The tryptophyl-glycine (Trp-Gly) and tryptophyl-glycyl-glycine (Trp-Gly-Gly) peptides have been studied by means of molecular dynamic simulations combined with high-level correlated ab initio quantum chemical and statistical thermodynamic calculations. The lowest energy conformers were localized in the free energy surface. The structures of the different Trp-Gly and Trp-Gly-Gly conformers coexisting in the gas phase have been for the first time reported and their scaled theoretical IR spectra unambiguously assigned and compared with previous gas-phase experimental results. Common geometrical features have been systematically observed for the sequence Trp, Trp-Gly, and Trp-Gly-Gly. In addition, the peptide backbone of Trp-Gly-Gly has been compared with that of the previously studied Phe-Gly-Gly (Reha, D. et. al. Chem. Eur. J. 2005, 11, 6803). From the observed systematic structural behavior between these peptide analogues, it is expected that the gas-phase conformers of other similar aromatic small peptides would present equivalent geometries. The DFT methodology failed to describe the potential energy surface of the studied peptides since the London dispersion energy (not covered in DFT) plays a significant role in the stabilization of most stable conformers.

The geometries of DNA hexamer (5'-GGAACC-3') and DNA 13-mer (5'-GCGTACACATGCG-3') have been determined by molecular dynamics (MD) simulations using an empirical force field. The central canonical base pair was replaced by a pair of nonpolar base analogues, 2,2'-bipyridyl and 3-methylisocarbostyril. The stabilization energy of the model system (model A) consisting of a central base pair (base-analogue pair) and two neighboring base pairs was determined by the RI-MP2 method using an extended aug-cc-pVDZ basis set. The geometry of the model was averaged from structures determined by MD simulations. The role of the solvent was covered by the COSMO continuum solvent model and calculations were performed for a larger model system (model B) which also contained a sugar-phosphate backbone. The total stabilization energies of the unperturbed system and the system perturbed by a base-analogue pair (model A) were comparable to the stability of both duplexes experimentally determined. This is due to large stacking interaction energy of the base-analogue self-pair which compensates for the missing hydrogen-bonding energy of the replaced adenine...thymine base pair. The selectivity of the base-analogue pair was reproduced (model B) when their desolvation energy was included with the interaction energy of both strands determined by the approximate SCC-DFTB-D method.

We investigated the potential-energy surface (PES) of the phenylalanyl-glycyl-glycine tripeptide in the gas phase by means of IR/UV double-resonance spectroscopy, and quantum chemical and statistical thermodynamic calculations. Experimentally, we observed four conformational structures and we recorded their IR spectra in the spectral region of 3000-4000 cm(-1). Computationally, we investigated the PES by a combination of molecular dynamics/quenching procedures with high-level correlated ab initio calculations. We found that neither empirical potentials nor various DFT functionals provide satisfactory results. On the other hand, the approximative DFT method covering the dispersion energy yields a reliable set of the most stable structures, which we subsequently investigated with an accurate, correlated ab initio treatment. The global minimum contains three moderately strong intramolecular hydrogen bonds and is mainly stabilized by London dispersion forces between the phenyl ring, the carboxylic acid group, and various peptide bonds. A proper description of the last type of interaction requires accurate correlated ab initio calculations, including the complete basis set limit of the MP2 method and CCSD(T) correction terms. Since in our beam experiments the conformations are frozen by cooling from a higher temperature, it is necessary to localize the most stable structures on the free-energy surface rather than on the PES. We used two different procedures (rigid rotor/harmonic oscillator/ideal gas approximation based on ab initio characteristics and evaluation of relative populations from the molecular dynamic simulations using the AMBER potential) and both yield four structures, the global minimum and three local minima. These four structures were among the 15 most energetically stable structures obtained from accurate ab initio optimization. The calculated IR spectra for these four structures agree well with the experimental frequencies, which validates the localization procedure.

Potential energy surfaces of monohydrated and dihydrated adenine-thymine and 9-methyladenine-1-methylthymine base pairs were examined by the molecular dynamics/quenching technique using the Cornell et al. force field (J. Am. Chem. Soc. 1995, 117, 5179). Long runs of molecular dynamics/quenching calculations allowed us to evaluate the free energy surface. The most stable and populated structures found were fully reoptimized at the correlated ab initio level employing the resolution of identity Møller-Plesset method. A systematic study of the base pairs' microhydration using both the empirical and the high-level correlated ab initio approaches is presented for the first time. We show that the occurrence of water molecules and their gradually increasing number as well as the methylation of the bases favor stacked structures over the planar hydrogen-bonded ones. These results based on the correlated ab initio calculations are in the excellent agreement with data obtained from our previous empirical potential molecular dynamics study (Kabelác et al. Chem.-Eur. J. 2001, 7, 2067).

Palladium-catalyzed Sonogashira cross-coupling reactions of halopurines 9-benzyl-6-chloropurine (2 a), 9-benzyl-8-bromoadenine (2 b), and 9-benzyl-2-chloroadenine (2 c) with ethynylferrocene (1) gave the corresponding (ferrocenylethynyl)purines 3 a-c in moderate to good yields. Catalytic hydrogenation of these alkynes over Pd/C afforded the respective saturated [2-(ferrocenyl)ethyl]purines 4 a-c. The crystal structures 3 a, 3 b, 4 a and 4 b as determined by X-ray diffraction show interesting solid-state interactions, markedly different for purines 3 a and 4 a on one hand and adenines 3 b and 4 b that possess a free amino group on the other. Electrochemistry of electrochemically labelled purines 3 and 4 has been studied by voltammetry and cyclic voltammetry on platinum disc electrode and the experimental oxidation potentials were confirmed and explained by ionization potentials from theoretical DFT calculations. Several compounds of this series exhibited a considerable cytostatic effect.

Properties of isolated intercalators (ethidium (E), daunomycin (D), ellipticine (EL), and 4,6'-diaminide-2-phenylindole (DAPI)) and their stacking interactions with adenine...thymine (AT) and guanine...cytosine (GC) nucleic acid base pairs were investigated by means of a nonempirical correlated ab initio method. All intercalators exhibit large charge delocalization, and none of them (including the DAPI dication) exhibits a site with dominant charge. All intercalators have large polarizability and are good electron acceptors, while base pairs are good electron donors. MP2/6-31G*(0.25) stabilization energies of intercalator...base pair complexes are large (E...AT, 22.4 kcal/mol; D...GC, 17.8 kcal/mol; EL...GC, 18.2 kcal/mol; DAPI...GC, 21.1 kcal/mol) and are well reproduced by modified AMBER potential (van der Waals radii of intercalator atoms are enlarged and their energy depths are increased). Standard AMBER potential underestimates binding, especially for DAPI-containing complexes. Because the DAPI dication is the best electron acceptor (among all intercalators studied), this difference is explained by the importance of the charge-transfer term, which is not included in the AMBER potential. For the neutral EL molecule, the standard AMBER force field provides correct results. The Hartree-Fock and DFT/B3LYP methods, not covering the dispersion energy, fail completely to reveal any energy minimum at the potential energy curve of the E...AT complex, and these methods thus cannot be recommended for a study of intercalation process. On the other hand, an approximate version of the DFT method, which was extended to cover London dispersion energy, yields for all complexes very good stabilization energies that are well comparable with referenced ab initio data. Besides the vertical dependence of the interaction, an energy twist dependence of the interaction energy was also investigated by a reference correlated ab initio method and empirical potentials. It is concluded that, despite the cationic (E +1, D +1, DAPI +2) or polar (EL) character of the intercalators investigated, it is the dispersion energy which predominantly contributes to the stability of intercalator...base pair complexes. Any procedure which does not cover dispersion energy is thus not suitable for studying the process of intercalation.