Hepsin is a membrane-anchored, trypsin-like serine protease with prominent expression in the human liver and tumours of the prostate and ovaries. To better understand the biological functions of hepsin, we identified macromolecular substrates employing a tetrapeptide PS-SCL (positional scanning-synthetic combinatorial library) screen that rapidly determines the P1-P4 substrate specificity. Hepsin exhibited strong preference at the P1 position for arginine over lysine, and favoured threonine, leucine or asparagine at the P2, glutamine or lysine at the P3, and proline or lysine at the P4 position. The relative activity of hepsin toward individual AMC (7-amino-4-methylcoumarin)-tetrapeptides was generally consistent with the overall peptide profiling results derived from the PC-SCL screen. The most active tetrapeptide substrate Ac (acetyl)-KQLR-AMC matched with the activation cleavage site of the hepatocyte growth factor precursor sc-HGF (single-chain HGF), KQLR downward arrowVVNG (where downward arrow denotes the cleavage site), as identified by a database analysis of trypsin-like precursors. X-ray crystallographic studies with KQLR chloromethylketone showed that the KQLR peptide fits well into the substrate-binding cleft of hepsin. This hepsin-processed HGF induced c-Met receptor tyrosine phosphorylation in SKOV-3 ovarian cancer cells, indicating that the hepsin-cleaved HGF is biologically active. Activation cleavage site mutants of sc-HGF with predicted non-preferred sequences, DPGR downward arrowVVNG or KQLQ downward arrowVVNG, were not processed, illustrating that the P4-P1 residues can be important determinants for substrate specificity. In addition to finding macromolecular hepsin substrates, the extracellular inhibitors of the HGF activator, HAI-1 and HAI-2, were potent inhibitors of hepsin activity (IC50 4+/-0.2 nM and 12+/-0.5 nM respectively). Together, our findings suggest that the HGF precursor is a potential in vivo substrate for hepsin in tumours, where hepsin expression is dysregulated and may influence tumorigenesis through inappropriate activation and/or regulation of HGF receptor (c-Met) functions.

No abstract

The contribution to genetic diversity of genomic segmental copy number variations (CNVs) is less well understood than that of single-nucleotide polymorphisms (SNPs). While less frequent than SNPs, CNVs have greater potential to affect phenotype. In this study, we have performed the most comprehensive survey to date of CNVs in mice, analyzing the genomes of 42 Mouse Phenome Consortium priority strains. This microarray comparative genomic hybridization (CGH)-based analysis has identified 2094 putative CNVs, with an average of 10 Mb of DNA in 51 CNVs when individual mouse strains were compared to the reference strain C57BL/6J. This amount of variation results in gene content that can differ by hundreds of genes between strains. These genes include members of large families such as the major histocompatibility and pheromone receptor genes, but there are also many singleton genes including genes with expected phenotypic consequences from their deletion or amplification. Using a whole-genome association analysis, we demonstrate that complex multigenic phenotypes, such as food intake, can be associated with specific copy number changes.

We have combined genetic algorithm (GA) and all paired (AP) support vector machine (SVM) methods for multiclass cancer categorization. Predictive features can be automatically determined through iterative GA/SVM, leading to very compact sets of non-redundant cancer-relevant genes with the best classification performance reported to date. Interestingly, these different classifier sets harbor only modest overlapping gene features but have similar levels of accuracy in leave-one-out cross-validations (LOOCV). Further characterization of these optimal tumor discriminant features, including the use of nearest shrunken centroids (NSC), analysis of annotations and literature text mining, reveals previously unappreciated tumor subclasses and a series of genes that could be used as cancer biomarkers. With this approach, we believe that microarray-based multiclass molecular analysis can be an effective tool for cancer biomarker discovery and subsequent molecular cancer diagnosis.