The genexpress IMAGE knowledge base of the human muscle transcriptome: a resource of structural, functional, and positional candidate genes for muscle physiology and pathologies.

Sequence, gene mapping, and expression data corresponding to 910 genes transcribed in human skeletal muscle have been integrated to form the muscle module of the Genexpress IMAGE Knowledge Base. Based on cDNA array hybridization, a set of 14 transcripts preferentially or specifically expressed in muscle have been selected and characterized in more detail: Their pattern of expression was confirmed by Northern blot analysis; their structure was further characterized by full-insert cDNA sequencing and cDNA extension; the map location of the corresponding genes was refined by radiation hybrid mapping. Five of the 14 selected genes appear as interesting positional and functional candidate genes to study in relation with muscle physiology and/or specific orphan muscular pathologies. One example is discussed in more detail. The expression profiling data and the associated Genexpress Index2 entries for the 910 genes and the detailed characterization of the 14 selected transcripts are available from a dedicated Web server at. The database has been organized to provide the users with a working space where they can find curated, annotated, integrated data for their genes of interest. Different navigation routes to exploit the resource are discussed.

The Genexpress IMAGE knowledge base of the human brain transcriptome: a prototype integrated resource for functional and computational genomics.

Expression profiles of 5058 human gene transcripts represented by an array of 7451 clones from the first IMAGE Consortium cDNA library from infant brain have been collected by semiquantitative hybridization of the array with complex probes derived by reverse transcription of mRNA from brain and five other human tissues. Twenty-one percent of the clones corresponded to transcripts that could be classified in general categories of low, moderate, or high abundance. These expression profiles were integrated with cDNA clone and sequence clustering and gene mapping information from an upgraded version of the Genexpress Index. For seven gene transcripts found to be transcribed preferentially or specifically in brain, the expression profiles were confirmed by Northern blot analyses of mRNA from eight adult and four fetal tissues, and 15 distinct regions of brain. In four instances, further documentation of the sites of expression was obtained by in situ hybridization of rat-brain tissue sections. A systematic effort was undertaken to further integrate available cytogenetic, genetic, physical, and genic map informations through radiation-hybrid mapping to provide a unique validated map location for each of these genes in relation to the disease map. The resulting Genexpress IMAGE Knowledge Base is illustrated by five examples presented in the printed article with additional data available on a dedicated Web site at the address http://idefix.upr420.vjf.cnrs.fr/EXPR++ +/ welcome.html.

The Genexpress Index: a resource for gene discovery and the genic map of the human genome.

Detailed analysis of a set of 18,698 sequences derived from both ends of 10,979 human skeletal muscle and brain cDNA clones defined 6676 functional families, characterized by their sequence signatures over 5750 distinct human gene transcripts. About half of these genes have been assigned to specific chromosomes utilizing 2733 eSTS markers, the polymerase chain reaction, and DNA from human-rodent somatic cell hybrids. Sequence and clone clustering and a functional classification together with comprehensive data base searches and annotations made it possible to develop extensive sequence and map cross-indexes, define electronic expression profiles, identify a new set of overlapping genes, and provide numerous new candidate genes for human pathologies.

Human TRE17 oncogene is generated from a family of homologous polymorphic sequences by single-base changes.

The tre oncogenic locus was identified in transformants receiving human DNA from Ewing's sarcoma cells EW1. Genetic elements of tre originate from chromosomes 5, 18, and 17. The TRE17 oncogene is consistently transcribed in various human cancer cells and proves oncogenic (onc) in expression vector-based assays. Here, the nucleotide sequence of TRE17 with defined noncoding and two coding exons (4,426 nucleotides) was compared with sequences cloned from placental DNA library or generated by polymerase chain reaction (PCR) from EW1 and independent healthy individuals. Cloned sequences displayed restriction site polymorphism, with different patterns for EW1 and normal tissues. Sequence analysis revealed that they originate from a family of homologous sequences alpha, beta, and gamma. TRE17 alpha and less frequent TRE17 beta (similarity score approximately equal to 88%) were found in both normal and EW1 cells. oncTRE17, classified as TRE17 beta, differed from the wild-type TRE17 beta, besides a few intronic changes, by a single-base frameshift insertion in one of the coding exons. TRE17 gamma, so far identified in EW1 but not in normal somatic cells, diverged from oncTRE17 by 6% nucleotide substitutions and by stop codons in each reading frame. The results are consistent with the possibility that TRE17 sequences other than oncTRE17 are translated if alternatively spliced. Expression of TRE17 in normal somatic cells was, however, not yet reported.

Mutagenicity of a unique apurinic/apyrimidinic site in mammalian cells.

Abasic sites are common DNA lesions produced either spontaneously or as a consequence of the action of some genotoxic agent. The mutagenic properties of a unique abasic site replicated in mammalian cells have been studied using a shuttle vector. A plasmid, able to replicate both in mammalian cells and in bacteria, carrying a unique abasic site chemically synthesized has been constructed. After replication in mammalian cells, plasmid DNA was recovered and used to transform bacteria. Mutants were screened without selection pressure by differential hybridization with a labelled oligonucleotide and their DNA was sequenced. A mutation frequency ranging from 1% to 3% was found, depending on the base originally inserted during the vector construction, opposite the abasic site. All the sequenced mutants correspond to single base-pair substitutions targeted at the abasic site. We observed a deficit in guanine incorporation opposite the abasic site, while the three other bases were incorporated with a similar efficiency. The mutational potency of abasic sites was observed without any voluntary preconditioning treatment of mammalian cells in order to induce "SOS" like conditions.

A novel transcriptional unit of the tre oncogene widely expressed in human cancer cells.

Tre is a recombinant gene isolated from NIH3T3 cells transfected with human Ewing's sarcoma DNA. It is composed of three major genetic elements derived, 5' to 3', from human chromosomes 5, 18 and 17. We report here on transcripts from the 3' domain of tre. The transcripts were cloned from a cDNA library of cytoplasmic poly(A)+ RNA from tre-transfected NIH3T3 tumor cells. The complete cDNA sequence, 8201 nucleotides, possessed an unusually long non-coding region and a translatable region with two open reading frames. In one cDNA clone, the presence of two insertions suggested the possibility of alternative splicing. The sequence mapped to the centromere-proximal region of 17q. Transfection-tumorigenicity assays with the open reading frames subcloned into expression vectors were positive for the reading frame adjacent to the 5' non-coding region and negative for the second, downstream, reading frame and the possible alternatively spliced versions of both reading frames. Analysis of the 786 amino acid sequence deduced from the 5' reading frame predicted a highly hydrophilic protein with two charge clusters suggesting nucleic acid-binding properties. When used as probe, the cloned sequence detected RNA transcripts in a wide variety of human cancer cells regardless of their lineage of origin from different tissues, but not in human cells from normal tissue.

Indefinite growth of mammalian cell autotrophs on amino acids, vitamins, and glucose.

Established mammalian cells segregate variants, named autotrophs, able to proliferate in the absence of hormones and proteins of any kind. We provided complete formulas for autotrophic media composed of amino acids, vitamins, glucose, and inorganic salts only and showed that such media are sufficient to sustain continuous propagation of Chinese hamster fibroblasts and human keratinocytes. The cells grew on bare polystyrene and divided every 32 to 44 hours reaching approximately 8 x 10(5) cells/cm2. The results showed that the autotrophs may replace conventional cultures in most experimental and industrial cell production systems.

Nucleotide sequence of mouse L19 ribosomal protein cDNA isolated in screening with tre oncogene probes.

A cDNA library was prepared from cytoplasmic poly(A)RNA from mouse NIH-3T3 cells carrying a transfected human tre oncogene. Screening with tre gene probes identified a tre cDNA clone 11-4 and a co-purifying weakly hybridizing cDNA clone 11-5. The 11-5-specific RNA was expressed in both nontransfected and tre-transfected NIH-3T3 cells, showing it is of mouse rather than tre gene origin. Its nucleotide sequence was 717 bp long and contained, starting from the first nucleotide, an open reading frame of 588 bp followed by a 3' noncoding region and 26 A residues at the 3' terminus. Comparison with the GenBank data base revealed 93.7% homology with cDNA encoding the rat L19 ribosomal protein. Furthermore, the 196-amino-acid polypeptide deduced from 11-5 was of the same length and contained only one amino acid difference compared with the rat L19 protein. Comparison with the weakly hybridizing tre gene probe showed stretches of homology that were, however, too short to be taken into consideration. We conclude that the 11-5 sequence encodes the mouse L19 ribosomal protein.

Isolation of mammalian cell autotrophs able to grow indefinitely in protein-free chemically defined medium.

A selection procedure is described for obtaining cell variants, called autotrophs, which are able to grow from low-molecular-weight metabolic precursors in the absence of macromolecules and serum replacement factors. These autotrophs can readily be isolated from a variety of cell lines, including normal (nontransformed) NIH3T3 cells, their spontaneous or tre oncogene-transfected tumor-producing derivatives, and HeLa cells. Tre-transfected autotrophs grew faster than their normal or spontaneously transformed counterparts, and much faster than autotrophic HeLa cells. Chinese hamster autotrophs, which have now been cultured for nearly two years, grow with a population doubling time of 30 h.

Autocrine factor-independent growth of mammalian fibroblasts established in fully synthetic medium: no v-onc requirement in establishment.

In a previous study Chinese hamster fibroblasts carrying a partially deleted v-src were established in a synthetic medium lacking macromolecular supplements and shown to possess a particular serum-free phenotype hereafter designated sf. In cloning efficiency assays, sf, unlike wild-type, fibroblasts required a threshold cell density to grow from single cells, suggesting autocrine stimulation. In the present study a conditioned medium harvested from sf cells was added to the same sf cells, and the resulting cloning density was found to markedly diminish rather than increase. Sf cells were found to be unable to grow at cloning density because of trypsin damage: sf cells seeded into trypsin inhibitor-containing medium cloned with no requirement for threshold cells and were therefore independent of autocrine secretion from neighboring cells. Their cloning efficiency reached 7.7%; this value could not be improved by subcloning the sf culture, and it diminished when selenium was not added to the assay medium. To determine whether v-src is involved in the sf phenotype, five clones of the parental Chinese hamster fibroblast line not infected with Rous sarcoma virus were explanted into serum-free cultures with no macromolecular additives as in the case of v-src-containing cells. Each clone gave rise to an sf cell line growing indefinitely in synthetic medium like the v-src-containing sf cells, showing that the v-src gene is not required either for the establishment or maintenance of the sf phenotype.

Ability of mammalian fibroblasts to grow in synthetic medium containing neither serum nor exogenously added macromolecules.

Rous sarcoma virus transformed Chinese hamster fibroblasts, clone CHR1-3, were established at high temperature, then subcloned. Six subclones with round and flat morphology harboring undeleted and partially deleted RSV proviruses, respectively, were seeded into serum-free synthetic medium with no macromolecular additives, and maintained for 2 mo. One flat subclone no. 14, fully designated sfCHR1-3.14 for its serum-free phenotype, was further propagated in the same medium. The cells grew exponentially in loosely attached monolayers and could be serially passaged on bare polystyrene with an average population doubling time of 46 h. Cell attachment could be improved by using collagen-coated polystyrene or by adding a methionine supplement to the culture medium. Furthermore, the sfCHR1-3.14 cells could be subcloned and further grown in nonselective medium. The reversion rate of the sf phenotype was estimated to be 1 to 2%/cell generation. Evidence for an autocrinal stimulation was obtained by cloning efficiency assays showing a requirement for a threshold cell density. Slight growth stimulation could also be detected in assays using conditioned medium from sfCHR1-3.14 cells and serum-restricted wild-type (wt)NIH3T3, but not wtCHR1-3.14, cells as indicator cells. Finally, wtNIH3T3 cells used in these assays were assayed for serum-free growth and found to be able to develop their own sf phenotype; in this respect they resemble the previously established sfCHR1-3.14 cells.

A rearranged transforming gene, tre, is made up of human sequences derived from chromosome regions 5q, 17q and 18q.

The transfection recombinant transforming gene, tre, originated from discontinuous human genetic elements after transfection of NIH3T3 cells with genomic DNA from a Ewing's sarcoma cell line. Probes for the three normally discontinuous human elements involved in the transfection recombinant were subcloned and used in conjunction with a panel of rodent-human hybrid cells to determine their normal location in the human genome. The leftmost (5') element derives from the long arm of chromosome 5 (5q); the internal fragment derives from human chromosome 18 proximal to the bcl-2 gene; and the rightmost (3') element derives from the long arm of chromosome 17 (17q) distal to an acute leukemia breakpoint at 17q21. In situ hybridization of the same probes to human metaphase chromosomes confirmed localization of these sequences to regions 5q23----q31, 18q12, and 17q12----q22.

Antimutagenic action of methionine in Chinese hamster fibroblasts: an opposite of methionine mutagenicity.

Serially propagated fibroblast clones were treated with 25-mM methionine and a thioguanine resistance assay was performed at each passage to measure mutagenesis at the HGPRT locus. The spontaneous frequency of thioguanine resistant mutants before treatment was less than 5 X 10(-6) in four clones and 22 X 10(-6) in one clone. These values dropped 10- to 100-fold and remained so until the methionine was withdrawn, then returned to, or overshot, the initial values. It is proposed that these fibroblasts, unlike their previously described RSV-transformed counterparts, possess a repair enzyme capable of an adaptive response to S-adenosylmethionine, the metabolite thought to be responsible for methionine mutagenicity in vivo.

Molecular cloning of a novel oncogene generated by DNA recombination during transfection.

Genomic DNA of a Ewing's sarcoma cell line and pKOneo were cotransfected into NIH3T3 cells and assayed for tumorigenicity in nude mice. Primary transfected tumors all contained human DNA. One transfected sequence was retained through secondary and tertiary tumors, and was cloned in four overlapping cosmids covering 51.5-kb human DNA flanked on both sides with mouse DNA. The cloned sequence, though continuous in transfected tumors, originated from three major genetic elements discontiguous in human cells and recombined during transfection. Accordingly there was one species of hybrid RNA transcripts of 9 kb in the transfected tumors and no transcripts in human cells, including Ewing's sarcoma cells. Evidence that the cloned sequence is a biologically active novel oncogene was provided by transfection of a mixture of two cosmid clones which generated tumors where they were transcribed into 9-kb RNA. The oncogene was named tre to designate its origin from transfection recombined DNA molecules.

Mutagenicity of methionine or its metabolic products in RSV-transformed Chinese hamster fibroblasts.

Methionine has been successfully used to control tumor progression in vivo and to induce reversions of transformed cells in vitro. In the present study, we measured mutations at the HGPRT locus of RSV-transformed cells serially propagated in methionine-supplemented medium and assayed at each passage for thioguanine resistance. The frequency of spontaneous mutants at this locus was 7.2 X 10(-5); this value gradually increased during the methionine treatment to as much as 9.2 X 10(-4), and returned to initial values when the methionine treatment was withdrawn. It is proposed that the mutants were induced by the methionine derivative, S-adenosylmethionine, and the resulting mutant frequency determined by equilibrium between mutagenic action of this metabolite and DNA repair.

Loss of the oncogene from human H-ras-1-transfected NIH/3T3 cells grown in the presence of excess methionine.

Mouse NIH/3T3 cells, transfected with T24 genomic DNA (J10 cells), carry the human H-ras-1 (hu-H-ras-1) oncogene stably integrated in the chromosomal DNA and display the transformed phenotype with a "criss-cross" morphology and capacity for anchorage-independent growth. When these cells were cultured in the presence of 25 mM DL-methionine, they lost their transformed phenotype; the characters of the normal phenotype (before transfection) reappeared, even though cell viability, as measured by the cloning potential on a solid support, was fully conserved. Southern blot analysis showed that cells continuously propagated in methionine-supplemented medium (J10met), unlike cells grown in normal medium, progressively segregated the hu-H-ras-1 gene. The progressively fewer cells still retaining the oncogene could be selected by inoculation of J10met cells into nude (nu/nu) mice of Swiss background. After prolonged methionine treatment, however, no such oncogene-positive cells could be detected, even when the excess methionine was withdrawn from the culture medium after 105 days. Conclusive evidence for the loss of the hu-H-ras-1 oncogene sequence was provided by subcloning the J10met cells. The oncogene sequence could no longer be found, and a fraction of subclones was no longer tumorigenic when assayed in nude mice. Other subclones generated late tumors negative for the hu-H-ras-1 transforming sequence, and a small fraction of subclones gave rise to rapidly growing early tumors that were also negative for the transforming sequence. The question of how methionine can induce the phenotypic and genetic change from malignancy to normality is discussed.

RSV provirus with same flanking sequences is found on different size classes of Chinese hamster chromosomes.

Rous sarcoma virus(RSV)-transformed Chinese hamster fibroblasts, containing approximately ten copies of the DNA domain comprising a single provirus and its flanking cellular sequences, were arrested in metaphase, and the chromosomes were fractionated by size in a sucrose gradient. The resolution of polymorphic ribosomal genes, the dihydrofolate reductase gene, and the c-src gene demonstrated that the gradient can distinguish between small, medium, and large chromosomes. The same DNA domain carrying the RSV provirus was found to be associated with chromosomes of all three size classes. Polymorphic copies of the domain in small and large chromosomes could be distinguished from those in medium-sized chromosomes because of the polymorphism in the XhoI and EcoRI sites on 5' and 3' adjacent cellular sequences, respectively. The presence of the same provirus domain on different chromosomes, together with the karyological data showing trisomies in the same chromosome size classes, suggest that the provirus domain, possibly the entire replicon, was duplicated and transferred to different nonhomologous chromosomes, and the transfer was followed by duplication of the target chromosomes. The possibility that proviral long terminal repeats might be involved in replicon transfer is discussed.

Isolation of a line of immortal chicken embryo fibroblasts after transfection with the nuclei of Rous sarcoma virus-transformed Chinese hamster cells.

Secondary cultures of chicken embryo fibroblasts were transfected with purified nuclei from lysed cells of a clonal line of temperature-sensitive Rous sarcoma virus (tsRSV)-transformed Chinese hamster fibroblasts. After propagation for 3 months an established cell line designated ChR32 was obtained in one chicken cell culture. The cells of this line have been propagated so far for 18 months, whereas normal chicken embryo fibroblasts died after 2 months. The established cells were heteroploid with a diploid modal number of macrochromosomes and two Z chromosomes. No Chinese hamster chromosomes could be identified. Southern blot analysis of DNA from the uncloned ChR32 cells and the clones provided evidence that these established cells were, in fact, clonal in origin and contained full-length RSV proviruses and no defective proviruses. Furthermore, they contained, at the 3' end proviral-cellular junction, Bg/II, HpaI, KpnI, SacI, and XbaI fragments of the same size as the Chinese hamster donor cells, suggesting that the cellular sequence adjacent to the provirus is of Chinese hamster origin. The cells after establishment were able to grow continuously at 37 degrees or 41 degrees C and produce a large amount of ts sarcoma virus particles. A corollary finding was that these virus particles were non-leaky for the transforming function at the non-permissive temperature.

Amplification of the provirus region in Rous sarcoma virus-transformed Chinese hamster cells and segregation of the amplified copies in somatic cell hybrids.

Four independent clones of RSV-transformed Chinese hamster fibroblasts were isolated. Southern blots and dot hybridization studies showed that in three out of four clones there were four to eight times as many integrated proviruses as in the fourth clone which contained at least one complete provirus. Restriction mapping studies showed that although the integration site varied from clone to clone, all the proviral copies in the same clone shared the same flanking cellular sequences. In one clone there are at least two polymorphic proviral variants A and B, one with and one without a BglI site. Further experiments were performed to see if the variants could be physically separated. RSV-Transformed Chinese hamster cells resistant to thioguanine were fused with mouse cells to give somatic hybrids which preferentially segregate Chinese hamster chromosomes. Ten out of eleven hybrids positive for virus rescue have lost up to 90% of the parental provirus copies. Four of these hybrids were found to contain the provirus variant A alone, one the variant B alone, and the rest contained both variants. All the proviruses retained in somatic hybrids shared the flanking cellular sequences of the parental provirus. Provirus segregation in somatic hybrids confirms that multiple (about ten) copies of the provirus region are present in the karyotype of parental RSV-transformed cells and, furthermore, suggests that the amplified copies of this region are translocated to different chromosomes.