John E. Collins

The development and application of automated gridding for efficient screening of yeast and bacterial ordered libraries

An automated gridding procedure for the inoculation of yeast and bacterial clones in high-density arrays has been developed. A 96-pin inoculating tool compatible with the standard microtiter plate format and an eight-position tablet have been designed to fit the Biomek 1000 programmable robotic workstation (Beckman Instruments). The system is used to inoculate six copies of 80 x 120-mm filters representing a total of approximately 20,000 individual clones in approximately 3 h. High-density arrays of yeast artificial chromosome (YAC) and cosmid clones have been used for rapid large-scale hybridization screens of ordered libraries. In addition, an improved PCR library screening strategy has been developed using strips cut from the high-density arrays to prepare row and column DNA pools for PCR analysis. This strategy eliminates the final hybridization step and allows identification of a single clone by PCR in 2 days. The development of automated gridding technology will have a significant impact on the establishment of fully versatile screening of ordered library resources for genomic studies.

Characterization of the human synaptogyrin gene family

Genomic sequencing was combined with searches of databases for identification of active genes on human chromosome 22. A cosmid from 22q13, located in the telomeric vicinity of the PDGFB (platelet-derived growth factor B-chain) gene, was fully sequenced. Using an expressed sequence tag-based approach we characterized human (SYNGR1) and mouse (Syngr1) orthologs of the previously cloned rat synaptogyrin gene (RATSYNGR1). The human SYNGR1 gene reveals three (SYNGR1a, SYNGR1b, SYNGR1c) alternative transcript forms of 4.5, 1.3 and 0.9 kb, respectively. The transcription of SYNGR1 starts from two different promoters, and leads to predicted proteins with different N- and C-terminal ends. The most abundant SYNGR1a transcript, the 4.5-kb form, which corresponds to RATSYNGR1, is highly expressed in neurons of the central nervous system and at much lower levels in other tissues, as determined by in situ hybridization histochemistry. The levels of SYNGR1b and SYNGR1c transcripts are low and limited to heart, skeletal muscle, ovary and fetal liver. We also characterized two additional members of this novel synaptogyrin gene family in human (SYNGR2 and SYNGR3), and one in mouse (Syngr2). The human SYNGR2 gene transcript of 1.6 kb is expressed at high levels in all tissues, except brain. The 2.2-kb SYNGR3 transcript was detected in brain and placenta only. The human SYNGR2 and SYNGR3 genes were mapped by fluorescence in situ hybridization to 17qtel and 16ptel, respectively. The human SYNGR2 gene has a processed pseudogene localized in 15q11. All predicted synaptogyrin proteins contain four strongly conserved transmembrane domains, which is consistent with the M-shaped topology. The C-terminal polypeptide ends are variable in length, display a low degree of sequence similarity between family members, and are therefore likely to convey the functional specificity of each protein.

A genome annotation-driven approach to cloning the human ORFeome

We have developed a systematic approach to generating cDNA clones containing full-length open reading frames (ORFs), exploiting knowledge of gene structure from genomic sequence. Each ORF was amplified by PCR from a pool of primary cDNAs, cloned and confirmed by sequencing. We obtained clones representing 70% of genes on human chromosome 22, whereas searching available cDNA clone collections found at best 48% from a single collection and 60% for all collections combined.

Comparative analyses of the Dominant megacolon-SOX10 genomic interval in mouse and human

Comprehensive linkage maps based upon microsatellite markers have been established for the mouse and human genomes (Dib et al. 1996; Dietrich et al. 1996). While these microsatellite linkage maps are useful for mapping within a species, gene-based maps are required for comparisons of genome organization between species (DeBry and Seldin 1996; Serikawa et al. 1998). Framework synteny maps provide localization of specific genes to regions within one genome, but are insufficient for ordering closely linked genes. Often, small inversions or deletions occur during evolution that become apparent only through construction of high-resolution comparative maps across syntenic regions (Botta et al. 1997; Cabin et al. 1998; Watkins-Chow et al. 1997). Comparative mapping between species not only facilitates resolution of syntenic regions, but dense, gene-based maps have proven useful for identifying genes defective in human diseases and mouse mutants that model disease syndromes. Hirschsprung disease (HSCR) is one phenotype in which animal models have been particularly relevant to the identification and analysis of human genes. For example, the elucidation of the genetic defect responsible for the Waardenburg syndrome type 4 (WS4 OMIM #277580) phenotypes of Dominant megacolon (Sox10) mice (Southard-Smith et al. 1998) led directly to identification of mutations in the human SOX10 ortholog of individuals with WS4 (Pingault et al. 1998; Southard-Smith et al., in press). The Sox10 mouse also mimics the variable penetrance and expressivity of the aganglionosis observed in HSCR/WS4 patients, making it a useful model system for analyzing modifier loci (Southard-Smith et al., in press). Initial characterization of theSox10 mutation determined that the alteration had occurred in the C57BL/6J allele of a C57BL/ 6J × C3HeB/FeJLea-a (B6C3H) F 1 background and mapped this locus to mid-distal mouse Chr 15 (Lane and Liu 1984). Subsequent comparative mapping studies established conserved synteny of the Sox10 interval with human Chr 22 (Pingault et al. 1997; Puliti et al. 1995). To generate a detailed map of this region, we initially performed linkage disequilibrium mapping with polymorphic microsatellite markers in pedigrees of mice generated during the construction ofSox10/+ congenic lines on C57BL/6J (B6) and C3HeB/FeJLea-a (C3H) backgrounds. (Southard-Smith et al., in press). Haplotypes were constructed with markers D15Mit122, D15Mit158, D15Mit28, D15Mit68, D15Mit2, D15Mit31, D15Mit189,andD15Mit107 (Fig. 1A). Evaluation of haplotypes from construction of our Sox10 Dom congenic line on the C3H background revealed the B6 alleles of markers proximal to D15Mit2and−31 were replaced by C3H alleles. In contrast, in the B6.Sox10 lines, the C3H allele of markers distal to D15Mit68 were replaced by B6 alleles at those loci. By following the inheritance of theSox10 locus in association with C57BL/6J alleles, we were quickly able to narrow the Sox10 critical interval to less than 2 cM between D15Mit68 and D15Mit2. A similar approach may be useful for other mutations that have not yet been cloned and do not yet have a high-resolution mapping cross generated. To further define the conserved linkage between mouse and human, we used 1716 meioses generated from crosses of Sox10 mice with C3H,M. castaneus/Ei andM. molossinus/Ei (Fig. 1B). Our mapping efforts were focused on reducing the 2-cM region located between markers D15Mit68 andD15Mit2 that we identified from the linkage disequilibrium mapping studies in our congenic pedigrees. Animals with demonstrated recombination events between these two markers (31/1716) were genotyped with additional markers, genes, and ESTs (Fig. 1B). As a consequence of genotyping recombinant animals, the genetic interval for the Sox10locus was reduced to a 0.6-cM region flanked by Smstr3and Pdgfb. Because bothSmstr3and Pdgfb map to human Chr 22, genes internal to these loci might also demonstrate conserved linkage and order between mouse and human. Crosses were assessed for differences in recombination frequency between males and females for this region of Chr 15. No significant differences in recombination were observed between crosses 1 and 2 ( p-value 0.30). No inconsistencies of the marker order were found by comparison with EUCIB crosses (http:// www.hgmp.mrc.ac.uk/MBx/MBxHomepage.html, December 1998). However, examination of the MGD database (www.informatics.jax.org, December 1998) revealed minor inconsistencies that could have arisen from comparison of data between different crosses. In comparison with the Chromosome 15 Committee report, our genetic mapping data place Smstr3proximal toD15Mit1, * Present address: University of Minnesota Medical School, St. Paul, Minnesota, USA.

A panel of human chromosome 22-specific sequence tagged sites

A panel of 29 sequence tagged sites (STSs) covering the long arm of chromosome 22 has been assembled. STS primer pairs were synthesized using available chromosome 22 sequence derived from the GenBank and EMBL DNA sequence databases, as well as published cDNA and genomic sequence, or from previously published and communicated primer pairs. Each STS was optimized for the polymerase chain reaction using a chromosome 22-only hybrid and human genomic DNA. Further STS content analysis on a panel of somatic cell hybrids that incorporated two chromosome 22 translocations resulted in the mapping of the X-box binding protein (XBP), D22S156, and transcobalamin II (TCN2) genes to 22q11-q13.1. The panel of STSs was used for the rapid determination of the STS content and thus the chromosomal DNA content of a new irradiation hybrid.

Genetic mapping of 14 short tandem repeat polymorphisms on human chromosome 22

We have constructed a linkage map of 14 short tandem repeat polymorphisms (11 with heterozygosity > 70%) on the long arm of human chromosome 22 using 23 non-CEPH pedigrees. Twelve of the markers could be positioned uniquely with a likelihood of at least 1,000:1, and distributed at an average distance of 6.62 cM (range 1.5–16.1 cM). The sex-combined map covers a total of 79.6 cM, the female map 93.2 cM and the male map 64.6 cM. Based on comparisons between physical maps and other genetic maps, we estimate that our map covers 70%–80% of the chromosome. The map integrates markers from previous genetic maps and uniquely positions one marker (D22S307). Data from physical mapping on the location of four genetic markers correlates well with our linkage map, and provides information on an additional marker (D22S315). This map will facilitate high resolution mapping of additional polymorphic loci and disease genes on chromosome 22, and act as a reference for building and verifying physical maps.

Facilitating Translational Research

Would a carpenter be asked to manage building a new housing development? Probably not. More likely, a real estate developer with the skills, experience and knowledge of the local market and trades, including carpentry, would manage such an undertaking. Good developers anticipate and address the challenges of building and selling homes. Delivering attractive, appropriately priced homes on time and on budget requires that developers use their knowledge, experience, and judgment to make numerous decisions that engage the right talent at the right time to balance development risks and costs.