MOLECULAR MODELING: INDISPENSABLE TOOL AT THE INTERFACE BETWEEN STRUCTURAL ANALYSIS AND MOLECULAR DESIGN

Christian Lehmann; Institute of Organic Chemistry, University of Lausanne, BCH-Dorigny, CH-1015 Lausanne, Switzerland

In view of the increasingly growing structural information available in protein as well as small molecule databases (Protein database [1] over 14'000 structures; Cambridge Crystallographic Database [2] over 200'000 entries), the visualization, understanding and treatment of these data by molecular modeling and structure analysis tools become ever more important. The potential of this information for the design of novel compounds is enormous, and by a number of selected structural paradigms, I will illustrate that the laws which govern 3-dimensional conformation are universal - no matter whether we are dealing with macromolecular or low molecular weight compounds. Comparisons of structures in their X-ray crystal or NMR derived solution conformation can be easily carried out with various superposition routines in a molecular modeling suite made recently available for the academic as well as industrial scientific community [3]. Pharmacophore modeling is enhanced by inclusion of a fast 3-d model builder, and conformational trajectories can be generated and analyzed with the ultimate goal to develop intrinsic molecular properties into enzyme inhibitors for therapeutic evaluation. De novo designed constitutions are amenable to examination by force fields calibrated against a large number of experimental structures. Important physical parameters such as the molecular dipole moment can be calculated and are congruent with experimental data. Starting from deductive molecular modeling to optimally reflect experimental observation, inductive molecular modeling will be described in fields as divergent as the elaboration of transcription factor mimetica from the development of new materials.

[1]  Protein Database: http://www.rcsb.org/pdb
[2]Cambridge Crystallographic Data Centre: http://www.ccdc.cam.ac.uk resp. its Zürich mirror:  http://sixray.ethz.ch/ csdeth.html
[3]  P. R. Gerber: MOLOC - A Molecular Design Software Suite: http://www.moloc.ch

Back

REPRESENTATION AND SEARCHING OF CONFORMATIONALLY FLEXIBLE MOLECULES BY EXTENDED CAST METHOD

Hiroko Satoh¹, Hiroyuki Koshino², Tadashi Nakata²;  ¹National Institute of Informatics and JST-PRESTO, ²RIKEN

Stereochemical information is important to treat molecular structures. Exact structure search with configurational and conformational information is essential for a database system including three-dimentional structural information. Computer-assisted synthetic design, reaction prediction, and NMR chemical shift prediction studies need to discriminate similarities and differences in stereochemistry around specific sites, such as strategic or reactive sites, and nuclei where NMR chemical shifts are to be predicted. Some coding methods for three-dimensional chemical structures have been developed and used for structural representation in some chemical software, however, an essential problem of exact search of structures considering configurational and conformational structural environments has been remained.

We developed a CAST (CAnonical-representation of STereochemistry) coding method (J. Chem. Inf. Comput. Sci., 40, 622-630 (2000); J. Chem. Inf. Comput. Sci., 41, 1106-1112(2001)) and extended the CAST method for structure search with stereochemical information of a database. Twelve types of CAST code are defined to twelve areas of the dihedral angle, where 360 degrees are equally divided into twelve areas and CAST codes are defined to the areas as the first two letters of numbers zero to eleven. For example, a CAST code of tw, si, or te is assigned to each atom in the staggered conformation and ze, fo, or ei is assigned to that in the eclipsed conformation. The basic CAST gives canonical codes based on the dihedral angles that are uniquely defined by a molecular tree structure. Two types of CAST notations for configurational and conformational structures can be flexibly used depending on the users together with CANOST notation for a planar structure. The extended CAST gives representation for structural search around a specific site. Using the CAST method, information on absolute/relative configurational and conformational, can be correctly retrieved from a database. The search by the CAST method is especially useful for getting partial structures with stereochemically similar environments from a database.

We are constructing a three-dimensional database using the CAST method for the structural representation, mainly including terpenoids, steroids, polycyclic ethers, polyketides, and some other organic natural compounds. Synthetic compounds are also used for construction of the database. Using the configurtational CAST, search of configurationally similar structures has been exactly and easily carried out, even for conformationally flexible compounds and for molecules having sequential stereocenters. Using the conformational CAST, search of conformationally similar structures has been also carried out. The similarities of conformation can be flexibly specified by the degree of the dihedral angles for definition of the CAST codes.

The database was applied to a database-oriented NMR chemical shift prediction, which requires exact representation of stereochemical information, and a CAST/CNMR system for accurate prediction of 13C-NMR chemical shifts by considering stereochemistry has been developed. The prediction is performed by matching with the database considering structures from an atom to be predicted to arbitrary level specified by the user. In the present study, structural environments including relative stereochemistry to g-level are mainly considered. Conformational information can be also used for the prediction. Predictions of 13C-chemical shifts by CAST/CNMR were examined for some organic natural products, and highly accurate 13C-NMR chemical shifts were predicted for them by considering stereochemistry.

The paper describes the CAST coding method and three-dimensional structural search for both conformationally rigid and flexible compounds. Results of the CAST/CNMR system are briefly demonstrated.

Back

EXPLORING THE ACHE GORGE WITH GALANTHAMINE-DERIVATIVES

G. Fels, E. Linnemann, E. Luttmann, C. Pilger;  University of Paderborn, D-33098 Paderborn, Germany

Acetylcholinesterase (AChE) plays a key role in the development of Alzheimer`s disease as this enzyme is responsible for cleavage of the neurotransmitter acetylcholine (ACh), and, according to recent investigations, also promotes aggregation of b-amyloid peptides, which causes plaque formation in synaptic areas of cholinergic neurotransmission. While the catalytic triad, which is responsible for ACh-cleavage, is located 20 Å deep in the enzyme at the end of narrow gorge, the b-amyloids seem to attach to the peripheral anionic site (PAS) centered around amino acid tryptophan (Trp279) at the mouth of the gorge.

We have investigated the potential of galanthamine, and some derivatives of this alkaloid with varying N-substituents, to interfere with both these AchE functions. Using three docking/scoring protocols (AutoDock, FlexX, QXP) we have independently derived at a correct predictions of the orientation and conformation of galanthamine in the active site of AChE from Torpedo californica (TcAChE). The resulting RMS deviation is about 0.5Å with respect to the crystal structure of the TcAChE-galanthamine complex, as published by C. Bartolucci and D. Lamba (International Centre for Genetic Engineering and Biotechnology (CNR), Rome and Trieste, Italy) and independently by the J. Sussman group (Weizmann Institute of Science, Rehovot, Israel).

The AchE-galanthamine complex structure was used as the starting point for the development of new galanthamine derivatives with improved inhibitory capacity towards esterase cleavage. To this end we have applied our docking strategies to galanthamine derivatives that carry a long side chain on the nitrogen atom. Particularly a piperidinopropyl-galanthamine was investigated in detail using QXP and FlexX and the results were compared to the crystal structure of the complex, which was characterized by the Lamba group (CNR, Rome and Trieste, Italy). The resulting structure has an almost identical orientation with respect to the rigid galanthamine ring system. The N-alkyl sidechain, however, does not show the expected extension of the N-substituent towards the upper part of the gorge, but rather folds back onto the galanthamine body yielding a scorpion like shape. Comparison of our docking result with the corresponding crystal structure again shows a close relationship of the predicted structure with the crystal complex.

Our galanthamine docking studies, furthermore, disclose a second binding site for galanthamine which is centered around tryptophan 279 at the peripheral anionic site (PAS). Based on these findings, we have performed a molecular modeling study to investigate bis-galanthamine derivatives that have two galanthamine moieties connected by a methylene spacer of varying length as dually acting AChE-ligands. The computational procedure involves docking of various bis-galanthamines in the AChE gorge and scoring functions that not only reflects energy criteria but also evaluate the hydrophobic interaction of the ligands particularly with the indole ring of Trp279. In addition we have compared the hydrophobicity values with geometrical parameters that describe the orientation and conformation of bis-galanthamine fragments interacting with the indole ringsystem.

Our results propose that these ligands indeed can simultaneously interfere with both biological functions of the enzyme, the neurotransmitter cleavage and the plaque formation and should therefore serve as the basis for a further development of bis-functional Alzheimer drugs.

Back

Last updated 21 January, 2005