A Novel Candidate Tumor Suppressor Gene from 10Q24.3

LAPSER1: a novel candidate tumor suppressor gene from 10q24.3

Yofre Cabeza-Arvelaiz1, Timothy C Thompson2, Jorge L Sepulveda3 and A Craig Chinault*,1

1Department of Molecular and Human Genetics, Baylor College of Medicine, Houston, Texas, TX 77030, USA; 2Department of Urology, Baylor College of Medicine, Houston, Texas, TX 77030, USA; 3Department of Pathology, University of Pittsburgh School of Medicine, Pittsburgh, Pennsylvania, PA 15261, USA

Numerous LOH and mutation analysis studies in Transfer of these portions of each chromosome into dierent tumor tissues, including prostate, indicate that cancer cells has provided further evidence indicating there are multiple tumor suppressor genes (TSGs) that these regions harbor genes that suppress tumor- present within the human chromosome 8p21 ± 22 and igenicity (Ichikawa et al., 1994; Murakami et al., 1996). 10q23 ± 24 regions. Recently, we showed that LZTS1 (or Several candidate prostate cancer genes have been FEZ1), a putative TSG located on 8p22, has the isolated from 8p21 ± 22, including PRLTS (Fujiwara et potential to function as a cell growth modulator. We al., 1995), N33 (MacGrogan et al., 1996), and FEZ1 report here the cloning, gene organization, cDNA (Ishii et al., 1999). The expression of the latter, which sequence characterization and expression analysis of has now been designated as the LZTS1 (leucine zipper, LAPSER1,anLZTS1-related gene. This gene maps putative tumor suppressor 1) gene, has been found to within a subregion of human chromosome 10q24.3 that be altered in many tumors and cell lines including has been reported to be deleted in various cancers, esophageal, breast and prostate (Ishii et al., 1999). We including prostate tumors, as frequently as the neighbor- have also recently demonstrated this gene's ability to ing PTEN locus. The complete LAPSER1 cDNA in¯uence cellular growth properties (Cabeza-Arvelaiz et sequence encodes a predicted protein containing various al., 2001), which, together with the previous results, domains resembling those typically found in transcription make it an excellent candidate for a TSG. factors (P-Box, Q-rich and multiple leucine zippers). The tumor suppressor gene PTEN, whose loss of LAPSER1 is expressed at the highest levels in normal function has been shown or suggested in multiple types prostate and testis, where multiple isoforms are seen, of cancer, including Cowden disease, Bannayan- some of which are either undetectable or dierentially Zonana syndrome, and glioma, kidney, breast and expressed in some prostate tumor tissues and cell lines. prostate tumors, is located at 10q23.3 (Li et al., 1997; Over-expression of LAPSER1 cDNA strongly inhibited Steck et al., 1997; Teng et al., 1997; Vlietstra et al., cell growth and colony-forming eciencies of most 1998). Actually, multiple neoplasms with recurrent cancer cells assessed. Together these data suggest that chromosomal aberrations in this region have been LAPSER1 is another gene involved in the regulation of described by the NCI Cancer Genome Anatomy cell growth whose loss of function may contribute to the Project (http://www.ncbi.nlm.nih.gov/CGAP/mitel- development of cancer. Oncogene (2001) 20, 6707 ± sum.cgi). Cancers with the unbalanced chromosomal 6717. abnormality del(10)(q24) include acute lymphoblastic and myeloid leukemias, astrocytomas, adenocarcinoma Keywords: prostate cancer; tumor suppressor; LZTS1; of various tissues (breast, large intestine, ovary, LAPSER1 prostate and stomach), bladder transitional cell carcinoma, undierentiated carcinoma of the prostate, chronic lymphoproliferative disorder, testis germ cell Introduction tumor, mesothelioma, non-Hodgkin's lymphoma, and uterus leiomyomas. Cancers with balanced chromoso- Loss of heterozygosity (LOH) and homozygous mal abnormalities that involve this region include acute deletions at human chromosomal regions 8p21 ± 22 lymphoblastic leukemias (t(10;14)(q24;q11), non- and 10q23 ± 24 are frequently found in prostate Hodgkin's lymphoma (t(10;14)(q24;q11), and adenocarcinomas, as well as other cancer tumors, and (t(10;14)(q24;q32), and thyroid adenocarcinoma suggest that multiple TSGs are present in each of these ((t(10;19)(q24;q13). This great variety of cancers with locations (Bova et al., 1996; Cairns et al., 1997; Carter apparent relevant genes in this region support the et al., 1990; Gray et al., 1995; Ishii et al., 1999; possibility that there may also be additional TSGs Ittmann, 1996; Kagan et al., 1995; Kim et al., 1998; Li located there that have not yet been characterized. et al., 1997; Petersen et al., 1998; Whang et al., 1998). We report here the cloning, cDNA sequencing, gene organization and predicted protein structure of LAPSER1, a novel LZTS1-related candidate TSG *Correspondence: AC Chinault; E-mail: [email protected] located on human chromosome 10q24.3 near the Received 10 May 2001; revised 20 July 2001; accepted 26 July 2001 PTEN locus. Expression analysis revealed that this The LAPSER1 gene Y Cabeza-Arvelaiz et al 6708 gene is expressed in most normal tissues, with the highest abundance found in prostate and testis. In addition, expression analysis revealed that multiple isoforms of this gene are dierentially expressed in dierent cell lines and tissues. Analysis of the eects of over-expression of LAPSER1 on colony formation and cell growth of several cancer cell lines suggests that LAPSER1 is another attractive cell growth control gene whose loss of function may be of relevance to the development of prostate cancer.

Results

Identification and mapping of the LAPSER1 gene LZTS1 (also known as FEZ1) has been suggested as a candidate TSG on the basis of its frequent deletion in various cancers (Ishii et al., 1999) and its recent functional identi®cation as a cell growth suppressing gene on 8p22 (Cabeza-Arvelaiz et al., 2001). When BLAST searches were performed with the LZTS1 gene sequence against the GenBank databases, we identi®ed three other related human genes of unknown function, as shown in the dendogram in Figure 1a. The sequence with highest overall sequence identity to LZTS1 corresponded to a gene region on contig AL133215 that we have provisionally designated as LAPSER1 (for its high content of the amino acids leucine (L), alanine (A), proline (P), serine (S), glutamic acid (E), and arginine (R)). This region has been mapped by FISH analysis to human chromosome 10q24.1 ± 24.33 (http://webace.sanger.ac.uk). To more precisely place the LAPSER1 gene relative to the PTEN locus, we compiled sequencing and mapping data from several sources, as described in Materials and methods, to construct a physical map encompassing 10q24, as shown in Figure 1b.

Cloning and organization of the LAPSER1 gene To study the LAPSER1 gene and to explore its possible involvement in the regulation of cell growth, we cloned its complete cDNA sequence, using a RT ± PCR based approach with normal human prostate and testis RNA. An initial nested PCR ampli®cation performed with gene speci®c primers that ¯anked the presumed LAPSER1 coding region Figure 1 (a) Phylogenetic analysis of LAPSER1 and related (inferred by comparison to the LZTS1 cDNA) human protein sequences. The dendogram depicting the relation- ship between the indicated protein sequences was compiled using yielded one major and two minor transcription ClustalW analysis from the Mac Vector 3.5 software package. products of approximately 2, 1.7, and 1.35 Kb, The phylogenetic distances between the sequences are indicated, designated I, II, and III, respectively, in Figure 2a. where a value of 0.1 corresponds to a dierence of 10% between PCR re-ampli®cation of the three major RT ± PCR the two sequences. (b) Chromosomal location of the LAPSER1 products with internal primer sequences from exons gene. An ideogram of chromosome 10 is shown along with the physical and transcript map of the 10q23.3 ± 24.3 sub-region 2, 3 or 4 paired with an exon 5 primer (¯anking the spanning *22 Mb between the PTEN and MXI1 genes. The coding sequence) suggested that the minor transcripts locations of two of the most frequently deleted segments of this were generated by alternative exon splicing (Figure sub-region in prostate cancer, based on analysis with the markers 2b). Transcript I contained the expected sequences shown, are enclosed by open boxes. The loci of several known genes that have been mapped to this sub-region are shown. The from all coding exons tested while transcripts II and estimated distances from the p-telomore are indicated in III lacked exon 4 (258 bp) and exon 3 (660 bp), megabases (Mb), based on the scale of the Human Genome respectively. Browser website (http://genome.ucsc.edu)

Oncogene The LAPSER1 gene Y Cabeza-Arvelaiz et al 6709

Figure 2 Cloning and genomic organization of the LAPSER1 gene. (a) RT ± PCR products of alternatively spliced transcripts of LAPSER1 using gene speci®c primers ¯anking the coding sequence are indicated by arrowheads on the left. (b) PCR analysis of the LAPSER1 mRNA isoforms detected in a, using primer sets designed to amplify coding sequences from either exon 2, exon 3 or exons 4 ± 5 (after the stop codon). The sizes of the PCR products in kilobases are indicated with arrows on the left. In (a) and (b), lanes marked M contain the size markers (100 bp ladder). (c) Schematic representation of the assembly of the full-length LAPSER1 cDNA sequence. The coding region (represented by black boxes), was obtained by sequencing the RT ± PCR-derived transcript I. Sequences from the indicated EST clones (represented by hatched bars and identi®ed by the GenBank accession number), were used to assemble a composite of the complete sequence of the LAPSER1 cDNA. Intron sequences are shown as dashed lines. (d) Comparison of the organization of the human LAPSER1 and LZTS1 genes. The coding region (indicated as closed boxes) is contained within exons 2 (451 bp), 3 (660 bp), 4 (258 bp), and 5 (1320 bp) of the LAPSER1 gene; thin connecting lines to the LZTS1 schematic indicate the orthologous exons. Potential alternative 5' exons for LAPSER1 are indicated by open boxes. The positions of possible initiator (ATG) codons and the locations of CpG islands are also indicated. The scale at the bottom of the diagram indicates the approximate position of the nucleotide residues in the LAPSER1 gene

Sequencing of the LAPSER1 cDNA derived from coding region encodes a predicted protein sequence the major 2-Kb transcript I (GenBank accession # of either 669 or 644 amino acids with a calculated AY029201) revealed an open reading frame (ORF) of molecular mass of 72.8 or 70.3 kD and estimated pI 1932 bp. BLAST searches with the LAPSER1 cDNA of 6.1 or 6.3, respectively. Besides the higher (9 ± 11%) and gene sequences against the dBEST database than average 5 ± 7.5%) content of the amino acids detected three EST clones that overlapped with the leucine, alanine, proline, serine, glutamic acid, and 5' end and several EST clones that overlapped with arginine, the content of both glycine and glutamine is the 3' end. The sequence of the presumed full-length also higher (8.21%) than average (6.8 and 4%, cDNA (including the 5' and 3' UTR) was obtained by respectively). combining the LAPSER1 cDNA coding sequence with We next ascertained the organization of the four of these EST sequences (accession #'s: BF310703, LAPSER1 gene and compared it to the LZTS1 gene BE887654, BF032717 and AI651618), as depicted in (Figure 2d) by aligning the genomic sequence to the Figure 2c. This composite sequence is composed of compiled LAPSER1 cDNA composite sequence. The the 1932 bp ORF of the sequenced cDNA, a 639-bp 3' 5' UTR or the LAPSER1 gene displays a simple UTR and a variable length 5' UTR of at least 156 ± pattern of selective splicing generating a 5' UTR of 206 bp. The LAPSER1 protein sequence inferred from variable length, as indicated by the two ESTs that are the cDNA sequence was identical to the combined alternatively spliced to exon 2. BF310703 has a 5' sequences of transcripts ENST00000189301 and exon (1a) of at least 31 bp long, located 5256 bp ENST00000065966 predicted from the BAC RP11- upstream of exon 2 in a region that contains predicted 107L8 sequence by the Ensembl system (http:// CpG islands; BE887654 has a 5' exon (1b) of at least www.ensembl.org). Depending on which starting 81 bp long, located 2541 bp upstream of exon 2, in a methionine residue is used, the complete cDNA region that also contains predicted CpG islands. Thus,

Oncogene The LAPSER1 gene Y Cabeza-Arvelaiz et al 6710 the length of the complete gene transcript is between 2802 and 2852 bp depending on which exon 1 is present. Another EST sequence (BF334338), which aligns to the region located between exon 1b and exon 2 (830 bp from exon 2) and is ¯anked by potential acceptor/donor splicing sites, may represent a third selectively spliced 5' exon (1c?). The second exon (451 bp) contains two potential translation initiation methionines, both of which are in the context of typical Kozak consensus sequences (GCCACCATGG or CCAGTCATGG) (Kozak, 1991). Using the second initiation methionine the proposed protein sequence contains an N-terminus that shows a high degree of identity to the reported sequence of LZTS1 (Figure 2a). However, it is conceivable that LZTS1 extends some 25 amino acids upstream of the reported sequence as suggested by the 2007 bp ORF in BAC RP-11 353K12. A polyadenylation site (ATTAAA) is located in the 3'- UTR, 639 nucleotides downstream Figure 3 Northern blot analysis of mRNA from multiple normal human tissues. (a) Two micrograms of poly(A)+ RNA of the stop codon and 16 nucleotides upstream of a from the various normal human tissues indicated were loaded per poly (dA) tail present in at least 20 dierent ESTs. lane and probed with a 660-bp cDNA fragment representing the Searches for core promoter elements and transcription entire exon 3 or LAPSER1.(b) The presence of relatively equal regulatory sequences in the 6-Kb region upstream of amounts of intact mRNA in all of the lanes was shown by stripping and rehybridizing the blot with a b-actin probe. The exon 2 predicted three potential promoter sequences position of the major transcript band is indicated by an arrow preceding exon 1b (3317, 1360, and 1980 nucleotides and size standards are indicated on the right in Kilobases (Kb) upstream with scores of 25, 7.5 and 22 respectively), and one preceding exon 1a (120 nucleotides upstream with a score 18). In addition to potential TATA boxes and putative binding motifs for various transcription lower and higher molecular weight were revealed factors, including Sp1, AP-1, AP-2, MYC, CDC25 when a probe comprising the complete cDNA was and p53, several CCAC- and CAAT-binding protein used (data not shown). This result further corrobo- boxes and ARE1.2 sites were detected (not shown). rated our previous RT ± PCR results (Figure 2a,b) The LAPSER1 gene spans a genomic area of at indicating that alternative splicing of coding exons least 10.6 Kb and is composed of at least ®ve exons. generates some of the observed isoforms. The All the intron/exon boundaries found and their approximately 3.8-Kb transcript detected in all surrounding sequences (GGTG..CAGG) closely agree tissues assessed may be generated by the presence with the consensus splice donor (GT)/acceptor (AG) of an additional 5' non-coding exon (exon 1c? in site rule (Mount, 1982). Comparison of the genomic Figure 2c,d). structure of the LAPSER1 and LZTS1 genes Northern analysis with prostate cell lines did not suggested that exons 3 and 4 of LAPSER1 may detect any LAPSER1 transcript in DU145 cells, while have been derived by division of LZTS1 exon 2. The relatively high levels of expression were detected in LAPSER1 gene is highly conserved among verte- LNCaP and PC3 cells (Figure 4a). The expression of brates, showing a 95% nucleotide sequence identity LAPSER1 and LZTS1 mRNAs was also examined in and a 98% amino acid sequence identity to a Macaca human cell lines and tumor tissues by RT ± PCR with fascicularis partial cDNA sequence (accession # primer sets that amplify the ORF of LAPSER1 or AB046013) and greater than 87% identity to several exon 3 of LZTS1. These included the prostate cancer ESTs from mouse (e.g. accession #'s AW911842, cell lines DU145, LNCaP, TSUPr1, PC3, PC3-m, BF182142, BB605928). PPC1, and ND1, the sarcoma cell line SaOs2, the lung cancer cell line H1299, the ®broblast cell line BUD-8, and cells from one primary prostate tumor. As shown Expression pattern of LAPSER1 in normal tissues and in Figure 5b, no expression of the LAPSER1 gene was cancer cell lines and tissues observed in the DU145 and ND1 cells, while LZTS1 Northern analysis of the LAPSER1 gene expression was expressed relatively ubiquitously and at equivalent in dierent tissues using an exon 3 probe detected a levels in all the cell lines assessed. The overall quantity major transcript of approximately 2.8 Kb in all of the LAPSER1 mRNA was not uniform between normal tissues assessed. This size corresponds well dierent cell lines and the ratio between the bands with that of the compiled consensus cDNA sequence corresponding to the various isoforms varied exten- (Figure 2c). However, the levels of mRNA were sively. The relative levels of type II and III isoforms variable, with prostate and testis expressing the appear to be elevated in the PC3, PC3-m, PPC, and highest levels and PBL having barely detectable H1299 cells (Figure 4b), suggesting that the dierent levels (Figure 3a). Multiple LAPSER1 transcripts of isoforms are dierentially expressed in each cell line.

Oncogene The LAPSER1 gene Y Cabeza-Arvelaiz et al 6711

Figure 4 Expression analysis of LAPSER1 mRNA in cancer cells. (a) Northern blot analysis of RNA from prostate cancer cell lines. Total RNA isolated from the indicated cancer prostate cell lines and normal peripheral blood lymphocytes (PBL) was analysed by hybridization with a 660-bp fragment derived from exon 3 of the LAPSER1 cDNA (top) or a b-actin cDNA probe (bottom). The intensity of the signal detected varied between samples but correlated with the amount of intact RNA in each lane as suggested by the control b-actin. (b) RT ± PCR analysis of LAPSER1 and LZTS1 mRNA expression in prostate and other cancers. Upper panel: products of RT ± PCR analysis of LAPSER1 mRNA in the indicated cell lines and tumor samples were analysed by agarose gel electrophoresis and ethidium bromide staining. Multiple isoforms are identi®ed with roman numerals and are indicated with arrowheads on the left. Bottom panel: agarose gel analysis of RT ± PCR-ampli®ed fragments from various LZTS1 isoforms. Primers used are located on exon 3 and revealed the expression of LZTS1 transcripts containing this exon. The expected size fragment (690 bp) is indicated with an arrowhead and truncated fragments are indicated with arrows on the left. Lane M includes the size marker used (100 bp ladder). DNA fragment sizes are indicated in Kb

Effects of over-expression of LAPSER1 on various CDKN1A (SDII/p21) cDNA. The noticeable dier- cancer cell lines ence in the number of cells in control-transfected cell To determine whether or not the LAPSER1 gene lines is probably due to dierence in the growth rates product acts as a cell-growth regulatory factor we of these cell lines. The smaller number of LNCaP cloned the complete LAPSER1 cDNA into the cells is probably due to the fact that, because of the pcDNA3.1 expression vector. This construct and the low transfection eciency attained with these cells, pEGFP reporter vectors were co-transfected into the number of cells seeded was only 1/4 of the logarithmically growing human cancer cell lines others; in addition LNCaP cells grow slower than the LNCaP, TRSUPr1, PC3, U2Os, HEK-293, the rat other cell lines used. Cell viability was estimated at cancer cell lines AT6.2 and normal rat ®broblast Rat- 495% in the control plates, but was lower in the 1 cells. The expression vector containing the cyclin- LAPSER1 and CDKN1A transfectant plates. How- dependent kinase inhibitor CDKN1A, which is a ever, the latter could not be measured reliably potent mammalian G1-G0 cell cycle-arrester, was because of extensive cell lysis in these plates after 3 included in the analysis as a control. The gene days in culture. transfer eciency achieved by lipofection was usually The LAPSER1 and the CDKN1A plasmids alone high (440% on day 1 for all cells except for LNCaP were also stably transfected into these cell lines and the which was generally lower). Nevertheless, to minimize results were compared to cells identically transfected contributions of untransfected cells on the analysis of with a negative control. A similar cell-growth arresting the eects of transient expression of the LAPSER1 eect was observed by long-term expression of the gene on these cell lines, the GFP-positive (GFP+) LAPSER1 and the CDKN1A genes on these cell lines. cells were isolated by sorting the selectively gated As illustrated in Figure 5b, after selection of the cells in green cell sub-population 30 h after co-transfection. antibiotic for 10 ± 14 days, a large number of colonies After plating 56105 GFP+ sorted cells (1.26105 for was observed with all cells transfected with the negative LNCaP cells), the cell numbers were assessed 96 h control vector PCNKLC, whereas fewer or no colonies later by counting cells from duplicate plates. As can were observed in most plates transfected with the be seen in Figure 5a, compared to cells co-transfected complete LAPSER1 gene or CDKN1A gene. It is with reporter vector and a negative control noteworthy that the eect of the LAPSER1 gene on (pCNKLC) using an identical approach, all LAP- the TSUPr1 and PC3 and U2Os cells appears to be less SER1-transfected cells grew signi®cantly more slowly. pronounced, allowing the formation of some colonies. A slightly stronger cell-growth inhibitory eect was This result indicates that the growth of these cells may observed in all cell lines transfected with the only be attenuated by the expression of this gene.

Oncogene The LAPSER1 gene Y Cabeza-Arvelaiz et al 6712

Figure 5 Cell growth proliferation and colony-forming eciency analysis after transfection of the LAPSER1 gene into human and rodent cancer cell lines. (a) Following co-transfection of the indicated cell lines with the reporter vector pEGFPN1 and either LAPSER1, CDKN1A (SDI1/p21), or control plasmid (pCNKLC), cells were sorted 30 h post-transfection as described in Materials and methods and used to determine the growth rate after 96 h in culture. The SDI1/p21 gene was included in the analysis as a control prototype TSG. Each bar represents cell counts from duplicate plates from two independent experiments, and error bars represent the standard error. (b) After transfection of the indicated cell lines with either the LAPSER1, CDKN1A, or control (pCNKLC) constructs as described in Materials and methods, G418 selection was started 48 h post-transfection and continued for 12 ± 14 days when plates were stained. Representative plates from various transfection experiments are shown for each cell line assessed

Discussion ization of LAPSER1, a novel candidate TSG located on 10q24.3 that is related to LZTS1 another In this study, we report the identi®cation, cloning, candidate TSG gene located on 8p22 (Ishii et al., tissue expression analysis and functional character- 1999; Cabeza-Arvelaiz et al., 2001). These two genes

Oncogene The LAPSER1 gene Y Cabeza-Arvelaiz et al 6713

Figure 6 Comparison of the predicted amino acid sequences and deduced structural functional domains of the LZTS1 (FEZ1) and LAPSER1 proteins. (a) ClustalW alignment depicting the degree of protein sequence identity between LZTS1 and LAPSER1. Darker gray background highlights identical amino acids and lighter gray background highlights conservative amino acid substitutions. Residue positions on each sequence are indicated on the right in italics. The leucine residues of potential leucine zipper motifs are indicated by asterisks on top of the LZTS1 or at the bottom of the LAPSER1 sequences; disruptions in these motifs in the LZTS1 sequence are indicated by X. Putative phosphorylation motifs of various kinases (cAMP-dependent protein kinase A (PKA), protein kinase C (PKC), cGMP-dependent kinase (PKG), CAM-dependent kinase II (CkinII), casein kinases I and II (CK1 and CK2), CDC2-type kinase (CDC2), autophosphorylation-dependent kinase (ADK), tyrosine kinase (TyrKin), and glycogen synthase kinase-3 (GSK3)), are indicated by underlining (LAPSER1) or over-lining (LZTS1). The predicted N-terminus sequence if a second potential initiation residue were used (not included in the numbering) of LAPSER1 is shown. (b,c) Top section: Schematic representation of the LAPSER1 (b) and LZTS1 (c) structural domains. Compositionally biased regions are indicated by the single letter code of the abundant amino acid(s) (PGS, SGP, SG, AP, and Q). The positions of a deca-serine (S10) and an octo-proline (P8) stretch in the LAPSER1 sequence are also indicated. The locations of putative phosphorylation boxes (P-Box) are indicated by cross-hatching. Potential leucine zipper (LZ) motifs are depicted by vertically striped boxes when intact or by hatched boxes when disrupted. The two glutamine-rich domains (Q1 and Q2) are shown as shaded gray boxes, one of which (Q1) overlaps with the LZ1 motif. Regions of relatively high charge density are indicated (+/7 or 7). Intron splice sites are indicated by downward arrows. Bottom part: Secondary structure prediction of the LAPSER1 and LZTS1 sequences by the Chou and Fasman algorithms. The predicted helical structures and b-sheet structures are shown as black solid boxes and gray shaded boxes respectively. Open boxes indicate predicted turns. The scale at the bottom of each diagram indicates the position of the amino acid residues in each protein molecule

Oncogene The LAPSER1 gene Y Cabeza-Arvelaiz et al 6714 are located in the most frequently deleted regions in tains dierent potential phosphorylation sites suggest- prostate cancer and may play a role in the ing that they could be dierentially phosphorylated. development of prostate cancer as well as other The LZ1 and LZ2 motifs in LZTS1 are disrupted by cancers. This possibility is supported by the results the substitution of leucine residues at two positions by described in this study. other amino acids (tyrosine, phenylalanine, valine or The LAPSER1 gene was originally identi®ed on isoleucine), which may aect its interaction with other the basis of homology with the LZTS1 gene, and, as proteins. shown in Figure 2, the two genes appear to be Although we currently do not know the precise closely related. Based on the coding sequences of the biochemical function of the novel protein family that two genes, we have also examined structures and includes LAPSER1 and LZTS1, several features of functional domains of the encoded proteins. The their genes and predicted protein sequences point to a alignment of the LAPSER1 and LZTS1 protein possible role in gene transcription modulation at sequences shown in Figure 6a illustrates the homol- certain point(s) during the cell division cycle. Both ogy between these two protein molecules. Overall, are composed of various protein domains (P-Box, Q- 38% of the amino acid residues were identical and rich, and LZ) and, by analogy to the bZIP family of 15% were conservative changes. Searches for motifs transcription factor, they can be regarded as modular and compositional analyses inside the LAPSER1 and inasmuch as both are composed of apparently similar LZTS1 protein sequences revealed the presence of functional domains. Indeed, the overall structural multiple potential phosphorylation sites for various topography of LAPSER1 and LZTS1 is reminiscent kinases, which are indicated for both the LZTS1 and of the structure of the bZip transcription factors such the LAPSER1 sequences in Figure 6a. Also, three as CREB/ATF and CREM, and other proteins that are leucine zipper (LZ) motifs (Figure 6a) were detected involved in gene transcription modulation and contain in the LAPSER1 sequence. LZ1 is encoded by exon 4 similar functional domains (Courey and Tjian, 1988; and LZ2 and LZ3 by exon 5. In addition, Foulkes and Sassone-Corsi, 1996; Habener, 1990; compositionally biased or low complexity regions Molina et al., 1993; Pirrotta et al., 1987) in which the (Wootton and Federhen, 1996), were found in both P-Box domain at the N-terminal is an activation protein sequences. These regions are predominantly domain and the leucine zippers at the C-terminal are composed of serine, glycine and proline residues as protein-interacting domains ¯anked by Q-rich activa- indicated in Figure 6b,c by the single letter code of tion domains. Some protein factors involved in the the enriched amino acid(s) in the order of their regulation of gene transcription interact directly with abundance. Thus, at its N-terminal half, LAPSER1 DNA regions (enhancers and promoters) whereas has a PGS-rich region and a SGP-rich region, which others are involved in protein ± protein interaction. are encoded by exons 2 and 3, respectively. Other Transfection of LAPSER1 cDNA into various interesting features at the N-terminus include the prostate and other cancer cells signi®cantly decreased presence of an octa-proline repeat and a deca-serine the growth rate of these cells, indicating that repeat in the SGP-rich region. At its C-terminus, LAPSER1, like LZTS1, is involved in the regulation LAPSER1 contains two glutamine-rich (Q1 and Q2) of cell growth. The eects of LAPSER1 expression on regions, which are encoded by exons 4 and 5, the growth properties of most cells was comparable to respectively, ¯anking the LZ motifs. Similar regions those observed with the prototype cell cycle arresting are observed in the LZTS1 protein sequence except gene CDKN1A. The rapid decrease in the number of that the proline content in the N-terminal regions is GFP+ cells during transient expression with both not as high. Interestingly, the adjacent basic region LAPSER1 and CDKN1A (Figure 6a) was probably due (DNA recognition sub-domain) that is usually found to extensive cell loss, as indicated by the larger number upstream of LZ motifs in transcription factors is not of observed lysed and apoptotic-like cells in the plates present in LAPSER1 or LZTS1. Instead, they are with these transfectants. This suggests, as with preceded by Q-rich or charged (+/7) stretches. CDKN1A (Gotoh et al., 1997), that LAPSER1 eects Secondary structure prediction algorithms (Chou may be mediated through inhibition of cell cycle and Fasman, 1978) detect helical structures at the division and/or by apoptosis. Interestingly, the eects C-termini of both proteins that correlate with the of LAPSER1 expression on cell growth properties were detected LZ domains and Q-rich regions (Figure found to be similar to those observed with ectopic 6b,c). The relatively unstructured N-termini of these expression of LZTS1 in some cell lines (e.g. AT6.2, proteins contain short stretches of sequences with a HEK-293, and U2Os), but dierent in others (e.g. Rat1 high potential for a-helical and b-sheet structure, a and PC3) (Cabeza-Arvelaiz et al., 2001, and unpub- prerequisite in transcription factors for protein lished data). Further comparative work is in progress interaction with other components of the transcrip- to address these issues. tional activation or repression machinery. The similarity of LAPSER1 to proteins involved in A careful examination of the LZTS1 protein gene transcription modulation is extended further by sequence reveals that it diers from the LAPSER1 the various mRNA isoforms produced by alternative sequence at several stretches. For instance, outside the splicing of various coding exons leading to changes in highly conserved C-terminal sub-domains, homology the structure and function of the resulting proteins. between the two decreases signi®cantly. LZTS1 con- The Northern blot and RT ± PCR analyses suggested

Oncogene The LAPSER1 gene Y Cabeza-Arvelaiz et al 6715 that the number of isoforms appear to dier in a cell (Frequently Rearranged in Advanced T-cell lympho- and tissue-speci®c manner. Although the LAPSER1 mas) locus. major transcript is abundantly expressed in normal PTEN has been implicated in the development of prostate its expression was absent in some prostate prostatic carcinomas and has been reported to be cancer cell lines, while in other prostatic cancer cell mutated in several prostate cell lines. However, results lines the ratio of major to minor transcripts was from several studies using primary prostate adenocar- signi®cantly altered. Assuming similar eciency of cinomas or low stage tumors detect few or no ampli®cation of the various isoforms between the mutations in most tumors, indicating either that PTEN dierent cell lines the results suggest that the dierent alteration is a late event in prostate cancer develop- isoforms are dierentially expressed in each cell line. ment or that another TSG critical in prostate cancer The RT ± PCR experiment (Figure 5b) suggested that may lie close to PTEN (Dong et al., 1998; Feilotter et down-regulation of expression of the major LAPSER1 al., 1998; Orikasa et al., 1998). Similar ®ndings from transcript or upregulation of the alternatively spliced studies with other types of cancers have been reported transcripts could represent a mechanism of attenuation (Chiariello et al., 1998; Yeh et al., 1999; Yokomizo et of LAPSER1 eects. al., 1998). The localization of the LAPSER1 to 10q24.3 Multiple screening studies for genetic alterations in near the PTEN locus and its likely involvement in cell chromosome 10q in various cancers have frequently growth or gene transcription modulation suggest that revealed two common minimally deleted regions in LAPSER1 may well represent the other TSG on this prostate tumors and other cancers (Cairns et al., 1997; region. Carter et al., 1990; Gray et al., 1995; Ittmann, 1996; Kim et al., 1998; Petersen et al., 1998). Table 1 summarizes the rates of LOH that have been detected with microsatellite markers within this region. Region Materials and methods 1, near the PTEN locus, is detected by probing for microsatellite markers D10S541, D10S608, or D10S185 Cell lines and tissue samples and region 2, more telomeric on 10q, is detected AT6.2, a rat prostate cell line was provided by Dr J Isaacs between markers D10S198 and D10S192. As shown in (Johns Hopkins University, Baltimore, MD, USA). Rat-1, a Figure 1b, these two frequently deleted sub-regions rat ®broblast cell line was obtained from American Type (boxed) are located at *90 ± 95 Mb (containing Culture Collection (ATCC, Rockville, MD, USA), as were PTEN)andat*102 ± 106 Mb (containing LAPSER1) the human prostate cancer cell lines (LNCaP, PC3, PC3-m, in a region that includes multiple genes related to cell PPC1, DU-145, TSUPr1, and ND1), the human osteoblast- division and DNA repair, including several transcrip- derived sarcoma cell lines (SaOs2 and U2Os), the human tion factor genes (HHEX, HOX11, HUG1, PAX2, embryonal kidney (HEK) cell line HEK293, the human non- CHUK (NFKBIKA), LBX1, PMX1, NFKB2 and small cell lung cancer cells (NSCLC) H1299, and the human MXI1). A more detailed mapping (Figure 1b) of the ®broblast cell line BUD-8. Cells were maintained in complete medium RPMI 1640 or D-MEM (low glucose) with 2 mML- LAPSER1 gene within the 102 ± 106 Mb sub-region glutamine (Life Technologies, Gaithersburg, MD, USA) indicates that this gene is located between the PAX2 supplemented with 10% fetal bovine serum (Life Technolo- (paired box gene 2) and the HOX11 (homeo box 11; or gies) in the presence of 0.4% antibiotic-antimycotic mixture, TCL3, T-cell leukemia 3 gene) loci. These two genes in a 5% CO2 environment at 378C. harbor breakpoints involved in 10q24 translocations associated with human diseases (Hatano et al., 1991; LAPSER1 cloning and molecular analyses Narahara et al., 1997). In addition, the region is near fragile sites, such as the papilloma virus integration site Primers ¯anking the coding exons 2 and 5 of the LAPSER1 HPV6AI1 (Kahn et al., 1994) and the FRAT1 gene were used to amplify LAPSER1 cDNA sequences from normal human prostate or testis tissues (Clontech) by reverse transcription RT ± PCR. This cDNA was cloned into the expression vector pcDNA3.1 (Invitrogen, Carlsbad, CA, USA) and sequenced using a primer designed against the Table 1 LOH rates in chromosome 10q* ¯anking T7 sequence and by primer walking with primers designed against internal sequences. Sequence analysis was Tissue Region 1 Region 2 Reference performed using a BigDye Terminator kit (Perkin-Elmer Prostate 2/7 6/7 Carter et al. (1990) Biosystems, Foster City, CA, USA) and an automated Prostate 19/23 12/23 Gray et al. (1995) sequencer (Perkin-Elmer) according to the manufacturer's Prostate 10/15 6/15 Ittmann (1996) protocols. For some Northern RNA expression analysis, a Prostate 23/26 19/26 Cairns et al. (1997) commercial ®lter (Human MTN Blot IV, Clontech, Palo Lung 7/19 15/19 Kim et al. (1998) Alto, CA, USA) containing 2 mg poly(A)+ RNA per lane Lung 9/17 6/17 Petersen et al. (1998) was used. In other cases, 10 ± 20 mg of total RNA isolated from cell lines or peripheral blood lymphocytes (PBL) using *The rate of LOH is expressed as the number of cases showing Trizol (Life Technologies) was run on a denaturing absence of at least one microsatellite marker at the indicated region formaldehyde agarose gel, transferred to a nylon membrane divided by the total number of cases that had LOH through the 10q arm. Region 1 (encompassing PTEN) is de®ned by markers D10S215, and ®xed. Northern ®lters were probed in Expresshyb D10S185, D10S541 and D10S564. Region 2 (encompassing LAP- solution (Clontech) with a LAPSER1 exon 3 fragment, a SER1) is de®ned by markers D10S198, D10S603 and D10S192 complete 2-Kb cDNA, or a human b-actin cDNA control

Oncogene The LAPSER1 gene Y Cabeza-Arvelaiz et al 6716 probe radio-labeled using the RadPrime DNA labeling a FACS-SORT ¯ow cytometer (Becton-Dickinson, San Jose, system (Life Technologies). Washes were performed at 658C CA, USA) equipped with an Argon ion laser tuned at with 0.16SSPE and 0.5% SDS. 488 nm. The GFP+ cells were monitored in the FL1 For expression analysis by RT ± PCR, 100 ng of normal emission channel using a standard 530+30 nm band pass. human prostate or testis poly(A)+ RNA (Clontech) or 2 mg The sort region used to gate GFP+ cells was set so as to of total RNA was used to synthesize ®rst-strand cDNA using exclude 98% of the non-transfected cells (Lampariello, 1994). the SuperScripttm pre-ampli®cation system (Life Technolo- The total population of GFP+ cells from two dierent plates gies) with an oligo(dT) primer. LAPSER1 transcripts were was sorted separately. Collected cells were plated, allowed to ampli®ed by thermal cycling for 35 cycles using the grow for an additional 96 h, and counted using a ELONGASEtm kit (Life Technologies), with the following hemocytometer. forward (F) and reverse (R) primers (5'?3'): LASPER1F-2a (caagggatcctcatgggtagtgtgagc) and LAPSER1R-5 (gagc- Nucleotide sequences, computer-assisted searches and gaattcctgagggcctagatctc). PCR of human LAPSER1 exon 3 alignments (695 bp), and LZTS1 exon 3 (730 bp) was performed using the primer (5'?3') sets: LAPSER1F-2b (ggcacctcatttcagaa- The LAPSER1 gene sequence was derived from BAC RP-11 catgg) and LAPSER1R-2 (ccatcattgcctctgaccacac); LZTS1F- 108L7 (accession # AL133215) reported by the Sanger Centre 3 (gtgggaggtgtgtgccagaagtc) and LZTS1R-3 (gatatcgc- (Hinxton, UK; http://webace.sanger.ac.uk/HGP/Chr10), and caggtccccagac). the LZTS1 sequences were obtained from GenBank (accession # AF123659 and AC025853). To identify genes, BACs, and ESTs signi®cantly related to LZTS1, BLAST Transient and stable transfections, cell sorting and cell analysis (Altschul et al., 1990) searches were performed with proliferation assays the LZTS1 gene and cDNA sequences against the non- The pCNLAPSER1 plasmid contains the wild type LAP- redundant (nr) GenBank database (http://www.ncbi.nlm.nih.- SER1 cDNA coding region cloned into pcDNA-3.1 (Invitro- gov/Genbank). The sequence most related to the LZTS1 gene gen). The pCNCDKN1A plasmid contains the wild type was a region of BAC RP11-108L7 from contig AL133215 CDKN1A cDNA cloned into the pCMVex expression vector (http://webace.sanger.ac.uk), which had tentatively been and was kindly provided by Dr JR Smith (Baylor College of predicted by the Ensembl system (http://www.ensembl.org) Medicine, Houston, TX, USA). The pCNKLC plasmid to contain two new genes, namely ENSG00000076052 and contains the human kinesin light chain cDNA from plasmid ENSG00000055948, which we have provisionally termed pLBL1 (Cabeza-Arvelaiz et al., 1993), cloned into LAPSER1. The sequence of the LAPSER1 gene was used pCDNA3.1. The GFP reporter vector pEGFP-N1 was to perform BLASTn searches against the dBEST database obtained from Clontech. The expression of the inserted gene (Genbank). To construct a precise map of the 10q24 region in all these plasmids is driven by the human cytomegalovirus sequencing and mapping data were compiled from several (CMV) immediate early gene promoter. Cells were grown in sources (GenBank accession # NT_008874; UDB, http:// 100 mm plates to 50 ± 60% con¯uency overnight and bioinformatics.weizmann.ac.il/cgi-bin/udb/; GDB, http:// transfected with 5 mg of plasmid DNA, puri®ed using the www.gdb.org; Ensembl system http://www.ensembl.org; the Endofree kit from QIAGEN (Valencia, CA, USA), and 15 ml Whitehead Institute/MIT Center for Genome Research, of 2 mg/ml lipofectamine reagent (Life Technologies). For co- http://carbon.wi.mit.edu:8000/cgi-bin/contig?contig=WC10.8; transfections 1 mg of reporter pEGFP-N1 plasmids and 5 mg and the Draft Genome Browser, http://genome.ucsc.edu/cgi- of the plasmid containing the complete LAPSER11 cDNA or bin/hgTracks?seqNAME=chr10). Searches for motifs and the control plasmid were used. Transfections were performed compositional analysis within the LAPSER1 and LZTS1 by incubating the cells in the transfection mixture plus serum protein sequences were performed by scanning the Prosite free medium for 6 h before refeeding with complete medium. databases using ScanProsite and Pro®leScan (http://www.is- For the study of stable colony formation G-418 (500 mg/ml) rec.isb-sib.ch/cgi-bin/PFSCAN). selection was started 36 h post-transfection and continued for 2 ± 3 weeks, changing the medium every other day until colonies were visualized by methylene blue/glutaraldehyde staining. For sorting at 30 h post-transfection cells were harvested, Acknowledgments washed, ®ltered through a 60 mm nylon mesh, and The GenBank accession number for the LAPSER1 cDNA resuspended at 16106 cells/ml in phosphate buered saline is AY029201. This study was supported by NIH SPORE plus 5% fetal bovine serum. Cell sorting was performed using Grant P50-CA58204.

References

Altschul SF, Gish W, Miller W, Myers EW and Lipman DJ. CairnsP,OkamiK,HalachmiS,HalachmiN,EstellerM, (1990). J. Mol. Biol., 215, 403 ± 410. Herman JG, Jen J, Isaacs WB, Bova GS and Sidransky D. Bova GS, MacGrogan D, Levy A, Pin SS, Bookstein R and (1997). Cancer Res., 57, 4997 ± 5000. Isaacs WB. (1996). Genomics, 35, 46 ± 54. Carter BS, Ewing CM, Ward WS, Treiger BF, Aalders TW, Cabeza-Arvelaiz Y, Shih LC, Hardman N, Asselbergs F, Schalken JA, Epstein JI and Isaacs WB. (1990). Proc. Bilbe G, Schmitz A, White B, Siciliano MJ and Lachman Natl. Acad. Sci. USA, 87, 8751 ± 8755. LB. (1993). DNA Cell Biol., 12, 881 ± 892. Chiariello E, Roz L, Albarosa R, Magnani I and Finocchiaro Cabeza-Arvelaiz Y, Sepulveda JL, Lebovitz RM, G. (1998). Oncogene, 16, 541 ± 545. Thompson TC and Chinault AC. (2001). Oncogene, Chou PY and Fasman GD. (1978). Annu. Rev. Biochem., 47, 20, 4169 ± 4179. 251 ± 276.

Oncogene The LAPSER1 gene Y Cabeza-Arvelaiz et al 6717 Courey AJ and Tjian R. (1988). Cell, 55, 887 ± 898. Molina CA, Foulkes NS, Lalli E and Sassone-Corsi P. DongJT,SipeTW,HyytinenER,LiCL,HeiseC, (1993). Cell, 75, 875 ± 886. McClintock DE, Grant CD, Chung LW and Frierson Jr. Mount SM. (1982). Nucleic Acids Res., 10, 459 ± 472. JF. (1998). Oncogene, 17, 1979 ± 1982. Murakami YS, Albertsen H, Brothman AR, Leach RJ and Feilotter HE, Nagai MA, Boag AH, Eng C and Mulligan White RL. (1996). Cancer Res., 56, 2157 ± 2160. LM. (1998). Oncogene, 16, 1743 ± 1748. Narahara K, Baker E, Ito S, Yokoyama Y, Yu S, Hewitt D, Foulkes NS and Sassone-Corsi P. (1996). Biochim. Biophys. Sutherland GR, Eccles MR and Richards RI. (1997). Acta., 1288, F101 ± F121. J.Med. Genet., 34, 213 ± 216. Fujiwara Y, Ohata H, Kuroki T, Koyama K, Tsuchiya E, Orikasa K, Fukushige S, Hoshi S, Orikasa S, Kondo K, Monden M and Makamura Y. (1995). Oncogene, 10, Miyoshi Y, Kubota Y and Horii A. (1998). J. Hum. Genet., 891 ± 895. 43, 228 ± 230. GotohA,KaoC,KoSC,HamadaK,LiuTJandChungLW. Petersen S, Rudolf J, Bockmuhl U, Gellert K, Wolf G, (1997). J. Urol., 158, 636 ± 641. Dietel M and Petersen I. (1998). Oncogene, 17, 449 ± 454. Gray IC, Phillips SM, Lee SJ, Neoptolemos JP, Weissenbach Pirrotta V, Manet E, Hardon E, Bickel SE and Benson M. J and Spurr NK. (1995). Cancer Res., 55, 4800 ± 4803. (1987). EMBO J., 6, 791 ± 799. Habener JF. (1990). Mol. Endocrinol., 4, 1087 ± 1094. Steck PA, Pershouse MA, Jasser SA, Yung WK, Lin H, Hatano M, Roberts CW, Minden M, Crist WM and Ligon AH, Langford LA, Baumgard ML, Hattier T, Davis Korsmeyer SJ. (1991). Science, 253, 79 ± 82. T, Frye C, Hu R, Swedlund B, Teng DH and Tavtigian SV. Ichikawa T, Nihei N, Suzuki H, Oshimura M, Emi M, (1997). Nat. Genet., 15, 356 ± 362. Nakamura Y, Hayata I, Isaacs JT and Shimazaki J. TengDH,HuR,LinH,DavisT,IlievD,FryeC,Swedlund (1994). Cancer Res., 54, 2299 ± 2302. B, Hansen KL, Vinson VL, Gumpper KL, Ellis L, El- Ishii H, Baa R, Numata SI, Murakumo Y, Rattan S, Inoue Naggar A, Frazier M, Jasser S, Langford LA, Lee J, Mills H, Mori M, Fidanza V, Alder H and Croce CM. (1999). GB, Pershouse MA, Pollack RE, Tornos C, Troncoso P, Proc. Natl. Acad. Sci. USA, 96, 3928 ± 3933. Yung WK, Fujii G, Berson A, Bookstein R, Bolen JB, Ittmann M. (1996). Cancer Res., 56, 2143 ± 2147. Tavtigian SV and Steck PA. (1997). Cancer Res., 57, Kagan J, Stein J, Babaian RJ, Joe YS, Pisters LL, Glassman 5221 ± 5225. AB, von Eschenbach AC and Troncoso P. (1995). Vlietstra RJ, van Alewijk DC, Hermans KG, van Steen- Oncogene, 11, 2121 ± 2126. brugge GJ and Trapman J. (1998). Cancer Res., 58, 2720 ± Kahn T, Turazza E, Ojeda R, Bercovich A, Stremlau A, 2723. Lichter P, Poustka A, Grinstein S and zur Hausen H. Whang YE, Wu X, Suzuki H, Reiter RE, Tran C, Vessella (1994). Cancer Res., 54, 1305 ± 1312. RL, Said JW, Isaacs WB and Sawyers CL. (1998). Proc. Kim SK, Ro JY, Kemp BL, Lee JS, Kwon TJ, Hong WK and Natl. Acad. Sci. USA, 95, 5246 ± 5250. Mao L. (1998). Oncogene, 17, 1749 ± 1753. Wootton JC and Federhen S. (1996). Methods Enzymol., 266, Kozak M. (1991). J. Cell Biol., 115, 887 ± 903. 554 ± 571. Lampariello F. (1994). Cytometry, 15, 294 ± 301. Yeh JJ, Marsh DJ, Zedenius J, Dwight T, Delbridge L, Li J, Yen C, Liaw D, Podsypanina K, Bose S, Wang SI, Puc Robinson BG and Eng C. (1999). Genes Chrom. Cancer., J, Miliaresis C, Rodgers L, McCombie R, Bigner SH, 26, 322 ± 328. Giovanella BC, Ittmann M, Tycko B, Hibshoosh H, Yokomizo A, Tindall DJ, Hartmann L, Jenkins RB, Smith Wigler MH and Parsons R. (1997). Science, 275, 1943 ± DI and Liu W. (1998). Int. J. Oncol., 13, 101 ± 105. 1947. MacGrogan D, Levy A, Bova GS, Isaacs WB and Bookstein R. (1996). Genomics, 35, 55 ± 65.

Oncogene