Variations of p53 in cultured fibroblasts of patients with lung cancer who have a presumed genetic predisposition.

The authors studied cultured skin fibroblasts of 64 patients with lung cancer for constitutive mutations of the p53 tumor suppressor gene by using polymerase chain reaction and single-strand conformation polymorphism covering the entire coding region. The patients were considered to be genetically predisposed because lung cancer had developed in 25 of them before age 46 and because 42 of them had at least one first-degree relative with lung cancer. One mutation was detected at position 235 coding for serine instead of asparagine in the conserved DNA binding domain. The Pro/Pro genotype at codon 72 of p53, considered to harbor an increased risk for lung cancer, could not be detected with increased frequency in this study's patients. From these data, the authors conclude that constitutive variations of the p53 gene do not represent a major genetic determinant for lung cancer.

Loss of epithelial polarity is accompanied by differential association of proteins with intracellular membranes.

Cellular membranes play an important role in the formation and maintenance of epithelial polarity, which is lost early during carcinogenesis. We set out to identify membrane proteins which are altered during loss of cell polarity in mammary epithelium. As a model system we used murine mammary epithelial cells expressing the conditional oncoprotein c-JunER, which induces a reversible loss of polarity upon beta-estradiol-driven activation [1]. When grown either in the absence or presence of hormone, these cells exhibit a polarized or unpolarized phenotype, respectively. Different membrane fractions of polarized or unpolarized cells were analyzed by two-dimensional electrophoresis (2-DE) and differentially expressed membrane proteins were identified. To distinguish between transmembrane orientation and peripheral attachment of these proteins, were performed extractions with carbonate at high pH or with Triton X-114. In addition, cytosolic proteins of both states were analyzed to investigate their differential association with distinct membrane fractions. We found ten protein spots preferentially or exclusively in polarized cells and 17 other proteins as being upregulated during loss of polarity. Some of the peripheral membrane proteins were identified by microsequencing. The resident Golgi protein nucleobindin and fructose-bisphosphate aldolase were preferentially associated with membranes of polarized cells, whereas alphaB crystallin was detected exclusively and in high amounts in unpolarized cells.

Molecular cloning and expression analysis of five novel genes in chromosome 1p36.

The human chromosome 1p36 region displays frequent nonrandom chromosomal deletions and translocations in a number of human malignancies; these are thought to inactivate tumor suppressor genes. To identify these putative tumor suppressors we employed exon trapping, cDNA selection, and zoo blot analysis to clone five new genes located in 1p36. Two of these represent novel genes and were designated C1orf1 and xylan 1,4-beta-xylosidase 1 (XBX1). Two further genes represented new members of known gene families: PTPRZ2 was a tyrosine phosphatase and FRAP2 represented a FKBP12-rapamycin-associated protein. The fifth gene identified, ENO1L1, was significantly homologous to c-myc promoter binding protein, MBP-1, and to enolase 1 (ENO1). It colocalized with alpha enolase (ENO1) on a single P1 clone. ENO1L1 differed from both ENO1 and MBP-1 in the organization of its 5' untranslated sequences. Second, MBP-1 contained two single-base insertions not present in either ENO1 or ENO1L1 sequences, which led to a shift in the MBP-1 reading frame. Expression analysis revealed two brain-specific transcripts of 7.9 and 6.5 kb for PTPRZ2. In contrast, C1orf1, FRAP2, ENO1L1, and XBX1 appeared to be expressed ubiquitously in the tissues tested, with transcript sizes of 4.5, 8.7, 1.75, and 4.5 kb, respectively. Using fluorescence in situ hybridization, we mapped the five novel genes relative to chromosome 1p36 breakpoints present in three established tumor cell lines and one nontumor cell line. The karyotypic abnormalities in these cell lines were exploited as chromosomal landmarks; we could thus show that the telomere to centromere gene order was PTPRZ2-(MBP-1/ENO1/ENO1L1)-(C1orf1/XBX1)-+ ++FRAP2. The localization of these genes to a chromosomal region that is prone to deletions in human cancers makes them potential candidate tumor suppressors.

Structure and chromosomal assignment of the mouse fra-1 gene, and its exclusion as a candidate gene for oc (osteosclerosis).

We have determined the genomic structure of the mouse fra-1 gene, which consists of four exons and three introns at positions also found in the other members of the fos gene family. Fra-1 is expressed rather highly in the brain and testes of adult mice, and at low levels in most other tissues. Absence of c-Fos leads to significantly reduced serum stimulation of fra-1 expression in gene targeted mouse fibroblasts, demonstrating that mitogen induction of fra-1 is partially mediated by c-Fos/AP-1. A polymorphic (CA)n microsatellite marker was found in intron 2 of fra-1 and used to map the gene to the centromeric region of mouse chromosome 19. Since fra-1 maps to the same genomic region as oc (osteosclerosis), an autosomal recessive disorder leading to the bone remodelling disease osteopetrosis, we tested it as a candidate gene for oc. The segregation of fra-1 in two different crosses of mice carrying oc and an allelism test between oc and a targeted disruption of fra-1 demonstrate that fra-1 and oc are two distinct genes rather than oc being a mutant allele of fra-1.

The complete DNA sequence and genomic organization of the avian adenovirus CELO.

The complete DNA sequence of the avian adenovirus chicken embryo lethal orphan (CELO) virus (FAV-1) is reported here. The genome was found to be 43,804 bp in length, approximately 8 kb longer than those of the human subgenus C adenoviruses (Ad2 and Ad5). This length is supported by pulsed-field gel electrophoresis analysis of genomes isolated from several related FAV-1 isolates (Indiana C and OTE). The genes for major viral structural proteins (Illa, penton base, hexon, pVI, and pVIII), as well as the 52,000-molecular-weight (52K) and 100K proteins and the early-region 2 genes and IVa2, are present in the expected locations in the genome. CELO virus encodes two fiber proteins and a different set of the DNA-packaging core proteins, which may be important in condensing the longer CELO virus genome. No pV or pIX genes are present. Most surprisingly, CELO virus possesses no identifiable E1, E3, and E4 regions. There is 5 kb at the left end of the CELO virus genome and 15 kb at the right end with no homology to Ad2. The sequences are rich in open reading frames, and it is likely that these encode functions that replace the missing El, E3, and E4 functions.

Cloning and characterization of CpG islands of the human chromosome 1p36 region.

CpG islands were identified and localized to chromosome 1p36 by means of pulsed-field gel blot hybridization with 1p36-specific microclone probes. Five CpG islands, designated CpG17, CpG28, CpG60, CpG112a, and CpG112b, were molecularly cloned from corresponding cosmids. All five islands are associated with transcribed sequences, as shown by RNA blot hybridizations. Screening of cDNA libraries with the island-specific genomic probes led to the isolation of two cDNA clones to date. These encode the human transcription factor E2F-2 and the dominant-negative helix-loop-helix gene ID3, respectively. Pulsed-field gel electrophoresis analysis also revealed that these two genes are located next to each other at a distance of about 25 kb.

Mutations in tumour suppressor gene p53 in two feline fibrosarcomas.

Exons 5-7 of the tumour suppressor gene p53 were investigated in genomic DNA of tumours of domestic cats. In one fibrosarcoma investigated we observed a mutation GAG-->AAG (glutamic acid-->lysine) in codon 180; in another there was a mutation CGG-->TGG (arginine-->tryptophane) in codon 248.

Conservation of a maternal-specific methylation signal at the human IGF2R locus.

The human IGF2R gene has been reported to be either biallelically or very rarely monoallelically expressed, in contrast to the maternally expressed mouse counterpart. We describe here an analysis of the 5' portion of the human IGF2R gene and show that it contains a maternally methylated CpG island in the second intron. A similar maternally methylated intronic element has been proposed to be the imprinting box for the mouse gene and although the relevance of this element has yet to be directly demonstrated, methylation has been reported to be essential to maintain allele-specific expression of imprinted genes. Allelic expression analysis of human IGF2R in 70 lymphoblastoid cell lines identified only one line showing monoallelic expression. Thus, in this tissue monoparental methylation of the IGF2R gene does not correlate with allele-specific expression. We also confirm here that the human IGF2R gene is located in an asynchronously replicating chromosomal region, as are all other imprinted genes so far analyzed. The mouse and human IGF2R intronic CpG islands both contain numerous large direct repeats that are methylated following maternal, but not paternal, transmittance. Thus features that attract maternal-specific methylation are conserved between the mouse and human genes. Since these intronic CpG islands share organizational rather than sequence homology, this suggests that secondary DNA structure may play a role in attracting a maternal methylation imprint.

Sequence of an exon of tumour suppressor p53 gene--a comparative study in domestic animals: mutation in a feline solid mammary carcinoma.

Partial sequence determinations were performed on exon 8 of tumour suppressor gene p53 of cattle, sheep, goat, horse and pig. High sequence homology between these species and other species including dog, cat, chicken and man is demonstrated. A mutation CGG-->TGG (arginine-->tryptophan) was detected in a feline solid carcinoma of the mammary gland.

Expression of three- and four-repeat tau isoforms in mouse liver.

Tau protein is a member of the family of microtubule-associated proteins, which support microtubule polymerization and stability. Under pathological conditions, tau is a major constituent of neurofibrillary tangles in nerve cells of patients with Alzheimer's disease. Neurofibrillary tangles share some morphological, biochemical and immunological properties with cytoplasmic inclusions associated with other diseases, such as Mallory bodies in the livers of patients with alcoholic hepatitis and in corresponding mouse models. Recently a Mallory body component was identified that in molecular mass and isoelectric point resembles the abnormally phosphorylated tau of neurofibrillary tangles. There has been, however, so far no report describing the occurrence of tau in normal liver. We now demonstrate the expression of two tau isoforms containing three and four repeats, respectively, of the microtubule-binding domains in normal mouse liver and kidney. This finding provides evidence for a physiological role of tau in the liver and, consequently, the basis for the involvement of tau in pathological situations.

Structure and expression of the chicken retinoblastoma gene.

The molecular cloning and initial characterization of the complementary DNA encoding the chicken homologue of the human retinoblastoma susceptibility gene product Rb are described. Chicken Rb (chRb) was found to be highly homologous to human, mouse, and Xenopus Rb in primary amino acid sequence, underlining its conserved function in cell cycle control in higher eukaryotes. The highest level of homology was found within the Rb "pocket domains" and at the Rb COOH terminus, whereas the extreme NH2 terminus of chRb is divergent. We also show that transcription of chRb initiates at multiple start sites within a DNA segment which represents a CpG island. ChRb is transcribed as a 4.7-kilobase (kb) mRNA species in a variety of normal chicken tissues and oncogene-transformed hematopoietic cells. However, in v-rel-transformed cell lines, a variant 3.1-kb Rb-specific mRNA was detected in addition to the 4.7-kb transcript. ChRb is expressed as a 104 kilodalton protein and is therefore slightly smaller than the mammalian Rb proteins.

Alternative splicing of Pax-8 gene transcripts is developmentally regulated and generates isoforms with different transactivation properties.

Pax-8, a member of the paired box-containing gene family, was shown to be coexpressed with Pax-2 in several human kidney carcinoma cell lines. Four different Pax-8 mRNA isoforms, a to d, were cloned from one of these cell lines by polymerase chain reaction amplification, and the Pax-8 gene was isolated from a human cosmid library. Analysis of the exon-intron structure of Pax-8 revealed that the four mRNA isoforms arise by alternative splicing, resulting in inclusion or exclusion of exon 7 and/or exon 8 sequences. All four Pax-8 proteins retain the paired domain as their DNA-binding motif and recognize DNA in the same manner as do the closely related Pax-2 and BSAP (Pax-5) proteins. The Pax-8a and Pax-8b isoforms end in a serine/threonine/tyrosine-rich sequence, while the C terminus of Pax-8c and Pax-8d is translated in a different, proline-rich reading frame. Transient transfection experiments revealed that Pax-8 isoforms a and b, but not c and d, strongly stimulate transcription from a promoter containing six copies of a paired-domain recognition sequence. The same four mRNA variants were also detected by RNase protection analysis in the mouse embryo and adult kidney, thus indicating evolutionary conservation of Pax-8 mRNA splicing. A different splice pattern was observed in the developing placenta, which expresses two new variants, Pax-8e and Pax-8f, instead of transcripts b to d. Expression of these mRNAs is high at embryonic day 9.5 and is gradually reduced until Pax-8a is the predominant transcript in the 12.5-day placenta. In the embryo, however, the synthesis of mRNAs b to d is initially low and then increases relative to that of Pax-8a. Hence, alternative splicing of Pax-8 gene transcripts not only generates six different Pax-8 variants but is also temporally and spatially regulated during early mouse development.

Mutually exclusive expression of a helix-loop-helix gene and N-myc in human neuroblastomas and in normal development.

We have isolated a novel human gene encoding a helix-loop-helix (HLH) protein by molecularly cloning chromosome 1p36-specific CpG islands. The gene termed heir-1 was localized to the neuroblastoma consensus deletion at 1p36.2-p36.12. Its predicted protein is 95.8% identical to the mouse HLH462 protein and has clear homology to the mouse Id and Drosophila emc proteins. Heir-1 does not encode a basic DNA binding domain as found in basic HLH proteins. The gene is expressed specifically at high abundance in adult lung, kidney and adrenal medulla, but not in adult brain. Despite prominent heir-1 expression in adrenal medulla, which is a prime target for neuroblastomas, 10 out of 12 neuroblastoma-derived cell lines revealed very low levels of heir-1 mRNA. Low heir-1 expression was generally found in tumor cell lines with N-myc overexpression, whereas the two cell lines displaying high heir-1 levels did not overexpress N-myc. Mutually exclusive expression of both genes was also found by in situ hybridization in developing mouse tissues, particularly in the forebrain neuroectoderm. We conclude that heir-1 expression is reduced specifically in the majority of neuroblastomas and suggest an inverse correlation between heir-1 and N-myc expression in neuroblastoma tumors and in embryonic development.