, the regions defined to 9p21-p24, 9q31, and 9q32-q34 all overlap with several known disease loci. In this aspect, the potential role of the CDKN2 gene at 9p21-p22 in ovarian carcinogenesis was assessed in an extended panel of ovarian tumors, 11 human ovarian carcinoma cell lines, and 1 cervical tumor cell line. With the use of comparative multiplex PCR, homozygous deletions were detected in 16 of 115 (14%) fresh tumors and 3 of 12 cancer cell lines. For those tumors demonstrating allelic loss for markers on 9p no somatic mutations were observed in the retained allele of CDKN2, as determined by single-strand conformation polymorphism analysis, but a mutation was observed in an additional cell line. Furthermore, CDKN2 mRNA levels were similar in the 9 cancer cell lines that retain CDKN2, as compared to normal human ovarian surface epithelial cell lines. Overall, our results suggest the potential involvement of a gene or genes on chromosome 9q and de-emphasize a significant role for the CDKN2 gene on 9p in the initiation and progression of ovarian cancer.

A common region of deletion on chromosome 17q in both sporadic and familial epithelial ovarian tumors distal to BRCA1.

Linkage analysis in familial breast and ovarian cancer and studies of allelic deletion in sporadic ovarian tumors have identified a region on chromosome 17q containing a candidate tumor-suppressor gene (referred to as BRCA1) of likely importance in ovarian carcinogenesis. We have examined normal and tumor DNA samples from 32 patients with sporadic and 8 patients with familial forms of the disease, for loss of heterozygosity (LOH) at 21 loci on chromosome 17 (7 on 17p and 14 on 17q). LOH on 17p was 55% (22/40) for informative 17p13.1 and 17p13.3 markers. When six polymorphic markers flanking the familial breast/ovarian cancer susceptibility locus on 17q12-q21 were used, LOH was 58% (23/40), with one tumor showing telomeric retention. Evaluation of a set of markers positioned telomeric to BRCA1 resulted in the highest degree of LOH, 73% (29/40), indicating that a candidate locus involved in ovarian cancer may reside distal to BRCA1. Five of the tumors demonstrating allelic loss for 17q markers were from individuals with a strong family history of breast and ovarian cancer. More important, two of these tumors (unique patient number [UPN] 57 and UPN 79) retained heterozygosity for all informative markers spanning the BRCA1 locus but showed LOH at loci distal to but not including the anonymous markers CMM86 (D17S74) and 42D6 (D17S588), respectively. Deletion mapping of seven cases (two familial and five sporadic) showing limited LOH on 17q revealed a common region of deletion, distal to GH and proximal to D17S4, that spans approximately 25 cM. These results suggest that a potential tumor-suppressor gene involved in both sporadic and familial ovarian cancer may reside on the distal portion of chromosome 17q and is distinct from the BRCA1 gene.

Western blotting and enzymatic activity analysis of cathepsin D in breast tissue and sera of patients with breast cancer and benign breast disease and of normal controls.

Increased total antigen amounts of cathepsin D in breast tissue have been reported to be associated with increased disease recurrence, more frequent metastasis, and increased mortality in breast cancer patients. In the present study, Western blotting analysis has been used for the first time to determine the relative amounts of precursor and processed forms of cathepsin D in sera and breast tissue of patients with breast cancer, benign breast disease, and normal controls. Sera gave similar blots for breast cancer patients and controls with two major forms of cathepsin D (M(r) 52,000 and 27,000). Malignant breast tissue contained the two forms of cathepsin D found in sera and an additional M(r) 31,000 form which was found in significantly increased (P < 0.001) relative amounts in breast tissue from 43 breast cancer patients [24 +/- 12% (SD)] when compared to 51 benign breast disease patients (13 +/- 8.9%) and 23 normal controls (1.8 +/- 4.4%). Preliminary analysis of subgroups of benign breast disease patients suggested no significant difference (P = 0.41) in relative amounts of the M(r) 31,000 form of cathepsin D between proliferative-type and non-proliferative-type fibrocystic breast disease. A cathepsin D assay has been optimized for human breast tissue and used to demonstrate for the first time significantly increased (P < 0.001) amounts of pepstatin-inhibitable, cathepsin D-specific activity in breast tissue from 36 breast cancer patients (2.2 +/- 1.4 units/mg of protein) when compared to 47 benign breast disease patients (0.63 +/- 0.43) and 23 normal controls (0.24 +/- 0.21). Preliminary analysis of subgroups of benign breast disease patients suggested no significant difference (P = 0.21) in pepstatin-inhibitable, cathepsin D-specific activity between proliferative-type and nonproliferative-type fibrocystic breast disease. The positive correlation (r = 0.82) of increased amounts of the M(r) 31,000 form of cathepsin D and increased pepstatin-inhibitable, cathepsin D enzymatic activity in malignant breast tissue suggests that the M(r) 31,000 form is the proteolytically active form of the enzyme which may be involved in the development and/or metastatic spread of breast cancer.