Identification of Plasmodium falciparum spermidine synthase active site binders through structure-based virtual screening.

Seven novel binders, binding in the active site of Plasmodium falciparum spermidine synthase, were identified by structure-based virtual screening. The binding of these compounds was experimentally verified by NMR techniques. Spermidine synthase, an enzyme involved in the polyamine pathway, has been suggested as a target for treating malaria. The virtual screening protocol combined 3D pharmacophore filtering, docking, and scoring, focusing on finding compounds predicted to form interactions mimicking those of a previously known binder. The virtual screen resulted in the selection of 28 compounds that were acquired and tested from 2.6 million starting structures. Two of the seven binders were predicted to bind in the amino substrate binding pocket. Both of these showed stronger binding upon addition of methylthioadenosine, one of the two products of the enzyme, and a known binder and inhibitor. The five other compounds were predicted to bind in the part of the active site where the other substrate, decarboxylated S-adenosylmethionine, binds. These five compounds all competed for binding with methylthioadenosine.

Are automated molecular dynamics simulations and binding free energy calculations realistic tools in lead optimization? An evaluation of the linear interaction energy (LIE) method.

An extensive evaluation of the linear interaction energy (LIE) method for the prediction of binding affinity of docked compounds has been performed, with an emphasis on its applicability in lead optimization. An automated setup is presented, which allows for the use of the method in an industrial setting. Calculations are performed for four realistic examples, retinoic acid receptor gamma, matrix metalloprotease 3, estrogen receptor alpha, and dihydrofolate reductase, focusing on different aspects of the procedure. The obtained LIE models are evaluated in terms of the root-mean-square (RMS) errors from experimental binding free energies and the ability to rank compounds appropriately. The results are compared to the best empirical scoring function, selected from a set of 10 scoring functions. In all cases, good LIE models can be obtained in terms of free-energy RMS errors, although reasonable ranking of the ligands of dihydrofolate reductase proves difficult for both the LIE method and scoring functions. For the other proteins, the LIE model results in better predictions than the best performing scoring function. These results indicate that the LIE approach, as a tool to evaluate docking results, can be a valuable asset in computational lead optimization programs.

Ligand bias of scoring functions in structure-based virtual screening.

A total of 945 known actives and roughly 10 000 decoy compounds were docked to eight different targets, and the resulting poses were scored using 10 different scoring functions. Three different score postprocessing methods were evaluated with respect to improvement of the enrichment in virtual screening. The three procedures were (i) multiple active site correction (MASC) as has been proposed by Vigers and Rizzi, (ii) a variation of MASC where corrections terms are predicted from simple molecular descriptors through PLS, PLS MASC, and (iii) size normalization. It was found that MASC did not generally improve the enrichment factors when compared to uncorrected scoring functions. For some combinations of scoring functions and targets, the enrichment was improved, for others not. However, by excluding the standard deviation from the MASC equation and transforming the scores for each target to a mean of 0 and a standard deviation of 1 (unit variance normalization), the performance was improved as compared to the original MASC method for most combinations of targets and scoring functions. Furthermore, when the molecular descriptors were fit to the mean scores over all targets and the resulting PLS models were used to predict mean scores, the enrichment as compared to the raw score was improved more often than by straightforward MASC. A high to intermediate linear correlation between the score and the number of heavy atoms was found for all scoring functions except FlexX. There seems to be a correlation between the size dependence of a scoring function and the effectiveness of PLS MASC in increasing the enrichment for that scoring function. Finally, normalization by molecular weight or heavy atom count was sometimes successful in increasing the enrichment. Dividing by the square or cubic root of the molecular weight or heavy atom count instead was often more successful. These results taken together suggest that ligand bias in scoring functions is a source of false positives in structure-based virtual screening. The number of false positives caused by ligand bias may be decreased using, for example, the PLS MASC procedure proposed in this study.

Mycobacterium tuberculosis ribose-5-phosphate isomerase has a known fold, but a novel active site.

Ribose-5-phosphate isomerases (EC 5.3.1.6) inter-convert ribose-5-phosphate and ribulose-5-phosphate. This reaction allows the synthesis of ribose from other sugars, as well a means for salvage of carbohydrates after nucleotide breakdown. Two unrelated types of enzyme are known to catalyze the isomerization. The most common one, RpiA, is present in almost all organisms. The second type, RpiB, is found in many bacterial species.Here, we demonstrate that the RpiB from Mycobacterium tuberculosis (Rv2465c) has catalytic properties very similar to those previously reported for the Escherichia coli RpiB enzyme. Further, we report the structure of the mycobacterial enzyme, solved by molecular replacement and refined to 1.88A resolution. Comparison with the E.coli structure shows that there are important differences in the two active sites, including a change in the position and nature of the catalytic base. Sequence comparisons reveal that the M.tuberculosis and E.coli RpiB enzymes are in fact representative of two distinct sub-families. The mycobacterial enzyme represents a type found only in actinobacteria, while the enzyme from E.coli is typical of that seen in many other bacterial proteomes. Both RpiBs are very different from RpiA in structure as well as in the construction of the active site. Docking studies allow additional insights into the reactions of all three enzymes, and show that many features of the mechanism are preserved despite the different catalytic components.

Improving structure-based virtual screening by multivariate analysis of scoring data.

Three different multivariate statistical methods, PLS discriminant analysis, rule-based methods, and Bayesian classification, have been applied to multidimensional scoring data from four different target proteins: estrogen receptor alpha (ERalpha), matrix metalloprotease 3 (MMP3), factor Xa (fXa), and acetylcholine esterase (AChE). The purpose was to build classifiers able to discriminate between active and inactive compounds, given a structure-based virtual screen. Seven different scoring functions were used to generate the scoring matrices. The classifiers were compared to classical consensus scoring and single scoring functions. The classifiers show a superior performance, with rule-based methods being most effective. The precision of correctly predicting an active compound is about 90% for three of the targets and about 25% for acetylcholine esterase. On the basis of these results, a new two-stage approach is suggested for structure-based virtual screening where limited activity information is available.

Crystal structure of the heterodimeric complex of LXRalpha and RXRbeta ligand-binding domains in a fully agonistic conformation.

The nuclear receptor heterodimers of liver X receptor (LXR) and retinoid X receptor (RXR) are key transcriptional regulators of genes involved in lipid homeostasis and inflammation. We report the crystal structure of the ligand-binding domains (LBDs) of LXRalpha and RXRbeta complexed to the synthetic LXR agonist T-0901317 and the RXR agonist methoprene acid (Protein Data Base entry 1UHL). Both LBDs are in agonist conformation with GRIP-1 peptides bound at the coactivator binding sites. T-0901317 occupies the center of the LXR ligand-binding pocket and its hydroxyl head group interacts with H421 and W443, residues identified by mutational analysis as critical for ligand-induced transcriptional activation by T-0901317 and various endogenous oxysterols. The topography of the pocket suggests a common anchoring of these oxysterols via their 22-, 24- or 27-hydroxyl group to H421 and W443. Polyunsaturated fatty acids act as LXR antagonists and an E267A mutation was found to enhance their transcriptional inhibition. The present structure provides a powerful tool for the design of novel modulators that can be used to characterize further the physiological functions of the LXR-RXR heterodimer.

Genome-based identification and analysis of collagen-related structural motifs in bacterial and viral proteins.

Collagens are extended trimeric proteins composed of the repetitive sequence glycine-X-Y. A collagen-related structural motif (CSM) containing glycine-X-Y repeats is also found in numerous proteins often referred to as collagen-like proteins. Little is known about CSMs in bacteria and viruses, but the occurrence of such motifs has recently been demonstrated. Moreover, bacterial CSMs form collagen-like trimers, even though these organisms cannot synthesize hydroxyproline, a critical residue for the stability of the collagen triple helix. Here we present 100 novel proteins of bacteria and viruses (including bacteriophages) containing CSMs identified by in silico analyses of genomic sequences. These CSMs differ significantly from human collagens in amino acid content and distribution; bacterial and viral CSMs have a lower proline content and a preference for proline in the X position of GXY triplets. Moreover, the CSMs identified contained more threonine than collagens, and in 17 of 53 bacterial CSMs threonine was the dominating amino acid in the Y position. Molecular modeling suggests that threonines in the Y position make direct hydrogen bonds to neighboring backbone carbonyls and thus substitute for hydroxyproline in the stabilization of the collagen-like triple-helix of bacterial CSMs. The majority of the remaining CSMs were either rich in proline or rich in charged residues. The bacterial proteins containing a CSM that could be functionally annotated were either surface structures or spore components, whereas the viral proteins generally could be annotated as structural components of the viral particle. The limited occurrence of CSMs in eubacteria and lower eukaryotes and the absence of CSMs in archaebacteria suggests that DNA encoding CSMs has been transferred horizontally, possibly from multicellular organisms to bacteria.

A triple mutant of the Drosophila ERR confers ligand-induced suppression of activity.

The steroid hormone (NR3) subfamily of nuclear receptors was until recently believed to be restricted to deuterostomes. However, a novel nuclear receptor belonging to the NR3 subfamily was recently identified in the Drosophila melanogaster genome, indicating the existence of an ancestor before the evolutionary split of deuterostomes and protostomes. This receptor, termed the Drosophila estrogen-related receptor (dERR), most closely resembles the human and mouse estrogen-related receptors (ERRs) in both the DNA binding domain (DBD) (approximately 85% identical) and the ligand binding domain (LBD) (approximately 35% identical). Here we describe the functional analysis and rational design of ligand responsive dERR mutants created by protein engineering of the LBD. On the basis of homology modeling, three amino acid residues in the LBD were identified and mutated to enable ligand-dependent suppression of transcriptional activity. Our results show that the Y295A/T333I/Y365L triple mutant is significantly suppressed by the known ERR inverse agonists 4-hydroxytamoxifen (OHT) and diethylstilbestrol (DES), in comparison to the wild-type dERR receptor, which was inefficiently suppressed by these substances. The coactivator mGRIP-1 (mouse glucocorticoid receptor interacting protein 1) was shown to significantly increase the activity of the triple mutant in transfection experiments, and the addition of OHT resulted in an efficient suppression of the activity. Accordingly, the ability to functionally interact with a coactivator is still maintained by the Y295A/T333I/Y365L mutant. These findings demonstrate the potential of using rational design and engineering of the LBD to study the function of a nuclear receptor lacking identified ligands.