Noise reduction method for molecular interaction energy: application to in silico drug screening and in silico target protein screening

We developed a new method to improve the accuracy of molecular interaction data using a molecular interaction matrix. This method was applied to enhance the database enrichment of in silico drug screening and in silico target protein screening using a protein-compound affinity matrix calculated by a protein-compound docking software. Our assumption was that the protein-compound binding free energy of a compound could be improved by a linear combination of its docking scores with many different proteins. We proposed two approaches to determine the coefficients of the linear combination. The first approach is based on similarity among the proteins, and the second is a machine-learning approach based on the known active compounds. These methods were applied to in silico screening of the active compounds of several target proteins and in silico target protein screening.     

Improvement of protein-compound docking scores by using amino-acid sequence similarities of proteins

The low accuracy of predicted docking scores is critical at in silico drug screening. In order to improve the accuracy of docking scores, we approximated the protein-compound binding free energy as a linear combination of the raw docking scores of a target compound with many different protein pockets. The coefficients of the linear combination were estimated by the similarities among proteins, simply by using the amino-acid sequence similarities or identities of the proteins. This method was applied to in silico screening of the active compounds of five target proteins, and it increased the hit ratio by approximately four to five times compared to that given only by the raw docking scores in every case. The hit ratio also became robust against differences of target proteins.     

Selection of in silico drug screening results for G-protein-coupled receptors by using universal active probes

We developed a new protocol for in silico drug screening for G-protein-coupled receptors (GPCRs) using a set of "universal active probes" (UAPs) with an ensemble docking procedure. UAPs are drug-like compounds, which are actual active compounds of a variety of known proteins. The current targets were nine human GPCRs whose three-dimensional (3D) structures are unknown, plus three GPCRs, namely β(2)-adrenergic receptor (ADRB2), A(2A) adenosine receptor (A(2A)), and dopamine D3 receptor (D(3)), whose 3D structures are known. Homology-based models of the GPCRs were constructed based on the crystal structures with careful sequence inspection. After subsequent molecular dynamics (MD) simulation taking into account the explicit lipid membrane molecules with periodic boundary conditions, we obtained multiple model structures of the GPCRs. For each target structure, docking-screening calculations were carried out via the ensemble docking procedure, using both true active compounds of the target proteins and the UAPs with the multiple target screening (MTS) method. Consequently, the multiple model structures showed various screening results with both poor and high hit ratios, the latter of which could be identified as promising for use in in silico screening to find candidate compounds to interact with the proteins. We found that the hit ratio of true active compounds showed a positive correlation to that of the UAPs. Thus, we could retrieve appropriate target structures from the GPCR models by applying the UAPs, even if no active compound is known for the GPCRs. Namely, the screening result that showed a high hit ratio for the UAPs could be used to identify actual hit compounds for the target GPCRs.     

Going Retro,Going Viral: Experiences and Lessons in Drug Discovery from COVID-19

The severity of the COVID-19 pandemic and the pace of its global spread have motivated researchers to opt for repurposing existing drugs against SARS-CoV-2 rather than discover or develop novel ones. For reasons of speed, throughput, and cost-effectiveness, virtual screening campaigns, relying heavily on in silico docking, have dominated published reports. A particular focus as a drug target has been the principal active site (i.e., RNA synthesis) of RNA-dependent RNA polymerase (RdRp), despite the existence of a second, and also indispensable, active site in the same enzyme. Here we report the results of our experimental interrogation of several small-molecule inhibitors, including natural products proposed to be effective by in silico studies. Notably, we find that two antibiotics in clinical use, fidaxomicin and rifabutin, inhibit RNA synthesis by SARS-CoV-2 RdRp in vitro and inhibit viral replication in cell culture. However, our mutagenesis studies contradict the binding sites predicted computationally. We discuss the implications of these and other findings for computational studies predicting the binding of ligands to large and flexible protein complexes and therefore for drug discovery or repurposing efforts utilizing such studies. Finally, we suggest several improvements on such efforts ongoing against SARS-CoV-2 and future pathogens as they arise.     

Two-dimensional turbulent flow chromatography coupled on-line to liquid chromatography–mass spectrometry for solution-based ligand screening against multiple proteins

We present herein a novel bioseparation/chemical analysis strategy for protein–ligand screening and affinity ranking in compound mixtures, designed to increase screening rates and improve sensitivity and ruggedness in performance. The strategy is carried out by combining on-line two-dimensional turbulent flow chromatography (2D-TFC) with liquid chromatography–mass spectrometry (LC–MS), and accomplished through the following steps: (1) a reversed-phase TFC stage to separate the protein/ligand complex from the unbound free molecules, (2) an on-line dissociation process to release the bound ligands from the complexes, and (3) a second mixed-mode cation-exchange/reversed-phase TFC stage to trap the bound ligands and to remove the proteins and salts, followed by LC–MS analysis for identification and determination of the binding affinities. The technique can implement an ultra-fast isolation of protein/ligand complex with the retention time of a complex peak in about 5 s, and on-line prepare the “clean” sample to be directly compatible with the LC–MS analysis. The improvement in performance of this 2D-TFC/LC–MS approach over the conventional approach has been demonstrated by determining affinity-selected ligands of the target proteins acetylcholinesterase and butyrylcholinesterase from a small library with known binding affinities and a steroidal alkaloid library composed of structurally similar compounds. Our results show that 2D-TFC/LC–MS is a generic and efficient tool for high-throughput screening of ligands with low-to-high binding affinities, and structure-activity relationship evaluation.     

泉生热袍菌结构基因组的选靶研究

对泉生热袍菌进行了结构基因组的选靶研究，从泉生热袍菌的蛋白组中挑选了20个蛋白质作为第一批进行结构测定的目标，以发现新的蛋白质折叠模式. 选靶研究主要使用了BLAST搜索， PSI-BLAST搜索和ProtoNet数据库搜索等方法. 另外，还用PredictProtein程序对选中的蛋白质进行了二级结构和外形预测. 选中的20个蛋白质中有8个被克隆、表达和纯化，其中2个得到了单晶并收集了X衍射数据. 实验结果和最近一些文献报道的结果表明，挑选的一些蛋白质具有新的折叠模式，表明了这种选靶策略的有效性.     

Towards the development of synthetic routes using theoretical calculations: an application of in silico screening to 2,6-dimethylchroman-4-one

This study describes an attempt to develop a synthetic route using theoretical calculations, i.e., in silico synthesis route development. The KOSP program created four potential synthetic routes for generating 2,6-dimethylchroman-4-one. In silico screening of these four synthetic routes was then performed. In silico screening involves theoretical analysis of synthetic routes prior to actual experimental work. A synthetic route using the Mitsunobu reaction had already been reported by Hoddgets et al. Theoretical investigations were also conducted on two S(N)Ar reactions as well as a Michael reaction before they were examined experimentally. In silico screening using DFT calculations indicated that only the Michael reaction was likely to produce the target. Experimental work confirmed that the target was obtained in a yield of 76.4% using the Michael reaction. The other two routes, except for the Mitsunobu reaction, failed to generate the target. Our results demonstrate that theoretical calculations can be used to narrow down the number of experiments that need to be conducted when developing novel synthetic routes.     

Identification of New Potential Inhibitors of Quorum Sensing through a Specialized Multi-Level Computational Approach

Biofilms are aggregates of microorganisms anchored to a surface and embedded in a self-produced matrix of extracellular polymeric substances and have been associated with 80% of all bacterial infections in humans. Because bacteria in biofilms are less amenable to antibiotic treatment, biofilms have been associated with developing antibiotic resistance, a problem that urges developing new therapeutic options and approaches. Interfering with quorum-sensing (QS), an important process of cell-to-cell communication by bacteria in biofilms is a promising strategy to inhibit biofilm formation and development. Here we describe and apply an in silico computational protocol for identifying novel potential inhibitors of quorum-sensing, using CviR—the quorum-sensing receptor from Chromobacterium violaceum—as a model target. This in silico approach combines protein-ligand docking (with 7 different docking programs/scoring functions), receptor-based virtual screening, molecular dynamic simulations, and free energy calculations. Particular emphasis was dedicated to optimizing the discrimination ability between active/inactive molecules in virtual screening tests using a target-specific training set. Overall, the optimized protocol was used to evaluate 66,461 molecules, including those on the ZINC/FDA-Approved database and to the Mu.Ta.Lig Virtual Chemotheca. Multiple promising compounds were identified, yielding good prospects for future experimental validation and for drug repurposing towards QS inhibition.     

Protein structure prediction and biomolecular recognition: from protein sequence to peptidomimetic design with the human beta3 integrin

Computational tools can bridge the gap between sequence and protein 3D structure based on the notion that information is to be retrieved from the databases and that knowledge-based methods can help in approaching a solution of the protein-folding problem. To this aim our group has implemented neural network-based predictors capable of performing with some success in different tasks, including predictions of the secondary structure of globular and membrane proteins, the topology of membrane proteins and porins and stable alpha-helical segments suited for protein design. Moreover we have developed methods for predicting contact maps in proteins and the probability of finding a cysteine in a disulfide bridge, tools which can contribute to the goal of predicting the 3D structure starting from the sequence (the so called ab initio prediction). All our predictors take advantage of evolution information derived from the structural alignments of homologous (evolutionary related) proteins and taken from the sequence and structure databases. When it is necessary to build models for proteins of unknown spatial structure, which have very little homology with other proteins of known structure, non-standard techniques need to be developed and the tools for protein structure predictions may help in protein modeling. The results of a recent simulation performed in our lab highlights the role of high performing computing technology and of tools of computational biology in protein modeling and peptidomimetic design.     

Comprehensive study of proteins that interact with microcystin-LR

We carried out a comprehensive study of proteins that exhibit specific interactions with a naturally occurring toxin, microcystin (MC)-LR, in order to gain insight into the unknown underlying mechanism of MC virulence. This audacious study employed a simple affinity test that used MC-LR immobilized on an original ethylene oxide based monolithic solid phase (Moli-gel), and swine liver lysate. Some of the proteins that interacted with MC-LR on this original affinity resin were separated by SDS-PAGE, measured by nano-LC/MS/MS after trypsin digestion, and identified using a Mascot database search. Protein sequence analyses revealed that glutathione S-transferase (GST) was one of the candidate target proteins for MC-LR. This protein was confirmed as a target protein for MC-LR based on the results of for the inhibition of an enzymatic reaction by Dhb-MC-LR. Moreover, L-3-hydroxyacyl coenzyme A dehydrogenase (HDHA) was shown to be one of the proteins that specifically interacts with MC-LR. Our results demonstrated that our analytical systems based on an original affinity resin and nano-LC/MS/MS were effective for target protein research.     

Dihydrofolate Reductase Inhibitors: The Pharmacophore as a Guide for Co-Crystal Screening

In this work, co-crystal screening was carried out for two important dihydrofolate reductase (DHFR) inhibitors, trimethoprim (TMP) and pyrimethamine (PMA), and for 2,4-diaminopyrimidine (DAP), which is the pharmacophore of these active pharmaceutical ingredients (API). The isomeric pyridinecarboxamides and two xanthines, theophylline (THEO) and caffeine (CAF), were used as co-formers in the same experimental conditions, in order to evaluate the potential for the pharmacophore to be used as a guide in the screening process. In silico co-crystal screening was carried out using BIOVIA COSMOquick and experimental screening was performed by mechanochemistry and supported by (solid + liquid) binary phase diagrams, infrared spectroscopy (FTIR) and X-ray powder diffraction (XRPD). The in silico prediction of low propensities for DAP, TMP and PMA to co-crystallize with pyridinecarboxamides was confirmed: a successful outcome was only observed for DAP + nicotinamide. Successful synthesis of multicomponent solid forms was achieved for all three target molecules with theophylline, with DAP co-crystals revealing a greater variety of stoichiometries. The crystalline structures of a (1:2) TMP:THEO co-crystal and of a (1:2:1) DAP:THEO:ethyl acetate solvate were solved. This work demonstrated the possible use of the pharmacophore of DHFR inhibitors as a guide for co-crystal screening, recognizing some similar trends in the outcome of association in the solid state and in the molecular aggregation in the co-crystals, characterized by the same supramolecular synthons.     

Recent Advances in In Silico Target Fishing

In silico target fishing, whose aim is to identify possible protein targets for a query molecule, is an emerging approach used in drug discovery due its wide variety of applications. This strategy allows the clarification of mechanism of action and biological activities of compounds whose target is still unknown. Moreover, target fishing can be employed for the identification of off targets of drug candidates, thus recognizing and preventing their possible adverse effects. For these reasons, target fishing has increasingly become a key approach for polypharmacology, drug repurposing, and the identification of new drug targets. While experimental target fishing can be lengthy and difficult to implement, due to the plethora of interactions that may occur for a single small-molecule with different protein targets, an in silico approach can be quicker, less expensive, more efficient for specific protein structures, and thus easier to employ. Moreover, the possibility to use it in combination with docking and virtual screening studies, as well as the increasing number of web-based tools that have been recently developed, make target fishing a more appealing method for drug discovery. It is especially worth underlining the increasing implementation of machine learning in this field, both as a main target fishing approach and as a further development of already applied strategies. This review reports on the main in silico target fishing strategies, belonging to both ligand-based and receptor-based approaches, developed and applied in the last years, with a particular attention to the different web tools freely accessible by the scientific community for performing target fishing studies.     

Suspect and nontarget screening approaches to identify organic contaminant records in lake sediments

Sediment cores provide a valuable record of historical contamination, but so far, new analytical techniques such as high-resolution mass spectrometry (HRMS) have not yet been applied to extend target screening to the detection of unknown contaminants for this complex matrix. Here, a combination of target, suspect, and nontarget screening using liquid chromatography (LC)-HRMS/MS was performed on extracts from sediment cores obtained from Lake Greifensee and Lake Lugano located in the north and south of Switzerland, respectively. A suspect list was compiled from consumption data and refined using the expected method coverage and a combination of automated and manual filters on the resulting measured data. Nontarget identification efforts were focused on masses with Cl and Br isotope information available that exhibited mass defects outside the sample matrix, to reduce the effect of analytical interferences. In silico methods combining the software MOLGEN-MS/MS and MetFrag were used for direct elucidation, with additional consideration of retention time/partitioning information and the number of references for a given substance. The combination of all available information resulted in the successful identification of three suspect (chlorophene, flufenamic acid, lufenuron) and two nontarget compounds (hexachlorophene, flucofuron), confirmed with reference standards, as well as the tentative identification of two chlorophene congeners (dichlorophene, bromochlorophene) that exhibited similar time trends through the sediment cores. This study demonstrates that complementary application of target, suspect, and nontarget screening can deliver valuable information despite the matrix complexity and provide records of historical contamination in two Swiss lakes with previously unreported compounds.     

An efficient protein preparation for proteomic analysis of developing cotton fibers by 2-DE

Preparation of high-quality proteins from cotton fiber tissues is difficult due to high endogenous levels of polysaccharides, polyphenols, and other interfering compounds. To establish a routine procedure for the application of proteomic analysis to cotton fiber tissues, a new protocol for protein extraction was developed by optimizing a phenol extraction method combined with methanol/ammonium acetate precipitation. The protein extraction for 2-DE was remarkably improved by the combination of chemically and physically modified processes including polyvinylpolypyrrolidone (PVPP) addition, acetone cleaning, and SDS replacement. The protocol gave a higher protein yield and vastly greater resolution and spot intensity. The efficiency of this protocol and its feasibility in fiber proteomic study were demonstrated by comparison of the cotton fiber proteomes at two growth stages. Furthermore, ten protein spots changed significantly were identified by MS/tandem MS and their potential relationships to fiber development were discussed. To the best of our knowledge, this is the first time that a protocol for protein extraction from cotton fiber tissues appears to give satisfactory and reproductive 2-D protein profiles. The protocol is expected to accelerate the process of the proteomic study of cotton fibers and also to be applicable to other recalcitrant plant tissues.     

Protein structure prediction and biomolecular recognition: From protein sequence to peptidomimetic design with the human β 3 integrin

Computational tools can bridge the gap between sequence and protein 3D structure based on the notion that information is to be retrieved from the databases and that knowledge-based methods can help in approaching a solution of the protein-folding problem. To this aim our group has implemented neural network-based predictors capable of performing with some success in different tasks, including predictions of the secondary structure of globular and membrane proteins, the topology of membrane proteins and porins and stable f -helical segments suited for protein design. Moreover we have developed methods for predicting contact maps in proteins and the probability of finding a cysteine in a disulfide bridge, tools which can contribute to the goal of predicting the 3D structure starting from the sequence (the so called ab initio prediction). All our predictors take advantage of evolution information derived from the structural alignments of homologous (evolutionary related) proteins and taken from the sequence and structure databases. When it is necessary to build models for proteins of unknown spatial structure, which have very little homology with other proteins of known structure, non-standard techniques need to be developed and the tools for protein structure predictions may help in protein modeling. The results of a recent simulation performed in our lab highlights the role of high performing computing technology and of tools of computational biology in protein modeling and peptidomimetic design.     

TMDIM: an improved algorithm for the structure prediction of transmembrane domains of bitopic dimers

\(\alpha\)-Helical transmembrane proteins are the most important drug targets in rational drug development. However, solving the experimental structures of these proteins remains difficult, therefore computational methods to accurately and efficiently predict the structures are in great demand. We present an improved structure prediction method TMDIM based on Park et al. (Proteins 57:577–585, 2004) for predicting bitopic transmembrane protein dimers. Three major algorithmic improvements are introduction of the packing type classification, the multiple-condition decoy filtering, and the cluster-based candidate selection. In a test of predicting nine known bitopic dimers, approximately 78% of our predictions achieved a successful fit (RMSD <2.0 Å) and 78% of the cases are better predicted than the two other methods compared. Our method provides an alternative for modeling TM bitopic dimers of unknown structures for further computational studies. TMDIM is freely available on the web at https://cbbio.cis.umac.mo/TMDIM. Website is implemented in PHP, MySQL and Apache, with all major browsers supported.     

A systematic evaluation of chip-based nanoelectrospray parameters for rapid identification of proteins from a complex mixture

HPLC-MS/MS is widely used for protein identification from gel spots and shotgun fractions. Although HPLC has well recognized benefits, this type of sample infusion also has some undesirable attributes: relatively low sample throughput, potential sample-to-sample carryover, time-varying sample composition, and no option for longer sample infusion for longer MS analyses. An automated chip-based ESI device (CB-ESI) has the potential to overcome these limitations. This report describes a systematic evaluation of the information-dependant acquisition (IDA) and sample preparation protocols for rapid protein identification from a complex mixture using a CB-ESI source compared with HPLC-ESI (gradient and isocratic elutions). Cytochrome c and a six-protein mixture (11-117 kDa) were used to develop an IDA protocol for rapid protein identification and to evaluate the effects of sample preparation protocols. MS (1-10 s) and MS/MS (1-60 s) scan times, sample concentration (50-500 fmol/microL), and ZipTipC(18) cleanup were evaluated. Based on MOWSE scores, protein coverage, experimental run time, number of identified proteins, and reproducibility, a 12.5 min experiment (22 cycles, each with one 3 s MS and eight 10 s MS/MS scans) was determined to be the optimal IDA protocol for CB-ESI. This work flow yielded up to 220% greater peptide coverage compared with gradient HPLC-ESI and provided protein identifications with up to a 2-fold higher throughput rate than either HPLC-ESI approach, whilst employing half the amount of sample over the same time frame. The results from this study support the use of CB-ESI as a rapid alternative to the identification of protein mixtures.     

Direct screening of a dynamic combinatorial library using mass spectrometry

A dynamic combinatorial library (DCL) screening approach is described that permits direct identification of the effective (from ineffective) combination of building blocks in the equilibrating DCL. The approach uses Fourier transform ion cyclotron resonance mass spectrometry (FTICR MS) together with sustained off-resonance irradiation collision activated dissociation (SORI-CAD) to detect noncovalent protein-DCL ligand complexes under native conditions. It was shown that in a single, rapid experiment one could concurrently identify all the ligands of interest from the DCL against a background of inactive DCL ligands while still in the presence of the target protein. This result has demonstrated that mass spectrometry may provide a fast preliminary screening approach to identify DCL candidates for later verification with more traditional but time-consuming analysis. The MS/MS enables DCL mixtures to be effectively deconvoluted without the need for either chromatography, synthesis of DCL sub-libraries, conversion of the DCL to a static library, or disruption of the protein-ligand complexes before analysis--all typically necessary for the current screening method for DCLs.