CBFA2T1,a Gene Rearranged in Human Leukemia, Is a Member of a Multigene Family

GENOMICS

52, 332–341 (1998) GE985429

ARTICLE NO.

CBFA2T1, a Gene Rearranged in Human Leukemia, Is a Member of a Multigene Family Franco Calabi1 and Vania Cilli Developmental Biology Unit, Cell and Molecular Biology Division, Institute of Child Health, 30 Guilford Street, London WC1N 1EH, United Kingdom Received March 6, 1998; accepted May 28, 1998

MTG8 (HGMW-approved symbol CBFA2T1) was originally identified as one of the loci involved in the t(8; 21)(q22;q22) of acute myeloid leukemia. We characterize two human MTG8-related genes, MTGR1 and MTGR2 (HGMW-approved symbols CBFA2T2 and CBFA2T3). The former is duplicated in mouse, one locus possibly being a retroposon. Multiple MTG8-related sequences are found in several vertebrate species, from fish to mammals, albeit not in a urodele. MTGR2 maps to 16q24 and, like MTG8 and MTGR1, is close to one of three loci encoding a syntrophin (dystrophin-associated proteins). Moreover, an alternative MTGR1 promoter/5* exon is contained within the a1syntrophin locus. Thus, the two classes of genes may define novel paralogous groups. MTGR1 is expressed mainly in brain, while MTGR2 is expressed in the thymus and possibly in monocytes. Like MTG8, MTGR1 is transcribed into a number of isoforms due to alternative splicing of different 5* exons onto a common splice acceptor site. Comparison of the three predicted human MTG8-related polypeptides to their Drosophila counterpart (nervy) highlights four separate regions of sequence conservation that may correspond to distinct domains. The most NH2-terminal of these is proportionately more conserved among the human polypeptides, presumably due to specific structural/ functional constraints. © 1998 Academic Press

common translocations in acute myeloid leukemia (Miyoshi et al., 1991, 1993; Nisson et al., 1992; Erickson et al., 1992; Look, 1997). CBFA2 is a human homologue of the Drosophila melanogaster segmentation gene runt and encodes a transcriptional regulator that binds directly to DNA via a novel DNA binding domain (Daga et al., 1992; Kagoshima et al., 1993). Evidence from gene knockout experiments shows that CBFA2 is essential for definitive hemopoiesis (Wang et al., 1996; Okuda et al., 1996). The MTG8 gene product has also been hypothesized to be a transcription factor, since its sequence spans both proline-rich regions and zinc finger-like motifs and is related to the polypeptide encoded by the D. melanogaster gene nervy, a target of the homeotic gene Ultrabithorax (Feinstein et al., 1995). A CBFA2/MTG8 chimeric gene such as that resulting from the t(8;21) blocks definitive hemopoiesis in a mouse model, presumably by acting as a dominant-negative mutant (Yergeau et al., 1997). It has also been found, by in vitro cotransfection experiments, both to interfere with the regulation of wildtype CBFA2 target genes and to activate distinctive targets (Frank et al., 1995; Meyers et al., 1995; Lenny et al., 1995; Rhoades et al., 1996). Overexpression of MTG8 has been reported to score positive in a transformation assay (Wang et al., 1997). Little, however, is known of the normal function of MTG8. In this paper we characterize two human homologues of MTG8.

INTRODUCTION

MATERIALS AND METHODS

The study of chromosomal translocations in tumors has often led to the identification of novel genes, mutations in which have been linked to tumorigenesis. The CBFA2 and MTG82 genes were isolated as the partner loci in the t(8;21)(q22;q22), one of the most Sequence data from this article have been deposited with the GenBank Data Library under Accession Nos. AF052210 to AF052220. 1 To whom correspondence should be addressed. Telephone: 44171-813 8492. Fax: 44-171-831 4366. E-mail [email protected]. 2 The HGMW-approved symbols for the genes described in this paper are CBFA2T1 (MTG8), CBFA2T2 (MTGR1), and CBFA2T3 (MTGR2). 0888-7543/98 $25.00 Copyright © 1998 by Academic Press All rights of reproduction in any form reserved.

Libraries. The following libraries were used: (i) a cDNA library from the human T cell line Molt-4 (F.C., unpublished); (ii) a human adult retina cDNA library (Nathans et al., 1986), a gift from David Nathans; (iii) a cDNA library from human fetal brain (21–26 weeks; Clontech product code HLL1149X); (iv) a mouse 129 genomic library, gift of Terry Rabbitts; and (v) the human RPCI1 PAC library (Ioannou and deJong, 1996), provided by the UK HGMP Resource Centre. Oligonucleotides. The following oligonucleotides were used (in parentheses, after the sequence, the corresponding locus and the position of the 59 base on the MTGR1 sequence in Fig. 1A): 490 (GCCTGTGGAAGTGAAGATA, MTGR1, 38), 491 (TAGGAGTGAAGGACTGG, MTGR1, 138), 492 (AATTGTCCACTGTTGAGTTC, MTGR1, 407), 495 (TGCAAAGCTTGTGTTGA, MTGR1, 594), 497 (ACTGTGCTGATGCCACGT, MTGR2, 215 nt 39 of oligo 498), 498

332

333

CBFA2T1, A MEMBER OF A MULTIGENE FAMILY (GAGCTCATCACCACGGAG, MTGR2, 1344), 505 (ACGCTCCGTGGTGATGAG, MTGR2, 1364), 506 (TCGTCGACCACTTCCGAGATG, MTGR2, 841), 507 (CTTGTCGACACAGCACCT, MTGR2, 125), and 508 (AGTAGGAGCTCTGAGGAGTC, MTGR2, 618). Somatic cell hybrids for mapping. DNA from a monochromosomal human 3 mouse somatic cell hybrid panel (Kelsell et al., 1995) and the GeneBridge 4 radiation hybrid panel (Gyapay et al., 1996) was obtained from the UK HGMP Resource Centre. Probes. The following probes were used: (A) For library screening: (i) a 392-nt SacI fragment corresponding to nt 275– 667 in the MTG8b sequence (Miyoshi et al., 1993), (ii) a cloned RT-PCR product generated with oligos 490 and 495 from human brain RNA (MTGR1/ 490-5), and (iii) an RT-PCR product generated with oligos 490 and 491 on clone MTGR1/490-5. (B) For Southern blots on PAC297-G8 and its subclones: (i) a PCR product generated with oligos 490 and 495 on clone rem3 (Table 1) and (ii) a fragment spanning from nt 1415 to the stop codon in the MTG8b sequence: (C) For nuclease protection experiments: (i) a cloned mouse MTGR1 RT-PCR product (clone MusMTGR1/490-2) corresponding to nt 38 to 407 in Fig. 1A; (ii) a PstI fragment from PAC297-G8, spanning the first 176 nt in MTGR2 exon 3 (clone 297/BX4.3/P2); (iii) a 307-nt fragment from clone CFB-1 (Table 1), spanning MTGR1 exon 1a and the first 138 nt of exon 2; (iv) a 271-nt fragment from clone RFB-2 (Table 1), spanning the last 133 nt of MTGR1 exon 1b and the first 138 nt of exon 2; and (v) a cloned cytoplasmic b-actin PCR product corresponding to nt 173–335 in GB: MMACTBR (Tokunaga et al., 1986). Other materials and protocols. Human embryos (of a range spanning 54 –91 days of age) were obtained from the MRC-funded Human Embryonic Tissue Bank maintained at the Institute of Child Health, London. The embryos were collected under full ethical permission and according to the Polkinghorne guidelines (Polkinghorne et al., 1989) from social termination of pregnancies using either RU486 or surgical methods. Human thymus tissue was obtained from cardiac surgery. The human hemopoietic cell lines used to test MTGR2 expression are characterized in Drexler and Minowada (1989). DNA and RNA manipulations were performed according to standard techniques (Sambrook et al., 1989).

RESULTS

Identification and Sequence Analysis of the Human MTGR1 and MTGR2 Loci Comparison of the MTG8 coding sequence to the EST database (August 1996) by the program BLAST (Altschul et al., 1990) revealed a number of hits that could be sorted into two classes: the first consisted of sequences 100% identical, the second of three sequences with 66 –76% homology. Two of the latter (H83692 and H83335) correspond to opposite ends of the same clone (ID 249228). The third (R13314, clone ID 27911) partly overlaps H83335, extending it in a 39 direction. It seemed therefore likely that the three ESTs defined a contig corresponding to a novel locus. To complete the sequence, and to verify independently the EST data, human and mouse genomic libraries and two human cDNA libraries were screened with an MTG8 probe. By restriction mapping and partial sequencing, a number of clones were found to be distinct from, though highly related to, MTG8 (Table 1). Altogether, they identify two loci. One of them corresponds to a recently published cDNA, named MTGR1 (for MTG8-related gene 1) (Kitabayashi et al., 1998). By analogy, we propose the name MTGR2 for the other.

TABLE 1 Summary of the DNA Clones Isolated and Characterized for the Present Work (cMTGR1 Refers to a Putative Retroposon) Clone

Species

Source

Locus

NHcDM3 rem3 CFB-1 CFB-11 CFB-12 RFB-2 RFB-6 RFB-11 RFB-4 ESMM1 ESMM7 ESMM3 ESMM8 297-G8

Human Human Human Human Human Human Human Human Human Mouse Mouse Mouse Mouse Human

Molt-4 cDNA Retina cDNA Fetal brain cDNA Fetal brain cDNA Fetal brain cDNA Fetal brain RT-PCR Fetal brain RT-PCR Fetal brain RT-PCR Fetal brain RT-PCR Genomic Genomic Genomic Genomic Genomic

MTGR1 MTGR1 MTGR1 (a) MTGR1 (c) MTGR1 (d) MTGR1 (b) MTGR1 (b) MTGR1 (b) MTGR1 (e) MTGR1 MTGR1 cMTGR1 cMTGR1 MTGR2

The human MTGR1 cDNA clones from the T cell line Molt-4 and from retina were sequenced. Further sequences were obtained from cDNA clones mapping to the 59 end of the transcripts (see below). A composite of the determined sequences is shown in Fig. 1A, starting from a position matching the beginning of MTG8 exon 2. Comparison of human cDNA to mouse genomic clones suggests that this position corresponds to the beginning of MTGR1 exon 2 (Fig. 1B) and that it is a common splicing acceptor for a number of alternative 59 exons. Starting at nt 3, an open reading frame extends through to nt 1751. In the retina cDNA, this is interrupted, at nt 1263, by a stretch of 121 nt harboring an Alu repeat that presumably corresponds to an unspliced intron. Two sets of mouse genomic MTGR1 clones that must correspond to separate loci were identified. Sequencing revealed a region at both loci that is highly related to the MTG8 and MTGR1 cDNAs. However, while the sequence from one locus (represented by clones ESMM1 and 7) matches a 242-nt-long stretch in the human MTGR1 cDNA that corresponds precisely to MTG8 exon 3 (F.C., unpublished), the other locus (clone ESMM3 and 8) matches nearly exactly a human cDNA (clone CFB-1) spanning an alternative 59 exon (MTGR1a, see below) and sequences corresponding both to MTG8 exons 2 and 3, without intervening introns. Thus the latter locus may represent a retroposon (Leibmosch and Seifarth, 1995). Human PAC 297-G8 contains a gene highly related to, yet different from, both MTG8 and MTGR1. Restriction fragments hybridizing to probes from different regions of the MTG8/MTGR1 cDNAs were subcloned and sequenced. These yielded four distinct contigs containing regions of homology to nt 311–552, 642– 824, 1433–1623, and 1624 –1778 in the MTG8b cDNA sequence. The former two correspond respectively to MTG8 exons 3 and 5 (F.C., unpublished). In all cases, the boundaries of the regions of homology match ca-

334

CALABI AND CILLI

CBFA2T1, A MEMBER OF A MULTIGENE FAMILY

335

FIG. 1. (A) Comparison of the coding sequences of MTGR1 and MTGR2. The first position shown of the former corresponds to nt 1 in exon 2 (i.e., the first exon common to the different isoforms). Nucleotide identities are marked by a colon. A downward-pointing arrowhead marks the position of an Alu-like repeat within the retina MTGR1 cDNA, presumably corresponding to an intron. Upward-pointing arrowheads mark the positions of exon boundaries in MTGR2. (B) Comparison of MTGR1 sequences from human cDNA and mouse genomic clones. A dash in the latter represents identity to human. Slashes indicate sequence divergence (due to introns at locus ESMM7). Nucleotide differences resulting in amino acid changes in the mouse are underlined.

FIG. 2. Block alignment of the three predicted MTG8-related polypeptides and of their Drosophila homologue (encoded by the gene nervy) as generated by the program GCG PILE-UP. The numbering refers to the sequence of MTG8b (Miyoshi et al., 1993). Dots indicate gaps introduced to maximize alignment. Residues shared by the three human sequences are boxed, those shared with the Drosophila protein are shaded. Four clusters of homology (NHR1– 4) (Kitabayashi et al., 1998) are indicated by bars underneath the sequence. The thicker segment within NHR1 corresponds to the region of homology to TAFs (Tanese et al., 1996; Mengus et al., 1997).

336

CALABI AND CILLI

TABLE 2 Percentage Identity in Pairwise Comparisons of MTG8 Family Members

MTG8 MTGR1 MTGR2

MTG8

MTGR1

MTGR2

— 86 92

72 — 85

75 69 —

Note. Bold numbers, codons (370); italics, nucleotides (1137).

nonical splice sites and therefore must be presumed to correspond to bona fide exons. Additional sequences were derived from the analyses of cloned RT-PCR products obtained from human thymus RNA with locusspecific primers based on the genomic sequence (oligos 505, 506, 507, and 508). By analogy to MTG8 and MTGR1, the determined MTGR2 sequence must correspond to nearly two-thirds of the coding sequence. An alignment of MTGR2 to MTGR1 is presented in Fig. 1A, and a block alignment of the three predicted human MTG8 family polypeptides, as well as their D. melanogaster homologue (encoded by the gene nervy; Feinstein et al., 1995), in Fig. 2. Pairwise percentage identities over aligned bases and codons for the three human genes are shown in Table 2. MTGR1 and 2 are more related to MTG8 than to each other. Chromosomal Mapping of MTGR2 EST H83335 had been previously mapped to human chromosome 20q11 (Banfi et al., 1996). The chromosomal location of MTGR2 was determined by analysis of two panels of somatic cell hybrids. Segregation of the MTGR2 locus was scored by a locus-specific PCR based on a pair of primers (oligos 497 and 498) mapping respectively to the exon encoding the sequence amino terminal to the zinc finger and to the following intron. Analysis of a set of monochromosomal human 3 mouse hybrids (Kelsell et al., 1995) yielded a positive result only with the line harboring human chromosome 16. Parallel testing of the GeneBridge 4 radiation hybrid panel (Gyapay et al., 1996) placed MTGR2 17.20 cR from D16S2976 at a LOD . 3.0, establishing unambiguous linkage to distal 16q24 (Table 3 and Fig. 3). Thus, MTGR2 maps relatively close to CBFB, the gene encoding the b subunit of CBF (core binding factor) (Liu et al., 1993).

FIG. 3. Framework map of human chromosome 16, with the position of reference GeneBridge 4 hybrids (Gyapay et al., 1996) and the determined location of MTGR2.

Phylogenetic Conservation of the MTG8 Family To investigate the existence of further members of the MTG8 family, and to determine its extent of phylogenetic conservation, Southern blotting analysis was carried out on DNA from a variety of vertebrate species with a short (284 nt) probe corresponding essentially to a single MTG8 exon (exon 3). Three fragments could be detected in human on both EcoRI and HindIII digests and two to four in mouse, dog, chicken, and fish (data not shown). The three human fragments can be ac-

TABLE 3 Results of the Screening of GeneBridge 4 Hybrids for the Presence of the Human MTGR2 Locus Positive 4A4, 4D7, 4DD5, 4DD8, 4F7, 4H9, 4H12, 4M5, 4O5, 4Q2, 4R2, 4R6, 4S12, 44T10, 4V7, 4Z6, 4Z9, 4Z12 Negative 4AA5, 4AA7, 4B3, 4BB1, 4BB6, 4BB8, 4C3, 4C11, 4CC8, 4D1, 4DD2, 4E2, 4E4, 4E6, 4E11, 4F6, 4F13, 4G1, 4G5, 4G6, 4G7, 4G11, 4H1, 4H8, 4I1, 4I4, 4J2, 4J5, 4J9, 4K5, 4K7, 4K9, 4K12, 4L3, 4L4, 4L6, 4M4, 4N3, 4N5, 4N6, 4N7, 4N12, 4P2, 4P9, 4P11, 4Q4, 4R1, 4R2, 4R3, 4R5, 4R10, 4S3, 4S6, 4S10, 4T3, 4T4, 4T11, 4U1, 4V2, 4V3, 4V8, 4W1, 4Y4, 4Y8, 4Y9, 4Z5, 4Z11 Not tested 4A5, 44B2, 4B9, 4BB10, 4BB12, 4K8, 4O10, 4R12, 4U3

CBFA2T1, A MEMBER OF A MULTIGENE FAMILY

337

FIG. 4. (A) Nuclease protection experiment on RNA (;10 mg/sample) from the indicated mouse tissues with probes spanning mouse MTGR1 exons 2 and 3 and cytoplasmic b-actin. tRNA was used as a negative control. Open arrowheads mark the position of residual undigested probe, filled arrowheads of protected fragments (top symbols, MTGR1; bottom symbols, actin). The positions of molecular markers are indicated on the left. (B) Nuclease protection experiment on RNA (;20 mg/sample) from human thymus and from the indicated human hemopoietic cell lines with a probe spanning the first 171 nt of human MTGR2 exon 3. tRNA was used as a negative control. An open arrowhead marks the position of residual undigested probe, a filled arrowhead of protected fragments. The positions of molecular markers are indicated on the left.

counted for by MTG8, MTGR1, and MTGR2, and it is therefore likely that all members of the human MTG8 family have now been identified. Curiously, newt DNA scored negative in this test, despite giving the expected signal with a homologous FGF-R2 probe. The result is likely to be due to whole gene deletion, since it was confirmed with a much larger probe spanning the first ;1.4 kb of the MTG8b cDNA. Pattern of Expression of MTGR1 and MTGR2 To define the pattern of expression of MTGR1, a mouse probe spanning exons 2 and 3 (generated by RT-PCR with oligos 490 and 492 from mouse brain) was used in a nuclease protection assay on RNA from a panel of mouse tissues. On the basis of the available sequence, this region shows the highest divergence among human MTG8 family members. As shown in Fig. 4A, a fragment of the expected size, showing presence of transcripts, was found in all tissues examined. However, expression was low, except in brain from adult and late fetal stage (i.e., E16). A parallel investigation of MTGR2 expression was performed with a human probe spanning the first 176 nt of exon 3 on a panel of RNA from hemopoietic cell lines and thymus (Fig. 4B). A clearly positive signal was obtained from the latter, as well as from tumor lines corresponding to common thymocytes (JM and CCRF-CEM) from the Burkitt lymphoma DAUDI and from the monocytic line U937. Northern blotting analysis of poly(A)1 RNA from human hemopoietic tissues with a probe from the

MTGR1 39 UTR reveals two major bands at ;7.5 and 6 kb and two minor ones at ;3 and 1.5 kb (data not shown). The two largest species are present at significantly higher levels in lymphoid organs than in bone marrow, fetal liver, or peripheral blood, suggesting that MTGR1 is transcribed in the lymphoid rather than in the myeloid lineages. Alternative 59 Ends of MTGR Transcripts To define the 59 end of the MTGR1 mRNA(s), two different strategies were resorted to. In the first, a cDNA library from human fetal brain was screened with a probe corresponding to MTGR1 exon 2 (generated by PCR with oligos 490 and 491 on clone MTGR1/ 490-5). Three non-MTG8 clones (CFB-1, -11, and -12) were isolated and partially sequenced. Each showed a different sequence spliced onto the 59 end of MTGR1 exon 2. In a separate approach, 59-terminal cDNA fragments were amplified following a variant of published procedures (Frohman, 1993). cDNA synthesis was primed on human fetal brain RNA with an oligonucleotide mapping immediately downstream of exon 3 (oligo 492). After tailing with dA and conversion to doublestranded form with a poly(dT) terminating primer, the cDNA was used as template in a PCR with primers mapping within MTGR1 exon 2 (oligo 491) and 59 of the oligo(dT) stretch. Upon cloning and screening with a probe for MTGR1 exon 2, six positive clones were recovered and sequenced. Two of these (RFB-1 and -12) contained just MTGR1 exon 2, three (RFB-2, -6, and

338

CALABI AND CILLI

FIG. 5. Nuclease protection experiment on RNA (;10 mg/sample) from human fetal brain and from the T cell line Molt-4 with probes spanning either MTGR1 exon 1a or 1b and the neighboring 138 nt of exon 2. tRNA was used as a negative control. Open arrowheads mark the position of residual undigested probe, filled arrowheads of protected fragments. Of the latter, the top ones correspond to exon 1a or 1b spliced onto exon 2, the bottom ones to exon 2 only. The positions of molecular markers are indicated on the right.

-11) matched a recent entry in dbEST (AA635096, submitted October 21, 1997), and the remaining one (RFB-4) showed a previously known sequence spliced onto MTGR1 exon 2 (see below). Probes spanning each alternative 59 end as well as the neighboring 138 nt of exon 2 were made from clones CFB-1, -11, and -12 and RFB-2 and tested for expression in RNA from human fetal brain in a nuclease protection assay. While each probe yielded a fragment of 138 nt, consistent with partial protection by coexisting alternative isoform(s) splicing onto exon 2, an additional full-length fragment was obtained only in the case of CFB-1 and RFB-2 (Fig. 5). We conclude that the corresponding transcripts must be the most abundant ones and designate the respective alternative 59 exons as MTGR1a and MTGR1b. Similar experiments on RNA from a human thymoma cell line show expression of MTGR1a only, as indicated by the absence both of a full-length protected fragment with the MTGR1b probe and of an exon 2 protected fragment with the MTGR1a

probe (Fig. 5). Thus expression of MTGR1b may be brain-specific. The sequence of CFB-1 (i.e., MTGR1a) and RFB-11 (the longest of the four clones corresponding to MTGR1b) is shown in Fig. 6 from nt 1 to the junction with the common MTGR1 exons. A sequence homologous to the former is also found in mouse clone ESMM8, a putative retroposon. Both human and mouse MTGR1a contain an open reading frame (ORF) in register with the main MTGR1 ORF and spanning most of their length, although the mouse sequence lacks a dC:dG-rich stretch of 39 nt. The only methionine codon is found at position 136 in the human sequence and is conserved in mouse. Its context would conform, albeit not optimally, to the criteria for translational starts (Kozak, 1996), i.e., a purine at position 24 and a G at position 14. In MTGR1b, a reading frame in register with the main MTGR1 ORF spans the last 100 nt. However, it does not contain any ATG and therefore the corresponding transcript, if translated, would encode a polypeptide starting from the methionine codon at position 28 in exon 2 (Fig. 1A). This codon is conserved in MTG8 and would also match the Kozak consensus (Kozak, 1996). Clone RFB-4, derived from PCR amplification of the MTGR1 cDNA 59 end, contains exon 2 spliced to a stretch of 96 nt identical to nt 1175–1270 from the noncoding strand of human a1-syntrophin. This gene has been recently mapped to 20q11.2, i.e., close to the chromosomal location of MTGR1 (Ahn et al., 1996). It is therefore likely that alternative MTGR1 promoter and exon(s) are contained within the a1-syntrophin locus. DISCUSSION

Sequence homology to MTG8 defines a small gene family widely conserved across evolution. We characterize in some detail two human loci, MTGR1 and MTGR2. Both are transcribed in a distinctive pattern and have the potential to encode functional polypeptides, based on the homology with MTG8. MTGR1 has at least two homologues in the mouse genome, one of which shows features of a retroposon. Multiple MTG8 family members are present in representatives of vertebrate species from fish to mammals, albeit not in a urodele. This suggests that gene duplication of the ancestral MTG8 sequence predated vertebrate radiation and that different, yet closely related MTG8-like genes have fundamental functions in the physiology of most vertebrate organisms. The three human MTG8 family members map to separate chromosomes, MTG8 to 8q22, MTGR1 to 20q11, and MTGR2 to 16q24. Each one is relatively close to one of three loci encoding another family of related polypeptides, the syntrophins, so called because of their association with dystrophin, the product of the Duchenne muscular dystrophy locus (Ahn et al., 1994, 1996). We have also found that an alternative

CBFA2T1, A MEMBER OF A MULTIGENE FAMILY

339

FIG. 6. Sequence of MTGR1a and 1b exons. The former is aligned with the sequence from mouse clone ESMM8, presumably corresponding to a retroposon. Dots indicate gaps introduced to maximize alignment, dashes indicate sequence identity. A putative initiator methionine is in bold. The first 2 nt of MTGR1 exon 2 at the end of each sequence are in italics.

MTGR1 transcript contains a 59 stretch derived from the noncoding strand of a1-syntrophin, suggesting that the two genes are partly overlapping and in opposite orientation. Thus, MTG8-related and syntrophin genes likely define a novel group of paralogous genes. The chromosomal location of MTGR2 suggests a second significant association with the map position of a chromosomal breakpoint in acute myeloid leukemia, t(16;21)(q24;q22). The reciprocal breakpoint in this translocation (21q22) falls within the CBFA2 locus (Shimada et al., 1997). It is tempting to speculate that, like MTG8, albeit at a lower frequency, MTGR2 may also be involved in a reciprocal translocation with CBFA2 leading to acute myeloid leukemia. Like MTG8, the MTGR1 gene is transcribed into a number of isoforms, due to alternative promoters/59 exons. We have identified five of the latter, and yet a different one has been found in a recently reported cDNA clone (Kitabayashi et al., 1998). Of the four variants tested, only two were found to be transcribed in significant amounts in brain, i.e., the tissue in which the locus is mainly expressed. One of these may be brain-specific. While four of the alternative 59 exons lack a potential start codon, in the remaining two this does not fall within the optimal context for translational initiation (e.g., they lack an A at position 24) (Kozak, 1996). It is thus possible that most of MTGR1 mRNA translation starts from a conserved AUG within exon 2. By analogy, a similar situation may also occur at the other MTG8-related loci. It has been previously observed that homologies between MTG8 and Drosophila nervy cluster into four regions, which have been named NHR1–4 (for nervy homology region) (Kitabayashi et al., 1998). This clustering is even more apparent in a block alignment of the

two sequences with MTGR1 and MTGR2 (Fig. 2). Of 370 codons that can be compared, 113 are shared by all four genes, and 102 of these fall within NHR1– 4. NHR1 includes a 60-amino-acid region related to two classes of TFIID-associated factors, dTAFII110/ hTAFII135 and hTAFII105 (Tanese et al., 1996; Mengus et al., 1997). Remarkably, differences among the three predicted MTG8-like polypeptides occur only at four positions within this stretch and consist in all cases of conservative replacements. NHR2, spanning just 22 amino acids, has been reported to be required both for transcriptional regulation and for control of proliferation/ differentiation by a chimeric CBFA2/MTG8 gene (Frank et al., 1995; Meyers et al., 1995; Lenny et al., 1995; Rhoades et al., 1996; Kitabayashi et al., 1998). It has also been recently shown to mediate a physical interaction between CBFA2/MTG8 and MTGR1 products (Kitabayashi et al., 1998). NHR4 contains a pattern of cysteine and histidine residues suggestive of zinc finger motifs and is weakly related to apoptosis-induced genes (Miyoshi et al., 1993). However, it is not required for any of the known effects of CBFA2/MTG8, while preliminary experiments have failed to show specific binding to DNA (F.C., unpublished). The three human MTG8-like polypeptides are most related over NHR4 and NHR1 (over 90% identities), less over NHR2 (82%), and even less over NHR3 (68%). In contrast, the proportion of residues shared by all human and Drosophila proteins is highest in NHR4 (71%), ranging between 46 and 54% for the other three regions. Thus, the human polypeptides are proportionately more related to each other over NHR1. This may reflect an additional degree of structural/functional constraint operating on this domain in humans.

340

CALABI AND CILLI

ACKNOWLEDGMENTS We are particularly grateful to the UK HGMP RC for providing library filters and hybrid panel DNAs, to Dr. Terry Rabbitts and David Nathans for the gift of libraries, to Dr. Jude Fitzgibbon for providing chicken and dog DNA, to the staff running the MRCfunded Human Embryonic Bank at ICH, and to Sally Walder for preparations of human fetal brain. This work was supported by MRC Program Grant PG9311737.

REFERENCES Ahn, A. H., Freener, C. A., Gussoni, E., Yoshida, M., Ozawa, E., and Kunkel, L. M. (1996). The 3 human syntrophin genes are expressed in diverse tissues, have distinct chromosomal locations, and each bind to dystrophin and its relatives. J. Biol. Chem. 271: 2724 –2730. Ahn, A. H., Yoshida, M., Anderson, M. S., Feener, C. A., Selig, S., Hagiwara, Y., Ozawa, E., and Kunkel, L. M. (1994). Cloning of human basic a1, a distinct 59-kDa dystrophin-associated protein encoded on chromosome 8q23–24. Proc. Natl. Acad. Sci. USA 91: 4446 – 4450. Altschul, S. F., Gish, W., Miller, W., Myers, E. W., and Lipman, D. J. (1990). Basic local alignment search tool. J. Mol. Biol. 215: 403– 410. Banfi, S., Borsani, G., Rossi, E., Bernard, L., Guffanti, A., Rubboli, F., Marchitiello, A., Giglio, S., Coluccia, E., Zollo, M., Zuffardi, O., and Ballabio, A. (1996). Identification and mapping of human cDNAs homologous to Drosophila mutant genes through EST database searching. Nat. Genet. 13: 167–174. Daga, A., Tighe, J. E., and Calabi, F. (1992). Leukemia—Drosophila homology. Nature 356: 484 – 484. Drexler, H. G., and Minowada, J. (1989). Morphological, immunophenotypical and isoenzymatic profiles of human leukemia cells and derived T-cell lines. Hematol. Oncol. 7: 115–125. Erickson, P., Gao, J., Chang, K. S., Look, T., Whisenant, E., Raimondi, S., Lasher, R., Trujillo, J., Rowley, J., and Drabkin, H. (1992). Identification of breakpoints in t(8;21) acute myelogenous leukemia and isolation of a fusion transcript, AML1/ETO, with similarity to Drosophila segmentation gene, runt. Blood 80: 1825– 1831. Feinstein, P. G., Kornfeld, K., Hogness, D. S., and Mann, R. S. (1995). Identification of homeotic target genes in Drosophilamelanogaster including nervy, a proto-oncogene homolog. Genetics 240: 573–586. Frank, R., Zhang, J., Uchida, H., Meyers, S., Hiebert, S. W., and Nimer, S. D. (1995). The AML1/ETO fusion protein blocks transactivation of the GM-CSF promoter by AML1b. Oncogene 11: 2667–2674. Frohman, M. A. (1993). Rapid amplification of complementary-DNA ends for generation of full-length complementary DNAs—Thermal RACE. Methods Enzymol. 218: 340 –356. Gyapay, G., Schmitt, K., Fizames, C., Jones, H., Vegaczarny, N., Spillett, D., Muselet, D., Prudhomme, J. F., Dib, C., Auffray, C., Morissette, J., Weissenbach, J., and Goodfellow, P. N. (1996). A radiation hybrid map of the human genome. Hum. Mol. Genet. 5: 339 –346. Ioannou, P. A., and deJong, P. J. (1996). Construction of bacterial artificial chromosome libraries using the modified P1 (PAC) system. In “Current Protocols in Human Genetics,” Wiley, New York. Kagoshima, H., Shigesada, K., Satake, M., Ito, Y., Miyoshi, H., Ohki, M., Pepling, M., and Gergen, P. (1993). The runt domain identifies a new family of heteromeric transcriptional regulators. Trends Genet. 9: 338 –341. Kelsell, D. P., Rooke, L., Warne, D., Bouzyk, M., Cullin, L., Cox, S., West, L., Povey, S., and Spurr, N. K. (1995). Development of a panel of monochromosomal somatic-cell hybrids for rapid genemapping. Ann. Hum. Genet. 59: 233–241.

Kitabayashi, I., Ida, K., Morohoshi, F., Yokoyama, A., Mitsuhashi, N., Shimizu, K., Nomura, N., Hayashi, Y., and Ohki, M. (1998). The AML1–MTG8 leukemic fusion protein forms a complex with a novel member of the MTG8(ETO/CDR) family, MTGR1. Mol. Cell. Biol. 18: 846 – 858. Kozak, M. (1996). Interpreting cDNA sequences—Some insights from studies on translation. Mamm. Genome 7: 563–574. Leibmosch, C., and Seifarth, W. (1995). Evolution and biological significance of human retroelements. Virus Genes 11: 133– 145. Lenny, N., Meyers, S., and Hiebert, S. W. (1995). Functional domains of the t(8;21) fusion protein, AML-1/ETO. Oncogene 11: 1761– 1769. Liu, P., Tarle, S. A., Hajra, A., Claxton, D. F., Marlton, P., Freedman, M., Siciliano, M. J., and Collins, F. S. (1993). Fusion between transcription factor CBFb/PEBP2b and a myosin heavy-chain in acute myeloid leukemia. Science 261: 1041–1044. Look, A. T. (1997). Oncogenic transcription factors in the human acute leukemias. Science 278: 1059 –1064. Mengus, G., May, M., Carre, L., Chambon, P., and Davidson, I. (1997). Human TAF(II)135 potentiates transcriptional activation by the AF-2s of the retinoic acid, vitamin D3, and thyroid hormone receptors in mammalian cells. Genes Dev. 11: 1381–1395. Meyers, S., Lenny, N., and Hiebert, S. W. (1995). The t(8;21)-fusion protein interferes with AML1b-dependent transcriptional activation. Mol. Cell. Biol. 15: 1974 –1982. Miyoshi, H., Kozu, T., Shimizu, K., Enomoto, K., Maseki, N., Kaneko, Y., Kamada, N., and Ohki, M. (1993). The t(8;21) translocation in acute myeloid leukemia results in production of an AML1-MTG8 fusion transcript. EMBO J. 12: 2715–2721. Miyoshi, H., Shimizu, K., Kozu, T., Maseki, N., Kaneko, Y., and Ohki, M. (1991). t(8;21) breakpoints on chromosome 21 in acute myeloid leukemia are clustered within a limited region of a single gene, AML1. Proc. Natl. Acad. Sci. USA 88: 10431–10434. Nathans, J., Thomas, D., and Hogness, D. S. (1986). Molecular genetics of human color vision—The genes encoding blue, green, and red pigments. Science 232: 193–202. Nisson, P. E., Watkins, P. C., and Sacchi, N. (1992). Transcriptionally active chimeric gene derived from the fusion of the AML1 gene and a novel gene on chromosome 8 in t(8;21) leukemic cells. Cancer Genet. Cytogenet. 63: 81– 88. Okuda, T., Vandeursen, J., Hiebert, S. W., Grosveld, G., and Downing, J. R. (1996). AML1, the target of multiple chromosomal translocations in human leukemia, is essential for normal fetal liver hematopoiesis. Cell 84: 321–330. Polkinghorne, J., Hoffenberg, R., Kennedy, I., and Macintyre, S. (1989). “Review of the Guidance on the Research Use of Fetuses and Fetal Material,” H. M. Stationery Office, London. Rhoades, K. L., Hetherington, C. J., Rowley, J. D., Hiebert, S. W., Nucifora, G., Tenen, D. G., and Zhang, D. E. (1996). Synergistic up-regulation of the myeloid-specific promoter for the macrophagecolony-stimulating factor-receptor by AML1 and the t(8;21) fusion protein may contribute to leukemogenesis. Proc. Natl. Acad. Sci. USA 93: 11895–11900. Sambrook, J., Frisch, E. F., and Maniatis, T. (1989). “Molecular Cloning: A Laboratory Manual,” Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. Shi, D. L., Launay, C., Fromentoux, V., Feige, J. J., and Boucaut, J. C. (1994). Expression of fibroblast growth factor receptor-2 splice variants is developmentally and tissue-specifically regulated in the amphibian embryo. Dev. Biol. 164: 173–182. Shimada, M., Ohtsuka, E., Shimizu, T., Matsumoto, T., Matsushita, K., Tanimoto, F., and Kajii, T. (1997). A recurrent translocation, t(16;21)(q24;q22), associated with acute myelogenous leukemia: Identification by fluorescence in situ hybridization. Cancer Genet. Cytogenet. 96: 102–105.

CBFA2T1, A MEMBER OF A MULTIGENE FAMILY Tanese, N., Saluja, D., Vassallo, M. F., Chen, J. L., and Admon, A. (1996). Molecular cloning and analysis of 2 subunits of the human TFIID complex—HTAF(II)130 and HTAF(II)100. Proc. Natl. Acad. Sci. USA 93: 13611–13616. Tokunaga, K., Taniguchi, H., Yoda, K., Shimizu, M., and Sakiyama, S. (1986). Nucleotide sequence of a full-length cDNA for mouse cytoskeletal b-actin mRNA. Nucleic Acids Res. 14: 2829 –2829. Wang, J. X., Wang, M., and Liu, J. M. (1997). Transformation properties of the ETO gene, fusion partner in t(8;21) leukemias. Cancer Res. 57: 2951–2955.

341

Wang, Q., Stacy, T., Binder, M., Marinpadilla, M., Sharpe, A. H., and Speck, N. A. (1996). Disruption of the CBFA2 gene causes necrosis and hemorrhaging in the central nervous system and blocks definitive hematopoiesis. Proc. Natl. Acad. Sci. USA 93: 3444 –3449. Yergeau, D. A., Hetherington, C. J., Wang, Q., Zhang, P., Sharpe, A. H., Binder, M., MarinPadilla, M., Tenen, D. G., Speck, N. A., and Zhang, D. E. (1997). Embryonic lethality and impairment of haematopoiesis in mice heterozygous for an AML1–ETO fusion gene. Nat. Genet. 15: 303–306.