Identification of a zinc-finger gene at 6q25: a chromosomal ... - Nature

Oncogene (1997) 14, 1973 ± 1979  1997 Stockton Press All rights reserved 0950 ± 9232/97 $12.00

Identi®cation of a zinc-®nger gene at 6q25: a chromosomal region implicated in development of many solid tumors Abbas Abdollahi1, David Roberts1,2, Andrew K Godwin1, David C Schultz1,2, Gonosuke Sonoda1, Joseph R Testa1 and Thomas C Hamilton1,2 1

Department of Medical Oncology, Fox Chase Cancer Center, Philadelphia, PA 19111 and 2Department of Chemistry, Lehigh University, Bethlehem, Pennsylvania 18015, USA

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 ®nger 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 ®nger 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 identi®ed 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 de®ned by band q25. Keywords: Cancer; ovary; transcription factor; zinc®nger

Introduction Unraveling the molecular genetic basis for ovarian cancer is proving to be dicult. These tumors generally arise from the modi®ed peritoneal mesothelium covering the ovarian surface (Bell, 1991). It is unclear whether benign precursor lesions derived from these cells precede the development of occult malignancy (Cheng et al., 1996; Salazar et al., 1996) and in most cases the disease is discovered in advanced stage (Ozols et al., 1992). Molecular analysis of these late stage

Correspondence: TC Hamilton 1 Current address: Department of Medical Oncology, Fox Chase Cancer Center, 7701 Burholme Ave; Philadelphia, PA 19111, USA Received 13 September 1996; revised 26 December 1996; accepted 13 January 1997

cancers reveal they contain abundant genetic alterations (Godwin et al., 1996; Tenaka et al., 1989). In fact, allotyping studies show some frequency of loss of heterozygosity on most chromosomes (Cliby et al., 1993; Osborne and Leech, 1994; Sato et al., 1991; Yang-Feng et al., 1993). This has complicated selection of chromosomal areas from which to initiate deletion mapping and positional cloning eects to identify ovarian cancer genes. As an alternative to positional cloning to identify ovarian cancer genes, we have developed a tissue culture model of the disease using growth stress to induce malignant transformation of rat ovarian surface epithelial cells. This has allowed the frequency of gene expression dierences between normal cells and independent malignant transformants to be examined in cells that initially had the identical genetic constitution (Godwin et al., 1992; Testa et al., 1994). Using a method referred to as dierential display (Johnson et al., 1996; Liang et al., 1993; Liang and Pardee, 1992), we identi®ed a transcript which was expressed in normal ovarian surface epithelial cells but not the related transformed cells. Using this starting point, we cloned Lot-1 (Lost on transformation-1, GenBank accession no. U72620). The product of this previously unidenti®ed gene is a protein containing seven zinc ®nger motifs of the C2H2 class (Abdollahi et al., 1997; Hamilton et al., 1996). Multispecies Southern blot analysis revealed that closely related genes likely exist in most species. We hoped to gain some insight as to the relevance of this gene to human disease by mapping its chromosomal location in the rat and inferring its location in human chromosomes based on proximity to known genes in the rat and synteny between these regions in the rat and human. Unfortunately, Lot-1 mapped to the short arm of chromosome one and to our knowledge is the only gene mapped to this chromosomal arm of the rat. Here, we describe cloning the human cognate of Lot-1. We show that LOT-1 (GenBank accession no. U72621) maps to human chromosome 6 at band q25. This general chromosomal region has been implicated in the genesis of ovarian, breast, kidney, and pleural mesothelial cancers (Fujii et al., 1996; Taguchi et al., 1993; Theile et al., 1996; Thrash-Bingham et al., 1995). To expand these data, we report allelic loss at this chromosomal location in a series of 47 ovarian cancer specimens. Furthermore, we show the gene is ubiquitously expressed in normal human tissues, but that its expression is frequently decreased or lost in human ovarian cancer cell lines and an in vitro transformation model of human ovarian cancer.

Potential target gene for 6q25 LOH A Abdollahi et al

1974

Results Identi®cation and characterization of the human homologue of rat Lot-1 Cloning the human homologue of Lot-1 was aided by comparison of rLot-1 to sequences in various nucleotide and protein databases. One match showed an 89.9% identity with 290 bp of sequence derived from an anonymous cDNA sequence (F08789; clone c2ah04) from a normal infant brain library submitted by the IMAGE (Integrated Molecular Analysis of the Human Genome and its Expression) Consortium Expressed Sequence Tag (EST) project. Sequencing this EST in combination with clone walking within a human fetal brain library produced a composite cDNA of a closely related human gene. The human cDNA contained an open reading frame of 1392 nucleotides bounded 5' by an initiation codon (ATG) and 3' by TAA (data not shown). It is noteworthy that the ®rst methionine codons of both the rat and human cDNAs are preceded by an in frame TGA stop codon and a nearly identical stretch of 12 bp (rat=TGAGAAGCAGAGGCCATGGC versus human=TGAGAAGCAAAGCCCATGGC). Additionally, the concluding 33 bases of both open reading frames provides for 11 identical deduced amino acids (Figure 1). Overall, human LOT-1 encodes a 464 amino acid protein with a molecular weight of 51 kDa. Comparison of the rLot-1 and hLOT-1 nucleotide and predicted amino acid

Figure 1 Amino acid sequence alignment of the predicted rat and human LOT-1 proteins. The ®gure shows the seven zinc ®nger domains (underlined and designated as ZF 1 to 7) that are shared in the N-terminal region of both human (hLOT1) and rat (rLOT1) proteins. The sequence comparison also reveals shared regions rich in proline and glutamine residues downstream of the ZF region and in the case of rLOT1 a glutamic acid-rich segment near the C-terminus. A high leucine content is also characteristic of both proteins (12% in the human and 9% in the rat). The single-letter abbreviations for the amino acid residues are as follows: A, Ala; C, Cys; D, Asp; E, Glu; F, Phe; G, Gly; H, His; I, Ile; K, Lys; L, Leu; M, Met; N, Asn, P, Pro; Q, Gln; R, Arg; S, Ser; T, Thr; V, Val; W, Trp; and Y, Tyr

sequences revealed a high degree of conservation: 76.4% percent similarity at the nucleotide level and 67.7% identity at the amino acid level (Figure 1). In Figure 2a, the seven zinc ®ngers of hLOT-1 are compared to each other. The amino acid sequence in this region shows 85.5% identity with the rat. The deduced amino acid sequence for the seventh zinc ®nger in hLOT-1 and rLOT-1 as well as the linker between ®nger six and seven are compared to one of the zinc ®nger motifs with adjoining amino terminal linkers contained in several known proteins in Figure 2b. Additional features of note are proline and glutamine rich regions in the carboxyl terminal portions of both the human (residues 220 ± 444) and rat (residues 220 ± 458) proteins. The deduced rat protein also contains clustered glutamic acids within residues 490 ± 540. Comparison of the nucleotide and deduced sequence data on hLOT-1 and rLot-1 to information in data bases did not reveal the existence of previously identi®ed homologous genes. However, numerous ESTs related to LOT-1 were present in the `GenBank'. Expression of LOT-1 in normal and malignant human tissues Examination of LOT-1 expression by Northern blotting using poly(A)+ RNA of normal human tissues revealed that the gene is ubiquitously expressed (Figure 3). Two prominent, transcripts, *3.9 kb and 2.8 kb in size, were seen in all tissues examined, i.e. spleen, thymus, prostate, testis, ovary, intestine, colon, leukocytes. In addition, large transcripts were most obvious in ovarian poly(A)+ RNA and to a lesser extent in the other tissues. These may be due to incompletely processed pre-mRNA. LOT-1 expression was not detected in four of 11 (36%) human ovarian cancer cell lines (Figure 4a). It is also noteworthy that when expression was evident it varied substantially between individual cell lines. When present, both the 3.9 kb (Figure 4a) and less abundant

Figure 2 Alignment of zinc ®nger region in LOT-1 protein. (a) Comparison of Cys- and His-rich repeats of the seven zinc ®ngers derived from the full-length amino acid sequence of the hLOT-1 protein as shown in Figure 4. (b) Homology of deduced amino acid sequence of rat and human LOT-1 to the zinc ®ngers of other proteins. The alignment shows the early growth response (EGR) genes, Wilms' tumor 1 gene (WT1) and hLOT-1 and rLOT-1 (boxed amino acid residues). The sequence shown is of the His portion of zinc ®nger (ZF) 6, the linker sequence, and complete zinc ®nger 7


Normal 26

HIO 135

Rat

Human

PEO 1

PEO 4

SKOV 3

OVCAR 10

OVCAR 8

OVCAR 7

OVCAR 5

OVCAR 4

OVCAR 3

OVCAR 2

b

Nutu 26

a

HIO 107

Figure 3 Human tissue Northern blot containing 2 mg of poly(A)+ RNA per lane. The probe (a-32PATP labeled) for hybridization was a 1.7 kbp cDNA fragment (ZFEST) encompassing the HindIII/NotI insert within the clone F0878 (Genome Systems, St Louis, MO)

HIO 117

GAPDH ➝

HIO 118

Leukocyte

Colon

Intestine

Ovary

2.4 ➝

28 S

18 S

28 S

➝

Testis

4.4 ➝

➝

Prostate

9.5 ➝ 7.5 ➝

We next mapped the chromosomal location of hLOT-1 by FISH. We examined 28 mitoses and found that 66.3% (59 of 89) of the signals detected were located on chromosome 6, band q25 (Figure 5). To determine the frequency of allelic loss in this chromosomal region in ovarian cancer, we evaluated 47 ovarian tumor/ normal tissue pairs for evidence of loss of heterozygosity (LOH) using ®ve highly polymorphic tandem repeat markers spanning a *22 cM region between bands 6q23 and 6q27. Overall, LOH was detected in 38% (18 of 47) of the tumors evaluated. The highest frequencies of allelic loss were observed for markers D6S1007 at q25-qter (33%) and D6S250 at q23 (32.5%) (Table 1). No clear cut correlation could be made between LOH and histopathological parameters such as subtype, stage, or grade due to the limited number of benign and early stage tumors evaluated. However, all but one of the ovarian tumors showing LOH for at least one marker was stage III or higher, and most were moderate to poorly dierentiated cancers. Only seven of the 18 tumors demonstrating LOH showed loss for all of the informative markers tested, indicating a high percentage of telomeric and/or interstitial deletions. In this aspect, two distinct minimum regions of allelic loss were identi®ed: an approximately 8.6 cM common region of loss located at 6q23-24, which is bounded by, but does not include the anonymous markers D6S250 and D6S441 (data not shown); a region in 6q25 de®ned by a recurrent

➝

Thymus

Chromosomal localization of hLOT-1 and allelic loss in ovarian cancer

kb

A2780

immortalized human ovarian surface epithelial cell line which was induced to undergo malignant transformation by growth stress (Figure 4b).

➝

Spleen

2.8 kb (data not shown) transcripts were seen. RNA from immortalized but non-tumorigenic human ovarian surface epithelial cells also contained 3.9 kb and 2.8 kb transcripts. Both transcripts were absent in one

18 S

3.9 Kbp

28 S

➝ ➝ ➝

18 S

Figure 4 Analysis of LOT-1 expression in human ovarian cancer cell lines. (a) Northern blots containing 15 mg of total RNA were hybridized with a cDNA fragment (1.7 kbp) of LOT-1 encompassing nucleotides 811 to 2536 and designated ZFEST (F08789, Genome Systems, St Louis, MO). Loading equivalency is shown by ethidium staining (lower panel). (b) Analysis of LOT-1 expression in total RNA (15 mg/lane) of four human ovarian surface epithelial cell strains (HIO cells). For comparison with the human transcript, rat total RNA (15 mg/lane) from a tumorigenic rat ovarian surface epithelial cell line (NuTu26) and the normal cells (Normal 26) from which the transformed cells arose are included. The probe was as is described in Figure 3

1975


1976

Figure 5 Digitized computer generated image of FISH with ¯uorecein-labeled LOT-1 probe to a normal human metaphase spread. Two pale/blue-green signals are seen against the darker blue chromosomes in the telomeric portion of the long arm of each copy of chromosome 6. The inset shows similar localization to copies of chromosome 6 from separate metaphases

Table 1 Loss of heterozygosity for chromosome 6q in ovarian cancer Distancea

Marker

Informative Cases

Cases showing LOH

% loss

8.6 4.8 2.0 5.7

D6S250 D6S441 D6S415 D6S1007 D6S1008

40 37 29 30 44

13 10 7 10 10

32.5 27.0 24.1 33.3 23.7

cM cM cM cM

a

Approximate sexed-averaged genetic distance between each marker based on a Genetic Location Database (University of Southampton)

adenocarcinoma (UPN 300) which shows loss for D6S415 and D6S1007 but retention of the ¯anking markers (i.e. D6S441 and D6S1008). Based on its mapped location (Figure 5) LOT-1 would be expected to reside within or very near to the latter region of recurrent deletions. Discussion The work presented here outlines the discovery and initial characterization of a previously unknown human gene which we refer to as LOT-1. This designation is based on its relationship to a recently described rat gene (Hamilton et al., 1996). Several facets of this endeavor and related work leading to the discovery of rat Lot-1 strongly infer the relevance of this gene's product in the development and/or clinical features of ovarian cancer and perhaps other solid tumors. Perhaps the most interesting aspect of the combined data relates to the utilization of rat ovarian surface epithelial cells and growth stress to induce their malignant transformation as the starting point of the work. Thus, the cell type from which most ovarian cancers arise (Bell, 1991) was induced to undergo malignant transformation by a process, i.e. the repeated requirement for growth as occurs with each ovulation, which much epidemiological evidence suggests is speci®cally relevant to ovarian cancer initiation (Fathalla, 1971; Fraceschi et al., 1991; Hamilton, 1992; Negri et al., 1991; Whittemore et al., 1992). Several additional data support the probable role of LOT-1 in ovarian carcinogenesis and/or some other

human disease(s). The fact that the human LOT-1 gene is expressed in all tissues examined and that there are numerous related ESTs derived from diverse cDNA libraries favors its fundamental role in homeostasis. It is the aberrant function of such homeostatic genes that have often been found to contribute to disease (Leach et al., 1993; Ohta et al., 1996; Schultz et al., 1996). It is noteworthy that even in the diverse genetic backgrounds which give rise to human ovarian cancer, which itself must arise through multiple pathways, we often see decreased or lost expression of LOT-1. Moreover, it is noteworthy that LOT-1 maps to chromosome 6q25. Several loss of heterozygosity studies have implicated the long arm of chromosome six as the probable site of one or more tumor suppressor genes involved in ovarian cancer (Orphanos et al., 1995; Saito et al., 1996; Shelling et al., 1995; Tibiletti et al., 1996; Wan et al., 1994; Zheng et al., 1991), but most interesting is a recent preliminary report re®ning a region of deletion to 4 cM at 6q25 (Mok et al., 1996). As described above, we have con®rmed this pattern of loss of heterozygosity in our panel of 47 ovarian cancer specimens, observing allelic losses from this region in 38% of cases, a frequency consistent with that observed by Mok and colleagues (Mok et al., 1996). Interestingly, genes that serve as templates for zinc ®nger proteins often are located in telomeric portions of chromosomes (Lichter et al., 1992) as is the case with LOT-1. The occurrence of somatic loss of genetic material as manifest by loss of heterozygosity is typical of the mechanism of inactivation of one allele of a tumor suppressor gene. Therefore, loss of heterozygosity at chromosome 6q25 suggests the existence of a tumor suppressor gene at this location. Of additional note, is the ®nding that microcell mediated transfer of a small chromosomal fragment containing a portion of 6q25 caused loss of the transformed phenotype in CAL51 breast cancer cells (Theile et al., 1996). As LOT-1 maps to 6q25, we suggest it to be a legitimate candidate for the 6q25 putative tumor suppressor gene. Additional evidence favoring this possibility is the pattern of lost or decreased expression of LOT-1 in ovarian cancers but not their normal progenitor cells. We should point out, however, that the region in question, although small in chromosomal terms, is suciently large that many genes may reside within it. Nonetheless, LOT-1 represents a worthy candidate based on the available data. Whether ongoing work solidi®es the role of LOT-1 in ovarian or other forms of cancer, its ubiquitous expression and its product's highly probable role as a zinc ®nger motif containing transcription factor makes it of intrinsic interest. Features in addition to the zinc ®nger motifs which are shared by the mRNAs of the rat and human genes are the high proline and glutamine contents of the translated region between the zinc ®nger segment and the carboxyl end of each molecule. Proline- and/or glutamine-rich regions have also been found in several DNA-binding proteins including EGR2/Krox-20 (Chavrier et al., 1988), KruÈppel gene product (Mermod et al., 1989), the Sp1 transcription factor (Courney and Tjian, 1988; Kadonaga et al., 1987), and WT1 (Gessler et al., 1990). Additionally, both transcripts contain AU-rich areas. Such sequences have been shown to markedly


decrease stability when present in the 3'-untranslated regions of many labile mammalian mRNAs such as those which encode cytokines and proto-oncogenes. (Caput et al., 1986; Peng et al., 1996; Shaw and Kamen, 1986; Vakalopoulon et al., 1991). Unique features of the rat compared to the human mRNA are a stretch of codons de®ning 42 glutamic acid residues found near the C-terminus and in the untranslated region 19 repeats 60 ± 70 bp each. With regard to this latter dierence, it would be interesting to speculate that this dierence results from alternate splicing between the two genes. This possibility, however, is excluded by failure to detect these sequences in human genomic DNA (Abdollahi et al., 1997). Hence, we must conclude that these sequences were eliminated from the human genome by some early homoid or prehomoid evolutionary selection. In summary, we have identi®ed a novel human gene based on its relationship to a recently discovered rat gene. We have shown it to be ubiquitously expressed in normal human tissues. Similar to the rat gene, the transcript is decreased or absent at high frequency in spontaneous and experimental human ovarian cancers. We have mapped the gene to human chromosome 6q25. This reveals a new syntenic region between rat and human. We have con®rmed that this chromosomal location is a frequent site of allelic loss in human ovarian cancer. Additionally, we have analysed the characteristics of the protein inferred by the deduced amino acid sequence and found its features to support its role as a transcription factor. Ongoing work will clarify whether LOT-1 is the tumor suppressor gene which LOH studies suggest resides at chromosome 6q25 and its role as a transcription factor.

Materials and methods Cells and cell culture Rat ovarian surface epithelial cells were obtained from the ovaries of adult female Fisher rats by selective trypsinization (Godwin et al., 1992; Testa et al., 1994). The cells were grown in vitro and induced to undergo malignant transformation as previously described (Godwin et al., 1992, 1995; Testa et al., 1994). Human ovarian cancer cells and ovarian surface epithelial cells were maintained in tissue culture as previously described (Auersperg et al., 1995; Johnson et al., 1994). The human ovarian surface epithelial cells studied were from ovarian cancer prone individuals and transfected with simian virus 40 (SV40) early genes (Dyck et al., 1996) to increase their life span. In this cell type, this process rarely was associated with true immortalization. Growth stress (Godwin et al., 1992; Testa et al., 1994) of one of the cell cultures, HIO 118, however, did result in immortalization and acquisition of the malignant phenotype, i.e. tumorigenicity in immunodeficient mice. Construction of cDNA libraries, screening, and analysis of clones To obtain the human cognate of rLot-1 a normal human fetal brain cDNA library in lZAPII vector (Stratagene) was used. For screening, several probes were generated. Most utilized was the rat DECCDNA-10 clone based on its homology to several Homo Sapiens EST clones identi®ed in

the GenBank/EMBL data base using the BLAST server in the NCBI program of the GCG package. One probe was based on the HindIII/NotI fragment within the EST clone identi®ed by the GenBank Accession Number F08789 (Genome System, St. Louis, MO). Sequences obtained on positive clones from lZAPII were aligned with ten additional available EST clones in the GenBank. The proteins and motifs of interest were analysed and identi®ed using SWISSprot, BEAUTY, and MOTIF databases. PCR Ampli®cation and Generation of Probes The ampli®cation reactions were carried out in a ®nal volume of 40 ml and contained 100 ng of template, 60 mM of dNTPs, 16PCR buer, 1 mM primer, 5% dimethyl sulfoxide, and 2.5 unit Amplitaq DNA Polymerase (Perkin Elmer). The cycling conditions consisted of 948C for 5 min, 19 cycles of 948C for 5 s, 658 ± 0.58C/cycle for 30 s, and 728C for 1 min, followed by an additional 25 cycles of 948C for 5 s, 558C for 30 s, 728C for 1 min, then extending at 728C for 5 min in a PTC-100 Thermal Cycler (MJ Research, Inc, Watertown, MA). PCR products were cloned, puri®ed, and labeled for use as probes by standard methods. RNA and DNA analysis Total RNA (Chomczynski and Mackey, 1995; Chomczynski and Sacchi, 1987) (15 mg/lane) was separated on 1% agarose gels containing 2.2 M formaldehyde. Transfer was to Nylon membranes (Micron Separations, Inc) by capillary action. The membranes were hybridized with a-32PdATP-labeled (DuPont-NEN, Boston, MA) cDNA probes by random priming. Prehybridization was performed at 428C in buer containing 50% formamide, 46SSC, 46Denhard's, 40 mM NaH2PO4 pH 6.5, 0.83% glycine, 160 mg/ml salmon sperm DNA, and 6.83% SDS adjusted to pH 7.4. The hybridization solution contained 50% formamide, 36SSC, 50 mM NaH2PO4, 112 mg/ml salmon sperm DNA, 1% SDS, 1.16Denhard's, and 10% dextran sulfate adjusted to pH 7.4. The membranes were washed for 1 h at 658C in 26SSC, 0.56SET (10% SDS, 0.05 M EDTA), 0.1 M Tris-base pH 7.5 (solution A) and for 1 h at 558C in 0.16SSC, 0.56SET, 0.1% sodium pyrophosphate pH 8 (solution B). Visualization was by autoradiography (X-OMAT, Kodak, Rochester, NY). Southern blot hybridization and cDNA library screening was done at 658C in Church buer (0.5 M NaPO4, pH 7.1, 2 mM EDTA, 7% SDS, 0.1% sodium pyrophosphate). The blots were washed in buer A at 658C for 1 h and buer B at 60 ± 658C for 1 h. Visualizations was as for Northern blots. Fluorescence in situ hybridization (FISH) Metaphase spreads were prepared by standard procedures (Testa et al., 1994). A cDNA clone (c-2ah04) was labeled with biotin 16-dUTP using a nick translation kit (Oncor). FISH and detection of immuno¯uorescence were performed basically according to the technique of Pinkel et al. (1986). The chromosome preparations were counter stained with 4',6'-diamidino-2-phenylindole (DAPI) in antifade solution (Oncor). Metaphase spreads were observed using a Zeiss Axiophot microscope and images were captured by a cooled CCD camera connected to a computer work station. Digitized images of DAPI staining and ¯uorescein signals were merged as described elsewhere (Altomare et al., 1996; Testa et al., 1992). By the hybridization and washing conditions utilized, nonspeci®c (background) hybridization typically is very low, i.e. usually less than two signals per metaphase (Bell et al., 1995; Bellacosa et al., 1993; Testa et al., 1992).

1977


1978

Loss of heterozygosity analysis Simple tandem repeat polymorphisms (STRPs) were typed in a PCR-based assay containing 15 ± 30 ng genomic DNA, 10 mM Tris-HCl (pH 8.3), 50 mM KCl, 1.5 mM MgCl2, 0.001% gelatin, 0.4 mM each primer, dCTP, dGTP, and TTP each at 16 mM, dATP at 2 mM, 0.65 mCi [a-35S]dATP (DuPont/New England Nuclear), 5% DMSO, and 0.25 units Amplitaq DNA polymerase (Perkin Elmer/Cetus) in a ®nal volume of 5 ml. Following an initial denaturation step at 948C for 4 min, DNA was ampli®ed through 20 cycles consisting of 5 s denaturing at 948C, 1 min annealing at 688C ± 0.58C/cycle, and 1 min extension at 728C. The samples were then subjected to an additional 25 cycles consisting of 5 s denaturation at 948C, 1 min at 588C, and 1 min extension at 728C and a ®nal extension of 728C for 5 min. PCR products were diluted 1:1 in 90% formamide, 20 mM EDTA, 0.3% bromophenol blue, 0.3% xylene cyanol, denatured at 948C for 5 min and 4 ml loaded onto

a 6% denaturing polyacrylamide gel, and electrophoresed at 90 W in 16TBE (16=0.09 M Tris, 0.09 M boric acid, and 0.002 M EDTA). After electrophoresis, gels were dried at 708C under vacuum and exposed to Kodak XAR-5 ®lm for 24 to 28 h. Polymorphisms used to map a minimum region of allelic loss are as follows: D6S250, D6S441, D6S415, D6S1007, and D6S1008.

Acknowledgements The work was supported by CA60643 and a Hoxie Harrison Smith grant to AKG. TCH was supported by CA56916 and The Evy Lessin Fund for Ovarian Cancer Research. The human ovarian tumor samples analysed were obtained from the Gynecological Oncology Group/ Cooperative Human Tissue Bank (Columbus, OH).

References Abdollahi A, Godwin A, Miller P, Getts L, Schultz D, Taguchi T, Testa J and Hamilton T. (1997). Cancer Res., submitted. Altomare D, Kozak C, Sonoda G and Testa J. (1996). Cytogenet. Cell Genet., 74, 248 ± 251. Auersperg N, Maines-Bandiera S, Booth J, Lynch H, Godwin A and Hamilton T. (1995). Amer. J. Obstet. Gynecol., 173, 558 ± 565. Bell D. (1991). Human Pathol., 22, 750 ± 762. Bell D, Taguchi T, Jenkins N, Gilbert D, Copeland N, Gilks C, Zweidler-McKay P, Grimes HL, Tsichlis P and Testa J. (1995). Cytogenet. Cell Genet., 70, 263 ± 267. Bellacosa A, Franke T, Gonzalez-Portal M, Datta K, Taguchi T, Gardner J, Cheng JQ, Testa JR and Tsichlis P. (1993). Oncogene, 8, 745 ± 754. Caput D, Buetler B, Hartog K, Thayer R, Brown-Shimer S and Cerami A. (1986). Proc. Natl. Acad. Sci. USA, 77, 4170 ± 4174. Chavrier P, Zerial M, Lemaire P, Almendral J, Bravo R and Charnay P. (1988). EMBO J., 7, 29 ± 35. Cheng P, Gosewehr J, Kim T, Velicescu M, Wan M, Zheng J, Felix J, Cofer K, Lou P, Biela B, Godorov G and Dubeau L. (1996). J. Natl. Cancer Inst., 88, 510 ± 518. Chomczynski P and Mackey K. (1995). Anal. Biochem., 225, 163 ± 164. Chomczynski P and Sacchi N. (1987). Analyt. Biochem., 162, 156 ± 159. Cliby W, Ritland S, Hartmann L, Dodson M, Halling K, Keeney G, Podratz K and Jenkins R. (1993). Cancer Research, 53, 2393 ± 2398. Courney A and Tjian R. (1988). Cell, 55, 887 ± 898. Dyck H, Hamilton T, Godwin A, Lynch H, Maines-Bandiera S and Auersperg N. (1996). Int. J. Cancer, 69, 429 ± 436. Fathalla M. (1971). Lancet, 2, 163. Franceschi S, Parazzini F, Negri E, Booth M, La Vecchia C, Beral V, Tzonou A and Trichopoulos D. (1991). Int. J. Cancer, 49, 61 ± 65. Fujii H, Zhou W and Gabrielson E. (1996). Genes Chromosomes & Cancer, 16, 35 ± 39. Gessler M, Poustka A, Cavenee W, Neve R, Dikin S and Bruns G. (1990). Nature, 343, 774 ± 778. Godwin A, Miller P, Getts L, Jackson K, Sonoda G, Schray K, Testa J and Hamilton T. (1995). Endocrinology, 136, 4640 ± 4649. Godwin A, Schultz D, Hamilton T and Knudson A. (1996). Gynecological Oncology: Principles and Practice. Hoskins W, Perez C and Young R. (eds). J.B. Lippincott Company: Philadelphia, pp 107 ± 148.

Godwin A, Testa J, Handel L, Liu Z, Vanderveer L, Tracey P and Hamilton T. (1992). J. Natl. Cancer Inst., 84, 592 ± 601. Hamilton T. (1992). Current Problems in Cancer, 16, 1 ± 57. Hamilton T, Abdollahi A, Getts L, Schultz D, Testa JR, Godwin A and Ozols R. (1996). Proceed Amer. Assoc. Cancer Res., 37, 623. Johnson S, Lissy N, Miller P, Testa J, Ozols R and Hamilton T. (1996). Analyt. Biochem., 236, 348 ± 352. Johnson SW, Swiggard PA, Handel LM, Brennan JM, Godwin AK, Ozols RF and Hamilton TC. (1994). Cancer Res., 54, 5911 ± 5916. Kadonaga T, Carner K, Masiarz E and Tjian R. (1987). Cell, 1079 ± 1090. Leach F, Nicolaides N, Papadopoulos N, Lui B, Jen J, Parsons R, PeltomaÈki P, Sistonen P, Aaltonen L, NystroÈm-Lahti M, Guan X, Zhang J, Meltzer P, Yu J, Kao F, Chen D, Cerosaletti K, Fournier R, Todd S, Lewis T, Leach R, Naylor S, Weissenbach J, Mecklin J, JaÈrvinen H, Petersen G, Hamilton S, Green J, Jass J, Watson P, Lynch H, Trent J, de la Chapelle A, Kinzler K and Vogelstein B. (1993). Cell, 75, 1215 ± 1225. Liang P, Averboukh L and Pardee A. (1993). Nucleic Acids Research, 21, 3269 ± 3275. Liang P and Pardee A. (1992). Science, 257, 967 ± 971. Lichter P, Bray P, Ried T, Dawid I and Ward D. (1992). Genomics, 13, 999 ± 1007. Mermod N, O'Neill E, Kelly T and Tjian R. (1989). Cell, 58, 741 ± 753. Mok S, Rodabaugh K, Colitti C, Welsh W and Berkowitz R. (1996). J. Soc. Gynecol. Invest., 3, 375A. Negri E, Franceschi S, Tzonou A, Booth M, La Vecchia C, Parazzini F, Beral V, Boyle P and Trichopoulos D. (1991). Int. J. Cancer, 49, 50 ± 56. Ohta M, Inoue H, Cotticelli M, Kastury K, Baa R, Palazzo J, Siprashvili Z, Mori M, McCue P, Druck, Croce C and Huebner K. (1996). Cell, 84, 587 ± 597. Orphanos V, McGown G, Hey Y, Thorncroft M, Santibanez-Koref M, Russell S, Hickey I, Atkinson R and Boyle J. (1995). Br. J. Cancer, 71, 666 ± 669. Osborne RJ and Leech V. (1994). Br. J. Cancer, 69, 429 ± 438. Ozols R, Rubin S, Dembo A and Robboy S. (1992). Gynecologic Ocology: Principles and Practice. Hoskins W, Perez C and Young R. (eds). Lippincott: Philadelphia, pp 731 ± 781. Peng S-Y, Chen C-Y and Shyu A-B. (1996). Mol. Cell Biol., 16, 1490 ± 1499.


Pinkel D, Straume T and Gray J. (1986). Proc. Natl. Acad. Sci. USA, 83, 2934 ± 2938. Saito S, Sirahama S, Matsushima M, Suzuki M, Sagae S, Kudo R, Saito J, Noda K and Nakamura Y. (1996). Cancer Res., 56, 5586 ± 5589. Salazar H, Godwin A, Daly M, Laub P, Hogan W, Rosenblum N, Boente M, Lynch H and Hamilton T. (1996). J. Natl. Cancer Inst., 88, 1810 ± 1820. Sato T, Saito H, Morita R, Koi S, Lee JH and Nakamura Y. (1991). Cancer Research, 51, 5118 ± 5122. Schultz D, Vanderveer L, Berman D, Hamilton T, Wong A and Godwin A. (1996). Cancer Res., 56, 1997 ± 2002. Shaw G and Kamen R. (1986). Cell, 46, 659 ± 667. Shelling A, Cooke I and Ganesan T. (1995). Brit. J. Cancer, 72, 521 ± 527. Taguchi T, Jhanwar S, Siegfried J, Keller S and Testa J. (1993). Cancer Res., 53, 4349 ± 4355. Tenaka K, Boice C and Testa J. (1989). Cancer Genet Cytogenet, 43, 1 ± 14. Testa J, Getts L, Salazar H, Liu Z, Handel L, Godwin A and Hamilton T. (1994). Cancer Research, 54, 2778 ± 2784. Testa J, Taguchi T, Knudson A and Hino O. (1992). Cytogenet Cell Genet, 60, 247 ± 249.

Theile M, Seitz S, Arnold W, Jandrig B, Frege R, Schlag P, Haensch W, Guski H, Winzer K-J, Barrett J and Scherneck S. (1996). Oncogene, 13, 677 ± 685. Thrash-Bingham C, Salazar H, Freed J, Greenberg R and Tartof K. (1995). Cancer Res., 55, 6189 ± 6195. Tibiletti M, Bernasconi B, Furlan D, Riva C, Trubia M, Buraggi G, Franchi M, Bolis P, Mariani A, Frigerio L, Capella C and Taramelli R. (1996). Cancer Res., 56, 4493 ± 4498. Vakalopoulon E, Schaak J and Shenk T. (1991). Mol. Cell Biol., 11, 3355 ± 3364. Wan M, Zweizig S, D'Ablaing G, Zheng J, Velicescu M and Dubeau L. (1994). Int. J. Oncol., 5, 1043 ± 1048. Whittemore A, Harris R, Itnyre J and the C.O.C.G. (1992). Amer. J. Epidem., 136, 1184 ± 1203. Yang-Feng T, Han H, Chen K, Li S-b, Claus E, Carcangiu M, Chambers S, JT C and Schwartz P. (1993). Int. J. Cancer, 54, 546 ± 551. Zheng J, Robinson W, Ehlen T, Yu M and Dubeau L. (1991). Cancer Research 51, 4045 ± 4051.

1979