Continuum of overlapping clones spanning the entire human chromosome 21q

Construction and validation of yeast artificial chromosome contig maps by RecA-assisted restriction endonuclease cleavage

RecA-assisted restriction endonuclease (RARE) cleavage is an "Achilles' heel" approach to restriction mapping whereby a RecA-protein-oligodeoxynucleotide complex protects an individual restriction site from methylation, thus limiting subsequent digestion to a single, predetermined site. We have used RARE cleavage to cut yeast artificial chromosomes (YACs) at specific EcoRI sites located within or adjacent to sequence-tagged sites (STSs). Each cleavage reaction produces two YAC fragments whose sizes are a direct measure of the position of the STS in the YAC. In this fashion, we have positioned 45 STSs within a contig of 19 independent YACs and constructed a detailed RARE-cleavage map that represents 8.4 Mbp of human chromosome 6p21.3-22. By comparing maps of overlapping YACs, we were able to detect seven internal deletions that ranged from approximately 75 kbp to approximately 1 Mbp in size. Thirteen pairs of EcoRI sites were targeted for double RARE cleavage in uncloned total human DNA. The excised fragments, up to 2 Mbp in size, were resolved by pulsed-field gel electrophoresis and were detected by hybridization. In general, the genomic RARE-cleavage results support the YAC-based map. In one case, the distance in uncloned DNA between the two terminal EcoRI sites of a YAC insert was approximately 1 Mbp larger than the YAC itself, indicating a major deletion. The general concept of RARE-cleavage mapping as well as its applications and limitations are discussed.

Physical and comparative mapping of distal mouse chromosome 16. 5 p5

Distal mouse Chromosome 16 (Chr. 16) includes a region of conserved linkage with human Chromosome 21 (Chr. 21). Mouse models of Down syndrome based on trisomy of distal Chr. 16 have several phenotypes similar to those seen in human patients and have proven useful for correlating dosage imbalance of specific genes with specific developmental anomalies. The degree to which such findings can be related to Down syndrome depends on how well the conserved synteny is maintained. Twenty-four genes have been mapped in both species and there are no discordancies, but the region could carry hundreds of genes. Comparative sequence represents the ultimate comparative map and will aid in identification of genes and their regulatory sequences. A physical map of the distal 4.5 Mb of Chr. 16 has been assembled as an essential step toward a map of sequence-ready templates. The map consists of 51 YACs and 15 BACs and includes 18 transcripts, 9 of which are mapped for the first time in mouse, and 3 of which are, for the first time, described in either species. YAC fragmentation was used to precisely localize the 49 markers on the map. Comparison of this physical map with that of the corresponding region on Chr. 21 shows conservation not only of gene order but of size in the 3 Mb from Cbr1 to Ets2; distal to Ets2, the human map is expanded.

Genomic organization and partial duplication of the human alpha7 neuronal nicotinic acetylcholine receptor gene (CHRNA7)

The human alpha7 neuronal nicotinic acetylcholine receptor gene (HGMW-approved symbol CHRNA7) has been characterized from genomic clones. The gene is similar in structure to the chick alpha7 gene with 10 exons and conserved splice junction positions. The size of the human gene is estimated to be larger than 75 kb. A putative promoter 5' of the translation start in exon 1 has been cloned and sequenced. The promoter region lacks a TATA box and has a high GC content (77%). Consensus Sp1, AP-2, Egr-1, and CREB transcription factor binding sites appear to be conserved between bovine and human genes. The alpha7 nAChR gene was found to be partially duplicated, with both loci mapping to the chromosome 15q13 region. A yeast artificial chromosome contig was constructed over a genetic distance of 5 cM that includes both alpha7 loci and the region between them. Four novel exons are described, located in genomic clones containing the partially duplicated gene. The duplicated sequences, including the novel exons, are expressed in human brain.

An improved approach for construction of bacterial artificial chromosome libraries

Presented here are improved methodologies that enable the generation of highly redundant bacterial artificial chromosome/P1-derived artificial chromosome libraries, with larger and relatively uniform insert sizes. Improvements in vector preparation and enhanced ligation conditions reduce the number of background nonrecombinant clones. Preelectrophoresis of immobilized high-molecular-weight DNA removes inhibitors of the cloning process, while sizing DNA fragments twice within a single gel effectively eliminates small restriction fragments, thus increasing the average insert size of the clones. The size-fractionated DNA fragments are recovered by electroelution rather than the more common melting of gel slices with subsequent beta-agarase treatment. Concentration of the ligation products yields a 6- to 12-fold reduction in the number of electroporations required in preparing a library of desirable size. These improved methods have been applied to prepare PAC and BAC libraries from the human, murine, rat, canine, and baboon genomes with average insert sizes ranging between 160 and 235 kb.

Mapping an endometrial cancer tumor suppressor gene at 10q25 and development of a bacterial clone contig for the consensus deletion interval

Frequent loss of chromosome 10q sequences in endometrial cancers suggests the involvement of a tumor suppressor gene. Previous loss-of-heterozygosity (LOH)studies have pointed to the 10q25-q26 region as the likely site of a tumor suppressor involved in endometrial tumorigenesis (S. L. Peiffer et al., 1995, Cancer Res. 55: 1922-1926; S. Nagase et al., 1996, Br. J. Cancer 74: 1979-1983; S. Nagase et al.,1997, Cancer Res. 57: 1630-1633). In an attempt to define further the localization of a tumor suppressor gene at 10q25, we screened a panel of 123 endometrioid adenocarcinomas for loss of heterozygosity of 10q25.3 sequences. Forty-three (35%) revealed LOH at one or more loci. The observed patterns of allelic loss define a minimum consensus region of deletion between D10S221 and D10S610. A sequence-ready bacterial clone contig and a long-range restriction map for a 1-Mb interval spanning the deletion region were developed as the first step in experiments directed toward the discovery the 10q25 tumor suppressor.

Two sequence-ready contigs spanning the two copies of a 200-kb duplication on human 21q: partial sequence and polymorphisms

Physical mapping across a duplication can be a tour de force if the region is larger than the size of a bacterial clone. This was the case of the 170- to 275-kb duplication present on the long arm of chromosome 21 in normal human at 21q11.1 (proximal region) and at 21q22.1 (distal region), which we described previously. We have constructed sequence-ready contigs of the two copies of the duplication of which all the clones are genuine representatives of one copy or the other. This required the identification of four duplicon polymorphisms that are copy-specific and nonallelic variations in the sequence of the STSs. Thirteen STSs were mapped inside the duplicated region and 5 outside but close to the boundaries. Among these STSs 10 were end clones from YACs, PACs, or cosmids, and the average interval between two markers in the duplicated region was 16 kb. Eight PACs and cosmids showing minimal overlaps were selected in both copies of the duplication. Comparative sequence analysis along the duplication showed three single-basepair changes between the two copies over 659 bp sequenced (4 STSs), suggesting that the duplication is recent (less than 4 mya). Two CpG islands were located in the duplication, but no genes were identified after a 36-kb cosmid from the proximal copy of the duplication was sequenced. The homology of this chromosome 21 duplicated region with the pericentromeric regions of chromosomes 13, 2, and 18 suggests that the mechanism involved is probably similar to pericentromeric-directed mechanisms described in interchromosomal duplications.

Fluorescence in situ hybridization analysis of 12;21 translocation in Japanese childhood acute lymphoblastic leukemia

Fluorescence in situ hybridization (FISH) analysis was applied to detect t(12;21) using two yeast artificial chromosome probes and cosmid probes covering the TEL(ETV6) and the AML1 gene to clarify the incidence of abnormality of t(12;21) in Japanese childhood acute lymphoblastic leukemia (ALL). We detected seven TEL/AML1 fusion positive patients (9.5%), all of whom were diagnosed as B-lineage ALL, among 74 childhood ALL. On the other hand, no TEL/AML1 fusion positive patients were found among 37 adult ALL. The incidence among Japanese seemed to be lower than that among other nations. Of the seven patients with the TEL/AML1 fusion, five exhibited normal karyotype, one was t(8;12)(q11;p13), i(21q) and the remaining one exhibited a near-triploid karyotype in conventional G-banding. The FISH method clearly demonstrated that all patients with the TEL/AML1 fusion had subpopulations of leukemic cells with deletion of the normal TEL allele, which is significant for understanding the progression of leukemia with t(12;21).

10 years of Genomics, chromosome 21, and Down syndrome

During the first 10 years of Genomics (1987-1997), the molecular structure of human chromosome 21 (HC21) has been intensively investigated. Due to its small size and involvement in Down syndrome, it continues to serve as a model in the development of "genomics technologies." Increasingly more detailed genetic, radiation hybrid, physical, and transcription maps, in addition to NotI restriction and chromosomal breakpoint maps, of HC21 have been developed, and approximately 10% of its genes have been cloned. These maps have been vital in the localization of loci for 15 monogenic disorders to HC21, and 10 of these genes have been identified and characterized. The genetic maps have aided in the detailed elucidation of the origin of the supernumerary HC21 in trisomy 21 from investigations of recombination and nondisjunction events. Mouse models of Down syndrome, with partial trisomy 16, the mouse chromosome principally syntenic to HC21, have been created and initially characterized. A substantial number of the above studies related to the molecular mapping, gene cloning, and infrastructure of HC21 were published in Genomics (e.g., approximately 30% of papers describing HC21 maps were published here). The future goals of genomic analysis of HC21 will be the determination of its complete nucleotide sequence and the identification and functional analysis of all of its genes. These advances will help to provide a molecular explanation of the pathophysiology of Down syndrome and aid in the identification of genes for monogenic and polygenic disorders that map on this chromosome. Novel therapeutic interventions for Down syndrome and the monogenic and polygenic disorders that map to HC21 will be designed and tried based on the knowledge of the disease pathogenesis resulting from the genomic analysis.

Physical mapping of the evolutionary boundary between human chromosomes 21 and 22 on mouse chromosome 10

Adjacent regions of mouse Chromosome 10 (MMU10) show conserved synteny with human chromosome 22 (HSA22) and the telomeric region of HSA21. Physical mapping on MMU10 using YAC fragmentation and PAC contig analyses demonstrates that Prmt2 has a position consistent with its human homolog, HRMT1L1, being telomeric to S100B on HSA21. This result establishes Prmt2 as the new proximal boundary of the region of conserved synteny between MMU10 and HSA21 and predicts that it is the most telomeric gene known on HSA21. Physical mapping refines the positions and order of HSA22 homologs Mmp11, Mif, and Ddt, demonstrates the orientation of S100b on the mouse chromosome, and localizes the junction of conserved synteny between HSA21 and HSA22 on MMU10. Comparative mapping in this region is important for defining gene structure and dosage imbalance in Down syndrome (DS), for developing animal models of DS, and for understanding processes of chromosome evolution.

Screening for overlapping bacterial artificial chromosome clones by PCR analysis with an arbitrary primer

In this article, we used PCR analysis with arbitrary primers (AP-PCR) to screen for overlapping bacterial artificial chromosome (BAC) clones and assembly of contigs. A rice BAC library with three genome equivalents was used to prepare pooled BAC DNA. Twenty-two arbitrary primers were used to survey the pooled BAC DNAs and individual BAC DNAs. Each primer identified 1-10 loci, and the average was 4.4 loci. There were 1-5 overlapping clones in each locus, and the average was 2.5 clones. A total of 245 BAC clones were identified as overlapping by AP-PCR and the identities were confirmed by DNA-DNA hybridization. The 245 BAC clones were then assembled into 80 contigs and 17 single-clone loci. The results indicated that PCR analysis with arbitrary primers is a powerful tool in screening for overlapping BAC clones with high accuracy and efficiency. The use of AP-PCR analysis should speed up the construction of physical maps of the plant and animal genomes, as well as the rice genome.

A detailed physical and transcriptional map of the region of chromosome 20 that is deleted in myeloproliferative disorders and refinement of the common deleted region

Acquired deletions of the long arm of chromosome 20 are the most common chromosomal abnormality seen in polycythemia vera and are also associated with other myeloid malignancies. Such deletions are believed to mark the site of one or more tumor suppressor genes, loss of which perturbs normal hematopoiesis. A common deleted region (CDR) has previously been identified on 20q. We have now constructed the most detailed physical map of this region to date--a YAC contig that encompasses the entire CDR and spans 23 cM (11 Mb). This contig contains 140 DNA markers and 65 unique expressed sequences. Our data represent a first step toward a complete transcriptional map of the CDR. The high marker density within the physical map permitted two complementary approaches to reducing the size of the CDR. Microsatellite PCR refined the centromeric boundary of the CDR to D20S465 and was used to search for homozygous deletions in 28 patients using 32 markers. No such deletions were detected. Genetic changes on the remaining chromosome 20 may therefore be too small to be detected or may occur in a subpopulation of cells.

High-fidelity digital hybridization screening

We developed a highly efficient screening method for minimizing the hybridization of high-density replica (HDR) filters and for allowing simultaneous use of numerous oligonucleotide probes for STS markers. We designated this method "digital hybridization (DH) screening," in which a binary n-bit ID number is given to each probe, and a series of probe mixtures is prepared in an arranged combination. The matrix pattern between probe mixtures and hybridization signals determines the relation between a particular STS marker(s) and the corresponding DNA clone(s). Here, we describe a successful DH screening of over 15,000 human BAC clones with 126 STS marker probes with 7-bit ID numbers, which required only 8 sets of HDR filter hybridizations. Our results indicate that DH screening can be performed with more than 1000 STS marker probes with only 10-bit ID numbers. The DH screening method is convenient, economical, and of high fidelity, and thus it should facilitate the construction of sequence-ready DNA contigs for the human genome as well as for genomes of various species. The principle of DH screening has various applications to the biological sciences.

Rearrangement of the human CDC5L gene by a t(6;19)(p21;q13.1) in a patient with multicystic renal dysplasia

Genetic studies have implicated the short arm of chromosome 6 in congenital hydronephrosis. In previous studies, we described a fetus carrying a t(6;19)(p21;q13.1) as the sole cytogenetic anomaly and suffering from bilateral multicystic renal dysplasia caused by a bilateral complete pelviureteric junction obstruction, resulting in a massive hydronephrosis. Characterization of the chromosome 19 breakpoint region revealed that the transcription factor-encoding USF2 gene is affected. In this report, we show that the CDC5L gene on chromosome 6p is rearranged in the cells of the fetus. CDC5L encodes a protein that is related to the product of the Schizosaccharomyces pombe Cdc5 gene, which exerts its effects at the G2/M transition during cell cycle progression. We have established the genomic organization of the CDC5L gene and found that it consists of at least 16 exons spanning approximately 50 kb of chromosome segment 6p21. Northern blot analysis indicated that the gene is ubiquitously expressed as a single mRNA of about 3.4 kb in both fetal and adult tissues. The translation product of the CDC5L gene has an electrophoretic mobility of about 100 kDa and is predicted to be a nuclear protein, since it contains a Myb-related DNA binding domain and potential nuclear localization signals in its aminoterminal region. Immunocytochemical analysis confirmed the nuclear localization of the CDC5L protein. CDC5L was also predicted to contain a hydrophilic, proline-rich region in its central part, which might function as a transcriptional activating domain. The chromosome 6 breakpoint was found in the intron between exons 9 and 10, indicating that, as a direct result of the 6;19 translocation, the Myb-related DNA binding domains and the nuclear localization signals are separated from the putative transactivating domain. Northern blot and RT-PCR experiments revealed that the other CDC5L allele is unaffected, and in Western blot experiments, expression of the 100-kDa protein was detected in fibroblasts of the fetus. Expression of a truncated or hybrid CDC5L transcript resulting from the CDC5L rearrangement could not be demonstrated.

Construction of a 2.5-Mb integrated physical and gene map of distal 21q22.3

The gene-rich telomeric region of 21q harbors several loci relevant to human diseases including autoimmune polyglandular disease type I, nonsyndromic deafness, Knobloch syndrome, holoprosencephaly, and bipolar affective disorder. A contig of genomic clones in this region would facilitate the isolation of these genes. However, distal 21q22.3 has yet been poorly mapped, presumably due to the presence of sequences that are underrepresented in yeast artificial chromosome (YAC) libraries. We generated a framework of YACs and used these clones as starting points for the isolation of a combination of bacterial artificial chromosome clones, P1-derived artificial chromosome clones, and cosmid clones by chromosome walking procedures. These studies resulted in the construction of a high-resolution contig map spanning the 2.5-Mb region from PFKL to the telomere, approximately 2 Mb of which are covered by ready-to-sequence contigs. Within this map we determined the location and relative distance of 21 markers. These include 9 established genetic markers, the order of which is cen-PFKL-D21S154-D21S170-D21S171-D21S1903- D21S1897- D21S112-D21S1446-D21S1575-tel. Moreover, we established the precise map position of 13 genes and 4 ESTs including the recently isolated genes C21ORF2, SMT3H1, RNA editing deaminase 1 (ADARB1), folate transporter (SLC19A1), COL18A1, lanosterol synthase (LSS-PEN), pericentrin (PCNT), and arginine methyltransferase (HRMT1L1). This integrated map provides a useful resource for the mapping and isolation of disease genes and for the construction of a complete transcription map of distal 21q as well as for large-scale sequencing efforts.

Bacterial contig map of the 21q11 region associated with Alzheimer's disease and abnormal myelopoiesis in Down syndrome

We present a high-resolution bacterial contig map of 3.4 Mb of genomic DNA in human chromosome 21q11-q21, encompassing the region of elevated disomic homozygosity in Down Syndrome-associated abnormal myelopoiesis and leukemia, as well as the markers, which has shown a strong association with Alzheimer's Disease that has never been explained. The map contains 89 overlapping PACs, BACs, or cosmids in three contigs (850, 850, and 1500 kb) with two gaps (one of 140-210 kb and the second <5 kb). To date, eight transcribed sequences derived by cDNA selection, exon trapping, and/or global EST sequencing have been positioned onto the map, and the only two genes so far mapped to this cytogenetic region, STCH and RIP140 have been precisely localized. This work converts a further 10% of chromosome 21q into a high-resolution bacterial contig map, which will be the physical basis for the long-range sequencing of this region. The map will also enable positional derivation of new transcribed sequences, as well as new polymorphic probes, that will help in elucidation of the role the genes in this region may play in abnormal myelopoiesis and leukemia associated with trisomy 21 and Alzheimer's Disease.

Identification and characterization of two putative human arginine methyltransferases (HRMT1L1 and HRMT1L2)

RNA-binding proteins such as heterogeneous nuclear ribonucleoproteins (hnRNPs), which contain the bulk of methylated arginine residues in eukaryotic cells, play many essential roles in the metabolism of nuclear pre-mRNA. Arginine methyltransferase activity has also been implicated in signal transduction events with components of the cellular growth and viral response pathways. We recently characterized a single yeast hnRNP methyltransferase (HMT1). We now present the identification and characterization of two putative human arginine methyltransferases termed HRMT1L1 and HRMT1L2. In addition to methyltransferase similarities, the N-terminal region of the HRMT1L1 protein contains an Src homology 2 domain. HRMT1L1 maps to a YAC containing the telomere of chromosome 21q. Three alternatively spliced HRMT1L2 transcripts with variable 5'-ends were observed, encoding proteins of 343, 347, and 361 amino acids, respectively. HRMT1L2 maps to human chromosome 19q. Recombinant HRMT1L2 protein encoded by the most common 5'-variant exhibited methyltransferase activity in vitro. Furthermore, in vivo activity was demonstrated by complementation of a yeast HMT1 mutant strain. The identification of highly conserved Hmt1p human homologues that function in yeast indicates that analyses of this class of enzymes in yeast may be directly applicable to higher eukaryotes. The possible roles of HRMT1L1 and HRMT1L2 in human disease are currently unknown.

High resolution physical mapping and identification of transcribed sequences in the Down syndrome region-2

The identification and mapping of genes within the Down syndrome region is an important step toward a complete understanding of the pathogenesis of this disorder. The objective of the present work is to identify and map genes within the Down syndrome region-2. Chromosome 21 cosmid clones corresponding to "cosmid pockets" 121-124 have been first used as a starting material for generation of a single high resolution integrated cosmid/PAC contig with full EcoRI/SmaI restriction map. The integrated contig has been further anchored to genetic and physical maps through the positioning of 6 markers in the following order: ACTL5-D21S3-684G2T7-D21S71-D21S343-D21S 268. The entire contig covers 342 kb of the Down syndrome region-2 of chromosome 21. Subsequently, we have isolated, identified, and mapped four novel cDNAs which we have named N143, N144, CHD/333, and 90/3H1 and a potentially transcribed genomic sequence (E05133T7). Additionally, we have accurately located a previously described gene, the WRB gene, between the markers ACTL5-D21S268 within this Down Syndrome Region-2.

Integrated YAC contig map of the Prader-Willi/Angelman region on chromosome 15q11-q13 with average STS spacing of 35 kb

Prader-Willi syndrome and Angelman syndrome are associated with parent-of-origin-specific abnormalities of chromosome 15q11-q13, most frequently a deletion of an approximately 4-Mb region. Because of genomic imprinting, paternal deficiency of this region leads to PWS and maternal deficiency to AS. Additionally, this region is frequently involved in other chromosomal rearrangements including duplications, triplications, or supernumerary marker formation. A detailed physical map of this region is important for elucidating the genes and mechanisms involved in genomic imprinting, as well as for understanding the mechanism of recurrent chromosomal rearrangments. An initial YAC contig extended from D15S18 to D15S12 and was comprised of 23 YACs and 21 STSs providing an average resolution of about one STS per 200 kb. To close two gaps in this contig, YAC screening was performed using two STSs that flank the gap between D15S18 and 254B5R and three STSs located distal to the GABRA5-149A9L gap. Additionally, we developed 11 new STSs, including seven polymorphic markers. Although several groups have developed whole-genome genetic and radiation hybrid maps, the depth of coverage for 15q11-q13 has been somewhat limited and discrepancies in marker order exist between the maps. To resolve the inconsistencies and to provide a more detailed map order of STSs in this region, we have constructed an integrated YAC STS-based physical map of chromosome 15q11-q13 containing 118 YACs and 118 STSs, including 38 STRs and 49 genes/ESTs. Using an estimate of 4 Mb for the size of this region, the map provides an average STS spacing of 35 kb. This map provides a valuable resource for identification of disease genes localized to this region as well as a framework for complete DNA sequencing.

Human NRD convertase: a highly conserved metalloendopeptidase expressed at specific sites during development and in adult tissues

We report the cloning of the human homologue of the rat metalloprotease N-arginine dibasic convertase (NRD convertase). This endopeptidase is responsible for the processing, at the Arg-Lys dibasic site on the N-terminal side of the arginine residue, of propeptides and proproteins. Comparisons of the human and rat full-length cDNAs show similarity and identity of 94 and 91%, respectively. In humans NRD convertase is predominantly expressed in heart, skeletal muscle, and testis. We have also studied the expression of this gene in mouse at various developmental stages and found that the neural tissue is the almost exclusive site of expression in early development (between E 10.5 and E 16.5). To gain information about the possibility that defects in this gene are linked to inherited neuromuscular disorders, we determined the chromosomal location of the human NRD convertase gene by FISH analysis, showing that the gene resides at 1p32.2.

Characterization of a human homolog (BACH1) of the mouse Bach1 gene encoding a BTB-basic leucine zipper transcription factor and its mapping to chromosome 21q22.1

Clinical interest in the genes on human chromosome 21, especially with respect to Down syndrome (DS), has provided a strong impetus for the creation of a transcript map of this chromosome. In an effort to identify new human genes on the basis of cDNA analysis, we found several cDNA clones that corresponded to chromosome 21-specific transcripts. One of these, ha2303, showed strong similarity to the murine transcription factor Bach1. We subsequently determined the entire nucleotide sequence of this cDNA clone and found it to contain the whole coding sequence. The gene, termed BACH1, encodes a 736-amino-acid polypeptide with 80.3% identity to the murine Bach1 protein and contains a Cap'n'collar (CNC)-type basic leucine zipper (bZip) domain and a protein interaction motif, the BTB domain. Northern blot analysis revealed that BACH1 was expressed in all tissues examined. Mapping using the NotI restriction map and the YAC contig map showed that the BACH1 gene is located at 21q22.1 between the NotI sites LA329 (D21S338) and LL60 (D21S389) and within approximately 400 kb of LA329. Both the prospective function and the chromosomal location suggest that this gene may be a DS candidate gene, contributing to certain DS phenotypes, and is possibly involved in certain features of monosomy 21.

High-resolution YAC-cosmid-STS map of human chromosome 13

We have assembled a high-resolution physical map of human chromosome 13 DNA (approximately 114 Mb) from hybridization, PCR, and FISH mapping data using a specifically designed set of computer programs. Although the mapping of 13p is limited, 13q (approximately 98 Mb) is covered by an almost continuous contig of 736 YACs aligned to 597 contigs of cosmids. Of a total of 10,789 cosmids initially selected from a chromosome 13-specific cosmid library (16,896 colonies) using inter-Alu PCR probes from the YACs and probes for markers mapped to chromosome 13, 511 were assembled in contigs that were established from cross-hybridization relationships between the cosmids. The 13q YAC-cosmid map was annotated with 655 sequence tagged sites (STSs) with an average spacing of 1 STS per 150 kb. This set of STSs, each identified by a D number and cytogenetic location, includes database markers (198), expressed sequence tags (93), and STSs generated by sequencing of the ends of cosmid inserts (364). Additional annotation has been provided by positioning 197 cosmids mapped by FISH on 13q. The final (comprehensive) map, a list of STS primers, and raw data used in map assembly are available at our Web site (genome1.ccc.columbia.edu/ approximately genome/) and can serve as a resource to facilitate accurate localization of additional markers, provide substrates for sequencing, and assist in the discovery of chromosome 13 genes associated with hereditary diseases.

Characterization of a novel gene, C21orf2, on human chromosome 21q22.3 and its exclusion as the APECED gene by mutation analysis

Exon trapping was performed from a partial cosmid, PAC, and P1 clone contig from human chromosome 21 between MX1 and 21qter to identify genes that may be involved in the pathogenesis of Down syndrome or several of the genetic diseases that map to chromosome 21q22.3. One 19-bp exon showed identity to three ESTs. The complete sequence of the EST clones, RT-PCR, and cDNA library screening were used to determine the full-length cDNA sequence of 2.2 kb with an open reading frame of 256-amino-acids. The putative 256-amino-acid peptide has homology with a hypothetical Caehorhabditis elegans protein of unknown function. Northern blot analysis of this gene, termed C21orf2 (chromosome 21 open reading frame 2), revealed two ubiquitously expressed mRNAs of 2.2 and 1.2 kb produced by use of alternative polyadenylation sites. Hybridization of the EST clones to a cosmid contig in chromosome 21q22.3 mapped C21orf2 just distal to PFKL, a critical mapping region for several genetic diseases. Comparison to publicly available genomic sequence, and additional data, revealed that the gene is split into seven exons over 10.5 kb, further refining the mapping position to only 1.2 kb distal to PFKL with the direction of transcription toward the centromere. The 5'UTR is contiguous with D21S400, and intron 2 contains a 52-bp VNTR polymorphism. Given its mapping position, C21orf2 is a candidate for involvement in disorders including autoimmune polyglandular disease type I (also called autoimmune polyendocrinopathy candidiasis ectodermal dystrophy or APECED) and the autosomal nonsyndromic deafness loci, DFNB8 and DFNB10. Mutation analysis using sequencing of RT-PCR and genomic DNA-derived PCR products, SSCP, and Southern and Northern blot analyses in APECED patients excluded C21orf2 as the gene for APECED.

Clonability and gene distribution on human chromosome 21: reflections of junk DNA content?

Data from transcriptional mapping of human chromosome 21 have been compiled from a number of sources. Regardless of the gene identification technique used, a consistent picture has developed: the centromere proximal half of 21q, which contains 50% of the DNA (20 Mb), harbors only 10% of the expressed sequences. Because of the variety of gene isolation techniques used, this result is unlikely to arise simply from methodological artefacts, biases in clonability or tissue specificity of expression. This region is known to be AT-rich and to contain APP, the largest gene (spanning 300 kb) currently analyzed on 21q. Interesting preliminary data from analysis of the Fugu rubripes homolog of APP has shown an unusually high, 50-fold, compaction of intron size, raising the intriguing possibility that >90% of the DNA in the human gene may be functionless. Thus, data from a variety of approaches suggest that a large part of 21q very likely has neither coding capacity nor associated regulatory function. By these criteria, it is a good candidate for a repository of junk DNA.

Towards the physical map of the Trypanosoma cruzi nuclear genome: construction of YAC and BAC libraries of the reference clone T. cruzi CL-Brener

Strategies to construct the physical map of the Trypanosoma cruzi nuclear genome have to capitalize on three main advantages of the parasite genome, namely (a) its small size, (b) the fact that all chromosomes can be defined, and many of them can be isolated by pulse field gel electrophoresis, and (c) the fact that simple Southern blots of electrophoretic karyotypes can be used to map sequence tagged sites and expressed sequence tags to chromosomal bands. A major drawback to cope with is the complexity of T. cruzi genetics, that hinders the construction of a comprehensive genetic map. As a first step towards physical mapping, we report the construction and partial characterization of a T. cruzi CL-Brener genomic library in yeast artificial chromosomes (YACs) that consists of 2,770 individual YACs with a mean insert size of 365 kb encompassing around 10 genomic equivalents. Two libraries in bacterial artificial chromosomes (BACs) have been constructed, BACI and BACII. Both libraries represent about three genome equivalents. A third BAC library (BAC III) is being constructed. YACs and BACs are invaluable tools for physical mapping. More generally, they have to be considered as a common resource for research in Chagas disease.

A sequence-ready physical map of a region of 12q24.1

We developed a sequence-ready map of a part of human chromosome 12q24.1. We utilized a number of sequence-tagged site (STS) markers from 12q24.1 to screen large insert bacterial chromosome libraries and a chromosome 12-specific cosmid library. The clones were assembled into contiguous sets (contigs) by STS-content analysis. Contigs were extended by obtaining end sequences of bacterial clones, generation of additional STSs, rescreening the libraries, and screening the additional clones for the presence of STSs. The resulting contig covers nearly 2 Mb of DNA and provides an average marker resolution of 16 kb. Based on the STS content, we developed fingerprints of a subset of clones. The STS content and fingerprint data allowed us to define a minimal tiling path of clones. These clones are being used to sequence this part of chromosome 12. This contig contains the Ataxin 2 gene, and it covers the interval harboring the gene responsible for Darier disease.

Homologous recombination of a flanking repeat gene cluster is a mechanism for a common contiguous gene deletion syndrome

Smith-Magenis syndrome (SMS), caused by del(17)p11.2, represents one of the most frequently observed human microdeletion syndromes. We have identified three copies of a low-copy-number repeat (SMS-REPs) located within and flanking the SMS common deletion region and show that SMS-REP represents a repeated gene cluster. We have isolated a corresponding cDNA clone that identifies a novel junction fragment from 29 unrelated SMS patients and a different-sized junction fragment from a patient with dup(17)p11.2. Our results suggest that homologous recombination of a flanking repeat gene cluster is a mechanism for this common microdeletion syndrome.

Generation of large insert yeast artificial chromosome libraries

The development of YAC cloning technology has directly enhanced the relationship among genetic, physical, and functional mapping of genomes. Because of their large size, YACs have enabled the rapid construction of physical maps by ordered clone mapping and contig building, and they complement other molecular approaches for mapping complex genomes. Large insert libraries are constructed by size fractionating large DNA embedded in agarose and protecting DNA from degradation with polyamines.

Cloning of the TMPRSS2 gene, which encodes a novel serine protease with transmembrane, LDLRA, and SRCR domains and maps to 21q22.3

To contribute to the development of the transcription map of human chromosome 21 (HC21), we have used exon trapping from pools of HC21-specific cosmids. Using selected trapped exons, we have identified a novel gene (named TMPRSS2) that encodes a multimeric protein with a serine protease domain. The TMPRSS2 3.8-kb mRNA is expressed strongly in small intestine and weakly in several other tissues. The full-length cDNA encodes a predicted protein of 492 amino acids that contains the following domains: (i) A serine protease domain (aa 255-492) of the S1 family that probably cleaves at Arg or Lys residues. (ii) An SRCR (scavenger receptor cysteine-rich) domain (aa 149-242) of group A (6 conserved Cys). This type of domain is involved in the binding to other cell surface or extracellular molecules. (iii) An LDLRA (LDL receptor class A) domain (aa 113-148). This type of domain forms a binding site for calcium. (iv) A predicted transmembrane domain (aa 84-106). No typical signal peptide was recognized. The gene was mapped to 21q22.3 between markers ERG and D21S56 in the same P1 as MX1. The physiological role of TMPRSS2 and its involvement in trisomy 21 phenotypes or monogenic disorders that map to HC21 are unknown.

Genomic organization, alternative splicing, and expression patterns of the DSCR1 (Down syndrome candidate region 1) gene

Down syndrome is a major cause of mental retardation and congenital heart defects and is due to the presence of three copies of human chromosome 21 in the affected individual. We have identified a gene, DSCR1 (HGMW-approved symbol), from the region 21q22.1-q22.2, which is highly expressed in human fetal brain and adult heart. Structural features of the conceptual protein encourage us to propose involvement of DSCR1 in the regulation of transcription and/or signal transduction. Higher expression of RNA in the brains of young rats compared to adults suggests a possible role for the gene in the development of the central nervous system. We have determined the genomic organization of DSCR1 and identified three additional alternative first exons by RACE and cDNA library screening. DSCR1 spans nearly 45 kb of genomic DNA and comprises seven exons, four of which (exons 1-4) are alternative first exons. All the exons are flanked by splice junctions that conform to the consensus AG-GT motif. We have studied the expression patterns of the alternative first exons. Exon 2 was detected in fetal brain and liver by RT-PCR. Both exons 1 and 4 were differentially expressed in fetal brain, lung, liver, and kidney and in all adult tissues tested by Northern analysis with two notable exceptions: exon 1 was not detected in adult kidney and exon 4 was not found in adult brain. The high level of expression of exon 1 in fetal brain suggests that this alternative form of DSCR1 has an important role in brain development. This information should help us to understand the possible relationship of DSCR1 with Down syndrome and aid in the development of animal models.

Search for DNA sequence variations using a MutS-based technology

The search for DNA sequence variations (DSV) is emphasized with genetic studies of a large number of multifactorial diseases. Saturation of regions of interest with diallelic polymorphisms will be an essential step to pinpoint, through association studies, predisposing genes. We have developed a solid-phase method based on the ability of mismatch binding protein MutS to recognize single nucleotide mismatches. This approach was applied to the study of 83 sequence-tagged sites (STSs) extracted from an eight centimorgans (cM) chromosome 21 region. One-third of tested STSs were found to be polymorphic leading to a frequency of one DSV every 822 base pairs (bp). Sequencing of analyzed STSs showed the high reliability of the MutS-based technology for mismatches up to 2 bp in DNA fragments ranging in size from 200 bp to 1 kilobase (kb). The entire assay which is performed in a solid-phase format without the need of electrophoresis or sequencing, will provide an efficient tool for new polymorphism detection.

A 2-Mb YAC contig and physical map covering the chromosome 8q12 breakpoint cluster region in pleomorphic adenomas of the salivary glands

Pleomorphic adenomas are benign epithelial tumors originating from the major and minor salivary glands. Extensive cytogenetic studies have demonstrated that they frequently show chromosome abnormalities involving chromosome 8, with consistent breakpoints at 8q12. In previous studies, we have shown that these breakpoints are located in a 9-cM interval between MOS/D8S285 and D8S260. Here, we describe directional chromosome walking studies starting from D8S260 as well as D8S285. Using the CEPH and ICRF YAC libraries, these studies resulted in the construction of two nonoverlapping YAC contigs of about 2 and 5 Mb, respectively. Initial fluorescence in situ hybridization (FISH) analysis suggested that the majority of 8q12 breakpoints clustered within the 2-Mb contig, which was mapped to the centromeric part of chromosome band 8q12. This contig has at least double coverage and consists of 34 overlapping YAC clones. The localization of the YACs was confirmed by FISH analysis. On the basis of mapping data of landmarks with an average spacing of 65 kb as well as restriction enzyme analysis, a long-range physical map was established for the chromosome region spanned by the 2-Mb contig. The relative positions of various known genes and expressed sequence tags within this contig were also determined. Subsequent FISH analyses of pleomorphic adenomas using YACs as well as cosmids revealed that all but two of the 8q12 breakpoints in the primary tumors tested mapped within a 300-kb interval between the MOS proto-oncogene and STS EM156. The target gene affected by the chromosome aberrations mapping within this interval was recently shown to be the PLAG1 gene, which encodes a novel zinc finger protein.

High-resolution physical and transcriptional mapping of the autoimmune polyendocrinopathy-candidiasis-ectodermal dystrophy locus on chromosome 21q22.3 by FISH

Autoimmune-polyendocrinopathy-candidiasis-ectodermal dystrophy (APECED, PGD type I) is an autosomal recessive disease enriched in the Finnish population. Previously, we have assigned APECED to a 2.6-cM interval on chromosome 21q22.3 by linkage analysis in 14 Finnish families. This subtelomeric region of 21q22.3 seems to have sequence features resulting in its under-representation in large insert genomic libraries, and only a few large insert clones have been available for positional cloning to date. Here, we report the refined localization of the APECED gene and a visual physical map of 800 kb covering the critical chromosomal region for the gene. In the construction of the physical map, the recently developed fiber FISH techniques were essential for the orientation of the cosmid PI, PAC, and BAC clones in relation to each other. We also localized two cDNAs within this genomic region by fiber FISH combined with the highly sensitive tyramide-based detection method. These data will facilitate the final cloning of the APECED gene and any other novel gene in this complex genomic region.

A compositional map of the cen-q21 region of human chromosome 21

A compositional map of the centromere and of the subcentromeric region of the long arm of human chromosome 21 was established by determining the GC levels (GC is the molar fraction of guanine+cytosine in DNA) of 11 YACs (yeast artificial chromosomes) covering this 13-14 Mb region which extends from the alpha-satellite sequences of the C(entromeric) band q11.1, through R(everse) band q11.2, to the proximal part of G(iemsa) band q21. The entire region is made up of GC-poor, or L, isochores with only one GC-rich H1 isochore, at least 2 Mb in size, located in band q21. The almost identical GC levels of the centromeric alpha-satellite repeats (38.5%), of R band q11.2 (39%), and of G bands (38-40%) provide a direct demonstration that base composition cannot be the only cause of the cytogenetic differences between C, G, and the majority of R bands, namely the H3- R bands (which do not contain the GC-richest H3 isochores). The results obtained also show that isochores may be as long as 6 Mb, at least in the GC-poor regions of the genome, and support previous observations suggesting that YACs from isochore borders are unstable and/or difficult to clone. Genes and CpG islands are very rare in the GC-poor region investigated, as expected from the fact that their concentration is proportional to the GC levels of the isochores in which they are contained.

Escape from X inactivation of two new genes associated with DXS6974E and DXS7020E

Most genes on the X chromosome undergo "inactivation," being transcribed from only one copy in female somatic cells, but several human genes have been shown to be expressed from both the active and the otherwise inactivated homologue. To assess further the fraction and location of genes that escape inactivation, we have analyzed the inactivation status of a set of 73 expressed sequence tags that were derived from the sequencing of cDNA collections and mapped to the X chromosome. Of 33 that were expressed in cultured cells, as assessed by reverse transcription and PCR, 4 (about 12%) were transcribed from both the active and the inactive X chromosome. Two, RPS4 and PCTAIRE1, are already known to escape inactivation; the other 2, of unknown function, include a short cDNA with a full open reading frame and a transcript with no detectable open reading frame. They map, respectively, to Xp11.3-p11.4 and Xp22.2; both regions were previously reported to encode sequences transcribed from the inactive X. Neither transcript has a corresponding sequence on the Y. Thus, they exhibit double dosage in females compared to males, and inactivation status may be inconsequential for these transcribed sequences.

A physical map of human chromosome 7: an integrated YAC contig map with average STS spacing of 79 kb

The construction of highly integrated and annotated physical maps of human chromosomes represents a critical goal of the ongoing Human Genome Project. Our laboratory has focused on developing a physical map of human chromosome 7, a approximately 170-Mb segment of DNA that corresponds to an estimated 5% of the human genome. Using a yeast artificial chromosome (YAC)-based sequence-tagged site (STS)-content mapping strategy, 2150 chromosome 7-specific STSs have been established and mapped to a collection of YACs highly enriched for chromosome 7 DNA. The STSs correspond to sequences generated from a variety of DNA sources, with particular emphasis placed on YAC insert ends, genetic markers, and genes. The YACs include a set of relatively nonchimeric clones from a human-hamster hybrid cell line as well as clones isolated from total genomic libraries. For map integration, we have localized 260 STSs corresponding to Genethon genetic markers and 259 STSs corresponding to markers orders by radiation hybrid (RH) mapping on our YAC contigs. Analysis of the data with the program SEGMAP results in the assembly of 22 contigs that are "anchored" on the Genethon genetic map, the RH map, and/or the cytogenetic map. These 22 contigs are ordered relative to one another, are (in all but 3 cases) oriented relative to the centromere and telomeres, and contain > 98% of the mapped STSs. The largest anchored YAC contig, accounting for most of 7p, contains 634 STSs and 1260 YACs. An additional 14 contigs, accounting for approximately 1.5% of the mapped STSs, are assembled but remain unanchored on either the genetic or RH map. Therefore, these 14 "orphan" contigs are not ordered relative to other contigs. In our contig maps, adjacent STSs are connected by two or more YACs in > 95% of cases. With 2150 mapped STSs, our map provides an average STS spacing of approximately 79 kb. The physical map we report here exceeds the goal of 100-kb average STS spacing and should provide an excellent framework for systematic sequencing of the chromosome.

HOSEpipe--a WWW-hosted data management and analysis system for STS content mapping projects

We have developed a data management system, 'HOSEpipe' (High Output STS Evaluation pipeline) to aid sample tracking and data analysis in STS content mapping projects. The system is based around a World Wide Web (WWW) server that provides a number of pages including forms for sample processing and data entry accessible via a standard WWW browser application. The system is split into two main modules: firstly, a sequence evaluation and annotation module that takes de novo sequence for a potential STS, screens it against existing STSs and DNA sequence databases, followed by appropriate primer sequence design; secondly, a module that handles YAC library STS screening and includes facilities for both sample tracking and experimental data analysis. We present the design and rationale of the HOSEpipe system and its development to support a whole chromosomal physical mapping project. This software and design approach is potentially applicable to physical mapping projects of varying sizes and resolution and to similar projects, such as sample sequencing and the construction of sequence-ready maps.

High-resolution physical mapping of a 6.7-Mb YAC contig spanning a region critical for the monosomy 21 phenotype in 21q21.3-q22.1

Deletion of genes from the chromosome 21 region between APP and SOD1 is a potential cause of some of the major phenotypic features of monosomy 21 patients. Fine physical mapping helps identify potential candidate genes. After selecting nonchimeric YACs by FISH analysis, we determined their marker contents by PCR and hybridization studies. Fifteen YACs were chosen and mapped by restriction enzyme analysis and labeling of end fragments. We localized 55 markers, including 31 STSs, 10 YAC ends, and 4 NotI linking clones, along a 6.7-Mb contig. This map facilitates transcriptional analysis of this region and construction of ready-to-sequence contigs. Furthermore, FISH mapping of two patients with partial monosomy 21 using YAC and cosmid clones allowed us to define more accurately the telomeric border of the critical region between markers S226 and S213.

Cloning of a novel human neural cell adhesion molecule gene (NCAM2) that maps to chromosome region 21q21 and is potentially involved in Down syndrome

To contribute to the development of the transcription map of human chromosome 21 (HC21), we have used exon trapping to identify portions of HC21 genes. One trapped exon showed strong homology with members of the neural cell adhesion molecule (NCAM) family of genes from different species. We subsequently cloned the complete coding sequence from a human fetal brain cDNA library and determined its nucleotide sequence and predicted amino acid sequence. The predicted polypeptide of this novel NCAM2 gene contains 837 amino acids and shows 62% similarity to the NCAM homologs. It contains five immunoglobulin-like domains, two fibronectin type III domains, a transmembrane domain and a cytoplasmic domain. The gene is expressed most strongly in human adult and fetal brain. Using somatic cell hybrids, we mapped NCAM2 to 21q21, between markers D21S18 and D21S282. Radiation hybrid mapping localized this novel gene between polymorphic markers D21S1914 and D21S265. NCAMs are members of the immunoglobulin superfamily and are essential in the formation and maintenance of tissue structure. To date there are no candidate human disorders on HC21 that could be associated with mutations in NCAM2. In addition, the role of NCAM2 in the pathophysiology of Down syndrome is unknown. However, it is a good candidate for involvement in certain Down syndrome phenotypes because a slight overexpression of NCAMs increases many-fold the homotypic adhesion properties of cells.

Molecular characterization of a complex chromosomal rearrangement in a pleomorphic salivary gland adenoma involving the 3'-UTR of HMGIC

Recently, the high mobility group protein gene, HMGIC, was identified as a common genetic denominator in benign tumors with chromosome 12q13-15 aberrations, such as lipomas, uterine leiomyomas, pleomorphic adenoma of the salivary glands, hamartomas of breast and lung, angiomyxomas, and endometrial polyps. In most cases, the rearrangements resulted in the separation of the three HMGIC DNA-binding motifs from the acidic carboxy-terminal tail. Here, we report about the molecular characterization of a case of pleomorphic adenoma carrying a t(1;12)(p22;q15). Studies were performed on a cell line derived from the primary tumor, i.e., cell line Ad-312/SV40. Although the chromosome 12 breakpoint was initially mapped more than 1 Mb distal to the HMGIC gene by fluorescence in situ hybridization (FISH) analysis, the present molecular studies reveal a more complex chromosomal rearrangement that directly affects the HMGIC gene. Using 3'-RACE analysis, a HMGIC fusion transcript was detected that contained the complete HMGIC, coding region but lacked the putative mRNA destabilizing AUUUA motifs that are normally present in the 3'-UTR of HMGIC. Wild-type HMGIC transcripts were also detected in the tumor cells. The results suggest that alterations in the 3'-noncoding region of HMGIC may have to be considered as pathogenetically relevant.

Human BAC library: construction and rapid screening

We have constructed a human genomic bacterial artificial chromosome (BAC) library using high molecular weight DNA from a pre-pro-B cell line, FLEB14-14, with a normal male diploid karyotype. This BAC library consists of 96,000 clones with an average DNA insert size of 110 kb, covering the human genome approximately 3 times. The library can be screened by three different methods. (1) Probe hybridization to 31 high-density replica (HDR) filters: each filter contains 3072 BAC clones which were gridded in a 6 x 6 pattern. (2) Probe hybridization to two Southern blot filters to which 31 HindIII digests of the pooled 3072 BAC clones were loaded. This identifies a particular HDR filter for which further probe hybridization is performed to identify a particular clone(s). (3) Two-step polymerase chain reaction (PCR). First, PCR is applied to DNA samples prepared from ten superpools of 9600 BAC clones each to identify a particular superpool and the second PCR is applied to 40 unique DNA samples prepared from the four-dimensionally assigned BAC clones of the particular superpool. We present typical examples of the library screening using these three methods. The two-step PCR screening is particularly powerful since it allows us to isolate a desired BAC clone(s) within a day or so. The theoretical consideration of the advantage of this method is presented. Furthermore, we have adapted Vectorette method to our BAC library for the isolation of terminal sequences of the BAC DNA insert to facilitate contig formation by BAC walking.

Functional screening of 2 Mb of human chromosome 21q22.2 in transgenic mice implicates minibrain in learning defects associated with Down syndrome

Using Down syndrome as a model for complex trait analysis, we sought to identify loci from chromosome 21q22.2 which, when present in an extra dose, contribute to learning abnormalities. We generated low-copy-number transgenic mice, containing four different yeast artificial chromosomes (YACs) that together cover approximately 2 megabases (Mb) of contiguous DNA from 21q22.2. We subjected independent lines derived from each of these YAC transgenes to a series of behavioural and learning assays. Two of the four YACs caused defects in learning and memory in the transgenic animals, while the other two YACs had no effect. The most severe defects were caused by a 570-kb YAC; the interval responsible for these defects was narrowed to a 180-kb critical region as a consequence of YAC fragmentation. This region contains the human homologue of a Drosophila gene, minibrain, and strongly implicates it in learning defects associated with Down syndrome.

A novel SH3-containing human gene family preferentially expressed in the central nervous system

The Src-homology-3 domain (SH3) is an evolutionarily conserved, 50- to 60-amino-acid module carried by intracellular proteins involved in the transduction of signals for cell polarization, motility, enzymatic activation, and transcriptional regulation. The SH3 drives protein-protein interactions through binding to proline-rich ligands. This function relies on the conserved secondary structure, whereas the SH3 primary structure is highly diverse. Taking advantage of the fact that the few conserved amino acids are clustered near the N- and C-terminal ends, we designed degenerate oligonucleotides spanning these two regions and screened by PCR a variety of normal and tumor tissues for the expression of SH3-containing transcripts. Using this strategy, we have identified a novel SH3-containing human gene family of six related transcripts that map to four different chromosomes. The SH3 domain lies at the C-terminal end and shows 56-50% amino acid homology to the C-terminal SH3 of Sem-5/Drk/GRB2. The N-terminal segment of this novel SH3GL (from SH3-containing Grb2-like) gene family does not resemble any known protein. Three of these transcripts are in-frame and show a peculiar tissue distribution: SH3GL2 is preferentially expressed in the brain, SH3GL3 in brain and testis, and SH3GL1 is ubiquitous.

An STS-based radiation hybrid map of the human genome

We have constructed a physical map of the human genome by using a panel of 83 whole genome radiation hybrids (the Stanford G3 panel) in conjunction with 10,478 sequence-tagged sites (STSs) derived from random genomic DNA sequences, previously mapped genetic markers, and expressed sequences. Of these STSs, 5049 are framework markers that fall into 1766 high-confidence bins. An additional 945 STSs are indistinguishable in their map location from one or more of the framework markers. These 5994 mapped STSs have an average spacing of 500 kb. An additional 4484 STSs are positioned with respect to the framework markers. Comparison of the orders of markers on this map with orders derived from independent meiotic and YAC STS-content maps indicates that the error rate in defining high-confidence bins is < 5%. Analysis of 322 random cDNAs indicates that the map covers the vast majority of the human genome. This STS-based radiation hybrid map of the human genome brings us one step closer to the goal of a physical map containing 30,000 unique ordered landmarks with an average marker spacing of 100 kb.

BAC and PAC contigs covering 3.5 Mb of the Down syndrome congenital heart disease region between D21S55 and MX1 on chromosome 21

Chromosome 21 is a model for the study of human chromosomal aneuploidy, and the construction of its physical and transcriptional maps is a necessary step in understanding the molecular basis of aneuploidy-dependent phenotypes. To identify the gene(s) responsible for Down syndrome congenital heart disease (DS-CHD), we constructed a physical map of the D21S55 to MX1 region. A bacterial artificial chromosome (BAC) library was screened using several YACs spanning the interval, and a P1-derived artificial chromosome (PAC) library was screened using radiolabeled STS PCR products and whole BACs in gap-filling initiatives. FISH confirmed the location of all BAC and PAC clones to 21q22.2-q22.3. Overlaps were established using clone-to-clone Southerns and 24 new STSs, generated from the direct sequencing of BAC and PAC ends, along with 35 preexisting STSs. Approximately 3.5 Mb of the 4- to 5-Mb D21S55 to MX1 interval is covered in 85 BACs and 24 PACs, representing fourfold coverage within the contigs. These BAC and PAC contigs are valuable reagents for isolating the genes for DS-CHD.

Cloning of a human RNA editing deaminase (ADARB1) of glutamate receptors that maps to chromosome 21q22.3

RED1 is a double-stranded RNA-specific editase characterized in the rat and is implicated in the editing of glutamate receptor subunit pre-mRNAs, particularly in the brain. Starting from human ESTs homologous to the rat RED1 sequence, we have characterized two forms of human RED1 cDNAs, one form coding for a putative peptide of 701 amino acids (similar to the shorter of two rat mRNAs) and a long form coding for a putative protein of 741 amino acids, the extra 120 bp of which are homologous to an AluJ sequence. Both forms were observed at approximately equal levels in cDNA clones and in seven different human tissues tested by RT-PCR. The human and rat short isoforms have 95 and 85% sequence identity at the amino acid and nucleotide levels, respectively. The human sequence (designated ADARB1 by the HGMW Nomenclature Committee) contains two double-stranded RNA-binding domains and a deaminase domain implicated in its editing action. Northern blot analysis detected two transcripts of 8.8 and 4.2 kb strongly expressed in brain and in many human adult and fetal tissues. ADARB1 maps to human chromosome 21q22.3, a region to which several genetic disorders map, including one form of bipolar affective disorder. Recently it was shown that heterozygous mice harboring an editing-incompetent glutamate receptor B allele have early onset fatal epilepsy. Since glutamate receptor channels are essential elements in synaptic function and plasticity and mediate pathology in many neurological disorders, and since RED1 is central in glutamate receptor channel control, ADARB1 is a candidate gene for diseases with neurological symptoms, such as bipolar affective disorder and epilepsy.

Cloning of a novel homeobox-containing gene, PKNOX1, and mapping to human chromosome 21q22.3

To contribute to the development of the transcript map of human chromosome 21 and to the understanding of the pathogenesis of Down syndrome, we have used exon trapping to identify portions of genes from pools of HC21-specific cosmids. More than 550 potential exons have been isolated to date. One such trapped exon, hmc37a09 (GenBank Accession No. X88106), was identical to a region of a human EST, L12425 (GenBank Accession No. D31072). Its predicted amino acid sequence was homologous to the homeodomain region of homeobox-containing genes. Using the trapped sequence and the EST as probes to screen human fetal brain and kidney cDNA libraries, we have cloned the corresponding full-length cDNA. This novel gene encodes a homeodomain-containing polypeptide of 436 amino acids. The most closely related sequence is that of the mouse Meis1, a PBX-like homeobox gene. The homeodomain of the novel gene is closely related to those of the mammalian PBX family and the plant Knotted1 family (involved in plant development). This gene is named PKNOX1 by the Human Nomenclature Committee. By PCR amplification, hybridization, and genetic linkage analysis using a (GT)n polymorphism in the 3'UTR, we have precisely localized PKNOX1 to chromosome 21q22.3 between markers D21S212 and D21S25 on YAC350F7. PKNOX1 is expressed in many human tissues tested by Northern blot analysis. The involvement of the PKNOX1 gene in Down syndrome and/or monogenic disorders associated with dysfunction of this gene is presently unknown. Targeted disruption of the PKNOX1 homolog in mice will enhance our understanding of its biological function in normal mammalian development.

Construction and characterization of a rad51rad52 double mutant as a host for YAC libraries

RAD52 or RAD51 recombination-deficient yeast strains stabilize otherwise unstable YACs containing ribosomal DNA or the human color vision locus (Kohno et al., 1994). Thus the RAD52RAD51 pathways(s) are apparently involved in the instability of YACs containing tandem repeat loci, presumably by promoting recombination-based deletion formation. Some other genomic loci are still unstable or unrecoverable in those strains, but we now find that greater stability is observed in a rad51rad52 double mutant strain that we have newly constructed. YACs containing a highly unstable region around DXS49 or centromeric regions throw off a variety of products in single mutants, but are much more stable in the rad51rad52 strain, which could therefore provide a better host for library construction and maintenance.

Coverage processes in physical mapping by anchoring random clones

The aim of this paper is to provide general results for predicting progress in a physical mapping project by anchoring random clones, when clones and anchors are not homogeneously distributed along the genome. A complete physical map of the DNA of an organism consists of overlapping clones spanning the entire genome. Several schemes can be used to construct such a map, depending on the way that clones overlap. We focus here on the approach consisting of assembling clones sharing a common random short sequence called an anchor. Some mathematical analyses providing statistical properties of anchored clones have been developed in the stationary case. Modeling the clone and anchor processes as nonhomogeneous Poisson processes provides such an analysis in a general nonstationary framework. We apply our results to two natural nonhomogeneous models to illustrate the effect of inhomogeneity. This study reveals that using homogeneous processes for clones and anchors provides an overly optimistic assessment of the progress of the mapping project.

CRF2-4: isolation of cDNA clones encoding the human and mouse proteins

To search for genes which encode human interferon receptor molecules, we applied the technique of exon trapping to human chromosome 21 yeast artificial chromosomes (YACs). A full length cDNA encoding human CRF2-4 was identified which is shown here to be widely expressed in human tissues. The human CRF2-4 cDNA was used as a probe to screen a mouse macrophage cDNA library. A mouse CRF2-4 cDNA was isolated which encodes a 349-amino-acid (aa) polypeptide that is 69% identical at the aa level with its human counterpart. Mouse CRF2-4 is predicted to be another member of the interferon receptor-class II cytokine receptor family.

The genomic structure of the gene defective in Usher syndrome type Ib (MYO7A)

Usher syndrome type Ib is a recessive autosomal disorder manifested by congenital deafness, vestibular dysfunction, and progressive retinal degeneration. Mutations in the human myosin VIIa gene (MYO7A) have been reported to cause Usher type Ib. Here we report the genomic organization of MYO7A. An STS content map was determined to discover the YAC clones that would cover the critical region for Usher syndrome type Ib. Three of the YACs (802A5, 966D6, and 965F10) were subcloned into cosmids and used to assemble a preliminary cosmid contig of the critical region. Part of the gene encoding human myosin VIIa was found in the preliminary cosmid contig. A cosmid, P1, PAC, and long PCR contig that contained the entire MYO7A gene was assembled. Primers were designed from the composite cDNA sequence and used to detect intron-exon junctions by directly sequencing cosmid, P1, PAC, and genomic PCR DNA. Alternatively spliced products were transcribed from the MYO7A gene: the largest transcript (7.4 kb) contains 49 exons. The MYO7A gene is relatively large, spanning approximately 120 kb of genomic DNA on chromosome 11q13.