It has previously been shown that mutations in the X-encoded human gene hATRX give rise to severe mental retardation associated with a-thalassemia, urogenital abnormalities, and a characteristic facial appearance (ATR-X syndrome; Gibbons et al. 1995). The predicted ATRX protein contains seven highly conserved colinear domains which classify it as a member of the helicase/ATPase superfamily. Further analysis of the entire helicase domain and the flanking regions of the protein has shown that ATRX represents a new subgroup of the SNF2-like family which contains proteins with a wide range of cellular functions including DNA recombination and repair (RAD16, RAD54, ERCC6), mitotic recombination (lodestar), and transcriptional regulation (hSNF2, Brahma, CHD1, MOT', BRG1; reviewed in Picketts et al. 1996). It has been suggested that these proteins may be variously incorporated into multicomponent complexes that utilize the energy from ATP hydrolysis to remodel chromatin and facilitate protein/DNA interactions (Kingston et al. 1996). The precise cellular role of ATRX is unknown. It seems likely, however, that ATRX has a role in transcription, since most mutations are associated with a down regulation of a globin gene expression causing a-thalassemia. There is good circumstantial evidence that, as for other members of the SNF-2 like family, ATRX also influences gene expression via an effect on chromatin. In particular, ATRX mutations act differently on the closely related a and 13 globin genes: even though both genes utilize a similar repertoire of transcription factors, it is clear that they reside in contrasting chromosomal environments and are regulated differently via their interactions with chromatin (Vyas et al. 1992; Craddock et al. 1995).

As a prelude to biochemical or experimental analysis, functionally important regions of large proteins, like ATRX (-280 kDa), may be efficiently localized by the analysis of natural mutants or by observing which regions of the molecule have been conserved during evolution. In this report, we describe a comparison of the entire open reading frames of the predicted human and mouse ATRX proteins, highlighting conserved regions of the molecule that may be critical to their normal role in vivo. The mATRX sequence was amplified by a combination of human primers and mouse primers derived from the partial mouse ATRX sequences. HPI-BP38 (Le Dourain et al. 1996) and mxnp (Gecz et al. 1994) in the public databases (Fig. 1). Comparison between the current sequence and the previously published partial sequence, mxnp (Gecz et al. 1994), shows a splice variation (see below) and 50 single nucleotide differences. Together, these corrections and the newly acquired sequence data enabled us to establish the complete ORF (7428 nt) of the mouse ATRX gene and predict the primary structure of this protein (see below). The nucleotide sequence of mATRX has been submitted to GenBank with the accession number AF026032. Xnp is presently