Leukemia (2000) 14, 2128–2134  2000 Macmillan Publishers Ltd All rights reserved 0887-6924/00 $15.00 www.nature.com/leu Human mortalin (HSPA9): a candidate for the myeloid leukemia tumor suppressor on 5q31 H Xie1,ZHu1, B Chyna1, SK Horrigan2 and CA Westbrook1

1Section of Hematology/Oncology, Department of Medicine, University of Illinois at Chicago, Chicago, IL; and 2Department of Pediatrics, Georgetown University, Washington DC, USA

Human mortalin (HSPA9) was originally identified by its close cDNAs for both the pancytosolic form (mot-1) and the peri- homology to murine mortalins, which play important roles in nuclear form (mot-2) of murine mortalin have been fully cellular senescence. The two murine , mot-1 and mot-2, cloned, and the genomic organization of mot-2 has been differ in only two amino acid residues, but have opposite func- 1,3 tions in cellular immortalization. HSPA9 was recently localized reported. Recent evidence suggests that the two mortalins to 5, band q31, a region that is frequently deleted are not isoforms, but instead originate from two distinct gen- in myeloid leukemias and myelodysplasia (MDS), making it a omic loci. Fluorescence in situ hybridization assigned the candidate tumor suppressor gene, which is consistent with the mortalin-related genes to mouse 18 and X.6 The biological function of its murine homologue. To evaluate mor- two forms of mortalin are variably present in different strains talin in this capacity, its expression in normal and leukemic cell lines was investigated, and its genomic structure was deter- of mouse; thus, both mot-1 and mot-2 are present in the mined in order to facilitate mutation detection. RT-PCR and Balb/c strain, while Swiss and C56BL/6 mice contain only Northern blot analysis revealed a broad distribution in normal mot-2, and C3H He mice contain only mot-1. It has been tissues and in leukemia cell lines, producing a single 2.8 kb suggested that this distribution of mortalin genes may explain transcript. Genomic characterization showed that the gene the different rates of spontaneous immortalization of cells spans 18 kb, and consisted of 17 exons with boundaries that found in different strains of mouse.5 were almost identical to its murine counterpart. Using intron- based primers to flank each exon, sequence of the complete Unlike murine mortalin, there appears to be only one -coding regions was obtained for three AML cell lines, human mortalin gene, HSPA9, which is highly homologous including two lines with loss (KG-1 and HL-60) to both mot-1 and mot-2, differing from them in amino acid and one without (AML-193) compared to normal DNA. No sequence by only 6 and 5 residues, respectively. HSPA9 has mutations were identified although one conservative nucleo- been assigned to chromosome 5 band q31.1.6 Recently, we tide sequence variant was observed in exon 16. We have shown refined the location of this gene to a 700-kb interval, which that mortalin is highly conserved in genomic structure as well we postulate to be the site of a tumor suppressor gene impli- as sequence, and the designed primers will be suitable for − future studies to detect mutations in clinical samples. Leukemia cated in myeloid malignancies manifesting 5 or del(5q), (2000) 14, 2128–2134. including acute myeloid leukemia (AML) and myelodysplasia Keywords: mortalin; HSPA9; genomic structure; mutation; tumor (MDS).7 The chromosomal location of the mortalin gene, its suppressor biologic function in murine cells, and its association with cellular immortalization, make it an appealing candidate for a tumor suppressor gene whose mutational loss or inactivation Introduction results in transformation of myeloid cells. Here, we report the genomic structure of human mortalin, which we found to be Human mortalin (also called peptide-binding protein 74 highly similar in organization to murine mot-2. Additionally, (PBP74), and heat shock 70kD protein 9 (HSPA9)) is an inter- we report intron-based primers, suitable for genomic sequen- esting gene that shows a high degree of nucleotide sequence cing, and their use to evaluate the mortalin coding to its murine counterparts mortalin-1 (mot-1) and in several myeloid leukemia cell lines with −5 or del(5q). mortalin-2 (mot-2). The mouse homologues were initially identified by their roles in cellular senescence, where they appear to have contrasting functions. In non-immortalized murine fibroblasts, mortalin displays a uniform pancytosolic distribution, whereas it manifests a perinuclear location in Materials and methods immortalized fibroblasts.1,2 Cytosolic mot-1 has been pro- posed to be a major determinant of fibroblast senescence, such that its transfection into NIH3T3 cells induces cellular Cell culture and bacterial clones senescence, whereas microinjection of anti-mortalin antibody can stimulate proliferation in senescent murine fibroblasts.3 On the other hand, over-expression of mot-2 can induce Cell lines used in this study were obtained from ATCC malignant transformation of NIH 3T3 cells.4 The cytosolic and (Rockville, MD, USA). Cells were cultured using rec- perinuclear forms of mortalin were later shown to arise from ommended conditions, or in RPMI-1640 medium (GIBCO the two related genes, mot-1 and mot-2, which differ from BRL, Grand Island, NY, USA) supplemented with 10% heat- each other by only two amino acid residues, respectively.5 inactivated fetal calf serum (Sigma, St Louis, MO, USA) at ° + 37 C in an atmosphere of 5% CO2 in air. CD34 cells were isolated from bone marrow obtained from normal donors with informed consent under the guidelines provided by the Uni- Correspondence: CA Westbrook, University of Illinois at Chicago, + Departments of Medicine and Genetics, M/C 734, 900 S Ashland, versity of Illinois. The CD34 cells were isolated and purified 8 Chicago, Illinois 60607–7170, USA; Fax: 312–413–7963 as previously described. BAC clone 15L17 was obtained from Received 22 May 2000; accepted 1 August 2000 Research Genetics (Huntsville, AL, USA). Mortalin as a candidate for tumor suppressor gene on 5q31 H Xie et al 2129 Preparation of DNA and RNA hybridized with a ␤-actin probe as an internal control of both the RNA integrity and amount. A multiple tissue Northern Genomic DNA and total cellular RNA were extracted using blot, MTN H, (Clontech, Palo Alto, CA, USA), which con- TRIZol Reagent (Gibco Technologies, Gaithersburg, MD, tained mRNAs from various human tissues including heart, USA) according to the recommended protocol from 107 brain, placenta, lung, liver, skeletal muscle, kidney and pan- CD34+ cells, or from cell lines that were harvested in the log- creas was used to detect expression of mortalin. A hybridiz- arithmic growth phase. BAC DNA was prepared according to ation probe was prepared from a 649-bp fragment of human standard protocols used in this laboratory.9 mortalin containing sequences from exon 9 to exon 11 and amplified by RT-PCR from human leukemia cell line KG-1. The pair of primers used consisted of the forward primer for Reverse-transcriptase PCR (RT-PCR) exon 9 (5Ј-CTGCTGAAAAGGCTAAATGTGAACTCTC-3Ј) and the reverse primer for exon 11 (5Ј-CACAGCCTCATCAG- RT-PCR was performed using either the Titan One Tube RT- GATTGAGAGCTT-3Ј). The RT-PCR product was labelled with PCR System (Boehringer Mannheim, Indianapolis, IN, USA) ␣-32P-dCTP by random primer labelling using Multiprime or Enhanced Avian RT-PCR kit (Sigma, St Louis, MO, USA) DNA Labeling System(Amersham Pharmacia Biotech, Piscata- according to the manufacturer’s protocol. Optimal Mg2+ con- way, NJ, USA). centrations and annealing temperatures were determined for each primer pair. All reactions were performed in duplicate, including control reactions which did not include reverse transcriptase or RNA. PCR reactions were performed in a Per- kin-Elmer (Foster City, CA, USA) 480 Thermocycler using the Results following parameters: initial denaturing at 94°C for 3 min, fol- lowed by 30 cycles of denaturation at 94°C for 40 s, annealing at 52–60°C, (optimized for each primer pair) for 1 min, and Genomic organization of human mortalin extension at 72°C for 1 min, with a final extension at 72°C for 7 min. Each 20 ␮␭ of PCR reaction contains 100 ng of template DNA, 10 mM Tris HCl, pH 8.0, 2.0 mM MgCl, 200 The intron–exon boundaries of human mortalin were pre- ␮M each of dATP, dTTP, dCTP and dGTP, 15 pmol of each dicted by alignment of the human cDNA sequence (GenBank primer and 0.5 units of Taq DNA polymerase (Bioline, IL, Accession No. L15189 ) with the published intron–exon USA). PCR products were separated by agarose gel electro- boundary sequences of murine mot-2 gene10 using Seqman phoresis and visualized with ethidium bromide, using a 100 software. Exon-based primers at the presumed intron–exon bp or 1 kb DNA ladder (GIBCO BRL) for size estimation. boundaries were then designed to amplify intron sequences Long-range PCR was performed using the Expand 20kbPlus between the exons (Table 1) using DNA from BAC clone PCR System (Boehringer Mannheim, Indianapolis, IN, USA) 15L17 as a template. This BAC clone was identified as con- according to the manufacturer’s instructions. taining the genomic sequence for mortalin by BAC typing in our laboratory (data not shown). In those instances in which the predicted PCR product appeared to be longer than 3.0 kb DNA sequencing and failed to amplify, such as intron 9 (5.0 kb), long-range PCR was attempted using the Expand 20 kbPlus PCR System The DNA sequencing reaction was performed using the ABI according to the manufacturer’s instruction (Boehringer Prism BigDye Terminator Sequencing kit (Perkin-Elmer). In Mannheim). The PCR products obtained from this exon-to- brief, PCR products amplified from each exon were purified exon amplification were separated by agarose gel electro- using Qiagen PCR purification kit (Qiagen, Valencia, CA, phoresis and the size of each intron was estimated by compar- USA). A typical 20 ␮l reaction containing 20–90 ng DNA tem- ing with a 1 kb DNA ladder (GIBCO BRL). These amplified plate from purified PCR products, 0.32 ␮M of forward or intron-specific sequences were then directly sequenced in reverse primer, and 8 ␮l of BigDye terminate mixture was both directions, and the exact positions of exon–intron bound- reamplified as follows: 24 cycles at 96°C for 30 s, 50°C for aries were identified by comparing the genomic sequences 15 s, and 60°C for 4 min. The reactions were purified and amplified from the BAC clone with human mortalin cDNA then sequenced using an ABI prism 377 DNA sequencer sequence. All boundary sequences contain consensus GT/AG (Perkin-Elmer). Analysis of the sequences was performed using sequences at donor and acceptor sites; these exon–intron Seqman software (DNASTAR, Madison, WI, USA). boundaries are given in Table 2. The genomic organization of the human mortalin is summa- rized and compared to murine mot-2 in Figure 1; there is no Northern blot analysis information available on the genomic organization of mot-1. The human and murine genes are remarkably similar in Ten micrograms of total RNA from each cell line were separ- organization, differing primarily in intron size. In the mouse, ated on 1.0% formaldehyde agarose gels and transferred to mot-2 is 16.5 kb in length and consists of 17 exons separated Hybond-N nylon filters (Amersham, Arlington Heights, IL, by 16 introns.5 The human mortalin gene also contains 17 USA), which were subsequently cross-linked with ultraviolet exons and 16 introns and spans over 18 kb; the intron light. After prehybridization at 65°C for 2 h, the blot was between exons 9 and 10 is the longest, being approximately hybridized to ␣-32P-dCTP-labelled cDNA probes for 18 h at 5 kb in length. All other introns are between 0.2 kb and 2.5 65°C, and washed with 2 × SSC + 0.1% SDS twice, followed kb in size. The intron–exon boundaries are almost identical by 1 × SSC + 0.1% SDS twice, and 0.1 × SSC + 0.1% SDS between the human and mouse genes and all boundary twice. The blots were exposed to X-ray films with an inten- sequences contain consensus GT/AG sequences at donor and sifying screen at −80°C. The same filters were subsequently acceptor sites (Figure 1 and Table 2).

Leukemia Mortalin as a candidate for tumor suppressor gene on 5q31 H Xie et al 2130 Table 1 Exon-based PCR primers to amplify introns

Primer Primer Sequence Amplimer Annealing name orientation 5Ј → 3Ј length (bp)a temperature (°C)

exon.1.for Forward CCATGATAAGTGCCAGCCGAGC 1250 57 exon2.rev Reverse CCTTGAAACAAGTCTAAAAGCCTCATGA exon2.for Forward TCATGAGGCTTTTAGACTTGTTTCAA 200 55 exon3.rev Reverse CCCAAATCAATACCAACAACTGCTC exon3.for Forward CCTGCGTGGCAGTTATGGAAGGTA 2350 58 exon4.rev Reverse CAACAAGTCGCTCACCATCTGCTGTAA exon4.for Forward TTTATGCTACCAAGCGTCTCATTGG 1200 57 exon5.rev Reverse AATACAATTTCCCATGAGCCTCAACCCC exon5.rev Forward ATGGGAAATTGTATTCTCCGAGTCAG 600 57 exon6.rev Reverse ATTTTTTGCTGTGTGCCCCAAGT exon6.for Forward GGGCACACAGCAAAAAATGC 1400 57 exon7.rev Reverse CACCCGAAGCACATTCAGTCCAGATAT exon7.for Forward GCTGCTGCTCTTGCCTATGGTCTA 400 55 exon8.rev Reverse TCACCTCAAATACTCCTTTCTGAATTTCCAGG exon8.for Forward CTTAGGTGGGGAAGACTTTGACCA 400 57 exon8.rev Reverse CACATTTAGCCTTTTCAGCAGCTTC exon9.for Forward CTGCTGAAAAGGCTAAATGTGAACTCTC 5000 57 exon10.rev Reverse GCACGGGTCAACTTCATATTCAAATGCT exon10.for Forward TCAGCAAGAGTGACATAGGAGAAGTGA 650 55 exon11.rev Reverse CACAGCCTCATCAGGATTGACAGCTT exon11.for Forward CCTGTCTCTGGGTATTGAAACTCTAGGAG 1250 58 exon12.rev Reverse CCTGACACACTTTAATTTCCACTTGCGT exon12.for Forward GAAAGAGAGATGGCTGGAGACAAC 550 57 exon13.rev Reverse TTGGCATCAATGTCAAATGTAACTTCAATCTGA exon13.for Forward GCCAATGGGATAGTACATGTTTCCTG 400 57 exon14.rev Reverse TTCAATATCATCTTTGCTTAATCCACCAGAAG exon15.for Forward ACGACACAGAAACCAAGATGGAAG 200 50 exon16.rev Reverse CCTCCTGTTTCGCTGTCTTTTCTAG exon16.for Forward CTCTTCAGCAGGCATCATTGAAG 550 50 exon17a.rev Reverse TTTGATCTTCCTTTTGTTCCCCAGTG

aSize based on estimation from PCR/agarose gel electrophoresis.

Expression patient with chromosome 5q deletion.12 MO7E is a megakar- yoblastic leukemia cell line representing a lineage that is often Human mortalin is known to encode an mRNA of 2.8 kb.11 pathologically involved in patients with MDS/AML. Among As expected, Northern blot analysis confirmed that human the 13 leukemia cell lines, five lines have chromosome 5q mortalin is widely expressed in multiple tissues (Figure 2a), deletion (KG-1, HL-60, MUTZ-1, RPMI-8226 and WSU-NHL), with the highest expression seen in heart; expression was one line has chromosome 5q addition (CMK), one has deriva- intermediate in skeletal muscle, kidney, liver and brain, and tive chromosome 5 (U-937) and six lines have normal chro- the lowest in lung, placenta and pancreas. Mortalin mosome 5.12–14 No correlation was observed between the expression was also investigated in human myeloid leukemia expression level of mortalin and the cell lineages or subtype cell lines. Using RT-PCR, the presence of mortalin transcript of the leukemia lines. was confirmed in three myeloid lines and in CD34+ cells derived from human bone marrow, which are thought to con- tain the earliest myeloid stem cell precursors (Figure 2b). Mutation analysis of mortalin in leukemia cell lines Northern blot evaluation of RNA derived from a panel of 13 leukemia cell lines, including 11 myeloid lines and one The coding sequence of mortalin was investigated for derived from B cell lymphoma (WSU-NHL) and another from mutations in three AML cell lines, including two lines with myeloma (RPMI-826) shows that mortalin is expressed in all chromosome 5 deletion (KG-1 and HL-60) and one without cell lineages (Figure 2c). Although the expression level varies (AML-193), by sequencing the complete coding regions of all between lines, each shows a single 2.8 kb transcript. AML cell 17 exons in genomic DNA using the intron-based primers line KG-1 and megakayroblastic cell line CMK displayed the given in Table 3. The sequence obtained was compared to highest level of mortalin transcript, while most of the cell lines genomic sequence from BAC DNA. No mutations were have a moderate expression of the message. The lowest detected in any of the cell lines, although a conservative expression was observed in cell lines MUTZ-1 and MO7E. nucleotide sequence variant was identified in exon 16 of the Interestingly, MUTZ-1 is a cell line established from a MDS three lines comparing to BAC DNA. As shown in Figure 3,

Leukemia Mortalin as a candidate for tumor suppressor gene on 5q31 H Xie et al 2131 Table 2 Intron–exon organization of the HSPA9 gene

Exon Positiona Exon sizeb Intron sizec Sequence at intron–exon junctiond (bp) (bp) 5Ј Splice donor 3Ј Splice acceptor

5Ј gtttccagaagcgctgcCGCCACCGCA 1(−72)–81 153 ෂ1230 CCGCCACCAGgtgaggcc ttttttagGATAGCTGGA 2 82–140 59 ෂ120 GGGATTATGCgtaagtat tgacacagATCAGAAGCA 3 141–228 88 ෂ2240 ACAAGCAAAGgtgactta gagtttagGTGCTGGAGA 4 229–411 182 ෂ1065 AGAAAGACATgtgagtaa tcttacagTAAAAATGTT 5 411–535 125 ෂ500 GAGACTGCAGgtgagtgg ttttccagAAAATTACTT 6 536–609 74 ෂ1280 GCAGAGACAGgtgaggtc ttctctagGCCACTAAAG 7 610–717 108 ෂ230 AAGACAAAGTgtaagtttg atttacagCATTGCTGTA 8 718–880 163 ෂ280 GAGAGAAGAGgtgagtta ctttatagACAGGGGTTG 9 881–972 92 ෂ4870 ATCTGTGCAGgtgagaca ctttgcagACTGACATCA 10 973–1182 210 ෂ530 GATGCCCAAGgtatggac atgttcagGTTCAGCAGA 11 1183–1411 228 ෂ1120 GAAGAGCCAGgtaagacc ttccctagGTATTCTCTA 12 1412–1516 105 ෂ440 AGTTTACTTTGgtaggtta tatttacATTGGAATTC 13 1516–1633 118 ෂ550 GAGCAGCAGAgtaagtca tactacagTTGTAATCCA 14 1634–1729 96 ෂ300 GCGAAAGAAGgtcattac tctctaagGAACGAGTTG 15 1730–1821 92 ෂ80 TGCTGATGAGgtgagttc tttttaagTGCAACAAGC 16 1822–1962 141 ෂ390 ATACAAAAAGgtacaagg cttaacagATGGCATGGC 17 1963–2725 763 3Ј TTTAAATAAAtacagaaa aThe nucleotide numbering starts with the ATG initiation codon (A = 1). bExon size was calculated from published cDNA sequence data and our exon–intron splicing sites. cIntron size was estimated from PCR/agarose gel electrophoresis. dExon sequences are represented in uppercase letters and intron sequences are represented in lowercase letters.

Figure 1 Genomic structure of human mortalin and comparison to murine mot-2. The genes are shown approximately to scale. The human and murine genes each contain 17 exons and 16 introns, which are represented as boxes and horizontal lines, respectively. Hatched boxes represent protein-coding sequences and open boxes represent untranslated region. The depiction of mouse mot-2 is based on the previous publication.10 the conservative sequence variant is CTG/TTG at nucleotide chemotherapy, lower remission rate (Ͻ30%), and a short position 1933, resulting in no amino acid changes in the median survival.18,19 It has been proposed that there is a tumor protein sequence. suppressor gene on 5q, whose mutational inactivation contrib- utes to myeloid malignancies. It is hypothesized that this gene, like other tumor suppressor genes which are associated with Discussion chromosomal deletions, will be functionally inactivated in AML cases with del 5q or −5. Thus, we expect in these situ- Deletion of chromosome 5, del(5q), or complete loss of the ations, to find that one allele of this gene is deleted and the entire chromosome, -5, is a recurring cytogenetic abnormality remaining allele mutated. observed in myeloid malignancies. It is observed in about As an initial step in the identification of this presumed 20% of MDS and 10% of AML arising de novo15,16 and up to tumor suppressor gene, we and others have attempted to 40% of AML arising after previous treatment with alkylating define a genomic region that is consistently deleted in cases agents and radiotherapy, namely therapy-related MDS or of del 5q, by comparing the extent of chromosomal loss in AML, or t-MDS/tAML.17 Del(5q) is among the worst prognostic clinical samples.9,20 These studies all suggest that the critical indicators in AML, being characterized by poor response to deletion interval is located at band 5q31. Recently, we further

Leukemia Mortalin as a candidate for tumor suppressor gene on 5q31 H Xie et al 2132 Figure 2 Expression of mortalin in normal and leukemia tissue. (a) Northern blot analysis of human multiple tissues. The human multiple tissue blot (Clontech, Palo Alto, CA, USA) contains mRNA (2 ␮g/lane) from several human tissues including heart, brain, placenta, lung, liver, skeletal muscle, kidney and pancreas, as indicated above each column. The blot was hybridized with a 32P-labelled probe amplified by RT-PCR from BAC clone 15L17 DNA as described in Materials and methods. A single transcript of about 2.8 kb (arrow) is apparent in each tissue, with the highest expression seen in heart, intermediate expression in skeletal muscle, kidney, liver and brain, and the lowest in lung, placenta and pancreas. (b) RT-PCR in CD34+ cells and myeloid leukemia cell lines. RT-PCR was performed on total RNA obtained from normal human bone marrow CD34+ cells and several cell lines using the primers: forward 5Ј-TCAGCAAGAGTGACTTAG- GAGAAGTGA-3Ј, and reverse 5Ј-CACAGCCTCATCAGGATTGA- CAGCTT-3Ј. The RT-PCR product was separated by agarose gel electrophoresis and visualized by ethidium bromide staining. RT-PCR and PCR reaction conditions are described in Materials and methods. The RT-PCR product size is indicated on the right in base pairs. The PCR results are displayed for each sample, with and without the addition of reverse transcriptase, indicated by (+)or(−), respectively. The upper panel shows the results obtained with mortalin. The lower ␤ panel shows the results obtained with 2-microglobulin primers which were used as the internal control for RT-PCR. (c) Northern blot analy- sis of human leukemia cell lines. Each lane contains total RNA (10 ␮g) from different cell lines, as indicated above each column. The blot was prepared as described in Materials and methods, and hybridized to a PCR amplified probe from BAC clone 15L17 DNA. A single transcript of 2.8 kb is observed in all 13 lines, as indicated by the arrow in the left margin. The approximate location of the riboso- mal 28S and 18S bands (as visualized by ethidium bromide staining) are indicated at the left. The same filter was subsequently hybridized with a ␤-actin probe as an internal control of both the RNA integrity and the same amount of RNA loaded, as shown in the bottom panel.

human mortalin in function. It is tempting to speculate that the loss of human mortalin may facilitate transformation of myeloid cells, functioning like mot-1. However, the biology of the human gene has not been as extensively studied as its murine counterpart, and its cellular localization and relation- ship to immortalization have not been clarified. Normal human cells display a pancytosolic distribution similar to that found in normal mouse cells, whereas four kinds of distri- bution patterns have been described in human transformed cells, including perinuclear or juxtanuclear patterns.2 For example, in bladder carcinoma, an overexpression of the non- pancytosolic protein was the predominant pattern, indicating that the expression of this form may play a role in malignant transformation.4 The evidence of elevated levels of mortalin expression in human brain tumors further suggests an involve- ment of nonpancytosolic mortalin in tumorigenesis in vivo.22 However, nothing is known about the expression and function of mortalin in myeloid cells, nor has the gene been examined delineated this minimal interval to the D5S500–D5S594 in any human systems as a possible tumor suppressor gene. region, an interval of about 700 kb, to which we have Thus, to further evaluate mortalin as a candidate tumor sup- localized several candidate genes and ESTs including pressor gene in AML, it will be necessary to examine clinical mortalin.7 samples for expression of the gene, and for mutations within Human mortalin is an appealing candidate for a tumor sup- it. Here we have begun this process by examining expression pressor gene because of both its role in cellular immortaliz- of mortalin in normal human myeloid cells and a panel of ation and its chromosomal localization within the common- human leukemia cell lines, and by designing intron-based deleted interval on 5q31.1 in AML/MDS. In mouse, there are PCR primers which are suitable for amplification and sequen- two closely related forms of mortalin, of which one (mot-1) cing of genomic DNA. Our data showed that mortalin is induces cellular senescence, whereas the other (mot-2) may ubiquitously expressed in human tissues as well as in leuke- facilitate cellular immortalization.3,5 Further evidence sup- mia cell lines, although the level of expression varies con- ports this viewpoint, as mot-2 co-localizes with and siderably. A single 2.8 kb transcript of mortalin was depicted represses p53-mediated transactivation, whereas mot-1 does by Northern blot in all of the 13 leukemia cell lines investi- not.21 Because human mortalin closely resembles both murine gated, suggesting that abnormal splicing or large deletions of mot-1 and mot-2 in amino acid sequence, it is not possible the mRNA are not frequent findings in myeloid leukemias. to predict which of these two isoforms most closely resembles To prepare PCR primers for genomic sequencing of exons,

Leukemia Mortalin as a candidate for tumor suppressor gene on 5q31 H Xie et al 2133

Figure 3 Nucleotide sequence variant detected in exon 16. The nucleotide sequence results obtained for human mortalin, from nucleotide 1915 to nucleotide 1947, is shown for BAC DNA 15L17 and three AML cell lines, AML-193, KG-1 and HL-60 (nucleotide numbering starts with the ATG initiation codon with A = 1). Nucleotide 1933 is highlighted in bold for all DNA sequences, demonstrating the difference in BAC DNA 15L17 (C) from the other cell lines (T). The amino acid sequences, translated from the corresponding nucleotide sequences, are given on the top line, demonstrating that the nucleotide substitution does not result in an amino acid change.

Table 3 Intron-based PCR primers to amplify exons

Amplified Primer Primer Sequence Amplimer Annealing exon name orientation 5Ј → 3Ј length temeprature (bp)c (°C)

1 exon1.fora Forward GGTTTCCAGAAGCGCCTGC 260 51 intron1.rev Reverse GCAGCGCTAACTATGGAAGG 2 intr1.for Forward TAATAGCGCCACGGTTGAGA 250 55 intr2.rev Reverse AGTGGTGTTTGGCTTAAGGG 3 intr2.for Forward TGTAGTATACCACTGTTAATTTTCCAA 400 54 intr3.rev Reverse TTGGGTCCCCCCTAAAA 4 intr3.for Forward AATTAGCTTGGTGTGGTGGT 580 54 intra4.rev Reverse CACTGCACTGGCCAAAAATGAA 5 intr4.for Forward GTGGGAATTTTGTCATGGTAA 330 55 intr5.rev Reverse TCATCTACAGGAGCTGGAGC 6 intr5.for Forward TCTTGGGTTTTTAACCAGTGTA 310 51 intr6.rev Reverse CACATTCAGTCCAGATATCTGGC 7 intr6.for Forward CTGACAGAATTCCATGATTGAA 500 54 intr7.rev Reverse ATTTAAGCTATTTCCCTGAGATATGT 8 intr7.for Forward TTTTGGGCCATGGTAATTTA 470 54 intr8.rev Reverse TTTATTACCCTTAATGTGTAGGG 9 intr.8for Forward AAGGGTAATAAAAAGTAATTCATGC 280 52 intr9.rev Reverse ACTCAACTGACATTAGGCCAAT 10 intr.9for Forward TTAGATCTCACCCTCCCTTTG 450 56 intr10.rev Reverse TTTTAGGTCACCCCAGCA 11 intr10.for Forward CATTTCGTACTTGGGTTCCTG 400 55 intr11.rev Reverse CCATGTGACAAGGTAGGAAG 12 intr11.for Forward CATTTTTTCTATAGAAATGGAGTAGG 200 52 intr12.rev Reverse ATGGCTTAGATATTATCTCACTTTAC 13 intr12.for Forward ACTTGTTTCTCCCAGGAACC 360 52 intr13.rev Reverse AGCAATCCACAGGCTCAGTG 14 intr13.for Forward TCCTTGGTTTCATTTTGCTTT 420 54 intr14.rev Reverse AAGTAGCTGGGACTGCAG 15 intr14.for Forward CCACTCTGATTTCTCCCTCA 300 53 intr15.rev Revese TCCTGTACCATTACAAGTAGAAG 16 intr15.for Forward TACTCTACTTGTAATGGTACAGGAG 220 52 intr16.rev Reverse GATCAAGTTAGAATCAAAATCCACT 17 Intr16.for Forward TGCAACAAAGAGCAAATGGT 750 54 exon17b.revb Reverse CATCCTTGATGATTGTTCTCT aPrimer was designed at position −72 to −56 in 5Ј untranslated region. The nucleotide numbering starts with the ATG initiation codon (A = 1). bPrimer was designed at position 2658 to 2682 in 3Ј untranslated region. cSize based on estimation from PCR/agarose gel electrophoresis.

Leukemia Mortalin as a candidate for tumor suppressor gene on 5q31 H Xie et al 2134 it was necessary to first ascertain the gene structure and the Y, Wadhwa R. Malignant transformation of NIH3T3 cells by over approximate intron–exon boundaries. The intron–exon expression of mot-2 protein. Oncogene 1998; 17: 907–919. boundaries were estimated to coincide with those of murine 5 Wadhwa R, Akiyama S, Sugihara T, Reddel RR, Mitsui Y, Kaul SC. Genetic differences between the pancytolic and perinuclear forms mot-2, so genomic sequencing proceeded from these sites of murine mortalin. Exp Cell Res 1996; 226: 381–386. along a human BAC clone which contains the gene in its 6 Kaul SC, Wadhwa R, Matsuda Y, Hensler PJ, Pereira-Smith OM, entirety. The human boundaries were almost identical to those Komatsu Y, Mitsui Y. Mouse and human chromosomal assign- of the murine gene, but the intron sequence and sizes differed, ments of mortalin, a novel member of the murine 70 family of as expected. The intron sequence was then used to design . FEBS Lett 1995; 361: 269–272. PCR primers flanking each exon which were suitable for gen- 7 Horrigan S, Arbieva Z, Xie H, Kravarusic J, Fulton N, Naik H, Lee TT, Westbrook CA. Delineation of a minimal interval and identifi- omic sequencing, which could be used for the detection of cation of 9 candidates for a tumor suppressor gene in malignant mutations within the protein coding regions of the gene. The myeloid disorders on 5q31. Blood 2000; 95: 2372–2377. primers were evaluated in a small series of leukemia cell lines, 8 Bruno E, Horrigan SK, Van Den Berg D, Rozler E, Fitting PR, Moss including two lines with chromosome 5 deletion (KG-1 and ST, Westbrook CA, Hoffman R. The Smad5 gene is involved in HL-60) and one without (AML-193), as compared to BAC gen- the intracellular signaling pathways which mediate the inhibitory omic DNA 15L17. All 17 exons of the mortalin gene were effects of transforming growth factor-beta on human hematopo- iesis. Blood 1998; 91: 1917–1923. sequenced in these DNA sources, but no deviations from the 9 Horrigan SK, Westbrook CA. Construction and use of YAC contigs normal sequence were found, except for a conservative nucle- in disease regions. In: Boultwood J (ed). Methods in Molecular otide sequence variant in exon 16 which does not change the Biology, Gene Isolation and Mapping Protocols. Humana Press: amino acid sequence and thus represents a single nucleotide Totowa, NJ, 1996; pp 123–136. polymorphism (SNP). Although we failed to identify mutations 10 Michikawa Y, Baba T, Arai Y, Sakakura T, Kusakabe M. Structure in these two cell lines with chromosome 5 deletions, we have and organization of the gene encoding a mouse mitochondrial stress-70 protein. FEBS Lett 1993; 336: 27–33. demonstrated that these primers are suitable for sequence 11 Bhattacharyya T, Karnezis AN, Murphy SP, Hoang T, Freeman BC, analysis of the protein-coding regions of mortalin, and will be Phillipa B, Morimoto RI. Cloning and subcellular localization of useful for additional analysis in clinical samples. human mitochondrial . J Biol Chem 1995; 270: 1705–1710. In summary, we report here the genomic organization, 12 Steube KG, Gignac SM, Hu ZB, Teepe D, Harms D, Kabisch H, expression analysis, mutation analysis, and intron-based pri- Gaedicke G, Hansen-Hagge T, Macleod RAF, Quentmeier H, mers for human mortalin, a candidate for the myeloid leuke- Drexler HG. In vitro culture studies of childhood myelodysplastic syndrome: establishment of the cell line MUTZ-1. Leuk Lym- mia tumor suppressor gene at 5q31. Application of these pri- phoma 1997; 25: 345–363. mers to two AML cell lines with del(5q) failed to reveal any 13 Gallagher R, Collins S, Trujillo J, McCredie K, Ahearn M, Tsai S, mutations, although the sample size is small. Additional Metzgar R, Aulakh G, Ting R, Ruscetti F, Gallo R. Characterization sequencing of a larger number of clinical specimens of leuke- of the continuous, differentiating myeloid cell line (HL-60) from mias with del(5q) is now in progress, to fully evaluate mortalin a patient with acute promyelocytic leukemia. Blood 1979; 54: for evidence of mutational inactivation. 713–733. 14 Drexler HG, Minowada J. History and classification of human leu- kemia–lymphoma cell lines. Leuk Lymphoma 1998; 31: 305–316. 15 Heim S, Mitelman F. Cancer Cytogenetics. Willy-Liss: New Acknowledgements York, 1995. 16 Nimer SD, Golde DW. The 5q− abnormality. Blood 1987; 70: We would like to thank Dr Ignatius Gomes and Dr Zarema 1705–1712. 17 Thirman RS, Larson RA. Therapy-related myeloid leukemia. Hem- Arbieva for their comments and discussions on the manu- atol Oncol Clin North Am 1996; 10: 293–320. script, Dr Hans G Drexler (DSMZ, Braunschweig, Germany) 18 Larson RA, Le Beau MM, Vardiman JW, Testa JR, Golomb HM, for kindly providing leukemia cell lines for the study, and Seby Rowley JD. The predictive value of initial cytogenetic studies in Edassery for his assistance in preparation of the figures for the 148 adults with acute nonlymphocytic leukemia: a 12-year study manuscript. This work has been supported by the PHS grants (1970–1982). Cancer Genet Cytogenet 1983; 10: 219–236. RO1-CA-72593 and PO1-CA-75606. 19 Schiffer CA, Lee EJ, Tomiyasu T, Wiernik PH, Testa JR. Prognostic impact of cytogenetic abnormalities in patients with de novo acute nonlymphocytic leukemia. Blood 1989; 73: 263–270. 20 Le Beau MM, Espinosa III R, Neuman WL, Stock W, Roulston D, References Larson RA, Reinanen M, Westbrook CA. Cytogenetic and molecu- lar delineation of the smallest commonly deleted region of chro- 1 Wadhwa R, Kaul SC, Mitsui Y, Sugimoto Y. Differential subcellular mosome 5 in myeloid leukemias. Proc Natl Acad Sci USA 1993; distribution of mortalin in mortal and immortal mouse and human 90: 5484–5488. fibroblasts. Exp Cell Res 1993; 207: 442–448. 21 Wadhwa R, Takano S, Robert M, Yoshida A, Nomura H, Reddel 2 Wadhwa R, Mitsui Y, Ide T, Kaul SC. The anti-proliferative aspects RR, Mitsui Y, Kaul SC. Inactivation of tumor suppressor p53 by of mortalin. Int J Oncol 1995; 7: 69–74. Mot-2, a 70 family member. J Biol Chem 1998; 273: 29586– 3 Wadhwa R, Kaul SC, Sugimoto Y, Mitsui Y. Induction of cellular 29591. senescence by transfection of cytosolic mortalin cDNA in NIH 3T3 22 Takano S, Washwa R, Yoshii Y, Nose T, Kaul SC, Mitsui Y. Elev- cells. J Biol Chem 1993; 268: 22239–22242. ated levels of mortalin expression in human brain tumors. Exp Cell 4 Kaul SC, Duncan EL, Englezou A, Takano S, Reddel RR, Mitsui Res 1997; 237: 38–45.

Leukemia