Search

Abstract A key factor in computational drug design is the consistency and reliability with which intermolecular interactions between a wide variety of molecules can be described. Here we present a procedure to efficiently, reliably and automatically assign partial atomic charges to atoms based on known distributions. We formally introduce the molecular charge assignment problem, where the task is to select a charge from a set of candidate charges for every atom of a given query molecule. Charges are accompanied by a score that depends on their observed frequency in similar neighbourhoods (chemical environments) in a database of previously parameterised molecules. The aim is to assign the charges such that the total charge equals a known target charge within a margin of error while maximizing the sum of the charge scores. We show that the problem is a variant of the well-studied multiple-choice knapsack problem and thus weakly $$\mathcal {NP}$$ NP -complete. We propose solutions based on Integer Linear Programming and a pseudo-polynomial time Dynamic Programming algorithm. We demonstrate that the results obtained for novel molecules not included in the database are comparable to the ones obtained performing explicit charge calculations while decreasing the time to determine partial charges for a molecule from hours or even days to below a second. Our software is openly available.

Despite decades of research, the mechanism of action of the ABC multidrug transporter P-glycoprotein (P-gp) remains elusive. Due to experimental limitations, many researchers have turned to molecular dynamics simulation studies in order to investigate different aspects of P-gp function. However, such studies are challenging and caution is required when interpreting the results. P-gp is highly flexible and the time scale on which it can be simulated is limited. There is also uncertainty regarding the accuracy of the various crystal structures available, let alone the structure of the protein in a physiologically relevant environment. In this study, three alternative structural models of mouse P-gp (3G5U, 4KSB, 4M1M), all resolved to 3.8 Å, were used to initiate sets of simulations of P-gp in a membrane environment in order to determine: a) the sensitivity of the results to differences in the starting configuration; and b) the extent to which converged results could be expected on the times scales commonly simulated for this system. The simulations suggest that the arrangement of the nucleotide binding domains (NBDs) observed in the crystal structures is not stable in a membrane environment. In all simulations, the NBDs rapidly associated (within 10 ns) and changes within the transmembrane helices were observed. The secondary structure within the transmembrane domain was best preserved in the 4M1M model under the simulation conditions used. However, the extent to which replicate simulations diverged on a 100 to 200 ns timescale meant that it was not possible to draw definitive conclusions as to which structure overall was most stable, or to obtain converged and reliable results for any of the properties examined. The work brings into question the reliability of conclusions made in regard to the nature of specific interactions inferred from previous simulation studies on this system involving similar sampling times. It also highlights the need to demonstrate the statistical significance of any results obtained in simulations of large flexible proteins, especially where the initial structure is uncertain.

Starch is a complex branched glucose polymer whose branch molecular weight distribution (the chain-length distribution, CLD) influences nutritionally important properties such as digestion rate. Chain-stopping in starch biosynthesis is by starch branching enzyme (SBE). Site-directed mutagenesis was used to modify SBEIIa from Zea mays (mSBEIIa) to produce mutants, each differing in a single conserved amino-acid residue. Products at different times from in vitro branching were debranched and the time evolution of the CLD measured by size-exclusion chromatography. The results confirm that Tyr352, Glu513, and Ser349 are important for mSBEIIa activity while Arg456 is important for determining the position at which the linear glucan is cut. The mutant mSBEIIa enzymes have different activities and suggest the length of the transferred chain can be varied by mutation. The work shows analysis of the molecular weight distribution can yield information regarding the enzyme branching sites useful for development of plants yielding starch with improved functionality.

ATP Binding Cassette (ABC) transporters couple the binding and hydrolysis of ATP to the transport of substrate molecules across the membrane. The mechanism by which ATP binding and/or hydrolysis drives the conformational changes associated with substrate transport has not yet been characterized fully. Here, changes in the conformation of the ABC export protein P-glycoprotein on ATP binding are examined in a series of molecular dynamics simulations. When one molecule of ATP is placed at the ATP binding site associated with each of the two nucleotide binding domains (NBDs), the membrane-embedded P-glycoprotein crystal structure adopts two distinct metastable conformations. In one, each ATP molecule interacts primarily with the Walker A motif of the corresponding NBD. In the other, the ATP molecules interacts with both Walker A motif of one NBD and the Signature motif of the opposite NBD inducing the partial dimerization of the NBDs. This interaction is more extensive in one of the two ATP binding site, leading to an asymmetric structure. The overall conformation of the transmembrane domains is not altered in either of these metastable states, indicating that the conformational changes associated with ATP binding observed in the simulations in the absence of substrate do not lead to the outward-facing conformation and thus would be insufficient in themselves to drive transport. Nevertheless, the metastable intermediate ATP-bound conformations observed are compatible with a wide range of experimental cross-linking data demonstrating the simulations do capture physiologically important conformations. Analysis of the interaction between ATP and its cofactor Mg(2+) with each NBD indicates that the coordination of ATP and Mg(2+) differs between the two NBDs. The role structural asymmetry may play in ATP binding and hydrolysis is discussed. Furthermore, we demonstrate that our results are not heavily influenced by the crystal structure chosen for initiation of the simulations.