Downloaded from genome.cshlp.org on September 28, 2021 - Published by Cold Spring Harbor Laboratory Press RESEARCH Transcriptional Mapping in a 700-kb Region Around the DXS52 Locus in Xq28: Isolation of Six Novel Transcripts and a Novel ATPase lsoform (hPMCAS) Nina S. Heiss, Ute C. Rogner, Petra Kioschis, Bernhard Korn, and Annemarie Poustkal Deutsches Krebsforschungszentrum, Abteilung Molekulare Genomanalyse, 69120 Heidelberg, Germany The chromosomal band Xq28 has been a focus of interest in human genetics because >20 hereditary diseases have been mapped to this region. However, about two-thirds of the disease genes remain uncloned. The region around the polymorphic DXS52 locus (ST14} within Xq28 lies in the candidate regions for several as-yet-uncloned disease genes. So far, only four melanoma antigen genes (MAGE) and the human biglycan (BGN} gene, have been mapped within the 700-kb stretch around DXS52, suggesting that more genes may reside in this region. By combining exon trapping and direct cDNA selection methods, we sought to identify novel transcripts around the DXS52 locus, in addition to recovering the MAGE and BGN genes, we isolated and mapped six putative novel genes (XAPI03-XAPI08), the caltractin gene, and a gene encoding a novel Ca2§ ATPase isoform (hPMCA5). The newly isolated sequences were considered as representing parts of putative genes if they contained at least one unique exon-trap product and/or at least one expressed sequence tag {EST) from sequence data bases and if, in addition, they showed evidence of expression by RT-PCR and/or Northern blot analysis. Our data facilitated the integration of the transcript map with the physical map around the DXS52 locus. Future analysis of the novel genes as candidates for Barth syndrome (BTHS) and chondrodysplasia punctata {CDPX2) is in progress. The most distal region of the long arm of the The 7.5-Mb region that spans Xq28, exhibits human X chromosome, especially band Xq28, drastic variations in chromosomal substructure has been a focus of interest with respect to gene that is exemplified by diverging distributions of identification. The correlation of an exception- Alu and L1 repetitive elements (Rogner et al. ally high CpG island density and GC-content 1994), CpG islands (Pilia et al. 1993), and gene with a high gene density in Xq28 (Bernardi 1989; content (Palmieri et al. 1994). Because the GC- Pilia et al. 1993), has made it the most intensively content was found to peak between the color vi- studied region of the human genome. Accord- sion and G6PD loci (Pilia et al. 1993), the hunt ingly, biological and medical interest has esca- for genes in this particular region was most in- lated because of the many disease loci that are triguing. Although the region around more linked to this region. The construction and inte- proximal loci such as DXS52 has been of lesser gration of physical (Poustka et al. 1991; Dietrich interest with respect to gene identification, it has et al. 1992; Kioschis et al. 1994; Palmieri et al. nonetheless deserved continuous attention be- 1994; Rogner et al. 1994) and transcript maps cause of its interesting chromosomal structure. (Korn et al. 1992; Bione et al. 1993; Sedlacek et al. The DXS52 loci, also referred to as the ST14 se- 1993), as well as the cloning of disease genes has quence family, belong to the most polymorphic progressed rapidly and is ongoing. Of the 22 ge- loci in the human genome and consist of three netic disorders with Xq28 linkage, however, 14 reiterating copies within 60 kb (Oberl~ et al. remain uncloned. 1985). It is suggested that this polymorphism arose owing to a variation in a number of tandem 1Corresponding author. repeats (VNTRs). Alternatively, rearrangements E-MAIL [email protected]; FAX (40) 6221 423454. or point mutations may have contributed to such 478 ~ GENOME RESEARCH 6:478-491 9 by Cold Spring Harbor Laboratory Press ISSN 1054-9803/96 $5.00 Downloaded from genome.cshlp.org on September 28, 2021 - Published by Cold Spring Harbor Laboratory Press TRANSCRIPT MAP OF DXS52 REGION IN Xo28 heterogeneity (Oberl6 et al. 1985). Furthermore, a tion of the disadvantages. In this report, we de- separate polymorphic locus, DXS15, is linked to scribe the isolation and mapping of known and the ST14 sequence cluster (Oberl6 et al. 1985; Pat- novel genes in a 700-kb region around the DXS52 terson et al. 1987). Despite the close proximity of loci between the DXS1104 (Chatterjee et al. the DXS52 and DXS15 markers, recombinational 1994) and ST35.638 (Palmieri et al. 1994) events were observed between these loci, indica- markers. tive of a hot spot for recombination in this region (Brown et al. 1988; Bell et al. 1989). Although the recombinational activity between DXS52 and RESULTS DXS15 was refuted by Patterson et al. (1989) and Strategies Employed for the Isolation of Feil et al. (1990), the data of Brown et al. (1988), Transcripts together with the problems associated with the subcloning of this region into yeast artificial Two strategies were employed for the isolation of chromosomes (YACs) (Rogner at al. 1994) and coding sequences in the DXS52 region. One ap- cosmids (P. Kioschis, pers. comm.), suggests an proach involved using a combination of the com- inherent, regional instability. Conceivably, the plex exon-trap and enriched cDNA-selection instability may contribute to a higher mutation products as probes on the region-specific cDNA rate and consequently to an altered disease geno- and exon-trap sublibraries, respectively. The sec- type. Although the significance of such polymor- ond approach involved the hybridization of spe- phisms with respect to gene density and expres- cific cosmid fragments onto the same subli- sion is unknown, the DXS52 polymorphic locus braries. was found to be highly conserved among mouse The complexities of the exon-trap and cDNA- and hamster, suggestive of a small cluster of a selection products were initially compared by hy- related, expressed gene family (Oberl6 et al. 1985; bridization onto the cosmids from which they Feil et al. 1990). were derived (Fig. 1). Most of the cosmid frag- In light of the many uncloned disease-caus- ments hybridizing strongly with the exon-trap ing genes such as Barth syndrome (BTHS) (Adds probe were also positive for the cDNA-selection et al. 1993), Waisman syndrome (Gregg et al. probe (Fig. 1). Although the complexity of the 1991), Happle syndrome (Traupe et al. 1992), cDNA-selection probe was higher in that many oto-palatodigital syndrome (Biancalana et al. more cosmid fragments were positive, hybridiza- 1991), Goeminne syndrome (Zuffardi et al. tion with a CA-positive probe (poly[d(A-C)] Phar- 1982), X-linked mental retardation (MRX3; macia) showed that most of these extra hybridiz- Gedeon et al. 1991; Nordstr6m et al. 1992), and ing bands were CA-repeat containing fragments dyskeratosis congenita (Connor et al. 1986), the (Fig. 1). Despite the CA-richness of the region, candidate regions of which span the DXS52 loci, the cDNA-selection products (Fig. 1) also hybrid- a considerably small amount of genes have been ized to fragments that were unique for the cDNA- mapped to this area. To date, only the melanoma selection probe only. For completeness, we there- antigen genes (MAGE)-2, -3, -6, and -12 (Rogner fore isolated these extra hybridizing cosmid et al. 1995) and the human biglycan genes (BGN) bands and utilized them as probes on the grid- (Traupe et al. 1991; Kioschis et al. 1994) have ded, enriched cDNA and exon-trap sublibraries. been mapped around the DXS52 loci. By combin- ing exon-trapping and cDNA-selection tech- Isolation of Known Genes in the DXS52 Region niques, we endeavoured to isolate novel genes in a 700-kb region around the DXS52 loci. By implementing the cosmid fragment hybrid- Although cDNA selection (Lovett et al. 1991; ization approach, two individual cDNA clones Parimoo et al. 1991; Korn et al. 1992) and exon and two unique exon-trap products that had trapping (Buckler et al. 1991; Church et al. 1993; high homologies to the MAGE-2, -3, -6, and -12 North et al. 1993) are sufficiently sensitive to iso- genes were isolated. Upon hybridizing a 3' end late rarely expressed transcripts from large chro- conserved MAGE probe (CH089; Rogner et al. mosomal regions, both methods have their draw- 1995) onto the cosmid digests, the MAGE genes backs when used in isolation (Brennan and were found to lie on cosmids LLNL K0238 [iso- Hochgeschwender 1995). A combination of the lated from the Lawrence Livermore National two methods, however, is synergistic, facilitating Laboratory (LLNL)-chromosome X cosmid li- complementation of the advantages and elimina- brary], Qc4G10, and QclD2, from which the cos- GENOME RESEARCHO 479 Downloaded from genome.cshlp.org on September 28, 2021 - Published by Cold Spring Harbor Laboratory Press HEISS ET AL. Figure 1 Cosmid digests showing the distal part of the contig extending from the most proximal of the three DXS52 markers to ST35.638. The top, middle and bottom panels show the positive bands on hybridizing with the complex exon-trap probe, the complex cDNA-selection probe, and the CA- repeat probe respectively. All of the strongly positive exon-trap bands (top) were in common with the strongly positive cDNA-selection bands (middle) and are marked by closed triangles. (bottom) Four of these exon-trap- and cDNA-selection-positive bands were CA-positive (T). Most of the extra hybridizing bands positive with the cDNA-selection probe were CA-containing fragments (V) (middle, bottom). The map positions of the novel genes and the BGN gene are indicated at the top of the panels. Cosmids la- beled Qc were isolated by hybridization screening of gridded clone libraries of an Xq28-specific cosmid library constructed from the hamster/human cell hybrid QIZ.