An anchored framework BAC map of mouse chromosome 11 assembled using multiplex oligonucleotide hybridization

Despite abundant library resources for many organisms, physical mapping of these organisms has been seriously limited due to lack of efficient library screening techniques. We have developed a highly efficient strategy for large-scale screening of genomic libraries based on multiplex oligonucleotide hybridization on high-density genomic filters. We have applied this strategy to generate a bacterial artificial chromosome (BAC) anchored map of mouse chromosome 11. Using the MIT mouse SSLP data, 320 pairs of oligonucleotide probes were designed with an "overgo" computer program that selects new primer sequences that avoid the microsatellite repeat. BACs identified by these probes are automatically anchored to the chromosome. Ninety-two percent of the probes identified positive clones from a 5.9-fold coverage mouse BAC library with an average of 7 positive clones per marker. An average of 4.2 clones was confirmed for 204 markers by PCR. Our data show that a large number of clones can be efficiently isolated from a large genomic library using this strategy with minimal effort. This strategy will have wide application for large-scale mapping and sequencing of human and other large genomes.

Tumour susceptibility and spontaneous mutation in mice deficient in Mlh1, Pms1 and Pms2 DNA mismatch repair

Germline mutations in the human MSH2, MLH1, PMS2 and PMS1 DNA mismatch repair (MMR) gene homologues appear to be responsible for most cases of hereditary non-polyposis colorectal cancer (HNPCC; refs 1-5). An important role for DNA replication errors in colorectal tumorigenesis has been suggested by the finding of frequent alterations in the length of specific mononucleotide tracts within genes controlling cell growth, including TGF-beta receptor type II (ref. 6), BAX (ref. 7) and APC (ref. 8). A broader role for MMR deficiency in human tumorigenesis is implicated by microsatellite instability in a fraction of sporadic tumours, including gastric, endometrial and colorectal malignancies. To better define the role of individual MMR genes in cancer susceptibility and MMR functions, we have generated mice deficient for the murine homologues of the human genes MLH1, PMS1 and PMS2. Surprisingly, we find that these mice show different tumour susceptibilities, most notably, to intestinal adenomas and adenocarcinomas, and different mutational spectra. Our results suggest that a general increase in replication errors may not be sufficient for intestinal tumour formation and that these genes share overlapping, but not identical functions.

Association of mouse fibrinogen-like protein with murine hepatitis virus-induced prothrombinase activity

Previously, we demonstrated induction of a unique macrophage prothrombinase during infection of BALB/cJ mice by mouse hepatitis virus strain 3 (MHV-3). By immunologic screening, a clone representing this prothrombinase was isolated from a cDNA library and sequenced. The sequence identified this clone as representing part of a gene, musfiblp, that encodes a fibrinogen-like protein. Six additional clones were isolated, and one clone, p11-3-1, encompassed the entire coding region of musfiblp. Murine macrophages did not constitutively express musfiblp but, when infected with MHV-3, synthesized musfiblp-specific mRNA. musfiblp mRNA induction was earlier and significantly greater in BALB/cJ than A/J macrophages. Prothrombinase activity was demonstrated when musfiblp was expressed from p11-3-1 in RAW 264.7 cells. These data suggest that musfiblp encodes the MHV-induced prothrombinase.

MHV-3 induced prothrombinase is encoded by musfiblp

Previously, we demonstrated induction of a unique macrophage prothrombinase, PCA, in MHV-3 infected BALB/cJ mice. By immunologic screening, a clone representing PCA was isolated from a cDNA library and sequenced. The sequence identified this clone as representing part of a gene, musfiblp, that encodes a fibrinogen-like protein. Six additional clones were isolated, and one clone, p11-3-1, encompassed the entire coding region of musfiblp. Murine macrophages did not constitutively express musfiblp, but when infected with MHV-3, synthesized musfiblp-specific mRNA. Musfiblp mRNA induction was earlier and significantly greater in BALB/cJ than A/J macrophages. Prothrombinase activity was demonstrate when musfiblp was expressed from p11-3-1 in RAW 264.7 cells. These data suggest that musfiblp encodes the MHV-induced prothrombinase.

Postsynthetic methylation of core histones in K562 cells is associated with bulk acetylation but not with transcriptional activity

The relationship between the postsynthetic modification of core histones by methylation and transcriptionally active chromatin was investigated in K562 erythroleukemia cells. Cells were incubated with L-[methyl-3H]methionine in the presence of cycloheximide. Under these conditions, only histones H3 and H4 were detectably methylated. Chromatin was fractionated by several methods, including low-salt elution, mononucleosome gel mobility shift using HPLC-purified HMG 17, and cesium chloride-guanidine equilibrium gradient centrifugation of formaldehyde-fixed chromatin. By these latter two methods, chromatin highly enriched for transcriptionally engaged or competent genes was isolated. A significant correlation was noted between postsynthetic modification of histones by methylation and by the slow-turnover form of acetylation. However, there was no enrichment of methylated histones in the transcriptionally competent fraction of chromatin isolated by HMG 17 binding. Moreover, only minor enrichment of methylated histone H3.1 and no enrichment of methylated histones H3.2 and H4 was detected in transcriptionally engaged chromatin isolated by gradient centrifugation. Chromatin soluble in low-salt buffer was found to be significantly enriched in methylated histones, but not in active genes. We conclude that histone methylation is associated with both transcriptionally active and transcriptionally inactive chromatin. The function of this modification, like that of bulk histone acetylation, remains to be determined.

Avian nuclear matrix proteins bind very tightly to cellular DNA of the beta-globin gene enhancer in a tissue-specific fashion

We have previously shown that a cloned 480 bp DNA fragment that spans the 3'-enhancer region of the avian beta-globin gene cluster can become very tightly, perhaps covalently, bound to protein in avian nuclear matrices in vitro [Zenk et al. (1990) Biochemistry 29, 5221-5226]. This binding was not tissue-specific and was probably not mediated by topoisomerase enzymes. In the present study, we have examined avian nuclear matrices (or scaffolds) for the presence of very tight cellular DNA-protein complexes in the region of the beta-globin gene enhancer and of several other avian genes. Nuclear matrices were prepared by both high- and low-salt methods, and protein-DNA complexes were isolated by SDS/K+ precipitation after restriction enzyme digestion. In adult reticulocytes, up to 30% of the intact 3800 bp HindIII-EcoRI fragment that encompasses the beta-globin enhancer element may be very tightly bound to nuclear matrix protein. In adult avian thymus nuclei, the beta-globin enhancer is neither matrix-associated nor tightly bound to protein. In contrast, a 5.0-kb HindIII fragment of the malic enzyme gene is very tightly bound to nuclear matrix-associated protein in thymus cells, but not reticulocytes. The malic enzyme gene is active in thymus cells, and not in reticulocytes. These results suggests that certain regions of cellular DNA are very tightly, perhaps covalently, attached to nuclear matrix-associated proteins. Attachment follows a tissue-specific pattern that is associated with transcriptional activity.

Binding of HMG 17 to mononucleosomes of the avian beta-globin gene cluster in erythroid and non-erythroid cells

The binding of HMG 17 to stripped core mononucleosomes containing DNA from the avian beta-globin gene cluster was examined to determine whether binding in vitro in this developmentally-regulated gene domain was associated with transcriptional activity or DNaseI-sensitivity in intact nuclei. Mononucleosomes were prepared from primitive and definitive stage embryonic red blood cells of chick embryos, adult reticulocytes, adult reticulocytes in which embryonic rho-globin transcription was induced, and adult thymus cells. Preferential binding by HMG 17 to mononucleosomes containing the beta-globin gene cluster was confined to erythroid-derived mononucleosomes that contain the embryonic rho-globin gene, the adult beta-globin gene, and DNA sequences located between these two genes, but not to those that contain the embryonic epsilon-globin gene. Comparison of these results to the known patterns of transcription and DNaseI-sensitivity within the beta-globin gene cluster shows that HMG 17 binding, although tissue-specific, does not correlate directly with either DNaseI-sensitivity or active gene transcription, but is dependent on other factors present in core mononucleosomes from this active gene domain.