Benchmark of four popular virtual screening programs: construction of the active/decoy dataset remains a major determinant of measured performance.

BACKGROUND: In a structure-based virtual screening, the choice of the docking program is essential for the success of a hit identification. Benchmarks are meant to help in guiding this choice, especially when undertaken on a large variety of protein targets. Here, the performance of four popular virtual screening programs, Gold, Glide, Surflex and FlexX, is compared using the Directory of Useful Decoys-Enhanced database (DUD-E), which includes 102 targets with an average of 224 ligands per target and 50 decoys per ligand, generated to avoid biases in the benchmarking. Then, a relationship between these program performances and the properties of the targets or the small molecules was investigated. RESULTS: The comparison was based on two metrics, with three different parameters each. The BEDROC scores with α = 80.5, indicated that, on the overall database, Glide succeeded (score > 0.5) for 30 targets, Gold for 27, FlexX for 14 and Surflex for 11. The performance did not depend on the hydrophobicity nor the openness of the protein cavities, neither on the families to which the proteins belong. However, despite the care in the construction of the DUD-E database, the small differences that remain between the actives and the decoys likely explain the successes of Gold, Surflex and FlexX. Moreover, the similarity between the actives of a target and its crystal structure ligand seems to be at the basis of the good performance of Glide. When all targets with significant biases are removed from the benchmarking, a subset of 47 targets remains, for which Glide succeeded for only 5 targets, Gold for 4 and FlexX and Surflex for 2. CONCLUSION: The performance dramatic drop of all four programs when the biases are removed shows that we should beware of virtual screening benchmarks, because good performances may be due to wrong reasons. Therefore, benchmarking would hardly provide guidelines for virtual screening experiments, despite the tendency that is maintained, i.e., Glide and Gold display better performance than FlexX and Surflex. We recommend to always use several programs and combine their results. Graphical AbstractSummary of the results obtained by virtual screening with the four programs, Glide, Gold, Surflex and FlexX, on the 102 targets of the DUD-E database. The percentage of targets with successful results, i.e., with BDEROC(α = 80.5) > 0.5, when the entire database is considered are in Blue, and when targets with biased chemical libraries are removed are in Red.

vSDC: a method to improve early recognition in virtual screening when limited experimental resources are available.

BACKGROUND: In drug design, one may be confronted to the problem of finding hits for targets for which no small inhibiting molecules are known and only low-throughput experiments are available (like ITC or NMR studies), two common difficulties encountered in a typical academic setting. Using a virtual screening strategy like docking can alleviate some of the problems and save a considerable amount of time by selecting only top-ranking molecules, but only if the method is very efficient, i.e. when a good proportion of actives are found in the 1-10 % best ranked molecules. RESULTS: The use of several programs (in our study, Gold, Surflex, FlexX and Glide were considered) shows a divergence of the results, which presents a difficulty in guiding the experiments. To overcome this divergence and increase the yield of the virtual screening, we created the standard deviation consensus (SDC) and variable SDC (vSDC) methods, consisting of the intersection of molecule sets from several virtual screening programs, based on the standard deviations of their ranking distributions. CONCLUSIONS: SDC allowed us to find hits for two new protein targets by testing only 9 and 11 small molecules from a chemical library of circa 15,000 compounds. Furthermore, vSDC, when applied to the 102 proteins of the DUD-E benchmarking database, succeeded in finding more hits than any of the four isolated programs for 13-60 % of the targets. In addition, when only 10 molecules of each of the 102 chemical libraries were considered, vSDC performed better in the number of hits found, with an improvement of 6-24 % over the 10 best-ranked molecules given by the individual docking programs.Graphical abstractIn drug design, for a given target and a given chemical library, the results obtained with different virtual screening programs are divergent. So how to rationally guide the experimental tests, especially when only a few number of experiments can be made? The variable Standard Deviation Consensus (vSDC) method was developed to answer this issue. Left panel the vSDC principle consists of intersecting molecule sets, chosen on the basis of the standard deviations of their ranking distributions, obtained from various virtual screening programs. In this study Glide, Gold, FlexX and Surflex were used and tested on the 102 targets of the DUD-E database. Right panel Comparison of the average percentage of hits found with vSDC and each of the four programs, when only 10 molecules from each of the 102 chemical libraries of the DUD-E database were considered. On average, vSDC was capable of finding 38 % of the findable hits, against 34 % for Glide, 32 % for Gold, 16 % for FlexX and 14 % for Surflex, showing that with vSDC, it was possible to overcome the unpredictability of the virtual screening results and to improve them.

Influence of the oxazole ring connection on the fluorescence of oxazoyl-triphenylamine biphotonic DNA probes.

On the basis of our previous work on DNA fluorophores derived from vinylpyridinium-triphenylamine, we explored the structure space around the electron-rich triphenylamine (TP) core by changing the vinyl bond to an oxazole ring. As 2,5-diaryloxazoles are known to be highly fluorescent and efficient two photon absorbers, we synthesized analogues with two different connections of the oxazole to the triphenylamine core: TP-Ox2Py and TP-Ox5Py sets. Since the benzimidazolium group was proven to be more effective in the TP series than the pyridinium, we also synthesized a TP-Ox5Bzim set. The TP-Ox5Py series retains the TP-Py properties: on/off behavior on DNA, good two-photon cross-section and bright staining of nuclear DNA by microscopy under both one or two-photon excitation. On the other hand, the TP-Ox2Py series does not display fluorescence upon binding to DNA. The TP-Ox5Bzim set is fluorescent even in the absence of DNA and displays lower affinity than the corresponding TP-Ox5Py. CD experiments and docking were performed to understand these different behaviors.

New aminopyrimidine derivatives as inhibitors of the TAM family.

In this study, we describe the synthesis of new pyrimidine analogs of BMS-777607, a potent and selective inhibitor of Met kinase. Inhibition of Met and Axl remained high whereas inhibition of Tyro3 and Mer decreased to some extend. The preferential moderate inhibition of the non-phosphorylated form of Abl1 of some derivatives suggests that they behave as type II inhibitors. This hypothesis was confirmed by docking studies into the structure of Met (3F82) and in a Tyro3 model where key interactions with the hinge region, the DFG-out motif and the allosteric pocket explain this inhibition.

DNA switches on the two-photon efficiency of an ultrabright triphenylamine fluorescent probe specific of AT regions.

We report on the design and synthesis of two-photon fluorescent triphenylamines bearing two or three vinyl branches terminated by a N-methyl benzimidazolium moiety. The new compounds (TP-2Bzim, TP-3Bzim) are light-up fluorescent DNA probes with a long wavelength emission (>580 nm). Compared to their pyridinium models, the TP-Bzim dyes exhibit a remarkable improvement of both their DNA affinity and fluorescence quantum yield, especially for the two-branch derivative (TP-2Bzim: ΦF = 0.54, Ka = 10(7) M(-1)), resulting in a large fluorescence emission turn-on ratio of up to 140. Concomitantly, the two-photon absorption cross-section of TP-2Bzim is dramatically enhanced upon DNA binding (δ = 1080 vs 110 GM for the free form). This effect of the DNA matrix on the nonlinear absorption is uncovered for the first time. This is attributed to a tight fit of the molecule inside the minor groove of AT-rich DNA which induces geometrical rearrangements in the dye ground state as supported by circular dichroism and molecular modeling data. Consequently, TP-2bzim displays an exceptional two-photon molecular brightness (δ×ΦF = 583 GM), a value unrivalled for a small biofluorophore. These properties enable to image nuclear DNA in fixed cells at submicromolar concentration ([TP-2Bzim] = 100 nM) and to visualize ultrabright foci of centromeric AT-rich chromatin. Finally TP-2Bzim exhibits a high photostability, is live-cell permeant, and does not require RNase treatment. This outstanding combination of optical and biological properties makes TP-2Bzim a bioprobe surpassing the best DNA stainers and paves the way for studying further nonlinear optical processes in DNA.

Inhibitors of the TAM subfamily of tyrosine kinases: synthesis and biological evaluation.

The TAM subfamily of Receptor Tyrosine Kinases (RTKs) contains three human proteins of therapeutical interest, Axl, Mer, and Tyro3. Our goal was to design a type II inhibitor specific for this family, i.e. able to interact with the allosteric pocket and with the hinge region of the kinase. We report the synthesis of several series of purine analogues of BMS-777607. The structural diversity of the designed inhibitors was expected to modify the interactions formed in the binding site and consequently to modulate their selectivity profiles. The most potent inhibitor 6g exhibits Kds of 39, 42, 65 and 200 nM against Axl, Mer, Met and Tyro3 respectively. Analysis of the affinity of 6g for active and inactive forms of Abl1, an RTK protein that does not belong to the TAM subfamily, together with the binding modes of 6g predicted by docking studies, indicates that 6g displays some selectivity for the TAM family and may act as a type II inhibitor.

Screening of a chemical library by HT-G4-FID for discovery of selective G-quadruplex binders.

Due to the lack of structural guidelines about G-quadruplex ligands, rational design cannot be the only approach to discover potent G4-ligands. As a complementary approach, screening of chemical library may provide interesting scaffolds known as hits provided that specific tools are available. In this work, the Institut Curie-CNRS chemical library was firstly screened by chemoinformatics methods. Similarity estimations by comparison with reference compounds (Phen-DC3, 360A, MMQ12) provided a set of molecules, which were then evaluated by high-throughput G4-FID (HT-G4-FID) against various G-quadruplex DNA. A full investigation of the most interesting molecules, using the HT-G4-FID assay and molecular modeling, supplied an interesting structure-activity relationship confirming the efficiency of this general approach. Overall, we demonstrated that HT-G4-FID coupled with screening of chemical libraries is a powerful tool to identify new G4-DNA binding scaffolds.

Recognition of G-quadruplex DNA by triangular star-shaped compounds: with or without side chains?

We report the synthesis of two new series of triangular aromatic platforms, either with three aminoalkyl side chains (triazatrinaphthylene series, TrisK: six compounds), or without side chains (triazoniatrinaphthylene, TrisQ). The quadruplex-DNA binding behavior of the two series, which differ essentially by the localization of the cationic charges, was evaluated by means of FRET-melting and G4-FID assays. For the trisubstituted triazatrinaphthylenes (TrisK), the length of the substituents and the presence of terminal hydrogen-bond-donor groups (NH(2)) were shown to be crucial for ensuring a high quadruplex affinity (ΔT(1/2) values of up to 20 °C at 1 µM for the best candidate, TrisK3-NH) and selectivity versus duplex DNA. Subsequently, comparison of data collected on both the telomeric- and c-myc-quadruplex showed that the nonsubstituted TrisQ is even more efficient than TrisK3-NH, both in terms of quadruplex affinity (ΔT(1/2)=26 °C in K(+) buffer) and selectivity versus duplex DNA. Structural considerations conducted with the c-myc quadruplex indicate that both TrisK3-NH and TrisQ stack well onto the G-quartet but in an offset position, which might be influenced by the formation of multiple hydrogen bonds with the target in the former case. Finally, the nonsubstituted TrisQ displays a binding profile very similar to some of the best quadruplex binders, BRACO-19 and bisquinolinium 360A, used herein as references, and thereby represents a highly promising novel molecular design for quadruplex recognition.

"One Ring to Bind Them All"-Part II: Identification of Promising G-Quadruplex Ligands by Screening of Cyclophane-Type Macrocycles.

A collection of 26 polyammonium cyclophane-type macrocycles with a large structural diversity has been screened for G-quadruplex recognition. A two-step selection procedure based on the FRET-melting assay was carried out enabling identification of macrocycles of high affinity (DeltaT(1/2) up to 30 degrees C) and high selectivity for the human telomeric G-quadruplex. The four selected hits possess sophisticated architectures, more particularly the presence of a pendant side-arm as well as the existence of a particular topological arrangement appear to be strong determinants of quadruplex binding. These compounds are thus likely to create multiple contacts with the target that may be at the origin of their high selectivity, thereby suggesting that this class of macrocycles offers unique advantages for targeting G-quadruplex-DNA.

"One ring to bind them all"-part I: the efficiency of the macrocyclic scaffold for g-quadruplex DNA recognition.

Macrocyclic scaffolds are particularly attractive for designing selective G-quadruplex ligands essentially because, on one hand, they show a poor affinity for the "standard" B-DNA conformation and, on the other hand, they fit nicely with the external G-quartets of quadruplexes. Stimulated by the pioneering studies on the cationic porphyrin TMPyP4 and the natural product telomestatin, follow-up studies have developed, rapidly leading to a large diversity of macrocyclic structures with remarkable-quadruplex binding properties and biological activities. In this review we summarize the current state of the art in detailing the three main categories of quadruplex-binding macrocycles described so far (telomestatin-like polyheteroarenes, porphyrins and derivatives, polyammonium cyclophanes), and in addressing both synthetic issues and biological aspects.

Macrocyclic DNA-mismatch-binding ligands: structural determinants of selectivity.

A collection of 15 homodimeric and 5 heterodimeric macrocyclic bisintercalators was prepared by one- or two-step condensation of aromatic dialdehydes with aliphatic diamines; notably, the heterodimeric scaffolds were synthesized for the first time. The binding of these macrocycles to DNA duplexes containing a mispaired thymine residue (TX), as well as to the fully paired control (TA), was investigated by thermal denaturation and fluorescent-intercalator-displacement experiments. The bisnaphthalene derivatives, in particular, the 2,7-disubstituted ones, have the highest selectivity for the TX mismatches, as these macrocycles show no apparent binding to the fully paired DNA. By contrast, other macrocyclic ligands, as well as seven conventional DNA binders, show lesser or no selectivity for the mismatch sites. The study demonstrates that the topology of the ligands plays a crucial role in determining the mismatch-binding affinity and selectivity of the macrocyclic bisintercalators.

3D Pharmacophore, hierarchical methods, and 5-HT4 receptor binding data.

5-Hydroxytryptamine subtype-4 (5-HT(4)) receptors have stimulated considerable interest amongst scientists and clinicians owing to their importance in neurophysiology and potential as therapeutic targets. A comparative analysis of hierarchical methods applied to data from one thousand 5-HT(4) receptor-ligand binding interactions was carried out. The chemical structures were described as chemical and pharmacophore fingerprints. The definitions of indices, related to the quality of the hierarchies in being able to distinguish between active and inactive compounds, revealed two interesting hierarchies with the Unity (1 active cluster) and pharmacophore fingerprints (4 active clusters). The results of this study also showed the importance of correct choice of metrics as well as the effectiveness of a new alternative of the Ward clustering algorithm named Energy (Minimum E-Distance method). In parallel, the relationship between these classifications and a previously defined 3D 5-HT(4) antagonist pharmacophore was established.

G-quadruplex recognition by quinacridines: a SAR, NMR, and biological study.

The synthesis of a novel group of quinacridine-based ligands (MMQs) is described along with an evaluation of their G-quadruplex binding properties. A set of biophysical assays was applied to characterize their interaction with DNA quadruplexes: FRET-melting experiments and equilibrium microdialysis were used to evaluate their quadruplex affinity and their ability to discriminate quadruplexes across a broad panel of DNA structures. All data collected support the proposed model of interaction of these compounds with G-quadruplexes, which is furthermore confirmed by a solution structure determined by 2D NMR experiments. Finally, the activity of the MMQ series against tumor cell growth is reported, and the data support the potential of quadruplex-interactive compounds for use in anticancer approaches.

Molecular modeling studies focused on 5-HT7 versus 5-HT1A selectivity. Discovery of novel phenylpyrrole derivatives with high affinity for 5-HT7 receptors.

The present study discusses the well-known 5-HT7/5-HT1A selectivity issue through a new series of phenylpyrrole derivatives. The first hits emerged from a virtual screening performed on a chemolibrary. Further study led to an optimization of a preliminary 5-HT7 pharmacophore model. The importance of each pharmacophoric feature is confirmed, but these characteristics have to be coupled to geometric constraints in order to achieve a 5-HT7 selectivity. Indeed, 5-HT1A affinity probably arises from extended conformations, whereas a bent one appears to be best suited for 5-HT7 selectivity.

Design, synthesis, and evaluation of a biomimetic artificial photolyase model.

Two new artificial photolyase models that recognize pyrimidine dimers in protic and aprotic organic solvents as well as in water through a combination of charge and hydrogen-bonding interactions and use a mimic of the flavine to achieve repair through reductive photoinduced electron transfer are presented. Fluorescence and NMR titration studies show that it forms a 1:1 complex with pyrimidine dimers with binding constants of approximately 10(3) M(-1) in acetonitrile or methanol, while binding constants in water at pH 7.2 are slightly lower. Excitation of the complex with visible light leads to clean and rapid cycloreversion of the pyrimidine dimer through photoinduced electron transfer catalysis. The reaction in water is significantly faster than in organic solvents. The reaction slows down at higher conversions due to product inhibition.

Sterically crowded bicyclo[1.1.0]butane radical cations.

The variability of carbon-carbon single bonds by steric and electronic effects is probed by DFT calculations of sterically crowded bicyclo[1.1.0]butanes and their radical cations. The interplay of sterics and electronics on the gradual weakening and breaking of bonds was studied by investigating bridgehead substitution in 1,3-di-tert-butylbicyclo[1.1.0]butane and 2,2',4,4'-tetramethyl-1,3-di-tert-butylbicyclo[1.1.0]butane and geminal substitution in 2,2'-di-tert-butylbicyclo[1.1.0]butane and 2,2',4,4'-tetra-tert-butylbicyclo[1.1.0]butane. Bridgehead substitution leads to a lengthening of the central bond, whereas bisubstitution on the geminal carbon leads to a shortening of this bond due to a Thorpe-Ingold effect. Although the character of the central bond can be modulated by substitution and electron transfer over a range of 0.35 A, the state forbidden ring planarization does not occur. Sterically crowded bicyclo[1.1.0]butane radical cations are therefore promising candidates for the investigation of extremely long carbon-carbon single bonds.

Isotope effects and the mechanism of an electron-transfer-catalyzed Diels-Alder reaction.

The electron-transfer-catalyzed Diels-Alder reaction of indole and 1,3-cyclohexadiene was studied by a combination of experimental and theoretical methods. The (13)C kinetic isotope effects were determined at natural abundance by NMR methodology. B3LYP/6-31G* calculations allow for the first time a quantitatively accurate description of the different possible pathways and provide the basis for an analysis of the experimentally observed isotope effects. The computational results, in conjunction with experimental observations, show that the reaction has a stepwise mechanism that is initiated by attack of the diene into the 3-position of the indole. Numerical simulation of the experimentally observed isotope effects shows that the first step is rate-determining and that the electron exchange in the reactant contributes partially to the overall isotope effect. The combination of electronic structure theory, experimental isotope effects, and numerical simulation thus allows a detailed analysis of a complex reaction mechanism.