Search

Heparin-binding growth-associated molecule (HB-GAM) (pleiotrophin) is a highly conserved extracellular matrix-associated protein implicated in a diverse range of developmental processes, including the formation and plasticity of neuronal connections. Using gene targeting, we have in the present study created HB-GAM-deficient mice that are viable and fertile and show no gross anatomical abnormalities. The hippocampal structure as well as basal excitatory synaptic transmission in the area CA1 appear normal in the mice lacking HB-GAM. However, hippocampal slices from HB-GAM-deficient mice display a lowered threshold for induction of long-term potentiation (LTP), which reverts back to the wild-type level by application of HB-GAM. HB-GAM expression in hippocampus is activity-dependent and upregulated in several neuropathological conditions. Thus, we suggest that HB-GAM acts as an inducible signal to inhibit LTP in hippocampus., (Copyright 2001 Academic Press.)

Proteins encoded by the fringe family of genes are required to modulate Notch signalling in a wide range of developmental contexts. Using a cell co-culture assay, we find that mammalian Lunatic fringe (Lfng) inhibits Jagged1-mediated signalling and potentiates Delta1-mediated signalling through Notch1. Lfng localizes to the Golgi, and Lfng-dependent modulation of Notch signalling requires both expression of Lfng in the Notch-responsive cell and the Notch extracellular domain. Lfng does not prevent binding of soluble Jagged1 or Delta1 to Notch1-expressing cells. Lfng potentiates both Jagged1- and Delta1-mediated signalling via Notch2, in contrast to its actions with Notch1. Our data suggest that Fringe-dependent differential modulation of the interaction of Delta/Serrate/Lag2 (DSL) ligands with their Notch receptors is likely to have a significant role in the combinatorial repertoire of Notch signalling in mammals.

Notch receptors function in highly conserved intercellular signalling pathways that direct cell-fate decisions, proliferation and apoptosis in metazoans. Fringe proteins can positively and negatively modulate the ability of Notch ligands to activate the Notch receptor. Here we establish the biochemical mechanism of Fringe action. Drosophila and mammalian Fringe proteins possess a fucose-specific beta1,3 N-acetylglucosaminyltransferase activity that initiates elongation of O-linked fucose residues attached to epidermal growth factor-like sequence repeats of Notch. We obtained biological evidence that Fringe-dependent elongation of O-linked fucose on Notch modulates Notch signalling by using co-culture assays in mammalian cells and by expression of an enzymatically inactive Fringe mutant in Drosophila. The post-translational modification of Notch by Fringe represents a striking example of modulation of a signalling event by differential receptor glycosylation and identifies a mechanism that is likely to be relevant to other signalling pathways.

The three members of the mammalian fringe gene family, Manic fringe (Mfng), Radical fringe (Rfng), and Lunatic fringe (Lfng), were identified on the basis of their similarity to Drosophila fringe (fng) and their participation in the evolutionarily conserved Notch receptor signaling pathway. Fringe genes encode pioneer secretory proteins with weak similarity to glycosyltransferases. Both expression patterns and functional studies support an important role for Fringe genes in patterning during embryonic development and an association with cellular transformation. We have now further characterized the expression and determined the chromosomal localization and genomic structure of the mouse Mfng, Rfng, and Lfng genes; the genomic structure and conceptual open reading frame of the human RFNG gene; and the refined chromosomal localization of the three human fringe genes. The mouse Fringe genes are expressed in the embryo and in adult tissues. The mouse and human Fringe family members map to three different chromosomes in regions of conserved synteny: Mfng maps to mouse Chr 15, and MFNG maps to human Chr 22q13.1 in the region of two cancer-associated loci; Lfng maps to mouse Chr 5, and LFNG maps to human Chr 7p22; Rfng maps to mouse Chr 11, and RFNG maps to human Chr 17q25 in the minimal region for a familial psoriasis susceptibility locus. Characterization of the genomic loci of the Fringe gene family members reveals a conserved genomic organization of 8 exons. Comparative analysis of mammalian Fringe genomic organization suggests that the first exon is evolutionarily labile and that the Fringe genes have a genomic structure distinct from those of previously characterized glycosyltransferases.

In both Drosophila wings and vertebrate limbs, signaling between dorsal and ventral cells establishes an organizer that promotes limb formation. Significant progress has been made recently towards characterizing the signaling interactions that occur at the dorsal-ventral limb border. Studies of chicks have indicated that, as in Drosophila, this signaling process requires the participation of Fringe. Studies of Drosophila have indicated that Fringe functions by inhibiting the ability of Notch to be activated by one ligand, Serrate, while potentiating the ability of Notch to be activated by another ligand, Delta. Recent studies of both Drosophila and vertebrates have also shed new light on the signaling activity of the dorsal-ventral boundary limb organizer, and have highlighted how this organizer is maintained by feedback mechanisms with neighboring cells.

The mouse Ulnaless locus is a semidominant mutation which displays defects in patterning along the proximal-distal and anterior-posterior axes of all four limbs. The first Ulnaless homozygotes have been generated, and they display a similar, though slightly more severe, limb phenotype than the heterozygotes. To create a refined genetic map of the Ulnaless region using molecular markers, four backcrosses segregating Ulnaless were established. A 0.4-cM interval containing the Ulnaless locus has been defined on mouse chromosome 2, which has identified Ulnaless as a possible allele of a Hoxd cluster gene(s). With this genetic map as a framework, a physical map of the Ulnaless region has been completed. Yeast artificial chromosomes covering this region have been isolated and ordered into a 2 Mb contig. Therefore, the region that must contain the Ulnaless locus has been defined and cloned, which will be invaluable for the identification of the molecular nature of the Ulnaless mutation.

The study of limb development has provided insight into pattern formation during vertebrate embryogenesis. Genetic approaches offer powerful ways to identify the critical molecules and their pathways of action required to execute a complex morphogenetic program. We have applied genetic analysis to the process of limb development by studying two mouse mutants, limb deformity (ld) and Strong's luxoid (lst). These mutations confer contrasting phenotypic alterations to the anteroposterior limb pattern. The six mutant ld alleles are fully recessive and result in oligosyndactyly of all four limbs. By contrast, the two mutant lst alleles result in a mirror-image polydactylous limb phenotype inherited in a semidominant fashion. Morphological and molecular analysis of embryonic limbs has shown that the ld and lst alleles affect the extent and distribution of two key signaling centers differentially: the apical ectodermal ridge and the zone of polarizing activity. Molecular characterization of the ld gene has defined a new family of evolutionarily conserved proteins termed the formins. The underlying molecular defect in the lst mutation has not been identified; however, both loci are tightly linked on mouse chromosome 2, suggesting the possibility that they may be allelic. In this study, we have used genetic analysis to examine the epistatic and allelic relationships of ld and lst. We observed that in + ld/lst + double heterozygotes, a single mutant ld allele is able to suppress the semi-dominant polydactylous lst limb phenotype. By segregating the lst and ld loci in a backcross, we observed that these loci recombine and are separated by a genetic distance of approximately 6 cM. Therefore, while our observations demonstrate a genetic interaction between ld and lst, it is probable that ld and lst are not allelic. Instead, lst and ld may be operating either in a linear or in a parallel (bypass) genetic pathway to affect the limb signaling centers.

Midkine (Mdk) and heparin-binding neurotrophic factor (Hbnf)/pleiotrophin (Ptn) comprise the Midkine family of developmentally regulated signaling molecules. We have determined the chromosomal localization of these genes in the mouse by use of single-strand conformation polymorphisms (SSCPs), which facilitated the typing of Mdk and Hbnf alleles in recombinant inbred (RI) strains and interspecific backcrosses. Mapping was performed relative to other cloned genes, as well as simple sequence length polymorphisms (SSLPs) in the interspecific backcrosses. Mdk maps to mouse Chromosome (Chr) 2, linked to the Hoxd gene cluster. Hbnf maps to proximal mouse Chr 6, linked to the Cftr and Cpa genes. Comparative mapping of human MDK and HBNF employing species-specific polymerase chain reaction (PCR) primers and human monochromosomal somatic cell hybrids assigns MDK to human Chr 11 and HBNF to human Chr 7q32-qter.

The murine myeloproliferative leukemia virus has previously been shown to contain a fragment of the coding region of the c-mpl gene, a member of the cytokine receptor superfamily. We have isolated cDNA and genomic clones encoding murine c-mpl and localized the c-mpl gene to mouse chromosome 4. Since some members of this superfamily function by transducing a proliferative signal and since the putative ligand of mpl is unknown, we have generated a chimeric receptor to test the functional potential of mpl. The chimera consists of the extracellular domain of the human interleukin-4 receptor and the cytoplasmic domain of mpl. A mouse hematopoietic cell line transfected with this construct proliferates in response to human interleukin-4, thereby demonstrating that the cytoplasmic domain of mpl contains all elements necessary to transmit a growth stimulatory signal. In addition, we show that 25-40% of mpl mRNA found in the spleen corresponds to a novel truncated and potentially soluble isoform of mpl and that both full-length and truncated forms of mpl protein can be immunoprecipitated from lysates of transfected COS cells. Interestingly, however, although the truncated form of the receptor possesses a functional signal sequence and lacks a transmembrane domain, it is not detected in the culture media of transfected cells.

The DNA sequence of the sulfate activation locus from Escherichia coli K-12 has been determined. The sequence includes the structural genes encoding the enzymes ATP sulfurylase (cysD and cysN) and APS kinase (cysC) which catalyze the synthesis of activated sulfate. These are the only genes known to reside in the sulfate activation operon. Consensus elements of the operon promoter were identified, and the start codons and open reading frames of the Cys polypeptides were determined. During this work, another gene, iap, was partially sequenced and mapped. The activity of ATP sulfurylase is stimulated by an intrinsic GTPase. Comparison of the primary sequences of CysN and Ef-Tu revealed that CysN has conserved many of the residues integral to the three-dimensional structure important for guanine nucleotide binding in Ef-Tu and RAS. nodP and nodQ, from Rhizobium meliloti, are essential for nodulation in leguminous plants. The Cys and Nod proteins are remarkably similar. NodP appears to be the smaller subunit of ATP sulfurylase. NodQ encodes homologues of both CysN and CysC; thus, these enzymes may be covalently associated in R. meliloti. The consensus GTP-binding sequences of NodQ and CysN are identical suggesting that NodQ encodes a regulatory GTPase.

Mutations of the mouse limb deformity locus, ld, map to Chromosome (Chr) 2 and result in defects in the morphogenesis and patterning of the limb and kidney. Complementation studies have defined the existence of five recessive ld alleles. Remarkably, two of these, ldTgHd and ldTgBri, are transgene-induced mutations. Recovery of the first transgene insertional allele, ldTgHd, facilitated the molecular cloning of a large (greater than 200 kb) candidate gene at the ld locus. This gene is broadly transcribed and encodes a set of novel protein isoforms, termed formins. Here we present characterization of the ldTgBri mutation that supports the molecular identification of the ld gene. We show that the ldTgBri fails to complement both the ldTgHd and the ldOR alleles and that it has undergone a genomic deletion that disrupts the cloned ld gene and its transcripts. Curiously, the ldTgBri deletion encompasses the same 11-kb interval in which the ldTgHd insertion occurred and in which a chromosomal rearrangement has been identified in a third allele, ldIn2. These findings suggest that this region of the ld gene is a preferential site for illegitimate recombination.

Genomic imprinting is a regulatory process that requires a cell to recognize the parental origin of alleles. To understand how these alleles are distinguished, we have assessed changes in the DNA methylation of an imprinted transgene as it switches from one inheritance pattern to another while moving through gametogenesis and embryogenesis. We find that both maternally and paternally inherited methylation patterns are erased in primordial germ cells and that distinctive patterns emerge during germ cell maturation. In the case of the maternal allele, the methylation pattern is fully acquired during oogenesis. In the case of the paternal allele, the methylation pattern found in sperm undergoes further modification during embryogenesis. Thus, the distinction between "erased" maternal and paternal alleles is first established during their residence in different germ cells and then may be maintained by the recognition of the distinctive patterns that each allele displays in the zygote.

Vertebrate limb formation is an evolutionarily conserved process programmed by an array of morphogenetic genes. As a result of transgene insertion, we previously identified a mutation at the mouse limb deformity (ld) locus that disrupts embryonic pattern formation, resulting in a reduction and fusion of the distal bones and digits of all limbs as well as variable incidence of renal aplasia. We have now characterized the ld locus at the molecular level. It contains evolutionarily conserved coding sequences that are transcribed in adult and embryonic tissues as a complex group of low abundance messenger RNAs created by alternative splicing and differential polyadenylation. The association of these transcripts with the gene responsible for the mutant phenotype was established by demonstrating that they are disrupted in two independently arising ld alleles. We have now deduced the structure of several novel proteins (termed formins) from the long open reading frames encoded by the various ld transcripts. The observation of these different RNA transcripts in different tissues suggests that the formins play a part in the formation of several organ systems.

The recent identification of a gene residing at the mouse limb deformity (ld) locus permits us to test the hypothesis that disruption of this gene is responsible for an inherited anomaly affecting embryonic pattern formation. The gene gives rise to alternatively processed messenger RNAs that can be translated as a family of related protein products, termed the formins. We have now analysed transcripts from this gene in four independently isolated mutant alleles. In two of these, the ldHd allele (created by insertion of a transgene) and the ldIn2 allele (created by a translocation-inversion involving mouse chromosomes 2 and 17), a common subset of ld transcripts is abolished, but others are apparently unaltered. The correlation of altered transcripts in two independent ld mutants strongly supports the notion that one or more altered formins is responsible for the observed phenotype. That the defect is limited to the limb and kidney, despite expression of ld mRNA in other unaffected organs, suggests that these mutant alleles represent only partial loss of ld function.

Genetic analysis provides an approach for identifying regulatory loci that govern the expression of specific genes within the context of the entire organism. Such analyses have defined two unlinked regulatory loci, termed raf and Rif, that modulate the levels of alpha-fetoprotein in liver. Of primary importance for the isolation and characterization of the raf product is to determine whether it is produced by the hepatocyte or whether it is produced by a different cell type. By means of analysis of alpha-fetoprotein expression in livers of embryo aggregation chimeras derived from mice of different raf genotypes it was possible to conclude that the product of the raf locus is expressed as a hepatocyte autonomous function that acts in trans to regulate the level of alpha-fetoprotein messenger RNA.

Multiple cis-acting regulatory elements consisting of three cellular enhancers and a proximal promoter element have been identified in the region upstream of the mouse alpha-fetoprotein (AFP) gene. We examined the role of these sequences during differentiation by the introduction of modified AFP genes into cells at different stages of commitment to its expression. Modified AFP genes introduced stably into F9 embryonal carcinoma stem cells by DNA transfection were silent until activated by treatment with retinoic acid to form visceral endoderm. Their activation required the presence of both the enhancer and proximal promoter domains. The introduced genes activated simultaneously with the endogenous AFP genes, but reached maximal levels of expression more rapidly, suggesting a greater initial accessibility to transcription factors. In contrast, when modified AFP genes were stably introduced into HepG2 cells, a human hepatoma cell line that constitutively expresses the AFP gene, the proximal promoter sequences were sufficient to direct a low level of expression. The absolute requirement for the AFP enhancers in F9 cells but not in HepG2 cells supports a model by which there is an obligate requirement for enhancers during differentiation in addition to their role in enhancing gene expression after differentiation.

To identify cis-acting DNA elements involved in the activation of the alpha-fetoprotein gene during differentiation, modified copies of the gene were introduced into murine F9 embryonal carcinoma cells. The differentiation of the transformants to either parietal or visceral endoderm was accompanied by induction of the exogenous template in a manner qualitatively, but not quantitatively, identical to that of the endogenous alpha-fetoprotein gene.

Cytogenetic assay systems based on the detection of sister chromatid exchanges (SCE) are widely advocated as a sensitive screening method for assessing genotoxic potential. While many agents have been examined for their ability to induce SCE's, complete dose-response information has often been lacking. We have reexamined the ability of one such compound-caffeine-to induce SCEs and also to inhibit cellular proliferation in human peripheral lymphocytes in vitro. An acute exposure to caffeine prior to the DNA synthetic period did not affect either SCE frequency or the rate of cellular proliferation. Chronic exposure to caffeine throughout the culture period lead to both a dose-dependent increase in SCEs (SCEd or doubling dose = 2.4 mM; SCE10 or the dose capable of inducing 10 SCE = 1.4 mM) and a dose-dependent inhibition of cellular proliferation (IC50 or the 50% inhibition concentration = 2.6 mM). The relative proportion of first generation metaphase cels, an assessment of proliferative inhibition, increased linearly with increasing caffeine concentrations. However, SCE frequency increased nonlinearly over the same range of caffeine concentrations. Examination of the ratio of nonsymmetrical to symmetrical SCEs in third generation metaphase cells indicated that caffeine induced SCEs in equal frequency in each of three successive generations. The dependency of SCE induction and cellular proliferative inhibition on caffeine's presence during the DNa synthetic period suggests that caffeine may act as an antimetabolite in normal human cells. The significance of these results in regard to both caffeine's genotoxic potential and to the reliability of the SCE assay system are discussed.