Toxicity of the bacterial luciferase substrate, n-decyl aldehyde, to Saccharomyces cerevisiae and Caenorhabditis elegans.

This study determined that the bacterial luciferase fusion gene (luxAB) was not a suitable in vivo gene reporter in the model eukaryotic organisms Saccharomyces cerevisiae and Caenorhabditis elegans. LuxAB expressing S. cerevisiae strains displayed distinctive rapid decays in luminescence upon addition of the bacterial luciferase substrate, n-decyl aldehyde, suggesting a toxic response. Growth studies and toxicity bioassays have subsequently confirmed, that the aldehyde substrate was toxic to both organisms at concentrations well tolerated by Escherichia coli. As the addition of aldehyde is an integral part of the bacterial luciferase activity assay, our results do not support the use of lux reporter genes for in vivo analyses in these model eukaryotic organisms.

Development and application of bioluminescent Caenorhabditis elegans as multicellular eukaryotic biosensors.

We describe a novel approach to assess toxicity to the free-living nematode Caenorhabditis elegans that relies on the ability of firefly luciferase to report on endogenous ATP levels. We have constructed bioluminescent C. elegans with the luc gene under control of a constitutive promoter. Light reduction was observed in response to increasing temperature, concentrations of copper, lead and 3,5-dichlorophenol. This was due to increased mortality coupled with decreased metabolic activity in the surviving animals. The light emitted by the transgenic nematodes gave a rapid, real-time indication of metabolic status. This forms the basis of rapid and biologically relevant toxicity tests.

Cadherin superfamily proteins in Caenorhabditis elegans and Drosophila melanogaster.

The ability to form selective cell-cell adhesions is an essential property of metazoan cells. Members of the cadherin superfamily are important regulators of this process in both vertebrates and invertebrates. With the advent of genome sequencing projects, determination of the full repertoire of cadherins available to an organism is possible and here we present the identification and analysis of the cadherin repertoires in the genomes of Caenorhabditis elegans and Drosophila melanogaster. Hidden Markov models of cadherin domains were matched to the protein sequences obtained from the translation of the predicted gene sequences. Matches were made to 21 C. elegans and 18 D. melanogaster sequences. Experimental and theoretical work on C. elegans sequences, and data from ESTs, show that three pairs of genes, and two triplets, should be merged to form five single genes. It also produced sequence changes at one or both of the 5' and 3' termini of half the sequences. In D. melanogaster it is probable that two of the cadherin genes should also be merged together and that three cadherin genes should be merged with other neighbouring genes. Of the 15 cadherin proteins found in C. elegans, 13 have the features of cell surface proteins, signal sequences and transmembrane helices; the other two have only signal sequences. Of the 17 in D. melanogaster, 11 at present have both features and another five have transmembrane helices. The evidence currently available suggests about one-third of the cadherins in the two organisms can be grouped into subfamilies in which all, or parts of, the molecules are conserved. Each organism also has a approximately 980 residue protein (CDH-11 and CG11059) with two cadherin domains and whose sequences match well over their entire length two proteins from human brain. Two proteins in C. elegans, HMR-1A and HMR-1B, and three in D. melanogaster, CadN, Shg and CG7527, have cytoplasmic domains homologous to those of the classical cadherin genes of chordates but their extracellular regions have different domain structures. Other common subclasses include the seven-helix membrane cadherins, Fat-like protocadherins and the Ret-like cadherins. At present, the remaining cadherins have no obvious similarities in their extracellular domain architecture or homologies to their cytoplasmic domains and may, therefore, represent species-specific or phylum-specific molecules.

Immunization with gastric H+/K(+)-ATPase induces a reversible autoimmune gastritis.

The gastric H+/K(+)-ATPase has been implicated as a major autoantigen in pernicious anaemia in humans and in thymectomy-induced autoimmune gastritis in mice. Here we have shown that autoimmune gastritis can be generated by direct immunization of non-thymectomized BALB/c mice with mouse gastric H+/K(+)-ATPase in complete Freund's adjuvant. The gastritis was characterized by infiltration of the gastric submucosa and mucosa with macrophages, CD4+ and CD8+ T cells, and B cells and by circulating autoantibodies to the H+/K(+)-ATPase. The mononuclear infiltrate within the gastric mucosa was accompanied by loss of parietal and zymogenic cells and accumulation of small immature epithelial cells. Splenocytes from gastritic mice adoptively transferred gastritis to naive recipients. Cessation of immunization resulted in decrease in autoantibody titre and regeneration of parietal and zymogenic cells. The results directly confirm that the gastric H+/K(+)-ATPase is the causative autoantigen in the genesis of autoimmune gastritis. Recovery of the lesion following cessation of immunization suggests that homeostatic mechanisms can reverse a destructive autoimmune process.

cdh-3, a gene encoding a member of the cadherin superfamily, functions in epithelial cell morphogenesis in Caenorhabditis elegans.

Several genes that encode members of the cadherin superfamily have been identified in Caenorhabditis elegans. Based on the roles of cadherins in vertebrates and Drosophila, it is expected that they function in the control of epithelial morphogenesis, an event which is poorly understood at the molecular level in C. elegans. Reporter genes under the control of upstream sequences from one of these genes, cdh-3, are expressed in developing epithelial cells, but also in a number of neuroectodermal cells that extend processes along some of these epithelial cells. We generated a loss-of-function mutation in cdh-3 by transposon-mediated deletion mutagenesis. This mutation affects the morphogenesis of a single cell, hyp10, which forms the tip of the nematode tail. The lack of detectable defects associated with the other cells expressing cdh-3 reporter constructs hints at the existence of other genes that can compensate for cdh-3 loss of function.

Repression of gene expression in the embryonic germ lineage of C. elegans.

The distinction between soma and germline was recognized more than a century ago: somatic cells form the body of an organism, whereas germ cells serve to produce future generations. In Caenorhabditis elegans, the separation of some and germline occurs through a series of asymmetrical divisions, in which embryonic germline blastomeres divide unequally to produce one somatic daughter and one germline daughter. Here we show that after each asymmetrical division, embryonically transcribed RNAs are detected in somatic, but not germline, blastomeres. This asymmetry depends on the activity of the germline specific factor, PIE-1. In the absence of PIE-1, embryonically transcribed RNAs are detected in both somatic and germline blastomeres. Furthermore, ectopic expression of PIE-1 in somatic blastomeres can significantly reduce the accumulation of new transcripts in these cells. Taken together, these results suggest that germ-cell fate depends on an inhibitory mechanism that blocks new gene expression in the early embryonic germ lineage.

From coiled tubules to a secretory canaliculus: a new model for membrane transformation and acid secretion by gastric parietal cells.

The acid-secreting gastric parietal cell has a unique secretory membrane system. This membrane system exists in an inactive (non-secreting) and an active (secreting) form. The current accepted model to explain the transformation events associated with the conversion of the non-secreting membrane to the secreting membrane, and vice versa, invokes membrane recycling of elongated vesicle structures. However, recent studies employing cryopreparation have shown that the non-secreting membrane in these cells is actually a complex network of helically coiled tubules. Here, we present an alternative model to explain how the membrane in parietal cells is activated to secrete HCl.

Fast freeze-fixation/freeze-substitution reveals the secretory membranes of the gastric parietal cell as a network of helically coiled tubule. A new model for parietal cell transformation.

The parietal cell of the gastric mucosa undergoes rapid morphological transformation when it is stimulated to produce hydrochloric acid. In chemically fixed cells, this process is seen as a reduction in number of cytoplasmic 'tubulovesicles' as the apical surface of the cell progressively invaginates to increase the secretory surface area. It is widely believed that the tubulovesicles represent stored secretory membrane in the cytoplasm of the unstimulated cell, which is incorporated into the apical membrane upon stimulation, because they share H+,K+-ATPase activity with the apical membrane. However, fusion of tubulovesicles with the apical membrane concomitant with parietal cell activation has never been convincingly demonstrated. We have used fast freeze-fixation and freeze-substitution to study stages of morphological transformation in these cells. Tubulovesicles were not seen in the cytoplasm of any of our cryoprepared cells. Instead, the cytoplasm of the unstimulated cell contained numerous and densely packed helical coils of tubule, each having an axial core of cytoplasm. The helical coils were linked together by connecting tubules, lengths of relatively straight tubule. Lengths of straight connecting tubule also extended from coils lying adjacent to the apical and canalicular surfaces and ended at the apical and canaliculus membranes. Immunogold labelling with alpha- and beta-subunit-specific antibodies showed that the gastric H+,K+-ATPase was localized to the membranes of this tubular system, which therefore represented the configuration of the secretory membrane in the cytoplasm of the unstimulated parietal cell. Stimulation of the cells with histamine and isobutylmethylxanthine lead to modification of the tubular membrane system, correlated with progressive invagination of the apical membrane. The volume of the tubule lumen increased and, as this occurred, the tight spiral twist of the helical coils was lost, indicating that tubule distension was accounted for by partial unwinding. This exposed the cores of cytoplasm in the axes of the coils as rod-shaped elements of a three-dimensional reticulum, resembling a series of microvilli in random thin sections. Conversely, treatment with the H2 antagonist cimetidine caused severe contraction of the tubular membrane system and intracellular canaliculi. Our results indicate that tubulovesicles are an artifact of chemical fixation; consequently, they cannot have a role in parietal cell transformation. From our findings we propose an alternative model for morphological transformation in the parietal cell. This model predicts cytoskeleton-mediated control over expansion and contraction of the tubular membrane network revealed by cryopreparation. The model is compatible with the localization of cytoskeletal components in these cells.

Medial-Golgi retention of N-acetylglucosaminyltransferase I. Contribution from all domains of the enzyme.

We have previously shown that the transmembrane domain and flanking residues of beta-1,2-N-acetylglucosaminyltransferase I (GnTI) can localize a hybrid molecule to medial-Golgi cisternae (Burke, J., Pettitt, J. M., Schachter, H., Sarkar, M., and Gleeson, P.A. (1992) J. Biol. Chem. 267, 24433-24440). Here, we have further examined the contribution of the cytoplasmic tail, transmembrane domain, and the catalytically active, luminal domain of GnTI in medial-Golgi localization, by analyzing the localization of hybrid molecules stably expressed in murine cells. In contrast to wild-type GnTI, which was efficiently localized to the medial-Golgi and not detected at the cell surface, hybrid molecules containing any two of the three domains of GnTI were localized to the medial-Golgi and were also present at low levels at the cell surface. Hybrid molecules containing only the transmembrane domain or the luminal domain of GnTI showed partial Golgi retention together with an increased level of cell surface expression compared with molecules containing two GnTI domains. The cytoplasmic tail independently was unable to retain reporter sequences to the Golgi but increased the ability of constructs containing either the luminal or transmembrane domain of GnTI to localize to the Golgi apparatus. Therefore, all three domains of GnTI contribute significantly to medial-Golgi localization. Furthermore, GnTI hybrid molecules showing increased cell surface expression were more readily extracted in a low salt buffer, suggesting that Golgi localized GnTI differs in physicochemical properties from cell surface GnTI. Based on an aggregation model of localization, we propose that Golgi retention of these hybrid molecules is mediated by the interaction of their GnTI domains with the corresponding domains of endogenous glycosyltransferase aggregates within the Golgi membranes of the transfected cell.

Developmentally regulated alternative splicing of a nematode type IV collagen gene.

A comparison of the genomic DNA sequence that encodes the Ascaris suum alpha 2(IV) collagen chain with the corresponding cDNA sequence led to the identification of a putative exon that was not expressed in the cDNA. The identification of this putative exon raised the possibility that transcripts of the alpha 2(IV) gene may undergo alternative splicing. We have used a reverse transcriptase-polymerase chain reaction assay to establish that such alternative splicing does indeed occur. Our results show that the A. suum alpha 2(IV) collagen gene produces at least two similar, but not identical, transcripts via the selection of two alternative exons. Furthermore, this alternative splicing appears to be developmentally regulated, suggesting that alternative splicing may be used in order to modify the properties of type IV collagen during nematode development.

Gastric parietal cell development: expression of the H+/K+ ATPase subunits coincides with the biogenesis of the secretory membranes.

The early development of the parietal cell in the embryonic murine gastric mucosa was investigated with particular attention paid to the biogenesis of the secretory membranes and the localization of the gastric H+/K+ ATPase alpha and beta subunits. Gastric glands were recognized in the day 18 foetus. However, at this stage in development no parietal cells could be distinguished ultrastructurally in the glands, and no immunoreactivity was detected with monoclonal antibodies to either the alpha or beta subunits of the gastric H+/K+ ATPase. In the 19 day embryo, parietal cells were recognizable morphologically by the presence of slender microvilli on the apical (lumenal) surface and differentiating intracellular canaliculi in the apical cytoplasm. Both subunits of the proton pump were found to be specifically associated with the apical and canalicular membranes and with the membranes of relatively large vesicles distributed in the subapical cytoplasm and the cytoplasm surrounding the canaliculi. In the parietal cells of the day 1 neonate, the intracellular canaliculi had extended basally to form the extensive compartments typical of parietal cells in the adult animal. Again, profiles of vesicles showing H+/K+ ATPase immunoreactivity were present in the pericanalicular cytoplasm. These results indicate that the intracellular canaliculi are formed by expansion of the apical surface and suggest that the delivery of newly synthesized gastric H+/K+ ATPase alpha and beta subunits to the apical plasma membrane is mediated by typical Golgi transport vesicles. The large immunoreactive vesicles that occur in the apical and pericanalicular cytoplasm of the developing cells may represent artefacts generated by fixation-induced fragmentation of the differentiating canalicular membrane system during preparation.

The transmembrane and flanking sequences of beta 1,2-N-acetylglucosaminyltransferase I specify medial-Golgi localization.

UDP-GlcNAc:alpha 3-D-mannoside beta 1,2-N-acetylglucosaminyltransferase I (GnTI) is an N(in)/C(out) (type II) membrane protein, localized in the medial-Golgi, that initiates the conversion of high mannose N-glycans to complex N-glycans. Anti-rabbit GnTI antibodies were generated using a purified, enzymatically active, bacterial recombinant fusion protein as immunogen. Rabbit GnTI was effectively retained in the Golgi complex of transfected COS-1 cells and murine L cells, as assessed by indirect immunofluorescence using the species-specific anti-GnTI antibodies; no surface expression of rabbit GnTI could be detected in the transfected cells. Rabbit GnTI, stably expressed in murine L cells, was localized by immunoperoxidase electron microscopy to the medial-cisternae of the Golgi stack. The role of the transmembrane domain of GnTI in Golgi localization was examined by generation of a hybrid construct containing the amino-terminal 31 amino acids of GnTI, corresponding to the 25-residue transmembrane (signal/anchor) domain and flanking hydrophilic sequences, fused with ovalbumin; this ovalbumin/GnTI hybrid molecule was retained in the Golgi complex of transfected COS cells and stably transfected murine L cells. No surface expression of ovalbumin/GnTI was detected. In contrast, ovalbumin fused to the equivalent domains of the human transferrin receptor, a type II cell-surface protein, was efficiently expressed on the cell surface of transfected cells. The ovalbumin/GnTI hybrid molecules in the transfected L cells were N-glycosylated, indicating an N(in)/C(out) membrane orientation, and were localized by immunoperoxidase electron microscopy to one or two cisternae of the medial-Golgi (90% of stained Golgi profiles showed medial-cisternae staining). These results show that a signal contained within the transmembrane domain and flanking residues of GnTI specifies medial-Golgi localization.

Human autoantibodies as reagents to conserved Golgi components. Characterization of a peripheral, 230-kDa compartment-specific Golgi protein.

We have used a serum from a patient with Sjögren's syndrome containing high titer (100,000) anti-Golgi autoantibodies and lower titer (20,000) anti-nuclear autoantibodies to characterize the Golgi complex. The Sjögren's syndrome serum immunoprecipitated a number of components of molecular mass 35-230 kDa from detergent extracts of [35S]methionine-labeled HeLa cells; at high dilution, the serum precipitated one major 230-kDa component. Using the Sjögren's syndrome serum, cDNA clones encoding the Golgi autoantigen were isolated from a lambda gt11 HeLa cell cDNA library. Autoantibodies from the Sjögren's syndrome serum, affinity purified from a recombinant bacterial fusion protein generated from one of the cDNA clones, showed Golgi staining of human, mouse, and chicken cells by immunofluorescence. The purified autoantibodies immunoprecipitated and immunoblotted a 230-kDa component. A rabbit antiserum raised to the recombinant fusion protein specifically stained the Golgi complex by immunofluorescence and reacted with a 230-kDa protein by immunoprecipitation and immunoblotting. The 230-kDa protein was recovered in both the 100,000 x g sedimentable and soluble fractions in cell lysates and in the aqueous phase of Triton X-114 extracts. The 230-kDa autoantigen was dissociated from the Golgi complex by 15-min brefeldin A treatment, dissociation kinetics similar to that of mannosidase II. However, unlike mannosidase II, autoantigen staining was markedly reduced after drug treatment. Removal of brefeldin A resulted in reassociation of the autoantigen with the Golgi complex. The epitopes recognized by the affinity purified human and rabbit antibodies were ultrastructurally localized to the cytosolic face of one side of the Golgi stack, probably the trans-face. Taken together, the 230-kDa protein is a conserved, peripheral membrane component specifically associated with one Golgi compartment. We suggest that this peripheral Golgi protein may have a role in the compartment-specific structural organization of the Golgi or in sorting and transport of proteins.

The pogo transposable element family of Drosophila melanogaster.

A 190 bp insertion is associated with the white-eosin mutation in Drosophila melanogaster. This insertion is a member of a family of transposable elements, pogo elements, which is of the same class as the P and hobo elements of D. melanogaster. Strains typically have many copies of a 190 bp element, 10-15 elements 1.1-1.5 kb in size and several copies of a 2.1 kb element. The smaller elements all appear to be derived from the largest by single internal deletions so that all elements share terminal sequences. They either always insert at the dinucleotide TA and have perfect 21 bp terminal inverse repeats, or have 22 bp inverse repeats and produce no duplication upon insertion. Analysis by DNA blotting of their distribution and occupancy of insertion sites in different strains suggests that they may be less mobile than P or hobo. The DNA sequence of the largest element has two long open reading frames on one strand which are joined by splicing as indicated by cDNA analysis. RNAs of this strand are made, whose sizes are similar to the major size classes of elements. A protein predicted by the DNA sequence has significant homology with a human centrosomal-associated protein, CENP-B. Homologous sequences were not detected in other Drosophila species, suggesting that this transposable element family may be restricted to D. melanogaster.

Primary sequence heterogeneity and tissue expression of glutathione S-transferases of Fasciola hepatica.

Glutathione S-transferases (GSTs) from Fasciola hepatica have been purified by glutathione affinity chromatography. Two closely migrating species of Mr 26,000 and 26,500 were identified by sodium dodecyl sulfate-polyacrylamide gel electrophoresis, and several species resolved by two-dimensional gel analysis, indicating substantial heterogeneity among the GSTs. N-terminal amino acid sequencing revealed one core sequence containing three polymorphisms, whereas the sequence of GST peptides implied a minimum of three different GSTs. The amino acid sequence data assigned the F. hepatica GSTs to the mu class of GSTs with high similarities to these proteins in other helminths and mammals. The native GSTs of F. hepatica appeared to behave as dimers as determined by molecular sieving chromatography. The observation that the GSTs of F. hepatica are heterogeneous in sequence and behave as dimers in the native state suggest that these isoenzymes may exhibit considerable functional heterogeneity which may be of importance to the parasite. Immunocytochemical studies suggest that the main source of GST in F. hepatica are the parenchymal cells and peripheral tissues of the parasite. Some extracellular GST is associated with the lamellae of the intestinal epithelium. The identification of an intestinal GST is unique among trematodes studied to date.

Isolation and developmental expression of Bcp1, an anther-specific cDNA clone in Brassica campestris.

Differential screening of a mature Brassica campestris pollen cDNA library has identified five cDNA clones that represent transcripts expressed exclusively, or at elevated levels, in pollen. We show here that the expression of one of these, clone Bcp1, is tissue specific and temporally regulated. The gene is activated during microspore development, as detected by in situ hybridization. Expression is enhanced at the time of pollen maturation and during pollen germination. In situ hybridization has also shown that Bcp 1 is activated in the tapetal cells in early anther development and continues to be expressed until tapetal dissolution. Homologous transcripts are present in pollen of other taxa of Brassicaceae including Arabidopsis, but not in pollen of any other families tested.

The complete primary structure of a nematode alpha 2(IV) collagen and the partial structural organization of its gene.

We have isolated and characterized cDNA and genomic DNA clones which encode an alpha 2(IV) collagen chain from the parasitic nematode Ascaris suum. In addition we have determined, by nucleic acid sequence analysis, the structural organization of approximately two-thirds of the gene. This analysis has shown that the gene contains at least 15 introns, and those that have been characterized range in size from 141 to 854 base pairs. The derived protein sequence contains 1763 amino acids and includes a putative 26-amino acid signal sequence. The collagenous triple-helical region contains 17 interruptions, many of which occur in the same positions as those in the human alpha 1(IV) and alpha 2(IV) chains. Comparison of the genomic DNA sequence with the cDNA sequence has revealed the presence of a sequence within the gene which appears to be an intact and normal exon that is not represented in our cDNA sequence. The presence of this putative exon raises the possibility that the A. suum alpha 2(IV) collagen gene may undergo alternative splicing.

Monoclonal antibodies specific for the core protein of the beta-subunit of the gastric proton pump (H+/K+ ATPase). An autoantigen targetted in pernicious anaemia.

The gastric H+/K(+)-transporting adenosine triphosphatase (H+/K+ ATPase) (proton pump) consists of a catalytic alpha-subunit and a recently proposed 60-90-kDa glycoprotein beta-subunit. Using dog gastric membranes as the antigen, we have produced two murine monoclonal antibodies, 4F11 (IgG1) and 3A6 (IgA), which are specific for the 60-90-kDa glycoprotein. The monoclonal antibodies (1) specifically stained the cytoplasm of unfixed and formalin-fixed dog gastric parietal cells; (2) specifically reacted by ELISA with gastric tubulovesicular membranes; (3) recognised epitopes located on the luminal face of parietal cell tubulovesicular membranes, the site of the proton pump, by immunogold electron microscopy; (4) immunoblotted a 60-90-kDa molecule from tubulovesicular membranes and a 35-kDa component from peptide N-glycosidase-F-treated membrane extracts; (5) immunoblotted the 60-90-kDa parietal cell autoantigen associated with autoimmune gastritis and pernicious anemia, purified by chromatography on parietal cell autoantibody- or tomato-lectin-Sepharose 4B affinity columns, and the 35-kDa protein core of this autoantigen; this autoantigen has amino acid sequence similarity to the beta-subunit of the related Na+/K(+)-transporting adenosine triphosphatase (Na+/K+ ATPase) [Toh et al. (1990) Proc. Natl Acad. Sci. 87, 6418-6422]; (6) co-precipitated a molecule of 95 kDa with the 60-90-kDa molecule from 125I-labelled detergent extracts of dog tubulovesicular membranes; and (7) co-purified the catalytic alpha-subunit of the H+/K+ ATPase with the 60-90-kDa molecule by immunoaffinity chromatography of tubulovesicular membrane extracts on a monoclonal antibody 3A6-Sepharose 4B column, indicating a physical association between the two molecules. These results provide further evidence that the 60-90-kDa glycoprotein is the beta-subunit of the gastric H+/K+ ATPase. We conclude that the monoclonal antibodies specifically recognise luminal epitopes on the 35-kDa core protein of the 60-90-kDa beta-subunit of the gastric proton pump, a major target molecule in autoimmune gastritis and pernicious anaemia. These monoclonal antibodies will be valuable probes to study the structure and function of this associated beta-subunit, as well as the ontogeny of the gastric proton pump.

Double lectin and immunolabelling for transmission electron microscopy: pre- and post-embedding application using the biotin-streptavidin system and colloidal gold-silver staining.

Pre- and post-embedding methods are described that can be used for consecutive localization of two intracellular cytoplasmic binding sites in cells and tissues embedded in acrylic plastic for transmission electron microscopy. Both applications make use of the biotin-streptavidin system with colloidal gold detector particles and involve silver staining of the first gold signal to a predetermined size. Silver augmentation effectively masked any free binding sites on the biotinylated molecule and on the streptavidin complex of the first labelling reaction, thereby allowing a second cycle with the same detection system. Excellent ultrastructural localization was obtained with silver lactate as the silver ion donor in the developing solution, and the enhancement treatment did not destroy or even visibly reduce target site reactivity for the subsequently applied probe. Using these methods it was possible to achieve specific double lectin and immunological labelling; they could, however, be adapted to dual or multiple-labelling procedures with any biotinylated molecules.