Conserved Synteny Between the Fugu and Human PTEN Locus and the Evolutionary Conservation of Vertebrate PTEN Function

Conserved synteny between the Fugu and human PTEN locus and the evolutionary conservation of vertebrate PTEN function

Wei-Ping Yu1, Catherine J Pallen*,1, Alice Tay1, Frank R Jirik2, Sydney Brenner1, YH Tan1 and Byrappa Venkatesh1

1Institute of Molecular and Cell Biology, 30 Medical Drive, Singapore 117609, Republic of Singapore; 2Center for Molecular Medicine and Therapeutics, British Columbia Research Institute for Children's and Women's Health, 950 West 28th Avenue, University of British Columbia, Vancouver, B.C. V5Z 4H4, Canada

Mutations of PTEN, which encodes a protein-tyrosine novel protein tyrosine phosphatase (TEP1) through and lipid phosphatase, are prevalent in a variety of database homology searching (Li and Sun, 1997). human cancers. The human genome `draft' sequence still Mutational analyses of a wide range of spontaneously lacks organization and much of the PTEN and adjacent arising tumors have revealed mutations of PTEN in loci remain unde®ned. The puer®sh, Fugu rubripes,by glioblastoma, endometrial cancer, metastatic prostate virtue of having a compact genome represents an cancer, malignant melanoma, mammary adenocarcino- excellent template for rapid vertebrate gene discovery. ma, and various other tumor types (Cantley and Neel, Sequencing of 56 kb from the Fugu pten (fpten) locus 1999). Heterozygous mutations of PTEN have also identi®ed four complete genes and one partial gene been identi®ed in autosomal dominant hamartomatous homologous to human genes. Genes neighboring fpten syndromes such as Cowden disease (Cantley and Neel, include a PAPS synthase (fpapss2) dierentially ex- 1999; Liaw et al., 1997; Marsh et al., 1997, 1999). pressed between non-metastatic/metastatic human carci- These hamartomatous syndromes are also associated noma cell lines, an inositol phosphatase (fminpp1) and an with the development of various types of malignancies, omega class glutathione-S-transferase (fgsto). We have for example, individuals with Cowden disease have an determined the order of human BAC clones at the increased incidence of breast and non-medullary hPTEN locus and that the locus contains hPAPSS2 and thyroid cancer. hMINPP1 genes oriented as are their Fugu orthologs. There is considerable experimental evidence showing Although the human genes span 500 kb, the Fugu genes that PTEN acts as a tumor suppressor (Cantley and lie within only 22 kb due to the compressed intronic and Neel, 1999; Di Cristofano and Pandol®, 2000; Stam- intergenic regions that typify this genome. Interestingly, bolic et al., 1998). Owing to the ability of PTEN to and providing striking evidence of regulatory element dephosphorylate the D3 position of PI(3,4,5)P3 conservation between widely divergent vertebrate species, (Maehama and Dixon, 1998; Myers et al., 1998), the compact 2.1 kb fpten promoter is active in human introduction of PTEN into cell lines lacking this cells. Also, like hPTEN, fpten has a growth and tumor phosphatase inhibits cell growth and anti-apoptotic suppressor activity in human glioblastoma cells, demon- signaling pathways that lie downstream of phosphati- strating conservation of protein function. Oncogene dylinositol 3-kinase (PI-3K). Consistent with its role as (2001) 20, 5554 ± 5561. a tumor suppressor, targeted disruption of the murine Pten gene results in an accelerated rate of tumorigen- Keywords: PTEN; phosphatase; tumor suppressor; esis in Pten+/7 mice (Di Cristofano et al., 1998; Fugu Podsypanina et al., 1999), with neoplasms that have been examined exhibiting spontaneous mutations of the second, untargeted allele (Suzuki et al., 1998). These in ONCOGENOMICS Introduction vivo results have suggested that loss of Pten phosphatase activity is penetrant with respect to tumor PTEN (phosphatase and tensin homolog deleted on development. chromosome 10) was identi®ed by positional cloning as Except for the report that PTEN is downregulated a candidate tumor suppressor gene located at chromo- by TGF-beta (Li and Sun, 1997), and evidence of some 10q23 that was mutated in multiple types of PTEN promoter silencing as a functionally important advanced cancers, hence its other designation as epigenetic event in subsets of tumors (Mutter et al., MMAC (Li et al., 1997a; Steck et al., 1997), and as a 2000; Zhou et al., 2000), little is known about the physiological regulation of PTEN expression. Further- more, there have been no reports characterizing the structural elements or functional regulation of the *Correspondence: CJ Pallen; E-mail: [email protected] Received 7 February 2001; revised 10 May 2001; accepted 10 May isolated PTEN gene promoter. Although PTEN 2001 activity may well be determined by post-translational Fugu PTEN gene locus W-P Yu et al 5555 modi®cations (Georgescu et al., 1999; Vazquez et al., genes. They are the multiple inositol polyphosphate 2000), PTEN gene transcription levels may also be an phosphatase (fminpp1) gene, 3'-phosphoadenosine 5'- important determinant of PTEN activity within normal phosphosulfate synthase 2 (fpapss2) gene, and the cells. PTEN protein levels may be closely regulated, as omega class glutathione-S-transferase (fgsto) gene. experimentally induced alterations in the levels of this The last three exons of another fminpp gene (desig- phosphatase appear to have eects on cell growth and nated fminpp2) are present at the 5' end of the locus. survival (Di Cristofano and Pandol®, 2000). For The identical exon-intron organization and the high example, Pten heterozygous mice develop a sponta- homology in the coding sequences (82% aa identity) neous non-neoplastic lymphoproliferative/autoimmune between the two fminpp genes suggest that they are the disease (Di Cristofano et al., 1999; Podsypanina et al., result of a tandem duplication of an ancestral fminpp 1999), implying that a *50% reduction in Pten protein gene. In addition to these known genes, we identi®ed level is capable of deregulating the murine immune three hypothetical genes (h1, h2 and h3) which were system over time. Thus the precise identi®cation of predicted by both GRAIL and Genscan. Hypothetical evolutionarily conserved cis-acting control elements gene h2 is a single exon gene that codes for a protein of within the PTEN promoter will be critical to under- 534 amino acids. Genes h1 and h3 comprise multiple standing how PTEN activity is normally regulated in exons and code for proteins of 230 and 599 amino vivo. acids respectively. No major repetitive sequences are Loss of heterozygosity (LOH) within and proximal found in this locus. to the PTEN locus is frequently found in cancers. The exon-intron structures of the Fugu genes, However, LOH does not appear to be invariably minpp1, papss2 and pten, are identical to those of their accompanied by PTEN mutations in tumors such as human homologs (Chi et al., 1999; Kurima et al., 1999; meningiomas, some glioblastomas, primary prostate GenBank Accession number AF067844). However, the tumors, and sporadic breast carcinomas, raising the Fugu genes are highly compressed compared to the possibility that other as yet unidenti®ed tumor human genes. The Fugu minpp1, papss2 and pten genes suppressor genes lie within the 10q region (Bostrom span 3.4 kb, 6.5 kb and 7.3 kb respectively, whereas et al., 1998 ; Feilotter et al., 1998, 1999; Fujisawa et al., their human homologs occupy 422 kb (Chi et al., 1999; Simpkins et al., 1998). Although 220 kb of the 1999), 485 kb (Kurima et al., 1999) and 102 kb (BAC human PTEN genomic locus has been characterized 265N13, GenBank Accession number AF067844), (human BAC 265N13), the organization and orienta- respectively. Thus, the Fugu genes are seven to 14 tion of other genes in the vicinity of this locus remain times smaller than the human genes. The dierence in to be fully elucidated. the sizes between the Fugu and human genes are The puer®sh, Fugu rubripes, has been proposed as a mainly due to the enormous introns found in the model vertebrate genome because of its compact size of human genes (Table 1). Fugu pten encodes a protein 400 Mb (Brenner et al., 1993). Its genome is devoid of with 89% amino acid identity to its human counterpart dispersed repetitive elements which constitute a sig- (Figure 2). The Fugu genes minpp1, papss2 and gsto ni®cant fraction of the non-coding sequence in the encode proteins with 38, 79, and 55% amino acid human genome. Thus the intergenic regions in the identity, respectively, to the human proteins (Figures Fugu genome are relatively short and considerably less 3±5). complex than their mammalian counterparts. Fugu is The human genes for MINPP, PAPSS2 and PTEN therefore an attractive model for analysing genomic have all been mapped to the long arm of chromosome organization and promoter sequences. As a ®rst step 10 at 10q23. Their exact position and orientation are towards further de®ning the human PTEN gene locus not known, although MINPP1 is estimated to lie and promoter structure, we have sequenced the Fugu within 1 Mb upstream of PTEN (Chi et al., 1999). We pten (fpten) gene locus. This sequence (56 kb) has allowed us to assemble human BAC clones to provide the order and orientation of three human genes, Table 1 Intron sizes of Fugu and human PTEN and PAPSS2 genes hPTEN, hMINPP, and hPAPSS2. Furthermore, we PTEN PAPPS2 show that the *2kbfpten promoter is active in human Fugu (kb) Human (kb) Fugu (kb) Human (kb) cells, thus constituting a useful tool to identify conserved regulatory elements. Lastly, we ®nd that Intron 1 3.00 28.6 1.50 449.0 Intron 2 0.08 31.4 0.08 3.8 the fpten phosphatase acts as a growth and tumor Intron 3 0.09 5.7 0.08 0.7 suppressor in human glioblastoma cells, demonstrating Intron 4 0.09 1.9 0.15 0.6 functional conservation with mammalian PTEN. Intron 5 0.07 18.8 1.67 0.1 Intron 6 1.60 5.6 0.24 0.6 Intron 7 0.16 2.9 0.16 5.9 Intron 8 0.06 4.2 ± 5.6* Results and discussion Intron 9 0.15 42.8 Intron 10 0.09 1.9 Intron 11 0.08 1.7 Fugu pten locus Intron 12 0.09 1.0

The Fugu pten locus1 (Figure 1a) contains four *Intron 8 in the human PAPSS2 gene results from an alternatively complete genes that show homology to known human spliced exon which is not present in the Fugu papss2 gene

Oncogene Fugu PTEN gene locus W-P Yu et al 5556

Figure 1 Schematic diagram of the Fugu and human PTEN loci. (a) The Fugu sequence was obtained from cosmids c141O7, c62M14 and c96J15. Human PTEN and PAPSS2 data are from GenBank (accession number AF067844) and Kurima et al. (1999), respectively. Arrows represent genes and indicate the direction of transcription. Exon-intron structure is indicated by the open boxes and thin line. Note the dierent scales for the Fugu and human loci. (b) Representation of the deduced ordering of human BAC clones containing sequences for hPTEN (RP11-165M8, 265N13), hMINPP1 (RP11-57C13), and hPAPSS2 (RP11-57C13, RP1177F13)

used the Fugu pten locus sequence as a guideline to the ®rst ®ve exons for the hPTEN gene. From these determine the order of the human genes. We ®rst did a data we deduce that the human genes for MINPP1, homology search of the non-redundant nucleotide PAPSS2 and PTEN lie within a region of about database and the human genome draft sequence 500 kb, and are organized in the same orientation as (`High-Throughput Genome Sequences') database of their Fugu homologs (Figure 1b). We also identi®ed the GenBank and identi®ed several BAC clones that three BACs (RP11-475M16, 373N18 and 99N20) that contain sequences for the three human genes. The contain sequences for the human GSTO gene. These complete sequence of only one of these BACs (265N13) three BACs also map to chromosome 10. However, is known, and the rest are in the `draft' form submitted none of them overlaps with the BACs we identi®ed for to GenBank as `unordered pieces'. BAC 265N13 hMINPP1, hPAPSS2 and hPTEN, and therefore the (218 kb) contains the complete sequence for the location and orientation of the human GSTO gene in hPTEN gene together with 22 kb upstream and 95 kb relation to the other three genes are unclear. downstream sequences. No other known gene is found Regarding the possibility that tumor suppressor on this BAC. Among other BACs, BAC RP11-57C13 genes other than PTEN, but located in the same (168 kb) contains sequences for all the exons for chromosomal region, are involved in some human hMINPP1 and the ®rst exon for hPAPSS2, indicating cancers (Bostrom et al., 1998; Feilotter et al., 1998, the proximity of these two genes. Another BAC, RP11- 1999; Fujisawa et al., 1999; Simpkins et al., 1998), it is 77F13 (153 kb), contains most of the exons for noteworthy that hPAPSS2 has a 20-fold higher hPAPSS2. This BAC overlaps with BAC RP11- expression in a non-metastatic human colon carcinoma 165M8 (169 kb) by about 23 kb and the latter contains cell line than in an isogenetic metastatic cell line

Oncogene Fugu PTEN gene locus W-P Yu et al 5557

Figure 2 Alignment of Fugu, human and Xenopus PTEN sequences. Alignment was generated using the CLUSTALW Figure 4 Alignment of the Fugu and human MINPP1 proteins. program (Thompson et al., 1994). Arrow heads indicate the The arrow heads indicate intron positions in the Fugu minpp1 position of introns in the Fugu and human PTEN genes. Human, gene. The Fugu minpp1 is 38% identical to human MINPP1 PTEN, GenBank accession number AAD38372; Xenopus pten, (GenBank accession number AAD09751) GenBank accession number AAD46165

Figure 5 Alignment of the Fugu and human GSTO proteins. The arrow heads indicate intron positions in the Fugu gsto gene. the Fugu gsto is 55% identical to the human GSTO (GenBank accession number P78417)

dyloepimetaphyseal dysplasia) (Kurima et al., 1998; ul Haque et al., 1998) that are likely due in part to reduced sulfation of chondrocyte extracellular matrix molecules. In view of the role of PTEN as a tumor suppressor being linked to its activity as a phosphati- dylinositol phosphatase, it is intriguing that the neighboring hMINPP1 gene also encodes a phosphatase with speci®city for phosphoinositides. The en- doplasmic reticulum-localized MINPP1 acts on inositol polyphosphates containing four or more phosphates Figure 3 Alignment of the Fugu and human PAPSS2 proteins. such as InsP5 (Ins(1,3,4,5,6)P5) and InsP6, and prefers The intron positions (arrow heads) are conserved between the to hydrolyze phosphate at the 3 and 6 positions (Chi et Fugu and human PAPSS2 genes. The Fugu papss2 protein is al., 2000; Craxton et al., 1997; Nogimori et al., 1991). 79% identical to human PAPSS2 (GenBank accession number While the preferred substrates of PTEN are AF074331) PtdIns(3,4,5)P3 and PtdIns(3,4)P2, it less optimally hydrolyzes Ins(1,3,4,5)P4 (Maehama and Dixon, 1998), an in vitro substrate in common with MINPP1 (Franzon et al., 1999). PAPSS2 is a disease-associated (Chi et al., 1999). The substrates of MINPP1 serve as gene, with mutations causing cartilage and skeletal binding sites for speci®c pleckstrin homology domains, defects in mice (brachymorphism) and humans (spon- are involved in several cell signaling pathways, and

Oncogene Fugu PTEN gene locus W-P Yu et al 5558 their hydrolysis has been associated with dierentiation promoter spans only 2.1 kb (from the polyadenyla- (Chi et al., 1999), indicating that mutations in this tion signal of fpapss2 to the transcription start site of phosphatase could have profound eects on cells. It is fpten) and would be predicted to contain all the key also interesting that, like PAPSS2, MINPP1 is elements involved in the regulation of fpten expres- associated with chondrocyte function as it is highly sion. We have identi®ed putative binding sites for a expressed in dierentiating chondrocytes (Reynolds et variety of transcription factors, such as GATA, v- al., 1996; Romano et al., 1998). Nevertheless, although Myb, CREB, AML-1, Nkx-2 and CdxA (based on a cellular InsP5 and InsP6 levels are elevated in Minpp1- search of the TRANSFAC database; Heinemeyer et null mice, the animals lack obvious defects and have al., 1998) within this promoter. The boundaries of normal chondrocyte and bone development (Chi et al., the human PTEN promoter is not known as no gene 2000). Also no germline mutations of MINPP1 were is found up to 22 kb upstream of this gene. A detected in patients with hamartomatous syndromes comparison of the fpten upstream promoter region and lacking germline mutations of PTEN (Dahia et al., and 10 kb of the proximal hPTEN promoter region 2000). identi®ed several 10 ± 11 bp long elements conserved The vertebrate lineages that led to Fugu and between the two promoters. While some of these humans diverged about 400 million years ago and elements correspond to putative binding sites for thus the extent to which gene order and synteny has Nkx-2, AML1 and CdxA, the signi®cance of others is been conserved between the Fugu and human unknown. genomes is still poorly understood. Sequencing of To determine whether the fpten promoter is several homologous regions in the Fugu and human functional in human cells, the activity of the 2.4 kb genomes, for example, has indicated that while the sequence ¯anked by the translation stop codon of the gene linkage has been totally conserved in some loci, upstream fpapss gene and the translation start codon of in other loci synteny but not gene order has been fpten was evaluated in the human glioblastoma cell line conserved, and in yet other loci, synteny has been U87-MG. In the sense orientation, the 2.4 kb sequence disrupted (Venkatesh et al., 2000). In the PTEN locus, drove expression of a luciferase reporter gene, whereas the gene order of three of the genes are totally it had negligible promoter activity in the reverse conserved between the Fugu and human genomes and orientation (Figure 7). The inclusion of an SV-40 a fourth gene found in the Fugu locus is found on the enhancer in the plasmid further increased promoter same human chromosome as the other three genes. activity 5 ± 6-fold, suggesting that the 2.4 kb sequence Furthermore, the three Fugu genes are located within contains a basic promoter which may be naturally a short stretch of 22 kb whereas their human regulated by other remote cis-elements. Thus, the DNA homologs are spread over a region of about 500 kb, upstream of fpten contains promoter elements which supporting the view that the expansion in the human are functionally recognized by the transcription genome has occurred predominantly within the components of a human cell. noncoding sequences (Brenner et al., 1993; Venkatesh et al., 2000).

Fugu Pten expression Fugu fpten transcripts were detected in all the adult tissues analysed in this study (Figure 6), indicating that it expresses ubiquitously like human PTEN (Li and Sun, 1997; Steck et al., 1997). The fpten

Figure 7 Fugu pten promoter activity in human cells. U87-MG cells were co-transfected with pCMV ± CAT and a plasmid Figure 6 Northern blot analysis of the Fugu pten expression containing the Fugu pten promoter sequence in the sense or pattern. Total RNA (10 mg each from brain and skin; 20 mg each antisense (anti) orientation upstream to a luciferase reporter gene, from heart and ovary; and 40 mg each from gills, intestine, kidney, and with or without the SV40 enhancer as indicated. Cell lysates liver, muscle and testis) was probed with a full length Fugu pten were assayed for luciferase and chloramphenicol (CAT) activity. cDNA clone (top panel). A Fugu actin probe was used to check Fugu pten promoter activity is presented as arbitrary units the quality and quantity of RNA (bottom panel) representing units of luciferase activity per unit CAT activity

Oncogene Fugu PTEN gene locus W-P Yu et al 5559 Fugu pten protein The Fugu pten gene codes for a protein of 412 residues which is 89% and 82% identical to the human and Xenopus PTEN, respectively (Figure 2). As a group, the vertebrate PTEN dier considerably from the PTEN orthologs identi®ed in invertebrates such as Drosophila (Huang et al., 1999), C. elegans (Rouault et al., 1999) and S. cerevisiae (Li et al., 1997b). Vertebrate PTEN sequences show only 15 ± 35% identity to their invertebrate counterparts. In fpten, the N-terminal phosphatase domain is highly conserved, with an active site and TI loop identical to human PTEN. The putative general acid Asp92 is present, as well as a residue which may bind the D5 phosphate of PI(3,4,5)P3, His93 (Lee et al., 1999). The C2-like phospholipid binding domain is also well conserved, and possesses the CBR3 loop and basic ca2 helix that in hPTEN may mediate phospholipid membrane binding (Lee et al., 1999). A portion of the C2-like domain spanning residues 286 ± 320 is not conserved. Here the fpten has a Gln/Pro-rich region distinct from hPTEN and an eight amino acid insert lacking in both human and Xenopus PTEN (Figure 2). The comparable region of hPTEN (aa 286 ± 311) is an unstructured or loosely folded protease-sensitive loop (Lee et al., 1999). Figure 8 Fugu pten has cell growth and tumor suppressor Residues which are hot spots for tumor-derived activities. Parent U87-MG cells (a,f), a cell line derived following hPTEN mutations are identical between wild-type transfection with empty plasmid and selection (U87-neg) (b,g), human and Fugu genes. and stable lines of U87-MG cells expressing similar levels of fpten (U87-7, -8, -9 [c,h]) were examined for their phenotypes (a±c), growth (d), and saturation densities (e) attained in liquid culture, Growth suppression by fpten in human cells and for their abilities to form colonies in soft agar (f±h) To test whether fpten is functionally equivalent to human PTEN, we examined its properties as a growth suppressor of the human glioma U87-MG cell line. Summary These cells are de®cient in hPTEN, and introduction of exogenous wild-type hPTEN reduces the tendency The compact genome of the Fugu, with its short of these cells to grow in clusters lacking cell ± cell promoters devoid of repetitive elements, is a useful tool contact inhibition, and suppresses proliferation and for completing the human genome `draft' sequence and anchorage-independent growth (Furnari et al., 1997; annotating it. We have indeed demonstrated the utility Georgescu et al., 1999). Stable U87-MG cell lines of the Fugu sequence for determining the order of the expressing GFP-fpten were generated. Compared to human BAC clones in the PTEN locus which should the parent cells or to those transfected with empty help in rapidly obtaining the contiguous sequence of plasmid, which grew in clusters of multicellular layers the human genome from this locus. The observation (Figure 8a,b), the fpten expressing cells grew pre- that the short *2 kb Fugu pten promoter is operative dominantly as a monolayer up to complete con¯uence in human cells, and the further de®nition of the (Figure 8c). This was re¯ected by the much lower operative elements, will allow comparison with the saturation density attained by the fpten-expressing long upstream sequence of hPTEN to identify and cells of about one-third that of the parent U87-MG analyse the homologous human cis-acting regulatory line (Figure 8d,e). The fpten cells also exhibited a elements. The ®nding that the fpten protein shows slower growth rate (Figure 8d) with an average functional conservation in human cells with hPTEN, doubling time of 35 h, in contrast to a doubling time supports the notion that the regulation of PTEN of about 27 h for the parent and vector transfected expression could also be conserved in the vertebrate cells. Consistent with a transformed phenotype, both lineage. Identi®cation of the compressed pten locus in the parent and vector transfected cells formed large Fugu also provides the opportunity to examine the colonies in soft agar (Figure 8f,g), but fpten theory that other tumor suppressor genes may be expressing cells formed much smaller colonies (Figure located near PTEN. For example, the comparative 8h). Together, these results indicate that fpten abilities of fpten cDNA under control of the Fugu pten functions as a growth and tumor suppressor in human promoter versus that of the larger region containing cells in a manner similar to hPTEN. Fugu pten and adjacent coding sequences to rescue

Oncogene Fugu PTEN gene locus W-P Yu et al 5560 Pten/PTEN-de®cient mice or human tumor cell lines, Assay of Fugu pten promoter activity or other cell lines with LOH in the region of chromosome 10q23 but lacking hPTEN mutations, The intergenic sequence from the translation stop codon of can now be tested. the upstream gene to the translation start codon of fpten was ampli®ed by PCR using the primers 5'-AACTGATATC- TTGGCACGGGAGTCAGCCAG-3' and 5'-CATCGATAT- CTGTAGCAGGTGACAGGAGTC-3'. The resulting 2.4 kb Materials and methods fragment was inserted in the sense or antisense orientation 5' to a ®re¯y luciferase reporter gene in plasmids possessing Cloning and sequencing of Fugu pten locus (pGL3-Enhancer, Promega) or lacking (pGL3-Basic, Prome- ga) the SV40 enhancer element. One mg of each plasmid was A fragment of the Fugu pten gene was ampli®ed from the co-transfected with the same amount of pCMV-CAT plasmid genomic DNA by PCR using degenerate primers comple- into U87-MG cells using lipofectamine (GIBCO ± BRL) mentary to the fourth (forward primer: 5'-CAY TAY GAY according to the manufacturer's protocol. Fire¯y luciferase CAN GCN AAR TT-3') and sixth exons (reverse primer: 5'- activity was assayed using a luciferase assay kit (Promega) GTY TCR AAC ATC ATY TTR TG-3') of the known according to the manufacturer's protocol. CAT activity was vertebrate PTEN sequences and cloned into pBluescript. This determined (Ausabel et al., 1995) and used to normalize the fragment was used as a probe to screen a gridded Fugu transfection eciency. Promoter activity is de®ned as ®re¯y cosmid library (G Elgar, UK-HGMP Resource Center) and luciferase activity (luminescence units) over CAT activity nine positive cosmids were isolated. Three of the overlapping (c.p.m.). The activity exhibited by the fpten promoter in the cosmids (96J15, 62M14 and 141O07) were selected for sense orientation in pGL3-Basic was assigned as one subcloning and sequencing. A total of 55.9 kb contiguous arbitrary unit and the activities of the other plasmids are sequence was obtained from these cosmids (Figure 1) by a given relative to this. combination of shotgun sequencing and primer walking on an ABI 377 DNA sequencer. Coding sequences of known genes were identi®ed by their homology to genes in other Generation and analysis of stable fpten-expressing cell lines vertebrates, and novel genes were predicted by using the U87-MG glioma cells (from ATCC) were transfected with gene-prediction programs GRAIL and Genscan. The coding pEGFP-C2-fpten or empty plasmid pEGFP-C2, and stable sequence and the transcription start site of the Fugu pten cell lines selected in medium containing G418 (250 mg/ml). gene were determined by sequencing cDNA clones generated The Fugu pten expression was examined by immunoblotting by RT ± PCR and 5'RACE (SMART RACE cDNA of cell lysates with anti-GFP antibody (Santa Cruz). For Ampli®cation Kit, Clontech). growth curve and saturation density determination, cells were cultured over a period of 2 weeks and the medium replaced Northern analysis every other day. Cells were counted at regular intervals using a haemocytometer. For the soft agar assay, cells were seeded Total RNA from various Fugu tissues was fractionated on a in soft agar and maintained as described (Zheng et al., 1992). 1.2% agarose gel containing formaldehyde, transferred to a After 3 weeks, colonies were visualized with a phase contrast Hybond-N nylon membrane (Amersham), and probed with microscope and photographed. a-32P-labeled Fugu pten cDNA.

Cloning and expression of Fugu pten cDNA Acknowledgments Forward and reverse primers (respectively 5'-TCACGAATT- We thank the UK-HGMP Resource Centre for providing CATGGCTGCTATTATAAAAGAAAATGG-3' with an the Fugu cosmids, and CH Ng and BH Tay for excellent added EcoRI site and 5'-AACAGGATCCTCACACTT- technical assistance. This work was supported by the TAGTGATTTCGC-3' with an added BamHI site) that ¯ank National Science and Technology Board of Singapore, the coding sequence of Fugu pten were used for PCR with a Industry Canada, the Province of British Columbia, and cDNA template prepared by reverse transcription of Fugu the Medical Research Council of Canada. brain RNA. Complete sequencing of the PCR product con®rmed its identity as fpten cDNA and excluded the presence of any mutations introduced by PCR. The PCR Accession number fragment was cloned into pEGFP-C2 (Clontech) so as to The nucleotide sequence reported in this paper has been allow expression of fpten tagged at the N-terminus with submittedtotheGenBankwithaccessionnumber GFP. AF325922.

References

AusabelFM,BrentR,KingstonRE,MooreDD,Seidman Cantley LC and Neel BG. (1999). Proc. Natl. Acad. Sci. JG,SmithJAandStruhlK.(1995).In:Current Protocols USA, 96, 4240 ± 4245. in Molecular Biology, Vol. 2, pp. 9.7.1 ± 9.7.11, John Wiley Chi H, Tiller GE, Dasouki MJ, Romano PR, Wang J, & Sons, Inc., New York, NY. O'Keefe RJ, Puzas JE, Rosier RN and Reynolds PR. Bostrom J, Cobbers JM, Wolter M, Tabatabai G, Weber (1999). Genomics, 56, 324 ± 336. RG, Lichter P, Collins VP and Reifenberger G. (1998). ChiH,YangX,KingsleyPD,O'KeefeRJ,PuzasJE,Rosier Canc. Res., 58, 29 ± 33. RN, Shears SB and Reynolds PR. (2000). Mol. Cell. Biol., Brenner S, Elgar G, Sandford R, Macrae A, Venkatesh B and 20, 6496 ± 6507. Aparicio S. (1993). Nature, 366, 265 ± 268.

Oncogene Fugu PTEN gene locus W-P Yu et al 5561 Craxton A, Carey JJ, Burkhart W, Safrany ST and Shears Mutter GL, Lin MC, Fitzgerald JT, Kum JB, Baak JP, Lees SB. (1997). Biochem. J., 328, 75 ± 81. JA, Weng LP and Eng C. (2000). J. Natl. Canc. Inst., 92, Dahia PM, Gimm O, Chi H, Marsh DJ, Reynolds PR and 861 ± 863. Eng C. (2000). J. Med. Genet., 37, 715 ± 717. Myers MP, Pass I, Batty IH, Van der Kaay J, Stolarov JP, Di Cristofano A, Pesce B, Cordon-Cardo C and Pandol® PP. Hemmings BA, Wigler MH, Downes CP and Tonks NK. (1998). Nat. Genet., 19, 348 ± 355. (1998). Proc. Natl. Acad. Sci. USA, 95, 13513 ± 13518. Di Cristofano A, Kotsi P, Peng YF, Cordon-Cardo C, Elkon Nogimori K, Hughes PJ, Glennon MC, Hodgson ME, KB and Pandol® PP. (1999). Science, 285, 2122 ± 2125. Putney Jr JW and Shears SB. (1991). J. Biol. Chem., 266, Di Cristofano A and Pandol® PP. (2000). Cell, 100, 387 ± 16499 ± 16506. 390. Podsypanina K, Ellenson LH, Nemes A, Gu J, Tamura M, Feilotter HE, Nagai MA, Boag AH, Eng C and Mulligan Yamada KM, Cordon-Cardo C, Catoretti G, Fisher PE LM. (1998). Oncogene, 16, 1743 ± 1748. and Parsons R. (1999). Proc. Natl. Acad. Sci. USA, 96, Feilotter HE, Coulon V, McVeigh JL, Boag AH, Dorion- 1563 ± 1568. Bonnet F, Duboue B, Latham WC, Eng C, Mulligan LM Reynolds SD, Johnston CJ, Leboy PS, O'Keefe RJ, Puzas and Longy M. (1999). Br.J.Canc.,79, 718 ± 723. JE, Rosier RN and Reynolds PR. (1996). Exp. Cell Res., Franzon VL, Gibson MA, Hatzinikolas G, Woollatt E, 226, 197 ± 207. Sutherland GR and Cleary EG. (1999). Int. J. Biochem. RomanoPR,WangJ,O'KeefeRJ,PuzasJE,RosierRNand Cell Biol., 31, 613 ± 626. Reynolds PR. (1998). J. Cell Sci., 111, 803 ± 813. Fujisawa H, Kurrer M, Reis RM, Yonekawa Y, Kleihues P Rouault JP, Kuwabara PE, Sinilnikova OM, Duret L, and Ohgaki H. (1999). Am. J. Pathol., 155, 387 ± 394. Thierry-Mieg D and Billaud M. (1999). Curr. Biol., 9, Furnari FB, Lin H, Huang H-JS and Cavenee WK. (1997). 329 ± 332. Proc. Natl. Acad. Sci. USA, 94, 12479 ± 12484. Simpkins SB, Peier-Schneider S, Mutch DG, Gersell D and Georgescu M-M, Kirsch KH, Akagi T, Shishido T and Goodfellow PJ. (1998). Gynecol. Oncol., 71, 391 ± 395. Hanafusa H. (1999). Proc. Natl. Acad. Sci. USA, 96, Stambolic V, Suzuki A, de la Pompa JL, Brothers GM, 10182 ± 10187. MirtsosC,SasakiT,RulandJ,PenningerJM,Siderovski Heinemeyer T, Wingender E, Reuter I, Hermjakob H, Kel DP and Mak TW. (1998). Cell, 95, 29 ± 39. AE, Kel OV, Ignatieva EV, Ananko EA, Podkolodnaya Steck PA, Pershouse MA, Jassar SA, Yung WKA, Lin H, OA, Kolpakov FA, Podkolodny NL and Kolchanov NA. Ligon AH, Langford LA, Baumgard ML, Hattier T, Davis (1998). Nuc. Acids Res., 26, 364 ± 370. T, Frye C, Hu R, Swedlund B, Teng DHF and Tavtigian HuangH,PotterCJ,TaoW,LiD-M,BrogioloW,HafenE, SV. (1997). Nat. Genet., 15, 356 ± 362. Sun H and Xu T. (1999). Development, 126, 5365 ± 5372. Suzuki A, de la Pompa JL, Stambolic V, Elia AJ, Sasaki T, Kurima K, Warman ML, Krishnan S, Domowicz M, del Barco Barrantes I, Ho A, Wakeham A, Itie A, Khoo Krueger Jr RC, Deyrup A and Schwartz NB. (1998). W, Fukumoto M and Mak TW. (1998). Curr. Biol., 8, Proc. Natl. Acad. Sci. USA, 95, 8681 ± 8685. 1169 ± 1178. Kurima K, Singh B and Schwartz NB. (1999). J. Biol. Chem., Thompson JD, Higgins DG and Gibson TJ. (1994). Nuc. 274, 33306 ± 33312. Acids Res., 22, 4673 ± 4680. LeeJ-O,YangH,GeorgescuM-M,DiCristofanoA, ul Haque MF, King LM, Krakow D, Cantor RM, Rusiniak Maehama T, Shi Y, Dixon JE, Pandol® P and Pavletich ME, Swank RT, Superti-Furga A, Haque S, Abbas H, NP. (1999). Cell, 99, 323 ± 334. Ahmad W, Ahmad M and Cohn DH. (1998). Nat. Genet., Li D-M and Sun H. (1997). Canc. Res., 57, 2124 ± 2129. 20, 157 ± 162. Li J, Yen C, Liaw D, Podsypanina K, Bose S, Wang SI, Puc Vazquez F, Ramaswamy S, Nakamura N and Sellers WR. J, Milaresis C, Rodgers L, McCombie R, Bigner SH, (2000). Mol. Cell. Biol., 20, 5010 ± 5018. Giovanella BC, Ittman M, Tycko B, Hibshoosh H, Wigler Venkatesh B, Gilligan P and Brenner S. (2000). FEBS Lett., MH and Parsons R. (1997a). Science, 275, 1943 ± 1947. 476, 3±7. Li L, Ernsting BR, Wishart MJ, Lohse DL and Dixon JE. Zheng XM, Wang Y and Pallen CJ. (1992). Nature, 359, (1997b). J. Biol. Chem., 272, 29403 ± 29406. 336 ± 339. Liaw D, Marsh DJ, Li J, Dahia PLM, Wang SI, Zheng Z, ZhouMP,GimmO,HampelH,NiemannT,WalkerMJand Bose S, Call KM, Tsou HC, Peacocke M, Eng C and Eng C. (2000). Am. J. Pathol., 157, 1123 ± 1128. Parsons R. (1997). Nat. Genet., 16, 64 ± 67. Maehama T and Dixon JE. (1998). J. Biol. Chem., 273, 13375 ± 13378. Marsh DJ, Dahia PLM, Zheng Z, Liaw D, Parsons R, Gorlin RJ and Eng C. (1997). Nat. Genet., 16, 333 ± 334. MarshDJ,KumJB,LunettaKL,BennettMJ,GorlinRJ, Ahmed SF, Bodurtha J, Crowe C, Curtis MA, Dasouki M, Dunn T, Feit H, Geraghty MT, Graham Jr JM, Hodgson SV, Hunter A, Korf BR, Manchester D, Miesfeldt S, Murday VA, Nathanson KL, Parisi M, Pober B, Romano C, Eng C et al. (1999). Hum. Mol. Genet., 8, 1461 ± 1472.

Oncogene