S. Banfi(1), G. Borsani(1), A. Bulfone(1), L.
Bernard(1), F. Rubboli(1), A. Marchitiello(1), M.
Zollo(1), O. Zuffardi(2,3), A. Ballabio(1,4).
1Telethon Institute of Genetics and Medicine (TIGEM) (2)Servizio di Citogenetica, San Raffaele Biomedical Science Park, Milan; (3)Cattedra di Biologia Generale, University of Pavia, Pavia; (4)Department of Molecular Biology, University of Siena, Siena, Italy
In spite of the considerable evolutionary distance, the mechanisms of eye development and vision present many common features in vertebrates and invertebrates. A large number of human and mouse genes concerned with the function or development of the eye have been shown to be highly conserved throughout evolution and to have closely similar orthologs in Drosophila melanogaster. In the course of a project aimed at the recognition of human cDNAs homologous to Drosophila mutant genes through EST database (dbEST) searching, we identified twelve human cDNAs homologous to Drosophila eye mutant genes. The degree of homology shared by these DRES (Drosophila-related expressed sequences) genes and the corresponding Drosophila eye mutant genes is very significant with a Pvalue at the protein level <10-13 in all cases.
The isolation and sequencing of the full-length human and murine transcript of these DRES is currently in progress. Furthermore, the expression pattern of DRES genes is being studied by RNA in situ hybridization in mouse embryo standard sections. The analysis of the data will provide useful information on the putative function of these genes in vertebrates and their possible involvement in human eye inherited disorders. Moreover, the correlation with the expression pattern of the corresponding Drosophila genes will be helpful to assess a conserved function of these genes during evolution.
Sarah Baxendale, Lynne Hampson, Ian N. Hampson and T.
Paterson Institute of Cancer Research, Department of Experimental Haematology, Christie Hospital, Manchester, United Kingdom
Haemopoietic stem cells have the ability to either self-renew or to differentiate into all the mature blood cell types. The mechanisms underlying these processes are, as yet, unclear but they are ultimately driven by changes in gene expression. Using the chemical cross linking subtraction technique CCLS (see also abstracts by J.M. Walter et al and improvements to CCLS by Hampson et al) we have now isolated a cohort of genes which down-regulate during myeloid differentiation of the well characterized, non-leukaemic murine multipotent stem cell line FDCP-Mix. Sequence analysis of these clones continues to identify both known and novel cDNAs. Future functional studies will allow determination of the precise role which the protein products of these genes have in haemopoietic stem cell self renewal and differentiation and should further our understanding of normal haemopoiesis and as a corollary, leukaemogenesis.
Brooksbank, R.A., Coffey, A.J., Cahn, A.P., Rhodes, S.,
Vaudin, M., Howell, G.R, Micklem, G., King, A., Durham,
J., Bye, J.M., Dunham, A., Ross, M.T. and Bentley, D.R.
The Sanger Centre, Wellcome Trust Genome Campus, Cambridge, United Kingdom
Now that the Human Genome Project has entered the large-scale sequencing stage, the identification of all the genes is a major goal. We are studying an approximately 2.5 Mb region of Xq25 delineated by a deletion present in a male patient suffering from X-linked lymphoproliferative syndrome (XLP), the gene responsible for which has not yet been identified. This region has been extensively mapped and a bacterial clone contig is being assembled across it. A minimal overlapping set of clones from this contig are being sequenced. The absence of any other symptoms in the patient implies that the 2.5 Mb region contains at most only a few genes, making it a potentially challenging region for transcript mapping. We are using exon-trapping, cDNA selection and direct screening of cDNA libraries with genomic clones to identify putative genes within the region. We have used an acedb database to collate the results of these experiments along with exon predictions and homologies to known sequences. The genomic sequence is then used to design PCR primers to screen cDNA libraries; positive clones are themselves sequenced and incorporated into the database. The abundance of repetitive elements in the region is a recurrent obstacle to all the methods that we are using. We can use the sequence data to visualize the distribution of repeats and to interpret our results.
Martin M. Corcoran, Roman Mullenbach, Rachel E. Ibbotson, Anne Gardiner and David G. Oscier Molecular Biology Laboratory, Department of Pathology, Royal Bournemouth General Hospital, Bournemouth, United Kingdom
A region of human chromosome 13Q14.3 commonly deleted in patients with Chronic Lymphocytic Leukemia has been subcloned into a cosmid and pac contig. The area covered, over 600 kb in total, including the markers D13s25 and D13s319 has been physically mapped and the positions of several CpG islands determined, possibly associated with a putative tumor suppressor gene at this locus. The basis for a transcriptional map of the region has been constructed by screening multiple tissue northern sources with cosmid inserts and a number of expressed transcripts have been identified. The cosmid and pac contig has been used as a source of microsatellite markers in order to identify informative polymorphisms and as a source of FISH probes leading to a narrowing down of the minimally deleted region in patient samples as an important step towards the identification and isolation of the candidate tumor suppressor gene or genes.
Roger D. Cox
Wellcome Trust Centre for Human Genetics, Oxfordshire, United Kingdom
My group is directly involved in gene identification on a number of polygenic disease projects within the Wellcome Trust Centre for Human Genetics, Oxford. In addition part of my group is working, within an international collaboration, on the cloning of maturity onset diabetes of the young 3 (MODY3) gene- a dominant disease gene.
We are using exon trapping, cDNA selection, direct screening of cDNA libraries with genomic DNA and more recently direct sequencing.
Our experience of these approaches and of characterizing transcripts identified will be discussed.
Treasa A. Creavin(1), Jaimie A, Greenham(1), Norman A.
Doggett(2), Sara E. Mole(1)
(1)Department of Paediatrics, The Rayne Institute, University College Medical School, London, United Kingdom (2)Life Science Division and Centre for Human Genome Studies, Los Alamos National Laboratory, Los Alamos, New Mexico, USA.
As part of a European consortium contributing to the physical and transcriptional mapping of the short arm of chromosome 16, we are constructing a transcriptional map of 16p12.3-12.2. At present this region consists of 68 cosmids which map to 5 hybrid intervals. Initially our strategy involved isolating unique restriction fragments from cosmids that do not hybridize to human placental DNA.
Each cosmid was analysed for the presence of a CpG island by digestion with NotI, AscI, EagI and BssHII. We have used these results to select cosmids which we have used in Direct cDNA Selection experiments. We are currently analysing over a thousand selected cDNA clones by sequencing and hybridization with chromosome 16 clones. To date seven cDNA clones have been shown to have high homology to ESTs in the database, one cDNA clone has high homology with a gene not previously mapped to chromosome 16 and another cDNA had no homology with any genes or ESTs currently deposited in the databases.
Laura de Gregorio, Manuela Gariboldi, Giacomo Manenti,
F. Stefania Falvella, Manuela Rota, Marco A. Pierotti,
and Tommaso A. Dragani
Instituto Nazionale Tumori, Milan, Italy
Susceptibility to hepatocarcinogenesis in mice represents a genetic model of polygenic inheritance of predisposition to cancer, because six Hepatocarcinogen susceptibility (Hcs) and two resistance (Hcr) loci have been mapped. With the aim of testing possible strain-specific pattern of gene expression that could be associated with genetic predisposition to liver cancer, we have analysed by mRNA differential display the normal liver tissue of different mouse strains. Two selenium binding liver protein genes (Lpsb1 and Lpsb2) were expressed in the normal liver of adult C57BL/6J and BALB/c mice (genetically resistant to hepatocarcinogenesis) at >10 fold higher levels than in the normal liver of C3H/He and CBA/J strains (susceptible to hepatocarcinogenesis). In F1 hybrid mice obtained from C3H/He strain crossed either with C57BL/6J or BALB/c mice, expression levels of Lpsb genes correlated with recessivity and dominance of resistance to hepatocarcinogenesis. The Lpsb genes are closely linked and map on Chromosome 3, in a region where no Hcs or Hcr loci have been localized.
University College, London Medical School, The Rayne Institute London, United Kingdom
As part of a European Consortium contributing to a physical and transcriptional map of the short arm of chromosome 16, we are constructing a transcriptional map of human chromosome 16p12.1. This region at present consists of 136 cosmids, representing about 50-60% coverage of the region. Our initial strategy involved identifying cosmids that contain rare-cutter CpG-rich restriction sites and isolating unique restriction fragments that do not hybridize to human placental DNA. These potential coding fragments have been used to screen multiple tissue northern blots and cDNA libraries to identify transcribed sequences. More recently direct cDNA selection and exon trapping have been carried out on a subset of these cosmids. Using a combination of these approaches we have mapped two known genes (PRKCB1 and PRKCB2), two novel cDNAs and three novel exons to the region.
In addition we have integrated other known genes (SCNN1B and SCNN1G) and IMAGE cDNAs that had previously only been localized to somatic cell hybrid intervals spanning the region. We have also identified a Fugu rubripes cosmid that we believe to contain PRKCB1, we aim to confirm this by shotgun sequencing and in the process identify other genes that may map within close proximity.
Peter C. Groot and Bernard A. van Oost
Department of Clinical Sciences of Companion Animals, University of Utrecht, Utrecht, The Netherlands
We have used Representational Difference Analysis on cDNA's (cDNA-RDA, Hubank and Schatz, Nucl. Acids Res. 22, 5640-8, 1994), as a means to directly isolate expressed sequences derived from human Xq28. To this end, the hamster cell-line Y21 was used as Driver and the Y21-derived hamster-human hybrid Q1Z cell-line (Warren et al., PNAS 87, 3856-60) as Tester. From both cell-lines cDNA was made, digested with the four-cutter DpnII (GATC) to obtain maximum representation, ligated to linkers and amplified by PCR. Two rounds of subtraction were used with Driver:Tester ratio's of 1:100 and 1:800, in the first and second round, respectively. To monitor the subtraction procedure, we used hybridization with a fragment of the human Xq28 derived QM gene, which was known to be highly expressed in Q1Z, as probe. After one round of subtraction we observed a smear of DNA after EtBr staining, with a number of discrete bands already clearly visible, indicating that the complexity of the DNA was already substantially reduced. After two rounds of subtraction only two major bands, with a very faint background, were seen after EtBr-staining, one of which was derived from the QM-gene. This result shows that the subtraction procedure in our hands has been surprisingly effective. To investigate the nature of the second major band and to establish if more human-derived cDNA's can be isolated using this procedure, we are currently characterizing the first and second round subtraction products by subcloning and sequencing.
Christian Gruber(1), Wu-Bo Li(1), Joel A. Jessee(1), Kevin G. Becker(2), Wayne D.L. Bentham(2), James W. Nagle(2), Ameer Gado(2), and William E. Biddison(2) (1)Life Technologies, Inc., Gaithersburg, MD, USA; (2)Neuroimmunology Branch and Laboratory of Neurogenetics, National Institute of neurological Diseases and Stroke, National Institutes of Health, Bethesda, MD, USA
A human brain cDNA expression library was greatly enriched for specific gene families with the GeneTrapper cDNA positive selection system coupled to motif specific degenerate oligonucleotides. Motif capture uses degenerate oligo-nucleotides coding for short highly conserved regions of gene families. Enriched sublibraries were isolated for C2H2-type zinc fingers, ribonucleoproteins (RNP1) SH2 domains, and CAG triplet repeats. These clones were filter arrayed creating family specific sets. Subsets of each family were EST sequenced to confirm the efficiency of cDNA family isolation and for identification of novel genes. Family specific filter sets are being analyzed for relative gene expression by tissue specific differential hybridization. These enriched plasmid sublibraries are being used in expression analyses linking large scale cDNA isolation with biological function. Similar family specific isolations are being performed in the yeast two-hybrid system and other specialty cDNA libraries. The motif capture approach describes a focused, general, rapid, non-random approach to genome transcript analysis. It is non-PCR based, results in large or full length cDNA clones, is easily linked to specific biological assays, and is rapidly adaptable to non-human systems.
K.W. Hasel(1), B. Hilbush(1), D. Seto(1), P. Miller(1),
and J.G. Sutcliffe(2)
(1)Digital Gene Technologies, La Jolla, CA, USA (2)Dept of Molecular Biology, The Scripps Research Institute, La Jolla, CA, USA
The TOGA (Total Gene expression Analysis) method when practiced manually is robust but cumbersome due to the large number of PCR reactions to be performed and analyzed so as to achieve complete coverage of gene space. The method has been optimized for each of the 256 TOGA primers. We have automated the method to increase throughput and quality control, and to facilitate automatic data collection. A series of robot-controlled steps has allowed consistent and rapid generation of TOGA data, as will be illustrated with experimental examples. The unique digital address for each mRNA, a combination of an 8-nucleotide sequence and fragment length, renders the data amenable for comparison and archiving into a relational database for query. This feature is easily merged with public genome databases, allowing the convergence of gene expression data, nucleic acid sequence data, gene mapping data and literature citations.
Simon Horvat, Roslin Institute, Edinburgh, Scotland, and Jaun F. Medrano, Dept. of Animal Science, University of California, Davis, Davis, CA 95616.
The high growth (hg) gene in mice produces a 30-50% increase in weight gain of homozygous individuals. Previous interval mapping and test crosses analyses positioned hg to ~ 5 cM interval around marker D10Mit41 on mouse chromosome 10 and demonstrated that hg is not an allele of Insulin-like growth factor I (Igf-1) or Decorin (Dcn), two closely linked candidates.
Recently, a microsatellite marker D10Mit69 was found to be deleted in high growth mice. This marker cosegregated with hg in a cross of congenic strains C57BL/6J-hghg X C57BL/6J (1051 F2 mice) indicating that a deletion of a region around D10Mit69 might be responsible for the high growth effect. Marker D10Mit69 has been utilized as an entry point to physical cloning of the hg-containing segment using Yeast Artificial Chromosome (YAC) and Bacterial Artificial Chromosome (BAC) clones. Exon trapping of the BACs is being used to uncover expressed sequences. The current status of candidate transcription units isolated from the hg region will be presented.
Heather A. McQueen, Giorgia Siriaco, Sally H. Cross,
and Adrian P. Bird
University of Edinburgh, Institute of Cell and Molecular Biology, Edinburgh, Scotland
We are engaged in the study of CpG islands and their organization in the chicken genome. CpG islands are short unmethylated CpG-rich sequences situated to the 5' end of many vertebrate genes and are particularly GC-rich in the chicken. The chicken karyotype comprises 39 chromosome pairs of which at least 29 are microchromosomes. Microchromosomes account for only about 25% of the genomic DNA and are cytologically indistinguishable. Due to technical limitation microchromosomal genes are poorly represented in current genome maps. However, we have used a CpG island library, prepared from chicken, to demonstrate that CpG islands are concentrated on the microchromosomes. Our results imply a higher gene density for chicken microchromosomes than is seen in mammalian species. We are currently testing this hypothesis using cloned segments of microchromosomal chicken DNA.
INSERM, Marseille, France
The Sxrb deletion interval of the mouse Y chromosome short arm is critical for normal spermatogenesis and expression of the male specific minor transplantation antigen H-Y. Mice deleted for this region manifest a block in spermatogenesis 3 days after birth, during the mitotic proliferation of the germ cells. We have established contigs of cosmids and yeast and bacterial artificial chromosomes covering over 1 Mb in this region. With the aim of defining a comprehensive transcriptional map of this interval, 400 kb of these contigs have so far been subjected to exon trapping and a preliminary exon map established. Exon-trapping has been carried out using the standard pSPL3 protocol on individual cosmids. Potential exons have been designated from among 30 distinct exon-trapped products on the basis of the following criteria: 1) the sequence shows no homology to a known repeat element, 2) the sequence does not hit several unrelated mouse sequences 3) one of the forward reading frames is open 4) the sequence is either homologous to the coding region of a known gene, or shows no identity to rodent sequences, 5) PCR from genomic DNA using primers specific for the product generates a male-specific fragment mapping to the Sxr region. In this way nine potential exons have been positioned on the map. The corresponding cDNAs have not yet been isolated for these fragments but reverse-transcriptase dependent products have been amplified from at least testis RNA for eight of these fragments. These preliminary results indicate that, as suggested by H-Y typing, the Sxrb deletion interval may be relatively complex at the genetic level.
S. Pickering, J. Morten, M. Finnegan, and R. Anand
Zeneca Pharmaceuticals, Cheshire, United Kingdom
We have been evaluating two methods which use chromosomal microdissection for the isolation of transcripts from a region of interest. The first uses DOP amplified microdissected material in direct selection (SHAC, Chen-Lui et al, Genomics 30, 388-392, 1996). The second relies on hybridizing adapter-tagged cDNA to the chromosomes followed by microdissection and recovery of the cDNA by PCR amplification (PrepISH, Hozier et al, Genomics 19, 441-447, 1994, Gracia et al, Hum. Mol. Genet. 5, 595-600, 1996). Both techniques provide a route for the isolation of transcripts from a 5-10 mb region without the need for physical cloning. Most of our work has been on Prep-ISH. To assess the utility of this technique, we have shown that representation is maintained following PCR from high and low copy template. We are currently assessing the degree of enrichment for transcripts from the microdissected region following different hybridization conditions.
Molecular Medicine, clinical Genetics Unit, Karolinska Hospital, Stockholm, Sweden
Meningiomas are brain tumors arising from the meninges covering the central nervous system. Extensive deletion mapping studies showed that monosomy or loss of large portions of human chromosome 22 are primary events in meningioma tumorigenesis. However, no restricted region of this chromosome arm is commonly deleted; leading to the conclusion that more than one gene on 22q is involved in tumor development.We constructed a cosmid contig covering the largest candidate region (estimated to 1Mbp) in 22q12. At present, as much as 50-60% of this contig is sequenced. In order to clone the transcribed genes contained in this available genomic sequence, we are combining sequence-based exon prediction (using the Grail computer program) and testing of human cDNA libraries. An update in the isolation of transcribed sequences from this region will be presented.
Susan Rhodes(1), Alan Ashworth(2), Laurent Baron(1),
Andrew Futreal(3), Pieter de Jong(4), Simon Gregory(1),
Gos Micklem(1), Catherine Rice(1), Michael Stratton(2),
Richard Wooster(1), and David Bentley(1)
(1)The Sanger Centre, Wellcome Trust Genome Campus, Hinxton, United Kingdom (2)Sections of Molecular Carcinogenesis and CRC Centre for Cell and Molecular Biology, Institute of Cancer Research, Haddow Laboratories, and Chester Beatty Laboratories, United Kingdom (3)Duke University Medical Centre, Durham, North Carolina, USA (4)Human Genetics Department, Rosewell Park Cancer Institute, Buffalo, New York, USA
The availability of genomic sequence provides the means to identify and characterize genes by a combination of computational and experimental techniques. A physical map of a 900 kb region of human chromosome 13q has been constructed using overlapping P1- artificial chromosome clones (PACs). Potential coding sequences were identified from a minimum overlapping set of clones by exon trapping, cDNA selection and direct screening of cDNA libraries. These sequences were added to an analysis from public databases for full-length cDNAs, protein matches and expressed sequence tags (ESTs) across the genomic sequence. All potentially coding sequences were aligned and displayed in ACEDB 4.3.
Few thousand expressed sequence tags have been subjected to a computer program to predict if they encode leader sequences for secreted proteins or receptors. A functional test has been devised to test these predictions. In this assay, the predicted signal sequences substitute for the leader sequence of a reporter gene. COS cells are transfected with the resulting hybrid DNA in a plasmid backbone. Functional substitution of the leader sequence of the reporter by the novel sequence is taken as the confirmation of the computer generated prediction.
S. Scherer(1), J. Huizenga(1), J. Herbrick(1), S.
Soder(1), G. Traverso(1), B. Beattie(2), J. Squire(2),
C. Wu(3), P. de Jong(3), Y.J. Kang(4), U. Sohn(4) and
(1)Department of Genetics and (2)Molecular Pathology, The Hospital for Sick Children, Toronto, Ontario, M5G 1X8, Canada; (3)Department of Human Genetics, Roswell Park Cancer Institute, Buffalo, NY, USA (4)Department of Genetic Engineering, Kyungpook National University, Taegu, Korea.
Our center was established as a national core facility to provide a service for Canadian investigators wishing to map genes to human chromosomes. Since the mapping reagents provided to us can be of a diverse origin (eg. human or mouse) and type (cDNA, genomic clone or DNA sequence) we have adopted a comprehensive approach in an effort to map the genes with the highest frequency, accuracy, and resolution. Preferably, both gene/cDNA probes and oligonucleotides designed from the DNA sequence are screened by hybridization and PCR, respectively, against the NIGMS somatic cell hybrid mapping panel (#2), genomic libraries (primarily the CEPH mega-YAC library and the RPCI-I total human PAC library), and radiation hybrid panels (Stanford-G3 and GeneBridge-G4 panels). The combination of these techniques usually allows for a precise mapping of the gene within at least one existing framework map. If it is not possible to determine the cytogenetic position of the gene from these experiments a cDNA clone or corresponding genomic clone is then FISH mapped against metaphase spreads. Using this approach we have successfully mapped over 270 genes across the genome. Since the majority of these genes have been characterized at the functional level the information is immediately useful for positional cloning experiments.
In collaboration with other groups we have also initiated large scale cDNA clone and EST mapping projects. One project involves the mapping of cDNAs and ESTs from a human thymus cDNA sequencing project. Briefly, 1,117 sequences were determined and following vigorous database searches, the DNA sequence of the 3'-UTR of the ESTs not already assigned to a chromosome are being converted into STSs. They are then screened by PCR against a monochromosomal somatic cell hybrid panel and a subset of the CEPH-megaYAC library. To expedite and economize the YAC library screening process we have used the most recent MIT-Whitehead whole genome map and other published maps of single chromosomes (chromosomes 12, 19, 21, and 22) as a guide to select 4000 clones which represent 2-fold coverage tiling paths of each chromosome. The same set of YAC clones is also being used in other projects for rapid screening of retinal and heart specific cDNAs. Progress on our whole-genome mapping project will be reported.
Supported by grants from the Canadian Genome Analysis and Technology Program.
Joanne M. Walter(1), Maxine Belfield(1), Ian N.
Hampson(2), Christopher A. Read(1).
(1)Amersham International plc, Amersham Laboratories, Amersham, United Kingdom (2)Paterson Institute for Cancer Research, Manchester, United Kingdom
Chemical Cross-linking Subtraction (CCLS)(1) is a novel approach to the problem of identifying and isolating differentially expressed genes. Here we describe the method by demonstrating its use in the initial isolation of specifically expressed genes from human foetal brain tissue. 10 micrograms of human heart poly A+ mRNA was hybridized with 400ng foetal brain cDNA in an 80% formamide buffer at 42oC to a Rot of 600. The common species in heart and brain RNA populations hybridized and were inactivated in the subsequent cross-linking step by 2,5 diaziridinyl-1,4-benzoquinone (DZQ). Generation of probes from the single stranded novel species took place using a random primer approach with Sequenase version 2.0 T7 DNA polymerase to prevent labeling of excess RNA and alpha 32P dCTP (Amersham:AA0005). The probe was used to screen approximately 30,000 clones from a directional Lambda MOSSlox foetal brain library and 12 of the potential positives were selected for further analysis by comparison to duplicate screens using unsubtracted heart cDNA. Secondary screening, automatic subcloning and Northern analysis was carried out resulting in 50% of the original primary selected clones showing differential expression. 40% of the clones failed to subclone and 10% were below the detection limits of the Northern analysis used. Having shown the utility of CCLS in this application further work is underway to isolate differentially expressed DNA repair genes after stimulation of a HeLa cell line with UV radiation.
CCLS represents a robust approach to the identification of differentially expressed genes which does not require multiple rounds of subtraction, physical separation of the subtracted cDNA or amplification by PCR. It is likely to find use in a variety of research applications, particularly when a library screening approach is favored over the running of numerous polyacrylamide gels and the system under study enables the generation of adequate levels of RNA.
Reference: HAMPSON, I. N. et al. Nucl. Acids Res., 20, pp 2899,1992.