Fully automated molecular mechanics based induced fit protein-ligand docking method.

We describe a method for docking a ligand into a protein receptor while allowing flexibility of the protein binding site. The method employs a multistep procedure that begins with the generation of protein and ligand conformations. An initial placement of the ligand is then performed by computing binding site hotspots. This initial placement is followed by a protein side-chain refinement stage that models protein flexibility. The final step of the process is an energy minimization of the ligand pose in the presence of the rigid receptor. Thus the algorithm models flexibility of the protein at two stages, before and after ligand placement. We validated this method by performing docking and cross docking studies of eight protein systems for which crystal structures were available for at least two bound ligands. The resulting rmsd values of the 21 docked protein-ligand complexes showed values of 2 A or less for all but one of the systems examined. The method has two critical benefits for high throughput virtual screening studies. First, no user intervention is required in the docking once the initial binding site selection has been made in the protein. Second, the initial protein conformation generation needs to be performed only once for a given binding region. Also, the method may be customized in various ways depending on the particular scenario in which dockings are being performed. Each of the individual steps of the method is fully independent making it straightforward to explore different variants of the high level workflow to further improve accuracy and performance.

LigScore: a novel scoring function for predicting binding affinities.

We present two new empirical scoring functions, LigScore1 and LigScore2, that attempt to accurately predict the binding affinity between ligand molecules and their protein receptors. The LigScore functions consist of three distinct terms that describe the van der Waals interaction, the polar attraction between the ligand and protein, and the desolvation penalty attributed to the binding of the polar ligand atoms to the protein and vice versa. Utilizing a regression approach on a data set of 118 protein-ligand complexes we have obtained a linear equation, LigScore2, using these three descriptors. LigScore2 has good predictability with regard to experimental pKi values yielding a correlation coefficient, r2), of 0.75 and a standard deviation of 1.04 over the training data set, which consists of a diverse set of proteins that span more than seven protein families.

LigandFit: a novel method for the shape-directed rapid docking of ligands to protein active sites.

We present a new shape-based method, LigandFit, for accurately docking ligands into protein active sites. The method employs a cavity detection algorithm for detecting invaginations in the protein as candidate active site regions. A shape comparison filter is combined with a Monte Carlo conformational search for generating ligand poses consistent with the active site shape. Candidate poses are minimized in the context of the active site using a grid-based method for evaluating protein-ligand interaction energies. Errors arising from grid interpolation are dramatically reduced using a new non-linear interpolation scheme. Results are presented for 19 diverse protein-ligand complexes. The method appears quite promising, reproducing the X-ray structure ligand pose within an RMS of 2A in 14 out of the 19 complexes. A high-throughput screening study applied to the thymidine kinase receptor is also presented in which LigandFit, when combined with LigScore, an internally developed scoring function, yields very good hit rates for a ligand pool seeded with known actives.

Nuclear Overhauser effect and computational characterization of the beta-spiral of the polypentapeptide of elastin.

The structure of the elastin polypentapeptide, poly(VPGVG), was studied by nuclear Overhauser effect experiments using perdeuterated Val1 and Val4 samples under the condition where intermolecular interactions are absent. More extensive interaction was found between the Val1 gamma CH and Pro2 beta CH protons than between the Val4 gamma CH and Pro2 beta CH protons. The Val1 gamma CH3-Pro2 beta CH interaction does not occur within the same pentamer as previously shown experimentally and as expected from steric considerations. The results are incompatible with the presence of a random chain network in poly(VPGVG) at room temperature but are readily explicable in terms of interturn interactions in a beta-spiral structure. More specifically, the results indicate that the beta-spiral conformation with 2.9 pentamers/turn is more prevalent than that with 2.7 pentamers/turn. Using conformations developed by molecular mechanics calculations, molecular dynamics simulations were carried out to compare the relative energies of these two variants of this class of beta-spiral structures. It was found in vacuo that the structure with 2.9 pentamers/turn is indeed more stable than that of 2.7 pentamers/turn by approximately 1 kcal/mole-pentamer.

Synthesis, characterization, and black lipid membrane studies of [7-L-alanine] gramicidin A.

With a view to study the relevance of side-chain orientation in the transport of cations through a gramicidin transmembrane channel and to identify an analogue with favorable characteristics, [L-Ala7] gramicidin A was synthesized, purified, verified, and characterized by high-performance liquid chromatography, by carbon-13 and proton magnetic resonance spectra, and by circular dichroism spectra in methanol. Complete incorporation as the channel state was achieved when packaged in lysolecithin-containing lipid bilayers. The single-channel conductance data in diphytanoyllecithin/n-decane membranes are presented along with those of synthetic gramicidin A (GA). [L-Ala7] GA exhibits the highest most probable single-channel conductance so far reported for an analogue occurring at 28 pS as compared to 21 pS for GA under similar conditions. Also, a dramatic reduction in the dispersity of conducting states is observed with about 76% of the events falling in a narrow 1.75-pS conductance window as compared to about 31% of the events for GA under identical conditions. Thus, with the above characteristics, [L-Ala7]GA appears to be a very good candidate for a thorough study of ionic mechanism. The present results indicate that elements intrinsic to the channel proper are rate-limiting for GA and that there is no interfacial polarization or diffusion-controlled association at 1 M KCl and a 100-mV applied potential.

Potassium-39 NMR of K+ interaction with the gramicidin channel and NMR-derived conductance ratios for Na+, K+ and Rb+.

A potassium-39 NMR study of potassium ion interaction with the gramicidin transmembrane channel in phospholipid bilayers at high ion activity is reported which allows determination of a weak binding constant, Kwb approximately equal to 8.3/M, and an off-rate constant for the weak site, kwoff approximately equal to 2.6 X 10(7)/sec. These values are interpreted with the aid of additional NMR data as the binding constant for formation of the doubly occupied channel state and the rate constant for an ion leaving the doubly occupied state. Considering the singly occupied channel state for the potassium ion to be "electrically silent" at 1 molar ion activity, as with the sodium ion, the single-channel conductance for 100 mV and 30 degrees C calculated to be 29 pS, and using the same approximation with previous NMR results on the sodium and rubidium ions, reasonable conductance ratios were calculated. Further experimental estimates of the other three constants with the experimental location of binding sites and Eyring rate theory to introduce voltage dependence allowed a more complete calculation of the two-site channel. The single-channel conductance for potassium ion is calculated to be 24 pS at 1 M activity and 26 pS at 0.6 M activity, which compares for diphytanoyl phosphatidylcholine membranes to an experimental most probable single-channel conductance of 25 pS and a mean channel conductance of 20 pS. The calculated conductance ratios using NMR-derived constants were gamma (K)/gamma (Na) = 2.0 and gamma (Rb)/gamma (Na) = 4.3.(ABSTRACT TRUNCATED AT 250 WORDS)

Carbon-13 NMR relaxation studies demonstrate an inverse temperature transition in the elastin polypentapeptide.

Carbon-13 NMR longitudinal relaxation time and line-width studies are reported on the coacervate concentration (about 60% water by weight) of singly carbonyl carbon enriched polypentapeptides of elastin: specifically, (L-Val1-L-[1-13C]Pro2-Gly3-L-Val4-Gly5)n and (L-Val1-L-Pro2-Gly3-L-Val4-[1-13C]Gly5)n. On raising the temperature from 10 to 25 degrees C and from 40 to 70 degrees C, carbonyl mobility increases, but over the temperature interval from 25 to 40 degrees C, the mobility decreases. The results characterize an inverse temperature transition in the most fundamental sense of temperature being a measure of molecular motion. This transition in the state of the polypentapeptide indicates an increase in order of polypeptide on raising the temperature from 25 degrees C to physiological temperature. This fundamental NMR characterization corresponds with the results of numerous other physical methods, e.g., circular dichroism, dielectric relaxation, and electron microscopy, that correspondingly indicate an increase in order of the polypentapeptide both intramolecularly and intermolecularly for the same temperature increase from 25 to 40 degrees C. Significantly with respect to elastomeric function, thermoelasticity studies on gamma-irradiation cross-linked polypentapeptide coacervate show a dramatic increase in elastomeric force over the same interval that is here characterized by NMR as an inverse temperature transition. The temperature dependence of mobility above 40 degrees C indicates an activation energy of the order of 1.2 kcal/mol, which is the magnitude of barrier expected for elasticity.

Shortened analog of the gramicidin A channel argues for the doubly occupied channel as the dominant conducting state.

A shortened analog of the gramicidin A transmembrane channel has been synthesized and its transport characterized in planar lipid bilayer membranes. General considerations of a shorter diffusional length and a shorter distance over which the voltage drop occurs (i.e., an increased electric field) would contribute to an increase in single-channel conductance. The finding of a decreased single-channel conductance supports the perspective that the dominant conducting state is the doubly occupied channel wherein distance-dependent repulsion due to the first ion in the channel impedes entry of the second ion in the shorter channel.

The source of the dispersity of gramicidin A single-channel conductances. The L X Leu5-gramicidin A analog.

The L X Leu5-gramicidin A analog has been designed, synthesized, and verified for the purpose of testing the perspective that different side-chain distributions are responsible for the dispersity of single-channel conductances. The structural analysis shows that the L . Leu5 analog would limit the possible number of side-chain distributions and that the dispersity of channel conductances would decrease. The finding is the expected decrease in less probable conductance states and the associated increase in the probability of the most probable conductance state.

Dispersity of des-L X Val7-D X Val8-Gramicidin A single channel conductances argues for different side chain orientations as basis.

Planar bilayer studies are reported on the channel activity of des-L X Val7-D X Val8-Gramicidin A. This analog is designed to provide more long-lived side chain distributions involving the Trp residues than occur with Gramicidin A. The carbonyls of these residues coordinate the permeant cation and the energetics of the coordination, which is proposed to depend on side chain orientation, determines the free energies of the rate limiting entrance-exit barriers and the binding sites. The finding of an increased dispersity of single channel conductance for the analog supports the perspective that dispersity derives from different side chain distributions on the same backbone conformation. Channel mechanism is not understood until dispersity is explained.

Temperature dependence of single channel currents and the peptide libration mechanism for ion transport through the gramicidin A transmembrane channel.

A study of the temperature dependence of gramicidin A conductance of K+ in diphytanoyllecithin/n-decane membranes shows the plot of In (single channel conductance) as a function of reciprocal temperature to be nonlinear for the most probable set of conductance states. These results are considered in terms of a series of barriers, of the dynamics of channel conformation, vis-a-vis the peptide libration mechanism, and of the effect of lipid viscosity on side chain motions again as affecting the energetics of peptide libration.

Calcium binding to a calcifiable matrix: 43Ca NMR binding studies on the polypentapeptide of elastin.

Calcium-43 nuclear magnetic resonance (NMR) has been used to characterize the binding of calcium ion to the polypentapeptide of elastin. The change in longitudinal relaxation time, T1, from 1000 msec to 15 msec at 38 mM CaCl2, on going from D2O to coacervate concentration at 20 degrees C, is dramatic. At 20 degrees C, the binding constant decreases with increasing calcium ion concentration over the range of 40 mM to 250 mM with apparent binding constants of Ktb congruent to 35 M-1 and Kwb congruent to 7M-1. On raising the temperature to 38 degrees C, there is a significant decrease in calcium interaction, as seen in the large increase in T1 from 15 msec at 20 degrees C to 350 msec at 38 mM CaCl2. This correlates with an inverse temperature transition for the polypentapeptide and results in a concentration dependence of the calcium interaction which shows positive cooperativity. Since the polypentapeptide has been shown to calcify massively when exposed to normal serum dialysates, these calcium interactions are of particular interest as the initiating process for calcification of this component of the elastic fiber.

Conformation characterization of cyclopentapeptide, L.Val-L.Pro-Gly-L.Val-Gly: a repeating analogue of elastin.

The cyclopentapeptide, L.Val1-L.Pro2-Gly3-L.Val4-Gly5, was synthesized and its conformational characterization was carried out using n.m.r. and theoretical energy calculations. The n.m.r. studies indicated the existence of a cis Val1-Pro2 peptide bond in water and a very strong intramolecular H-bond between the val1 NH and Gly3 C=O groups. This H-bond forms a beta-turn (type II) placing Val4 and Gly5 residues within the turn. Two minimum energy conformations were derived, one of which agrees very well with the solution conformation.

The malonyl gramicidin channel: NMR-derived rate constants and comparison of calculated and experimental single-channel currents.

Malonyl gramicidin is incorporated into lysolecithin micelles in a manner which satisfies a number of previously demonstrated criteria for the formation of the transmembrane channel structure. By means of sodium-23 nuclear magnetic resonance, two binding sites are observed: a tight site and a weak site with binding constants of approximately 100 M-1 and 1 M-1 respectively. In addition, off-rate constants from the two sites were estimated from NMR analyses to be kofft congruent to 3 X 10(5)/sec and koffw congruent to 2 X 10(7)/sec giving, with the binding constants, the on-rate constants, kont congruent to 3 X 10(7)/Msec and konw congruent to 2 X 10(7)/Msec. Five different multiple occupancy models with NMR-restricted energy profiles were considered for the purpose of calculating single-channel currents as a function of voltage and concentration utilizing the four NMR-derived rate constants (and an NMR-limit placed on a fifth rate constant for intrachannel ion translocation) in combination with Eyring rate theory for the introduction of voltage dependence. Using the X-ray diffraction results of Koeppe et al. (1979) for limiting the positions of the tight sites, the two-site model and a three-site model in which the weak sites occur after the tight site is filled were found to satisfactorily calculate the experimental currents (also reported here) and to fit the experimental currents extraordinarily well when the experimentally derived values were allowed to vary to a least squares best fit. Surprisingly the "best fit" values differed by only about a factor of two from the NMR-derived values, a variation that is well within the estimated experimental error of the rate constants. These results demonstrate the utility of ion nuclear magnetic resonance to determine rate constants relevant to transport through the gramicidin channel and of the Eyring rate theory to introduce voltage dependence.

Rate theory calculation of gramicidin single-channel currents using NMR-derived rate constants.

By means of 23Na NMR, two ion binding sites were observed in phospholipid-packaged gramicidin channels and the four associated rate constants were approximated. Limits also were placed on a fifth rate constant for an intrachannel ion translocation. By using Eyring rate theory to introduce voltage dependence, these rate constants were used in steady-state-current equations for calculation of gramicidin single-channel currents for two- and three-site models. Calculated single-channel currents are compared with previously published experimental single-channel currents obtained by electrical measurements on Na+ transport across gramicidin-doped planar lipid bilayers. The calculated results for the two- and three-site models compare favorably with the experimental results. Accordingly, it is demonstrated that NMR-derived rate constants can be coupled with Eyring rate theory to calculate currents through a transmembrane channel and to do so within levels of variation that compare with the differences obtained on planar lipid bilayers formed with different lipids.

Free-energy calculations of the interactions of helical poly(L-proline) with water.

Conformational energies, consisting of nonbonded and electrostatic interactions, have been calculated for poly(L-proline) I and II helices, of six-residues length, with a water molecule hydrogen-bonded to each imide group. For polyproline I, the Traub-Shmueli prolyl-ring geometry was used, whereas for polyproline II, calculations were done for the 2 ring geometries of Leung and Marsh and, in some cases, for the several ring geometries of Ramachandran, et al. The chain torsion angles omega(N-C') and Psi(C(alpha)-C') were varied, as were the two angles that specify the orientations of the water molecules. By summation over the latter, it was possible to calculate a free energy at each omega,Psi that incorporated the conformational entropy of the water molecules. Such omega,Psi maps reveal that binding of water can cause changes in the equilibrium conformation of polyproline, as well as in the energetics of the chain with respect to omega and Psi. This has important consequences on the statistics of the chain. Considerations of prolyl-ring self-energies, as well as the conformational entropies of the chains, show that polyproline II is more stable in water than is polyproline I, and determine which ring geometry is favored for polyproline in water.