Predicting the Metabolic Sites by Flavin-Containing Monooxygenase on Drug Molecules Using SVM Classification on Computed Quantum Mechanics and Circular Fingerprints Molecular Descriptors.

As an important enzyme in Phase I drug metabolism, the flavin-containing monooxygenase (FMO) also metabolizes some xenobiotics with soft nucleophiles. The site of metabolism (SOM) on a molecule is the site where the metabolic reaction is exerted by an enzyme. Accurate prediction of SOMs on drug molecules will assist the search for drug leads during the optimization process. Here, some quantum mechanics features such as the condensed Fukui function and attributes from circular fingerprints (called Molprint2D) are computed and classified using the support vector machine (SVM) for predicting some potential SOMs on a series of drugs that can be metabolized by FMO enzymes. The condensed Fukui function fA- representing the nucleophilicity of central atom A and the attributes from circular fingerprints accounting the influence of neighbors on the central atom. The total number of FMO substrates and non-substrates collected in the study is 85 and they are equally divided into the training and test sets with each carrying roughly the same number of potential SOMs. However, only N-oxidation and S-oxidation features were considered in the prediction since the available C-oxidation data was scarce. In the training process, the LibSVM package of WEKA package and the option of 10-fold cross validation are employed. The prediction performance on the test set evaluated by accuracy, Matthews correlation coefficient and area under ROC curve computed are 0.829, 0.659, and 0.877 respectively. This work reveals that the SVM model built can accurately predict the potential SOMs for drug molecules that are metabolizable by the FMO enzymes.

Anti-proliferative effect on a colon adenocarcinoma cell line exerted by a membrane disrupting antimicrobial peptide KL15.

The antimicrobial and anticancer activities of an antimicrobial peptide (AMP) KL15 obtained through in silico modification on the sequences of 2 previously identified bacteriocins m2163 and m2386 from Lactobacillus casei ATCC 334 by us have been studied. While significant bactericidal effect on the pathogenic bacteria Listeria, Escherichia, Bacillus, Staphylococcus, Enterococcus is exerted by KL15, the AMP can also kill 2 human adenocarcinoma cells SW480 and Caco-2 with measured IC50 as 50 µg/ml or 26.3 µM. However, the IC50 determined for KL15 on killing the normal human mammary epithelial cell H184B5F5/M10 is 150 µg/ml. The conformation of KL15 dissolved in 50% 2,2,2-trifluroroethanol or in 2 large unilamellar vesicle systems determined by circular dichroism spectroscopy appears to be helical. Further, the cell membrane permeability of treated SW480 cells by KL15 appears to be significantly enhanced as studied by both flow cytometry and confocal microscopy. As observed under a scanning electron microscope, the morphology of treated SW480 cells is also significantly changed as treating time by 80 µg/ml KL15 is increased. KL15 appears to be able to pierce the cell membrane of treated SW480 cells so that numerous porous structures are generated and observable. Therefore, KL15 is likely to kill the treated SW480 cells through the necrotic pathway similar to some recently identified AMPs by others.

Antimicrobial peptide m2163 or m2386 identified from Lactobacillus casei ATCC 334 can trigger apoptosis in the human colorectal cancer cell line SW480.

Ribosomal synthesized antimicrobial peptides (AMPs) are widely distributed in nature and are toxic to certain microorganisms. Some of these AMPs are found to exhibit cytotoxic activity against the growth of cancer cells and thus have obvious anticancer potential. Here, we have studied the antiproliferation on the human colorectal cancer cell line SW480 of two AMPs, namely m2163 and m2386, identified by us from a lactic acid bacterium Lactobacillus casei ATCC 334 previously. A half maximal inhibitory concentration (IC50) of 40 µg/ml is determined first using the MTT (3-(4, 5-cimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide) assay for either peptide m2163 or m2386. The apoptosis in treated SW480 cells by either peptide m2163 or m2386 is analyzed using flow cytometry with annexin V-fluorescein isothiocyanate (FITC) and propidium iodide double staining. These analyses show that a substantial population of treated SW480 cells can undergo apoptosis by either peptide m2163 or m2386. The real-time quantitative polymerase chain reaction (qPCR) and Western blot analyses are subsequently used to study how the apoptosis is induced in the treated SW480 cells by either peptide m2163 or m2386. While m2163 is found to induce the expression of Fas and TRAILR1, the expression of Fas, TNFR1, and TRAILR1 death receptors on the cell surface of treated SW480 cells is found to be induced by m2386. Further, the expression of some mitochondria-related apoptosis proteins such as Smac is found to be also induced, suggesting that either peptide m2163 or m2386 can trigger both the extrinsic and intrinsic apoptosis pathways. The cell membrane permeability is greatly enhanced upon treatment with either peptide m2163 or m2386 as analyzed by the flow cytometry using both FITC-labeled peptides. The flow cytometry is also used to analyze the fluorescence intensity given by FITC-m2163 in either the mitochondria or cytoplasm fraction of the treated and fractionated SW480 cells. It is found that the detected fluorescence intensity of the mitochondria fraction is much weaker than that of the cytoplasm one, suggesting that most of the FITC-m2163 peptides are located in the cytoplasm rather than the mitochondria. This is further confirmed by a confocal microscopy study that either peptide m2163 or m2386 can localize on the cell membrane for a substantial length of time and then penetrate into the cell cytoplasm to induce the apoptosis.

Constructing a linear QSAR for some metabolizable drugs by human or pig flavin-containing monooxygenases using some molecular features selected by a genetic algorithm trained SVM.

A genetic algorithm optimized and feature selectable support vector machine (GFSVM) was designed for classifying some 71 and 151 substrates of human and pig flavin-containing monooxygenases (FMOs; EC 1.14.13.8) collected from the literatures. While a novel fitness function was designed, a feature mask for selecting (represented by bit 1) or masking (represented by bit 0) a feature was also implemented in the chromosomes generated during the evolution process. The feature selection was performed according to the ranked accumulated |w| values computed from several preliminary runs. Some numbers of top ranked features were then selected and gradually increased in a multiple linear regression process employed for building a linear quantitative structure-activity relationships (QSARs) for both human and pig FMOs. Each of these preliminary QSAR models generated was judged by both a conventional and 10 fold cross-validation statistics computed for choosing the best set of top ranked features for building the best linear QSAR model. The best linear QSAR thus constructed for human or pig FMOs was from 89 or 145 top ranked features selected, respectively. Moreover, these two linear QSARs were also found to be specific to their own top ranked features computed and selected. These two linear QSARs constructed may be useful for predicting whether or not a drug is metabolizable by human or pig FMOs if the same feature computation and ranking scheme has been applied on it beforehand.

Virtual Screening of Some Active Human Macrophage Migration Inhibitory Factor Antagonists.

Macrophage migration inhibitory factor (MIF) is an autocrine- and paracrine-acting cytokine that is involved in several inflammatory, autoimmune, infectious, and oncogenic diseases. Clinical data have shown that inhibition of MIF, especially its tautomerase activity, with small compounds has been beneficial in some disease models. A virtual screening (VS) experiment is conducted for searching some active compounds to inhibit the tautomerase activity of MIF from the ZINC database. By using an x-ray-determined structure OXIM-11 as the query and an in-house developed two-dimensional scaffold comparing method designated as Sfilter, we have screened out some 1500 compounds for ranking by our previously published docking method ADDock. After further ranking by ADDock on 119 compounds screened, we have decided to choose 17 of them for measuring their inhibitory activity IC50 against the MIF tautomerase experimentally. The IC50's are measured using both human monocytic THP-1 cell lysate and purified recombinant human MIF protein. We have found that the IC50's measured for three searched compounds (namely, ZINC02693801, ZINC00141102, and ZINC12368346) are better than that determined for ISO-1, a known MIF tautomerase inhibitor and standard used throughout our VS experiment. Moreover, the scaffolds of most of our active compounds searched are also quite different from those searched and published by others previously.

Interaction between trehalose and MTHase from Sulfolobus solfataricus studied by theoretical computation and site-directed mutagenesis.

Maltooligosyltrehalose trehalohydrolase (MTHase) catalyzes the release of trehalose by cleaving the α-1,4-glucosidic linkage next to the α-1,1-linked terminal disaccharide of maltooligosyltrehalose. Computer simulation using the hydrogen bond analysis, free energy decomposition, and computational alanine scanning were employed to investigate the interaction between maltooligosyltrehalose and the enzyme. The same residues that were chosen for theoretical investigation were also studied by site-directed mutagenesis and enzyme kinetic analysis. The importance of residues determined either experimentally or computed theoretically were in good accord with each other. It was found that residues Y155, D156, and W218 of subsites -2 and -3 of the enzyme might play an important role in interacting with the ligand. The theoretically constructed structure of the enzyme-ligand complex was further validated through an ab initio quantum chemical calculation using the Gaussian09 package. The activation energy computed from this latter study was very similar to those reported in literatures for the same type of hydrolysis reactions.

Stable integration and expression of heterologous genes in several lactobacilli using an integration vector constructed from the integrase and attP sequences of phage ΦAT3 isolated from Lactobacillus casei ATCC 393.

An integration vector capable of stably integrating and maintaining in the chromosomes of several lactobacilli over hundreds of generations has been constructed. The major integration machinery used is based on the ΦAT3 integrase (int) and attP sequences determined previously. A novel core sequence located at the 3' end of the tRNA(leu) gene is identified in Lactobacillus fermentum ATCC 14931 as the integration target by the integration vector though most of such sequences found in other lactobacilli are similar to that determined previously. Due to the lack of an appropriate attB site in Lactococcus lactis MG1363, the integration vector is found to be unable to integrate into the chromosome of the strain. However, such integration can be successfully restored by cotransforming the integration vector with a replicative one harboring both attB and erythromycin resistance sequences into the strain. Furthermore, the integration vector constructed carries a promoter region of placT from the chromosome of Lactobacillus rhamnosus TCELL-1 which is used to express green fluorescence and luminance protein genes in the lactobacilli studied.

Characterization of putative class II bacteriocins identified from a non-bacteriocin-producing strain Lactobacillus casei ATCC 334.

Several putative class II bacteriocin-like genes were identified in Lactobacillus casei ATCC 334, all of which might encode peptides with a double-glycine leader. Six peptides encoded by these genes were heterologously expressed in Escherichia coli and then partially purified in order to test their bacteriocin activity. The results revealed that the mature LSEI_2163 peptide was a class IId bacteriocin that exhibited antimicrobial activity against some lactobacilli and several Listeria species. Similarly, mature LSEI_2386 was a putative pheromone peptide that also had significant bacteriocin activity against several Listeria species. The activities of both peptides tolerated 121°C for 30 min but not treatment with proteinase K or trypsin. The two Cys residues located at positions 4 and 24 in the mature LSEI_2163 peptide were shown by mass spectrometry to form a disulfide bridge, which was required for optimal antibacterial activity. However, replacement of one or both Cys with Ser would cause significant reduction of the antibacterial activity, the reduction being greater when only one of the Cys residues (C4S) was replaced than when both (C4S/C24S) were replaced.

Developing consensus 3D-QSAR and pharmacophore models for several beta-secretase, farnesyl transferase and histone deacetylase inhibitors.

Three consensus 3D-QSAR (c-3D-QSAR) models were built for 38, 34, and 78 inhibitors of ß-secretase, histone deacetylase, and farnesyltransferase, respectively. To build an individual 3D-QSAR model, the structures of an inhibitor series are aligned through docking of a protein receptor into the active site using the program GOLD. CoMFA, CoMSIA, and Catalyst are then performed for the training set of each structurally aligned inhibitor series to obtain a 3D-QSAR model. Since the consensus in features identified is high for the same pharmacophore features selected for building a 3D-QSAR model by a 3D-QSAR method, a c-3D-QSAR model for each inhibitor series is constructed by combining the pharmacophore features selected for building the 3D-QSAR model using the SYBYL spread sheet and PLS module. Each c-3D-QSAR pharmacophore model built was examined visually and compared with that obtained by simultaneous mapping of the corresponding 3D-QSAR pharmacophores built onto a selected inhibitor structure. It was found that the c-3D-QSAR model built for an inhibitor series improves not only the overall prediction statistics for both training and test sets but also the prediction accuracy for some less active inhibitors of the series.

Theoretical study on the alkaline hydrolysis of methyl thioacetate in aqueous solution.

A base-catalyzed hydrolysis reaction of thiolester has been studied in both gas and solution phases using two ab initio quantum mechanics calculations such as Gaussian09 and CPMD. The free-energy surface along the reaction path is also constructed using a configuration sampling technique, namely, the metadynamics method. While there are two different reaction paths obtained for the potential profile of the base-catalyzed hydrolysis reaction for thiolester in the gas phase, a triple-well reaction path is computed for the reaction in the solution phase by two quantum mechanics calculations. Unlike the S(N)2 mechanism (a concerted mechanism) found for the gas-phase reaction, a nucleophilic attack from the hydroxide ion on the carbonyl carbon to yield a tetrahedral intermediate (a stepwise mechanism) is observed for the solution-phase reaction. Moreover, the energy profiles computed by these two theoretical calculations are found to be very comparable with those determined experimentally.

Deriving the phylogenetic information from some physicochemical properties of protein sequences computed.

The evolutionary relationships of organisms are traditionally delineated by the alignment-based methods using some DNA or protein sequences. In the post-genome era, the phylogenetics of life could be inferred from many sources such as genomic features, not just from comparison of one or several genes. To investigate the possibility that the physicochemical properties of protein sequences might reflect the phylogenetic ones, an alignment-free method using a support vector machine (SVM) classifier is implemented to establish the phylogenetic relationships between some protein sequences. There are two types of datasets, namely, the "Enzymatic" (assigned by an EC accession) and "Proteins" used to train the SVM classifiers. By computing the F-score for feature selection, we find that the classification accuracies of trained SVM classifiers could be significantly enhanced to 84% and 80%, respectively, for the enzymatic and "proteins" datasets classified if the protein sequences are represented with some top 255 features selected. These show that some physicochemical features of amino acid sequences selected are sufficient for inferring the phylogenetic properties of the protein sequences. Moreover, we find that the selected physicochemical features appear to correlate with the physiological characteristic of the taxonomic classes classified.

HDAPD: a web tool for searching the disease-associated protein structures.

BACKGROUND: The protein structures of the disease-associated proteins are important for proceeding with the structure-based drug design to against a particular disease. Up until now, proteins structures are usually searched through a PDB id or some sequence information. However, in the HDAPD database presented here the protein structure of a disease-associated protein can be directly searched through the associated disease name keyed in. DESCRIPTION: The search in HDAPD can be easily initiated by keying some key words of a disease, protein name, protein type, or PDB id. The protein sequence can be presented in FASTA format and directly copied for a BLAST search. HDAPD is also interfaced with Jmol so that users can observe and operate a protein structure with Jmol. The gene ontological data such as cellular components, molecular functions, and biological processes are provided once a hyperlink to Gene Ontology (GO) is clicked. Further, HDAPD provides a link to the KEGG map such that where the protein is placed and its relationship with other proteins in a metabolic pathway can be found from the map. The latest literatures namely titles, journals, authors, and abstracts searched from PubMed for the protein are also presented as a length controllable list. CONCLUSIONS: Since the HDAPD data content can be routinely updated through a PHP-MySQL web page built, the new database presented is useful for searching the structures for some disease-associated proteins that may play important roles in the disease developing process for performing the structure-based drug design to against the diseases.

A molecular dynamics study on binding recognition between several 4,5 and 4,6-linked aminoglycosides with A-site RNA.

A molecular dynamics (MD) simulation has been performed for two sets of aminoglycoside antibiotics bound with an RNA duplex corresponding to the aminoacyl-tRNA decoding site of the 16S rRNA to characterize the energetics and dynamics of binding for several aminoglycosides. The binding free energy, essential dynamics and hydration analysis have been conducted to characterize the dynamics' properties associated with the binding recognition between each set of antibiotics and the RNA duplex. We have built several dynamic models with reasonable binding free energies showing good correlation with the experimental data. We have also conducted a hydration analysis on some long residency water molecules detected as W8 and W49 sites around the U1406 . U1495 pair and which are found to be important in binding recognition and in causing some apparent stretch variations of this pair during the dynamic studies. In addition, we also find that the hydration sites with long residence time identified between the ring III of two 4,6-linked antibiotics (tobramycin and kanamycin) and phosphate oxygen atoms of G1405/U1406 may be worthy of further exploration for rational drug design.

Specific point mutations in Lactobacillus casei ATCC 27139 cause a phenotype switch from Lac- to Lac+.

Lactose metabolism is a changeable phenotype in strains of Lactobacillus casei. In this study, we found that L. casei ATCC 27139 was unable to utilize lactose. However, when exposed to lactose as the sole carbon source, spontaneous Lac(+) clones could be obtained. A gene cluster (lacTEGF-galKETRM) involved in the metabolism of lactose and galactose in L. casei ATCC 27139 (Lac(-)) and its Lac(+) revertant (designated strain R1) was sequenced and characterized. We found that only one nucleotide, located in the lacTEGF promoter (lacTp), of the two lac-gal gene clusters was different. The protein sequence identity between the lac-gal gene cluster and those reported previously for some L. casei (Lac(+)) strains was high; namely, 96-100 % identity was found and no premature stop codon was identified. A single point mutation located within the lacTp promoter region was also detected for each of the 41 other independently isolated Lac(+) revertants of L. casei ATCC 27139. The revertants could be divided into six classes based on the positions of the point mutations detected. Primer extension experiments conducted on transcription from lacTp revealed that the lacTp promoter of these six classes of Lac(+) revertants was functional, while that of L. casei ATCC 27139 was not. Northern blotting experiments further confirmed that the lacTEGF operon of strain R1 was induced by lactose but suppressed by glucose, whereas no blotting signal was ever detected for L. casei ATCC 27139. These results suggest that a single point mutation in the lacTp promoter was able to restore the transcription of a fully functional lacTEGF operon and cause a phenotype switch from Lac(-) to Lac(+) for L. casei ATCC 27139.

Using support vector regression to model the correlation between the clinical metastases time and gene expression profile for breast cancer.

OBJECTIVE: Recently, the microarray analysis has been an important tool used for studying the cancer type, biological mechanism, and diagnostic biomarkers. There are several machine-learning methods being used to construct the prognostic model based on the microarray data sets. However, most of these previous studies were focused on the supervised classification for predicting the clinical type of patients. In this study, we investigate whether or not the expression level of some significant genes identified can be used to predict the clinical metastases time of patients. MATERIALS AND METHODS: We have used a regression method to remodel the data set of breast cancer published in 2002. Some significant genes were ranked and selected based on a wrapper method with 10-fold cross-validation procedure and the selected genes were used to fit the support vector regression (SVR) model. This method could model the relationship between the significant gene expression value and the clinical metastases time of breast cancer. RESULTS: 44 significant genes are selected for building the regression model and the corresponding cross-validated correlation coefficient obtained is 0.82 which is much superior to those reported previously by others using some different data sets. Moreover, there are two breast cancer related genes (the ligand 14 of the chemokine C-X-C motif (CXCL14) and estrogen receptor gene (ER)) selected in the gene set and one of them is never been included in the other data sets. CONCLUSION: In this report, we have shown that the expression level of some significant genes identified could strongly correlate with the clinical metastases time of breast cancer patients. The 44 selected genes may be used as a benchmark to evaluate the risk of recurrence of breast cancer.

An anchor-dependent molecular docking process for docking small flexible molecules into rigid protein receptors.

A molecular docking method designated as ADDock, anchor-dependent molecular docking process for docking small flexible molecules into rigid protein receptors, is presented in this article. ADDock makes the bond connection lists for atoms based on anchors chosen for building molecular structures for docking small flexible molecules or ligands into rigid active sites of protein receptors. ADDock employs an extended version of piecewise linear potential for scoring the docked structures. Since no translational motion for small molecules is implemented during the docking process, ADDock searches the best docking result by systematically changing the anchors chosen, which are usually the single-edge connected nodes or terminal hydrogen atoms of ligands. ADDock takes intact ligand structures generated during the docking process for computing the docked scores; therefore, no energy minimization is required in the evaluation phase of docking. The docking accuracy by ADDock for 92 receptor-ligand complexes docked is 91.3%. All these complexes have been docked by other groups using other docking methods. The receptor-ligand steric interaction energies computed by ADDock for some sets of active and inactive compounds selected and docked into the same receptor active sites are apparently separated. These results show that based on the steric interaction energies computed between the docked structures and receptor active sites, ADDock is able to separate active from inactive compounds for both being docked into the same receptor.

Formation of an inverted repeat junction in the transposition of insertion sequence ISLC3 isolated from Lactobacillus casei.

An insertion sequence, ISLC3, of 1351 bp has been isolated from Lactobacillus casei. Formation of IS circles containing a 3 bp spacer (complete junction) or deletion of 25 bp at the left inverted repeat (IRL) between the abutted IS ends of the ISLC3 junction region (deleted junction) was also discovered in the lactobacilli and Escherichia coli system studied. We found that the promoter formed by the complete junction P(jun) was more active than that formed by the 25 bp deleted junction P(djun) or the indigenous promoter P(IRL). The corresponding transcription start sites for both promoter P(jun) and P(IRL) as well as P(djun) were subsequently determined using a primer extension assay. The activity of transposase OrfAB of ISLC3 was also assayed using an in vitro system. It was found that this transposase preferred to cleave a single DNA strand at the IRR over the IRL end in the transposition process, suggesting that attack of one end by the other was oriented from IRR to IRL.

Exploring the P2 and P3 ligand binding features for hepatitis C virus NS3 protease using some 3D QSAR techniques.

Several three-dimensional quantitative structure-activity relationship (3D-QSAR) models have been constructed using the comparative molecular field analysis (CoMFA), comparative molecular similarity indices analysis (CoMSIA), and catalyst pharmacophore feature building programs for a series of 26 truncated ketoacid inhibitors designed particularly for exploring the P2 and P3 binding pockets of HCV NS3 protease. The structures of these inhibitors were built from a structure template extracted from the crystal structure of HCV NS3 protease. The structures were aligned through docking each inhibitor into the NS3 active site using program GOLD. The best CoMSIA model was identified from the stepwise analysis results and the corresponding pharmacophore features derived were used for constructing a pharmacophore hypothesis by the catalyst program. Pharmacophore features obtained by CoMFA and CoMSIA are found to be in accord with each other and are both mapped onto the molecular 5K surface of NS3 active site. These pharmacophore features were also compared with those obtained by the catalyst program and mapped onto the same NS3 molecular surface. The pharmacophore building process was also performed for 20 boronic acid based NS3 inhibitors characterized by a long hydrophobic side chain attached at position P2. This latter pharmacophore hypothesis built by the catalyst program was also mapped onto the molecular surface of NS3 active site to define a second hydrophobic feature at position P2. The possibility of using the pharmacophore features mapped P2 and P3 binding pocket to design more potent depeptidized NS3 inhibitors was discussed.

Predicting the MHC-peptide affinity using some interactive-type molecular descriptors and QSAR models.

The ligand-receptor interaction between some peptidomimetic inhibitors and a class II major histocompatibility complex (MHC)-peptide presenting molecule, the HLA-DR4 receptor, can be modeled using some 3D quantitative structure-activity relationship (QSAR) methods such as the comparative molecular field analysis (CoMFA) and some molecular descriptors using the Cerius2 program. The structures of these peptidomimetic inhibitors can be generated theoretically, and the conformations used in the 3D QSAR studies can be defined by aligning them against the known structure of HLA-DR4 receptor through a least-square fitting procedure. The best CoMFA models can be constructed using the aligned structures of the best fitting result. The principal components analysis (PCA) module of the Cerius2 program can be used to trim outliers of the CoMFA columns generated. Procedures for a direct QSAR analysis using the Cerius2 descriptors and regression analysis by the genetic function module are also presented

Modeling the ligand-receptor interaction for a series of inhibitors of the capsid protein of enterovirus 71 using several three-dimensional quantitative structure-activity relationship techniques.

The structure of enterovirus 71 (EV71) capsid protein VP1 has been constructed by using homology modeling and molecular dynamics simulation techniques. The ligand structures were a series of EV71 VP1 inhibitors synthesized by Shia et al. in 2002 and Chern et al. in 2004. The training set was selected by the VOLSURF4.1/PCA program and the IC50 values varied from 0.06 to 10.83 microm. Then, the training set was analyzed by the following three-dimensional quantitative structure-activity relationship techniques: CoMFA, CoMSIA, CATALYST4.9, and VOLSURF4.1/PCA. The model generated by a two-stage flexible docking procedure and without any structural alignment has far more significant statistics. Highly accurate activities for the test sets were then predicted by the top hypothesis of the CATALYST program and were compared with those predicted by CoMFA, CoMSIA, and VOLSURF. These studies identified some important clues for searching or making more potent inhibitors against the EV71 infection.