Oncogene (2000) 19, 3739 ± 3749 ã 2000 Macmillan Publishers Ltd All rights reserved 0950 ± 9232/00 $15.00 www.nature.com/onc Cloning and expression of human B cell-speci®c BACH2 mapped to 6q15

Shinya Sasaki1, Etsuro Ito*,1, Tsutomu Toki1, Taira Maekawa3, Rika Kanezaki1,2, Takamichi Umenai1, Akihiko Muto4,5, Hirokazu Nagai7, Tomohiro Kinoshita7, Masayuki Yamamoto4, Johji Inazawa8, Makoto M Taketo9, Tatsutoshi Nakahata3,6,10, Kazuhiko Igarashi5,11 and Masaru Yokoyama1

1Department of Pediatrics, School of Medicine, Hirosaki University, Hirosaki 036-8563, Japan; 2Department of Biology, Faculty of Sciences, Hirosaki University, Hirosaki 036-8563, Japan; 3Department of Transfusion Medicine and Cell Therapy, The Institute of Medical Science, The University of Tokyo, Minato-ku, Tokyo 113-0033, Japan; 4Center for Tsukuba Advanced Research Alliance and Institute of Basic Medical Sciences, University of Tsukuba, Tsukuba 305-8575, Japan; 5Department of Biochemistry, Tohoku University School of Medicine, Sendai 980-8575, Japan; 6Department of Clinical Oncology, The Institute of Medical Science, The University of Tokyo, Minato-ku, Tokyo 113-0033, Japan; 7First Department of Internal Medicine, Nagoya University School of Medicine, Nagoya 466-8560, Japan; 8Department of Molecular Genetics, Division of Genetics, Medical Research Institute, Tokyo Medical and Dental University, Bunkyo-ku, Tokyo 113-8519, Japan; 9Graduate School of Pharmaceutical Sciences, The University of Tokyo, Bunkyo-ku, Tokyo 113-0033, Japan

The transcription factor Bach2, a member of the BTB- Introduction basic region (bZip) factor family, binds to a 12-O-tetradecanoylphorbol-13-acetate (TPA)-respon- The Maf family transcription factors possess a sive element and the related Maf-recognition element conserved basic region-leucine zipper (bZip) domain (MARE) by forming homodimers or heterodimers with which mediates /protein interaction and DNA Maf-related transcription factors. Bach2 regulates tran- binding (Nishizawa et al., 1989; Kataoka et al., scription by binding to these elements. To understand the 1993). While c-Maf, MafB and NRL contain function in hematopoiesis, we isolated a cDNA clone for putative transcription activation domains (Kataoko human Bach2 (BACH2) encoding a protein of 841 amino et al., 1994a; Nishizawa et al., 1989; Swaroop et acid residues with a deduced amino acid sequence having al., 1992), MafF, MafK and MafG lack canonical 89.5% identity to mouse homolog. Among human transactivation domains (Andrews et al., 1993b; hematopoietic cell lines, BACH2 is expressed abundantly Blank et al., 1997; Fujiwara et al., 1993; Igarashi only in some B-lymphocytic cell lines. RT ± PCR analysis et al., 1995; Kataoka et al., 1995; Toki et al., of hematopoietic cells revealed that BACH2 mRNA is 1997). MafF, MafG and MafK are essentially expressed in primary B-cells. Enforced expression of composed of bZip domains and are collectively BACH2 in a human Burkitt cell line, RAJI that does not referred to as the family . express endogenous BACH2, resulted in marked reduc- Various dimeric combinations of Maf family tion of clonogenic activity, indicating that BACH2 proteins bind in vitro to a DNA sequence motif G possesses an inhibitory e€ect on cell proliferation. By called T-MARE (TGCTGA /CTCAGCA) that con- G ¯uorescent in situ hybridization, the BACH2 was tains a TPA-responsive element (TRE; TGA /CT- localized to chromosome 6q15. Because deletion of the CA) (Kataoka et al., 1994a,b, 1995). The small Maf long arm of (6q) is one of the commonest proteins form heterodimers with a family of bZip chromosomal alterations in human B-cell lymphoma, we proteins that show remarkable homology to two examined for the loss of heterozygosity (LOH) of the invertebrate proteins, CNC and skn-1, which BACH2 gene in human B-cell non-Hodgkin's lymphomas regulate embryonic development in Drosophila (NHL). Among 25 informative cases, ®ve (20%) showed melanogaster and Caenorhabditis elegans, respectively LOH. These results indicate that BACH2 plays (Mohler et al., 1991, 1995; Bowerman et al., 1992). important roles in regulation of B cell development. This family of factors, referred to as the CNC Oncogene (2000) 19, 3739 ± 3749. family, include p45 NF-E2, Nrf1/LCR-F1/TCF-11, Nrf2, and Nrf3, all of which are similar in the Keywords: Bach2; Maf; non-Hodgkin's lymphoma; bZip domain and form heterodimers with small CNC family; NF-E2 Maf proteins (Andrews et al., 1993a,b; Chan et al., 1993a,b; Ney et al., 1993; Caterina et al., 1994; Moi et al., 1994; Igarashi et al., 1994, 1995; Marini et al., 1997; Johnsen et al., 1996; Toki et al., 1997; Kobayashi et al., 1999). Gene targeting has revealed distinct roles for several CNC family members in mammalian gene regulation. Homozygous disruption *Correspondence: E Ito of p45 NF-E2 gene in mouse results in defective Current addresses: 10Department of Pediatrics, Kyoto University, megakaryopoiesis and profound thrombocytopenia, Faculty of Medicine, Kyoto 606-8501, Japan; 11Department of leading to postnatal death (Shivdasani et al., 1995). Biochemistry, Hiroshima University School of Medicine, Hiroshima Nrf-1-null mutant mice are anemic due to a non- 734-8551, Japan Received 19 January 2000; revised 25 April 2000; accepted 31 May cell autonomous defect in de®nitive erythropoiesis 2000 and die in utero (Chan et al., 1998). Nrf-2-null Cloning and expression of BACH2 transcription factor S Sasaki et al 3740 mutant mice are viable but impaired in the The murine Bach proteins (Bach1 and Bach2) were xenobiotic inductive response of phase II detoxifying ®rst isolated as molecules that interact with MafK by enzyme gene (Itoh et al., 1997). yeast two hybrid screenings and were found to possess

A

Oncogene Cloning and expression of BACH2 transcription factor S Sasaki et al 3741 a CNC-related and less conserved bZip domain (Oyake To understand the function of Bach factors in et al., 1996). In addition to the bZip domain, Bach human cells, we isolated a cDNA clone for human proteins share a signature motif called the Bric-a-brac, Bach2 (BACH2) and determined the chromosomal Tramtrack and Broad complex (BTB) domain which is position for BACH2 and its expression in various involved in protein ± protein interactions (Albagli et al., tissues and hematopoietic cell lines. Among various 1995). Heterodimers of small Maf proteins and the hematopoietic cell lines, BACH2 was expressed CNC-related factors bind to the NF-E2 consensus site abundantly only in some B-lymphocytic cell lines. G T (TGCTGA /CTCA /C) which is related to MARE Enforced expression of BACH2 in a human Burkitt (Caterina et al., 1994; Chan et al., 1993a,b; Johnsen cell line, RAJI that does not express endogenous et al., 1996; Marini et al., 1997; Moi et al., 1994; Oyake BACH2, resulted in a marked reduction of clonogenic et al., 1996; Toki et al., 1997). Among these CNC eciency. The BACH2 gene is mapped to chromo- family proteins, the Bach proteins take a unique some 6q15, in a region where minimal molecular position in that they function as transcription deletion (RMD) is frequently detected in B-cell NHLs, repressors rather than activators in transient transfec- and we have found LOH of BACH2 in ®ve out of 25 tion assays (Oyake et al., 1996). However, the precise B-NHL cases (20%). These results have led to role of these potential heterodimers during develop- speculation that BACH2 may play an important role ment and cell di€erentiation has remained elusive. in B cell development.

B

Figure 1 (a) The nucleotide and deduced amino acid sequences of the BACH2 cDNA. Nucleotide sequences are numbered on the left and amino acid sequences numbered on the right. (b) Comparison of amino acid sequence of BACH2 with previously characterized murine Bach2 amino acid sequence. Asterisks indicate identical amino acids. The BTB domain, CNC domain and Serine-rich region are boxed. The bZip domain is underlined

Oncogene Cloning and expression of BACH2 transcription factor S Sasaki et al 3742 found another well-conserved region, adjacent to the Results bZip domain, including the so-called serine-rich Molecular cloning of BACH2 cDNA region (Oyake et al., 1996): 184 residues (94%) out of a total of 195 were found to be identical. Such a A screening of 16106 independent clones of a high degree of conservation suggests functional human endothelial cell cDNA library under relaxed importance for these domains. hybridization conditions using a mouse bach2 cDNA as a probe resulted in isolation of seven positive BACH2 is abundantly expressed in some B cell lines clones. Upon Southern hybridization and sequence analysis, clone #61 was found to encode BACH2 To determine expression of BACH2 in various based on a similarity to its mouse homolog. Since tissues, we analysed a panel of human tissues by clone #61 was found to be a partial cDNA, we Northern hybridization analysis. As shown in Figure screened another cDNA library from human ery- 2a, the human BACH2 mRNA (approximately 11 kb throleukemia cell line K562 using the mouse Bach2 in length) was essentially restricted to the thymus, cDNA probe and isolated a full-length clone (Figure spleen, and leukocytes. Relatively low levels of 1a). A comparison of the deduced amino acid BACH2 expression were also detected in the small sequences of mouse and human Bach2/BACH2 is intestine and brain. Because the expression sites of shown in Figure 1b. A high degree of conservation BACH2 were hematopoietic tissues, we next exam- along the entire ORF is evident. As is the case for ined BACH2 expression in various hematopoietic cell mouse Bach2, the BACH2 sequence contained a lines representing di€erent cell lineages (Figure 2b). BTB domain (from aa 1 ± 120) and a bZip domain The BACH2 mRNA was found in T lymphoid (from aa 491 ± 585). Out of the 120 amino acid (CEM) and erythro-megakaryocytic (K562) cells, residues in the BTB domain, 117 residues (97.5%) whereas it was barely detectable in B-lymphoid were found to be identical between the two species. (RAJI), myeloid-monocytoid (KG-1, HL60, THP1 The bZip regions showed a 96.8% identity. In and U937), and other erythro-megakaryocytic addition to these known functional domains, we (CMK86, CMK11-5, MGS, MEG01, KU812F,

Figure 2 BACH2 expression in human tissues and hematopoietic cell lines. (a) BACH2 mRNA expression in human adult tissues. Hybridization analysis was carried out with multiple human tissues. Northern blots (Clontech) containing approximately 2 mg aliquots of poly(A)+ RNA from each tissue. RNA size markers are shown on the left. (b) Northern blot analysis with poly(A)+ RNA derived from human leukemic cell lines. CMK86, CMK11-5, MGS, Meg01, KU812F, K562, HEL and YN-1 are erythroid- megakaryocytic cell lines. CEM and RAJI are lymphoid cell lines. THP-1 and U937 are monocytic cell lines. HL60 and KG-1 are myeloid cell lines. Integrity of RNA was veri®ed by hybridization with a glyceraldehyde-3 phosphate dehydrogenase (GAPDH) probe. (c) Northern blot analysis with poly(A)+ RNA derived from human lymphoid cell lines. CEM, KOPT-K1 and THP-6 are T cell lines. NALL-1, NALM-6, NALM-17 and RAJI are pre-B cell lines. NAMALWA and BALM-2 are immature B cell lines and RPMI 8226 is a myeloma cell line. (d) Immunoblot analysis of expression of BACH2 protein in human B cell lines. Fifteen mg aliquots of whole cell extracts prepared from each B cell line were separated with SDS-polyacrylamide gel, transferred onto membranes, and reacted with anti-Bach2 antibody. Positions of molecular markers are shown at the left side

Oncogene Cloning and expression of BACH2 transcription factor S Sasaki et al 3743 HEL and YN-1) cell lines. Because BACH2 Forced expression of BACH2 resulted in reduced colony expression was most abundant in a T lymphoid cell formation potential in a human B-lymphoma cell line, line, we next examined various lymphoid cell lines RAJI cells (Figure 2c). Unexpectedly, very high levels of expression of the BACH2 mRNA were detected in Mouse Bach2 forms heterodimers with the Maf some B-lymphoid (NALL1, NALM6, NALM17, oncoprotein family (Oyake et al., 1996). Thus, BACH2 NAMALWA) cell lines, although it was lacking in may be involved in regulation of cell proliferation. To three others (BALM2, RAJI and RPMI18226). The test this possibility, we examined the e€ects of BACH2 levels of expression of BACH2 in T lymphoid over-expression in the Burkitt lymphoma cell line (CEM, KOPT-K1, THP-6, KOPY-K1) cell lines RAJI, which does not express endogenous BACH2. were found to be relatively low compared with RAJI cells were infected with the retrovirus harboring some of the B-lymphoid cell lines. To con®rm these the human BACH2 cDNA, and stably transduced results, immunoblotting analysis with anti-BACH2 RAJI clones were selected in the presence of G418. As antiserum was performed (Figure 2d). BACH2 was control lines, stable RAJI clones, which were trans- found to be expressed in all B-precursor cell lines duced with the same vector lacking the BACH2 cDNA, (NALL1, NALM6, and NALM17), three out of ®ve were also established. Expression levels of BACH2 Burkitt cell lines (NAMALWA, RAMOS, and protein were veri®ed by a blot analysis using anti- Daudi) and two out of ®ve multiple myeloma cell Bach2 polyclonal antibodies. Six BACH2 RAJI clones lines (KMS-12-PE and HS-Sultan). However, abundantly expressed BACH2 protein, whereas the BACH2 expression was lacking in two Burkitt cell control clone lacked its expression (Figure 4a). Two lines (BALM2, RAJI) and three myeloma (or clones that over-expressed BACH2 at the highest level plasmacytoma) cell lines (RPMI 1788, IM9, were studied further. Generations of MARE-binding RPMI8226). The expression pro®le of the BACH2 activities within these cells were then examined in protein in di€erent B cell lines correlated well with electrophoretic mobility shift (EMSA) assay. As shown the presence of BACH2 mRNA. in Figure 4b, a prominent MARE-binding activity was detected in the BACH2 over-expression lines but not in the control lines (compare lanes 2 ± 5). The speci®city BACH2 is expressed in primary B cells was con®rmed by the inhibition of the complex To examine whether BACH2 is expressed in primary formation in the presence of excess unlabeled MARE B-cells in a stage-speci®c manner, we ®rst separated oligonucleotide (data not shown). The complex con- human cord blood cells into two fractions by cell tained both BACH2 and small Maf protein, since anti- sorting using antibodies against CD34 and CD10. BACH2 monoclonal antibodies super-shifted the com- The two fractions were (1) CD34+, CD10-(hemato- plex (lanes 6 and 7) and anti-small Maf antiserum poietic progenitor cells), and (2) CD34+, CD10+ inhibited the MARE binding activity (lanes 8 and 9). (progenitor B cells at early stage). To obtain more mature progenitor B-cells or precursor B-cells, we used a long-term culture system using the murine bone marrow stromal cells MS-5 (Nishihara et al., 1998). Fractionated CD34+ hematopoietic progenitor cells from human cord blood were cultured on MS-5 in the presence of rhSCF and rhG-CSF for 5 weeks. Flow cytometric analysis was performed on adherent cells with the gate for the lymphocyte fraction. About 95% of the cells showed phenotypes of CD10+/19+. These cells did not express CD34 or IgM on the cell surface. A small part of the cells expressed cytoplasmic m chain, suggesting that these cells are progenitor B cells or precursor B cells. BACH2 expression in each fraction was analysed by RT ± PCR. Only the CD10+/19+ fraction showed BACH2 mRNA expression, whereas the CD34+/ CD107 and CD34+/CD10+ fraction did not express BACH2 (Figure 3a). To analyse BACH2 expression in the later stage of B-cell development, we next separated immature B-cells and mature B-cells from adult peripheral blood cells by cell sorting using the B-cell speci®c antibodies. The CD19+/CD21+/IgD7 fraction and the CD19+/ CD21+/IgD+ fraction represent immature B-cells and mature B-cells, respectively. As shown in Figure Figure 3 Expression of BACH2 during primary B-cell di€er- 3b, BACH2 was expressed in both immature B-cells entiation. (a) Hematopoietic cells were fractionated from human and mature B-cells, and the expression level was cord blood (CB) cells and CD34+ CB cells cultured on MS-5 comparable to that of some B-cell lines. These results cells depending on the surface marker expression. BACH2 established that BACH2 is expressed in primary B-cells expression in each fraction was analysed by RT ± PCR analysis. (b) Immature B-cells and mature B-cells were fractionated from at the progenitor B, precursor B, immature B and adult peripheral blood and BACH2 expression was determined by mature B stages. RT ± PCR

Oncogene Cloning and expression of BACH2 transcription factor S Sasaki et al 3744 These results indicated that over-expressed BACH2 be consistently and signi®cantly reduced (10 ± 40%) generated the MARE binding complex together with (Figure 5b). Furthermore, the average size of the one or other of the small Maf proteins within RAJI colonies from BACH2 expressing RAJI clones was cells. smaller than those from control clones (Figure 5a). Having established RAJI cell lines that over- These results suggested that BACH2 is a negative expressed BACH2, we examined its e€ect on cell regulator of clonogenic activity. growth in two kinds of assays. One was to monitor growth curves, and the other was to monitor colony Isolation of a genomic clone and chromosome location of formation activity. The latter assay is more stringent in BACH2 gene terms of cell proliferation. The BACH2 over-expressing lines showed roughly similar curves to those of the The human BACH2 was shown to be a unique gene by control lines when monitored in liquid cultures (data Southern hybridization analysis (data not shown). not shown). On the other hand, the forced expression Human BACH2 cDNA probes were then used to of BACH2 caused signi®cant e€ects on the growth screen a commercial P1 human genomic library properties in RAJI cells in methylcellulose colony (Genomic Systems, St Louis, MO, USA) and a BACH2 assay. The plating eciency of control clones was genomic clone was isolated. Metaphase cells were about 40 ± 60%. In contrast, the plating eciencies of examined by FISH using the P1 clone as a probe. all RAJI clones which expressed BACH2 was found to Examination of these chromosomal spreads allowed us to localize the human BACH2 gene to 6q15 (Figure 6). Hybridization eciency for the BACH2 probe was approximately 95%. No consistent hybridization signal was observed at any other site.

LOH of BACH2 gene was found in B-NHL patients with a relatively high frequency Deletions of chromosome 6 is commonly detected in human B-cell lymphoma by both cytogenetic and LOH analyses (Guan et al., 1996). As BACH2 was expressed

Figure 4 Expression of BACH2 in BACH2 transfectants. (a) Expression of BACH2 protein in various RAJI clones was examined by Western blot analysis. Thirty mg aliquots of whole cell extracts prepared from each clone were separated with SDS- polyacrylamide gel, transferred onto membranes, and reacted with anti-Bach2 antibody. Position of molecular markers is shown at the left side. (b) MARE-binding complex in RAJI cell extracts. EMSA was carried out with oligonucleotide probe which contains Figure 5 Comparative analysis of plating eciency of BACH2 MARE in the HS3/Ca3'E (Muto et al., 1998), and extracts were transfectants. Methylcellulose colony formation assay was prepared from the indicated cell lines. Antibodies were anti-Bach2 performed as described in Materials and methods. As a control, monoclonal antibody (Muto et al., 1998; lanes 6 and 7), anti- stable RAJI clones, which were transduced with the pDL+ vector MafK rabbit serum that reacts with the small Maf proteins (lanes lacking the BACH2 cDNA, were used. (a) Colonies were 8 and 9), or preimmune serum (lanes 10 and 11). Bach2/small photographed 2 weeks after plating. (b) The relative plating Maf complex is indicated with arrow eciency was examined 11 days after plating

Oncogene Cloning and expression of BACH2 transcription factor S Sasaki et al 3745

Figure 7 LOH analysis of the BACH2 gene in human B-cell NHLs using microsatellite marker BACH2/CA-1. The PCR products of the tumor sample (T) are shown to the left and of the normal sample (N) are shown

human B-lymphoma cell line, RAJI. Furthermore, FISH analysis showed the BACH2 gene to be localized to chromosome 6q15, where alterations have been often reported in human B-cell lymphomas (Guan et al., 1996). Indeed, we detected LOH of the BACH2 gene in ®ve out of 25 patients with B-cell NHL. These data, coupled with the observation that BACH2 is expressed in primary B cells, have led to speculation that BACH2 may play an important regulatory role in B cell development. In the mouse, Bach2 is expressed in pro-B, pre-B, immature B and mature B cell lines, and it is absent in Figure 6 Mapping of the human BACH2 gene by FISH. plasmacytoma cell lines (Muto et al., 1998). In human Photograph of human metaphase hybridized with B-lymphoid cell lines, Northern hybridization and the human BACH2 gene fragment as a probe (the upper) and corresponding G-banded chromosomes (the lower). Arrows point Western blot analyses revealed abundant expression to the site of hybridization on both chromosomes 6 in band q15 in three pre-B cell lines, NALM-1, NALM-6 and NALM-17, and three immature B cell lines, NAMAL- WA, RAMOS and Daudi. In contrast, BACH2 at high levels in B-lymphocytic lineage, forced expression was not detected in three multiple myeloma expression of BACH2 in a B-cell line resulted in cell lines (or plasma cell lines), RPMI 8226, IM9 and reduced growth potential, and its gene was located at RPMI 1788. These results support the notion that 6q15, we investigated whether it is altered in B-cell BACH2 is expressed in B cells in a stage-speci®c lymphomas. To examine LOH, a polymorphic (CA)n manner. However, two Burkitt lymphoma cell lines, repeat microsatellite marker (BACH2/CA-1, see Mate- RAJI and BALM2, which are in pre-B and immature- rials and methods) was ®rst isolated from the human B cell stages, respectively, did not express BACH2. BACH2 P1 phage genomic clone. LOH was then Considering the fact that BACH2 is expressed in analysed in pairs of normal DNA (peripheral blood) primary human B cells at corresponding stages, these and tumor DNA from 30 B-NHL cases using the results suggest that BACH2 expression is deregulated BACH2/CA-1 microsatellite marker. The average size in these tumor cell lines due to abnormalities in either of the PCR product is about 170 bp, which depends on trans- or cis-factors. the number of CA repeats. The BACH2/CA-1 Two multiple myeloma cell lines abundantly ex- microsatellite marker was found to be heterozygous pressed BACH2. Multiple myeloma is characterized by and informative for LOH analysis in 25 out of 30 extensive in®ltration of the bone marrow by mono- cases. Representative LOH results obtained with this clonal plasma cells. However, the self-renewing stem marker are shown in Figure 7. LOH was found in ®ve cell in myeloma is likely to be a B-lineage cell that is patients out of the 25 informative patients (20%). less mature than the end-stage plasma cells in the bone Seven cases had LOH in another microsatellite marker marrow (Caligaris-Cappio et al., 1985). High level of D6S292, located on chromosome 6q16.3-q23.2. All ®ve BACH2 expression indicates that these two cell lines patients with LOH of marker BACH2/CA-1 were also established from multiple myeloma patients may not found to have the LOH at D6S292. These results represent plasma cells. Alternatively, BACH2 may be indicate that LOH of BACH2 is associated with NHL deregulated in these myeloma cell lines. Because with a relatively high frequency. multiple myeloma is a heterogeneous disorder, BACH2 expression might be a useful marker for classi®cation of multiple myelomas. Discussion Human tumors have been shown to progress by accumulation of genetic abnormalities, with frequent In the present study, we have described for the ®rst recurrent chromosomal changes which cause altera- time the isolation and characterization of human tions in the expression of oncogenes or tumor cDNAs encoding the BACH2 protein. The results of suppressor (Fearon and Vogelstein, 1990). Northern hybridization and Western blot analyses Non-random loss of heterozygosity (LOH) of chro- indicated that BACH2 was expressed at a high level mosomal regions has often been observed in various in human B cell lines. Over-expression of BACH2 human tumors, and these areas are likely to harbor resulted in reduced colony formation potential in the tumor suppressor genes whose mutations are respon-

Oncogene Cloning and expression of BACH2 transcription factor S Sasaki et al 3746 sible for the tumorigenesis. Cytogenetic studies have Cell culture shown that deletion of the long arm of chromosome 6 (6q) is one of the commonest chromosomal changes in Hematopoietic cell lines were maintained in RPMI1640 all non-Hodgkin's lymphomas (NHL) (Ot et al., (Nissui) medium supplemented with 10% fetal bovine serum atmosphere. NALL-1, 1993, 1994; Gaidano et al., 1992) and acute lympho- (FBS), at 378C and in 5% CO2 NALM-6, BALM-2, NAMALWA and RPMI 8226 cells were blastic leukemias (ALL) (Hayashi et al., 1990). Along gifts of Dr K Ohtani, Fujisaki Cell Center of Hayashibara 6q, several regions of convergent deletions have been Bichemical Laboratories, Inc. (Okayama). KOPT-K1, THP-6 reported: (1) two regions of minimal molecular and NALM-17 cells were provided by Dr Y Hayashi, the deletion (RMDs) termed RMD1 (localized to 6q25 ± University of Tokyo. MGS cells were donated by Dr S 27) and RMD2 (localized to 6q21 ± 23) determined by Yokoyama, Yamagata University, and KG-1, HL60, U937, LOH studies (Ot et al., 1993), (2) two common THP1, RAMOS, Daudi, RPMI1788, KMS-12-PE, IM-9 and regions of deletions at 6q23.2 ± 27 and 6q14 ± 21 in HS-Sultan were obtained from the Japanese Cancer Re- ALL detected by ¯uorescent in situ hybridization sources Bank (Tokyo). (FISH) of yeast arti®cial chromosome clones mapped by 6q (Menasce et al., 1994), (3) RMD encompassing Immunoblot analysis 6q11 ± 21 in B-cell lymphoma found by Micro-FISH Whole-cell extracts were prepared using triple-detergent lysis techniques (Guan et al., 1996), and (4) RMD of less bu€er consisting of 50 mM Tris-HCl, 150 mM NaCl, 0.02% than 500 kb in chromosome 6q16 ± 21 in ALL sodium azide, 0.1% SDS, 1% NP-40, 0.5% sodium revealed by LOH studies (Merup et al., 1998). These deoxycholate and 0.6 mM PMSF, and were separated on observations suggest that the 16q region may contain SDS-polyacrylamide gels. Proteins were transferred onto multiple tumor suppressor genes. However, candidate polyvinylidene di¯uoride membranes (Waters) and processed tumor suppressor genes in this chromosomal region for reaction with the antiserum against Bach2 (F69-2; Oyake have yet to be described. The present FISH results et al., 1996). Secondary anti-rabbit IgG antibodies were revealed that BACH2 is located in the middle of the conjugated with horseradish peroxidase, and detection of RMD encompassing 6q11 ± 21 in B-cell lymphoma. peroxidase activity was carried out with the enhanced chemiluminescence system (Amersham). The present LOH analysis revealed a relatively high frequency of LOH of BACH2 gene (20%) in human B-cell lymphomas. However, we cannot exclude the Northern hybridization possibility that the BACH2 gene may be deleted just Poly(A)+ RNAs were isolated from human hematopoietic cell as a bystander of some unknown tumor suppressor lines with a Quickprep micro mRNA puri®cation kit genes of B cell tumorigenesis localized on chromosome (Pharmacia Biotec). Approximately 2 mg aliquots of Poly(A)+ 6q. Twenty-seven out of 30 patients analysed had RNAs were separated on 1% agarose-formaldehyde gels and di€use large cell type lymphoma, and the remaining transferred to Hybond-N+ membranes (Amersham) as three included di€use mixed, small cleaved and described previously (Ito et al., 1993). Northern blots marginal zone types. All patients with LOH of the containing multiple human tissue RNA samples were purchased from Clontech, each tissue sample containing BACH2 gene belonged to the di€use large cell group. 2 mg of poly(A)+ RNA. Hybridization with a 32P-labeled Three out of the ®ve patients survived more than 2.5 human BACH2 cDNA fragment containing the CNC domain years, indicating that LOH of the BACH2 gene may and bZip domain, or rat glyceraldehyde-3-phosphate dehy- not be a factor that presupposes a poor prognosis. drogenase (GAPDH) cDNA, and subsequent washings were Although the number of the cases included in the carried out as described previously (Toki et al., 1997). present study was small, the possibility of an association between LOH of BACH2 and development Primary hematopoietic cell culture of B-cell malignancies is worth further investigation in a large series. Human umbilical cord blood (CB) was collected from healthy, full-term neonates according to institutional guide- lines immediately after vaginal delivery. Mononuclear cells (MNCs) were separated by Ficoll density gradient centrifuga- Materials and methods tion after depletion of phagocytes with silica (IBL) as described elsewhere (Imai et al., 1991). CB CD34+ cells Isolation of human BACH2 cDNA clones were isolated from MNCs by using Dynabeads M-450 Human endothelial cell and K562 erythroleukemic cell cDNA conjugated with CD34 monoclonal Ab (mAb) and DETA- libraries constructed in the lgt11 vector were plated on CHaBEAD CD34 (Dynal) as described previously (Sui et al., 150 mm Petri dishes at a density of 56104 plaque-forming 1995). CB CD34+107, CD34+10+ cells were puri®ed from units (pfu) per plate. One6106 pfu were screened with a CB CD34+ cells by cell sorting on FACS Vantage (Becton mouse Bach2 cDNA fragment (Oyake et al., 1996). Dickinson). Puri®ed CB CD34+ cells were plated at a Duplicated ®lters were made from each plate and processed density of 16103 cells/ml onto the MS-5 feeders in a-MEM for hybridization to the probe as described previously (Ito et supplemented with 10% FCS, recombinant human (rh) stem al., 1993). This was carried out in the presence of 66SSC cell factor (rhSCF) and granulocyte-colony-stimulating factor and 0.25% skim milk at 558C overnight. Filters were washed (rhG-CSF) as described before (Nishihira et al., 1998). All in 26SSC/0.1% SDS at room temperature and 0.56SSC/ cultures were conducted in a 378C, humidi®ed incubator 0.1% SDS at 558C. Positive clones were puri®ed by two containing 5% CO2 in air, and the cultures were fed every additional plaque hybridization screenings. Inserts of positive week by removing half of medium and replacing it with fresh phage clones were then subcloned into the EcoRI site of the medium. Cytokines were also added at the time of feeding. pBluescript KS(+) plasmid (Stratagene) and sequenced on After 2 ± 6 weeks, cells were harvested from the dishes by both strands using a Sequenase version 2.0 DNA sequencing using 0.05% trypsin (Gibco) plus 0.02% ethylenediaminete- kit (Amersham) or an ABI PRISM Cycle sequencing kit tracetic acid (Wako), and the number of viable cells per well (Applied Biosystems). was determined.

Oncogene Cloning and expression of BACH2 transcription factor S Sasaki et al 3747 Gene transduction into RAJI cells was performed as Fluorescent antibodies and phenotype analysis follows. One ml of RAJI cells at a concentration of 105 ml Cyto¯uorometric studies were performed using ¯uorescent in RPMI1640 containing 10% FCS and 1 ml of the virus isothiocyanate (FITC)-, phycoerythrin (PE)- and PE- supernatants were placed into the RetroNectionTM Dish cyanin5 isothiocyanate (PECy5)-conjugated mAbs. PE- (TaKaRa), which is a precoated culture dish (35 mm) with CD19 (Leu-12) and FITC-CD34 (HPCA-2) mAbs were a recombinant human ®bronectin fragment RetroNectinTM purchased from B&D. FITC and PE-CD10 (SS2/36) mAbs (TaKaRa). Stable transfectants were selected in the medium were purchased from DAKO. FITC-conjugated goat anity containing 1 mg ml of G418, and then single cell clones were puri®ed F(ab')2 Ab to human mH chain was used to detect established by the limiting dilution method. human Cm and Sm, and FITC-conjugated goat anity puri®ed F(ab')2 to rabbit Ig was used as a negative control. Electrophoretic mobility shift assay (EMSA) Both were purchased from Tago. Biotin-conjugated goat polyclonal Ab to mouse Ig (DAKO) and Tricolor (TC)- An oligonucleotide probe was from the IgH HS3 (5'- conjugated streptavidin (Caltag Lab.) were used to stain GCTGAGGGCAGCTGAGTCATCCTGAGCCT-3'). The these puri®ed Abs. For all Abs, negative controls were double-stranded oligonucleotide was labeled with [g- provided by Ig of the same isotype but unrelated speci®city. 32P]ATP and T4 Polynucleotide Kinase (TOYOBO). Incuba- The surface phenotype of culture cells was determined by tion of B cell extracts with the DNA probe was carried out at two or three-color ¯ow cytometry on a FACS (Cytoron 378C for 10 min under the conditions described previously Absolute). Culture cells were incubated for 30 min at room (Muto et al., 1998). Where indicated, antibodies were added temperature with 0.5% human g-globulin (Sigma) or 0.5% to the binding reaction at dilution of 1/10 and chilled for normal mouse serum (DAKO) to block nonspeci®c binding. 30 min on ice before addition of DNA. The reaction products After washing, cells were then stained with ¯uorescent- were electrophoresed on 4% polyacrylamide gels. Nuclear conjugated Abs. extracts from RAJI cells were prepared from 107 cells as described previously (Andrews and Faller, 1991). RT ± PCR assays Colony growth in methylcellulose culture One6105 cells of fractions (CD34+10+ and CD34+107) in CB cells and fraction (CD10+19+cells) in generated B The clonogenic activity of each RAJI cell subclone was cells from CB CD34+ cells on MS-5 were sorted by FACS evaluated by methylcellulose culture. Five6102 RAJI cells Vantage. The total RNA was isolated following procedures were placed in 1 ml of RPMI medium containing 0.88% described by the manufacturer (Qiagen). Genomic DNA in methylcellulose and 10% FCS in triplicate, and incubated total RNA was digested by DNAase I (Life Technologies) under fully humidi®ed conditions in 6-well culture dish and DNAase I was inactivated by heat for 10 min. The (IWAKI) at 378C in 5% CO2 atmosphere. Colonies were synthesis of ®rst strand cDNA (reverse transcription (RT)) counted under an inverted microscope on days 7 and 11 after was performed for 60 min at 428C using oligo (dT)12 ± 18 plating. and Superscript II reverse transcriptase (Life Technologies). The cDNa solution was ampli®ed using EX Taq polymerase Fluorescence in situ hybridization (FISH) mapping (Takara) and each oligonucleotide as primer; humna G3PDH: sense, 5'-ACCACAGTCCATGCCATCAC-3', and The human BACH2 cDNA clone was used to screen a antisense, 5'-TCCACCACCCTGTTGCTGTA-3', at 608C; commercial P1 phage human genomic library (Genomic human S14: sense, 5'-GGCAGACCGAGATGAATCCTCA- Systems), and a human BACH2 genomic clone was isolated. 3', antisense, 5'-CAGGTCCAGGGGTCTTGGTCC-3' at Detailed analysis of the genomic structure will be reported 558C; human BACH2: sense, 5'-TCCTTGCCACAGAACAT- elsewhere. FISH analysis was performed on metaphase cells CAGGAAC-3' and antisense, 5'-TGGATGTCTCGG- obtained from normal human lymphocyte cultures as CAAACTTCCTGG-3', at 558C. To insure that PCR described earlier (Inazawa et al., 1993). products were not derived from DNA, primers were de®ned from discrete exons for each transcript. The ®rst cycle was Isolation of (CA)n repeat microsatellite marker in BACH2 run as follows: denaturation at 948C for 3 min, annealing at genomic clone temperatures indicated above for 30 s, and synthesis at 728C for 30 s. The next 29 cycles were run similarly and the To identify the polymorphic marker (microsatellite) in the P1 synthesis of last cycle was 5 min. The ampli®ed products were phage genomic clone, a Southern blot analysis was performed resolved by electrophoresis in 2% agarose gel. The products, using the (CA)20 oligonucleotide as a probe, and selected obtained with each set of primers, were sequenced using the fragments were subcloned into the Bluescript KS(+) plasmid Ready Reaction Dye Determinator (Perkin-Elmer Corp., (Stratagene). DNA sequences were determined on both Norwalk) and an automated DNA sequencer (model 373; strands by dideoxynucleotide methods using an ABI PRISM Applied Biosystems, Inc.). Cycle sequencing kit (Applied Biosystems). The microsatellite marker thus obtained was termed BACH2/CA-1, and had 25 repetititions of the CA sequence. BACH2 retrovirus and gene transduction into RAJI cells A BamHI and SalI fragment containing the human BACH2 Tissue samples cDNA was inserted into the unique corresponding cloning sites of pDL+ retrovirus vector (Nakao et al., 1995). The Thirty matched samples of normal and tumor DNA were amphotrophic packaging cell line PAMP51 that was kindly obtained from consenting patients with B-cell di€use NHL provided by Drs K Ikenaka and T Yoshimatsu (National (di€use large cell type, 27 cases; di€use mixed cell type, 1 Institute of Physiological Science, Okazaki, Japan) was case; di€use small cleaved cell type, 1 case; marginal zone transfected with pDL+-BACH2 plasmid by using Lipofect- type, 1 case). B cell phenotype was evaluated by immuno- AMINE Reagent (GIBCO ± BRL) as described previously histochemical analysis using anti-CD20 monoclonal antibody. (Yoshimatsu et al., 1998). The stable amphotrophic producer We used biopsied lymph nodes involved by tumor and an cells were established after selection in the medium contain- extranodal lesion (one case from thyroid) as tumor samples. ing 1 mg/ml of G418 (GIBCO ± BRL). The virus super- Mononuclear cells separated from peripheral blood or bone natants were harvested from the producer cells cultured at marrow, which were microscopically determined to have no 328C. contamination by malignant cells, were used as normal

Oncogene Cloning and expression of BACH2 transcription factor S Sasaki et al 3748 samples. DNA extraction was performed using the method stained by Silver Stain Plus (BIO-RAD) according to the described previously (Uchida et al., 1997). manufacturer's recommendations.

Loss of heterozygosity (LOH) analysis and patients Acknowledgments The primer sequences for the microsatellite BACH/CA-1 were We would like to dedicate this paper to Professor M as follows: Sense, 5'-CTGCAGGGAATAGGTAACTGT-3'; Yokoyama who passed away on 11 March 2000. We thank antisense, 5'-AAAGTGCAGGACCCCTCTGA-3'. Fifty ng Dr K Ohtani, Fujisaki Cell Center of Hayashibara aliquots of genomic DNA were subjected to PCR, performed Biochemical Laboratories, Inc., Dr Y Hayashi for various in 25 ml containing 100 mM of each dNTP, 1 mM of each cell lines and helpful discussions, and Dr T Ikenaka for primer and 12.5 ml of Premix taq (TaKaRa). The reaction retrovirus vector and cell lines. This work was supported mixture was heated to 948C for 5 min followed by 30 cycles by Grants-in-Aid for Scienti®c Research, Grants-in-Aid for of 30 s at 948C, 30 s at 508C, 1 min at 728C, followed by a Scienti®c Research on Priority Areas and a Grant-in-Aid 5 min ®nal extension at 728C with a Gene Amp PCR System for Encouragement of Young Scientists from the Ministry (Perkin-Elmer). The PCR products were electrophoretically of Education, Science, Sports and Culture and grants from separated on 10% denaturing polyacrylamide gels and then Uehara Memorial Foundation and Naito Foundation.

References

Albagli O, Dhordain P, Deweindt C, Lecocq G and Leprince Johnsen O, Skammelsrud N, Luna L, Nishizawa M, Prydz H D. (1995). Cell Growth Di€er., 6, 1193 ± 1198. and Kolsto AB. (1996). Nucl. Acids Res., 24, 4289 ± 4297. Andrews NC and Faller DV. (1991). Nucl. Acid. Res., 19, Kataoka K, Nishizawa M and Kawai S. (1993). J. Virol., 67, 2499. 2133 ± 2141. Andrews NC, Erdjument-Bromage H, Davidson MB, Kataoka K, Fujiwara KT, Noda M and Nishizawa M. Tempst P and Orkin SH. (1993a). Nature, 362, 722 ± 728. (1994a). Mol. Cell. Biol., 14, 7581 ± 7591. Andrews NC, Kotkow KJ, Ney PA, Erdjument-Bromage H, Kataoka K, Noda M and Nishizawa M. (1994b). Mol. Cell. Tempst P and Orkin SH. (1993b). Proc. Natl. Acad. Sci. Biol., 14, 700 ± 712. USA, 90, 11488 ± 11492. Kataoka K, Igarashi K, Itoh K, Fujiwara KT, Noda M, Blank V, Kim MJ and Andrews NC. (1997). Blood, 89, Yamamoto M and Nishizawa M. (1995). Mol. Cell. Biol., 3925 ± 3935. 15, 2180 ± 2190. Bowerman B, Eaton BA and Priess JR. (1992). Cell, 68, Kobayashi A, Ito E, Toki T, Kogame K, Takahashi S, 1061 ± 1075. Igarashi K, Hayashi N and Yamamoto M. (1999). J. Biol. Caligaris-Cappio F, Bergui L, Tesio L, Pizzolo G, Malavasi Chem., 274, 6443 ± 6452. F, Chilosi M, Campana D, van Camp B and Janossy G. Marini MG, Chan K, Casula L, Kan YW, Cao A and Moi P. (1985). J. Clin. Invest., 76, 1243 ± 1251. (1997). J. Biol. Chem., 272, 16490 ± 16497. Caterina JJ, Donze D, Sun CW, Ciavatta DJ and Townes Menasce LP, Orphanos V, Santibanez-Koref M, Boyle JM TM. (1994). Nucl. Acids Res., 22, 2383 ± 2391. and Harrison CJ. (1994). Gene Chromosome Cancer, 10, Chan JY, Han XL and Kan YW. (1993a). Proc. Natl. Acad. 26 ± 29. Sci. USA, 90, 11366 ± 11370. Merup M, Moreno TC, Hayman M, Ronnberg K, Grander Chan JY, Han XL and Kan YW. (1993b). Proc. Natl. Acad. D, Detlofsson R, Rasool O, Liu Y, Soderhall JS, Juliusson Sci. USA, 90, 11371 ± 11375. G, Gahrton G and Eihorn S. (1998). Blood, 91, 3397 ± Chan JY, Kwong M, Lu R, Chang J, Wang B, Yen TSB and 3400. Kan YW. (1998). EMBO J., 17, 1779 ± 1787. Mohler J, Vani K, Leung S and Epstein A. (1991). Mech. Fearon EF and Volgelstein B. (1990). Cell, 61, 759. Dev., 34, 3 ± 9. Fujiwara K, Kataoka K and Nishizawa M. (1993). Mohler J, Maha€ey J, Deutsch E and Vani K. (1995). Dev. Oncogene, 8, 2371 ± 2380. Suppl., 121, 237 ± 247. Gaidano G, Hauptschein RS, Parsa NZ, Ot K, Rao PH, Moi P, Chan K, Asunis I, Cao A and Kan YW. (1994). Proc. Lenoir G, Knowles DM, Chaganti RSK and Dalla-Favera Natl. Acad. Sci. USA, 91, 9926 ± 9930. R. (1992). Blood, 82, 1781 ± 1787. Muto A, Hoshino H, Madisen L, Yanai N, Obinata M, Guan XY, Horsman D, Zhang HE, Parsa NZ, Meltzer PS Karasuyama H, Hayashi N, Nakauchi H, Yamamoto M, and Trent JM. (1996). Blood, 88, 1418 ± 1422. Groudine M and Igarashi K. (1998). EMBO J., 17, 5734 ± Hayashi Y, Raimondi SC, Look AT, Behn FG, Kitcheng- 5743. man GR, Pui CH, Rivera GK and Williams DL. (1990). Nakao J, Yamada M, Kagawa T, Kim SU, Miyao Y, Blood, 76, 1626 ± 1630. Shimizu K, Mikishiba K and Ikenaka K. (1995). J. Igarashi K, Kataoka K, Itoh K, Hayashi N, Nishizawa M Neurochem., 64, 2396 ± 2403. and Yamamoto M. (1994). Nature, 367, 568 ± 572. Ney PA, Andrews JC, Jane SM, Safer B, Purucker ME, Igarashi K, Itoh K, Motohashi H, Hayashi N, Matuzaki Y, Weremowicz S, Morton CC, Go€ SC, Orkin SH and Nakauchi H, Nishizawa M and Yamamoto M. (1995). J. Nienhuis AW. (1993). Mol. Cell. Biol., 13, 5604 ± 5612. Biol. Chem., 270, 7615 ± 7624. Nishihara M, Wada Y, Ogami K, Ebina Y, Ishii T, Tsuji K, Imai T, Koike K, Kudo T, Kikuti T, Amano Y, Takagi M, Ueno H, Asano S, Nakahata T and Maekawa T. (1998). Okumura N and Nakahata T. (1991). Blood, 78, 1969 ± Eur. J. Immunol., 28, 855 ± 864. 1974. Nishizawa M, Kataoka K, Goto N, Fujiwara KT and Kawai Inazawa J, Saito H, Ariyama T, Abe T and Nakamura Y. S. (1989). Proc. Natl. Acad. Sci. USA, 91, 9926 ± 9930. (1993). Genomics, 17, 153 ± 162. Ot K, Parsa NZ, Gaidano G, Filippa DA, Luie D, Pan D, Ito E, Toki T, Ishihara H, Ohtani H, Gu L, Yokoyama M, Jhanwar SC, Dalla-Favera R and Chaganti RSK. (1993). Engel JD and Yamamoto M. (1993). Nature, 362, 466 ± Blood, 82, 2157 ± 2162. 468. Ot K, Luie DC, Parsa NZ, Filippa D, Gangi M, Siebert R Itoh K, Chiba T, Takahashi S, Ishii T, Igarashi K, Katoh Y, and Chaganti RSK. (1994). Blood, 83, 2611 ± 2618. Oyake T, Hayashi N, Satoh K, Hatayama I, Yamamoto M Oyake T, Itoh K, Motohashi H, Hayashi N, Hoshino H, and Nabeshima Y. (1997). Biochem. Biophys. Res. Nishizawa M, Yamamoto M and Igarishi K. (1996). Mol. Commun., 236, 313 ± 322. Cell. Biol., 16, 6083 ± 6095.

Oncogene Cloning and expression of BACH2 transcription factor S Sasaki et al 3749 Shivdasani RA, Rosenblatt MF, Zucker-Flanklin D, Jack- Toki T, Itoh J, Kitazawa J, Arai K, Hatakeyama K, Akasaka son CW, Hunt P, Saris CJM and Orkin SH. (1995). Cell, J, Igarashi K, Nomura N, Yokoyama M, Yamamoto M 81, 695 ± 704. and Ito E. (1997). Oncogene, 14, 1901 ± 1910. Sui X, Tsuji KR, Tanaka R, Tajima S, Muraoka K, Ebina Y, Uchida T, Kinoshita T, Nagai H, Nakahata Y, Saito H, Ikebuchi K, Yasukawa K, Taga T, Kishimoto T and Hotta T and Murata Y. (1997). Blood, 90, 1403 ± 1409. Nakahata T. (1995). Proc. Natl. Acad. Sci. USA, 92, Yoshimatsu T, Tamura M, Kuriyama S and Ikenaka K. 2859 ± 2863. (1998). Hum. Gene Ther., 9, 161 ± 172. Swaroop A, Xu J, Pawar H, Jackson A, Scolnick C and Agarwal N. (1992). Proc. Natl. Acad. Sci. USA, 89, 266 ± 270.

Oncogene