J. Ross Miller

RE-SEQUENCING OF DNA FROM A DIVERSE PANEL OF CATTLE REVEALS A HIGH LEVEL OF POLYMORPHISM IN BOTH INTRON AND EXON

Abstract. In order to assess the extent of DNA sequence variation in cattle, introns and exons from both the leptin and Amyloid Precursor Protein (APP) genes have been sequenced in a panel of DNAs derived from 22 diverse animals. Direct DNA sequencing of PCR products was used; thus, 44 chromosomes were studied. Polymorphisms were identified by manual scanning of sequence chromatograms and computerized sequence analysis. Twenty Single Nucleotide Polymorphisms (SNPs) were detected in 1788 bp sequenced from the leptin gene, giving a frequency of 1 SNP per 89 bp. Twenty-four SNPs were detected in a 458-bp fragment of the APP gene; 23 of the polymorphisms were contained in a 302-bp intron 16 fragment. This equates to an SNP frequency of 1 per 13 bp for the intron. We can thus conclude that this portion of the bovine APP gene constitutes a hypermutable region. Nucleotide sequence diversity values of 0.019 and 0.0026 were obtained for APP and leptin respectively.

A 5x genome coverage bovine BAC library: Production, characterization, and distribution

Abstract. A bovine large-insert DNA library has been constructed in a Bacterial Artificial Chromosome (BAC) vector. The source DNA was derived from lymphocytes of a Jersey male. High-molecular-weight DNA fragments were produced by treatment with EcoRI/EcoRI methylase and cloned into the EcoRI site of pBACe3.6. In total, 157,240 individual BACs have been picked into 384-well plates. Approximately 190 randomly chosen clones have been characterized by Pulsed Field Gel Electrophoresis (PFGE) and have an average insert size of 105 kb, suggesting library coverage representing 5–6 genome equivalents. The frequency of clones without inserts is 4%. The chromosomal location of 51 BACs was studied by FISH; 3 showed more than one signal, indicating a chimerism frequency of roughly 6%. Approximately 50% of the clones in the library contain Simple Repeat Sequences (microsatellites), and 4% of the clones contain centromeric repeats. Insert stability was assessed by restriction digestion of DNA prepared from 20 clones after serial culture for one and three nights. Only one clone showed any evidence of an altered restriction pattern. Clones from 360 × 384-well plates (138,240 colonies) were gridded onto high-density membranes, and PCR superpools were produced from the same set of clones. Both membranes and superpools are available from the RZPD, Berlin (http://www.rzpd.de). PCR 4-D superpools have been prepared from an additional 23,000 clones. The library has been screened for a total of 24 single-copy sequences; positive clones have been obtained in all cases.

The SNF2-family member Fun30 promotes gene silencing in heterochromatic loci.

Chromatin regulates many key processes in the nucleus by controlling access to the underlying DNA. SNF2-like factors are ATP-driven enzymes that play key roles in the dynamics of chromatin by remodelling nucleosomes and other nucleoprotein complexes. Even simple eukaryotes such as yeast contain members of several subfamilies of SNF2-like factors. The FUN30/ETL1 subfamily of SNF2 remodellers is conserved from yeasts to humans, but is poorly characterized. We show that the deletion of FUN30 leads to sensitivity to the topoisomerase I poison camptothecin and to severe cell cycle progression defects when the Orc5 subunit is mutated. We demonstrate a role of FUN30 in promoting silencing in the heterochromatin-like mating type locus HMR, telomeres and the rDNA repeats. Chromatin immunoprecipitation experiments demonstrate that Fun30 binds at the boundary element of the silent HMR and within the silent HMR. Mapping of nucleosomes in vivo using micrococcal nuclease demonstrates that deletion of FUN30 leads to changes of the chromatin structure at the boundary element. A point mutation in the ATP-binding site abrogates the silencing function of Fun30 as well as its toxicity upon overexpression, indicating that the ATPase activity is essential for these roles of Fun30. We identify by amino acid sequence analysis a putative CUE motif as a feature of FUN30/ETL1 factors and show that this motif assists Fun30 activity. Our work suggests that Fun30 is directly involved in silencing by regulating the chromatin structure within or around silent loci.

Using yeast to place human genes in functional categories

The availability of the draft sequence of the human genome has created a pressing need to assign functions to each of the 35,000 or so genes that it defines. One useful approach for this purpose is to use model organisms for both bioinformatic and functional comparisons. We have developed a complementation system, based on the model eukaryote Saccharomyces cerevisiae, to clone human cDNAs that can functionally complement yeast essential genes. The system employs two regulatable promoters. One promoter, tetO (determining doxycycline-repressible expression), is used to control essential S. cerevisiae genes. The other, pMET3 (which is switched off in the presence of methionine), is employed to regulate the expression of mammalian cDNAs in yeast. We have demonstrated that this system is effective for both individual cDNA clones and for cDNA libraries, permitting the direct selection of functionally complementing clones. Three human cDNA libraries have been constructed and screened for clones that can complement specific essential yeast genes whose expression is switched off by the addition of doxycycline to the culture medium. The validity of each complementation was checked by showing that the yeast cells stop their growth in the presence of doxycycline and methionine, which represses the expression of the yeast and mammalian coding sequence, respectively. Using this system, we have screened 25 tetO replacement strains and succeeded in isolating human cDNAs complementing six essential yeast genes. In this way, we have uncovered a novel human ubiquitin-conjugating enzyme, have isolated a human cDNA clone that may function as a signal peptidase and have demonstrated that the functional segment of the human Psmd12 proteosome sub-unit contains a PINT domain.

A Natural Hypomorphic Variant of the Apoptosis Regulator Gimap4/IAN1

The Gimap/IAN family of GTPases has been implicated in the regulation of cell survival, particularly in lymphomyeloid cells. Prosurvival and prodeath properties have been described for different family members. We generated novel serological reagents to study the expression in rats of the prodeath family member Gimap4 (IAN1), which is sharply up-regulated at or soon after the stage of T cell-positive selection in the thymus. During these investigations we were surprised to discover a severe deficiency of Gimap4 expression in the inbred Brown Norway (BN) rat. Genetic analysis linked this trait to the Gimap gene cluster on rat chromosome 4, the probable cause being an AT dinucleotide insertion in the BN Gimap4 allele (AT(+)). This allele encodes a truncated form of Gimap4 that is missing 21 carboxyl-terminal residues relative to wild type. The low protein expression associated with this allele appears to have a posttranscriptional cause, because mRNA expression was apparently normal. Spontaneous and induced apoptosis of BN and wild-type T cells was analyzed in vitro and compared with the recently described mouse Gimap4 knockout. This revealed a “delayed” apoptosis phenotype similar to but less marked than that of the knockout. The Gimap4 AT(+) allele found in BN was shown to be rare in inbred rat strains. Nevertheless, when wild rat DNA samples were studied the AT(+) allele was found at a high overall frequency (∼30%). This suggests an adaptive significance for this hypomorphic allele.

SWI/SNF-like chromatin remodeling factor Fun30 supports point centromere function in S. cerevisiae.

Budding yeast centromeres are sequence-defined point centromeres and are, unlike in many other organisms, not embedded in heterochromatin. Here we show that Fun30, a poorly understood SWI/SNF-like chromatin remodeling factor conserved in humans, promotes point centromere function through the formation of correct chromatin architecture at centromeres. Our determination of the genome-wide binding and nucleosome positioning properties of Fun30 shows that this enzyme is consistently enriched over centromeres and that a majority of CENs show Fun30-dependent changes in flanking nucleosome position and/or CEN core micrococcal nuclease accessibility. Fun30 deletion leads to defects in histone variant Htz1 occupancy genome-wide, including at and around most centromeres. FUN30 genetically interacts with CSE4, coding for the centromere-specific variant of histone H3, and counteracts the detrimental effect of transcription through centromeres on chromosome segregation and suppresses transcriptional noise over centromere CEN3. Previous work has shown a requirement for fission yeast and mammalian homologs of Fun30 in heterochromatin assembly. As centromeres in budding yeast are not embedded in heterochromatin, our findings indicate a direct role of Fun30 in centromere chromatin by promoting correct chromatin architecture.

The relative merits of the tetO2 and tetO7 promoter systems for the functional analysis of heterologous genes in yeast and a compilation of essential yeast genes with tetO2 promoter substitutions

We have generated a collection of yeast strains, each of which has an essential yeast gene under the control of the tetracycline‐responsive, tetO, promoter. Screens using first‐generation promoter‐swap strains uncovered the non‐specific responsiveness of the tetO7 promoter to a known human transcription factor (hIRF‐1). Non‐specific regulation was not observed with the tetO2 promoter. Reporter assays have been used to demonstrate this phenomenon. Subsequent efforts to generate a collection of tetracycline‐regulatable strains have focused on the tetO2 promoter. These strains are available to the yeast community and can be used for functional genomics studies. Copyright

Molecular characterization of the novel rat NK receptor 1C7

A novel receptor, named 1C7 or NKp30 and involved in natural cytotoxicity, was recently identified. This receptor is encoded by the 1C7 gene, which is located within the class III region of the human MHC, HLA. It is a member of the immunoglobulin gene superfamily (IgSF) and, remarkably, is expressed at the mRNA level as six different splice variants in human. Recent investigations have indicated that the 1c7 gene of the mouse is silenced by in‐frame stop codons. In this study, the molecular characterization of the rat 1c7 gene is described. cDNA derived from this gene encode a protein of 192 amino acid residues predicted to contain a single IgV‐set domain in the extracellular region and a positively charged residue in the transmembrane region. Expression of the gene was detected in freshly isolated rat Natural Killer (NK) and T splenocytes. Transfection of rat 1C7 into the NK cell line RNK‐16 induced cytolytic activity against glioma as well as lymphoma tumor cells. In addition, binding of a r1C7‐Fc fusion protein by a panel of target cells correlated with susceptibility to killing by RNK‐16–1C7 effector cells. These results indicate that the r1C7 molecule could function as an NK activating receptor as previously reported for the human NKp30 receptor molecule.

A bovine YAC library containing four- to five-fold genome equivalents.

Successful positional-cloning strategies in human and mouse have relied upon the wide availability of numerous large-fragment libraries (Collins 1995). Although genomic analysis of livestock species is not as advanced as that in human and mouse, it is progressing rapidly. In cattle there are now over 746 loci assigned to the genetic linkage map, with an average marker interval of 5.3cM (Barendse et al. 1997; Kappes et al. 1997; Ma et al. 1996). The current bovine linkage map is thus of sufficient density to begin the identification of economically important trait loci. Once a region of interest has been defined on a linkage map, this information can be used to construct a long-range physical map. In the case of economically important trait loci (ETL), this is potentially a region covering more than 1 cM, which is most easily covered by a YAC contig where inserts are 500–1000 kb compared with BACs where inserts are typically less than 300 kb. Identification of overlapping large fragment clones allows the construction of a contig across the region from which new markers can be recovered and, potentially, identification of the gene(s) accomplished. Although some large-fragment libraries have been made for livestock, including cattle (Libert et al. 1993; Smith et al. 1996; Takeda et al. 1998), sheep (Broom and Hill 1994), pigs (Alexander et al. 1997; Rogel-Gaillard et al. 1998) and chicken (Toye et al. 1998), for successful contig assembly and gene cloning in these species, a greater depth of genome coverage is necessary to obtain clones from given regions than is represented in an individual library. Construction of high-quality, large-fragment libraries is laborious and technically complex. Yeast artificial chromosomes (YACs) and P1/bacterial artificial chromosomes (PACs/BACs) have been developed to enable cloning of large genomic fragments. The YAC system can accommodate the largest range of insert sizes (100 to 2000 kb), while bacteria-based systems offer a lower range of insert sizes (50 to <300 kb). Employing the two library types together has proven a valuable approach enabling sequences that are not represented by one vector type to be recovered in the other. Here, we describe the production of a bovine genomic YAC library with a fourto five-fold genome coverage and an average insert size of 500 kb. This library was constructed with the cloning vector pYAC4, allowing interruption of thesup-4o gene such that canavanine resistance is conferred to recombinants in ade-101deficient yeast strains (Burke et al. 1987). High quality genomic DNA was purified from peripheral blood mononuclear cells (PBM) of a Holstein-Friesian bull (“10814”) and has been characterized at the MHC locus in our laboratory. Cells were embedded in agarose digesting in Proteinase K (1 mg/ ml) in 1% sodium dodecyl sulfate, 0.5 M EDTA pH 8.0 at 50°C for 24 h, followed by exhaustive exchanging against TE buffer. Embedded DNA was partially digested with EcoRI:EcoRI methylase (typically 1:200 u) as described by Larin and coworkers (1991), then size selected by electrophoresis in a 1% Sea Plaque agarose gel with a Bio-Rad CHEF pulsed field gel apparatus (set at 30 s switch time, 4.7 v/cm at 14°C in 0.5 × TBE for 18 h). The agarose slice containing the unresolved compression zone (>400 kb in these conditions) was recovered and exhaustively exchanged against TE to remove inhibitory borate ions. Ligation to pYAC4 vector arms, prepared by EcoRI andBamHI digestion followed by dephosphorylation, was performed as described by Larin and associates (1991). High molecular weight ligation products were purified by PFGE under the conditions described above, and the compression zone was recovered and treated as before. The ligated DNA was equilibrated in 1M Sorbitol, 10 mM Tris-Cl pH 7.5, 10 mM EDTA prior to transformation. The agarose was cut into small cubes, melted at 70°C for 10 min, then digested with AgarACE (Promega) at 42°C for 1 h. Spheroplasts were prepared from the yeast strains AB1380 or J570, and transformations were carried out as described by Albertsen and colleagues (1990), employing Zymolyase 20-T (ICN). Initial selection allowed for Ura + transformants, which were picked after 5 days’ growth at 30°C, although in later experiments canavanine was also included at 60 mg/ml. These were transferred to AHC (Ura − Trp) plates to select for Ura Trp recombinants (both YAC arms) and insertcontaining red clones (interrupted sup-4otRNA gene) were transferred to 96-well plates containing SD/Uracontaining Hogness modified freezing medium (Werner et al. 1997) supplemented with antibiotics (ampicillin 50mg/ml, kanamycin 30mg/ml, and tetracycline 5mg/ml). After growing at 30°C for 48 h, the plates were stored frozen at −80°C. Construction of YAC libraries is notoriously difficult and success or failure due to many factors. All primary transformants (Ura) were transferred to AHC plates (Ura − Trp Ade) to allow additional selection for the Trp marker and red coloration of recombinant colonies from the loss of sup-4o function where adenine metabolism is uninterrupted. The yeast strain AB1380 was used for the first 21,984 clones; more than 90% of these were recovered in eight transformations. The final 6048 clones were generated in two transformations with J570 as the host strain. Indeed, the final 9312 YACs were recovered after canavanine selection, included in all J570 transformations. This strain is reported to transform at a considerably higher efficiency than AB1380 (Haldi et al. 1996) with identical selection criteria, although in the transformations performed here this was not apparent. The average size of inserts in the library was determined by analyzing 81 randomly selected clones. The YAC clones are sized on Southern blots of total yeast chromosomes following PFGE hybridized to a pBR322 plasmid probe specific for the YAC arms. The average insert size was found to be 500 kbp, ranging from 100 kbp to 1600 kbp (Fig. 1). This library will afford a 99% chance for Correspondence to: D. Hills Mammalian Genome 10, 837–838 (1999).