An enhanced workflow for variant interpretation in UniProtKB/Swiss-Prot improves consistency and reuse in ClinVar.

Personalized genomic medicine depends on integrated analyses that combine genetic and phenotypic data from individual patients with reference knowledge of the functional and clinical significance of sequence variants. Sources of this reference knowledge include the ClinVar repository of human genetic variants, a community resource that accepts submissions from external groups, and UniProtKB/Swiss-Prot, an expert-curated resource of protein sequences and functional annotation. UniProtKB/Swiss-Prot provides knowledge on the functional impact and clinical significance of over 30 000 human protein-coding sequence variants, curated from peer-reviewed literature reports. Here we present a pilot study that lays the groundwork for the integration of curated knowledge of protein sequence variation from UniProtKB/Swiss-Prot with ClinVar. We show that existing interpretations of variant pathogenicity in UniProtKB/Swiss-Prot and ClinVar are highly concordant, with 88% of variants that are common to the two resources having interpretations of clinical significance that agree. Re-curation of a subset of UniProtKB/Swiss-Prot variants according to American College of Medical Genetics and Genomics (ACMG) guidelines using ClinGen tools further increases this level of agreement, mainly due to the reclassification of supposedly pathogenic variants as benign, based on newly available population frequency data. We have now incorporated ACMG guidelines and ClinGen tools into the UniProt Knowledgebase (UniProtKB) curation workflow and routinely submit variant data from UniProtKB/Swiss-Prot to ClinVar. These efforts will increase the usability and utilization of UniProtKB variant data and will facilitate the continuing (re-)evaluation of clinical variant interpretations as data sets and knowledge evolve.

Gene Ontology annotations and resources.

The Gene Ontology (GO) Consortium (GOC, http://www.geneontology.org) is a community-based bioinformatics resource that classifies gene product function through the use of structured, controlled vocabularies. Over the past year, the GOC has implemented several processes to increase the quantity, quality and specificity of GO annotations. First, the number of manual, literature-based annotations has grown at an increasing rate. Second, as a result of a new 'phylogenetic annotation' process, manually reviewed, homology-based annotations are becoming available for a broad range of species. Third, the quality of GO annotations has been improved through a streamlined process for, and automated quality checks of, GO annotations deposited by different annotation groups. Fourth, the consistency and correctness of the ontology itself has increased by using automated reasoning tools. Finally, the GO has been expanded not only to cover new areas of biology through focused interaction with experts, but also to capture greater specificity in all areas of the ontology using tools for adding new combinatorial terms. The GOC works closely with other ontology developers to support integrated use of terminologies. The GOC supports its user community through the use of e-mail lists, social media and web-based resources.

Ectopic expression of the beta-cell specific transcription factor Pdx1 inhibits glucagon gene transcription.

AIMS/HYPOTHESIS: The transcription factor Pdx1 is required for the development and differentiation of all pancreatic cells. Beta-cell specific inactivation of Pdx1 in developing or adult mice leads to an increase in glucagon-expressing cells, suggesting that absence of Pdx1could favour glucagon gene expression by a default mechanism. METHOD: We investigated the inhibitory role of Pdx1 on glucagon gene expression in vitro. The glucagonoma cell line InR1G9 was transduced with a Pdx1-encoding lentiviral vector and insulin and glucagon mRNA levels were analysed by northern blot and real-time PCR. To understand the mechanism by which Pdx1 inhibits glucagon gene expression, we studied its effect on glucagon promoter activity in non-islet cells using transient transfections and gel-shift analysis. RESULTS: In glucagonoma cells transduced with a Pdx1-encoding lentiviral vector, insulin gene expression was induced while glucagon mRNA levels were reduced by 50 to 60%. In the heterologous cell line BHK-21, Pdx1 inhibited by 60 to 80% the activation of the alpha-cell specific element G1 conferred by Pax-6 and/or Cdx-2/3. Although Pdx1 could bind three AT-rich motifs within G1, two of which are binding sites for Pax-6 and Cdx-2/3, the affinity of Pdx1 for G1 was much lower as compared to Pax-6. In addition, Pdx1 inhibited Pax-6 mediated activation through G3, to which Pdx1 was unable to bind. Moreover, a mutation impairing DNA binding of Pdx1 had no effect on its inhibition on Cdx-2/3. Since Pdx1 interacts directly with Pax-6 and Cdx-2/3 forming heterodimers, we suggest that Pdx1 inhibits glucagon gene transcription through protein to protein interactions with Pax-6 and Cdx-2/3. CONCLUSION/INTERPRETATION: Cell-specific expression of the glucagon gene can only occur when Pdx1 expression extinguishes from the early alpha cell precursor.

The pancreatic beta-cell-specific transcription factor Pax-4 inhibits glucagon gene expression through Pax-6.

AIMS/HYPOTHESIS: The paired-homeobox genes pax-4 and pax-6 are crucial for islet development; whereas the null mutation of pax-6 results in the nearly absence of glucagon-producing alpha cells, pax-4 homozygous mutant mice lack insulin and somatostatin-producing beta and delta cells but contain an increased number of alpha cells suggesting that alpha cells could develop by a default mechanism. METHODS: To investigate whether beta-cell specific factors act negatively on glucagon gene transcription, we ectopically expressed pax-4 in glucagon producing InR1G9 cells; Pax-4 inhibited basal transcription of the glucagon gene promoter by 60%. To assess the mechanism of this inhibition, we cotransfected the non-islet cell line BHK-21 with Pax-4 and various transcription factors present in alpha cells. RESULTS: In addition to a general repressor activity on basal glucagon gene promoter activity of 30-50%, a specific 90% inhibition of Pax-6 mediated transactivation was observed. In contrast, Pax-4 had no effect on Cdx-2/3 or HNF3alpha mediated transcriptional activation. Pax-4 showed similar affinity to the Pax-6 binding sites on the glucagon gene promoter compared to Pax-6, but varying with KCl concentrations. CONCLUSION/INTERPRETATION: Pax-4 impairs glucagon gene transcription specifically through inhibition of Pax-6 mediated transactivation. Transcriptional inhibition seems to be mediated by direct DNA binding competition with Pax-6 and potentially additional mechanisms such as protein-protein interactions and a general repressor activity of Pax-4. Glucagon gene expression in alpha cells could thus result from both the presence of islet cell specific transcription factors and the absence of Pax-4.

The paired homeodomain transcription factor Pax-2 is expressed in the endocrine pancreas and transactivates the glucagon gene promoter.

Glucagon gene expression is controlled by at least four DNA elements within the promoter; G2, G3, and G4 confer islet-specific expression, while G1 restricts glucagon transcription to alpha cells. Two islet-specific complexes are formed on G3, the insulin-responsive element of the glucagon gene; one of these corresponds to the paired homeodomain protein Pax-6, a major glucagon gene transactivator that plays a crucial role in alpha cell development. We describe here the identification of the second complex as Pax-2, another member of the paired box family. Pax-2 is known to be crucial for the development of the urogenital tract and of the central nervous system, but its presence in the endocrine pancreas has not been reported. We detected Pax-2 gene expression by RT-PCR; in islets, Pax-2 is present as two alternative splicing isoforms, Pax-2A and Pax-2B, whereas in the glucagon- and insulin-producing cell lines alphaTC1 and Min6, a distinct isoform, Pax-2D2, is found in addition to Pax-2B. Both islet-specific isoforms bind to the enhancer element G3 and to the alpha-specific promoter element G1 that also interacts with Pax-6. Pax-2A and Pax-2B dose-dependently activate transcription from the G3 and the G1 elements both in heterologous and in glucagon-producing cells. Our data indicate that Pax-2 is the third paired domain protein present in the endocrine pancreas and that one of its roles may be the regulation of glucagon gene expression.

Prognostic significance of p53 mutation in breast cancer: frequent detection of non-missense mutations by yeast functional assay.

p53 status was tested in 180 patients with primary breast cancer using a yeast functional assay. Mutations were identified in 32% of cases. Only half were point missense mutations; the remainder were nonsense, insertion, deletion and splice site mutations. Twenty-two percent of mutations were located outside exons 5-8. For a median follow-up of 88 months, survival analysis showed that p53 mutation conferred a worse prognosis in the whole population and the node-positive subgroup but not in node-negative patients. p53 status, tumour size >2 cm, axillary lymph node metastasis and high histological grade were major adverse risk factors in univariate analysis. Multivariate analysis of 153 patients for whom full data were available showed that p53 status contributed prognostic information when tumour size and lymph node status were taken into account but not when histological grade was included. p53 status thus contributes only limited new prognostic information in breast cancer when established prognostic factors are taken into account. Int. J. Cancer (Pred. Oncol.) 84:587-593, 1999.

Increased apoptosis induction by 121F mutant p53.

p53 mutants in tumours have a reduced affinity for DNA and a reduced ability to induce apoptosis. We describe a mutant with the opposite phenotype, an increased affinity for some p53-binding sites and an increased ability to induce apoptosis. The apoptotic function requires transcription activation by p53. The mutant has an altered sequence specificity and selectively fails to activate MDM2 transcription. Loss of MDM2 feedback results in overexpression of the mutant, but the mutant kills better than wild-type p53 even in MDM2-null cells. Thus the apoptotic phenotype is due to a combination of decreased MDM2 feedback control and increased or unbalanced expression of other apoptosis-inducing p53 target genes. To identify these genes, DNA chips were screened using RNA from cells expressing the apoptosis-inducing mutant, 121F, and a sequence-specificity mutant with the reciprocal phenotype, 277R. Two potential new mediators of p53-dependent apoptosis were identified, Rad and PIR121, which are induced better by 121F than wild-type p53 and not induced by 277R. The 121F mutant kills untransformed MDM2-null but not wild-type mouse embryo fibroblasts and kills tumour cells irrespective of p53 status. It may thus expand the range of tumours which can be treated by p53 gene therapy.

Pax-6 and Cdx-2/3 interact to activate glucagon gene expression on the G1 control element.

The promoter element G1, critical for alpha-cell-specific expression of the glucagon gene, contains two AT-rich sequences important for transcriptional activity. Pax-6, a paired homeodomain protein previously shown to be required for normal alpha-cell development and to interact with the enhancer element G3 of the glucagon gene, binds as a monomer to the distal AT-rich site of G1. However, although the paired domain of Pax-6 is sufficient for interaction with the G3 element, the paired domain and the homeodomain are required for high affinity binding to G1. In addition to monomer formation, Pax-6 interacts with Cdx-2/3, a caudal-related homeodomain protein binding to the proximal AT-rich site, to form a heterodimer on G1. Both proteins are capable of directly interacting in the absence of DNA. In BHK-21 cells, Pax-6 activates glucagon gene transcription both through G3 and G1, and heterodimerization with Cdx-2/3 on G1 leads to more than additive transcriptional activation. In glucagon-producing cells, both G1 and G3 are critical for basal transcription, and the Pax-6 and Cdx-2/3 binding sites are required for activation. We conclude that Pax-6 is not only critical for alpha-cell development but also for glucagon gene transcription by its independent interaction with the two DNA control elements, G1 and G3.

Use of transcription reporters with novel p53 binding sites to target tumour cells expressing endogenous or virally transduced p53 mutants with altered sequence-specificity.

p53 triple mutants (120N/121G/277H, 120H/121G/ 277H, 120S/121G/277H and 120H/121G/277Y) have altered sequence specificity in bandshift assays in vitro and transcription assays in vivo. These mutants activate transcription from the site TTT CATG AAA but not from wild type sites. The triple mutants activate more strongly than p53 with a single 277Y mutation. The TTT site matches the wild type p53 consensus at only 4/10 positions and is not recognised by wild type p53. 277Y mutations have been described in human tumours, and Ewing tumour cells expressing this mutant from the endogenous p53 locus selectively activate transcription from transfected luciferase reporters regulated by TTT-mutant p53 binding sites. p53 mutants with altered sequence specificity have potential advantages for cancer gene therapy: if used to activate transcription of conditionally toxic genes they would allow tumour-targeting by p53, which acts as a sensor for the malignant state, but place control over cell killing in the hands of the clinician. Rare tumours expressing such mutants from the endogenous p53 locus could be targeted directly with p53-regulated suicide vectors, but for most tumours both the p53 mutant and the reporter would need to be encoded by the virus.

Field cancerisation and polyclonal p53 mutation in the upper aero-digestive tract.

Field cancerisation of the aerodigestive tract is caused by chronic exposure to alcohol and tobacco, but the nature of the genetic alterations preceding overt malignancy is unknown. To identify potential field changes we have used a functional assay which tests the transcriptional competence of human p53 expressed in yeast. To increase the sensitivity and reliability of the technique for samples containing under 20% mutant p53, the 5' and 3'-ends of the p53 cDNA were examined separately. With this split form of the assay the tissue p53 mRNA acts as its own control for RNA quality. Mutations were detected in 87% (46/53) of tumours, reflecting the high sensitivity of the technique. Multiple biopsies of histologically normal tissue from the upper aero-digestive tract were tested and clonal p53 mutations were identified in 76% (38/50) of biopsies from patients presenting with multiple tumours compared with 32% (38/117) of biopsies from patients presenting with single tumours (P<0.000001). All patients (16/16) presenting with multiple tumours had at least one positive biopsy, compared with only 53% (19/36) of patients presenting with single tumours (P <0.001). This defines expansion of multiple clones of mutant p53-containing cells as an important biological mechanism of field cancerisation, and provides a means to identify patients likely to benefit from intensive screening for the development of new head and neck tumours.

Inverse correlation between RER+ status and p53 mutation in colorectal cancer cell lines.

The high point mutation rate of replication error-prone (RER+) cells could theoretically lead to inactivation of the p53 gene by polyclonal mutations, which might explain the conflicting results that have been published on the p53 status of RER+ colon cancers. To address this issue, we tested the p53 status of 21 human colorectal cancer cell lines, including four showing microsatellite instability (RER+ phenotype). Denaturing gradient gel electrophoresis (DGGE) followed by sequencing showed that all four RER+ cell lines were wild type for p53 while 15 of the 17 RER- cell lines contained p53 mutations (P=0.001). Eight cell lines (four RER+ and four RER-) were analysed using three complementary methods to test more rigorously the polyclonal mutation hypothesis. (i) Of 87 single-cell clones (seven to 14 per cell line) examined by DGGE, only those derived from known p53 mutant cell lines showed altered profiles. (ii) Antibody DO-7 stained more than 80% of nuclei from the p53 mutant cell lines, but only 15% of nuclei from the RER+ cell lines. (iii) A yeast functional assay which can simultaneously detect polyclonal mutations at over 500 different sites in the p53 cDNA scored all four RER+ cell lines as containing only transcriptionally active p53. These data thus do not support the polyclonal mutation hypothesis and instead suggest that mismatch repair deficiency provides a p53-independent pathway for development of colorectal cancers.

Germ cell tumors of the testis overexpress wild-type p53.

Several recent studies have suggested that testicular germ cell tumors express high levels of wild-type p53 protein. To clarify and confirm this unexpected result, we have investigated seminomatous and nonseminomatous germ cell tumors at the genomic, mRNA, and protein levels. Thirty-five tumors were examined for p53 overexpression using antibodies directed against the p53 (PAb1801, PAb240, and CM1), mdm2 (IF2), and p21Waf1/Clp1 (EA10) proteins. Thirty-two tumors were screened for p53 mutations by single-strand conformation polymorphism analysis. Eighteen tumors were screened with a functional assay that tests the transcriptional competence of human p53 protein expressed in yeast. On frozen sections, 100, 65, 35, 73, and 0% of tumors reacted with the CM1, PAb240, PAb1801, IF2, and EA10 antibodies, respectively. No p53 mutations were detected by single-strand conformation polymorphism or by functional assay. The fact that many tumors overexpress wild-type p53 but not mdm2 rules out mdm2 overexpression as a general explanation for the presence of wild-type p53 in these tumors. The absence of p21 overexpression suggests that p53 may be unable to activate transcription of critical target genes, which may explain why the presence of wild-type p53 is tolerated in this tumor type, although the mechanism for this transcriptional inactivity remains to be established.

Low molecular-weight proteins in human gingival crevicular fluid.

Samples of gingival crevicular fluid (GCF) were collected in 30 volunteers with inflamed gingiva, using either capillary tubes (cGCF) or Durapore strips (sGCF). They were examined, together with samples of serum from the same patients, by sodium dodecyl sulphate-polyacrylamide gel electrophoresis and/or by two-dimensional electrophoresis, followed by silver staining. The results confirmed that the distribution of the major proteins in GCF is similar to that found in serum. However, an 8.5 kDa protein was found in gingival fluid but not in serum. The low molecular-weight protein appears to decrease with time of fluid sampling with both techniques, and does not originate either from blood or from the cellular fraction of GCF. Two-dimensional electrophoretic analysis suggested that it may consist of several polypeptides.

Clonality and stability of the p53 gene in human astrocytic tumor cells: quantitative analysis of p53 gene mutations by yeast functional assay.

Mutation of the p53 gene is found in about one third of astrocytic brain tumors, and expansion of tumor cell clones containing mutant p53 has been implicated in astrocytic tumor progression. However, admixture of normal cells in astrocytic tumor specimens limits the power of traditional studies of tumor cell clonality. To address this problem we have employed a yeast p53 functional assay that scores the content of mutant p53 alleles in tumors and cell lines quantitatively. We have analyzed 17 cases where matching tumor material and derived cell lines were available. The yeast assay gave > 20% red (i.e., mutant p53-containing) yeast colonies in 7 out of 17 cases. One case had no mutations in the primary tumor but gave 76% red colonies in a recurrence, clearly demonstrating tumor overgrowth by a mutant clone. During early passages of cultured tumor cells, mutant p53 content increased rapidly with passage due to outgrowth of mutant clones from a heterogeneous starting population. In addition, de novo p53 mutations appeared during culture in 2 cases. This indicates that there is stronger selective pressure for mutation during the establishment of cell lines in vitro than during tumor growth in vivo. Our results demonstrate the utility of the p53 functional assay for studies of clonality and support the hypothesis of clonal progression of brain tumors in vivo.

The human tumour suppressor gene p53 is alternatively spliced in normal cells.

Alternative splicing affecting the p53 carboxy-terminus has previously been described in mouse but not in normal human cells. We report here the detection in normal human lymphocytes of an alternatively spliced form of human p53 mRNA containing an additional 133 bp exon derived from intron 9. This splice variant encodes a truncated protein of 341 amino-acids including 10 new amino-acids derived from the novel exon. The truncated protein, which lacks part of the p53 tetramerization domain, fails to bind DNA in vitro and has a transcriptional defect in vivo in both yeast and mammalian cells. Quantitative RT-PCR experiments suggest that the alternatively spliced form is only present in significant amounts in quiescent cells. Considering the numerous functions ascribed to the carboxy-terminus of the p53 protein, this splice variant may have important implications for the biological role of p53 in normal cells.

Polarized secretion of urokinase-type plasminogen activator by epithelial cells.

Numerous epithelial cell types produce and secrete plasminogen activators (PAs) and/or PA inhibitors (PAIs). When epithelial cells were grown on polycarbonate filters and their apical and basolateral secretion products analyzed, PA activity accumulated in a highly polarized fashion; depending upon the cell line, the compartment of PA accumulation was either apical (MDCK I cells and HBL-100 cells) or basolateral (LLC-PK1, CaCo-2, and HeLa cells). By contrast, PAI-1 was recovered in roughly equal amounts in both compartments. Basolateral accumulation of urokinase-type plasminogen activator (uPA), but not its apical targeting, required an acidic compartment and the integrity of the cytoskeleton. Polarity of uPA accumulation did not result from removal of the free enzyme from the opposite compartment through its binding to the cell surface. Transfection with wild-type or mutated murine uPA demonstrated that neither the "growth factor" domain nor the kringle domain is required for the appropriate sorting of the protein. We propose that polarized secretion of PAs is one mechanism whereby cells spatially control extracellular proteolysis.

The receptor for urokinase type plasminogen activator polarizes expression of the protease to the leading edge of migrating monocytes and promotes degradation of enzyme inhibitor complexes.

Receptor-bound urokinase-type plasminogen activator (uPA) remains associated to the surface of human monocytes for many hours. Monocytes induced to migrate in a chemotactic gradient of f-Met-Leu-Phe rapidly polarize their uPA receptors to the leading front of the cells. Receptor-bound enzyme can be inhibited by plasminogen activator inhibitor 2 (PAI-2), with a kinetics comparable to that determined for the free enzyme, and uPA/PAI-2 complexes can bind to the uPA receptor. In contrast to the active enzyme, the uPA/PAI-2 complex is rapidly cleared from the monocyte cell surface; this involves an initial cleavage of the complex at the cell surface, followed by endocytosis and degradation. These results indicate that the uPA receptor can function both to focus plasmin-mediated extracellular matrix degradation in front of migrating cells, and to target uPA/PAI-2 enzyme/inhibitor complexes for degradation; they suggest that this receptor is a key determinant in the control of uPA-catalyzed extracellular proteolysis.

Characterization of the cellular binding site for the urokinase-type plasminogen activator.

Human urokinase-type plasminogen activator (uPA) binds rapidly and with high affinity to a number of human cell types; this localizes plasmin generation to the close environment of the cell surface. uPA binding to HeLa and U937 cells is mediated by a single class of sites with an affinity of 3.4 +/- 1.3 x 10(-10) M. Binding is abolished by treatment of the cells with trypsin. Chemical cross-linking of Mr 55,000 125I-uPA to the surface of HeLa and U937 cells with disuccinimidyl suberate or with formaldehyde results in the formation of a labeled complex of Mr 100,000, suggesting a Mr of 45,000 +/- 5,000 for the receptor or a subunit thereof. When cells solubilized in Triton X-114 are subjected to heat-induced phase separation, unoccupied receptor, receptor-bound 125I-uPA, and cross-linked 125I-uPA-receptor complex all partition in the detergent phase, whereas the unbound ligand remains in the aqueous phase; similar phase partitioning is observed with endogenous uPA-receptor complexes from cultured human and murine cells. Thus, uPA bound at the cell surface is tightly associated with an amphiphilic membrane protein. Interaction of uPA with this plasma membrane receptor is species-specific, since human uPA fails to bind to murine cells, and murine uPA does not bind to human cells. Finally, incubation of HeLa cells in the presence of epidermal growth factor or phorbol 12-myristate 13-acetate results, over a period of 24 h, in a progressive change in uPA binding: an approximately 10-fold increase in the number of sites is accompanied by a 10-fold decrease in their affinity. Cross-linking and phase partitioning of 125I-uPA bound to epidermal growth factor- or phorbol 12-myristate 13-acetate-treated cells indicate that, as in control conditions, it is associated with a Mr 45,000 cell surface amphiphilic polypeptide.