Subcellular localization of the oncoprotein MTG8 (CDR/ETO) in neural cells N Sacchi 1,2 , F Tamanini 3 , R Willemsen 3 , S Denis-Donini 2,4 , S Campiglio 2 and AT Hoogeveen 3 1 Department of Biology and Genetics, School of Medicine and 2 Department of Biology, School of Biological Sciences, University of Milan, Milan, Italy; 3 Department of Cli nical Genetics, Medical School, Erasmus University, Rotterdam, The Netherlands; 4 CNR Center of Cytopharmacology, Milan, Italy The t(8;21) translocation associated with acute myeloid leukemia (AML) disrupts two genes, the AML1 gene also known as the core binding factor A2 (CBFA2) on chromosome 21, and a gene on chromosome 8, hereafter referred to as MTG8, but also known as CDR and ETO. Extensive information is available on AML1, a member of the CBF family of transcription factors, containing a highly conserved domain, the runt box, of the Drosophila segmentation gene runt. This gene is essential for the hematopoietic development and is found disrupted in several leukemias. In contrast, the function of the MTG8 gene is poorly understood. The predicted protein sequence shows two unusual, putative zinc-®ngers, three proline-rich regions, a PEST domain and several phosphorylation sites. In addition, we found a region encompassing aa 443 ± 514 predicted to have a signi®cant propensity to form coiled coil structures. MTG8 displays a high degree of similarity with nervy, a homeotic target gene of Drosophila, expressed in the nervous system. Human and mouse wild-type MTG8 are also highly expressed in brain relative to other tissues. For these reasons, we set out to investigate the expression and subcellular localization of the MTG8 protein in neural cells. Immunohistochemical experiments in a 12.5-day- old mouse embryo clearly showed that the protein was expressed in the neural cells of the developing brain and the spinal cord. In primary cultures of hippocampal neurons of 2 ± 3 day-old mice, MTG8 was found in the nucleus, in the cytoplasm and as ®ne granules in the neurites. Cytoplasmic localization of the protein was observed in Purkinje cells of both human and mouse cerebellum. The molecular mass of MTG8 in total human and mouse brain was analysed by immunoblotting and determined to be between 70 and 90 kDa. Isoforms with the same molecular mass were demonstrated in synaptosomes isolated from mouse forebrain. The evidence of MTG8 in the nucleus and cytoplasm of neural cells suggests a speci®c mechanism regulating the subcellular localization of the protein. Keywords: MTG8 protein; neural cells; cellular localization Introduction The (8;21) translocation associated with acute myeloid leukemia is one of the chromosome rearrangements involving the core binding factor (CBF), also known as the polyomavirus-enhancer-binding-protein 2 (PEBP2) (for reviews see Speck and Stacy, 1995; Nucifora and Rowley, 1995; Meyers and Hiebert, 1995; Ito, 1996). CBF is a heterodimeric transcription factor composed of two subunits, a and b (Kagoshima et al., 1993). The a subunit, with DNA binding properties, contains a region of homology with the runt gene of Drosophila (Daga et al., 1992) and the b subunit stabilizes the binding of the a subunit to DNA. Recent work has demonstrated that both subunits are essential for a correct hematopoiesis (Okuda et al., 1996; Wang et al., 1996a,b; Castilla et al., 1996; Sasaki et al., 1996; Yergeau et al., 1997; Niki et al., 1997) and that their disruption may be crucial for leukemogenesis. Other genes are found involved together with the CBF genes in dierent leukemia subtypes. Speci®cally, in the t(8;21) positive acute myelogenous leukemia, AML1 (Miyoshi et al., 1991) encoding for a CBFa subunit was found fused to an unknown gene on chromosome 8, independently isolated in dierent laboratories named ETO (Erickson et al., 1992), CDR (Nisson et al., 1992) and MTG8 (Miyoshi et al., 1993). A clue to the possible function of the gene came with the cloning of two full-length brain cDNAs, MTG8a and b, contain- ing almost identical open reading frames (ORFs). Since our work is based on the information provided by these ORFs, hereafter we will use for this protein the designation MTG8. The predicted ORFs of MTG8a and b encode two proteins of 577 and 604 amino acids respectively. MTG8 contains a few relevant motifs suggesting a role as a transcription factor (Miyoshi et al., 1993). MTG8 shows a high degree of sequence similarity to nervy, a homeotic target gene of Drosophila that is expressed in the nervous system (Feinstein et al., 1995), it is a remarkably conserved between human and mouse (Erickson et al., 1994) and its expression seems to be very high in brain relative to other tissues (Miyoshi et al., 1993). These overall observations point to a possible role of the wild-type MTG8 in the dierentiation/development of the neural system. Analysis of the transcripts of t(8;21) leukemic cells demonstrated that MTG8 was fused either in-frame (Erickson et al., 1992; Nisson et al., 1992; Miyoshi et al., 1993; Nucifora et al., 1993; Era et al., 1995a,b) or out-of-frame (Tighe and Calabi, 1994; van de Locht et al., 1994; Saunders et al., 1996) with the 5' region of AML1. Thus, in leukemic cells MTG8 is expressed as a part of the chimeric AML1/MTG8 protein(s). Chimeric proteins were detected in a t(8;21) positive cell line and patients' blasts (Sacchi et al., 1996; Erickson et al., 1996). Since some t(8;21) negative leukemic cell types also express high levels of wild-type MTG8 transcripts Correspondence: N Sacchi Received 10 July 1997; revised 17 December 1997; accepted 19 December 1997 Oncogene (1998) 16, 2609 ± 2615  1998 Stockton Press All rights reserved 0950 ± 9232/98 $12.00 http://www.stockton-press.co.uk/onc