Construction of a human chromosome 1 physical and transcription map

Simon Gregory

Sanger Centre, Hinxton, Cambridgeshire, UK

The Sanger Centre has been funded to generate the complete map and sequence of chromosome 1, the largest human chromosome. The strategy of generating a chromosome specific sequence map provides a means whereby the international collaboration to determine the sequence of the whole genome can be coordinated, along with the development of a physical and transcription map in association with the chromosome 1 community.

The physical map is being built using landmark based mapping and restriction digest fingerprinting. 5486 sequence tagged sites (STSs), 88% RH mapped, have been used to identify 30700 PAC or BAC clones. Fluorescent restriction digest fingerprinting of 23000 clones and incorporation of 14000 clones from the GSC, St Louis, has produced an estimated 95% (227 Mb) sequence contig coverage of the euchromatic region of chromosome 1. Gaps between contigs are sized using fibre FISH and closed by (i) the generation of de novo STSs from clones at the ends of contigs or (ii) using end sequences to probe large insert clone libraries.

A 'working draft' of chromosome 1 is being produced as part of the Human Genome Project. The working draft comprises genomic sequence of each bacterial clone in the physical map, determined at an average 3X depth of coverage. The draft sequence is an intermediate step prior to the 'finishing'. Finishing involves the generation of additional shotgun sequence, when necessary, and directed additional sequencing and checking to close all gaps and resolve ambiguities. Sequence produced from a finished clone provides sequence accuracy of 99.99%. As of 5th of June, the Sanger Centre has generated 131 Mb of draft sequence and 29 Mb of finished sequence.

The Sanger Centre chromosome 1 project is providing a detailed manual annotation and experimental analysis on 'finished' sequence clones. Genomic sequence analysis incorporates both in silico gene prediction and experimental homology screening. To date, analysis of finished chromosome 1 genomic sequence has identified 433 genes from 256 finished PAC/BAC sequences. Sanger Centre chromosome 1 mapping, sequence and analysis data is released freely in the public domain via http://www.sanger.ac.uk/HGP/Chr1/.

Tamara Kucaba¹, Brian Berger¹, Sara Mackerly¹, Rudy Marcelino¹, Irina Koroleva¹, Hehuang Xie¹, Sergey Malchenko¹ Jeffrey Kasperski¹, Jamila Somani¹, Ning Wu¹, Lankai Guo¹, Jennifer Laffin¹, Shereen Chang¹, Ivana Sunjevaric¹, Micca Donohue¹, Greg Doonan¹, Robert Brown¹, Christy Smith¹, Brad Johnson¹, Keith Crouch¹, Ryan Kinkaid¹, Vladan Miljkovic¹, Gretel Beck¹, Jack Gardiner¹, Chad Roberts², Clay Birkett², Kang Liu², Maria de Fatima Bonaldo¹, Todd Scheetz¹, Val Sheffield^1,4, Tom Casavant², Marcelo Bento Soares^1,3

Departments of ¹Pediatrics, ²Electrical and Computer Engineering, ³Physiology and Biophysics, and ⁴Howard Hughes Medical Institute, The University of Iowa, Iowa City, Iowa 52242 USA

Our gene discovery program relies on a strategy that we have developed and successfully implemented named "Serial Subtraction of Normalized Libraries", which is an iterative process whereby each new set of about 5-10,000 arrayed cDNAs is pooled and used as a driver in a subtractive hybridization with the library from which it was derived. The latter typically comprises a complex mixture of individually tagged normalized and/or subtracted libraries. As a result, redundant identification of ESTs is greatly minimized and the representation of rare mRNAs, likely to be missed in more random approaches, is increased after each subtraction. Central to this strategy is the fact that all starting libraries are individually tagged to enable computational identification of tissue/library of origin of each EST within a mixture. We have applied this strategy in three gene discovery projects that we are conducting [http://genome.uiowa.edu/]: (1) Rat Gene Discovery and Mapping Program, (2) Mouse Brain Molecular Anatomy Project (BMAP), and (3) Cancer Genome Anatomy Project (CGAP). To date we have identified 50,316 unique rat 3'ESTs, 28,179 unique mouse BMAP 3'ESTs, and 24,886 unique human CGAP 3'ESTs. These clones have been re-arrayed and 5'ESTs have been generated from the resulting non-redundant collections of rat, mouse and human cDNAs, respectively. All sequences and clones have been made publicly available.

Ying Zhang Chen^1,2, Eiichi Soeda¹, Jian Guo Wu^1,2, Ei-ichiro Takaoka¹, Kohei Maekawa¹, Yasuhide Hayashi², Naoki Watanabe³, Johji Inazawa³, Fumie Hosoda⁴, Yasuhiko Arai⁴, Hiroshi Mizushima⁵, Aiko Morohashi⁶, Miki Ohira⁶, Akira Nakagawara⁶, Si-Yuan Liu⁷, Masato Hoshi⁷, Akira Horii⁷

¹Gene Bank, Tsukuba Institute, Institute of Physical and Chemical Research (RIKEN), 3-1-1 Koyadai, Tsukuba 305-0074, Japan; ²Department of Pediatrics, Faculty of Medicine, University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo, 113-8655, Japan; ³Department of Molecular Cytogenetics, Division of Genetics, Medical Research Institute, Tokyo Medical and Dental University, 1-5-45 Yushima, Bunkyo-ku, Tokyo 113-8519, Japan; ⁴Cancer Gemomics Division, National Cancer Center Research Institute, 5-1-1 Tsukiji, Chuou-ku, Tokyo 104-0045, Japan; ⁵Cancer Information and Epidemiology Division, National Cancer Center Research Institute, 5-1-1 Tsukiji, Chuou-ku, Tokyo 104-0045, Japan; ⁶Division of Biochemistry, Chiba Cancer Center Research Institute, 666-2 Nitona, Chuoh-ku, Chiba 260-8717, Japan; ⁷Department of Molecular Pathology, Tohoku University School of Medicine, 2-1 Seiryo-machi, Aoba-ku, Sendai, 980-8575, Japan

Distal end of human chromosome 1p is frequent target for deletion in many human malignant tumors such as neuroblastoma, melanoma, and cancers of the colorectum, stomach, breast, liver, pancreas, and lung. Moreover, this region is also recognized as harboring other genes that are responsible and/or closely related to various human diseases. Using nearly 1000 STS and EST markers, we have constructed a BAC-based contig in 1p36-p35 consisting of nearly 1400 BAC clones. This region is estimated to be nearly 35-Mb in length, and our contig covered the most of the region. The longest contig is estimated to be nearly 20-Mb. This contig map is available from our web site, aiming to be of great help in the final stage of human genome sequencing project.

John D. Carpten for the Prostate Investigation Group

NHGRI/NIH, Bethesda, MD

Our group recently published a detailed physical and transcription map of the region on 1q24-q31 encompassing the HPC1 locus (Carpten et al., Genomics, 2000). In addition to the HPC1 locus, hyperparathyroidism-jaw tumor syndrome has been genetically mapped to this region of the genome. The positional candidate approach is being used to identify the HPC1 gene, as well as the HPT-JT gene.

With the recent advancements of the Human Genome Project, we have abandoned traditional physical mapping techniques and have adopted methods aimed at utilizing the large amounts of draft human genome sequence deposited in the NCBI database. To identify sequence contigs in our regions of interest InterAlu PCR of YACs was performed to identify novel sequence tagged sites (STS). These STSs are mapped back to the YAC contig for relative order information. These sequences are then searched against the high-throughput genome sequence database (NCBI) using the BLAST algorithm.

To aid in the analysis of newly identified genome sequence contigs, we are using a bioinformatics tool called Genemachine, designed in the Genome Technology Branch, NHGRI. Genemachine is a suite of genome analysis programs developed to aid in the identification of both genetic and transcriptional information from large genomic sequence contigs. Genemachine runs RepeatMasker to prepare sequence for BLAST sequence alignments to search for STSs and ESTs contained within these genomic contigs. Genemachine also runs a simple sequence repeat identification program called Sputnik to facilitate the generation of novel polymorphisms to be used for genetic refinement of candidate loci. Also contained within Genemachine are the FGENES, GENSCAN, MZEF, and GRAIL exon prediction algorithms. Taken together, this suite of programs is a very informative and comprehensive genome analysis tool. Genemachine data outputs are in the ASN.1 file format, which allows for sequence and graphical viewing of the data, using Sequin (NCBI). We are using the strategy described above to aid in the identification of the HPC1 and HPT-JT genes.

Miki Ohira¹, Hajime Kageyama¹, Motohiro Mihara¹, Shigeyuki Furuta¹, Taiichi Machida¹, Tomotane Shishikura¹, Hajime Takayasu¹, Ashraful Islam¹, Yohko Nakamura¹, Masato Takahashi¹, Nobumoto Tomioka¹, Shigeru Sakiyama¹, Yasuhiko Kaneko², Atsushi Toyoda³, Masahira Hattori³, Yoshiyuki Sakaki⁴, Misao Ohki⁵, Akira Horii⁶, Eiichi Soeda⁷, Yasuhide Hayashi⁸, Johji Inazawa⁹, Naohiko Seki¹⁰, Hidekazu Kuma¹¹, Iwao Nozawa¹¹, Akira Nakagawara¹

¹Division of Biochemistry, Chiba Cancer Center Research Institute, Chiba 260-8717; ²Department of Cancer Chemotherapy, Saitama Cancer Center Hospital, Saitama 362-0806; ³RIKEN Genomic Sciences Center, Sagamihara, Kanagawa 228-8555; ⁴Human Genome Center, Institute of Medical Science, University of Tokyo, Minato-ku, Tokyo 108-8639; ⁵Cancer Genomics Division, National Cancer Center Research Institute, Chuoh-ku, Tokyo 104-0045; ⁶Department of Molecular Pathology, Tohoku University School of Medicine, Sendai 980-8575; ⁷RIKEN Gene Bank, Tsukuba, Ibaraki 305-0074; ⁸Department of Pediatrics, Faculty of Medicine, University of Tokyo, Bunkyo-ku, Tokyo 113-8655; ⁹Department of Molecular Cytogenetics, Division of Genetics, Medical Research Institute, Tokyo Medical and Dental University, Bunkyo-ku, Tokyo 113-8519; ¹⁰Biological Technology Laboratory, Helix Research Institute, Inc., Kisarazu, Chiba 292-0812; ¹¹New Technology Development Department, Central Research Center, Hisamitsu Pharmaceutical Co., Inc., Tsukuba Research Laboratories, Tsukuba, Ibaraki 305-0856; Japan

The frequent deletions of the distal part of 1p in many human cancers indicate the presence of a tumor suppressor gene(s) on this chromosome region. In neuroblastoma (NBL), loss of 1p36.2-36.3 is frequently occurred and significantly associated with MYCN amplification and poor prognosis. Here, we present the identification of a homozygous deletion at the marker D1S244 within the smallest region of overlap at 1p36.2-p36.3 in two NBL cell lines, although our genotyping has suggested the possibility that both lines are derived from the same origin. This is the first instance of a homozygous deletion in1p36 region in NBL. The 800-kb PAC contig covering the entire region of homozygous deletion was constructed and partially sequenced (about 80%). The estimated length of the deleted region was approximately 500-kb. We have so far identified 6 genes within the region which include three known genes as well as three other genes which have been reported during processing our present project for the last 3 years and a half. These genes are related to apoptosis, glucose metabolism, ubiquitin-proteasome pathway, a neuronal microtubule-associated motor molecule and biogenesis of peroxisome. Semi-quantitative RT-PCR indicated that at least three genes were differentially expressed at high levels in favorable NBLs and at low levels in unfavorable subsets of primary NBLs. Although RT-PCR-SSCP analysis has demonstrated infrequent mutation of the genes so far identified, those differentially expressed genes could be the new members of the candidate NBL suppressor, because some of the 1p distal region is reported to be imprinted. Full-sequencing and gene prediction for the region of homozygous deletion would elucidate more detailed structure of this region and might lead to discovery of additional candidate genes.

Frida Abel, Tommy Martinson

Department of Clinical Genetics, Sahlgrenska Univ. Hosp.-East, S-41685 Gothenburg, Sweden (tommy.martinsson@clingen.gu.se)

Caspase-9, a key player of the apoptotic pathway, was recently mapped to chromosome region 1p36.2-3 (HGM locus: CASP9). This region has been alleged to involve one or several tumor suppressor genes in human neuroblastoma tumors. In the current study we intended to investigate whether a coding mutation of the CASP9 gene could be one of the events involved in development or progression of primary neuroblastoma tumors. Forty-seven primary tumors were screened for mutation by sequence analysis of the CASP9 gene. Three polymorphisms could be detected in the coding region, and two of them caused an amino acid substitution. One rare polymorphism, detected in 4 out of 47 cases, gave rise to an exchange of a polar to a non-polar amino acid. One tumor was homozygous for the rare allele of this polymorphism. However, this polymorphism could be detected with nearly the same frequency in normal controls (2 out of 49) in the heterozygous form. CASP9 and the alternative isoform CASP9S was expressed in all stages of neuroblastomas. A slightly lower expression level could be seen in high stage neuroblastoma tumors (stage 3 and 4) The present study shows that coding mutations of the CASP9 gene is not a frequent event in neuroblastoma tumors.

T. Carling^1,2, W. Fang¹, P. Correa¹, S. Huang¹

¹ The Burnham Institute, La Jolla, Ca; ²Dept. of Surgery, Uppsala University Hospital, Uppsala, Sweden

Most molecular genetic abnormalities contributing to the development of pheochromocytomas and parathyroid tumors remain unknown. However, allelic loss at chromosome 1p has been found to be the dominant genetic alteration in these tumors, suggesting that a hitherto unidentified tumor suppressor gene is located at this locus. To help with positional cloning of the candidate suppressor, we here studied whether the locus of the candidate tumor suppressor RIZ may lie within a minimal region of deletion in these tumors. Thirteen parathyroid adenomas, previously known to display loss of heterozygosity (LOH) at chromosome 1, and 23 pheochromocytomas (20 sporadic and 3 familial cases; 19 benign and 4 malignant) were included into the study. Four highly polymorphic microsatellite markers in the immediate vicinity of the RIZ (PRDM2) gene (D1S434, D1S489, D1S228, and D1S507), as well as the intragenic Pro704 deletion and CA repeat polymorphisms were analyzed for LOH. In total, 8/13 (62 %) of the parathyroid tumors and 9/23 (39%) of the pheochromocytomas displayed allelic loss at atleast one of the investigated markers. Interestingly, in four of the parathyroid adenomas, LOH at the intragenic Pro704 deletion polymorphism was the sole genetic abnormality, thus mapping a notably small region of minimal deletion to the RIZ locus. The Pro704 deletion polymorphism was frequently deleted also in the pheochromocytomas, but the minimal deleted region in these tumors is yet to be mapped. No association between malignancy and LOH at 1p36 was seen in the pheochromocytomas. Our results confirm previous finding that loss of chromosome 1p36 is frequent in tumors of the parathyroid and adrenal medulla, and suggest that the RIZ gene is located within a minimal deleted region of parathyroid tumors. Whether the non-deleted allele of RIZ is inactivated by genetic or epigenetic events remains to be investigated.

J.M. McDonald, G.N. Fuller, W. Zhang

Cancer Genomics Laboratory, Dept. of Pathology, University of Texas M.D. Anderson Cancer Center, Houston, Texas, U.S.A

Gliomas are the most prevalent type of primary brain neoplasm, accounting for greater than 40% of all primary brain tumors. There are several currently recognized glioma subtypes, including low grade oligodendroglioma, anaplastic oligodendroglioma, low grade astrocytoma, anaplastic astrocytoma, and glioblastoma. Classification of the gliomas is currently based on pattern recognition of various characteristic morphologic features, such as tumor cell size, shape, and cytopasmic configuration, and grading is based on additional features such as the presence or absence of pleomorphism, mitotic figures, vascular proliferation, and necrosis. This diversity suggests multiple genetic etiologies. One type of genetic alteration, loss of heterozygousity (LOH), has been observed for chromosomal regions 1p36 and 19q in 80-90% of oligodendrogliomas. This finding is of clinical significance; oligodendrogliomas with 1p36 and 19q LOH respond significantly better to combination chemotherapy compared to oligodendrogliomas that retain heterozygousity at 1p36 and 19q. There is strong evidence for tumor suppressor genes in these chromosomal regions that have developmental effects for some gliomas. 1p36 and 19q LOH has also been seen in anaplastic astrocytoma and glioblastoma samples, but at much lower frequencies. This suggests that, in addition to classification based on morphology, the genetic signatures of individual tumors are potentially useful as adjuncts for tumor diagnosis and treatment.

One of the major objectives of this project is to identify genetic markers on 1p36 for glioma subclassification. Sequencing of 1p36 is underway as part of the Human Genome Project and there are working drafts and completed clone sequences for this region. BLAST searches of clone sequences against various databases (dbEST, nr, etc.) are being performed in order to construct a microarray that contains genes and ESTs from the region of interest. Following microarray construction, RNA samples from the recognized glioma subtypes will be used for gene expression profiling. Genetic markers may thus be obtained that will help to predict tumor response to various forms of therapy.

Additionally, the genetic markers used for glioma subclassification will also serve as candidate tumor suppressor loci. DNA extracted from the same samples used in the gene expression profiling studies can be used in mutation analyses of these candidates. Observation of alterations in tumor DNA at these loci will provide further evidence for a suppressor gene role. By identifying a better screening process for tumors that will or will not respond to various forms of therapy, and by identifying candidate tumor suppressor loci, we hope to improve treatment and overall long-term survival of glioma patients.

C. Malchoff, M. Sarfarazi, B. Tendler, F. Forouhar, V. Joshi, A. Arnold, D. Malchoff

University of Connecticut Health Center, Farmington, CT 06030-1110

Papillary thyroid carcinoma (PTC) usually is sporadic, but may occur in a familial form. The familial PTC (fPTC) susceptibility gene(s) constitutes the earliest change of thyroid tumorigenesis. The fPTC clinical phenotype has not been described completely, and the susceptibility gene(s) is unknown. A large fPTC kindred was investigated to more fully characterized the clinical phenotype, to determine the chromosomal location of the disorder, and to evaluate candidate genes within the linkage region. Of 28 living available subjects in 3 generations, 8 were affected and 2 were obligate PTC carriers. Inheritance was autosomal dominant with partial penetrance. In addition to the known association with benign thyroid nodular disease, there was an association with the rare entity of papillary renal neoplasia (PRN). One PRN subject was affected with PTC and the other, with multifocal PRN, was an obligate PTC carrier. Linkage analysis of specific candidate genes and a genome wide screen excluded MET, the protooncogene of isolated familial PRN, and genes known to cause familial tumor syndromes enriched in PTC. The fPTC/PRN syndrome was mapped to a chromosome 1 region that spans the centromere. A maximum three-point LOD score was 3.58 for markers D1S2342 and D1S2345 and for markers D1S2343 and D1S305. Critical recombination events limited the linkage region to 22 cM between markers D1S3009 and D1S2721. Candidate genes within the linkage region include the neuroblastoma ras oncogene homolog (NRAS), the papillary renal cell carcinoma gene (PRCC) and the high affinity neurotrophin receptor gene (NTRK1). NRAS becomes tumorigenic in thyroid, when it acquires activating mutations. PRCC becomes tumorigenic in kidney when fused to TFE3, a nuclear factor on the X chromosome. NTRK1 becomes tumorigenic in thyroid, when gene fusion effects illicit expression of this tyrosine kinase in thyroid follicular cells. Analysis of karyotype, genomic DNA sequence, and protein expression in premalignant thyroid follicular cells excluded these candidates as the fPTC/PRN gene. We conclude that the fPTC/PRN syndrome is a distinct familial tumor syndrome. The fPTC/PRN susceptibility gene is within a 22 cM chromosome 1 region, and is a gene not previously implicated in thyroid or renal tumorigenesis.

A. Forus¹, L.A. Meza-Zepeda¹, A.B. Dahlberg¹, L. Godager¹, A. South ², D. Nizetic², I. Marenholz³, M. Lioumi⁴, I. Ragoussis⁴, G.M. Maelandsmo¹, V.A. Florenes⁵, J. Henriksen¹, J. Nesland⁵, M. Tarkkanen⁶, S. Knuutila⁶, M. Serra⁷, D. Mischke³, O. Myklebost¹

¹Dept. of Tumour Biology, The Norwegian Radium Hospital, Oslo, Norway; ²Centre for Applied Mol. Biology, School of Pharmacy, University of London, UK; ³Institut für Immungenetik, Charite Humboldt- Universität, Berlin, Germany; ⁴Division of Medical and Molecular Genetics, Guy's Hospital, London, UK; ⁵Dept. of Pathology, The Norwegian Radium Hospital, Oslo, Norway; ⁶Department of Medical Genetics, Haartman Institute, University of Helsinki, Helsinki, Finland; ⁷Laboratory of Oncology Research, Istituti Ortopedici Rizzoli, Bologna, Italy

Recurrent gains of different regions on the long arm of chromosome 1 are among the most common alterations in human malignancies. These gains are particularly frequent in bladder cancer, hepatocellular carcinomas, invasive cervical carcinomas, ovarian carcinomas and different subtypes of sarcomas and breast carcinomas, among others. In sarcomas, the gains are localised to 1q21-22. We have previously identified the most frequently amplified part of 1q21, and identified a highly amplified marker (Yeast Artificial Chromosome, YAC). The amplified YAC was used to capture highly expressed cDNAs from an osteosarcoma cell line with high-level amplification of 1q21-q22. This approach resulted in the cloning of three novel cDNAs from this region, COAS 1, 2 and 3 (Chromosome One Amplified Sequences). Sequencing of the cDNA clones revealed that COAS1 is a large gene with many repeats, but we found no homology between COAS1 and any known gene. COAS2 was a novel member of a family of proteins involved in stress response, chemotherapy resistance and differentiation of muscle, bone and adipocytes. Interestingly, another amplified member of the COAS2 family have been cloned from the 20q13 amplicon in breast carcinomas. The third gene, COAS3, showed homology to profilin 1. Members of this protein family are associated with the cytoskeleton and may influence processes like migration, invasion or metastasis. The three genes were transcribed in multiple normal tissues, were highly expressed and amplified in sarcomas and breast carcinomas, but generally showed lower amplification levels in breast cancer samples. This is not surprising since gain of the whole 1q is the most frequent finding in breast cancer, whereas the local 1q21-q22 amplification is occasionally observed. COAS1, 2 and 3 are so far the best candidate target genes for the 1q21 amplifications. Further studies are warranted to predict their role in tumour progression and/or development.

V. Dyomin, W. Chen, N. Palanisamy, K. Lloyd, K. Dyomina, S. Jhanwar, J. Houldsworth, R.S.K. Chaganti

Memorial Sloan-Kettering Cancer Center, New York

The band 1q21 is one of the most frequent sites affected by chromosomal translocations in lymphomas as well as in tumors of other lineages. In non-Hodgkin's lymphoma (NHL), translocations and duplications affecting this chromosomal site are usually seen in association with primary chromosomal abnormalities suggesting a role for 1q21 rearrangements in tumor progression. Three different translocation breakpoints at 1q21 have been cloned so far accounting for less than 10% of all breakpoints. The heterogeneity of the breakpoints suggests that multiple genes are involved. As a part of an effort to identify all 1q21 genes affected in NHL we report here the mapping and cloning of two different breakpoints in two cases of Diffuse Large B-cell Lymphoma with t(1;14)(q21;q32) translocations. The breakpoints on the der(1) and der(14) chromosomes were mapped by FISH and Southern blot analysis and subsequently cloned using genomic phage library screening with IGH probes or inverse PCR technique. In one of the cases the translocation linked the IGH sequences to MUC1 gene, leading MUC1 transcriptional unit intact and leading to its ectopic activation by the IGH enhancer. MUC-1 mucin has previously been shown to be frequently overexpressed in human epithelial cancers and associated with tumor progression and poor clinical outcome. In the other case the translocation mapped to the FCg receptor cluster region, resulting in activation of an aberrant isoform of FCGIIB. Screening of a panel of B-cell lymphomas by Southern blot analysis showed that both breakpoints are recurrent in NHL. Strategies for a utilizing the chromosome 1 mapping and sequencing data to precisely map all 1q21 breakpoints in NHL and to identify all 1q21 genes involved in lymphomagenesis will be discussed.

Mary Callanan, Patricia Le Baccon, Samuel Duley, Danielle Marais, Jean Jacques Sotto, Dominique Leroux

Lymphoma Research Group, Institut Albert Bonniot and the Departments of Cancer Genetics and Clinical Hematology, Grenoble University Hospital, La Tronche 38706, France

1q rearrangement is a frequent secondary chromosomal change in non-Hodgkin's lymphoma (NHL) and may thus be of pathological significance in tumor progression. The target genes of these rearrangements remain largely unknown. In this setting, we have used karyotyping and fluorescence in situ hybridization (FISH) to search for recurrent 1q breakpoints in a consecutive series of 31 NHL patients and 4 B-cell tumor lines with 1q abnormalities. By karyotyping a total of 43 1q12-q24 breaks were detected. FISH mapping of these breaks (whole chromosome painting, 1q21-24 CEPH YACs and a 1q12-specific probe) revealed the majority to localize to three novel breakpoint sites, heterochromatin band 1q12 (40% of breaks), and two adjacent sites (cluster I and II) within a 2.5 Mb region of proximal 1q21 (30% of breaks). The 1q12 breaks (mostly mis-diagnosed as 1q21 breaks by karyotyping) were frequent in both follicular and diffuse large cell lymphoma and placed 1q12 heterochromatin adjacent to potentially coding sequence on either chromosome 1 (dup1q) or other partner chromosomes. This is the first evidence of a pathological role for 1q12 constitutive heterochromatin (1q12 gain and/or heterochromatin/euchromatin junctions) in lymphoma progression. Potential pathogenic mechanisms may be linked to the ability of constitutive heterochromatin to modify gene expression (transcriptional silencing).

The two remaining novel cluster sites fell within an intensively mapped region of 1q21 that has previously been shown by karyotyping and FISH to be rearranged/amplified in a variety of tumor types. The first 1q21 breakpoint cluster (cluster I), which appears to border the 1q12 heterochromatin band, mapped to YAC 934G9 (1.2Mb; spans markers D1S3620 to WI5663, (1,2)) and was frequent in high grade lymphoma. The second 1q21 breakpoint hotspot (cluster II) localized to an interval of approximately 500 kb that was flanked by YACs 882B3 / 776H9 (span markers WI-5663 to D1S3623, (1,2)). Cluster II breaks were associated with low grade lymphoma and potentially target BCL9 and AF1q both of which lie centromeric to the S100/EDC locus and MUC1. Southern blotting detected a single case of AF1q rearrangement and no BCL9 rearrangements thus raising the possibility of at least one other oncogene target in this region.

Greg Denomme, Kimberly Matheson

Department of Pathology and Laboratory Medicine, Mount Sinai Hospital, Department of Laboratory Medicine and Pathobiology, University of Toronto, and Canadian Blood Services, Toronto, Canada

The RH blood group locus is comprised of two highly homologous genes; RHD and RHCE. In the early 1990's, Cherif-Zahar and MacGeoch independently assigned the locus to region 1p36.2-34 using cytogenetic analysis. The unique feature of this locus is that the RhD-negative phenotype is characterized by the complete deletion of the RHD gene. Wagner and Flegel used homology analysis from genomic sequences combined with long PCR to determine the relative orientation and position of these genes. RHD and RHCE are oriented with their 3' poly-A tails towards one another separated by ~30 kilobasepair of intervening DNA. The intervening region contains the unrelated gene SMP1. The 3'-end of SMP1 overlaps with the 3'-end of RHCE by 58 basepair. Complete RHD gene deletion is likely due to recombination of the 'Rh boxes' that flank the gene (Blood 2000;95:3662). Now, we extended these observations using NCBI Map View along with contiguous segments and STS markers to determine the chromosomal orientation of the RH locus. A query for RHD shows that contigs NT_001684 and NT_002478 are assigned to this region. NT_001684 and NT_002478 contain the STS markers stSG3884 and stSG34182, respectively, and other additional markers. Electronic PCR for these STS markers indicates that stSG3884 is contained in genomic sequence AL031432, which was sequenced from clone 465N24 that contains a region homologous with the Rh boxes. stSG34182 is found in genomic sequence AL031284, which was sequenced from clone RP3-469D22. Wagner and Flegel did not assigned RP3-469D22 to either the RHD or RHCE gene. However, a portion of RP3-469D22 is 100% homologous with the 5' promoter region of RHCE but contains 3 basepair differences with RHD confirming the origin of the clone as RHCE. Therefore, from cloned sequence data, mapped contigs NT_001684 and NT_002478 and the relative orientation of RHD and RHCE, the presumed chromosomal orientation of this region is p-ter-RHD-SMP1-RHCE-cent.

E. Eden¹, R. Naumova ², X.-M. Sun¹, A.K. Soutar¹

¹Lipoprotein and ²Molecular Medicine Groups, MRC Clinical Sciences Centre, Imperial College School of Medicine, Hammersmith Hospital, Ducane Road, London W12 ONN

Background: Familial hypercholesterolaemia (FH) is an autosomal codominant disorder characterised clinically by high levels of low density lipoprotein (LDL), the major cholesterol-transport lipoprotein in human plasma. This leads to accelerated atherosclerosis and thus an increased risk of coronary heart disease. FH is generally caused by defects in genes for the LDL receptor or apolipoprotein B (apoB). Two unrelated clinically homozygous patients, both offspring of consanguineous unions, have been identified whose cells in culture exhibit negligible degradation of LDL, yet have no defect in the LDL receptor or apoB genes. LDL receptors have been shown to be present on the cell surface and to bind LDL normally but appear not to be internalised (Norman et al., J Clin Invest, 104, 619-628).

Objective: To identify the chromosomal localisation of the defect in these families using homozygosity mapping.

Methods: Fluorescence-labelled primer pairs flanking polymorphic markers were used to amplify genomic DNA in 10 members of the largest family (pedigree1). PCR products were sized and analysed using an ABI 377 sequencer with Genescan and Genotyper software. Allele frequencies in the population were estimated in unaffected unrelated individuals of the same racial origin. Linkage analysis was performed using Genehunter software.

Results: A genome-wide scan using 500 markers spaced at approximately 10cM intervals revealed only one region of shared homozygosity exclusively in the affected individuals. Genotyping of a second, unrelated family with the same phenotype provided further evidence for linkage to this locus, mapping the disease-causing gene to chromosome 1p36-p35 with a combined LOD score of 4.0 in these families.

Conclusions: We have localised the defect for the recessive FH phenotype in these families to a 10.5-cM interval at 1p36-p35. Although this region contains approximately 100 known genes and 100 ESTs, there are no obvious candidate genes, but the full sequence is not yet available.

Erik P. Sulman¹, Peter S. White^1,2, Jan P. Dumanski³, John M. Maris^1,2, Garrett M. Brodeur^1,2

¹Division of Oncology, Children's Hospital of Philadelphia, PA; ²Dept. of Pediatrics, University of Pennsylvania School of Medicine, Philadelphia, PA; ³Dept. of Genetics and Pathology, Uppsala University, Sweden

Meningioma is a frequently recurrent, common tumor of the meninges surrounding the central nervous system. Hemizygous deletion of the short arm of chromosome 1 (1p) is second only to monosomy of chromosome 22 as the most frequently observed chromosomal alteration in these tumors. We conducted loss of heterozygosity (LOH) analysis of 157 matched normal and tumor DNAs and observed LOH in 35% of tumors, identifying a smallest region of overlapping deletion (SRO) spanning 1.5 centimorgans on 1p32. 1p LOH was predictive of tumor recurrence and correlated with loss of chromosome 22 and mutation of the NF2 tumor suppressor gene (TSG). Deletion of this region has been observed in other tumor types, such as oligodendroglioma and neuroblastoma. We hypothesized that the 1p SRO harbors one or more TSGs important for tumor development or progression. CompView, a high-resolution comprehensive map of chromosome 1, was used as a basis for constructing a physical map of the region. Localization of unmapped genes to the region was aided by the use of a somatic cell hybrid panel containing translocation chromosomes involving 1p whose breakpoints map to the boundaries of the SRO. Currently, we have collected 80 large-insert clones assembled into 4 contigs spanning approximately 3 megabases (Mb). Forty of these clones have unfinished sequence available in Genbank, with finished sequence available for four clones. A total of approximately 6 Mb of sequence is currently available. Preliminary analysis of the sequence has identified additional, previously unmapped genes in the SRO. To identify candidate TSGs by expression profiling, we are constructing a low resolution, glass slide cDNA microarray containing all mapped transcripts in the SRO, as well as other genes previously proposed to play a role in meningioma tumorigenesis, such as NF2, LIF, and ADTB1. Analysis of the expression differences between a normal meninges cell line, LTAg2B, and a meningioma cell line, KT21MG1, using Atlas Cancer Gene blots identified additional genes of interest which are included on the microarray. Ultimately, we hope to obtain a panel of candidate transcripts exhibiting tumor-specific changes in expression that can be examined further for genomic rearrangements or for mutations in tumors with characteristic deletions in the SRO.

M Varret¹, M Abi-Fadel¹, L Villéger¹, J-P Rabès¹, M Devillers¹, M Krempf², M Martinez³, C Junien¹, C Boileau¹

¹INSERM U383, Paris; ²CHU Hôtel Dieu, Nantes; ³INSERM U358, Paris, France

Autosomal Dominant Hypercholesterolemia (ADH), one of the most frequent hereditary disorders, is characterized by an isolated elevation of LDL particles that leads to premature mortality from cardiovascular complications. It is generally assumed that mutations in the LDLR and APOB genes account for ADH. However, we have shown that ADH is genetically more heterogeneous than conventionally accepted. We identified 13 Caucasian ADH families in which we excluded linkage to the LDLR and APOB genes thus demonstrating the implication of a new locus we named « FH3 ». Genetic linkage was obtained in 4 pedigrees for chromosome 1 markers, localizing the FH3 locus in a 9 cM interval at 1p32-p34.1. By radiation hybrid mapping, four candidate genes (FABP3, SCP2, APOER2 and EPS15) at 1p32-p34.1 were located outside the critical region, demonstrating no identity with the FH3 gene. Subsequently, Hunt et al. (Arterioscler Thromb Vasc Biol, 2000, 20: 1089-93) confirmed this localization by linkage analyses in a large Utah pedigree. Finally, we performed heterogeneity tests that estimated that only 27% of our non-LDLR/non-APOB ADH families were linked to the FH3 locus. These results indicated the implication of a fourth locus called « FH4 » that we are currently mapping by linkage analysis. To refine the localization of the FH3 gene, we have tested other regional polymorphic markers in the 4 originally linked pedigrees. In the HC2 family, the FH3 gene is now linked to a conserved haplotype defined by markers tel-D1S2892-D1S2722-D1S2645-cen that span an area of 2.8 cM. We are currently testing 10 new ADH families for which linkage was excluded to the LDLR and APOB genes. This should enable us to further shorten the genetic interval in which the FH3 gene is located. In parallel, we are analysing available sequence data for this interval, predicting exon contents and identifying regional candidate genes.

Francesca Capon¹, Sabrina Semprini¹, Bruno Dallapiccola², Giuseppe Novelli¹

¹Dept. of Biopathology, "Tor Vergata" University of Rome, Italy; ²Dept. of Experimental Medicine "La Sapienza" University of Rome, Italy

Psoriasis is a chronic skin disorder affecting approximately 2% of the Caucasian population. Family clustering of the disease is well established and non-parametric linkage analyses have mapped disease susceptibility loci on chromosomes 6p (PSORS1) and 17q (PSORS2). Non confirmed evidence for linkage is also available for chromosomes 2q 3q, 4q (PSORS3), 8q, 16q and 20p. We mapped an additional susceptibility locus on chromosome 1q21 (PSORS4), within the region of the Epidermal Differentiation Complex, including more than 25 genes involved in epithelial growth and differentiation. In this study, we have carried out a linkage disequilibrium statistical analysis, in order to achieve a finer localization of the susceptibility locus. We have recruited 79 Italian triads and typed them at five loci spanning the 1.6 Mb region generating the highest multipoint lod scores in our previous linkage study. We have observed significant evidence for association with D1S2346 marker (p = 0.004). Moreover we have identified two haplotypes including markers D1S2346 and D1S2715 and showing a preferential association with the disease (p = 0.002). Besides, we have typed markers D1S2346 and D1S2715 in a sample of 30 patients and 50 healthy controls, originating from the genetic isolate of Sardinia. Thus, we have observed p values of 0.02 with both markers. To obtain further confirmation of LD data, we have undertaken the isolation of additional markers from the examined region. The BAC clone 140J1 containing D1S2346 was obtained from the HGMP, and a vectorette library was prepared. So far, one polymorphic marker has been isolated, using the "microsatellite rescue" method. A preliminary analysis of a sample of control individuals has demonstrated an heterozigosity of 0.65. The typing of this marker in our PS data-set is currently in progress.

Work funded by the Italian Ministry of Health and MURST. The authors wish to thank the HGMP for providing BAC clones.

Takashi Shiina¹, Asako Ando¹, Yumiko Suto², Fumio Kasai², Atsuko Shigenari¹, Nobusada Takishima¹, Eri Kikkawa¹, Kyoko Iwata¹, Yuko Kuwano¹, Yuka Kitamura¹, Yumiko Matsuzawa¹, Masahiro Nogami¹, Hisako Kawata¹, Yasuhito Fukuzumi³, Masaaki Yamazaki³, Hiroyuki Tashiro³, Gen Tamita¹, Atsushi Kohda⁴, Katsuzumi Okumura⁴, Toshimichi Ikemura⁵, Eiichi Soeda⁶, Nobuhisa Mizuki⁷, Minoru Kimura¹, Seiamak Bahram⁸, Hidetoshi Inoko¹

¹Department of Genetic Information, Division of Molecular Life Science,Tokai University School of Medicine, Bohseidai, Isehara, Kanagawa 259-1193, Japan; ²Department of Biological science, Graduate School of Science, The University of Tokyo, 7-3 Hongo, Bunkyo-ku, Tokyo 113-0033, Japan; ³Bioscience Research Laboratory, Fujiya Co., Ltd., Soya, Hadano, Kanagawa 257-0031, Japan; ⁴Faculty of Bioresources, Mie University, Tsu, Mie 514-0008, Japan; ⁵Department of Evolutionary Genetics, National Institute of Genetics, 1111 Yata, Mishima, Shizuoka 411-0801, Japan; ⁶Tukuba, Life Science Center, The Institute of Physical and Chemical Research (RIKEN), 3-1-1 Koyadai, Yatabe-choh, Tsukuba, Ibaraki 305-0861, Japan; ⁷Centre de Recherche d'Immunologie et d'Hˆ©matologie, Rue Kirschleger, 6708 5, Strasbourg, France; ⁸Department of Ophthalmology, Yokohama City University School of Medicine, 3-9, Fuku-ura, Kanazawa-ku, Yokohama, Kanagawa 236-0004, Japan

A series of paralogous gene clusters were localized in four extensive regions on human chromosomal bands, 1q21-q25, 6p21.3 (HLA), 9q33-q34, and 19p13.1-p13.4. These four regions were supposed to be created by two rounds of large scale duplication around the time of vertebrate emergence. Among them, it is suggested that the 1q21-25 region is most evolutionally related to the 6p21.3 HLA region, because this is the only paralogous region that includes the HLA genes (non-classical MHC class I genes, CD1 and MR1), outside the chromosome 6 HLA region. Here, in order to clarify genetic structure of this chromosome 1 paralogous region and also to investigate how similar gene organization is to that of other three paralogous regions, we constructed a 3.7 megabase (Mb) YAC, BAC, and PAC contig in the chromosome 1q22-23. The contig includes the five CD1 genes between the +P5 and FASL in which a 1.1 Mb bp BAC and PAC contig between CD1D and FCER1A was subjected to nucleotide sequence determination. A total of 66 genes (40 expressed and 26 pseudo genes; 39 known genes and 27 new genes) were mapped in this 3.7 Mb region. Among them, 31 were new genes elucidated by genome sequencing, including 20 olfactory receptor genes. Further, 33 genes such as CD1A-E, SPTA1, FCERIA, APCS, CD48, LY9, PBX1, RXRG and POU2F1 along with 20 olfactory receptor genes belong to multi-gene families, which have paralogous genes in other three duplicated regions (6p21.3, 9q33-q34, and 19p13.1-p13.4). Furthermore, it is noteworthy that 23 of the 40 expressed genes in the 1q22-q23 region around the CD1 genes are of immunological importance, such as CD1D, CD1A, CD1C, CD1B, CD1E, SPTA1, MNDA, IFI-16, AIM2, FY, BL1A, FCERIA, CRP, APCS, LYAM, CD48, LY9, SCM, CD3Z, ATAC, ELAM1, LNHR and FASL as similarly observed in the HLA region. These genes around the CD1 genes may be dedicated to some specialized immune function associated with the CD1 mediated lipid or sugar presentation pathway.

Sumera Hasham, Shao-Qing Kuang, Martin D. Phillips, David Wolf, Ying Wan, Perumal Thiagarajan, Dianna M. Milewicz

Department of Internal Medicine, University of Texas-Houston Medical School, Houston, TX 77030

A large east Texas family with autosomal dominant inheritance of a novel bleeding disorder has been identified. The disorder is characterized clinically by easy bruising, life threatening bleeding with trauma or surgery, and menorrhagia in affected women. Laboratory studies demonstrated prolongation of the prothrombin time (PT) and activated partial thromboplastin time (aPTT) in affected individuals. Paradoxically, assays of known coagulation factors are all within normal limits. To determine the molecular basis of this disease, we undertook a candidate gene linkage analysis in these kindred. We initially hypothesized that the cause of the disease in this family could be an antithrombin III mutation that resulted in a constitutively active antithrombin III in the absence of heparin binding. Linkage studies using DNA from the family and an intragenic polymorphic marker within AT3 showed that the disease mapped to this locus. The coding region and intron/exon junctions of AT3 were sequenced using the proband's DNA but this analysis failed to identify a mutation. Further family members were recruited for the study and 16 polymorphic markers around the AT3 gene were analyzed. Using 2 recombinants, we have narrowed the critical interval for the defective gene to approximately 1.5 cM, centromeric to AT3. Based on chromosome 1 contigs 13 & 195 (Sanger Center, http://webace.sanger.ac.uk/cgi-bin/ace/simple/1ace) and the human gene map fingerprint database (Washington University and Baylor College of Medicine), we have constructed a contig for our region. Multiple gene prediction software, including GENESCAN and FGENEH were used to identify known and novel genes in this region. The Factor V gene (FV) mapped into the disease interval and was sequenced but no disease causing mutations were found. There are no known candidate genes in this region and characterization of the predicted genes is in progress. Elucidation of the genetic defect causing the bleeding disorder in this family may reveal a novel protein involved in the coagulation cascade.

B.C. Bjork¹, B. C. Schutte^1,2, Y. Watanabe², S. Gregory³, E.H. Howard⁴, M.I. Malik², M. Dixon⁴, J.C.Murray^1,2

¹Genetics PhD Program and ²Department of Pediatrics, The University of Iowa, Iowa City, IA 52242; ³The Sanger Centre, Cambridge, U.K.; ⁴University of Manchester, Manchester, U.K.

VWS is an autosomal dominant form of clefting with bilateral lower lip pits. The VWS gene was localized to a 1.6 cM region of 1q32 between D1S491 and D1S205. The sequence for 1.1 Mb surrounding the 350 kb VWS critical region (VWSCR) is available in two sequences of 721 kb and 269 kb, separated only by a 100 kb gap. Gene identification was performed using a suite of similarity searches and gene prediction programs included in a modified version of Genotator. Five gene prediction programs were used to identify putative exons. RepeatMasker2 was used to identify and "mask" repetitive sequences before performing BLAST homology searches against various publicly available nucleotide, est and protein databases. These analyses revealed 4 'known' genes, 14 'related' genes, 20 'predicted' genes, 3 pseudogenes and an additional 49 exons predicted by at least 2 gene prediction programs. 75 of 94 exons (80%) from 11 'known' and 'related' genes were predicted by 2 or more gene prediction programs. In addition, from these analyses we estimated values for the following genomic characteristics: minimum gene density (37 per Mb), gene size (min., 900 bp; max., >157 kb; avg., 36 kb), exons per gene (min., 2; max., 23; avg., 10), exon size (min., 30 bp; max., 6177 bp; avg., 275 bp), intron size (min., 66 bp; max., 73.2 kb; avg., 3.89 kb) and intergenic space (min., 5 kb; max., 81 kb; avg., 36 kb) all of which compare reasonably to other estimates derived for the human genome. We have also generated a contig of mouse clones from the RPCI-23 BAC library that spans the syntenic region at chromosome 1H. The Sanger Centre (U.K.) selected seven of these clones as a minimum tiling path and generated a "working draft" sequence for each clone. Comparative genomic analysis is in progress to identify new exons and regulatory sequences. To identify etiologic mutations, SSCV analysis and direct sequencing was performed on 106 exons on a panel of 90 affected individuals. A total of 25 sequence variants were identified but none of these are etiologic mutations. We continue to prioritize all 'related' and 'predicted' genes for mutation screening by obtaining cDNA sequence and expression data.

P. Fernandez¹, H. Moolman-Smook¹, P. Brink², A. Christoffels³, W. Hide³, V. Corfield¹

¹ US/MRC Centre for Molecular and Cellular Biology, University of Stellenbosch Medical School; ²Department of Internal Medicine and Tygerberg Hospital, ³South African National Bioinformatics Institute (SANBI), Cape Town, South Africa

Progressive familial heart block type II (PFHBII) is an autosomal dominant inherited cardiac conduction disorder often complicated by dilated cardiomyopathy (DCM), segregating in a South African family. Clinical features progress from atrioventricular nodal disturbances to complete heart block. Left or bi-ventricular DCM manifest late in adulthood. We mapped another conduction disorder in a South African family to chromosome (C) 19, while several international groups mapped disorders displaying conduction defects and/or DCM to various chromosomal positions. These were all PFHBII-causative candidate loci.

The aims of the study were 1) to use linkage analysis to determine the candidacy of previously described loci as cause for PFHBII and 2) to search for the causative gene within a positively identified region, using molecular and novel computer-based strategies.

Methods: Blood was drawn and DNA extracted from 8 affected, 11 unaffected and 7 clinically equivocal or non-assessed individuals. Linkage analysis at candidate loci was performed. Genetic fine mapping was pursued to reduce the target interval. A cloned contig (BAC/PAC) covering the critical region was used to generate an integrated genetic and physical map. Database searches identified known genes and expressed sequence tags (ESTs) within the critical area. Bioinformatic tools (STACK and NCBI-BLAST) were used to extend and process ESTs.

Results: Linkage analysis mapped the PFHBII-causative gene to a locus on C1q32.1-q41 (Zmax=3.05, q=0.0 at D1S205). Previously mapped candidate loci were excluded as a cause of PFHBII. Fine mapping reduced the search area to a ~3cM region, telomeric to a previously described DCM-causative locus. EST sequences were placed onto clone sequences using a modification of EST2genome. Concomitantly, sequence tag sites on ESTs were riveted onto contig clones using PCR amplification. Three plausible candidate genes (known) were identified by database searches. Additionally, ESTs (potentially novel genes) were mapped to the target area.

Conclusions: The PFHBII gene mapped to a novel disease-causing locus. An adjacent locus was excluded by recombination events. Known and novel candidate genes identified generated a comprehensive transcript map of the search area. Candidates identified in this manner will be prioritised and screened for possible PFHBII-causative mutations.

K. Hoffmann^1,2, H. Toennies², H. Karl³, R. Kaps³, D. Mueller⁴, G. Nuernberg¹, A. Reis¹, K. Sperling⁵

¹Gene Mapping Center, Max Delbrueck Center for Molecular Medicine, Berlin; ²Franz Volhardt Clinic at the Charité, Humboldt University, Berlin; ³General practitioners, Chemnitz and Gelenau; ⁴Department of Pediatrics, Klinikum Chemnitz; ⁵Department of Medical Genetics, Charité at the Humboldt University, Berlin, Germany

Purpose: The Pelger Huët anomaly is an autosomal dominant benign disorder altering granulocyte morphology. Its prevalence varies from one country to another, but usually affects about 1 in 5000 individuals. However, we found a small town in south eastern Germany where its frequency is much higher and feasible for the identification of a Pelger Huët locus by linkage analysis.

Methods: Light microscopy screening of blood smears from 4386 individuals revealed one homozygous and 53 heterozygous probands with Pelger Huët anomaly in the village Gelenau (total number of inhabitants 6530). From that pool, 49 persons from 9 obviously unrelated families have been selected for genetic analysis. The sample includes one homozygous and 30 heterozygous affected individuals. A total genome scan with 389 microsatellite markers was performed followed by fine mapping and multipoint parametric linkage analysis.

Results: Region 1q41-43 showed evidence for linkage to Pelger Huët anomaly with a lod score of 5.2. Combined linkage and haplotype data confirmed the critical region between markers D1S229 and D1S2847 (241.6 cM to 247.6). A shared trait haplotype could be identified.

Conclusion: The increased frequency of this anomaly in the village Gelenau seems to be due to a founder effect by a common ancestor. Cloning the gene might give insight to altered cell development processes that are also involved in other disorders. Acquired Pelger like granulocyte morphology is observed in some forms of cancer (e.g. leukemia), infections (e.g. tuberculosis, infectious mononucleosis) and intoxications (e.g. long term valproate therapy).

K. Ewens¹, J. Tisdall^1,2, K. O'Brien¹, S. Gutin¹, A. Bruzel², S. Gregory³, P. Concannon⁴, R.S. Spielman¹

¹Dept Genetics, Univ Pennsylvania School of Medicine, Philadelphia, PA; ²Dept Pediatrics, Univ Pennsylvania School of Medicine, Philadelphia, PA; ³The Sanger Centre, Wellcome Trust Genome Campus, Hinxton, UK; ⁴Molecular Genetics, Virginia Mason Research Center, Seattle, WA and Univ Washington School of Medicine, Seattle, WA

In spite of exhaustive studies of the genetic factors contributing to susceptibility to type 1 (insulin dependent) diabetes mellitus (T1DM or IDDM), the HLA region on 6p and IDDM2 near the insulin gene on 11p remain the only well-established T1DM susceptibility genes. A genome-wide screen of 428 multiplex families (467 ASPs) confirmed these two linkages and provided suggestive evidence for linkage of T1DM to a region of chromosome 1 (1q42-q43) centered on the polymorphic marker D1S1617 (lod=3.31) (Nature Genet, 19:292-296, 1998). In order to identify candidate genes in this region, we generated a high-density physical map of the 7-10 Mb region between D1S439 and D1S459, which flank D1S1617. Initially 104 BAC clones were isolated by screening the RPCI-11 BAC library with 40 STS markers mapping to this region. An additional 309 PAC clones were mapped to this region by the Sanger Centre. These clones were assembled into contiguous sets (contigs) by STS content analysis using a total of 111 STS markers. The order of clones within the contigs was verified by further STS content mapping using end-sequences derived from the BAC and PAC clones and by fingerprint mapping.

This BAC/PAC genomic contig is being used for sequencing and further characterization of the genomic structure of this region. Both molecular and electronic gene-finding techniques are being applied to identify the genes in this region and mutation screens will be carried out in diabetic families on any genes that appear to be likely candidates for T1DM.

H.M. Hoffman¹, F.A. Wright⁴, D.H. Broide¹, A.A. Wanderer³, R.D. Kolodner^1,2

¹ Department of Medicine, University of California, San Diego, La Jolla, CA; ²Ludwig Institute of Cancer Research and Cancer Center, University of California, San Diego, La Jolla, CA; ³Division of Human Cancer Genetics and Cancer Center, Ohio State University, Columbus, OH; ⁴Department of Pediatrics and Allergy, University of Colorado Health Sciences Center, Denver, CO

Familial cold urticaria (FCU - OMIM 120100) is a rare autosomal dominant inherited inflammatory disease characterized by intermittent episodes of rash, fever, arthralgia, and conjunctivitis following generalized exposure to cold environments. We have established linkage for FCU to a 10 cM region on chromosome 1q44 between markers AFMB358WG1 and D1S2682. In order to further characterize this region, several databases were searched to locate additional STR polymorphic markers for genetic mapping. Exact location and order of these markers were confirmed using the GB4 and TNG radiation hybrid sets. PCR conditions were optimized for these markers and genotyping was performed on members of 5 FCU families. An incomplete physical map of the region was constructed with BACs and PACs by chromosome walking utilizing oligonucleotide hybridization and end sequencing of clones. Three contigs of varying sizes were formed in the region and these clones were submitted to the Sanger Centre for fingerprinting and comparison with existing contigs. Sequencing the clones in these contigs will provide candidate genes for analysis.