Toward Developing Techniques─Agnostic Machine Learning Classification Models for Forensically Relevant Glass Fragments.

Glass fragments found in crime scenes may constitute important forensic evidence when properly analyzed, for example, to determine their origin. This analysis could be greatly helped by having a large and diverse database of glass fragments and by using it for constructing reliable machine learning (ML)-based glass classification models. Ideally, the samples that make up this database should be analyzed by a single accurate and standardized analytical technique. However, due to differences in equipment across laboratories, this is not feasible. With this in mind, in this work, we investigated if and how measurement performed at different laboratories on the same set of glass fragments could be combined in the context of ML. First, we demonstrated that elemental analysis methods such as particle-induced X-ray emission (PIXE), laser ablation inductively coupled plasma mass spectrometry (LA-ICP-MS), scanning electron microscopy with energy-dispersive X-ray spectrometry (SEM-EDS), particle-induced Gamma-ray emission (PIGE), instrumental neutron activation analysis (INAA), and prompt Gamma-ray neutron activation analysis (PGAA) could each produce lab-specific ML-based classification models. Next, we determined rules for the successful combinations of data from different laboratories and techniques and demonstrated that when followed, they give rise to improved models, and conversely, poor combinations will lead to poor-performing models. Thus, the combination of PIXE and LA-ICP-MS improves the performances by ∼10-15%, while combining PGAA with other techniques provides poorer performances in comparison with the lab-specific models. Finally, we demonstrated that the poor performances of the SEM-EDS technique, still in use by law enforcement agencies, could be greatly improved by replacing SEM-EDS measurements for Fe and Ca by PIXE measurements for these elements. These findings suggest a process whereby forensic laboratories using different elemental analysis techniques could upload their data into a unified database and get reliable classification based on lab-agnostic models. This in turn brings us closer to a more exhaustive extraction of information from glass fragment evidence and furthermore may form the basis for international-wide collaboration between law enforcement agencies.

The CFTR P67L variant reveals a key role for N-terminal lasso helices in channel folding, maturation, and pharmacologic rescue.

Patients with cystic fibrosis (CF) harboring the P67L variant in the cystic fibrosis transmembrane conductance regulator (CFTR) often exhibit a typical CF phenotype, including severe respiratory compromise. This rare mutation (reported in <300 patients worldwide) responds robustly to CFTR correctors, such as lumacaftor and tezacaftor, with rescue in model systems that far exceed what can be achieved for the archetypical CFTR mutant F508del. However, the specific molecular consequences of the P67L mutation are poorly characterized. In this study, we conducted biochemical measurements following low-temperature growth and/or intragenic suppression, which suggest a mechanism underlying P67L that (1) shares key pathogenic features with F508del, including off-pathway (non-native) folding intermediates, (2) is linked to folding stability of nucleotide-binding domains 1 and 2, and (3) demonstrates pharmacologic rescue that requires domains in the carboxyl half of the protein. We also investigated the "lasso" helices 1 and 2, which occur immediately upstream of P67. Based on limited proteolysis, pulse chase, and molecular dynamics analysis of full-length CFTR and a series of deletion constructs, we argue that P67L and other maturational processing (class 2) defects impair the integrity of the lasso motif and confer misfolding of downstream domains. Thus, amino-terminal missense variants elicit a conformational change throughout CFTR that abrogates maturation while providing a robust substrate for pharmacologic repair.

Stability Prediction for Mutations in the Cytosolic Domains of Cystic Fibrosis Transmembrane Conductance Regulator.

Cystic Fibrosis (CF) is caused by mutations to the Cystic Fibrosis Transmembrane Conductance Regulator (CFTR) chloride channel. CFTR is composed of two membrane spanning domains, two cytosolic nucleotide-binding domains (NBD1 and NBD2) and a largely unstructured R-domain. Multiple CF-causing mutations reside in the NBDs and some are known to compromise the stability of these domains. The ability to predict the effect of mutations on the stability of the cytosolic domains of CFTR and to shed light on the mechanisms by which they exert their effect is therefore important in CF research. With this in mind, we have predicted the effect on domain stability of 59 mutations in NBD1 and NBD2 using 15 different algorithms and evaluated their performances via comparison to experimental data using several metrics including the correct classification rate (CCR), and the squared Pearson correlation (R2) and Spearman's correlation (ρ) calculated between the experimental ΔTm values and the computationally predicted ΔΔG values. Overall, the best results were obtained with FoldX and Rosetta. For NBD1 (35 mutations), FoldX provided R2 and ρ values of 0.64 and -0.71, respectively, with an 86% correct classification rate (CCR). For NBD2 (24 mutations), FoldX R2, ρ, and CCR were 0.51, -0.73, and 75%, respectively. Application of the Rosetta high-resolution protocol (Rosetta_hrp) to NBD1 yielded R2, ρ, and CCR of 0.64, -0.75, and 69%, respectively, and for NBD2 yielded R2, ρ, and CCR of 0.29, -0.27, and 50%, respectively. The corresponding numbers for the Rosetta's low-resolution protocol (Rosetta_lrp) were R2 = 0.47, ρ = -0.69, and CCR = 69% for NBD1 and R2 = 0.27, ρ = -0.24, and CCR = 63% for NBD2. For NBD1, both algorithms suggest that destabilizing mutations suffer from destabilizing vdW clashes, whereas stabilizing mutations benefit from favorable H-bond interactions. Two triple consensus approaches based on FoldX, Rosetta_lpr, and Rosetta_hpr were attempted using either "majority-voting" or "all-voting". The all-voting consensus outperformed the individual predictors, albeit on a smaller data set. In summary, our results suggest that the effect of mutations on the stability of CFTR's NBDs could be largely predicted. Since NBDs are common to all ABC transporters, these results may find use in predicting the effect and mechanism of the action of multiple disease-causing mutations in other proteins.

A Comparison between Enrichment Optimization Algorithm (EOA)-Based and Docking-Based Virtual Screening.

Virtual screening (VS) is a well-established method in the initial stages of many drug and material design projects. VS is typically performed using structure-based approaches such as molecular docking, or various ligand-based approaches. Most docking tools were designed to be as global as possible, and consequently only require knowledge on the 3D structure of the biotarget. In contrast, many ligand-based approaches (e.g., 3D-QSAR and pharmacophore) require prior development of project-specific predictive models. Depending on the type of model (e.g., classification or regression), predictive ability is typically evaluated using metrics of performance on either the training set (e.g.,QCV2) or the test set (e.g., specificity, selectivity or QF1/F2/F32). However, none of these metrics were developed with VS in mind, and consequently, their ability to reliably assess the performances of a model in the context of VS is at best limited. With this in mind we have recently reported the development of the enrichment optimization algorithm (EOA). EOA derives QSAR models in the form of multiple linear regression (MLR) equations for VS by optimizing an enrichment-based metric in the space of the descriptors. Here we present an improved version of the algorithm which better handles active compounds and which also takes into account information on inactive (either known inactive or decoy) compounds. We compared the improved EOA in small-scale VS experiments with three common docking tools, namely, Glide-SP, GOLD and AutoDock Vina, employing five molecular targets (acetylcholinesterase, human immunodeficiency virus type 1 protease, MAP kinase p38 alpha, urokinase-type plasminogen activator, and trypsin I). We found that EOA consistently outperformed all docking tools in terms of the area under the ROC curve (AUC) and EF1% metrics that measured the overall and initial success of the VS process, respectively. This was the case when the docking metrics were calculated based on a consensus approach and when they were calculated based on two different sets of single crystal structures. Finally, we propose that EOA could be combined with molecular docking to derive target-specific scoring functions.

A defective flexible loop contributes to the processing and gating defects of the predominant cystic fibrosis-causing mutation.

People with the genetic disease cystic fibrosis (CF) often carry a deletion mutation ΔF508 on the gene encoding the CF transmembrane conductance regulator (CFTR) Cl- channel. This mutation greatly reduces the CFTR maturation process and slows the channel opening rate. Here, we investigate whether residues near F508 contribute to these defects in ΔF508-CFTR. Most deletion mutations, but not alanine substitutions, of individual residues from positions 503 to 513 impaired CFTR maturation. Interestingly, only protein processing of ΔY512-CFTR, like that of ΔF508-CFTR, was greatly improved by low-temperature culture at 27°C or small-molecule corrector C18. The 2 mutant Cl- channels were equally slow to open, suggesting that they may share common structural flaws. Studies on the H3-H4 loop that links residues F508 and Y512 demonstrate that G509A/V510G mutations, moving G509 1 position backward in the loop, markedly enhanced ΔF508-CFTR maturation and opening rate while promoting protein stability and persistence of the H3 helix in ΔF508 nucleotide-binding domain 1. Moreover, V510A/S511A mutations noticeably increased ΔY512-CFTR maturation at 27°C and its opening rate. Thus, loop abnormalities may contribute to ΔF508- and ΔY512-CFTR defects. Importantly, correcting defects from G509 displacement in ΔF508-CFTR may offer a new avenue for drug discovery and CF treatments.-Chen, X., Zhu, S., Zhenin, M., Xu, W., Bose, S. J., Wong, M. P.-F., Leung, G. P. H., Senderowitz, H., Chen, J.-H. A defective flexible loop contributes to the processing and gating defects of the predominant cystic fibrosis-causing mutation.

Evaluation of QSAR Equations for Virtual Screening.

Quantitative Structure Activity Relationship (QSAR) models can inform on the correlation between activities and structure-based molecular descriptors. This information is important for the understanding of the factors that govern molecular properties and for designing new compounds with favorable properties. Due to the large number of calculate-able descriptors and consequently, the much larger number of descriptors combinations, the derivation of QSAR models could be treated as an optimization problem. For continuous responses, metrics which are typically being optimized in this process are related to model performances on the training set, for example, R2 and QCV2. Similar metrics, calculated on an external set of data (e.g., QF1/F2/F32), are used to evaluate the performances of the final models. A common theme of these metrics is that they are context -" ignorant". In this work we propose that QSAR models should be evaluated based on their intended usage. More specifically, we argue that QSAR models developed for Virtual Screening (VS) should be derived and evaluated using a virtual screening-aware metric, e.g., an enrichment-based metric. To demonstrate this point, we have developed 21 Multiple Linear Regression (MLR) models for seven targets (three models per target), evaluated them first on validation sets and subsequently tested their performances on two additional test sets constructed to mimic small-scale virtual screening campaigns. As expected, we found no correlation between model performances evaluated by "classical" metrics, e.g., R2 and QF1/F2/F32 and the number of active compounds picked by the models from within a pool of random compounds. In particular, in some cases models with favorable R2 and/or QF1/F2/F32 values were unable to pick a single active compound from within the pool whereas in other cases, models with poor R2 and/or QF1/F2/F32 values performed well in the context of virtual screening. We also found no significant correlation between the number of active compounds correctly identified by the models in the training, validation and test sets. Next, we have developed a new algorithm for the derivation of MLR models by optimizing an enrichment-based metric and tested its performances on the same datasets. We found that the best models derived in this manner showed, in most cases, much more consistent results across the training, validation and test sets and outperformed the corresponding MLR models in most virtual screening tests. Finally, we demonstrated that when tested as binary classifiers, models derived for the same targets by the new algorithm outperformed Random Forest (RF) and Support Vector Machine (SVM)-based models across training/validation/test sets, in most cases. We attribute the better performances of the Enrichment Optimizer Algorithm (EOA) models in VS to better handling of inactive random compounds. Optimizing an enrichment-based metric is therefore a promising strategy for the derivation of QSAR models for classification and virtual screening.

Discovery and Design of Novel Small Molecule GSK-3 Inhibitors Targeting the Substrate Binding Site.

The serine/threonine kinase, GSK-3, is a promising drug discovery target for treating multiple pathological disorders. Most GSK-3 inhibitors that were developed function as ATP competitive inhibitors, with typical limitations in specificity, safety and drug-induced resistance. In contrast, substrate competitive inhibitors (SCIs), are considered highly selective, and more suitable for clinical practice. The development of SCIs has been largely neglected in the past because the ambiguous, undefined nature of the substrate-binding site makes them difficult to design. In this study, we used our previously described structural models of GSK-3 bound to SCI peptides, to design a pharmacophore model and to virtually screen the "drug-like" Zinc database (~6.3 million compounds). We identified leading hits that interact with critical binding elements in the GSK-3 substrate binding site and are chemically distinct from known GSK-3 inhibitors. Accordingly, novel GSK-3 SCI compounds were designed and synthesized with IC50 values of~1-4 µM. Biological activity of the SCI compound was confirmed in cells and in primary neurons that showed increased ß-catenin levels and reduced tau phosphorylation in response to compound treatment. We have generated a new type of small molecule GSK-3 inhibitors and propose to use this strategy to further develop SCIs for other protein kinases.

Ligand binding to a remote site thermodynamically corrects the F508del mutation in the human cystic fibrosis transmembrane conductance regulator.

Many disease-causing mutations impair protein stability. Here, we explore a thermodynamic strategy to correct the disease-causing F508del mutation in the human cystic fibrosis transmembrane conductance regulator (hCFTR). F508del destabilizes nucleotide-binding domain 1 (hNBD1) in hCFTR relative to an aggregation-prone intermediate. We developed a fluorescence self-quenching assay for compounds that prevent aggregation of hNBD1 by stabilizing its native conformation. Unexpectedly, we found that dTTP and nucleotide analogs with exocyclic methyl groups bind to hNBD1 more strongly than ATP and preserve electrophysiological function of full-length F508del-hCFTR channels at temperatures up to 37 °C. Furthermore, nucleotides that increase open-channel probability, which reflects stabilization of an interdomain interface to hNBD1, thermally protect full-length F508del-hCFTR even when they do not stabilize isolated hNBD1. Therefore, stabilization of hNBD1 itself or of one of its interdomain interfaces by a small molecule indirectly offsets the destabilizing effect of the F508del mutation on full-length hCFTR. These results indicate that high-affinity binding of a small molecule to a remote site can correct a disease-causing mutation. We propose that the strategies described here should be applicable to identifying small molecules to help manage other human diseases caused by mutations that destabilize native protein conformation.

Novel inhibitors of leukocyte transendothelial migration.

Leukocyte transendothelial migration is one of the most important step in launching an inflammatory immune response and chronic inflammation can lead to devastating diseases. Leukocyte migration inhibitors are considered as promising and potentially effective therapeutic agents to treat inflammatory and auto-immune disorders. In this study, based on previous trioxotetrahydropyrimidin based integrin inhibitors that suboptimally blocked leukocyte adhesion, twelve molecules with a modified scaffold were designed, synthesized, and tested in vitro for their capacity to block the transendothelial migration of immune cells. One of the molecules, namely, methyl 4-((2-(tert-butyl)-6-((2,4,6-trioxotetrahydropyrimidin-5(2H)-ylidene) methyl) phenoxy) methyl) benzoate, (compound 12), completely blocked leukocyte transendothelial migration, without any toxic effects on immune or endothelial cells (IC50 = 2.4 µM). In vivo, compound 12 exhibited significant therapeutic effects in inflammatory bowel disease (IBD)/Crohn's disease, multiple sclerosis, fatty liver disease, and rheumatoid arthritis models. A detailed acute and chronic toxicity profile of the lead compound in vivo did not reveal any toxic effects. Such a type of molecule might therefore provide a unique starting point for designing a novel class of leukocyte transmigration blocking agents with broad therapeutic applications in inflammatory and auto-immune pathologies.

Nrf2 Activation by SK-119 Attenuates Oxidative Stress, UVB, and LPS-Induced Damage.

BACKGROUND/AIMS: The Nrf2 signaling pathway plays a pivotal role in neutralizing excess reactive oxygen species formation and therefore enhancing the endogenous cellular protection mechanism. Thus, activating this pathway may provide therapeutic options against oxidative stress-related disorders. We have recently applied a computer-aided drug design approach to the design and synthesis of novel Nrf2 enhancers. The current study was aimed at investigating the potential beneficial impact of (E)-5-oxo-1-(4-((2,4,6-trihydroxybenzylidene)amino)phenyl)pyrrolidine-3-carboxylic acid (SK-119) in skin oxidative damage models. METHODS: SK-119, tested initially in PC-12 cells, attenuated oxidative stress-induced cytotoxicity concomitantly with Nrf2 activation. The potential impact of this compound was evaluated in skin-based disease models both in vitro (HaCaT cells) and ex vivo (human skin organ culture). RESULTS: The data clearly showed the marked anti-inflammatory and photoprotection properties of the compound; SK-119-treated cells or tissues displayed a reduction in cytokine secretion induced by lipopolysaccharides (LPS) in a manner comparable with dexamethasone. In addition, topical application of SK-119 was able to block UVB-induced oxidative stress and attenuated caspase-mediated apoptosis, DNA adduct formation, and the concomitant cellular damage. CONCLUSION: These results indicate that SK-119 is an Nrf2 activator that can be used as a prototype molecule for the development of novel treatments of dermatological disorders related to oxidative stress.

Structural stability of purified human CFTR is systematically improved by mutations in nucleotide binding domain 1.

The Cystic Fibrosis Transmembrane Conductance Regulator (CFTR) is an ABC transporter containing two transmembrane domains forming a chloride ion channel, and two nucleotide binding domains (NBD1 and NBD2). CFTR has presented a formidable challenge to obtain monodisperse, biophysically stable protein. Here we report a comprehensive study comparing effects of single and multiple NBD1 mutations on stability of both the NBD1 domain alone and on purified full length human CFTR. Single mutations S492P, A534P, I539T acted additively, and when combined with M470V, S495P, and R555K cumulatively yielded an NBD1 with highly improved structural stability. Strategic combinations of these mutations strongly stabilized the domain to attain a calorimetric Tm > 70 °C. Replica exchange molecular dynamics simulations on the most stable 6SS-NBD1 variant implicated fluctuations, electrostatic interactions and side chain packing as potential contributors to improved stability. Progressive stabilization of NBD1 directly correlated with enhanced structural stability of full-length CFTR protein. Thermal unfolding of the stabilized CFTR mutants, monitored by changes in intrinsic fluorescence, demonstrated that Tm could be shifted as high as 67.4 °C in 6SS-CFTR, more than 20 °C higher than wild-type. H1402S, an NBD2 mutation, conferred CFTR with additional thermal stability, possibly by stabilizing an NBD-dimerized conformation. CFTR variants with NBD1-stabilizing mutations were expressed at the cell surface in mammalian cells, exhibited ATPase and channel activity, and retained these functions to higher temperatures. The capability to produce enzymatically active CFTR with improved structural stability amenable to biophysical and structural studies will advance mechanistic investigations and future cystic fibrosis drug development.

Visualization of Solar Cell Library Space by Dimensionality Reduction Methods.

Visualizing high-dimensional data by projecting them into a two- or three-dimensional space is a popular approach in many scientific fields, including computer-aided drug design and cheminformatics. In contrast, dimensionality reduction techniques have been far less explored for materials informatics. Nevertheless, similar to their usefulness in analyzing the space of, e.g., drug-like molecules, such techniques could provide useful insights on materials space, including an intuitive grasp of the overall distribution of samples, the identification of interesting trends, including the formation of materials clusters and the presence of activity cliffs and outliers, and rational navigation through this space in the search for new materials. Here we present the first application of four dimensionality reduction techniques, namely, principal component analysis (PCA), kernel PCA, Isomap, and diffusion map, to visualize and analyze a part of the materials space populated by solar cells made of metal oxides. Solar cells in general and metal-oxide-based solar cells in particular hold the promise of contributing to the world's search for clean and affordable energy resources. With the exception of PCA, these methods have seldom been used to visualize chemistry space and almost never been used to visualize materials space. For this purpose, we integrated five metal-oxide-based solar cell libraries into a uniform database and subjected it to dimensionality reduction by all four methods, comparing their performances using various criteria such as maintaining the local environment of samples and the clustering structure in the low-dimensional space. We also looked at the number of outliers produced by each method and analyzed common outliers. We found that PCA performs best in terms of the ability to correctly maintain the local environment of samples, whereas Isomap does the best job of assigning class membership on the basis of the identities of nearest neighbors (i.e., it is the best classifier). We also found that many of the outliers identified by all of the methods could be rationalized. We suggest that the methods used in this work could be extended to study other types of solar cells, thereby setting the ground for further analysis of the photovoltaic (PV) space as well as other regions of materials space.

Rescoring of docking poses under Occam's Razor: are there simpler solutions?

Ligand affinity prediction from docking simulations is usually performed by means of highly empirical and diverse protocols. These protocols often involve the re-scoring of poses generated by a force field (FF) based Hamiltonian to provide either estimated binding affinities-or alternatively, some empirical goodness score. Re-scoring is performed by so-called scoring functions-typically, a reweighted sum of FF terms augmented by additional terms (e.g., desolvation/entropic penalty, hydrophobicity, aromatic interactions etc.). Sometimes, the scoring function actually drives ligand positioning, but often it only operates on the best scoring poses ranked top by the initial ligand positioning tool. In either of these rather intricate scenarios, scoring functions are docking-specific models, and most require machine-learning-based calibration. Therefore, docking simulations are less straightforward when compared to "standard" molecular simulations in which the FF Hamiltonian defines the energy, and affinity emerges as an ensemble average property over pools of representative conformers (i.e., the trajectory). Paraphrasing on Occam's Razor principle, additional model complexity is only acceptable if demonstrated to bring a significant improvement of prediction quality. In this work we therefore examined whether the complexity inherent to scoring functions is indeed justified. For this purpose we compared sampler for multiple protein-ligand entities, a general purpose conformation sampler based on the AMBER/GAFF FF, complemented with continuum solvation terms, with several state of the art docking tools that rely on calibrated scoring functions (Glide, Gold, Autodock-Vina) in terms of its ability to top-rank the actives from large and diverse ligand series associated with various proteins. There is no clear winner of this study, where each program performed well on most of the targets, but also failed with respect to at least one of them. Therefore, a well-parameterized force field with a simple, energy-based ligand ranking protocol appears to be an as effective docking protocol as intricate rescoring strategies based on scoring functions. A tool that can sample the conformational space of the free ligand, the bound ligand and the protein binding site using the same force field may avoid many of the approximations common to contemporary docking protocols and allow e.g., for docking into highly flexible active sites, when current scoring functions are not well suited to estimate receptor strain energies.

A novel image-based high-throughput screening assay discovers therapeutic candidates for adult polyglucosan body disease.

Glycogen storage disorders (GSDs) are caused by excessive accumulation of glycogen. Some GSDs [adult polyglucosan (PG) body disease (APBD), and Tarui and Lafora diseases] are caused by intracellular accumulation of insoluble inclusions, called PG bodies (PBs), which are chiefly composed of malconstructed glycogen. We developed an APBD patient skin fibroblast cell-based assay for PB identification, where the bodies are identified as amylase-resistant periodic acid-Schiff's-stained structures, and quantified. We screened the DIVERSet CL 10 084 compound library using this assay in high-throughput format and discovered 11 dose-dependent and 8 non-dose-dependent PB-reducing hits. Approximately 70% of the hits appear to act through reducing glycogen synthase (GS) activity, which can elongate glycogen chains and presumably promote PB generation. Some of these GS inhibiting hits were also computationally predicted to be similar to drugs interacting with the GS activator protein phosphatase 1. Our work paves the way to discovering medications for the treatment of PB-involving GSD, which are extremely severe or fatal disorders.

Specific stabilization of CFTR by phosphatidylserine.

The Cystic Fibrosis Transmembrane Conductance Regulator (CFTR, ABCC7) is a plasma membrane chloride ion channel in the ABC transporter superfamily. CFTR is a key target for cystic fibrosis drug development, and its structural elucidation would advance those efforts. However, the limited in vivo and in vitro stability of the protein, particularly its nucleotide binding domains, has made structural studies challenging. Here we demonstrate that phosphatidylserine uniquely stimulates and thermally stabilizes the ATP hydrolysis function of purified human CFTR. Among several lipids tested, the greatest stabilization was observed with brain phosphatidylserine, which shifted the Tm for ATPase activity from 22.7±0.8°C to 35.0±0.2°C in wild-type CFTR, and from 26.6±0.7°C to 42.1±0.2°C in a more stable mutant CFTR having deleted regulatory insertion and S492P/A534P/I539T mutations. When ATPase activity was measured at 37°C in the presence of brain phosphatidylserine, Vmax for wild-type CFTR was 240±60nmol/min/mg, a rate higher than previously reported and consistent with rates for other purified ABC transporters. The significant thermal stabilization of CFTR by phosphatidylserine may be advantageous in future structural and biophysical studies of CFTR.

Molecular Dynamics Flexible Fitting Simulations Identify New Models of the Closed State of the Cystic Fibrosis Transmembrane Conductance Regulator Protein.

Cystic fibrosis (CF) is a lethal, genetic disease found in particular in humans of European origin which is caused by mutations in the cystic fibrosis transmembrane conductance regulator (CFTR) chloride channel. The search for CF therapies acting by modulating the impaired function of mutant CFTR will be greatly advanced by high resolution structures of CFTR in different states. To date, two medium resolution electron microscopy (EM) structures of CFTR are available (one of a distant zebrafish (Danio rerio) CFTR ortholog and one of human CFTR). The two models are nearly identical to one another, and both correspond to the inward-facing, nucleotide binding domains (NBDs) separated, closed state of the channel. In addition, lower resolution structural data are available for human CFTR in an alternative conformation which likely features associated NBDs and thus geometrically resembles the conducting state of the channel. Multiple homology models of human CFTR in multiple states have been developed over the years, yet their correspondence to the existing structural information is unexplored. In this work we use molecular dynamics flexible fitting (MDFF) simulations to refine two previously described CFTR models based on the available cryo-EM map of the human protein. This map was recorded in the absence of ATP and consequently represents closed-state CFTR yet its features likely correspond to an NBD associated conformation of the protein. Accordingly, the resulting models feature dimerized NBDs yet with no membrane traversing pore. Moreover, the open probability of the new models as deduced from the MDFF trajectories is significantly lower than that deduced from control MD trajectories initiated from the starting models. We propose that the new models correspond to a CFTR conformation which to date was largely unexplored yet is one that is relevant to the gating cycle of the protein. In particular this conformation may participate in rapid channel opening and closing through small allosteric movements controlled by nucleotide binding and dissociation events. Analyzing the resulting trajectories (and not only the final models as is usually the case), we demonstrate that the refined models have good stereochemical properties and are also in favorable agreement with multiple experimental data. Moreover, despite different starting points, the final models share many common features. Finally, we propose that the combination of high resolution cryo-EM maps, which are currently emerging from multiple laboratories, and MDFF simulations will be of value for the development of yet more reliable CFTR models as well as for the identification of binding sites for CFTR modulators.

Potentiators exert distinct effects on human, murine, and Xenopus CFTR.

VX-770 (Ivacaftor) has been approved for clinical usage in cystic fibrosis patients with several CFTR mutations. Yet the binding site(s) on CFTR for this compound and other small molecule potentiators are unknown. We hypothesize that insight into this question could be gained by comparing the effect of potentiators on CFTR channels from different origins, e.g., human, mouse, and Xenopus (frog). In the present study, we combined this comparative molecular pharmacology approach with that of computer-aided drug discovery to identify and characterize new potentiators of CFTR and to explore possible mechanism of action. Our results demonstrate that 1) VX-770, NPPB, GlyH-101, P1, P2, and P3 all exhibited ortholog-specific behavior in that they potentiated hCFTR, mCFTR, and xCFTR with different efficacies; 2) P1, P2, and P3 potentiated hCFTR in excised macropatches in a manner dependent on the degree of PKA-mediated stimulation; 3) P1 and P2 did not have additive effects, suggesting that these compounds might share binding sites. Also 4) using a pharmacophore modeling approach, we identified three new potentiators (IOWH-032, OSSK-2, and OSSK-3) that have structures similar to GlyH-101 and that also exhibit ortholog-specific potentiation of CFTR. These could potentially serve as lead compounds for development of new drugs for the treatment of cystic fibrosis. The ortholog-specific behavior of these compounds suggest that a comparative pharmacology approach, using cross-ortholog chimeras, may be useful for identification of binding sites on human CFTR.

k-Nearest neighbors optimization-based outlier removal.

Datasets of molecular compounds often contain outliers, that is, compounds which are different from the rest of the dataset. Outliers, while often interesting may affect data interpretation, model generation, and decisions making, and therefore, should be removed from the dataset prior to modeling efforts. Here, we describe a new method for the iterative identification and removal of outliers based on a k-nearest neighbors optimization algorithm. We demonstrate for three different datasets that the removal of outliers using the new algorithm provides filtered datasets which are better than those provided by four alternative outlier removal procedures as well as by random compound removal in two important aspects: (1) they better maintain the diversity of the parent datasets; (2) they give rise to quantitative structure activity relationship (QSAR) models with much better prediction statistics. The new algorithm is, therefore, suitable for the pretreatment of datasets prior to QSAR modeling.

REMD Simulations Reveal the Dynamic Profile and Mechanism of Action of Deleterious, Rescuing, and Stabilizing Perturbations to NBD1 from CFTR.

Cystic Fibrosis (CF) is a lethal, genetic disease caused by mutations to the CFTR chloride channel. The most common CF causing mutation is the deletion of F508 from the first Nucleotide Binding Domain (F508del-NBD1). This mutation leads to a thermally unstable domain and a misfolded, nonfunctioning CFTR. Replica Exchange MD simulations were used to simulate seven NBD1 constructs including wt and F508del-NBD1 both alone and in the presence of known rescuing mutations as well as F508del-NBD1 in complex with a known small (ligand) stabilizer. Analyzing the resulting trajectories suggests that differences in the biochemical properties of the constructs result from local and coupled differences in their dynamic profiles. A comparative analysis of these profiles as well as of the resulting trajectories reveals how the different perturbations exert their deleterious, rescuing, and stabilizing effects on NBD1. These simulations may therefore be useful for the design and mechanism-of-action analysis of new NBD1 stabilizers.

Toward Focusing Conformational Ensembles on Bioactive Conformations: A Molecular Mechanics/Quantum Mechanics Study.

The identification of bound conformations, namely, conformations adopted by ligands when binding their target is critical for target-based and ligand-based drug design. Bound conformations could be obtained computationally from unbound conformational ensembles generated by conformational search tools. However, these tools also generate many nonrelevant conformations thus requiring a focusing mechanism. To identify such a mechanism, this work focuses on a comparison of energies and structural properties of bound and unbound conformations for a set of FDA approved drugs whose complexes are available in the PDB. Unbound conformational ensembles were initially obtained with three force fields. These were merged, clustered, and reminimized using the same force fields and four QM methods. Bound conformations of all ligands were represented by their crystal structures or by approximations to these structures. Energy differences were calculated between global minima of the unbound state or the Boltzmann averaged energies of the unbound ensemble and the approximated bound conformations. Ligand conformations which resemble the X-ray conformation (RMSD < 1.0 Å) were obtained in 91%-97% and 96%-98% of the cases using the ensembles generated by the individual force fields and the reminimized ensembles, respectively, yet only in 52%-56% (original ensembles) and 47%-65% (reminimized ensembles) as global energy minima. The energy window within which the different methods identified the bound conformation (approximated by its closest local energy minimum) was found to be at 4-6 kcal/mol with respect to the global minimum and marginally lower with respect to a Boltzmann averaged energy of the unbound ensemble. Better approximations to the bound conformation obtained with a constrained minimization using the crystallographic B-factors or with a newly developed Knee Point Detection (KPD) method gave lower values (2-5 kcal/mol). Overall, QM methods gave lower energy differences than force field methods. These energy thresholds could be used for focusing conformational ensembles on bound conformations. For example, when using energy cutoffs which corresponded to retaining 50% and 70% of the ensembles, QM methods and CHARMm offer 60-65% and 80-84% probability of obtaining the bound conformation, respectively. In contrast, none of the structural criteria considered in this work was able to differentiate between bound and unbound conformations.