Genomic Structure of Human OKL38 Gene andIts Differential Expression in Kidney Carcinogenesis*

Received for publication, August 6, 2003, and in revised form, October 10, 2003Published, JBC Papers in Press, October 21, 2003, DOI 10.1074/jbc.M308668200

Choon Kiat Ong‡, Chuan Young Ng§, Caine Leong‡, Chee Pang Ng‡, Keong Tatt Foo§,Puay Hoon Tan¶, and Hung Huynh‡�

From the ‡Laboratory of Molecular Endocrinology, Division of Cellular and Molecular Research,National Cancer Center, Singapore 169610 and the Departments of §Urology and ¶Pathology, Singapore GeneralHospital, Outram Road, Singapore 169608

We previously demonstrated the growth inhibitoryproperty of OKL38 and its possible roles in mammarycarcinogenesis. To further understand the regulationand roles of OKL38 in tumorigenesis we proceeded toclone and characterize the human OKL38 gene andthree of its variants with transcripts of 1.9, 2.2, and 2.4kb. The human OKL38 gene spans �18 kb and contains 8exons and 7 introns with exon size ranging from 92 to1270 bp. RT-PCR and sequence analysis suggest thatdifferent transcripts were arrived through differentialpromoter usage and alternate splicing. Multiple TissueExpression array (MTE) and Multiple Tissue Northernblot (MTN) indicated that OKL38 was ubiquitously ex-pressed in all tissues with high expression in liver, kid-ney, and testis. The cancer profiling array (CPA) ofpaired normal/tumor cDNA showed that OKL38 mRNAwas down-regulated in 70% (14 of 20) of kidney tumors.Western analysis revealed that the OKL38 protein wasundetectable in 78% (7 of 9 pairs) of kidney tumor tis-sues. Immunohistological analysis showed that 64% (14of 22) of kidney tumors were either lost or underex-pressed OKL38 protein compared with the adjacent nor-mal tissue. A transfection study using OKL38-eGFP re-combinant construct showed that overexpression of the52 kDa OKL38 protein in A498 cells resulted in growthinhibition and cell death. This study demonstrates thecomplex genomic structure of the OKL38 gene and itsgrowth inhibitory and cytotoxic properties. Our datasuggest the potential use of OKL38 in diagnosis, progno-sis, and/or treatment of kidney cancer.

Renal cell carcinoma (RCC)1 is the most common malignanttumor of the adult kidney, accounting for around 3% of human

malignancies (1). The incidence of RCC is increasing, and it isestimated that RCC accounts for 95,000 deaths per year world-wide (2). Complete surgical resection is considered to be theonly effective treatment for patients with clinically localizedRCC. However, the disease recurs postoperatively in 20–40%of patients who undergo potentially curative nephrectomy (1).Accurate prediction of long term cancer-free survival immedi-ately after resection of clinically localized diseases, therefore,will be valuable for planning follow-up protocols and for iden-tifying patients with a high risk of recurrence and for patientswho would benefit the most from adjuvant therapy. Althoughconventional prognosis factors such as pathologic tumor stageand grading are useful, other novel prognostic parameters,including clinical, laboratory, and biomolecular factors, will beneeded to provide additional predictive value (3).

OKL38 has previously been cloned and partially character-ized in our laboratory as a pregnancy-induced growth inhibi-tory gene (4). This gene is ubiquitously expressed in all rattissues with the highest levels detected in the ovary, kidney,and liver. OKL38 expression is increased during pregnancyand lactation in the rat mammary gland; however, low levels ofOKL38 transcripts are observed in various human breast can-cer cell lines and barely detectable in DMBA-induced rat mam-mary tumors. Transfection of MCF-7 cells with OKL38 cDNAresulted in growth inhibition in vitro and reduction in tumorformation in vivo, suggesting that OKL38 may play a vital rolein the growth regulation and differentiation of breast epithelialcells during pregnancy and probably in tumorigenesis (4). Themolecular mechanisms by which OKL38 exerts its role ingrowth inhibition and differentiation are still unknown.

As a first step in understanding function and regulation,herein we report the cloning, sequencing, and genomic organi-zation of the human OKL38 gene. Three novel OKL38 cDNAswere cloned and characterized. Because OKL38 has been im-plicated in tumorigenesis, we investigated the expression ofOKL38 transcripts and protein in normal and tumor kidney.Beside growth and differentiation, the transfection studyshows that OKL38 may also play an important role in celldeath.

EXPERIMENTAL PROCEDURES

Probe Labeling—All probes used for library screening as well as forSouthern and Northern blot analyses were radioactively labeled with[�-32P]deoxy-CTP (ICN, Costa Mesa, CA) using the Rediprime II DNALabeling System (Amersham Biosciences) as described by the manufac-turer. Unincorporated nucleotides were removed using a nucleotidepurification kit (Qiagen, GmbH, Hilden, Germany).

Screening, Subcloning, and Sequencing of the OKL38 Gene—Twocosmid clones containing human OKL38 gene were isolated from thehuman cosmid library (Clontech) using the previously reported humanOKL38 cDNA (4) as the probe. Restriction digestion and Southernanalysis were used to estimate the insert size and confirm their iden-tity. Shotgun strategy was adopted to subclone the cosmid insert for

* This work was supported by National Cancer Centre Tissue Repos-itory and Grant LS/00/019 from A*STAR-BMRC (to H. H.). The costs ofpublication of this article were defrayed in part by the payment of pagecharges. This article must therefore be hereby marked “advertisement”in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

The nucleotide sequence(s) reported in this paper has been submittedto the GenBankTM/EBI Data Bank with accession number(s) HuOKL38:AF334780, HuOKL38-1a: AY258068, HuOKL38-2a: AY258067, andHuOKL38-2b: AY258066).

� To whom correspondence should be addressed: Laboratory of Molec-ular Endocrinology, Division of Cellular and Molecular Research, Na-tional Cancer Centre, Singapore 169610, Singapore. E-mail:cmrhth@nccs.com.sg.

1 The abbreviations used are: RCC, renal cell carcinoma; DMBA, 7,12-dimethylbenz(a)anthracene; RACE, rapid amplification of cDNAends; ORF, open reading frame; MTE, multiple tissues expression;CPA, cancer profiling array; eGFP, enhanced green fluorescence pro-tein; LOH, loss of heterozygosity; UTR, untranslated region; pyr-redox,pyridine nucleotide-disulfide oxidoreductase; Inr, initiator; DPE, down-stream promoter element; GFP, green fluorescent protein; MTN, mul-tiple tissue Northern blot.

THE JOURNAL OF BIOLOGICAL CHEMISTRY Vol. 279, No. 1, Issue of January 2, pp. 743–754, 2004© 2004 by The American Society for Biochemistry and Molecular Biology, Inc. Printed in U.S.A.

This paper is available on line at http://www.jbc.org 743

by guest on Novem

ber 21, 2020http://w

ww

.jbc.org/D

ownloaded from

sequencing. In brief, the cosmid clone was digested with various restric-tion enzymes, single or double digestion. The digested products wereseparated on an agarose gel and products of less than 6 kb were clonedinto pBluescript®SK (Stratagene, La Jolla, CA) vector with the sub-cloning map shown in Fig. 1A. Fragments with the expected size of thecosmid vector were excluded. The longer clones were fully sequenced bydirect cosmid sequencing and using the Template Generation System™(Finnzymes) as recommended by the manufacturer. The Genome-Walker™ kit (Clontech) was used to clone the missing 3�-end of theOKL38 gene using the GW-F forward primers (Table II) as described bythe manufacturer. All the sequencing was performed using automatedsequencing via dideoxy chain termination using the BigDye™ version3.0 (Applied Biosystems, Foster City, CA). All the regions were se-quenced at least three times.

Rapid Amplification of cDNA Ends (RACE) Analysis—To establishthe full-length cDNA of human OKL38, 5�-RACE was performed usingthe SMART™ RACE cDNA amplification kit (Clontech) according to themanufacturer’s direction with the following modification: The second-ary PCR of the primary RACE products was performed using Advan-tage® Genomic polymerase (Clontech). The gene-specific primer,GSP1R (Table II) was designed based on the 5�-end sequence of thepreviously isolated human OKL38 cDNA (4), while GSP2R was de-signed based on the 5�-end sequence of the GSP1R RACE fragment (Fig.3). The RACE products were identified using Southern blot analysisand subsequently cloned into the pCR®-Blunt-II-TOPO vector (Invitro-gen) and sequenced using T7 and SP6 primers. Sequence analysis wascarried out using Laser gene sequence analysis software (Dnastar Inc.).

Primer Extension Study—2 �M 32P-labeled primer (Pext1 on exon 1and Pext3 on exon 3) (Table II) were allowed to hybridized with 200 ngof human liver, kidney, and prostate poly(A)� mRNA (Clontech). Thetouchdown method was employed to increase the stringency of hybrid-ization with the following parameter: 85 °C, 15 min; 84 °C, 90 s; �0.1 °Cfor every subsequent cycle to dock at a final temperature of 50 °C. Thehybridization mix was added to the reverse transcription mix to a totalvolume of 100 �l containing 5� Superscript buffer, 0.1 M dithiothreitol,10 �M dNTP, 250 units of Superscript II enzyme (Invitrogen), Rnasin(Promega, Madison, WI), DEPC-H2O, and followed by extension for 1 h.The reaction was terminated by heating to 70 °C for 15 min, ethanol-precipitated, dried, redissolved in formamide-dye mix, and run on a 6%denaturing gel. The fmol™ cycle sequencing kit (Promega) was used togenerate a DNA ladder with subclone P-3.4 and S-5.5 (Fig. 1A) astemplate, and Pext1 and Pext3 as primers, respectively.

Cloning cDNA of HuOKL38-1a, -2a, and -2b cDNAs via PCR—Clon-ing of the full-length cDNA of the HuOKL38-1a transcript was achievedusing two PCR primers, forward primer 1F (nucleotide positions 7–33)designed at the 5�-end and reverse primer 1R (nucleotide positions1863–1889) (Table II) designed at the 3�-end of the HuOKL38-1a cDNA.To clone the HuOKL38-2a and -2b cDNAs, forward primer 2F (nucleo-tide positions 52–75 of the -2a transcript) (Table II) was used instead ofthe 1F primer (Fig. 3A). In all cases, the forward primer carried aHindIII and the reverse primer an EcoRI restriction site for directionalcloning. To generate the cDNA template for PCR cloning, 1 �g of humanliver mRNA was reversed-transcribed using SuperScript™II reversetranscriptase (Invitrogen) as recommended by the manufacturer. Thetemplate was amplified by PCR under the following conditions: dena-turation step at 94 °C for 2 min, followed by 35 cycles of 94 °C for 1 min,60 °C for 1 min, 72 °C for 2 min and 30 s, and a final extension at 72 °Cfor 5 min. PCR products of the correct size were cloned intopBluescript®SK (Stratagene) for sequencing.

In Vitro Transcription and Translation (TNT) Study—To verify thepredicted molecular weight of the putative ORF of OKL38, the TNT®-coupled reticulocyte lysate system (Promega) was used to transcribeand translate the HuOKL38-1a, -2a, -2b and the previously cloned

1.6-kb cDNA according to the manufacturer’s instructions. Briefly, thecloned cDNAs in pBluescript®SK (Stratagene) were digested withEcoRI (New England Biolabs, Beverly, MA) and purified twice by phe-nol/chloroform extraction. The DNA was precipitated with ethanol,washed with 70% ethanol, air-dried, and resuspended in H2O. 1 �g ofthe digested plasmid was used for the TNT reaction. The synthesizedprotein was labeled with L-[35S]methionine (ICN), and 5 �l of thereaction were loaded and electrophoresed on a 10% SDS-PAGE. Thesynthesized protein was transferred onto a nitrocellulose membraneand exposed to film. To confirm the identity of the in vitro synthesizedprotein, Western blot analysis was performed on the TNT productsusing rabbit anti-OKL38 antibody.

Multiple Tissues Expression (MTE) Array, Northern Blot Analysis,and Cancer Profiling Array (CPA)—The MTE and CPA were purchasedfrom Clontech and the human Northern blots were from Invitrogen.Both the arrays and blots were hybridized with a 700-bp SmaI humanOKL38 probe (936–1685 nt of HuOKL38-1a) that detects all differentvariants. Ubiquitin (provided by the manufacturer) and glyceralde-hyde-3-phosphate dehydrogenase (GAPDH; American Type CultureCollection, Manassas, VA) were used for normalizing arrays and North-ern blots, respectively. Pre-hybridization of the arrays and blots wereperformed at 50 °C for 1 h in ExpressHyb™ (Clontech) with shearedsalmon testis DNA added to a final concentration of 0.1 mg/ml. Thedenatured radiolabeled probe was mixed directly into the pre-hybrid-ization solution and hybridized overnight at 50 °C. After hybridization,the arrays were washed three times at 55 °C in solution 1 (2� SSC,0.5% SDS) for 30 min each time, repeated with solution 2 (0.2� SSC,0.5% SDS), and a final rinse in 2� SSC at room temperature. Blots wereexposed to a phosphorimager overnight. For the Northern blot analysis,washing was performed at 45 °C using solutions 1 and 2. mRNA levelswere determined by densitometric scanning of autoradiographs andnormalized to the level of GAPDH.

Semi-quantitative RT-PCR of OKL38 Variants—1 �g of total mRNAwas used as template for One Step RT-PCR (Qiagen, GmbH, Hilden,Germany) adopting the procedure recommended by the manufacturer.To study the expression of the HuOKL38-1a transcript, variant-specificforward primer PCR-1F (nucleotide positions 1–28) and a commonreverse primer PCR-1R (nucleotide positions 339–443 of HuOKL38-1a)(Table II) were designed for RT-PCR. To determine the expression ofHuOKL38-2a, -2b, and -2c transcripts, RT-PCR was performed usingthe forward primer PCR-2F (nucleotide positions 387–413) and reverseprimer PCR-1R (Table II). The specificity of the designed primers andPCR conditions were optimized using the three cloned cDNAs,HuOKL38-1a, -2a, and -2b as templates. A pair of tubulin primers,TubF (5�-AACGTCAAGACGGCCGTGT-3�) and TubR (5�-GACAGAG-GCAAACTGAGCAC-3�), which amplify a 400-bp fragment of tubulincDNA, was used for normalization. The One Step RT-PCR was per-formed as follows: 50 °C for 30 min; 95 °C for 15 min; followed by 38cycles of 94 °C for 1 min, 60 °C for 1 min, 72 °C for 1 min, and finalextension at 72 °C for 5 min. The amplified products were separated ona 2.0% agarose gel.

Patients and Tissue Samples—Tissue samples were obtained intra-operatively from tumors and adjacent non-tumor kidney tissues duringresection for kidney tumor at the Singapore General Hospital. Thesamples were snap frozen in liquid nitrogen and stored at �80 °C untilanalysis. A similar set of samples was fixed in 10% formalin andparaffin-embedded. Prior written informed consent was obtained fromall patients, and the study received ethics board approval at the Na-tional Cancer Centre of Singapore as well as the Singapore GeneralHospital.

OKL38 Antibody and Western Blot Analysis—Rabbit polyclonalOKL38 antibody was raised against the OKL38-specific peptide:CAVEWGTPDPSSCGAQ (amino acid positions 200–214). Affinity-pu-

TABLE IExon intron junctions of OKL38 gene

Exon no. Exon size 5� Splice donor Intron size 3� Splice acceptor

bp bp

1 440 CAGACAAGAGgtacgtcggc 1599 ccctctgtagGTTCCTGCTA2 160 CAGGACGAGGgtgaggaggg �2500 atccccacagGGTAATGGGT3 109 TGCCTGTCAGgtgagtgtcc 4270 cccactccagGTCCGCTGCC4 99 ATCATTGTGGgtgagtgtca 1521 ctctccccagGTAACGGCCC5 137 CCTGGACCAGgtgggtcagc 1172 ccccctccagGACCTGGACT6 192 AGCCTGGCACgtgagtgggg 220 ttccctgcagTCCATCGAAG7 92 AGAAGCGAAGgtgaggccgc 3990 tccccaacagAGGTCTTCGC8 1270

Genomic Structure of OKL38 Gene and Kidney Cancer744

by guest on Novem

ber 21, 2020http://w

ww

.jbc.org/D

ownloaded from

rified rabbit anti-human OKL38 antibody was diluted in Tris-bufferedsaline (TBS, 20 mM Tris, 200 mM NaCl, pH 7.6) containing 0.1%Tween-20 (TBST) at a final concentration of 1 �g/ml. Western analysiswas performed by incubating the blots with 1:2000 anti-OKL38 anti-body or anti-OKL38 antibody preadsorbed with 50� antigen peptide(control for antibody specificity) overnight at 4 °C and washed threetimes with TBST, 15 min each. Subsequently, the blots were incubatedwith 1:7500 horseradish peroxidase-conjugated donkey anti-rabbit sec-ondary antibody for 1 h. After washing three times with TBST, 15 mineach, the blots were then visualized with a chemiluminescent detectionsystem (Amersham Biosciences) as described by the manufacturer.

Immunohistochemical Analysis—5-�m thick sections were cut, de-waxed in xylene, and then rehydrated as described (5). Antigen retrivalwas performed by boiling the slides in 10 mM citrate buffer pH 6.0 for 20min. Endogenous peroxidase activity was blocked by 3% hydrogen per-oxide in methanol for 30 min. After two washes of TBS, the sectionswere preincubated with 5% skim milk in TBS containing 0.1%Tween-20 for 15 min to reduce nonspecific background staining. Thesection were washed twice with TBST for 5 min and incubated over-night at 4 °C, with purified primary antisera against human OKL38 orantisera preadsorbed with 50� antigen peptide to serve as a control forantibody specificity. Immunohistochemistry was performed using thestreptavidin-biotin peroxidase complex method according to the manu-facturer’s instructions (Lab Vision, Fremont, CA) using AEC as thechromogen.

Generating OKL38-eGFP-pcDNA3.0 Construct—The HuOKL38-1acDNA contained an open reading frame (ORF) of 477 amino acids. Tofuse this OKL38 protein to the eGFP via PCR, 4 primers namely 477-F,477-eGFP-R, eGFP-F, and eGFP-R (Table II) were designed. Two sep-arate PCR reactions were performed using the primer 477-F and 477-eGFP-R to amplify the ORF of OKL38 and primers eGFP-F and eGFP-Rto amplify the ORF of eGFP. The amplified products from the two PCRreactions were then mixed and reamplified using primers 477-F andeGFP-R. The PCR reaction was performed as follows: 95 °C for 5 min;followed by 25 cycles of 94 °C for 1 min, 55 °C for 1 min, 72 °C for 2 min,and final extension at 72 °C for 5 min. The recombinant products werecloned into pCR®-Blunt-II-TOPO vector (Invitrogen), screened for ori-entation and subsequently cloned into pcDNA3.0 (Invitrogen) mamma-lian expression vector. The OKL38-eGFP-pcDNA3.0 construct was fullysequenced, and endotoxin-free plasmid for transfection was preparedusing the Maxi-prep kit (Qiagen, GmbH, Hilden, Germany). The posi-tive control eGFP-pcDNA3.0 was constructed with the same strategyusing only the eGFP-F and eGFP-R primers for PCR cloning.

Cell Culture and Transfection—Human kidney cancer A498 cells weremaintained as monolayer cultures at 37 °C (5% CO2) in �-modified Eagle’smedium (�-MEM) plus phenol red, supplement with 10% fetal bovineserum, 1% penicillin, and streptomycin (Invitrogen). A498 cells wereseeded at 2 � 105 in 100-mm culture dishes containing a coverslip,swabbed with ethanol, and grown to 70% confluence prior to transfection.Cells were transfected with 10 �g of either eGFP-pcDNA3.0 orOKL38-eGFP-pcDNA3.0 plasmid DNA and 12 �l of LipofectAMINE rea-

gent (Invitrogen) following the manufacturer’s recommendations. Eachcoverslip was removed at 24, 48, 72, 96, and 120 h post-transfection usinga sterile forceps. The coverslip with cells was fixed with 10% formalin,washed with phosphate-buffered saline, and mounted onto a slide forobservation using a microscope (Olympus) equipped with epifluorescenceoptics and appropriate filters for fluorescein isothiocyanate.

Computational Analysis—Sequence identity and ORF prediction weredone using analysis software from the National Center for BiotechnologyInformation (NCBI). The ClustalW v1.82 program (EMBL) was used toperform the multiple sequence alignment. The predicted amino acid se-quences were analyzed using: (i) cPfam CDS-Conserved Domain Search(NCBI) for conserved domain detection and (ii) SignalP (6) for analyzing thepresence of signal peptide. The putative promoter region was analyzed usingthe MatInspector V2.2 (transfac.gbf.de/cgi-bin/matSearch/matsearch.pl),CpG Island Searcher (www.uscnorris.com/cpgislands/cpg.cgi) and RepeatMasker (ftp.genome.washington.edu/RM/RepeatMasker.html).

RESULTS

Cloning, Sequencing, Genomic Organization, and SequenceAnalysis of the Human OKL38 Gene—Two cosmid clones con-taining the human OKL38 gene were isolated with the longerclone containing an insert of �35 kb. Both clones contained theOKL38 gene as determined by Southern blotting and sequenc-ing. Direct cosmid sequencing of the insert ends indicate that�100 bp were missing from the 3�-end of OKL38 exon 8. A2.5-kb fragment containing this missing end was cloned usingthe GenomeWalker™ kit. The cloning map was shown in Fig.1A. The sequence information was deposited in GenBankTM

with the accession number AF334780.Comparing the sequenced genomic clone with the sequences

of the cloned cDNAs, the exon/intron junctions of the OKL38gene were identified. Each of the 5�-donor and 3�-acceptorsplice sites conformed to the consensus sequences with thehighly conserved, invariable GT/AG dinucleotides present atthe immediate exon/intron boundaries (Table I). The humanOKL38 gene spanned a genomic region of �18 kb and contains8 exons with sizes ranging from 92 to 1270 bp (Fig. 1B).

In the process of sequencing, an �850 bp CCCT-rich regionresiding in the subclone Bs/X-1.1 (Fig. 1A), of which 300 bpcould not be sequenced even after the SequenceRx enhancersolution F (Invitrogen) had been used, was identified. Thisregion was part of the intron 2, located 1.2-kb upstream of thepromoter P1 transcriptional start site (Fig. 2A). Another inter-esting region spanning �500 bp was identified with the RepeatMasker program as simple CA repeats. Careful investigationshowed that this region contained 11 alternate repeats of 26

TABLE IISequences of oligonucleotides used for RT-PCR primer, cloning, primer, sequencing primer, variant-specific primer, and 5�RACE primer

Oligonucleotide Sequence

Pext1 5�-CTGCTGACCCTGACCTTGTTCTAGA-3�Pext3/GSP2R 5�-GGTCAGGGGAACACGGATCACAGAGTCC-3�GSP1R/PCR-1R 5�-CGTAGAAGGGCATCAAAGAGCAGGG-3�1F 5�-AAGCTTGGATCCCCACAGGGTAATGGGTGT-3�1R 5�-GAATTCTGGAAGGCGCAGGGCTGCAGGTCT-3�2F 5�-AAGCTTGGCAGGGAGGAAAGTCCACGTCT-3�PCR-1R 5�-CGTAGAAGGGCATCAAAGAGCAGGG-3�PCR-2F 5�-GGGAAGTGGAGACTGAGAGGCTGCTGC-3�477-F 5�-CACCATGAGCTCCTCCAGAAAGGACCA-3�477-eGFP-R 5�-CTCGCCCTTGCTCACCATGGGTGGCTTCCTGGTCTC-3�eGFP-F 5�-ATGGTGAGCAAGGGCGAG-3�eGFP-R 5�-TTACTTGTACAGCTCGTCCA-3�Seq1Ra 5�-CTGGTCAGGAAGCCGCTCACCTGGAAGA-3�Seq2Ra 5�-CAGTCTTCCTTGAAGCACAGCAGCTGGT-3�Seq3Ra 5�-GATGGCGTGCTCCTTCCGGTGCTTCC-3�Seq4F/GW-Fa 5�-GAGATGACCACATCCCTGCTGGATGC-3�Seq5Fa 5�-GTGTTCAACCAGCTGCCCAAGATGC-3�P1F 5�-CTGAGGAAGAGGGAGGCAAGAGACAGAG-3�P1R 5�-GCTGATATGGGGAACTGTGGGGCGAGAC-3�V-1F 5�-AGGCATCACCTCAGGCACAA-3�V-1R 5�-CTGACAGGCAGTGGTGCAGGA-3�

a Sequencing primer.

Genomic Structure of OKL38 Gene and Kidney Cancer 745

by guest on Novem

ber 21, 2020http://w

ww

.jbc.org/D

ownloaded from

and 28 nucleotides, interrupted by three conserved 26-nucleo-tide repeats. These repeats could be arranged to form four longrepeats with repeats 2 and 3, each 160 bp, aligned perfectly,while 1 and 4 are partly conserved. This region contained�70% cytosine and adenine, and was localized 140-bp up-stream of exon 2 (Fig. 2A). The functional significance of theseunique regions is unknown at the present time.

BLAST analysis performed using the present sequence dataagainst the published draft human sequence (Genomic contig:NT_024797.13) (NCBI) revealed that the OKL38 gene waslocalized to chromosome 16 (16q23.3), which was prone to lossof heterozygosity (LOH) in a variety of tumors (7–10). Compar-ing our sequenced OKL38 gene with the published humangenome using ClustalW delineated two different regions. A�20-bp random deletion in the CA repeats region and a 550-bp

longer CCCT-rich region were observed in our sequencedOKL38 gene (Fig. 2B). To verify the presence of this longerCCCT-rich region in our cosmid clone, PCR was performedusing a forward primer, V1-F, which resides immediatelydownstream of exon 2 and reverse primer, V1-R, which residesin exon 3. A PCR product of �2.5 kb was synthesized instead ofa 1.95-kb fragment, suggesting that the 850 bp CCCT-richregion in our clone is not an artifact of cloning. These differ-ences suggest regions of DNA instability.

Sequence Analysis of the Two Putative Promoter Regions—Two putative promoters were identified while comparing the5�-RACE fragments with the cloned OKL38 genomic sequence(Fig. 2A). Analyzing the proximal 300-bp region upstream ofpromoter P1 using the MatInspector program identified severalpotential sites for the Ikaros factor (IK2), homeodomain factor

FIG. 1. Cloning strategies and genomic structure of the human OKL38 gene. A, a schematic diagram showing the contig map of theOKL38 gene derived by restriction digestion with B, BamHI; P, PstI; S, SacI; Bs/X, BstXI/XhoI; N/H, NotI/HindIII. The GW-2.5 fragment wascloned using the genome walk strategy, while P1F and P1R were primers designed for direct cosmid sequencing. The subclone Bs/X-1.1 containsan 850-bp region of CCCT repeats, which was cloned from subclone P-4.3. B, the genomic structure of human OKL38 showing differential splicingat the 5� region of the gene. The various transcripts derived from differential splicing of exons 2 and 3 are shown.

FIG. 2. Sequence comparison andanalysis of the OKL38 gene. A, detailedmap of the first four exons of the humanOKL38 gene. The distance between twoputative promoters, and the distance be-tween promoter P1 and the CCCT repeatsare shown. Arrows pointing down indicatetranscriptional start site. P1 and P2 indi-cate the positions of putative promoter 1and 2, respectively. Note that the tran-scription start site is located 22-bp up-stream of the exon 3 splice junction. B,our sequenced OKL38 gene showed twodifferent regions when compared withthat published by the human genome se-quencing project. A longer CCCT repeatregion and a deletion of 20 bp of the CArepeats are shown. Primers V-1F andV-1R were designed to verify the presenceof longer CCCT repeats in our sequencedOKL38 gene.

Genomic Structure of OKL38 Gene and Kidney Cancer746

by guest on Novem

ber 21, 2020http://w

ww

.jbc.org/D

ownloaded from

Nkx-2.5/Csx (NKX25), upstream stimulating factors (USF),myeloid zinc finger protein (MZF1), Sp1, and AP-1, -2, and -4.Using the CpG Island Searcher (lower limit values of 200 bp forlength, 55% for GC content, and 0.65 for observed CpG/ex-pected CpG ratio), a CpG island of 215 bp located 746-bpupstream of the promoter P1 transcriptional start site wasidentified. No TATA or CAAT boxes were observed upstream ofthis initiation site. However, an alternative core promoter mo-tif called the initiator (Inr) (11), which encompasses the initi-ation site, was identified. Interestingly, the sequence aroundthis start site, TCAGAGC matched the Inr consensus se-quences except for the penultimate G, indicating that this coresequence is likely responsible for driving transcription fromthis site. In support of the presence of the Inr, a putativedownstream promoter element (DPE) at positions �27 to � 33was identified (11). Four of the seven nucleotides of this ele-ment were conserved.

None of the basal transcription elements, such as TATA,CAAT, Inr, or DPE were observed in promoter P2. But severalpotential sites for IK2, USF, myb, and, many AP-1 and GATA1were observed in 300 bp of the proximal promoter region. ACpG island of 278 bp, located �2.5-kb upstream of the putativeinitiation site was identified.

5�-RACE and Primer Extension Study—To establish the full-length transcripts, 5�-RACE was performed using poly(A)�

mRNA derived from human kidney tissues. Using GSP1Rprimer residing in exon 6, a PCR fragment of 443 bp, whichconsists of exons 3, 4, and 5 was amplified (Fig. 3A). A GSP2R

primer, residing in exon 3 amplified a 494-bp fragment thatestablished exon 1 (Fig. 3A). This DNA fragment containingexon 1 could only be detected via Southern blot analysis. Nofragment containing exon 2 was cloned in the 5�-RACE exper-iment. To determine the possibility that the transcriptionalstart site was located in exons 1 and 3, a primer extensionstudy was performed using Pext1 and Pext3 primer residing inthe respective exons. However, only the transcription start sitepreceding exon 3 could be mapped in liver, kidney, and prostate(Fig. 3B) suggesting the presence of a cryptic promoter P1. Aminor transcriptional start site at position �10 was alsodetected.

Cloning and Sequence Analysis of HuOKL38-1a, -2a, and -2bcDNAs—Sequence comparison of the GSP1R and GSP2RRACE products, and the cloned genomic sequence showed thatthe first 22 bp (positions �1 to �22 at promoter P1) did notbelong to exon 3 and was unique to the HuOKL38-1a variant.To clone the full-length of the above cDNA, forward primer 1F(Table II) that resided mostly within the 22-bp unique region ofthe GSP1 RACE fragment and a common reverse primer, 1R(Table II) were designed for RT-PCR cloning of theHuOKL38-1a transcript. For cloning of the HuOKL38-2a and-2b transcripts, forward primer 2F residing in exon 1 was usedinstead of 1F (Fig. 4A). Three PCR products of �1.9, 2.2, and2.4 kb were cloned and five clones of each cDNA were se-quenced, which enabled us to establish three novel humanOKL38 cDNAs of 1930, 2240, and 2400 bp, respectively. The1.9-kb cDNA was annotated as HuOKL38-1a (GenBankTM ac-cession no. AY258068), the 2.2-kb cDNA as HuOKL38-2a (Gen-BankTM accession no. AY258067), and the 2.4-kb cDNA asHuOKL38-2b (GenBankTM accession no. AY258066).

The 5�-RACE analysis, one-step RT-PCR, and Southern blotanalysis indicated the presence of another minor transcript(HuOKL38-2c), which contained exons 1, 3, 4, 5, and 6. Thistranscript most likely also contains exons 7 and 8. However,due to the low expression of this transcript in the liver andkidney RNA used, we have not managed to clone the full-lengthsequence of this cDNA. The cloning and one-step RT-PCR re-sults confirmed the existence of OKL38 variants.

Sequence analysis indicated that differential promoter usageand alternative splicing at the 5�-region of OKL38 gene wereresponsible for the variants. The HuOKL38-1a transcript wasarrived as a result of promoter P1 usage, while -2a, -2b, and -2cexpression might be regulated by promoter P2. The HuOKL38-2a, -2b, and -2c were derived by alternative splicing of exons 2and 3, exon 3 and exon 2, respectively (Fig. 4A). TheHuOKL38-1a and -2a variants harbored an ORF of 1431 basepairs. The conceptual translation of these cDNAs predicted aprotein of 477 amino acids, with a calculated molecular mass of52 kDa (Fig. 4B) and a predicted pI of 6.6. Two other proteinisoforms with a molecular weight of 61 and 59 kDa were alsopredicted from HuOKL38-2b and -2c, respectively. In vitrotranscription and translation study using the three cDNAs,namely HuOKL38-1a, -2a, and -2b resulted in the synthesis of�52 and 61 kDa proteins (Fig. 5A), and their identity wasconfirmed by Western blot analysis using anti-OKL38 antibod-ies (Fig. 5B). The previously cloned OKL38 cDNA (GenBankTM

accession no. AF191740) showed the expression of a 38 kDaprotein. Western blot analysis also detected low levels of the 38kDa protein from these larger cDNA constructs suggesting thatinternal translation start codon (ATG) might exist (Fig. 5B).

Similar to the previously cloned human OKL38 protein, noputative signal peptide was found in all the three humanOKL38 protein isoforms using the publicly available SignalPprogram (6). Computational analysis of the predicted aminoacid sequence using Pfam CDS-Conserved Domain Search

FIG. 3. Schematic illustration of 5�-RACE and primer exten-sion analysis. A, the 1.6-kb cDNA has previously been cloned, whichconsists of only exons 6, 7, and 8. Gene-specific primer, GSP1R used in5�-RACE experiment established exons 3, 4, and 5, while GSP2R estab-lished exon 1. Both reverse primers, Pext1 and Pext3 were designed forprimer extension analysis. B, primer extension study performed usingLi, liver; Ki, kidney; Pr, prostate mRNA as templates and reverseprimer Pext3. The nucleotides G and T, indicate a major and a minortranscriptional start site, respectively. Note that Pext1 failed to obtainany extended product.

Genomic Structure of OKL38 Gene and Kidney Cancer 747

by guest on Novem

ber 21, 2020http://w

ww

.jbc.org/D

ownloaded from

(NCBI) detected a putative domain belonging to that of thepyridine nucleotide-disulfide oxidoreductase (Pyr-redox) in allthe three OKL38 isoforms, and the domain was encoded byexon 8.

Expression of OKL38 in Various Human Tissues—We havepreviously shown the differential expression of OKL38 in rattissues, but not in the human (4). MTE array was adopted todetermine the global tissue distribution of human OKL38 (Fig.6A). The MTE array showed that OKL38 was ubiquitouslyexpressed in all the tissues investigated, with high expressiondetected in the kidney (7:A), skeletal muscle (7:B), testis (8:F),liver (9:A), adrenal gland (9:C), and fetal liver (11:D) (Fig. 6B).Northern blot analysis showed that transcripts of �2.0–2.4 kbwere ubiquitously expressed in all 15 tissues investigated (Fig.6, C and E). High expression of these transcripts in liver,kidney, and testis confirmed the MTE array analysis. Largerputative transcripts ranging from 4.0 to 7.0 kb were also ob-served in the liver (Fig. 6C).

Preliminary Expression Study of Human OKL38 Variants inKidney, Liver, and Ovarian Tissue and Cell Line via Semi-quantitative One Step RT-PCR—Because the OKL38 has beenimplicated in regulating cell growth, differentiation (4), andtumorigenesis, the distribution of human OKL38 variants invarious human normal/tumor tissues and cancerous cell lineswere determined via One Step RT-PCR using variant-specificprimers (Fig. 4A). Four RT-PCR products of 443, 373, 533, and482 bp corresponding to the HuOKL38-1a, -2a, -2b, and -2ctranscripts, respectively, were amplified (Fig. 7, A–D). South-

ern blot analysis (Fig. 7C) was performed using exons 4 and 5as probes, and sequencing verifies the identity of the amplifiedproducts. All the four variants were expressed in all tissuesinvestigated. Liver appeared to express the highest level of allthe transcripts compared with kidney and ovary. In general thenormal kidney, liver, and ovarian tissues seemed to expresshigher levels of OKL38 transcripts compared with the adjacenttumor tissues. All the cancer cell lines showed lower expressionof OKL38 transcripts, except the A498 kidney cancer cell line.However, more paired tissue samples and cell lines are neededin order to determine the significant differences in the expres-sion level of each variant in normal and cancerous tissues.

Expression of OKL38 Protein in Paired Normal/Tumor Kid-ney—To determine the expression of OKL38 mRNA transcriptsin paired normal/tumor kidney, CPA was used. The blot wasprobed with 700-bp human OKL38 SmaI fragment, whichcould detect all OKL38 variants. Fig. 8 shows that 70% (14/20)of the kidney tumors expressed lower levels of OKL38 tran-scripts compared with the adjacent normal tissues.

The OKL38 mRNA transcripts were generally down-regu-lated in the kidney tumors, and we proceeded to investigate thepossibility of reduction of OKL38 protein in the tumors. West-ern blot analysis was performed using rabbit anti-humanOKL38 antibody on nine paired normal/tumor kidney tissues.The results showed that OKL38 protein was undetected in 78%(7/9) of the tumor kidney tissues compared with their adjacentnormal tissues. The endogenous 61, 52, and 38 kDa isoforms ofhuman OKL38 protein were also detected by Western blot

FIG. 4. Splicing of various RNAtranscripts of the human OKL38 gene.A, schematic diagram showing the variousOKL38 mRNA isoforms. HuOKL38-1a, -2a,and -2b were cloned via PCR using primers1F, 2F, and 1R, while HuOKL38-2c was ver-ified via one-step RT-PCR using primerPCR-2F and PCR-1R. Primer 1F wasalso used for specific amplification ofHuOKL38-1a in one-step RT-PCR. Ar-rows indicate the translational start andtermination sites. The dotted lines indi-cate cDNA region has not been cloned inthis study. B, putative protein isoformsderived from the mRNA mentionedabove. The splicing events have gener-ated three different protein isoforms andregions of protein encoded by differentexons, indicated by different shading.

Genomic Structure of OKL38 Gene and Kidney Cancer748

by guest on Novem

ber 21, 2020http://w

ww

.jbc.org/D

ownloaded from

analysis, which is in agreement with the TNT results. Lowlevels of OKL38 protein were detected in the A498 kidneycancer cell line, which is contrary to the RT-PCR results (Fig.7). The 61 and 38 kDa proteins were differentially down-regu-lated in the kidney tumor tissues (Fig. 9A). When the OKL38antibody was preadsorbed with 50� excess OKL38 antigenpeptide, immunoreactive signal was drastically reduced or lost(Fig. 9, C and G)

Immunohistochemical study was performed on 22 pairednormal/tumor kidney tissues using the same OKL38 antibod-ies. OKL38 protein was undetected in 64% (14/22) of the kidneytumors examined, which supported the observation in Westernblot analysis. All normal kidney epithelial cells were positivefor OKL38 protein, except the Bownman capsules (arrow) andthe stroma (Fig. 9D), while no immunostaining of the tumorcell was observed (Fig. 9E).

Overexpression of OKL38 Protein Was Lethal to A498 Cells—Because the evidence suggests that OKL38 functions as a tu-mor suppressor protein, we proceeded to investigate its growthinhibitory function. Initially, the full-length cDNAHuOKL38-1a was cloned into pcDNA3.0 and transfected intoA498 cells to generate stable cell lines. However, the trans-fected cells failed to propagate and eventually disintegrated(data not shown). Thus, we could not characterize these trans-fected cells. To study the function of the OKL38 protein, wegenerated an OKL38-eGFP-pcDNA3.0 construct, which fusedthe reporter gene eGFP to the C-terminal of OKL38 protein.Our preliminary transfection study shows that OKL38-eGFPrecombinant protein formed an aggregation in the A498 cells asearly as 24 h post-transfection (Fig. 10). None of the cellsexpressing the recombinant protein survived 96 h post-trans-fection compared with the eGFP-positive control cells. Theresults suggest that over-expression of OKL38 protein waslethal to A498 cells.

DISCUSSION

Our previous study showed that OKL38 was involved intumorigenesis of the breast (4). To understand the functionsand regulation of OKL38 gene, we cloned, sequenced, and char-acterized the human OKL38 gene. The human OKL38 genereported in this study was cloned and sequenced before thehuman genome sequence was published. Our sequence datashowed two regions of mismatch, and we have verified that ourclone contained a 550-bp longer CCCT repeat compared withthe published human sequence. We presume that the 300-bpunsequenced region of these repeats also contains CCCT re-peats. The functional importance of the CCCT repeats in thehuman OKL38 is unknown at the present time. However, theinstability of CCCT repeats is highlighted in several reports,suggesting that the different repeat lengths are correlated withDNA instability (12, 13). Fiskerstrand et al. (13) have alsoshown that the specific region in intron 2 of the rat prepro-tachykinin-A gene contained 128 bp of the CCCT tandem re-peat domain. This domain is able to support reporter geneexpression in mouse embryonic stem (ES) cells that have beeninduced to differentiate, but not in undifferentiated ES cells,suggesting that this region is a highly restrictive enhancer andmay be associated with differentiation (13). Similarly, CCCTrepeats located in intron 2 of the OKL38 gene may associatewith differentiation, as we have postulated in our previousreport (4). Since the CCCT repeats are located 1.2-kb upstreamof the promoter P1 initiation site, it is likely that this regionfunctions as an enhancer for the P1 promoter (Fig. 2A). Exper-iments are underway to examine this possibility.

An �500-bp dinucleotide CA repeat is localized to intron 1 ofthe OKL38 gene (Fig. 2). CA repeats are widespread throughoutthe human genome, representing �0.25% of the human genome.They are the most common dinucleotide polymorphism amongthe microsatellite DNA fraction (14). Human genes with poly-morphic intronic CA repeats include the human interferon-�gene (15), the epidermal growth factor receptor gene (16), and thecystic fibrosis cellular stress response (CFTR) gene (17). A�20-bp random deletion in the CA repeat region is found in oursequenced OKL38 gene suggesting polymorphism in this region.Hui et al. (18) reported a recent discovery that dinucleotide CArepeats of variable length found in intron 13 of the human eNOSgene promote intron removal. They have shown that the ribonu-cleoprotein (hnRNP) L, a member of the hnRNP family, is themajor factor that binds specifically and in a length-dependentmanner to the CA repeat enhancer. Although the polymorphicnature of this region in the human OKL38 gene has not beenestablished, it is possible that the polymorphic intronic CA re-peats in intron 1 of human OKL38 gene may serve as dockingsites for the hnRNP L factor to bind and regulate differentialsplicing of this gene. This is also supported by the fact thatalternative splicing occurs at the 5�-end of the OKL38 gene,where the CA repeats are present.

The genomic organization of the human OKL38 gene and itssplicing pattern have been established in this study. The splic-ing pattern was very similar to that of the human and mousethioredoxin reductase 1 (TXNRD1) gene (19). Differential pro-moter usage and alternative splicing at the 5�-region of theTXNRD1 are highly similar to that of the OKL38 gene, exceptthat the gene contains 16 exons compared with the OKL38gene, which consists of 8 exons. The similarity in splicingpattern of the TXNRD1 and OKL38 genes suggests that thetwo genes may be regulated in a similar fashion, and they maybelong to the same family. This postulation is further sup-ported by the presence of a pyridine nucleotide-disulfide oxi-doreductase (Pyr-redox) domain found in both genes.

Regulation of the OKL38 gene at promoter P1 could involve

FIG. 5. Transcription and translation analysis of OKL38cDNAs. Radiolabeled OKL38 proteins produced from cloned cDNAsvariants using the TNT system as described under “Experimental Pro-cedures.” The synthesized proteins were separated on a PAGE gel andtransferred onto a nitrocellulose membrane. Blots were exposed to thefilms, and the radiolabeled OKL38 proteins are shown in A. The sameblots were blotted with rabbit anti-OKL38 antibodies and the expectedsize of OKL38 protein is shown (B).

Genomic Structure of OKL38 Gene and Kidney Cancer 749

by guest on Novem

ber 21, 2020http://w

ww

.jbc.org/D

ownloaded from

FIG. 6. Tissue distribution of human OKL38 transcripts. A, MTE array was probed with a 700-bp human OKL38 SmaI-radiolabeled probeas described under “Experimental Procedures.” The identity of each cDNA dot is represented in the grid (B). C, MTN blot analysis was performedusing the same probe, which recognizes all the OKL38 variants. The tissues are: He, heart; Br, brain; Li, liver; Pa, pancreas; Pl, placenta; Lu, lung;Mu, muscle; Ut, uterus; Bl, bladder; Ki, kidney; Sp, spleen; Ce, cervix; Ov, ovary; Te, testis; Pr, prostate. The amount in each lane is normalizedwith GAPDH (D). The relative level of OKL38 expression after GAPDH normalization is shown in E.

Genomic Structure of OKL38 Gene and Kidney Cancer750

by guest on Novem

ber 21, 2020http://w

ww

.jbc.org/D

ownloaded from

several factors. A CpG island of 215 bp, located 746 bp precedesthis TATA-less promoter upstream of promoter P1 initiationsite. CpG islands are frequently associated with 5�-regulatorysequences of housekeeping genes but also of genes with a moretissue-restricted pattern of expression (20). Putative Initiator(Inr) with the consensus Py-Py-A�1-N-T/A-Py-Py where A�1 isthe transcriptional start site (21) is identified in promoter P1suggesting that this core sequence is likely responsible fordriving transcription at this site. Additional support for thisconclusion is the presence of a putative DPE that is found inabout 20% of the TATA-less promoters containing Inr (21).Transcription factor binding sites identified within promoterP1 are slightly different than those observed in the promoter

P2, suggesting that they may be differentially regulated. Inter-estingly, no typical promoter elements are identified in theputative promoter P2 suggesting that other novel sequencesmay regulate this promoter. Further characterization of thesepromoters is needed to establish their regulatory functions.

In this study, we have cloned the full-length of three OKL38variants. The HuOKL38-1a and HuOKL38-2a mRNA isoforms,which are expressed from different promoters; therefore, theypossess different 5�-UTR (Fig. 4). These two mRNA isoformsencode for the same ORF (477 amino acids) verified by TNTanalysis (Fig. 5). The lethal effect of this OKL38 protein iso-form suggests that this protein has to be regulated in a preciseand controlled manner. Therefore, the use of 2 different pro-

FIG. 7. Differential expression of human OKL38 variants in normal/tumor tissues and cell lines. A, the specificity of each pair ofprimers was verified by PCR using the cloned OKL38 cDNAs. PCR was performed on 3 of the 4 variants of OKL38 cDNAs using a combinationof 3 variant-specific primers illustrated in Fig. 3. Lanes 1 and 4, HuOKL38-2b cDNA; lanes 2 and 5, HuOKL38-2a cDNA; lanes 3 and 6,HuOKL38-1a cDNA; lane 7, H2O (negative control). Reactions in lanes 1–3 contained both PCR-2F and PCR-1R primers, while lanes 4–6 containedthe 1F and PCR-1R primers (A). Note that primers were specific for detection of all three variants. One step RT-PCR was performed using totalRNA extracted from: Li, liver; Ki, kidney; Ov, ovary and cell line; A498, kidney; HepG2, liver; SL, a subline of HepG2; NIH, ovarian (B–E). FourRT-PCR products of 443, 373, 533, and 482 bp corresponding to the HuOKL38-1a, -2a, -2b, and -2c transcripts, respectively. RT-PCR was performedusing primers, 1F and PCR-1R (B), primers, PCR-2F and PCR-1R (C), and a pair of tubulin primers (E) as internal control. The amplified productswere separated on a 2.0% agarose gel and transferred onto nylon membrane. Southern blot analysis was performed using probe derived from exons4, 5, and 6 (D).

Genomic Structure of OKL38 Gene and Kidney Cancer 751

by guest on Novem

ber 21, 2020http://w

ww

.jbc.org/D

ownloaded from

moters to produce 2 mRNA isoforms coding for the same pro-tein is far from redundant. The intrinsic regulation of OKL38protein may be regulated at the translational level with possi-ble regulatory elements in the 5�-UTR.

Sequence analysis of the cloned OKL38 variants show thatall the 1st AUG on the transcripts conform to the Kozak se-quence (22), except HuOKL38-2b and -2c. We have noticed inthis study and the previous study (4) that there are severalOKL38 protein isoforms (61, 52, and a doublet 38 kDa) detect-able by Western blot analysis (Figs. 5 and 9). The 61-kDaOKL38 isoform may have been expressed by HuOKL38-2b or-2c transcripts, while the 52 kDa protein is most probablyderived from HuOKL38-1a or 2a transcripts. The 38-kDa dou-blet may have derived from usage of an internal translationstart site. An internal AUG translation initiation codon at 139(nucleotide positions 574–576 of HuOKL38-1a), which con-forms to the Kozak sequence, was identified and verified byTNT. However, the post-translational process could also beresponsible for OKL38 protein isoforms observed in the presentstudy. The various isoforms of OKL38 may serve differentfunctions in cells. We are currently determining the functionsof these isoforms.

The Pry-redox family of proteins consists of both class I andclass II oxidoreductase, as well as NADH oxidases and peroxi-dases. This family of proteins includes the well-characterizedglutathione reductase and thioredoxin reductase proteins,which carry this domain in addition to a pyridine dimerizationdomain (23). It has been shown that the thioredoxin (Trx) andglutathione (GSH) system are involved in a variety of redox-dependent processes such as DNA synthesis, antioxidant de-fense, and regulation of cellular redox state (24–26). In addi-tion, the Trx and GSH systems regulate the activities ofvarious transcription factors, kinases and phosphatases, andthey are implicated in the redox control of cell growth anddeath, transcription, cell signaling, and other processes (27,28). The cell death-associated function of OKL38 demonstratedin this study and the presence of a pyr-redox domain in OKL38

protein suggest that the gene may belong to the larger family ofpyridine nucleotide-disulfide oxidoreductases and may sharesimilar functions to this family of proteins.

In our previous study, we have shown that OKL38 mRNAtranscripts are down-regulated in several breast cancer celllines. In this study, we have investigated the expression ofOKL38 transcripts and protein in kidney cancer. Our resultsshow that OKL38 is down-regulated in the kidney tumor atboth the mRNA and protein levels. The down-regulation ofOKL38 expression may possibly be caused by point mutation inthe ORF or region of regulatory elements. The silencing ofOKL38 expression could also be due to hypermethylation at theCpG islands in the promoter region, which have been known tosilence tumor suppressor genes (29). The OKL38 gene is local-ized to chromosome 16q23.3, which is susceptible to LOH in avariety of tumors (7–10). This mechanism may also be thecause of OKL38 down-regulation. The results suggest that theOKL38 protein may play a growth inhibitory or differentiationrole in normal kidney epithelial cells.

The OKL38 cDNA HuOKL38-1a cloned in this study is lon-ger and encoded a longer protein of 52 kDa compared with thepreviously cloned cDNA (4). Overexpression of this proteincauses growth inhibition and also is lethal to the A498 cells.The observation is similar to that of the p53 protein, as abnor-

FIG. 8. Cancer profiling array of human OKL38 in normal/tumor kidney. The CPA was probed with a 700-bp human OKL38SmaI probe as described under “Experimental Procedures.” Note thatthe OKL38 transcript was generally down-regulated in kidney tumorsamples (A). The relative density of OKL38 expression in tumor andadjacent normal tissues was compared and expressed as a percentage oftotal sample examined (B). The legend shown on the right indicatesup-regulation, down-regulation, and no observable change.

FIG. 9. Expression of OKL38 protein in kidney tumor and ad-jacent normal tissue. A, Western blot analysis showing the expres-sion of OKL38 proteins in A498 and paired normal (N)/tumor (T) kidneytissues. The loading amount was normalized to the level of �-tubulinshown in B. Note that three OKL38 protein isoforms with molecularweights of 61, 52, and 38 were detected. Immunohistochemical per-formed on normal (D) and kidney tumor (E) tissues showed differentialstaining indicating loss of OKL38 protein in tumor tissue. Note that theBownman capsules (arrow), the stroma, and the tumor cells were notstained. The specificity of anti-OKL38 antibody is shown in C withWestern analysis and with immunohistochemistry (F and G). The ni-trocellulose membrane and slide with liver total protein and kidney,respectively, were blotted with OKL38 antibody (C and F) and OKL38antibody preadsorbed with 50� excess antigen peptide (C (�) and G).Note that signal was reduced to background levels when preadsorbedantibody was used indicating that the antibody for OKL38 was specific.

Genomic Structure of OKL38 Gene and Kidney Cancer752

by guest on Novem

ber 21, 2020http://w

ww

.jbc.org/D

ownloaded from

mal regulation of this gene causes the cells to become cancerousor lethal to its own survival (30). These results supported ourprevious postulation that OKL38 functions as a tumor suppres-sor gene and that this protein may regulate cell homeostasisthrough cell death. We are currently investigating the mecha-nisms of OKL38-induced cell death.

In conclusion, we have cloned, sequenced, and revealed thegenomic organization of the human OKL38 gene. Several impor-tant regions that may play important roles in the regulation ofthe OKL38 gene are identified. In addition, we have cloned threeof the novel OKL38 cDNAs and partially characterized them.Our present data show that OKL38 protein is down-regulated ina high proportion of kidney tumors, suggesting its potential useas a diagnostic or prognostic markers for kidney cancer. Wepostulate that OKL38 regulates the differentiation and prolifer-ation of normal cells through the regulation of cell death. Loss ofOKL38 protein disturbs the balance between cell growth, differ-entiation, and cell death in normal tissue, resulting in uncon-

trolled growth and formation of tumors. The present study pro-vides the basic foundation for future studies on the role of OKL38in differentiation, growth, and tumorigenesis.

REFERENCES

1. Belldegrun, A. and deKernion, J. B. (1998) Campbell’s Urology, Vol. 3, pp.2283–2326, W. B. Saunders Co., New York

2. Vogelzang, N. J., and Stadler, W. M. (1998) Lancet 352, 1691–16963. Mejean, A., Oudard, S., and Thiounn, N. (2003) J. Urol. 169, 821–8274. Huynh, H., Ng, C. Y., Ong, C. K., Lim, K. B., and Chan, T. W. (2001) Endo-

crinology 142, 3607–36155. Huynh, H., Chow, P. K., Ooi, L. L., and Soo, K. C. (2002) Cell Growth Differ. 13,

115–1226. Nielsen, H., and Krogh, A. (1998) Prediction of Signal Peptides and Signal

Anchors by a Hidden Markov Model. Proceedings of the Sixth InternationalConference on Intelligent Systemsfor Molecular Biology (ISMB 6)TMpred,AAAI Press, Menlo Park

7. Latil, A., Cussenot, O., Fournier, G., Driouch, K., and Lidereau, R. (1997)Cancer Res. 57, 1058–1062

8. Sato, M., Mori, Y., Sakurada, A., Fukushige, S., Ishikawa, Y., Tsuchiya, E.,Saito, Y., Nukiwa, T., Fujimura, S., and Horii, A. (1998) Genes Chrom.Cancer 22, 1–8

9. Kawakami, M., Staub, J., Cliby, W., Hartmann, L., Smith, D. I., and Shridhar,

FIG. 10. Transfection study of OKL38-eGFP recombinant. A498 cells were transfected with controlled eGFP-pcDNA3.0 (A, C, and E) andthe OKL38-eGFP-pcDNA3.0 recombinant construct (B, D, and F), as described under “Experimental Procedures.” The transfected cells wereharvested at 24 h (A and B), 48 h (C and D), and 72 h (E and F) post-transfection. Cells expressing OKL38-eGFP protein were visualized usingmicroscope equipped with epifluorescence optic. Magnification �800.

Genomic Structure of OKL38 Gene and Kidney Cancer 753

by guest on Novem

ber 21, 2020http://w

ww

.jbc.org/D

ownloaded from

V. (1999) Int. J. Oncol. 15, 715–72010. Sheu, J. C., Lin, Y. W., Chou, H. C., Huang, G. T., Lee, H. S., Lin, Y. H., Huang,

S. Y., Chen, C. H., Wang, J. T., Lee, P. H., Lin, J. T., Lu, F. J., and Chen,D. S. (1999) Br. J. Cancer 80, 468–476

11. Javahery, R., Khachi, A., Lo, K., Zenzie-Gregory, B., and Smale, S. T. (1994)Mol. Cell. Biol. 14, 116–127

12. Kosteas, T., Palena, A., and Anagnou, N. P. (1997) Hum. Genet. 100, 441–44513. Fiskerstrand, C. E., Lovejoy, E., Gerrard, L., and Quinn, J. P. (1999) Neurosci.

Lett. 263, 141–14414. International Human Genome Sequencing Consortium (2001) Nature 409,

860–92115. Awad, M., Pravica, V., Perrey, C., El Gamel, A., Yonan, N., Sinnott, P. J., and

Hutchinson, I. V. (1999) Hum. Immunol. 60, 343–34616. Han, H. J., and Nakamura, Y. (1998) J. Hum. Genet. 43, 212–21317. Mateu, E., Calafell, F., Bonne-Tamir, B., Kidd, J. R., Casals, T., Kidd, K. K.,

and Bertranpetit, J. (1999) Hum. Hered. 49, 15–2018. Hui, J., Stangl, K., Lane, W. S., and Bindereif, A. (2003) Nat. Struct. Biol. 10,

33–3719. Osborne, S. A., and Tonissen, K. F. (2001) BMC. Genomics 2, 10

20. Gardiner-Garden, M., and Frommer, M. (1987) J. Mol. Biol. 196, 261–28221. Burke, T. W., and Kadonaga, J. T. (1997) Genes Dev. 11, 3020–303122. Kozak, M. (2002) Mamm. Genome 13, 401–41023. Williams, Jr. C. H. (1992) in Chemistry and Biochemistry of Flavoenzymes

(Muller, F., ed) pp. 121–211, CRC Press, Boca Raton24. Holmgren, A., Arner, E. S., Aslund, F., Bjornstedt, M., Zhong, L., Ljung, J.,

Nakamura, H., and Nikitovic, D. (1998) in Oxidative Stress, Cancer, AIDS,and Neurodegenerative Diseases (Montagnier, L., Oliver, R., and Pasquier,C., eds) pp. 229–246, Dekker, New York

25. Halliwell, B. (1999) Free Radic. Res. 31, 261–27226. Sies, H. (1999) Free Radic. Biol. Med 27, 916–92127. Rhee, S. G. (1999) Exp. Mol. Med 31, 53–5928. Finkel, T. (2000) FEBS Lett. 476, 52–5429. Toyooka, K. O., Toyooka, S., Virmani, A. K., Sathyanarayana, U. G., Euhus,

D. M., Gilcrease, M., Minna, J. D., and Gazdar, A. F. (2001) Cancer Res. 61,4556–4560

30. Smith, N. D., Rubenstein, J. N., Eggener, S. E., and Kozlowski, J. M. (2003)J. Urol. 169, 1219–1228

Genomic Structure of OKL38 Gene and Kidney Cancer754

by guest on Novem

ber 21, 2020http://w

ww

.jbc.org/D

ownloaded from

Hoon Tan and Hung HuynhChoon Kiat Ong, Chuan Young Ng, Caine Leong, Chee Pang Ng, Keong Tatt Foo, Puay

Kidney Carcinogenesis Gene and Its Differential Expression inOKL38Genomic Structure of Human

doi: 10.1074/jbc.M308668200 originally published online October 21, 20032004, 279:743-754.J. Biol. Chem. 

  10.1074/jbc.M308668200Access the most updated version of this article at doi:

 Alerts:

  When a correction for this article is posted• 

When this article is cited• 

to choose from all of JBC's e-mail alertsClick here

  http://www.jbc.org/content/279/1/743.full.html#ref-list-1

This article cites 27 references, 5 of which can be accessed free at

by guest on Novem

ber 21, 2020http://w

ww

.jbc.org/D

ownloaded from