The Rat Genome Database (RGD) facilitates genomic and phenotypic data integration across multiple species for biomedical research

Model organism research is essential for discovering the mechanisms of human diseases by defining biologically meaningful gene to disease relationships. The Rat Genome Database (RGD, ( https://rgd.mcw.edu )) is a cross-species knowledgebase and the premier online resource for rat genetic and physiologic data. This rich resource is enhanced by the inclusion and integration of comparative data for human and mouse, as well as other human disease models including chinchilla, dog, bonobo, pig, 13-lined ground squirrel, green monkey, and naked mole-rat. Functional information has been added to records via the assignment of annotations based on sequence similarity to human, rat, and mouse genes. RGD has also imported well-supported cross-species data from external resources. To enable use of these data, RGD has developed a robust infrastructure of standardized ontologies, data formats, and disease- and species-centric portals, complemented with a suite of innovative tools for discovery and analysis. Using examples of single-gene and polygenic human diseases, we illustrate how data from multiple species can help to identify or confirm a gene as involved in a disease and to identify model organisms that can be studied to understand the pathophysiology of a gene or pathway. The ultimate aim of this report is to demonstrate the utility of RGD not only as the core resource for the rat research community but also as a source of bioinformatic tools to support a wider audience, empowering the search for appropriate models for human afflictions.

Analysis of disease-associated objects at the Rat Genome Database

The Rat Genome Database (RGD) is the premier resource for genetic, genomic and phenotype data for the laboratory rat, Rattus norvegicus. In addition to organizing biological data from rats, the RGD team focuses on manual curation of gene–disease associations for rat, human and mouse. In this work, we have analyzed disease-associated strains, quantitative trait loci (QTL) and genes from rats. These disease objects form the basis for seven disease portals. Among disease portals, the cardiovascular disease and obesity/metabolic syndrome portals have the highest number of rat strains and QTL. These two portals share 398 rat QTL, and these shared QTL are highly concentrated on rat chromosomes 1 and 2. For disease-associated genes, we performed gene ontology (GO) enrichment analysis across portals using RatMine enrichment widgets. Fifteen GO terms, five from each GO aspect, were selected to profile enrichment patterns of each portal. Of the selected biological process (BP) terms, ‘regulation of programmed cell death’ was the top enriched term across all disease portals except in the obesity/metabolic syndrome portal where ‘lipid metabolic process’ was the most enriched term. ‘Cytosol’ and ‘nucleus’ were common cellular component (CC) annotations for disease genes, but only the cancer portal genes were highly enriched with ‘nucleus’ annotations. Similar enrichment patterns were observed in a parallel analysis using the DAVID functional annotation tool. The relationship between the preselected 15 GO terms and disease terms was examined reciprocally by retrieving rat genes annotated with these preselected terms. The individual GO term–annotated gene list showed enrichment in physiologically related diseases. For example, the ‘regulation of blood pressure’ genes were enriched with cardiovascular disease annotations, and the ‘lipid metabolic process’ genes with obesity annotations. Furthermore, we were able to enhance enrichment of neurological diseases by combining ‘G-protein coupled receptor binding’ annotated genes with ‘protein kinase binding’ annotated genes.

Database URL:http://rgd.mcw.edu

Specific targeting of EGP-2+ tumor cells by primary lymphocytes modified with chimeric T cell receptors

A promising strategy for cancer treatment is adoptive immunotherapy with gene-modified lymphocytes expressing a chimeric T cell receptor (cTCR) that directs tumor targeting and stimulates T cell effector functions. In this study, the activities of two novel cTCR molecules (GAgamma and GAHgamma) were investigated. Both encode a single-chain variable fragment (scFv) derived from the monoclonal antibody (MAb) GA733.2, which binds the epithelial glycoprotein 2 (EGP-2) overexpressed on a variety of human carcinomas. In the GAgamma cTCR, the scFv is directly fused to the transmembrane/cytoplasmic portions of the immunoglobulin Fc receptor (Ig FcRI) gamma subunit, which mediates T cell signaling. GAHgamma possesses an extracellular spacer composed of the CD8alpha immunoglobulin hingelike domain inserted between the scFv and gamma chain. Activated T cells (ATCs), stimulated ex vivo using anti-CD3 MAb, were derived from either healthy donors or patients and transduced with recombinant retrovirus encoding the respective GA cTCR molecules. After culture expansion for 14 days, GAgamma-modified ATCs demonstrated enhanced targeting and lysis of EGP-2+ colon cancer cells and increased cytokine secretion. Cells transduced with the GAHgamma cTCR displayed specific lytic activity that was about twofold greater than that of GAgamma-ATCs and produced significantly more cytokine. In addition, reactivation of GAHgamma-ATC with anti-CD3 MAb prior to addition to EGP-2+ tumor target induced a further increase in lytic activity. Because the activation status influences T cell antitumor functions, our data suggest that reactivation prior to adoptive transfer would improve the clinical efficacy of GAHgamma-modified ATCs.

Tissue factor activity is increased in human endothelial cells cultured under elevated static pressure

We tested the hypothesis that elevated blood pressure, a known stimulus for vascular remodeling and an independent risk factor for the development of atherosclerotic disease, can modulate basal and cytokine-induced tissue factor (TF; CD 142) expression in cultured human endothelial cells (EC). Using a chromogenic enzymatic assay, we measured basal and tumor necrosis factor-alpha (TNF-alpha; 10 ng/ml, 5 h)-induced TF activities in human aortic EC (HAEC) and vena cava EC (HVCEC) cultured at atmospheric pressure and at 170 mmHg imposed pressure for up to 48 h. Basal TF activities were 22 +/- 10 U/mg protein for HAEC and 14 +/- 9 U/mg protein for HVCEC and were upregulated in both cell types >10-fold by TNF-alpha. Exposure to pressure for 5 h induced additional elevation of basal TF activity by 47 +/- 16% (P < 0.05, n = 6) for HAEC and 17 +/- 5% (P < 0.05, n = 3) for HVCEC. Pressurization also enhanced TF activity in TNF-alpha-treated cells from 240 +/- 28 to 319 +/- 32 U/mg protein in HAEC (P < 0.05, n = 4) and from 148 +/- 25 to 179 +/- 0.8 U/mg protein (P < 0.05, n = 3) in HVCEC. Cytokine stimulation caused an approximately 100-fold increase in steady-state TF mRNA levels in HAEC, whereas pressurization did not alter either TF mRNA or cell surface antigen expression, as determined by quantitative RT-PCR methodology and ELISA. Elevated pressure, however, modulated the EC plasma membrane organization and/or permeability as inferred from the increased cellular uptake of the fluorescent amphipathic dye merocyanine 540 (33 +/- 7%, P < 0.05). Our data suggest that elevated static pressure modulates the hemostatic potential of vascular cells by modifying the molecular organization of the plasma membrane.

Tissue-specific alternative mRNA splicing of phenylethanolamine N-methyltransferase (PNMT) during development by intron retention

The expression of phenylethanolamine N-methyl transferase (EC 2. 1.1.2.8, PNMT), the final enzyme in the cascade of catecholamine synthesis, is differentially regulated in adrenergic neurons in the brain and in adrenal chromaffin cells. Using reverse transcription-polymerase chain reaction-based techniques, we detected in the prenatal developing rat brainstem, two species of PNMT mRNA which were produced by a rare alternative splicing mechanism known as intron retention. The spliced, intronless message was downregulated postnatally, while the intron-retained mRNA species continued to be constitutively expressed through adulthood. By contrast in the adrenals, at all stages of development examined, only the intronless message was expressed. In line with previous reports on the failure of glucocorticoids to induce PNMT expression in the brain, the pattern of PNMT splicing in brainstem explants was not affected by the presence of the synthetic glucocorticoid dexamethasone. Undifferentiated sympathoadrenal PC12 pheochromocytoma cells expressed very low basal levels of both mRNA variants, accompanied by a very low basal PNMT enzymatic activity. Exposure of PC12 cells to dexamethasone resulted in the upregulation of only the spliced mRNA variant concomitant with a 3-fold increase in PNMT enzymatic activity. In contrast, treatment of PC 12 cells with nerve growth factor (NGF) enhanced the expression of both the intron-retained and the intronless mRNA species without changes in the basal enzyme activity. This latter result suggests that the translation of the intronless mRNA species may be regulated by the intron-retained mRNA species, which by itself may yield a truncated, yet enzymatically functional translational product. Our data suggest that the tissue-specific regulation of PNMT expression is based on a rare alternative splicing mechanism termed intron retention, and that in the adrenal, but not in the brain, this mechanism is sensitive to regulation by glucocorticoids. Thus, this system is uniquely suited for studying the hormonal control of tissue-specific splicing in the nervous system.

Comparison of ICAM-1 and VCAM-1 expression in various human endothelial cell types and smooth muscle cells

The diversity in the local manifestation of inflammatory vascular lesions might be partially attributable to heterogenous cell adhesion molecule (CAM) expression among endothelial cells (EC) derived from different anatomical locations. We compared basal and tumor necrosis factor-alpha (TNFalpha, 0-100 ng/ml, 0-48 h)-induced intercellular adhesion molecule-1 (ICAM-1) and vascular cell adhesion molecule-1 (VCAM-1) expression in cultured human aortic EC (HAEC), vena cava EC (HVCEC), dermal microvascular EC (HMVEC), and vena cava smooth muscle cells (HVCSM), using a fluorescent ELISA and the competitive quantitative RT-PCR. We found marked differences in basal ICAM-1 expression, both at the protein and mRNA levels, such that HAEC>HVCEC approximately equal to HMVEC>>HVCSM. Basal VCAM-1 mRNA levels were significantly lower in HVCEC than in HAEC and HVCSM, while protein levels were indistinguishable. TNFalpha-induced ICAM-1 and VCAM-1 levels in all EC were similar and significantly higher than in HVCSM (2.5- and 5-fold, respectively). Dissimilar levels of basal and TNFalpha-induced CAM expression in vascular cells may explain the varied predisposition of different blood vessels to developing certain vasculopathies.

Simulated microgravity conditions enhance differentiation of cultured PC12 cells towards the neuroendocrine phenotype

We are studying microenvironmental cues which contribute to neuroendocrine organ assembly and tissue-specific differentiation. As our in vitro model, we cultured rat adrenal medullary PC12 pheochromocytoma cells in a novel cell culture system, the NASA rotating wall vessel (RWV) bioreactors. This "simulated microgravity" environment in RWV bioreactors, characterized by randomizing gravitational vectors and minimizing shear stress, has been shown to favor macroscopic tissue assembly and to induce tissue-specific differentiation. We hypothesized that the unique culture conditions in the RWV bioreactors might enhance the in vitro formation of neuroendocrine organoids. To test our hypothesis, we evaluated the expression of several markers of neuroendocrine differentiation in cultures of PC12 cells maintained for up to 20 d in the slow turning lateral vessel (STLV) type RWV. PC12 cell differentiation was assessed by morphological, immunological, biochemical and molecular techniques. PC12 cells, cultured under "simulated microgravity" conditions, formed macroscopic, tissue-like organoids several millimeters in diameter. Concomitantly, the expression of phenylethanolamine-N-methyl transferase (PNMT), but not of other catecholamine synthesizing enzymes, was enhanced. Increased PNMT expression, as verified on both the gene and protein level, was accompanied by an increase in the specific activity of the enzyme. Furthermore, after 20 d in culture in the STLV, we observed altered patterns of protein tyrosine phosphorylation and prolonged activation of c-fos, a member of the AP-1 nuclear transcription factor complex. We conclude that culture conditions in the RWV appear to selectively activate signal transduction pathways leading to enhanced neuroendocrine differentiation of PC12 cells.

Thrombin modulates vectorial secretion of extracellular matrix proteins in cultured endothelial cells

We have identified a novel cellular action of thrombin on cultured rat adrenal medullary endothelial cells (RAMEC). Five-minute incubation of RAMEC with physiological concentrations of thrombin (<1 U/ml) caused within 3 h an increase in the basolateral deposition of the extracellular matrix (ECM) proteins fibronectin, laminin, and collagens IV and I, concomitant with a corresponding decrease in the apical release of these proteins into the medium. This shift in vectorial secretion of ECM proteins, quantitated with enzyme-linked immunoassays, was time dependent. Maximal stimulation of ECM protein deposition was observed after incubation of cells with thrombin for 5-15 min. Prolonged exposure (>1 h) to thrombin resulted in loss of proteins from the ECM. Thrombin-stimulated ECM protein deposition exhibited a bell-shaped dose dependence, peaking for all proteins at 0.25 U/ml of thrombin, and was independent of de novo mRNA or protein synthesis. Maximal amounts of deposited proteins increased between 2.5-fold (fibronectin) and 4-fold (collagen I) over baseline values. Similar results were obtained with thrombin receptor agonist peptide (TRAP), proteolytically active gamma-thrombin, and, to a lesser extent, other serine proteases such as trypsin and plasmin. A scrambled TRAP, proteolytically inactive PPACK-thrombin, DIP-thrombin, and type IV collagenase were ineffective. Together, these results suggest that the thrombin effects are mediated by proteolytic activation of the thrombin receptor. Possible involvement of the phospholipase C-signaling pathway in thrombin-mediated ECM protein deposition was also investigated. Inhibition or downregulation of protein kinase C (PKC) and chelation of intracellular or extracellular Ca2+ did not suppress, but rather enhanced, basal and thrombin-stimulated ECM protein deposition. Quantitative differences in augmentation of basolateral deposition by these treatments suggest differential regulatory pathways for individual ECM proteins. Our data indicate that, in cultured RAMEC, short-term activation of the thrombin receptor causes an increase in amounts of deposited ECM protein by a cellular signaling pathway that is independent of PKC activation and/or elevation of intracellular Ca2+.

Sequence and analysis of an aldose (xylose) reductase gene from the xylose-fermenting yeast Pachysolen tannophilus

A xylose reductase gene was isolated from the xylose-fermenting yeast Pachysolen tannophilus as a cDNA clone by selecting clones that hybridized specifically to xylose-inducible messenger RNA. Use of the cDNA clone as a probe in Northern hybridizations identified a xylose-inducible mRNA species large enough to encode a 36 kDa xylose reductase protein known to be produced by this yeast. A corresponding genomic clone was isolated as a 3 kb EcoRI fragment that specifically hybridized to the cDNA clone. The sequence of the cDNA and the largest open reading frame of the genomic clone are identical. The predicted translation product exhibits: (1) significant sequence identity with a previously published N-terminal amino acid sequence from purified P. tannophilus xylose (aldose) reductase protein exhibiting NADH/NADPH-dependent activities (aldose reductase, EC 1.1.1.21); (2) identity with a protein composed of 317 amino acid residues with a calculated molecular mass of 36.2 kDa, equivalent to that reported for purified P. tannophilus xylose reductase; and (3) considerable sequence similarity to, and features of, a superfamily of oxidoreductases.

The nucleotide and deduced amino acid sequences of a peptidyl-prolyl cis-trans isomerase from Arabidopsis thaliana

The nucleotide sequence of a cDNA clone from Arabidopsis thaliana ecotype Columbia was determined, and the corresponding amino acid sequence deduced. The predicted sequence is highly homologous to those of members of the cyclophilin A-family of peptidyl-prolyl cis-trans isomerases. The tryptophan residue necessary for interaction with cyclosporin A is located at approximately the same position in the plant sequence as it is in mammalian sequences. The carboxy-terminus of the plant prolyl isomerase sequence contains a cryptic prenylation motif.

Isolation and sequence analysis of a gene from the linear DNA plasmid pPacl-2 of Pichia acaciae that shows similarity to a killer toxin gene of Kluyveromyces lactis

The toxin-encoding linear plasmid systems found in Pichia acaciae and Kluyveromyces lactis yeasts appear to be quite similar, both in function and structural organization. By Southern hybridization, a linear plasmid of P. acaciae, pPacl-2, was found to hybridize to the second open reading frame (ORF2) of K. lactis plasmid pGKL1, known to encode the alpha and beta subunits of the K. lactis toxin. A 1.7 kbp segment of pPacl-2 DNA was cloned, sequenced and shown to contain four regions of strong homology to four similarly oriented regions of K. lactis ORF2. This 1.7 kbp fragment also contained an ORF of 1473 bp that could encode a protein of approximately 55.8 kDa. Like the alpha subunit gene of K. lactis ORF2, a very hydrophobic region occurs at the N-terminus, perhaps representing a signal sequence for transport out of the cell. Unlike K. lactis ORF2, however, the encoded polypeptide is much smaller and lacks a recognizable domain common to chitinases. The structure of a toxin that includes the translation product of this P. acaciae ORF would likely be quite different from that of the K. lactis toxin. Analysis of the upstream region of the P. acaciae ORF revealed an upstream conserved sequence identical to that found before ORFs 8 and 9 of pGKL2. A possible hairpin loop structure, as has been described for each of the four K. lactis pGKL1 ORFs, was found just upstream of the presumed start codon. The similarity of the promoter-like elements found in the linear plasmid genes of these diverse yeasts reinforces the idea of the existence of a unique, but highly conserved, expression system for these novel plasmids. The sequence has been deposited in the GenBank data library under Accession Number U02596.

Movement of shuttle plasmids from Escherichia coli into yeasts other than Saccharomyces cerevisiae using trans-kingdom conjugation

Transfer of bacteria/yeast shuttle plasmids from Escherichia coli into the yeast species Kluyveromyces lactis, Pichia angusta (Hansenula polymorpha), and Pachysolen tannophilus has been accomplished, presumably through inter-kingdom conjugal transfer. Plasmid pEK2 was transferred into a K. lactis mutant to complement trp auxotrophy, while plasmid YEp13 was mobilized into and complemented P. angusta and P. tannophilus Leu- auxotrophs. Plasmid DNA in the recipient strains was detected by transformation of E. coli with crude yeast cell extracts. Freely replicating plasmids without detectable alterations as well as plasmids with rearrangements were recovered from yeast transconjugants.

Genetic analysis of the agrocinopine catabolic region of Agrobacterium tumefaciens Ti plasmid pTiC58, which encodes genes required for opine and agrocin 84 transport

The acc region, subcloned from pTiC58 of classical nopaline and agrocinopine A and B Agrobacterium tumefaciens C58, allowed agrobacteria to grow using agrocinopine B as the sole source of carbon and energy. acc is approximately 6 kb in size. It consists of at least five genes, accA through accE, as defined by complementation analysis using subcloned fragments and transposon insertion mutations of acc carried on different plasmids within the same cell. All five regions are required for agrocin 84 sensitivity, and at least four are required for agrocinopine and agrocin 84 uptake. The complementation results are consistent with the hypothesis that each of the five regions is separately transcribed. Maxicell experiments showed that the first of these genes, accA, encodes a 60-kDa protein. Analysis of osmotic shock fractions showed this protein to be located in the periplasm. The DNA sequence of the accA region revealed an open reading frame encoding a predicted polypeptide of 59,147 Da. The amino acid sequence encoded by this open reading frame is similar to the periplasmic binding proteins OppA and DppA of Escherichia coli and Salmonella typhimurium and OppA of Bacillus subtilis.

Opine catabolism and conjugal transfer of the nopaline Ti plasmid pTiC58 are coordinately regulated by a single repressor

The Ti plasmids of Agrobacterium tumefaciens are conjugal elements whose transfer is strongly repressed. Transfer is induced by the conjugal opines, a group of unique carbon compounds synthesized in crown gall tumors. The opines also induce Ti plasmid-encoded genes required by the bacteria for opine catabolism. We have cloned and sequenced a gene from the Ti plasmid pTiC58, whose product mediates the opine-dependent regulation of conjugal transfer and catabolism of the conjugal opines, agrocinopines A and B. The gene, accR, is closely linked to the agrocinopine catabolic locus. A spontaneous mutant Ti plasmid, pTiC58Trac, which constitutively expresses conjugal transfer and opine catabolism, was complemented in trans by a clone of wild-type accR. Comparative sequence analysis identified a 5-base-pair deletion close to the 5' end of the mutant accR allele from pTiC58Trac. Analysis of lacZ fusions in conjugal transfer and opine catabolic structural genes demonstrated that the accR-encoded function is a transcriptional repressor. accR can encode a 28-kDa protein. This protein is related to a class of repressor proteins that includes LacR, GutR, DeoR, FucR, and GlpR that regulate sugar catabolic systems in several bacterial genera.

Linear DNA plasmids of Pichia inositovora are associated with a novel killer toxin activity

Pichia inositovora, strain NRRL Y-18709, which contains three linear double-stranded DNA plasmids, pPinl-1, pPinl-2 and pPinl-3, was cured of these plasmids both by growing the strain in the presence of 50 micrograms/ml bisbenzimide, and by exposure to ultraviolet light. Both cured and uncured strains were tested for growth on a variety of carbon sources. No differences in growth response were detected, indicating no discernible involvement of the linear plasmids in the catabolism of these compounds. Culture supernatants of Pichia inositovora were shown to contain a substance larger than 100 kDa that is toxic to Saccharomyces cerevisiae, strain GS 1688. Toxin activity was optimal in YEPD assay plates containing 50 mM citrate buffer with a pH between 3.4 and 4.2. Culture supernatants from P. inositovora were also weakly active against Cephaloascus albidus, strain NRRL Y-18710, and Citeromyces matritensis, strain NRRL Y-18711. Concentrated supernatants from cured P. inositovora strains did not exhibit these activities, consistent with the hypothesis that this toxic activity is linear plasmid-encoded. Unlike the well-known Kluyveromyces lactis system, or the newly identified P. acaciae system, P. inositovora strains cured of their linear plasmids do not become detectably sensitive to toxin produced by the wild-type strain suggesting a nonplasmid-encoded immunity function.

Agrobacterium plasmids encode structurally and functionally different loci for catabolism of agrocinopine-type opines

Agrobacterium tumefaciens strains C58, T37, K827 and J73, A. rhizogenes strains A4 and 15834, and A. radiobacter strain K299 were all susceptible to agrocin 84 and this sensitivity was enhanced in each case by addition of agrocinopines A and B. Analysis of transconjugants showed that sensitivity of strain A4 to agrocin 84 was encoded by pArA4a and not by the rhizogenic plasmid, pRiA4. The acc region of the A. tumefaciens nopaline-type Ti plasmid pTiC58, contained on the recombinant plasmid pTHH206, hybridized strongly to restriction fragments of plasmids from strains T37, K827, J73 and K299. Hybridizing fragment patterns generated with BamHI and EcoRI were identical among the four Ti plasmids while pAtK299 showed restriction fragment length polymorphisms at acc with the two enzymes. At moderate stringency, the pTiC58 acc region hybridized weakly to a single restriction fragment from the Ar plasmid of A. rhizogenes strain A4, but not to pTiBo542, which encodes catabolism of the closely related opines agrocinopines C and D. Plasmid pAtK84b of A. radiobacter strain K84 is induced for conjugal transfer by agrocinopines A and B. However, no hybridization was detected between this plasmid and acc from pTiC58 under conditions of moderate stringency. Like pTiC58, pAtK84b conferred transport of agrocinopines A and B on its host bacteria despite the absence of detectable sequence homology with the pTiC58-derived acc probe. However, unlike pTiC58, pAtK84b failed to confer sensitivity to or uptake of agrocin 84 on its bacterial host. These results indicate that at least four distinguishable systems exist for catabolism of the two agrocinopine opine families with the prototype locus, exemplified by acc from pTiC58, being strongly conserved among nopaline-type Ti plasmids.

Characterization and mapping of the agrocinopine-agrocin 84 locus on the nopaline Ti plasmid pTiC58

Overlapping segments of pTiC58 inserted into cosmid vectors were used to characterize the agrocinopine-agrocin 84 locus from the nopaline/agrocinopine A and B Agrobacterium tumefaciens strain C58. All of the clones conferring agrocin 84 sensitivity on agrobacteria also conferred uptake of agrocin 84 and agrocinopines A and B. Transposon Tn3-HoHo1 insertion mutations of one such clone were generated that simultaneously abolished agrocin 84 sensitivity and transport of agrocinopines A and B and agrocin 84. Such insertions were found to cluster within a 4.4-kilobase region. Analysis of beta-galactosidase activity in these insertion mutants suggested a single transcriptional unit regulated at the transcriptional level by agrocinopines A and B. The smallest DNA fragment subcloned from the region to confer all three activities was 8.5 kilobases long. This subclone was still properly regulated, indicating that the regulatory gene is closely linked to the locus. The data are consistent with a single operon encoding catabolism of agrocinopines A and B and conferring sensitivity to agrocin 84. Based on these results, we support the locus name acc, for agrocinopine catabolism.