oncogenomics - Nature

ã

Oncogene (2001) 20, 5378 ± 5392 2001 Nature Publishing Group All rights reserved 0950 ± 9232/01 $15.00 www.nature.com/onc

A novel member of the WD-repeat gene family, WDR11, maps to the 10q26 region and is disrupted by a chromosome translocation in human glioblastoma cells Olga B Chernova1, Aaron Hunyadi1, Eda Malaj2, Haquin Pan2, Carol Crooks1, Bruce Roe2 and John K Cowell*,1

ONCOGENOMICS

1 2

Center for Molecular Genetics, Lerner Research Institute /ND40, Cleveland Clinic Foundation, Cleveland, Ohio, USA; Department of Chemistry, University of Oklahoma, Norman, Oklahoma, USA

Allelic deletions of 10q25 ± 26 and 19q13.3 ± 13.4 are the most common genetic alterations in glial tumors. We have identi®ed a balanced t(10;19) reciprocal translocation in the A172 glioblastoma cell line which involves both critical regions on chromosomes 10 and 19. In addition, loss of an entire copy of chromosome 10 has occurred in this cell line suggesting that the translocation event may provide a highly speci®c critical inactivating event in a gene responsible for tumorigenesis. Positional cloning of this translocation breakpoint resulted in the identi®cation of a novel chromosome 10 gene, WDR11, which is a member of the WD-repeat gene family. The WDR11 gene is ubiquitously expressed, including normal brain and glial tumors. WDR11 is composed of 29 exons distributed over 58 kilobases and oriented towards the telomere. The translocation resulted in deletion of exon 5 and consequently fusion of intron 4 of WDR11 to the 3' untranslated region of a novel member, ZNF320, of the KruÈppel-like zinc ®nger gene family. Since ZNF320 is oriented toward the centromere of chromosome 19, both genes appeared on the same derivative chromosome der(10). The chimeric transcript encodes the WDR11 polypeptide, which is truncated after the second of six WD-repeats. ZNF320 is also expressed in A172 cells, although it is not clear if the translocation aects the expression of the altered gene because of the presence of another unrearranged gene on chromosome 19. We suggest that, because of its localization in a region frequently showing LOH and the observation of inactivation of this gene in glioblastoma cells, WDR11 is a candidate gene for the frequently proposed tumor suppressor gene in 10q25 ± 26 which is involved in tumorigenesis of glial and other tumors showing frequent alterations in the distal 10q region. Oncogene (2001) 20, 5378 ± 5392. Keywords: brain tumors; gene cloning; WD40 repeat; translocation; chromosome 10

*Correspondence: JK Cowell, Department of Cancer Genetics, Roswell Park Cancer Institute, Elm and Carlton Streets, Bualo, New York, NY 14263, USA Received 6 March 2001; revised 17 May 2001; accepted 31 May 2001

Introduction Diuse gliomas are the most common central nervous system tumors in adults, with almost 15 000 cases diagnosed annually in the United States and with a mortality approaching 80% during the ®rst year of diagnosis (Deen et al., 1993). Cytogenetic and molecular studies have identi®ed several recurrent, nonrandom genetic abnormalities associated with glial tumors. Both oncogene activation and tumor suppressor gene inactivation have been shown to be a part of the multistep process of glial tumorigeness and tumor progression. Inactivation of the tumor suppressor genes TP53, RB1, CDKN2A, PTEN, and DMBT1 through point mutations, deletions, or both, have been detected in glial tumors (for review see Louis and Gusella, 1995; Nagane et al., 1997; Smith and Jenkins, 2000). One of the most frequent genetic alterations in glial tumors, especially in GBMs, is heterozygous loss of chromosome 10 which has been associated with malignant progression (Louis and Gusella, 1995; Ohgaki et al., 1995; Saxena et al., 1999; Cheng et al., 1999; Fujisawa et al., 1999). Losses of regions on 10q and 10p have been found at relatively low frequency in low-grade astrocytomas (von Deimling et al., 1992; Fults et al., 1990) and oligodendrogliomas (Maier et al., 1997, 1998). In GBMs, however, loss of chromosome 10 is the most frequently reported genetic event which has been detected in 70 ± 93% of tumors. In most cases this loss involved an entire copy of chromosome 10 (Rasheed et al., 1995; Fults and Pedone, 1993; Sonoda et al., 1996). The distinct pattern of allele loss on both the long and short arms of chromosome 10 suggest the involvement of multiple tumor suppressor genes (Ichimura et al., 1998). Recently, three putative tumor suppressor genes, PTEN, DMBT1, and LGI1, were isolated from 10q23, 10q26, and 10q24, respectively (Li et al., 1997; Steck et al., 1997; Mollenhauer et al., 1997; Chernova et al., 1998). Rearrangements of each of these genes, however, were reported in only 30% or less of GBMs, suggesting that there may be additional tumor suppressor genes on 10q. Deletion of 19q is another common genetic alteration shared by all diuse glioma subtypes (von Deimling et al., 1992; Bello et al., 1995; Maintz et al., 1997; Smith et al., 1999a). This alteration is also

WDR11 rearrangement in gliomas OB Chernova et al

associated with the frequent progression from lowgrade astrocytoma to secondary glioblastoma (von Deimling et al., 1994; Ohgaki et al., 1995; Louis and Gusella, 1995; Nakamura et al., 2000). Importantly, in oligodendroglioma patients, deletion of 19q is associated with prolonged overall survival (Cairncross et al., 1998; Smith et al., 2000). The fact that the deletions of the terminal portions of 10q and 19q were found in low-grade astrocytomas and oligodendrogliomas ± two histologically distinct entities of gliomas ± suggests the existence of a putative suppressor gene(s) involved in the early glial tumorigenesis. Loss of heterozygosity (LOH) studies have played a pivotal role in identifying the regions of the chromosomes which are repeatedly deleted in particular subgroups of the glial tumors, indicating the position of the genes whose inactivation is critical for tumorigenesis. However, identi®cation of the critical gene from the LOH region is often a dicult task, especially when the frequently deleted region spans several megabases. Consequently, the target genes are still unknown in many of the regions showing frequent loss in gliomas. Molecular cloning of the translocation breakpoints that occur in the critical glioma-speci®c regions of the involved chromosomes allows the position of the critical gene to be pinpointed. This approach has been used recently to identify a putative tumor suppressor gene, LGI1, from the t(10;19)(q24;q13.1) translocation in the T98G glioblastoma cells (Chernova et al., 1998). Using FISH analysis and whole chromosome 10 and 19-speci®c probes we have now identi®ed another apparently balanced, reciprocal chromosomal translocation t(10;19)(q26;q13.4) in A172 glioblastoma cells. In addition, these cells have lost one entire copy of chromosome 10 but retained an apparently normal copy of chromosome 19. The breakpoints on both chromosomes occurred in the regions which are frequently deleted in GBMs. We have used a positional cloning approach to isolate and characterize a novel gene from chromosome 10, WDR11, encoding a WD-repeat containing protein. The coding region of the WDR11 gene was disrupted by this translocation and fused to the 3' UTR of a novel KruÈppel-like zinc ®nger protein, ZNF320, on chromosome 19. The translocation resulted in complete inactivation of WDR11, although its eect on the expression of ZNF320 is less clear due to the presence of an unrearranged copy of chromosome 19. Results To establish the nature of chromosome rearrangements involving chromosomes 10 and 19 in a series of human glioma cell lines (GB1, CCF4, CCF3, U87, U373, A172), we used a combination of chromosome 10 and chromosome 19-speci®c painting probes in FISH experiments. From this analysis we identi®ed an apparently balanced, reciprocal chromosomal translocation t(10;19)(q26;q13.4) in A172 cells. The breakpoints on both chromosomes occurred in regions which show frequent loss of heterozygosity (LOH) in GBMs.

An example of the structural chromosome rearrangement seen in these cells is shown in Figure 1. The A172 cells are near tetraploid and carry four copies of the der10, two copies of the der19, two apparently normal copies of chromosome 19 and no normal copy of chromosome 10. To de®ne the position of the breakpoints more accurately, we used a series of YAC clones which mapped to a distal regions of 10q and 19q in a FISH analysis in combination with chromosome-speci®c paints. Using this approach, we were able to establish that the chromosome 10-speci®c YAC, 937a6, and the chromosome 19-speci®c YAC, 790c10, cross the breakpoints (Figure 2). To build a more detailed map of the chromosome 10 breakpoint region, and so de®ne its position more accurately, we generated a set of fragmented YACs derived from the approximately 1.7 Mbp 937a6 YAC using the pBCL8.1 vector as described previously (Chernova and Cowell, 1998; Roberts et al., 1998). The size of the fragmented YACs was determined by using PFGE with subsequent hybridization with total human DNA. In this way we generated a series of smaller YACs ranging in size from 200 ± 1400 kb (Figure 2a). FISH analysis of the fragmented YACs demonstrated that Fr17 (570 kb) lay entirely above the breakpoint, whereas both Fr14 (660 kb) and Fr12 (690 kb) crossed the breakpoint (data not shown). These data suggested that the breakpoint region lay within an approximately 100 kb non-overlapping region within Fr14. To de®ne the position of the breakpoints more accurately, a series of somatic-cell hybrids were generated between A172 cells and mouse tk-de®cient 3T3 cells. PCR analysis of individual hybrid clones identi®ed two hybrids, AF41 which contained both the der(19) and normal 19 chromosomes, and AF2 which contained the der(10) but neither of the homologous normal chromosomes. The presence of these derivative chromosomes in these hybrid clones was con®rmed using FISH with paints for human chromosomes 10 and 19 (data not shown). Primers were then generated from the ends of fragmented YACs Fr17 and Fr14 and used in PCR analysis of the somatic cell hybrid clones to determine their presence/absence (Figure 2a). This analysis allowed us to determine the position of the breakpoints more accurately relative to the physical map of the region. We found that genetic marker SHGC35172 and the right end of YAC Fr17 (Fr17R) both lay above the breakpoint and the right end of YAC Fr14 (Fr14R) and marker CHLCATA29C03 lay below the breakpoint. All these data indicated that the breakpoint was located between PCR markers Fr17R and Fr14R, which are separated by approximately 90 kb. To obtain smaller clones crossing the translocation breakpoint, we used PCR primers for the markers Fr17R and Fr14R to screen a human BAC library (Research Genetics) and isolated one BAC for each marker, 571A20 and 216J18, respectively (Figure 2a). FISH analysis demonstrated that BAC 216j18, which actually contained both end-clone markers, crosses the breakpoint but BAC 571a20, which contains only

5379

Oncogene


5380

Figure 1 FISH analysis of chromosome rearrangements in A172 glioblastoma cells. (a ± c) Hybridization of whole chromosome 10 (red) and 19 (green) paints against metaphase chromosomes revealed the reciprocal t(10;19) translocation in A172 cells (b). The same image was digitally processed to reveal chromosome DAPI banding (a). The majority of the A172 cells contain four copies of der(10), two copies of der(19), and two copies of chromosome 19. Enlarged images of the chromosomes involved are shown in c. (d ± f) Hybridization with BAC 216J18 revealed that signal is split between both the derivative chromosomes (e and f). Karyotype DAPI banding is shown in d

Fr17R, apparently did not because it hybridized exclusively to the (der)10 (data not shown). FISH analysis of normal chromosomes con®rmed that BAC 216j18 localizes to 10q26 and is not chimeric. Our FISH experiments indicated that the chromosome 19 breakpoint lies within the chromosome region de®ned by YAC 790c10. We used the restriction map of the chromosome 19 which was constructed by the LLNL Human Genome Center to identify the BAC clones overlapping YAC 790c10 (see partial map of this region in Figure 2b). FISH analysis demonstrated that non-overlapping BACs 2287B18 and 310315 hybridize above and below the breakpoint, respectively. FISH analysis with BAC clones 3099C6, 2101P23, and 2331H12, gave confusing results since they all showed hybridization on both sides of the breakpoint. Limited analysis of the DNA sequences from the ends of these BACs identi®ed several zinc ®nger-containing (ZNF) genes in all three clones. These data are in agreement with earlier observations that this part of 19q13.4 carries large numbers of tandemly clustered ZNF genes which have apparently evolved through repeated duplication of the ancestral genes (Carver and Stubbs, 1997). Cross-hybridization between conserved regions of the ZNF genes was the most likely reason for the ambiguous FISH results. At that time this Oncogene

region had been prepared for complete sequencing by the LLNL Human Genome Center; therefore we concentrated our eort on sequencing chromosome 10 BAC 216j18 in an attempt to identify any gene involved in the chromosome translocation. Identification of the WDR11 gene Since the A172 breakpoint clearly occurred within BAC 216j18, it was subjected to high-throughput sequencing. The complete sequence comprising 128 769 bp (AC013357) is now available. To determine which genes were present in this BAC the sequence was submitted, online, to the Genome Channel Pipeline at the Oak Ridge National Laboratories. The output indicated that this clone contains several predicted exons and three EST clusters (Figure 3a), all of which appear to be parts of the same larger gene. Using PCR primer pairs from distal ends of the genomic sequence and DNA from somatic cell hybrids, we were able to demonstrate that the 10q26 breakpoint occurred within this gene. A comparison of the EST sequences with the genomic sequence of the BAC 216j18 allowed us to reconstruct the structure of the gene which spans 58.1 kb and consists of 29 exons. The exon ± intron boundaries are shown in Table 1. Sequences of the three large cDNAs (acc #AB037772, AK001368, and


5381

Figure 2 Summary maps of the chromosome regions encompassing the 10q26 and 19q13.4 breakpoint regions. In (a), the physical map as it relates to YAC 937A6 which crosses the 10q26 translocation breakpoint is shown. The derivative fragmented YACs (Fr) and BACs (b) are shown below. Genetic markers contained within the YACs are shown as solid vertical lines. The plus or minus symbol above YAC 937A6 indicates the presence or absence of the marker in somatic cell hybrid AF2 which contains only the der(10). The fragmented clones Fr12 and 14 crossed the breakpoint by FISH analysis. The fragmented YAC-derived markers Fr17R and Fr14R, which were used for BAC library screening, are shown in italic and their location is indicated by stippled lines. BAC 216J18 was shown to cross the breakpoint using both FISH and PCR analysis of somatic cell hybrids. In (b) a map of the BAC contig that overlaps the YAC, 790C10, which spans the translocation breakpoint on 19q13.4 is shown. FISH analysis demonstrated that BACs 2287B18 and 310315 lay above and below the breakpoint, respectively

AL137699) which were recently deposited into GeneBank, con®rmed the accuracy of the reconstructed sequence of the WDR11 cDNA (Figures 3b and 4) which currently consists of 4520 bp and correlates well with the size of the transcript (see below). This sequence contains a 3675 bp open reading frame which encodes a putative 1224 amino acid protein with the initiation methionine at position 34. The 3' untranslated region is 817 bp long (Figure 4). At the 5' end of this gene is a 400 bp CpG-island that starts upstream (7250 bp) and continues into the 5' UTR. The promoter prediction programs at BCM (TSSG and TSSW: http://dot.imgen.bcm.tmc.edu:9331/gene-®nder/ gf.html) revealed a promoter immediately upstream of cDNA sequence with a transcription start site at position 740 bp. The predicted promoter contains two putative TATA-boxes (Figure 4). Both the similarity in size between the predicted gene sequence and the transcript seen on Northern blots, as well as the presence of a putative promoter upstream of the gene strongly support the conclusion that the cDNA sequence is complete.

The predicted protein sequence produces a 1224 amino acid polypeptide sequence with a calculated molecular mass of 137 kD (Figure 4). Protein sequence motif searching and pattern matching analysis using a series of Internet-based software programs [PROSITE (http://expasy.cbr.nrc.ca), SMART (http://smart.emblheidelberg.de), PSORT (http://psort.nibb.ac.jp/ form2.html), and MOTIF (http://motif.genome.ad.jp)] revealed the presence of six putative WD40-repeats. The WD-repeat unit is a loosely conserved motif of approximately 40 amino acids that typically contains the GH dipeptide 11 ± 24 residues from its N-terminus and the WD dipeptide at the C-terminus (reviewed by Neer et al., 1994; Smith et al., 1999b). Several WD repeats fold into highly symmetrical, propeller-like structure with the conserved region of each repeat forming a part of the propeller blade. This structure can form complexes with other proteins and, thus, mediates protein ± protein interactions (discussed in Smith et al., 1999b). In WDR11 the ®rst two and the last three WR-repeats are tandemly repeated at the very N-terminus and in the C-terminal half, respecOncogene


5382

Figure 3 Organization of the WDR11 gene relative to the translocation breakpoint. (a) Diagrammatic representation of the 60 kb segment of BAC 216J18 encompassing the WDR11 gene showing the 29 exons (indicated by closed boxes). The three homologous EST clusters identi®ed by gene prediction software are shown below. (b) Relationship between the individual WDR11 cDNA clones (sequence accession numbers: WDR11, AF320223; 1, AL137699; 2, AB037772; 3, AK001368; 4, AK000579; 5, BE242873; 6, BE160256, 7, AA446765; 8, AA706241) used to compile the full-length cDNA sequence. Clone 1 did not contain exon 9 (hatched region) and had an unspliced intron (broken line) between exons 27 and 28. Clones 5 and 6 represent only partial DNA sequences, whereas other clones were either completely sequenced by us (clones 7 and 8) or their complete sequence was found in GenBank

Table 1 Summary of the sequences found at the splice junctions of the 29 exons of the WDR11 gene Exon 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29

Length

5'-boundary

3'-boundary

119 111 153 173 194 165 114 195 103 176 84 106 75 108 124 147 106 114 171 108 62 65 177 45 192 97 145 79 962

GGTCAGCGCCCAGGTCCT ctgttttgctagGGGCTG ttcttgatacagGTTAAA gggtgtttccagATGTTC ttctcaatatagTGCTTAC tacttttttcagTGCTGAA atctatttaaagGTAATA ttttatgtgaagATCCAG atgtctttgcagTAGTTC attacaaatcagGGCAAA tcattttgaaagGTACAA ttcctattatagGGGTAT ttgcaattaaagGTAGGA tgtttattttagGCAGTA tttgctttacagGAGTGG actatcccaaagCTTGCT gttgttttctagGGAAGT ctttgtcttcagAGGAAT tcactcttgaagGTTCAG tttgttgtctagAGCCTG ttttctttacagTGACTA tttcttcacagTGACAT catccttgccagGCTCTA ttccattctagAAATTT ttctcttactagACAGAC tttgtctttcagAGGGCG cgtctttctcagGTCCGT tctgttgagcagCATGAG tatggtttacagAAACTC

GGACTGgtgaggacgccg GTAAACgtaagtaaaaatc TCCAGGgtgaggaaagtc TAACTTgtgagtaacagt ACCTAGgtaagttacaag TTACAGgtatctacaatt AACCAGgtgagtttctgt GAACAGgtaaatgaatca TGATTGgtaagctttttt CTGTTGgtgagtatttga AGTCAAgtaagtatgtca CAACAGgtttgtttttaa TTTGAAgtaagtgtcaac GCTTTGgtaagttacaat GATCAGgtacagtacagt CCAGATgtgagtacaacc ATCCAGgtataagccaag AAAGAGgtaggccctctc TAACCGgtatggaatcct TCATGTgtaagtttttca GTCTAAgtaagcactcca TTCAAGgtaatattgttt TTTCAGgtagtctgcttc GGTCAAgtatgtcagttt TGGCAGgtaaggcacact GCAAAAgtaggtggttcc TCACAGgtaaaccgagag ACACAGatatcctttgca AATAAATTAAAAATTCCAtccaggcctttgt

tively. A transmembrane region was predicted at the Cterminal end (residues 1128 ± 1144), although no signal peptide was found. A tyrosine kinase phosphorylation site was predicted at amino acid residue 1075. Oncogene

WDR11 expression analysis To determine size of the transcript(s) and expression pattern of the WDR11 gene the 5' and 3' ends of WDR11


5383

Figure 4 Primary structure of the WDR11 gene. Sequences of the putative promoter region and the cDNA, including the 5' UTR, the deduced amino acid sequence of WDR11, and the 3' UTR are shown. Sequence of the WDR11 cDNA (accession number AF320223) is shown in capitals. The two putative TATA boxes, transcription start site, the ATG start codon, the six internal WDrepeat domains (amino acids 51 ± 99; 103 ± 145; 554 ± 596; 698 ± 821), and the translation stop codon are shaded. The CpG island and polyadenylation signal are underlined

were hybridized to human multiple tissues Northern blots purchased from Clontech (Figure 5). In normal tissues the 5'-speci®c probe detected three major transcripts of approximately 4.5, 2.7, and 2.0 kb (Figure 5) which were found in all tissues tested. When the same blots were reprobed with the 3'-speci®c probe, representing the last 10 exons, the strongest signal represented the 4.5 kb transcript with a weaker 2.0 kb transcript (Figure 5). These observations strongly suggest the possibility of alternative splicing and/or polyadenylation sites in WDR11 transcripts. Hybridization of the same probes to Northern blot containing RNA from the A172 cells (Figure 5) revealed that the 5'-speci®c probe detected only an aberrant transcript of approximately 3.0 kb and no speci®c transcripts were detected in these cells using the 3'-speci®c probe indicating that WDR11 transcrip-

tion is aberrant and that this gene is the target of the translocation rearrangement. Expression of the truncated 5' portion of the WDR11 gene in the A172 cells has been con®rmed by RT ± PCR using exon-speci®c primers (Figure 6a). A PCR product spanning exons 3 and 4 was clearly detected in the A172 cells, but no PCR product could be generated from exon 5 or any of the more distal exons in these cells. Apparently, therefore, the translocation breakpoint interrupted the WDR11 gene between exons 4 and 5. To investigate this further, a series of exonspeci®c primer pairs were designed from within the sequences ¯anking the ®rst 10 exons. PCR analysis of genomic DNA then demonstrated that exons 4 and 6 with ¯anking sequences are intact in A172 cells, but exon 5 and the ¯anking intronic sequences were not, Oncogene


5384

Figure 5 Expression pattern of WDR11. RNA samples from adult tissues (Clontech, Human I) (a,b) and cell lines (c) were hybridized with the probes spanning exons 1 ± 5 (a) or exons 26 ± 29 (b and c) of the WDR11 cDNA. Cell lines: A172 and U87, glioblastoma multiforme; CAPANI, pancreatic carcinoma. In (d) Northern blot analysis of cell lines normal tissues reveals a prominent band at 3.0 kb in A172 cells which is the size of the predicted chimeric gene in these cells. The probe used was derived from a short region spanning the 5' exons which required low stringency washing to detect the 3.0 kb band. Under these conditions a background of splice variants is also seen

thus demonstrating a homozygous deletion of exon 5 in A172 cells as a result of the translocation. The RT ± PCR analysis in a series of ®ve low grade gliomas, three anaplastic gliomas and three glioblastoma multiforme using primer pairs from exons 3 and 4, 16 ± 19 and 26 ± 29 revealed the expression of the WDR11 at similar levels in tumors and normal brain and showed no altered transcripts or splicing variants (Figure 6b). To analyse the rearrangement of the WDR11 gene in glioblastoma cells further, we hybridized a cDNA probe overlapping exons 3 to 9 to a Southern blot of DNA from four glioblastoma cell lines (Figure 7) digested with dierent enzymes. Rearrangements of WDR11 were seen in two cell lines, A172 and U87. Cell lines U373-MG and GB-1 showed no abnormal bands. The rearrangement in the A172 cells con®rmed that WDR11 is a target for the translocation in these cells. The rearranged fragment in U87 cells was dierent from that seen in A172 cells and could either be a true rearrangement, or a polymorphism. RACE through the breakpoint To identify the exact position of the translocation breakpoints on chromosomes 10 and 19, we performed Oncogene

3' RACE using cDNA from A172 cells. Since the breakpoint lay downstream of intron 4, we used a combination of a forward primer from within exon 4 and the reverse primer which was a modi®ed oligo(dT) oligonucleotide from the SMART RACE cDNA Ampli®cation kit (Clontech). The resulting PCR products were cloned into the pCR2.1 vector using the TA-cloning system (Invitrogen). Individual plasmids were then isolated from six representative colonies which were between 1.5 ± 1.8 kb. When these clones were sequenced, all six were shown to overlap and contained sequences from both chromosome 10 and chromosome 19. The sequence of the RACE-products crossing the breakpoint is shown in Figure 8. We had thus demonstrated the position of the breakpoint and con®rmed that a chimeric transcript was produced in the A172 cells. Alignment of the sequences of 3'-RACE products and BAC 216j18 showed that the breakpoint on chromosome 10 occurred in the middle of the fourth intron of the WDR11 gene (898 bp downstream of exon 4) within a LINE2 repeat. BLAST search of the 3' RACE clones indicated that the chromosome 19 breakpoint occurred within the BAC clone CTD2331h12 (accession number AC010487) which had been identi®ed earlier by FISH as one of the clones


5385

Figure 6 Expression of the WDR11 gene in the GBM cell lines and glial tumor tissue. (a) Ampli®cation of speci®c exons from cDNA and genomic DNA templates (1, human brain tissue; 2 and 3, the glioblastoma cell lines, A172 and U87MG, respectively; cno template control). Exons 3 and 4 are transcribed in human brain tissue, and both GBM cell lines (cDNA, ex3 and 4). Exon 5 and following downstream exons cells (cDNA: ex5, ex6, ex4 ± 7, ex6 and 7, and ex8 ± 10), ampli®ed individually or as a group, are present in brain and the U87-MG cells but absent in A172. Ampli®cation of exons 4, 5, and 6 from genomic DNA (below) shows that exons 4 and 6 are intact in the A172 cells (genomic DNA: ex4 and ex6 were ampli®ed from coding regions, and ex6i was ampli®ed from its introns), however, exon 5, ampli®ed from inside or outside its coding region (ex5 and ex5i, respectively), is absent in the A172 cells, indicating that exon 5 and ¯anking intronic sequences had been deleted as a result of the translocation event. Sequences of the PCR primers are described in Material and methods. (b) WDR11 expression in glial tumors and normal brain tissue. The results of RT ± PCR using the primer pairs that amplify exons 26 ± 29 are shown. LGO ± low-grade oligodendroglioma; AA ± anaplastic astrocytoma; GBM ± glioblastoma multiforme

which crossed the translocation breakpoint (Figure 2b). No protein coding sequences were found immediately around the chromosome fusion point. A comparison of the 3'-RACE product and the genomic sequences revealed a stretch of As at the chromosome 19 end of the RACE product, although the absence of a strong polyadenylation signal suggested that the oligo(dT)-primer used for the RACE experiment may have annealed at an internal site. A block of overlapping poly(A) signals was found, however, 553 bp downstream at the 3' end of Alu repeat. A series of primers were then designed from sequences on both sides of this putative poly(A) site and were used in combination with a primer (5'-CTTACGGTTAGCTTCTGCTGATG-3') derived from within exon 3 of the WDR11 gene in an RT ± PCR reaction. Because of the 1.3 kb intron between exons 3 and 4 of the WDR11, gene ampli®cation of unspliced transcripts or contaminating genomic sequences was prevented. The primer annealing immediately before the putative poly(A) signal produced an RT ± PCR product of the predicted size (2450 bp) in A172 cells (data not shown), but the primer annealing after the poly(A) site did not. As expected, no RT ± PCR product was generated using RNA from normal brain. This observation strongly supports the suggestion that

the polyA signal for the fused transcript lies 533 bp downstream of the end of the RACE product and extended the size of the fusion gene to 2739 bp, which correlates well (after addition of poly(A)-tail) with the 3.0 kb transcript observed in Northern hybridization. Identification of a new ZNF320 gene In attempt to predict genes within BAC 2331h12 which might be associated with the translocation, we analysed its sequence online using the Genome Channel Pipeline web server. At the time of analysis the sequence of chromosome 19 BAC 2331h12 was still un®nished, and consisted of six contigs with the breakpoint within the largest, 30.2 kb, contig. The resulting data strongly suggest that BAC 2331h12 contains three new zinc®nger genes which have similar KRAB-domain motifs and a tandem array of the twelve C2H2 zinc-®ngers. The translocation breakpoint occurred at the 3' end of one of these genes, named ZNF320, 1498 bp from the 3'-end of the coding region. Since the un®nished sequence of the BAC 2331h12 had gaps in it, we were only able to identify the two last exons of the ZNF320 gene, one encoding the KRAB-domain (128 bp) and a downstream exon (1388 bp of coding region) encoding a zinc-®ngers domain. The deduced amino acid Oncogene


5386

Figure 7 Southern blot analysis of the WDR11 gene in glioblastomas. Genomic DNA (7 mg) from GBM cell lines (1 ± T98G; 3 ± A172; 4 ± GB1; 5 ± U87-MG; 6 ± U373) and peripheral blood (2 ± normal control) were digested with BglII, PstI, EcoRV, and StuI, and probed with a fragment of the WDR11 cDNA containing exons 3 to 9. Asterisks indicate the lanes containing abnormal bands in A172 and U87-MG DNA digests. For A172 abnormal bands are shown by the arrows for all four dierent enzyme digests. The appearance of the new bands in these samples often accompany loss of normal bands. For U87 a novel band (lane 5) was noted only for EcoRV digests (arrow)

sequence of the KRAB-domain is highly homologous, up to 90 ± 98%, to the KRAB-domains of other proteins of this family (data not shown). The sequence of the zinc-®nger domain has a spacer in front of the ®ngers, which is common for this class of proteins, followed by twelve C2H2 zinc-®ngers separated by the short KruÈppel-type linker peptide TGEKP (Figure 9) indicating that ZNF320 belongs to a growing family of KruÈppel-like zinc-®nger proteins involved in activation or repression of transcription. The amino acid sequence of the zinc-®ngers demonstrates 62% identity and 73% similarity to ZNF83, although the spacer sequence upstream of the zinc-®ngers is unique, indicating that this gene is a new member of the family. The similarity and number of the ZNF genes in this region hinder their hybridization analysis because of cross-hybridization between the genes. To estimate the size of the ZNF320 transcript we synthesized the unique 351 bp spacer DNA segment generated by PCR (primers 5'ACTTAATTGGAAACCTATTGGTGTT-3' and 5'CCTATGTCTTCGCAAATGCA-3') and hybridized this probe to a human multiple tissues Northern blot (Figure 10). A single transcript, with a size of approximately 4.5 kb, was detected at highest levels of expression in placenta, pancreas, kidney, and heart tissue, while lower levels of expression were detected in liver, skeletal muscle and brain. Expression of ZNF320 in a variety of human cancer cell lines was studied using RT ± PCR with primers derived from zinc-®nger domain and unique portion of the 3'UTR. Data shown in Figure 10 demonstrate that ZNF320 was expressed in the most tested cell lines including glioblastoma (A172, U118, and U87-MG), neuroblastoma (CLB-Ba, NBLW, TR14) and prostate cancer (LN-Cap). Because Oncogene

of the presence of unrearranged chromosome 19 in the A172 cells, it is not clear at the moment whether the t(10;19) translocation aected the expression of the altered ZNF320 gene. Alignment of the 3'-RACE products to the BAC sequences indicated that, on the derivative chromosome 10, both WDR11 and ZNF320 genes are oriented in opposite directions, tail-to-tail (Figure 8a). To investigate whether the translocation breakpoint disrupted the 3'UTR of the ZNF320 gene, we searched dbEST at NCBI. The 3' ¯anking region of ZNF320 contains several repetitive elements (three Alu-family members, one LINE1 and one Mariner transposon-like sequences) which had complicated the assembly of the transcript. Using the BAC 2331h12 sequence we assembled a contig from overlapping EST clones, which crossed the translocation breakpoint and continued on another side of the translocation on der(19)over 1360 base pairs. The polyadenylation signal was found 20 base pairs from the end of this extended ZNF320 3'UTR contig. The ZNF320 3'UTR sequence located above the translocation breakpoint on der(19) in shown in Figure 8b. Analysis of sequences around the translocation breakpoint The translocation breakpoints on both chromosomes occurred in non coding regions which frequently contain transposons and other genetic elements which may facilitate recombination events. Analysis of the sequences around the translocation breakpoint using RepeatMasker (http://ftp.genome.washington.edu/cgibin/RepeatMasker) revealed a presence of several mobile transposable elements. A truncated LINE2 element has been found immediately upstream of the breakpoint on chromosome 10 in the fourth intron. Exon 5 and the immediate ¯anking sequences have been deleted in the A172 cells. Deletions of the DNA sequences, which are adjacent to the breakpoints, are frequently found at the translocation breakpoints. Another breakpoint on chromosome 10 occurred in the ®fth intron between two inverted MIR repeats (Figure 11). The corresponding region of chromosome 19 is very rich in interspersed repeated elements and 1000 bp around the breakpoint consist of three types of transposons: LINE1, Alu, and the less frequent mariner2. The breakpoint occurred at the 5'-end of the mariner2 element, which has insertion of two Alurepeats above the breakpoint. Immediately below mariner2 lies LINE1 retroposon. The position and orientation of the repetitive elements is depicted in Figure 11. Mariner2 (Hsmar2) is a recently discovered subfamily of the ancient DNA transposable elements of an insect origin (Oosumi et al., 1995; Robertson and Martos, 1997). All three interspersed repetitive elements, mariner2, Alu and LINE, were identi®ed at the recombination hot spots in several human diseases (Kiyosawa and Chance, 1996; Reiter et al., 1996; Chen et al., 1997; Purandare and Patel, 1997; Kazazian and Moran, 1998; Suminaga et al., 2000) and recombination between repeats has been proposed as a mechan-


5387

Figure 8 Chromosome 10 and 19 sequences around the translocation breakpoints. (a) Partial sequence of the WDR11 fusion transcript. Protein coding sequences are shown in capitals and primers are indicated by arrows. The cctg sequence (1070 ± 1073, boxed) is the position of the translocation breakpoint; this sequence is present at the fusion point in both chromosome 10 and 19 normal genomic sequences (accession numbers AC013357 and AC010487, respectively). A 3' RACE reaction was performed using the forward WDR11 exon 4-derived primer Ex4F and the reversed modi®ed oligo(dT) primer from Smart RACE kit. The 3' RACE PCR products ended at A-rich sequence (RACE end) at 1869 bp below the break. Primers R1 and R2 were used to con®rm the position of poly(A) signal (shown in bold, FT polyA) in the fused transcript (b). Sequence of the ZNF320 3' UTR on der(19) above the breakpoint. The cctg sequence at the fusion point is boxed and the poly(A) signal is shown in bold. The sequence is numbered from the breakpoint

ism of DNA rearrangements including inversions, gene fusions, duplications or deletions. Interestingly, the four base pair sequence CCTG, which has been found at the fusion point between chromosomes 10 and 19, may belong to either chromosome since it is present in WDR11 intron 4 and ZNF320 3' UTR.

Discussion Constitutional chromosome translocations in patients with hereditary cancers have repeatedly been shown to interrupt critical genes and aect their function. In these cases the translocations represent the critical Oncogene


5388

Figure 9 Primary structure of the deduced ZNF320 protein. The KRAB-domain is underlined and the twelve zinc-®nger motifs are shown in bold

Figure 11 Schematic representation of repeated elements at the translocation breakpoint. The t(10;19) translocation in A172 cells results from a reciprocal recombination between the fourth intron of the WDR11 gene and 3' UTR of the ZNF320 gene which is extremely rich in repeated sequences. The Mariner transposon has an insertion of Alu-like sequence. The repeats were identi®ed using RepeatMasker web server (http://ftp.genome.washington.edu/cgi-bin/RepeatMasker) Figure 10 Expression pattern of ZNF320. (a) RNA samples from adult tissues (Clontech, Human I) were hybridized with the probe spanning the ZNF320 spacer region. The weaker, upper band may either represent a splice variant of this gene or be produced as a result of cross hybridization with another member of this highly homologous family of genes. (b) RT ± PCR analysis of ZNF320 in cultured tumor cell lines: glioblastoma (U87MG and A172), prostate cancer (LN-Cap), and neuroblastoma (CLBBa, NBLW, TR14). Again the smaller band probably represents cross hybridization with related sequences

mutations which are responsible for tumorigenesis in a variety of dierent cancer types. In some tumors, recurrent chromosome translocations have been described where, again, the breakpoints represent mutations that aect speci®c genes which are presumed to be involved in tumorigenesis (Rabbitts, 1994, Still and Cowell, 1998). One of the long standing questions in tumor cytogenetics has been whether the frequent ideopathic chromosome translocations represent signi®cant events in the evolution of the tumor or whether they simply represent the consequences of an unstable genome characteristic of fast growing tumor cells. In our ®rst report of an ideopathic translocation in glioma cells involving a t(10;19)(q24;q13) rearrangement, we demonstrated that the chromosome 10q24 translocation breakpoint indeed interrupted and inactivated a gene which was normally expressed predominantly in the brain. The location of this gene, LGI1, in 10q coincided with one of the regions which showed Oncogene

frequent loss of heterozygosity during the transition of brain tumors from a low grade to a high grade malignancy (Chernova et al., 1998). In this report we present another example where a t(10;19)(q26;q13) chromosome translocation involving also interrupts genes expressed in the normal brain. In this example, it appears that genes on both chromosomes 10 and 19 may also be aected by the rearrangement. These observations tend to support the idea, at least for chromosome 10 in gliomas, that these translocation breakpoints serve as highly speci®c inactivating mutations of genes in the regions of LOH which create cells lacking the activity of these genes. These events ful®ll the prediction of the two-hit hypothesis (Knudson, 1971), although, unlike LOH studies alone, the chromosome translocations provided the means to identify the genes involved. Although we were unable to characterize the breakpoint on chromosome 19 involved in the inactivation of LGI1, in A172 cells we were able to show that the translocation event not only resulted in disruption of the WDR11 gene but also generated an intragenic deletion. As more and more chromosomes translocations of this type are described, it is becoming frequently clear that deletions often accompany the translocation events. Furthermore the presence of repetitive sequences on both sides of the translocation points to inappropriate recombination events between


homologous sequences as a mechanism for the generation of these translocations. What is curious, and as yet not explained, is the fact that a chimeric mRNA can be identi®ed in A172 cells which retains only the ®rst two WD repeats and apparently uses a polyadenylation signal within the 3' UTR of the chromosome 19 gene, ZNF320. As far as we were able to determine the exons distal to the second WD repeat in WDR11 are not detected in A172 cells suggesting that the reciprocal fusion gene is not translated. Since the ZNF320 gene is expressed o the normal homologous gene in A172 it is not possible to determine the consequences of the translocation on the expression of the homologue which is involved in the translocation. The WDR11 gene is located close to the genetic marker D10S209 (see Figure 2) in the 10q25 ± 26 region which is a target of allelic loss in the majority of glioblastomas (Karlbom et al., 1993; Rasheed et al., 1995). Although two other putative tumor suppressor genes, PTEN (10q23) and DMBT1(10q26), have been cloned from chromosome 10, there is growing evidence for the involvement of other genes on this chromosome in tumorigenesis. Both PTEN and DMBT1 were only altered in 25 ± 30% of glioblastomas, and in a signi®cant number of tumors with heterozygous deletions of 10q, alterations in either PTEN or DMBT1 could not be found (Liu et al., 1997; Tohma et al., 1998; Maier et al., 1998; Somerville et al., 1998; von Deimling et al., 2000). This lack of mutations in these genes was also demonstrated for other tumors which show frequent loss of distal regions of 10q, such as carcinomas of the respiratory tract (Petersen et al., 2000), melanoma (Herbst et al., 1999), endometrial cancers (Simpkins et al., 1998), hepatocellular carcinomas (Fujiwara et al., 2000; Yeh et al., 2000) and prostate carcinoma (Dong et al., 1998). Since Northern blot analysis indicates that WDR11 is expressed in a wide variety of human tissues, it may be a target for inactivation in these tumors as well. RT ± PCR analysis of a limited set of gliomas showed that WDR11 is expressed in all cases. Clearly, loss of expression is not a common ®nding although a more detailed study is required to determine whether any of these transcripts carry change-of-function mutations or not. Given that a number of dierent regions of LOH have been identi®ed on chromosome 10, it is possible that only a proportion would involve 10q26, and so a more extensive study of expression in gliomas will be needed to establish the extent to which functional inactivation is found. WD-repeat proteins belong to a superfamily of proteins which are found in all eukaryotes and have been implicated in a wide variety of crucial functions. Members of this family contain 4 ± 16 copies of the WD-repeat and include proteins that regulate signal transduction, apoptosis, transcription, pre-mRNA splicing, cell cycle progression, cytoskeletal organization, and vesicular fusion (reviewed by Neer et al., 1994; Smith et al., 1999b). A list of more that 250 human WD-repeat proteins has been compiled at the

SMART web page (http://smart.embl-heidelberg.de). The crystal structure of the best-characterized WDrepeat protein, the û-subunit of G proteins, has recently been described (Wall et al., 1995; Sondek et al., 1996; Neer and Smith, 1996) which reveals that the seven repeating WD motifs form a circular, propellerlike structure with seven blades each made up of four beta strands. This structure can form reversible complexes with several proteins, thus coordinating sequential and/or simultaneous interactions involving several sets of proteins (discussed in Smith et al., 1999b). The û-propeller structures are unlikely to contain fewer than four blades and a single polypeptide containing two or three WD-repeats may potentially dimerize with other WD-polypeptide to form a propeller. Since WDR11 polypeptide has two predicted WD repeats at the N-terminus and another cluster of three repeats at the C-terminus, it may form a propeller as a monomer, or each WD-cluster may dimerize with another protein. Clearly, identifying the binding partners for WDR11 will be important to obtain clues as to the function of this protein. The long arm of chromosome 19 also shows frequent loss in brain tumors, especially in oligodendrogliomas and oligoastrocytomas where LOH for 19q markers occurs in 70 ± 80% of tumors (Kraus et al., 1995; von Deimling et al., 1994; Smith et al., 1999a). The frequency of LOH on 19q in GBM, however, is smaller. Because of this observation, the question arises whether the t(10;19) translocation in A172 cells also interrupts a gene on chromosome 19 which may be important for tumorigenesis. The minimal, commonly deleted region has been assigned to 19q13.3 and has apparently been re®ned to within a few hundred kilobases (Rosenberg et al., 1996; Smith et al., 2000b). Despite continuous eorts, the 19q glioma tumor suppressor gene from this region, however, remains to be identi®ed. The chromosome 19 translocation breakpoint in A172 cells occurred more distally in the 19q13.4 region and disrupted a novel zinc-®nger gene, ZNF320, which belongs to a growing family of KruÈppel-like zinc-®nger proteins involved in activation or repression of transcription. The breakpoint region contains a large cluster of characterized and uncharacterized zinc-®nger genes, presumably re¯ecting a history of tandem in situ duplications starting from a single ancestral gene (Carver and Stubbs, 1997; Shannon et al., 1998). Indeed, the Genome Channel Pipeline software predicted that BAC 2331h12, which crosses the breakpoint in A172 cells, contains three novel Kruppel-associated box (KRAB)-containing ZNF genes. BLAST searches of the human genomic sequences (htgs database: http://www.ncbi.nlm.nih.gov/ blast/blast.cgi) established that the overlapping BACs 2101p23 and 3099c6, proximal to BAC 2331h12 also contain homologous zinc-®nger genes. The translocation event disrupted the 3' UTR of the ZNF320 gene, 1.5 kb downstream of the coding region, which potentially might negatively aect the transcription of the rearranged gene. However, since expression of ZNF320 in A172 cells was at similar levels to that seen

5389

Oncogene


5390

in other tumor cell lines, it is not clear at the moment whether the t(10;19) translocation aects the expression of the rearranged ZNF320 gene or ZNF320 transcription occurred from the remaining unrearranged chromosome 19. In summary, we describe a novel WD-repeat gene, WDR11, which was disrupted by the t(10;19) translocation in glioblastoma cells resulting in homozygous inactivation. Its location in the 10q25 ± 26 region, which is frequently deleted in gliomas and tumors of other tissues, strongly suggests that it is a candidate gene for the tumor suppressor locus in this region. In addition, this observation further supports a role for ideopathic chromosome translocations as one of the mechanisms of selectively inactivating genes in cancer cells.

Materials and methods Tumor samples and cell lines Primary tumor samples were obtained from an unselected series of patients treated at Cleveland Clinic Foundation. Tumors were graded according to the revised WHO classi®cation (Kleihues et al., 1993). Established cell lines were obtained from ATCC, except for GB1 and NBLW cells, which were a gift from Dr B Barna and Dr S Cohn, respectively. Preparation of YAC DNA and YAC fragmentation CEPH YAC clones were obtained from Research Genetics, Inc. YAC culture, preparation of yeast genomic DNA containing the YAC and YAC fragmentation were carried as described previously by Roberts et al. (1998). Essentially, the linearized pBCL8.1 vector DNA was transfected into the yeast cells carrying YAC 937A6, and host cells containing fragmented YACs were identi®ed by their inability to grow on Ura+ selective media. BACs and cDNA isolation BAC clones were obtained by screening the Research Genetics Human BAC Library Pools (Release IV) with primers of interest. Individual BACs and I.M.A.G.E Consortium (LLNL) cDNA clones (ESTs) were purchased from Research Genetics, Inc. (Huntsville, AL, USA). Plasmid DNA was prepared using Qiagene plasmid kits. BAC DNA was prepared for restriction enzyme analysis and endsequencing was achieved directly from the linearized BAC using T7 and SP6 vector primers. Fluorescence in situ hybridization Metaphase chromosome spreads were prepared using standard air-drying techniques. The FISH analysis was carried out basically as described by Chernova and Cowell (1998) with minor modi®cations. YAC-derived Alu-PCR products and BAC DNA were labeled with biotin-14-dATP using the Bio-nick kit (Gibco) or the digoxygenin-11-dUTP using the Dig-Nick kit (Boehringer). The chromosome 10 and 19 paints were purchased from AL Technologies, Arlington, VA, USA. The biotinylated probes were detected with Texas Redconjugated antibiotin antibodies (Rockland ImmunochemOncogene

icals, Inc., Gilbertsville, PA, USA), digoxygenin-labeled probes were detected with FITC-conjugated sheep antidigoxygenin Fab fragments (8 mg/ml; Roche). Chromosome spreads were viewed using a CCD camera (Photometrics) and Smartcapture software (Vysis, Chicago, IL, USA). Expression analyses Northern blots containing the RNA from the tissues of interest were obtained from Clontech. Hybridizations were carried out at 688C according to the ExpressHyb-protocol provided by the manufacturer. Expression analyses in brain tumors and cell lines was done by RT ± PCR. Total RNA was isolated from primary tumor tissues and cultured cells using the Tryzol reagent (Life Technologies). The ®rst-strand cDNA was generated from total RNA using random primers and Superscript revertase (Life Technologies) according to the manufacturer's protocol. Ten per cent of the ®rst-strand cDNA was ampli®ed with gene-speci®c primers. WDR11 cDNA-speci®c primer pair sequences were as follows: Exons 3 and 4, F, 5'CTTACGGTTAGCTTCTGCTGATG3', R, 5'AGAAAGAATGTTATCTGCATAGCTC3'; Exons 4 ± 7, F, 5'CCAAATTACATTGTGCTCTGGA3', R, 5'CACGCTGAAAGCAGGGTATTA3'; Exon 5, F, 5'AGCGAGGGTATTGTTTTCATC3', R, 5'TATTTAGAGCTTTCTTGGCACCT3'; Exon 6, F, 5'TGCCTTCAGTTGGCATACCT3', R, 5'CTGTAAAAATGGAACTCCTGTGC3'; Exons 6 and 7, F, 5'TGATTGCCTTCAGTTGGCA3', R, 5'CGAACACGTAAAGTTATACAACCAT3'; Exons 8 ± 10, F, 5'TGTGATGCAATCAGGGTGAC3', R, 5'GCAGTGGCTGATACATCTTGA3'. After an initial denaturation (948C for 3 min), 35 cycles of ampli®cation (948C for 1 min, 578C for 30 s, 728C for 1 min) were carried out, followed by a ®nal extension for 5 min at 728C. For a normalization control, 1/20 fraction of the same samples was ampli®ed using primers speci®c for the a-actin cDNA (Ra et al., 1997). Southern blot analysis Samples containing 7 mg of genomic DNA were digested with dierent restriction enzymes and electrophoresed on 0.8% agarose gels, transferred to nylon membranes, and hybridized with the 32P-labeled cDNA probes following standard protocols. Hybridization was carried out in 56SSC at 658C, with a ®nal wash at 0.26SSC at 658C. 3'-RACE The 3'-RACE experiment was performed using the SMART RACE cDNA Ampli®cation kit (Clontech) according to the manufacturer's protocol. Brie¯y, RACE-ready cDNA from the A172 cells was synthesized using the modi®ed oligo(dT) primer provided. The RACE-PCR reaction was performed using a WDR11 exon 4-speci®c primer (5'-CCAAATTACATTGTGCTCTGGA-3') and the universal primers provided with the kit. DNA sequencing DNA sequencing of all cDNA clones was performed on both strands using an ABI377 automated DNA sequencer (Applied Biosystems, Foster, CA, USA) by the CCF sequencing core facility. Information for EST clones and expressed sequence tags was obtained from public databases (dbEST) using the network service at the National Center for Biotechnology Information (NCBI: http://www.ncbi.nlm.nih.gov). cDNA and


predicted protein sequences were analysed using a variety of generally available prediction programs, in particular, BLAST (Altschul, 1990), BLOCK, SMART, MOTIF. Large-scale sequencing The detailed procedures for cosmid DNA isolation, random shot-gun cloning, ¯uorescent-based DNA sequencing and subsequent analysis have been described previously (Pan et al., 1994; Bodenteich et al., 1993; Chissoe et al., 1995; Roe et al., 1996) (http://www.genome.ou.edu/proto.html). Brie¯y, in this present work, BAC DNA was isolated free from host genomic DNA via a cleared lysate-acetate precipitation-based protocol, randomly sheared and made blunt-ended. After kinase treatment and gel puri®cation, fragments in the 1 ± 3 kb range were ligated into SmaI-cut, bacterial alkaline phosphatase (BAP)-treated pUC18 (Pharmacia) and Escherichia coli, strain XL1BlueMRF' (Stratagene), was transformed by electroporation. A random library of approximately 1200 colonies were picked from each transformation, grown in Terri®c Broth_(TB) medium supplemented with 100 micrograms of ampicillin (Bodenteich et al., 1993; Roe et al., 1996). Sequencing reactions were performed as previously described (Chissoe et al., 1995) using Thermus aquaticus (Taq) DNA polymerase, the Perkin-Elmer Cetus Fluorescent-labeled Big Dye Taq terminators. The reactions were incubated for 60 cycles in a Perkin-Elmer Cetus DNA Thermocycler 9600 and after removal of unincorporated dye terminators by ®ltration through Sephadex G-50, the ¯uorescent-labeled nested fragment sets were resolved by

electrophoresis on an ABI 3700 Capillary DNA Sequencer. After base calling with the ABI Analysis Software, the analysed data was transferred to a Sun Workstation Cluster, and assembled using Phil Green's Phred and Phrap programs (Ewing et al., 1998; Ewing and Green, 1998) Overlapping sequences and contigs were analysed using the graphical tools developed by Gordon et al. (1998). Gap closure and proof-reading was performed using either custom primer walking or using PCR ampli®cation of the region corresponding to the gap in the sequence followed by sub-cloning into pUC18 and cycle sequencing with the universal pUC-primers via Taq terminator chemistry. In some instances, additional synthetic custom primers were necessary to obtain at least threefold coverage for each base. Following manual proof-reading by viewing the data with Consed, the sequence was analysed on either a Sun workstation or Dec VAX computer with the programs contained within the GCG package (Genetics Computer Group, 1994) as well as the BLAST (Altschul et al., 1990), BEAUTY (Worley et al., 1995) and BLOCKS (Heniko and Heniko, 1994) programs.

5391

Computational sequence analysis Genomic sequences were analysed online using the Genome Analysis Pipeline (Oak Ridge National Laboratory: http:// grail.lsd.ornl.gov/GP/), which predicts both genes and exons using GRAIL (Xu et al., 1994) and GENSCAN (Burge and Karlin, 1997), as well as providing integrated GRAIL annotated features and BLAST and BEAUTY analysis.

References Altschul SF, Gish W, Miller W, Myers EW and Lipman DJ. (1990). J. Mol. Biol., 215, 403 ± 410. Bello MJ, Leone PE, Vaquero J, de Campos JM, Kusak ME, Sarasa JL, Pestana A and Rey JA. (1995). Int. J. Cancer, 64, 207 ± 210. Bodenteich A, Chissoe S, Wang YF and Roe BA. (1993). Automated DNA Sequencing and Analysis Techniques Venter JC (ed). Academic Press: London pp. 42 ± 50. Burge C and Karlin S. (1997). J. Mol. Biol., 268, 78 ± 94. Cairncross JG, Ueki K, Zlatescu MC, Lisle DK, Finkelstein DM, Hammond RR, Silver JS, Stark PC, Macdonald DR, Ino Y, Ramsay A and Louis DN. (1998). J. Natl. Cancer Inst., 90, 1473 ± 1479. Carver EA and Stubbs L. (1997). Genome Res., 12, 1123 ± 1137. Chen KS, Manian P, Koeuth T, Potocki L, Zhao Q, Chinault AC, Lee CC and Lupski JR. (1997). Nat. Genet., 17, 154 ± 163. Cheng Y, Ng HK, Zhang SF, Ding M, Pang JC, Zheng J and Poon WS. (1999). Hum Pathol., 30, 1284 ± 1290. Chernova O and Cowell JK. (1998). Cancer Genet. Cytogenet., 105, 60 ± 68. Chernova OB, Somerville RP and Cowell JK. (1998). Oncogene, 17, 2873 ± 2881. Chissoe SL, Bodenteich A, Wang YF, Wang YP, Burian D, Clifton SW, Crabtree J, Freeman A, Iyer K and Jian L. (1995). Genomics, 27, 67 ± 82. Deen DF, Chiarodo A, Grimm EA, Fike JR, Israel MA, Kun LE, Levin VA, Marton LJ, Packer RJ, Pegg AE et al. (1993). J. Neurooncol, 16, 243 ± 2472.

Dong JT, Sipe TW, Hyytinen ER, Li CL, Heise C, McClintock DE, Grant CD, Chung LW and Frierson Jr HF. (1998). Oncogene, 17, 1979 ± 1982. Ewing B and Green P. (1998). Genome Res., 8, 186 ± 194. Ewing B, Hillier L, Wendl MC and Green P. (1998). Genome Res., 8, 175 ± 185. Fujisawa H, Kurrer M, Reis RM, Yonekawa Y, Kleihues P and Ohgaki H. (1999). Am. J. Pathol., 155, 387 ± 394. Fujiwara Y, Hoon DS, Yamada T, Umeshita K, Gotoh M, Sakon M, Nishisho I and Monden M. (2000). Jpn. J. Cancer Res., 3, 287 ± 292. Fults D, Pedone CA, Thomas GA and White R. (1990). Cancer Res., 50, 5784 ± 5789. Fults D and Pedone C. (1993). Genes Chromosomes Cancer, 7, 173 ± 177. Gordon D, Abajian C and Green P. (1998). Genome Res., 8, 195 ± 202. Herbst RA, Podewski EK, Mommert S, Kapp A and Weiss J. (1999). Arch. Dermatol. Res., 10, 567 ± 569. Heniko S and Heniko JG. (1994). Genomics, 19, 97 ± 107. Ichimura K, Schmidt EE, Miyakawa A, Goike HM and Collins VP. (1998). Genes Chromosomes Cancer, 22, 9 ± 15. Karlbom AE, James CD, Boethius J, Cavenee WK, Collins VP, Nordenskjold M and Larsson C. (1993). Hum. Genet., 92, 169 ± 174. Kazazian Jr HH and Moran JV. (1998). Nat. Genet., 19, 19 ± 24. Kiyosawa H and Chance PF. (1996). Hum. Mol. Genet., 5, 745 ± 753.

Oncogene


5392

Oncogene

Kleihues P, Burger PC and Scheithauer BW. (1993). Brain Pathol., 3, 255 ± 268. Knudson AG. (1971). Proc. Natl. Acad. Sci. USA, 68, 820 ± 823. Kraus JA, Koopmann J, Kaskel P, Maintz D, Brandner S, Schramm J, Louis DN, Wiestler OD and von Deimling A. (1995). J. Neuropath Exp. Neurol., 54, 91 ± 95. Li J, Yen C, Liaw D, Podsypanina K, Bose S, Wang SI, Puc J, Miliaresis C, Rodgers L, McCombie R, Bigner SH, Giovanella BC, Ittmann M, Tycko B, Hibshoosh H, Wigler M and Parsons R. (1997). Science, 275, 1943 ± 1947. Liu W, James CD, Frederick L, Alderete BE and Jenkins RB. (1997). Cancer Res., 57, 5254 ± 5257. Maintz D, Fiedler K, Koopmann J, Rollbrocker B, Nechev S, Lenartz D, Stangl AP, Louis DN, Schramm J, Wiestler OD and von Deimling A. (1997). J. Neuropathol Exp. Neurol., 56, 1098 ± 1104. Mollenhauer J, Wiemann S, Scheurlen W, Korn B, Hayashi Y, Wilgenbus KK, von Deimling A and Poustka A. (1997). Nat. Genet., 17, 32 ± 39. Louis DN and Gusella JF. (1995). Trends Genet., 11, 412 ± 415. Maier D, Comparone D, Taylor E, Zhang Z, Gratzl O, Van Meir EG, Scott RJ and Merlo A. (1997). Oncogene, 15, 997 ± 1000. Maier D, Zhang Z, Taylor E, Hamou MF, Gratzl O, Van Meir EG, Scott RJ and Merlo A. (1998). Oncogene, 16, 3331 ± 3335. Nagane M, Huang HJ and Cavenee WK. (1997). Curr. Opin. Oncol., 9, 215 ± 222. Nakamura M, Yang F, Fujisawa H, Yonekawa Y, Kleihues P and Ohgaki H. (2000). J. Neuropathol. Exp. Neurol., 59, 539 ± 543. Neer EJ, Schmidt CJ, Nambudripad R and Smith TF. (1994). Nature, 371, 297 ± 300. Neer EJ and Smith TF. (1996). Cell, 84, 175 ± 178. Ohgaki H, Schauble B, zur Hausen A, von Ammon K and Kleihues P. (1995). Virchows Arch., 427, 113 ± 118. Oosumi T, Belknap WR and Garlick B. (1995). Nature, 378, 672. Pan HQ, Wang YP, Chissoe SL, Bodenteich A, Wang Z, Iyer K, Clifton SW, Crabtree JS and Roe BA. (1994). Genet. Anal. Tech. Appl., 11, 181 ± 186. Petersen S, Rudolf J, Bockmuhl U, Deutschmann N, Dietel M and Petersen I. (2000). Int. J. Cancer, 88, 71 ± 76. Purandare SM and Patel PI. (1997). Genome Res., 8, 773 ± 786. Rabbitts TH. (1994). Nature, 372, 143 ± 149. Ra T, van der Giet M, Endemann D, Wiederholt T and Paul M. (1997). Biotechniques, 23, 456 ± 460. Rasheed BK, McLendon RE, Friedman HS, Friedman AH, Fuchs HE, Bigner DD and Bigner SH. (1995). Oncogene, 10, 2243 ± 2246. Reiter LT, Murakami T, Koeuth T, Pentao L, Muzny DM, Gibbs RA and Lupski JR. (1996). Nat. Genet., 12, 288 ± 297. Roberts T, Chernova O and Cowell JK. (1998). Hum Molec. Genet., 7, 1169 ± 1178. Robertson HM and Martos R. (1997). Gene, 205, 219 ± 228. Roe B, Crabtree J and Khan A. 1996. DNA Isolation and Sequencing: Essential Techniques Series. J. Wiley and Sons: New York. Rosenberg JE, Lisle DK, Burwick JA, Ueki K, von Deimling A, Mohrenweiser HW and Louis DN. (1996). Oncogene, 13, 2483 ± 2485.

Saxena A, Shriml LM, Dean M and Ali IU. (1999). Oncogene, 18, 1385 ± 1390. Shannon M, Kim J, Ashworth L, Branscomb E and Stubbs L. (1998). DNA Seq., 8, 303 ± 315. Simpkins SB, Peier-Schneider S, Mutch DG, Gersell D and Goodfellow PJ. (1998). Gynecol Oncol., 3, 391 ± 395. Smith JS, Alderete B, Minn Y, Borell TJ, Perry A, Mohapatra G, Hosek SM, Kimmel D, O'Fallon J, Yates A, Feuerstein BG, Burger PC, Scheithauer BW and Jenkins RB. (1999a). Oncogene, 18, 4144 ± 4152. Smith TF, Gaitatzes CG, Saxena K and Neer EJ. (1999b). TIBS, 24, 181 ± 185. Smith JS and Jenkins RB. (2000). Front Biosci, 5, D213 ± D231. Smith JS, Perry A, Borell TJ, Lee HK, O'Fallon J, Hosek SM, Kimmel D, Yates A, Burger PC, Scheithauer BW and Jenkins RB. (2000a). J. Clin. Oncol., 18, 636 ± 645. Smith JS, Tachibana I, Lee HK, Qian J, Pohl U, Mohrenweiser HW, Borell TJ, Hosek SM, Soderberg CL, von Deimling A, Perry A, Scheithauer BW, Louis DN and Jenkins RB. (2000b). Genes Chromosomes Cancer, 29, 16 ± 25. Somerville RPT, Shoshan Y, Eng C, Barnett G, Miller D and Cowell JK. (1998). Oncogene, 17, 1755 ± 1757. Sondek J, Bohm A, Lambright DG, Hamm HE and Sigler PB. (1996). Nature, 379, 369 ± 374. Sonoda Y, Murakami Y, Tominaga T, Kayama T, Yoshimoto T and Sekiya T. (1996). Jap. J. Cancer Res., 87, 363 ± 367. Steck PA, Pershouse MA, Jasser SA, Yung WKA, Lin H, Ligon AH, Langford LA, Baumgard ML, Hattier T, Davis T, Frye C, Hu R, Swedlund B, Teng DHF and Tavtigian SV. (1997). Nature Genet., 15, 356 ± 362. Still IH and Cowell JK. (1998). Blood, 92, 1456 ± 1458. Suminaga R, Takeshima Y, Yasuda K, Shiga N, Nakamura H and Matsuo M. (2000). J. Hum. Genet., 45, 331 ± 336. Tohma Y, Gratas C, Biernat W, Peraud A, Fukuda M, Yonekawa Y, Kleihues P and Ohgaki H. (1998). J. Neuropathol. Exp. Neurol., 57, 684 ± 689. von Deimling A, Louis DN, von Ammon K, Petersen I, Wiestler OD and Seizinger BR. (1992). Cancer Res., 52, 4277 ± 4279. von Deimling A, Bender B, Jahnke R, Waha A, Kraus J, Albrecht S, Wellenreuther R, Fassbender F, Nagel J and Menon AG. (1994). Cancer Res., 54, 1397 ± 1401. von Deimling A, Fimmers R, Schmidt MC, Bender B, Fassbender F, Nagel J, Jahnke R, Kaskel P, Duerr EM, Koopmann J, Maintz D, Steinbeck S, Wick W, Platten M, Muller DJ, Przkora R, Waha A, Blumcke B, Wellenreuther R, Meyer-Puttlitz B, Schmidt O, Mollenhauer J, Poustka A, Stangl AP, Lenartz D and von Ammon K. (2000). J. Neuropathol. Exp. Neurol., 59, 544 ± 558. Wall MA, Coleman DE, Lee E, Iniguez-Lluhi JA, Posner BA, Gilman AG and Sprang SR. (1995). Cell, 83, 1047 ± 1058. Xu Y, Mural R, Shah M and Uberbacher E. (1994). Genet. Eng. (N Y)., 16, 241 ± 253. Worley KC, Wiese BA and Smith RF. (1995). Genome Res., 5, 173 ± 184. Yeh KT, Chang JG, Chen YJ, Chen ST, Yu SY, Shih MC, Perng LI, Wang JC, Tsai M and Chang CP. (2000). Cancer Invest., 18, 123 ± 129.