National health-care expenses and R&D expenditure of pharmaceutical industry have risen in Japan, while the economic growth and the extension of total sales of pharmaceuticals have shrunk in recent years. On the other hand, the amount of the sales and potentiality of discovering innovative drugs in the western industry are growing year by year. Most recently, some of the domestic companies have started to discover the innovative drugs using genome technology. Thus, these circumstances surrounding the pharmaceutical industry are drastically changing in Japan. These situations will be described in this paper. Generally speaking, pharmaceutical information is generated from broad fields which start from drug-discovery research to post marketing surveillance (PMS). Thus, we can call " information accumulated compound " to the drugs. Among them, the PMS information is usually called "drug information", which is the most important knowledge for the patient's safety during the drug use. Here, the speaker will survey the activities of the drug information in Japan. >From the viewpoint of the medicinal chemistry, the package inserts-information including the PMS would be useful knowledge for the drug design. The speaker will talk about a concept "Spiral Circulation of Drug Information" for the drug design.

View full abstract

The fundamental questions a chemist is faced with are: 1) What structure do I need for a certain property? (Structure - Property Relationships), 2) How can I synthesize this structure? (Synthesis Design), 3) What is the structure of my reaction product or of the compound that I have isolated in the environment? (Structure Elucidation). Chemists have addressed these problems by learning from prior experience, by making a series of experiments and then ordering their observations into a mental framework, making a model about the features that are important for a given system and then make predictions by analogy. In recent years powerful computer methods have been developed to mimic the information processing of the human brain by artificial neural networks. Neural networks need experimental data to be analyzed in an inductive learning process.

View full abstract

Having a huge amount of genome sequence data, we have aimed to analyze the data and provide reliable information extracted for biological scientists. Focusing on the protein structure prediction, we have so far analyzed 80 organisms whose complete genomes were sequenced. The main tool employed was PSI-BLAST, a homology search method much more powerful than the conventional ones such as FASTA/BLAST. The fraction of all ORFs in a genome, predictable of their 3D structure, turned out to be as high as 40 - 50 %. All the data analyzed were compiled in a database called GTOP (http://spock.genes.nig.ac.jp/~genome/gtop.html). As an application study of GTOP, a way to identify a significant number of pseudogenes in E. coli is also presented.

View full abstract

The rapid progress of computer and network technologies makes it easy to collect and store a large amount of unstructured texts or semi-structured data such as webpages and HTML/XML archives. Thus, there are increasing demands for efficient tools for extracting useful information from large amount of semi-structured data. In this paper, we consider the problems of finding interesting patterns from large amounts of texts and semi-structured data, and present efficient data mining methods for the tasks. The topics to be discussed are: efficient text mining using optimized pattern discovery, information extraction from Web, and efficient substructure discovery from semi-structured data.

View full abstract

Physicochemical properties of solvent give significant influence on the equilibrium, rate, and yield of chemical reactions. While ago, we have proposed a statistical mechanics of molecular liquids (extended RISM theory), which can account for chemical specificity of solvent in molecular level, and have combined the theory with the ab initio MO theory to describe chemical reactions in solutions. We have applied the hybridized theory (RISM-SCF) to a variety of chemical reactions in solution including, acid-base equilibria, SN2, isomerization, and redox reactions to clarify solvent effect on those reactions. My lecture in the meeting concerns some of those reactions in solutions as well as the latest development made in our group.

View full abstract

Quantitative Structure Activity Relationship (QSAR) is widely studied in the field of Chemistry. In 3D-QSAR analyses such as comparative molecular field analysis (CoMFA), proper superimposition of molecules is required. Since appropriate superimposition is an important factor to construction of predictive model, various methodologies for molecular alignment have been proposed. In this study, the novel molecular alignment method using Hopfield Neural Network (HNN) is proposed. The alignment method was applied to two QSAR data sets for validation. First, the result of 3D-QSAR analysis of Cyclooxygenase-2 (COX-2) inhibitors which consist of three different types of skeleton, is reported. The structures of COX-2 inhibitors were aligned using our HNN method and analyzed by CoMFA method. The robust PLS model (R2=0.922, Q2=0.653) was obtained and it was validated by contour map of the regression coefficients and X-ray crystal structure of COX-2. Second, the result of 3D-QSAR analysis of human epidermal growth factor receptor-2 (HER2) inhibitors which consist of two different types of skeleton, is reported. The structures of HER2 inhibitors were automatically aligned using HNN and analyzed by CoMFA method. The robust PLS model (R2=0.805, Q2=0.701) was obtained and it was validated by contour map of the regression coefficients of the PLS model.

View full abstract

The correlation between the molecular electrostatic potential (MEP) on a benzofuran inhibitor and the N-myristoyltransferase (Nmt) inhibitory activity is investigated by three-dimensional quantitative structure-activity relationship (3D-QSAR) approach. To each compound, Kohonen neural network (KNN) is used to project the MEP values on Connolly surface into a two-dimensional map. Each node in the KNN map is coded by the associated MEP value of the occupying sampling point and all the coding nodes are collected together to define chemical descriptors. The best partial least squares (PLS) model with the selected descriptors is searched among the possible candidate models by genetic algorithm (GA). The resulting best model is significantly more predictive than the one with all chemical descriptors. The model has given the additional information about the binding of a benzofuran inhibitor and Nmt. *This study was complied and submitted to Bulletin of the Chemical Society of Japan.

View full abstract

Three-Dimensional Structure Characterization of Drug Molecules by Self-Organization-Map. Kentaro Hayashi(University of Tokushima), Yoneo Yano(University of Tokushima), Kazuhide Kanenishi(University of Tokushima), Chang Feng(University of Tokushima), Hiroshi Chuman(University of Tokushima), Because there is an enormous number of conformations, it is difficult to estimate the pharmacophore and to do the docking-study with the receptor using the 3D structure information that acquired from search of drug molecules' all stable conformation. So, there is the need to narrow down the comparison of conformation. In this research, we visualize the property-vector of the pharmaceutical's 3D stable conformation using the Self-Organization Map(SOM: Kohonen map) that is a kind of the neural network. The feature of SOM data mining is the ability to visualize correlation of many variables. Property-vector visualization enables a conformation classification and a discovery of the structural similarity. This time, we tried the our analysis to three kinds of HIV virus protease Inhibitors.

View full abstract

It is important problem in a protein science to detect active sites on protein structures. Recently some automatic detection methods were presented; finding a similar structure to known protein active sites, finding pockets and/or cavities on protein surfaces, and so on. We already presented a detection method for amino acid residues at function site using the similarity in 3-D structure of same function proteins. In that investigation two same functional proteins were compared and the amino acid residues including active site of the proteins were extracted. For the improvement of the finding method now we proposed to compare three or more same function proteins. We applied this method to three serine proteases of different subgroup and obtained the 10 amino acid residues as active site residues within reasonable time. The result indicates the effectiveness of this improved method.

View full abstract

The electron transfer and transport in DNA have recently attracted considerable interest in view of their possible use in molecular electronics. Through the ACT-JST project entitled "Ab Initio Approach to Nanoscale Dynamics in DNA" we have been developing theoretical and computational tools for describing the charge (electron/hole) transfer in DNA on the bases of molecular orbital, molecular dynamics and reaction dynamics calculations. Essential ingredients in the calculations include the electronic site energies, transfer integrals, and solvent reorganization energies. We have thus analyzed the charge transfer in DNA duplexes in aqueous solution and that in electrochemical DNA chip. In the former a considerable nuclear quantum effect on the rate constant has been identified in the inverted region in agreement with experiment. In the latter the possibility of electrochemically detecting the mismatched base pairs has been shown quantitatively. In the light of our analyses recent advances in DNA molecular electronics will be reviewed and its forthcoming developments will also be suggested.

View full abstract

This paper deals with the development of CD-ROM for referring and predicting physico-chemical property data like phase equilibria, solubilities and etc. The books "Physico-Chemical Properties for Chemical Engineering (edited by SCEJ)" have been used for obtaining the secondary informations. The numerical information has been inputted using the values in original references. The predicted part is based on the ASOG group contribution method. Now we can get the predicted results for low-pressure vapor-liquid equilibria.

View full abstract

An NMR spectral database for natural products, SDBS-NMR-NP, has been built as the development of the SDBS-NMR for general organic compounds. The chemical information, NMR data and the spectral assignments have been compiled from the literatures (Tetrahedron, Tetrahedron Letts., Magn. Reson. Chem., J. Antibiotics, J. Natural Products, J. Org. Chem., Eur. J. Org. Chem. and others). The input of chemical information is made by a ISIS/Base and the spectral information such as the shifts and J are compiled by a Microsoft Excel. The data are displayed on the web system where 13CNMR spectral patterns were shown by a Java applet. The intensities are generated from the number of the carbons for each line. The software to transform original FID patterns to the database-spectral format has been prepared for JEOL, Varian and Bruker instruments. At present, about 270 data were collected and stored in the database from the literatures in 2002.

View full abstract

Polyhedra, especially regular and semi-regular polyhedra are found in nature and artificial objects, such as architecture and designs, in which important basic concepts of chemistry and physics are involved. However in modern school education, students learn only a few things about polyhedra as a part of a unit of space objects. Moreover, in the new Course of Study settled by the government, 3-dimensional objects to be taught are greatly reduced. According to their explanation for this reduction is it is not easy for students to understand polyhedra only from textbooks and models. On the other hand, computers are very useful tools for learning polyhedra. A computer-aided-education system available on the World Wide Web (WWW) for studying polyhedra was developed. The system has the following characteristics: easy operation by the use of a mouse and ample menu icons, and potential transferability for extension and development in various fields, such as mathematics, chemistry, architecture, and industrial design. For improvement of the system, this software was tested in a few classes of university students of non-science major. The results are introduced.

View full abstract

Covalent bond of hydrogen is formed when one pair of valence electrons are shared between two hydrogen atoms, resulting in stable molecular orbital. In order to interpret covalent bond formation by scientific visualization, we developed a novel animation. It begins a display of two atomic orbitals of hydrogen atoms placed too far away to interact. As these two orbitals come near to one another, gradual formation of bonding molecular orbital (MO) is displayed. The MOs are drawn as a function of atomic distance, together with difference electron densities and the corresponding position on a potential energy curve. When the atomic distance comes near to the optimized geometry of hydrogen molecule, i. e., 74 pm, electron rich region between two nuclei of hydrogens is displayed, indicating minus charged electron cloud attracts plus charged hydrogen nuclei to form chemical bond. The CAVE is a projection-besed virtual reality system that surrounds the viewer with 4 screens (each 3m*3m). The animation developed here is also displayed in the CAVE to give stereo-views from various viewpoints.

View full abstract

According to the increase of concern about ecological environment, it is interested to calculate the yields of byproducts quantitatively. The program for exhaustive reaction networks was developed, where the concept of imaginary transition structure proposed by one of author was applied to generate main products and byproducts. In order to evaluate multiple reactions, program for time-evolving simulation was also developed.

View full abstract

We have been constructing a program system, called Data Base for Transition State Geometry, which fuses the computational chemistry and information chemistry. The program was intended to examine whether or not synthesis routes obtained from TOSP or AIPHOS, the synthesis route design system developed by Funatsu et al, is feasible. The program uses both the data in the system and new MO calculations to rank the synthesis routes by comparing activation energies and other information calculated from molecular orbital calculations. We tried to clarify that the data base is useful for developing new synthesis routes. In the present study, we performed MO calculations as well as experiments which related to a synthesis route of 2,3-substituted-furane-2,3-dione obtained from the TOSP system. The B3LYP/6-31G* level of calculations showed existence of a reaction path via aldol condensation. Chemics program and GIAO calculations were performed to analyze an intermediate from experiments for aldol condensation. The intermediate obtained was an unexpected product without a five-membered ring. The experimental conditions before beginning experiments should be revised according to the present experimental results.

View full abstract

The density-functional theory (DFT) for molecular fluids [J. Chem. Phys. 115, 6653 (2001)] is extended to the case of polymer liquids. A system consisting of the idea chains is employed as a reference system for the DFT, where many-body effects are considered as an effective field that acts on each site of the ideal chains. We derived a relation between the site-site pair distribution functions and the site-density distribution functions under a mean field arising from a single polymer molecule. An integral equation for the site-site pair distribution functions is obtained by the DFT, where the external field is taken to be the mean field. We propose an approximate expression of the intramolecular correlation functions for isolated single-polymer chains to take account for the excluded volume effects arising from intramolecular interactions. The intramolecular correlation function considering the intramolecular excluded volume effects was in qualitative agreement with simulated intramolecular correlation functions for freely jointed tangent-soft-core chain liquids. The site-density integral equation under the mean field, using the intramolecular correlation function, provides accurate predictions for the simulated site-site pair distribution functions of the freely jointed tangent-soft-core chain liquids.

View full abstract

Of all the theoretical concepts that constitute the rational basis of modern organic chemistry, the concept of aromaticity is one of the most general, but at the same time one of the most vaguely defined. In this conference of the last year, the aromaticity of six-membered rings with six π electrons was studied by CiLC (CI/LMO/CASSCF) analysis on the basis of ab initio molecular orbital methods. The aromaticity was defined for equivalent electronic states for each bond and the gap between the weights of the singlet coupling and polarization terms. These weights were calculated for the aromatic structure (D6h) of C6H6, Si6H6, B6, Al6, N6, and P6, and it was revealed that the aromaticity is closely related to the electronegativity of the composed atoms. Our definition of aromaticity corresponds to resonance concepts. In the present conference, the definition of the aromaticity is extended to more general compounds as C4H4, C8H8, and C10H10. The results obtained are related with the 4n+2 rule in Huchel molecular orbital theory.

View full abstract

In a water cluster, water molecules are connected by H-bonds and form a H-bond network. Given the number of water molecules, several H-bond patterns are possible to connect them and to form a cluster. By enumerating the directed graphs, possible H-bond patterns in water trimer is found to be 5. To understand thermodynamic properties of a system, we need to know a whole phase space of the system. At a finite temperature the trajectory locates in various places in the phase space. The whole phase space of water clusters is multidimensional, but each point can be classified according to the H-bond pattern to which the point attribute. The phase space can be divided into 5 areas, in the case of the water trimer. In the ensemble of water trimers at a temperature T, the elements distribute among the 5 areas according to T. Its distribution may vary depending on T and correspond to the ratio of the existence which may be observed. We adopted the Monte Carlo sampling method to sample points of the phase space of the water trimer. We executed this procedure at various temperatures, and assesed the ratio of the existence of water trimers.

View full abstract

One of the authors previously reported that gangliosides inhibited the activity of an enzyme NAD+ glycohydrolase (CD38), and that those with a tandem sialic acid in the sugar chain had the largest inhibitory effect. The objective of our investigation is to clarify the cause of difference in the inhibitory effect in terms of three-dimensional structures and electronic structures of gangliosides by computer simulation. We assessed the similarity of the calculated structures of NAD+ and tandem sialic acid by conformational analyses and molecular orbital method, on the assumption that CD38, an NAD+ reacting enzyme, cross-reacts with the tandem sialic acid in gangliosides. A strong correlation was found between the orbital energy of HOMO by PM3 calculations and the extent of the inhibitory effect. Solvation effects were considered to interpret the substrate recognition mechanisms in the biological system. ONIOM method applied to the gangliosides gave the results similar to those for the isolated molecules. A novel mechanism in substrate recognition of enzymes to the substrate is proposed.

View full abstract

The presence of solvents is essential in all steps of pharmaceutical processes (reaction, separation, and formulation). For toxicological reasons, drug manufacturers are increasingly required to minimize the number of solvents employed in pharmaceutical processes. In practice, however, the selection of solvents and antisolvents for pharmaceuticals mostly relies on experience and experimental testing. Our work aims to systematically investigate the solubility of pharmaceutical compounds and their intermediates in a wide range of solvents. We describe the development of solubility databases of mono- and bifunctional compounds that are based on the core molecular fragments of common drugs and containing substituents of gradually increasing polarity. The core fragments can be identified by a classification of drugs into pharmaceutical categories. Each of the functionalized solutes is studied in a series of pharmaceutical solvents of varying polarity and hydrogen bonding tendency. The utility of an example database is shown which uses the benzene ring as the core fragment and contains activity coefficient data of 13 monofunctional benzene derivatives in 17 solvents. The database was analyzed by the Modified UNIFAC (Dortmund) group contribution method and quantum mechanical/COSMO calculations. The testing shows the importance of polarization effects which are common between drug-like molecules and small polar solvents. The proposed approach can be useful to provide a systematic insight into solvent-solute interactions and for testing of new methods of solubility prediction.

View full abstract

Integrated Spectral Database System for Organic Compounds (SDBS) has a long history of development since 1981. The operation of SDBS inherited to National Metrology Institute of Japan, National Institute of Advanced Science and Technology (AIST) from 2001 when AIST was established. The SDBS system is opened to public through the Internet freely since April 1997. By the end of August 2002 the total number of access exceeded 35 million. The large number of accesses and the remarks from all over the world has proved the importance and necessity of SDBS. It is extremely important to maintain simple and easy data management in adding and replacing data for keeping the quality of the database. We have been developing a new input tool of the compound dictionary and NMR spectral data in order to simplify and reducing mistakes in the process of data input. Here we present outline of the new input tool developed with Microsoft Access along with Accord for Access from Accelrys. Furthermore, current activities of SDBS are also reported.

View full abstract

We have been constructing a proton NMR spectral database for over twenty years. The revised management system for the database has been developed and reported at the last conference. The database is now opened on WWW. In the system, however, only the minimum function required for maintenance of the database has been implemented. Then, we have developed various tools for maintenance work such as adding, updating, deleting files. These tools were developed using Java language. The first one is the tool which performs an addition and renewal data records by batch processing. It reads data files at the designated directory by command line, then updates the tables under the control of Oracle DBMS. Next tool has the interactive GUI and is used for addition, renewal, and deletion of a single record. This function actually works with a combination of the following two components, i. e., Java Applet which provide interactive GUI working on WWW browser, and Java Servlet which provide table updating service working on WWW server.

View full abstract

The UCSI (Unit Subduced Cycle Index) method has been applied to enumerate edge configurations on an octahedral skeleton, where we take account of multidentate ligands. The numbers of isomers are obtained in an itemized manner with respect to their formulae and point-group symmetries. The UCSI method provides a common tool to systematize inorganic and organic stereochemistry as well as to accomplish combinational enumerations of isomers.

View full abstract

AUTOCHEMMAPPER has been developed to map atoms corresponding between reactants and products. Species, unsaturation degree and equivalency of atom as well as the number of atoms adjacent to each atom are utilized to atom-to-atom mapping. Each two atoms equivalent between reactant and product are mapped to be a common atom in the course of reaction. Atoms equivalent in reactants or product simplify the mapping procedure. Starting atom to be mapped is selected by a priority rule and mapping is achieved from a starting atom to the adjacent atom along the bond until the bond terminates or meets an atom of different species or with different properties about unsaturation degree or the number of adjacent atom. In the latter cases, an algorithm is applied if the change is caused by a transform in the reaction, and the mapping of atom candidates from the starting atom along the bonds is continued or abondoned to restart mapping from the next starting atom. More than 99.5% correctly mapped scheme is achieved about ca. 6000 reaction data.

View full abstract

We have developed three different types of organic synthesis design systems (AIPHOS, KOST and TOSP), which are complementary each other. In this research an unified graphic user interface for these systems was provided to operate these systems flexibly. At the same time client/server type configuration was adopted. Server OS of this unified system is Linux, the corresponding client OS is MS-Windows. Multi-task and multi-user environment was realized.

View full abstract

The rate constants for the hydrogen abstraction reaction of hydrofluoroethers (HFEs) and hydrofluorocarbons (HFCs) with OH radicals at 298 K were estimated by using new empirical estimation method, which consists of two steps. In the beginning, we constructed an artificial neural network to calculate the bond dissociation enthalpies (BDEs) of the breaking C-H bond at 298 K by using topological descriptors defined only from molecular formula. Here we utilized the BDEs of 93 C-H sites, which are computed by (RO)B3LYP/6-311G** level, as training data. Next, the reaction rate constants were evaluated by using the estimated BDEs and a modified Heiklen's empirical equation. Consequently, the combination of these two empirical methods could estimate the rate constants of HFEs and HFCs with sufficient accuracy.

View full abstract

In semiconductor industries, CVD (Chemical Vapor Deposition) is used as one of the most important ultra-fine manufacturing processes. The technique of evaluating and developing the new CVD processes rapidly and costlessly is strongly demanded. Therefore, we developed the novel system, which analyzes the experimental data of various reactors in the CVD processes in integration and identifies the reaction model of them automatically using GA (Genetic Algorithms). The data consisted of both the profiles of the thickness of the deposited films and the detailed structures in the minute domains of the films. While the detailed structures were reproduced by Monte Carlo simulations, the enormous costs of calculations obstructed analysis of the data using GA. Therefore we modeled the correlation between the sticking probabilities of the deposition species and the stepcoverage, which is the typical index of the detailed structure of the films, and replaced the simulations in the system with the modeled functions. We could reduce the calculation time dramatically and make the system practical.

View full abstract

The chemometrics software package "Chemish" has been originally developed. The fundamental calculation method of chemometrics such as MLR, PLS, QPLS and PCA are incorporated to this package. This package also includes methods of PLS and QPLS with variable selection enhanced by genetic algorithm and a method of MLR with variable selection enhanced by STEPWIZE. Furthermore, neural networks, hierarchical cluster analysis and quality control are available in Chemish. For back propagation the number of neurons of hidden layer and the number of learning can be optimized automatically by cross validation. In addition, reverse analysis of given models is possible by using Chemish. The user-interface of Chemish is focused on user-friendly operation. Thus, data input, display of the results, and procedure of analysis are smooth and easy. Finally, the tutorial which assists the user to understand how to operate Chemish, has been also developed. Chemish will be demonstrated on notebook-type personal computer.

View full abstract

Various programs that support the design of new drug structure heretofore, have been developed. LigBuilder were developed as one of the programs for structure-based drug design. Under the structural constraints of the target protein, LigBuilder can build up ligands step by step using a library of structural fragments. However, the problem is the case that the information on the pocket of the target protein is not given. In many cases, we have no information on the structure of the target protein in developing new drugs. Thereupon we propose a new method that can constructs new drug structure, by connecting appropriate fragments based on an initial fragment structure. First, an initial fragment structure is given and a genetic algorithm that is one of an optimization method is used in mechanism of selection and connection of fragment structure to complete drug structure. At this time structure-activity relationship models by CoMFA etc. as the evaluation criteria. The overview of program for automated construction of drug candidates and the application of this method to the actual data are reported.

View full abstract

Representation of stereochemical environments by FRAU system is investigated. All chair form conformers of diastereomers of cyclohexanol, 1,2-cyclohexa-diol, 1,3-cyclohexa-diol, 1,4-cyclohexa-diol, and quercitol are used as model structures. FRAU system characterizes a molecule based on electrostatic and steric interactions with pseudo-reactant and gives three kinds of FRAU's features (FF) representing electrostatic interactions (FF-electro), steric interactions (FF-steric), and extent of reaction field (FF-field). In the study, correlation between FF-steric and FF-field for oxygen atoms in the model structures and their stereochemical environments are analyzed. Oxygen atoms at equatorial and axial positions formed different clusters. In the cluster for axial positions, three clusters were found for mono-axial, 1,3-diaxial, and 1,3,5-diaxial positions, respectively. In each of the cluster, oxygens are classified based on more detailed similarities and differences of stereochemical environments. Furthermore, distances between each of the oxygen atoms and the other atoms in the molecules are plotted are referred in the analyses, and it shows that the FF values represent similarities and differences of stereochemical environments about within 4 A from the oxygen atoms. The results demonstrate good correlation between FF values and stereochemical environments.

View full abstract

Dibenzofuran is one of the most dangerous organic chlorinated compounds. Effect on the human health is serious problem now. There are 135 isomers of dibenzofuran. The structures of isomers should be correlated with toxicity. Here we report the theoretical estimation of toxicity of dibenzofuran by computational calculations. Frontier orbital densities by ab initio calculations are used to perform multivariate statistical analysis. Principal component analysis predicted that 136 molecules can be classified by their scores. We have found that the fifth principal component strongly correlated with toxicity.

View full abstract

The study of weakly bound molecular complexes is a useful step toward gaining a basic understanding of the dynamic behaviour of molecules in various environments. The Van der Waals complexes of molecular species bound to rare gas atoms have been of interest as the atom can be regarded as a structureless probe for the anisotropy of intermolecular potentials. Recently, the experimental detection and characterization of Van der Waals trimers is a subject of growing interest. In the present study, high-level ab initio calculations [MP2] using correlation consistent basis sets of Dunning [aug-cc-pVTZ] were carried out to study the structures and stability of Ar2/NH3 Van der Waals trimer. For the calculation of interaction energies, the basis set superposition errors (BSSE) were estimated by counterpoise method. One minimum isomer was found and characterized. The most stable isomer exhibits a X-shaped geometry, in which both the bond axis of Ar2 molecule and the C3 axis of NH3 molecule are perpendicular and in which a value of about 3.2 angstrom is obtained for the Ar2-NH3 distance. The details of the calculated potential energy surface and the calculated vibrational frequencies are also discussed.

View full abstract

Fullerene molecules are spherical in shape. Higher fullerenes have larger ratios of six- to five-membered rings. Therefore, they must be similar in electronic structure not only to electron gas on the surface of a sphere but also to a graphite sheet. We analyzed Hueckel pi-orbital energies of many fullerenes from this viewpoint. For example, C100 fullerene has 450 isolated-pentagon isomers. Distribution of all pi-orbital energies for all these isomers represents a kind of average density of states (DOS) for C100. The average DOS was found to be very small near the energy of alpha. This reflects the electronic structure of graphite, being compatible with the fact the average HOMO-LUMO gap of C100 fullerene isomers is ca. two times as large as that expected by assuming that all pi-orbitals are arranged with equal spacings. There are several distinct peaks on the low-energy side of the average DOS, which correspond to the energy levels of free electrons on the surface of a sphere. C108 Leapfrog isomers were found to have a band gap in the average DOS. It is now clear that pi-electrons in fullerene molecules behave not only like those in graphite but also like free electrons on the spherical surface.

View full abstract

The series of MO electronic spectral calculations at different level of theory (ZINDO/S, TD-DFT, TD-DFT with SCRF) are presented for the merocyanine dyes newly synthesized via the reaction of N,N'-dialkylanilines and other nucleophilic reagents with 4-methythiomalemides which serve as a versatile building block in heterocyclic synthesis, in aiming to theoretically characterize their first intense π-π* absorption maxima in the visible region which plays a decisive role in their color appearances. The introduction of cyano- and carboxymethyl group to 4-position of maleimide ring - in spite of their opposite electronic substituent effects - results in a moderate bathochromic shift. For some of these compounds, thioimidation to one of the two carbonyl group leads to the corresponding new blue dyes with the considerable bathochromic shift of more than 70 nm from original ones, shifting their absorption maxima into exceeding 650 nm wavelength regions. Our TD-DFT calculations combined with SCRF model qualitatively reproduced this substantial red shift although somewhat notable discrepancies (in the range of 52 to 88nm) still remained between theoretical absorption peaks and experimental ones. Concerning the deviation, the general performance of TD-DFT is discussed for the spectral calculations of merocyane class dyes.

View full abstract

The regioselectivities in [2+2]photocycloadditions of 2-cyclohexenone(1) to methyl cycloalkenecarboxylates(2-4) give drastic different experimental results depending on the ring size of the cycloalkenecarboxylates, that is to say that a gradual reversal of regioselectivity from head-to-head(hh) to head-to-tail(ht) adducts with the increasing ring size. We have located the roots of the regioselective difference by performing the frontier molecular orbital(FMO) and transition state(TS) analysis using MOPAC. We could explain that for methyl cyclobutenecarboxlate(2) TS energy differences of the second-step and back-reaction as well as the first-step has effects for the regioselectivities and for methyl cyclohexenecarboxylate(4) TS energy of the first-steps is the most important factor deciding the regioselectivities. In addition, We also report the results of the ab initio calculations in these TS geometrical optimizations.

View full abstract

A new method for flexible docking in explicit water using multicanonical and annealing algorithms was developed. The conformational sampling by the multicanonical molecular dynamics from 400 to 700 K was carried out, and then the conformation at 300 K was obtained by the simulated annealing from 400 to 300 K. The method was applied to the lysozyme and tri-N-acetyl-D-glucosamine (NAG) complex. We picked 127 conformations from the canonical ensemble at 400 K, and the conformational ensemble at 300 K was obtained by the 500-ps annealing and sampling simulations. Four clusters, A - D, were found in the ensemble at 300 K. The conformations of tri-NAG in cluster A exhibited the similar conformations to the X-ray structure (root-mean-square deviation = 2 - 4Å).

View full abstract

Oligonucleotides containing phenylsulfide derivative of 2-amino- vinylpurine nucleoside analogue are activated within duplex to form cross-linking toward cytidine selectively at the target site. It was confirmed that this reaction does not proceed in organic solvent. We have already studied the mechanism of the reaction by using the model system and our theoretical calculations at the B3LYP/6-31G* level of theory is consistent with the experimental results, i.e., the activation energy of the reaction was calculated to be more than 30 kcal/mol and the reaction hardly proceed under normal conditions. Therefore, the geometry of the DNA duplex or the electronic effect of the base pair of 2-amino-vinylpurine is strongly related to why the cross-linking occurs in the duplex. In the present study, we investigated theoretically effects of the base pair and the steric effect of the duplex for activation of the phenylsulfide fragment. It was confirmed that the existence of the base pair does not change the activation energies very much for the activation process of the phenylsulfide fragment. Therefore, the steric hindrance in the duplex is important for the cross-linking in DNA.

View full abstract

We have newly developed QM/MM (Quantum mechanical/Molecular Mechanics) simulation system for bio-macromolecules. The system is an extension of AMOSS (Ab-initio Molecular Orbital System for Supercomputers), now combined with classical molecular dynamics (MD) simulation programs, with proper account of the polarization effect. In order to smoothly link various programs developed in different fields, we propose UDS (Universal Data Sets) protocols. As an application, we performed QM/MM calculations on an influenza virus enzyme complexed with its inhibitors.

View full abstract

DNA-binding proteins recognize specific base sequences within DNA. Although the experimental structural analyses have presented valuable structural data on DNA-protein interactions, the molecular mechanism of the DNA-protein recognition is not well understood yet. To examine the origin of the specificity, we calculate the interaction free energy (ΔΔG), the interaction enthalpy (ΔΔH), the interaction entropy (TΔΔS) and the minimum interaction energy (ΔΔEmin) landscapes for base sequences (GAG and GAA) with Asn side chain and their components from each base pair using the ab initio potentials, by means of extensive conformational sampling. Ab initio potentials have the parameters which reproduce well the interaction energies calculated with the ab initio MO method. It is found that the effect on the interaction between the base sequences (GAG and GAA) with Asn from the GA sequence is bigger than that from the other base sequence (AG or AA). This indicates that the interaction of Asn with DNA depends on not only the base pair but also the base sequence.

View full abstract

The Fragment MO method combined with the Density Functional Theory was proposed. In this method, we used the exchange functional of the Perdew-Wang 91 with adjustable parameter. For investigating the accuracy of this method, the total energies of α-helix and β-strand glycine peptides were calculated by both the FMO/DFT and conventional DFT methods.

View full abstract

To elucidate a common feature of OmpR family, we investigated stable structures and electronic properties of DNA binding sites of OmpR proteins by semi-empirical MO method. The results clarify that the HOMO is localized on the 2nd aspartic acid for almost all families. This amino acid exists near the backbone of DNA, when OmpR binds to DNA. Therefore, it seems that the 2nd aspartic acid of DNA binding site of OmpR family is essential for the binding between OmpR and DNA. The effect of amino-acid mutation on the structures and electronic properties was also invesgated.

View full abstract

We studied two subjects associated with orbital resonance energy ORE for monocyclic conjugated systems. One of them is the alternation in signs of ORE values, which is necessary for the validity of Huckel rule for ionic conjugated molecules. The other is the relationship between topological resonance energy TRE and relative hardness. Bonds which do not participate in cyclic conjugation of pi-electrons are called "acyclic" bond. For various types of monocyclic conjugated molecules with "acyclic" bonds, which were classified into five types A - E, the alternation in signs of ORE values were studied and the following results were obtained; the alternation in signs of ORE values is valid for annulenes, radialenes and monocyclic conjugated molecules belonging to group D and not valid for monocyclic conjugated molecules belonging to groups C and E. Next, the relationship between topological resonance energy TRE and relative hardness were discussed for alternant monocyclic conjugated molecules. The following results were obtained; there are good correlations between TRE and relative hardness for alternant annulenes, [4n]-radialenes and monocyclic conjugated molecules belonging to group E, but there are no good correlations between TRE and relative hardness for [4n+2]-radialenes and monocyclic conjugated molecules belonging to group D.

View full abstract

Carbon nanotube has been a matter of great interest since the discovery in 1991 by Iijima. It has been predicted that the electronic feature such as band gap depends on the diameter and chirality of the tube. It was based on the tubes of semi-permanent length. A lot of works on nanotubes having no chirality, i.e., armchair-type and zigzag-type tubes, have been reported. We have, however, little information concerning the chiral nanobubes with the chiral angle between 0 and 30 degrees. In this paper we present computational results on the electronic structure of several series of chiral nanotubes of single layer with open edges. Total energies, band gaps, dipole moments, and frontier molecular orbitals are compared and discussed the dependency of the chiral angle and the tube length. The computational methods used in this study are Huckel, Extended Huckel, and PM3 methods.

View full abstract

Caged (photolabile) compounds consist of a biologically active molecule and a light-sensitive protective group. Upon irradiation with light of appropriate wavelength, the biologically active molecule itself will be isolated. It is important to control the wavelength or energy of light in order to design functional molecules. In this paper the substituent effect is discussed on the photodissociation process of o-nitrobenzyl-based type photosensitive phosporamide mustard. The maximum absorption wavelength max with oscillator strengths of several substituted caged phosphoramide mustards were calculated for the substiuents, OMe, OH, NH2, COOH, and glycine by semi-empirical CNDO/S CI, ab initio CIS, and TDDFT calculations. The results of TDDFT calculations are in good agreement with the experiment. A type of di-substitution and two types of mono-substitution in aromatic ring were examined from the viewpoint of the position of the substitution and the kind of substituents. The characteristic of the molecular orbitals will be also discussed.

View full abstract

Fuji Xerox' MD Engine II (MDE-II) is a high performance accelerator board which was designed to speedup for molecular dynamics simulation by exclusively computing the non-bonded interactions of the target molecule or ensemble. In order to improve the performance of conformation search for extra-large and flexible molecules, MDE-II was introduced to our conformational space search method CONFLEX. Benchmarks for conformation searches of polyalanine octamer to 98-mer suggested that while the maximum performance of the CONFLEX/MDE-II calculations in the non-bonded interactions is almost 4.3 times faster than those of the normal CONFLEX calculation without MDE-II, the speedup for the total energy calculation using the CONFLEX accelerated by MDE-II reaches the only 2.3 times. This inefficiency is caused by the difference of the computation times per one interaction between non-bonded pairs and the others combinations.

View full abstract

In order to elucidate the possibility of hydrogen transfer in the cyclic conjugated hydrogen-bonding system of malonaldehyde enols (the n-meric cyclamers of malonaldehyde, CMAn with n=3-10) DFT calculations were performed using B3LYP functionals with the 6-31G** basis sets. Two stable geometries with forms I and II about the hydrogen bonds in the cyclamers are defined as "[syn-anti]anti" and "[anti-syn]anti," respectively, on the basis of the CC-OH and CC=O...H configurations and the CO...OC conformation. Smaller cyclamers prefer form I, whereas larger ones (n>8) prefer form II. The energy difference between two forms is the smallest for CMA8, where the barrier height for the proton transfer is 9.06 kcal/mol, suggesting that the proton transfer should be possible. In order to probe the accuracy of the present calculations the geometry and the C-13 NMR chemical shifts for the CMA6 clathrate with benzene were examined. Comparison the calculated values with the corresponding experimental ones for the hexameric cyclamer of 1,3-cyclohexanedione enclathrating a benzene molecule yielded satisfactory results.

View full abstract

Since many organic reactions proceed in solvent, it is important to determine theoretical sigma + values for solution phase. For this purpose, it is necessary to know appropriate solvation models which reproduce the stabilities for the standard system of t-cumyl cations in solution. In this point of view, stabilities of t-cumyl cation coordinated by water molecule(s) (n=1-4) at cationic center were determined by means of hydride transfer reactions between ring substituted and unsubstituted cations. Energies for all species were calculated at the MP2/6-31G*//RHF/6-31G* level using Gaussian 98 program. The Yukawa-Tsuno analyses for the stabilities of hydrated t-cumyl cation with theoretically determined gas phase sigma values gave excellent linear correlations. While the r value didn't change so much, the absolute rho value decreased significantly with the number (n) of the water molecule down to 70% for the n = 4 system. When the substituent effect for this system was correlated with Brown and Okamoto's sigma + which is determined in solvolysis, deviations from the correlation line were observed for m-MeO and p-OH, indicating that specific solvation to the particular substituents is also important as a factor of appropriate solvation model.

View full abstract

In order to elucidate the resonance effect between boron cation and aryl group, the substituent analysis was carried out on the basis of the equilibria of hydride transfer reaction. The optimized structures and energies of all boron species were calculated at B3LYP/6-31G* and HF/6-31G* levels by using Gaussian 98. The substituent effect on the equilibria were correlated with only the Yukawa-Tsuno equation. The results showed that the electronic demand, r values, become small with the increase of stability of boron cation and are less than one-half the r values of the corresponding carbocations.

View full abstract