Genetic algorithm optimization of a molecular mechanics force field for technetium

Re(I) and Tc(I) complexes for targeting nitric oxide synthase: influence of the chelator in the affinity for the enzyme

Aiming to design (99m) Tc complexes for probing nitric oxide synthase (NOS) by SPECT, we synthesized conjugates (L4-L6) comprising a NOS-recognizing moiety connected to a diamino-propionic acid (dap) chelating unit. The conjugates led to complexes of the type fac-[M(CO)3 (ĸ(3) -L)] (M = Re/(99m) Tc; Re4/Tc4: L = L4; Re5/Tc5: L = L5; Re6/Tc6: L = L6). Enzymatic studies showed that L4 and L5, but not L6, gave complexes (Re4 and Re5) that are less potent than the conjugates. To rationalize these results, we performed docking and molecular dynamics simulations. The high affinity of L4 and L5 is due to the strong interactions between the dap chelator and polar residues of the binding cavity. These interactions are hampered by metallation resulting in complexes with lower affinity. The higher potency of Re5 compared to Re4 was assigned to the increased bulkiness of Re5 and the presence of additional anchoring groups that better fit the active site and provide more extensive contacts. In turn, Re6 is too bulky and its organometallic tail is oriented toward the peripheral pocket of iNOS, leading to loss of contacts and a lower affinity. These results were compared with our previous results obtained with analogue complexes stabilized by a pyrazolyl-diamine chelating unit.

Optimization of a molecular mechanics force field for type-II polyoxometalates focussing on electrostatic interactions: a case study

In this study, we have focussed on type-II polyanions such as [M(7)O(24)](6-), and we have developed and validated optimized force fields that include electrostatic and van der Waals interactions. These contributions to the total steric energy are described by the nonbonded term, which encompasses all interactions between atoms that are not transmitted through the bonds. A first validation of a stochastic technique based on genetic algorithms was previously made for the optimization of force fields dedicated to type-I polyoxometalates. To describe the new nonbonded term added in the functional, a fixed-charged model was chosen. Therefore, one of the main issues was to analyze that which partial atomic charges could be reliably used to describe these interactions in such inorganic compounds. Based on several computational strategies, molecular mechanics (MM) force field parameters were optimized using different types of atomic charges. Moreover, the influence of the electrostatic and van der Waals buffering constants and 1,4-interactions scaling factors used in the force field were also tested, either being optimized as well or fixed with respect to the values of CHARMM force field. Results show that some atomic charges are not well adapted to CHARMM parameters and lead to unrealistic MM-optimized structures or a MM divergence. As a result, a new scaling factor has been optimized for Quantum Theory of Atoms in Molecules charges and charges derived from the electrostatic potential such as ChelpG. The force fields optimized can be mixed with the CHARMM force field, without changing it, to study for the first time hepta-anions interacting with organic molecules.

Searching for globally optimal functional forms for interatomic potentials using genetic programming with parallel tempering

Molecular dynamics and other molecular simulation methods rely on a potential energy function, based only on the relative coordinates of the atomic nuclei. Such a function, called a force field, approximately represents the electronic structure interactions of a condensed matter system. Developing such approximate functions and fitting their parameters remains an arduous, time-consuming process, relying on expert physical intuition. To address this problem, a functional programming methodology was developed that may enable automated discovery of entirely new force-field functional forms, while simultaneously fitting parameter values. The method uses a combination of genetic programming, Metropolis Monte Carlo importance sampling and parallel tempering, to efficiently search a large space of candidate functional forms and parameters. The methodology was tested using a nontrivial problem with a well-defined globally optimal solution: a small set of atomic configurations was generated and the energy of each configuration was calculated using the Lennard-Jones pair potential. Starting with a population of random functions, our fully automated, massively parallel implementation of the method reproducibly discovered the original Lennard-Jones pair potential by searching for several hours on 100 processors, sampling only a minuscule portion of the total search space. This result indicates that, with further improvement, the method may be suitable for unsupervised development of more accurate force fields with completely new functional forms.

Molecular modeling of hexakis(areneisonitrile)technetium(I), tricarbonyl eta5 cyclopentadienyl technetium and technetium(V)-oxo complexes: MM3 parameter development and prediction of biological properties

Genetic algorithms (GA) were used to develop specific technetium metal-ligand force field parameters for the MM3 force field. These parameters were developed using automated procedures within the program FFGenerAtor from a combination of crystallographic structures and ab initio calculations. These new parameters produced results in good agreement with experiment when tested against a blind validation set. To illustrate the utility of these new force field parameters, quantitative structure-activity relationship (QSAR) models were developed to predict the P-glycoprotein uptake (log10 VI) of a series of hexakis(areneisonitrile)technetium(I) complexes and to predict their biodistribution. The log10 VI QSAR model, built using a training set of 16 Tc(I) isonitrile complexes, exhibited a correlation between the experimental log10 VI and 5 simple descriptors as follows: r2 = 0.94, q2 = 0.93. When applied to an external test set of six Tc(I) isonitrile complexes, the QSAR preformed with great accuracy q2 = 0.78 based on a leave-one-out cross-validation analysis. Further QSAR models were developed to predict the biodistribution of the same set of Tc(I) isonitrile complexes; a QSAR model to predict hepatic uptake exhibited a correlation between the experimental log10(Blood/Liver) with six simple descriptors as follows: r2 = 0.97, q2 = 0.96. A QSAR model to predict renal uptake exhibited a correlation between the experimental log10(Blood/Kidney) and six simple descriptors as follows: r2 = 0.85, q2 = 0.82. When applied to the external test set the QSAR models preformed with great accuracy, q2 = 0.78 and 0.56, respectively.

De novo structural prediction of transition metal complexes: application to technetium

De novo structural prediction of transition metal complexes is investigated. Technetium complexes are chosen given their importance in medical imaging and nuclear waste remediation and for the chemical diversity they display. A new conformational searching algorithm (LIGB) for transition metals is described that allows one to search for different conformational and geometric isomers within a single simulation. In the preponderance of cases, both conformational searching techniques (LIGB and high-temperature molecular dynamics/simulated annealing) provide comparable results, while LIGB is superior for macrocyclic complexes. A genetic algorithm-optimized PM3(tm) parametrization for Tc is compared with the standard implementation and found to yield a significant improvement in predictive ability for the most prevalent Tc structural motifs. The utility of a coupled molecular mechanics-semiempirical quantum mechanics protocol is demonstrated for very rapid, efficient, and effective de novo prediction of transition metal complex geometries.

MM3 parametrization of four- and five-coordinated rhenium complexes by a genetic algorithm--which factors influence the optimization performance?

Genetic algorithms (GA) were used to solve one of the multidimensional problems in computational chemistry, the optimization of force field parameters. The correlation between the composition of the GA, its parameters (p(c), p(m)) and the quality of the results were investigated. The composition was studied for all combinations of a Simple GA/Steady State GA with a Roulette Wheel/Tournament Selector using different values each for crossover (0.5, 0.7, 0.9) and mutation rates (0.01, 0.02, 0.05, 0.10, 0.20). The results show that the performance is strongly dependent on the GA scheme, where the Simple GA/Tournament Selector yields the best results. Two new MM3 parameters were introduced for rhenium compounds with coordination number four (204) and coordination number five (205), the formal oxidation states of rhenium ranging from +V to +VII. A manifold of parameters (Re-C, N, O, S) was obtained by using a diverse set of CSD structures. The advantage of the GA vs. UFF calculations is shown by comparison of several examples. The GA optimized parameters were able to reproduce the geometrical data of the X-ray structures.