Targeted covalent compounds or drugs have good potency as they can bind to a specific target for a long time with low doses. Most currently known covalent ligands were discovered by chance or by modifying existing noncovalent compounds to make them covalently attached to a nearby reactive residue. Computational methods for novel covalent ligand binding prediction are highly demanded. We performed statistical analysis on protein complexes with covalent ligands attached to cysteine residues. We found that covalent modified cysteine residues have unique features compared to those not attached to covalent ligands, including lower pKa, higher exposure, and higher ligand binding affinity. SVM models were built to predict cysteine residues suitable for covalent ligand design with prediction accuracy of 0.73. Given a protein structure, our method can be used to automatically detect druggable cysteine residues for covalent ligand design, which is especially useful for identifying novel binding sites for covalent allosteric ligand design.

Genetically encoded covalent peptide tagging technology, such as the SpyTag–SpyCatcher reaction, has emerged as a unique way to do chemistry with proteins. Herein, we report the reactivity engineering of SpyTag–SpyCatcher mutant pairs and show that distinct reactivity can be encrypted for the same reaction based on protein sequences of high similarity. Valuable features, including high selectivity, inverse temperature dependence and (nearly) orthogonal reactivity, could be achieved based on as few as three mutations. This demonstrates the robustness of the SpyTag–SpyCatcher reaction and the plasticity of its sequence specificity, pointing to a family of engineered protein chemistry tools.

•The recent progress in ligand-binding protein design is reviewed.•Combining computational design with experimental screen gives successful results.•Success rate of computational design alone is still limited.•Designing ligands for specific function is more challenging than binding design.•Methods for drug target identification can be used in ligand-binding protein design.Custom-designed ligand-binding proteins with novel functions hold the potential for numerous applications. In recent years, the developments of computational methods together with high-throughput experimental screening techniques have led to the generation of novel, high-affinity ligand-binding proteins for given ligands. In addition, naturally occurring ligand-binding proteins have been computationally designed to recognize new ligands while keeping their original biological functions at the same time. Furthermore, metalloproteins have been successfully designed for novel functions and applications. Though much has been learned in these successful design cases, advances in our understanding of protein dynamics and functions related to ligand binding and development of novel computational strategies are necessary to further increase the success rate of computational protein–ligand binding design.

•Receptor-TNF and mutant binding kinetics changes lead to functional selectivity.•R1antTNF is a fast associating and fast dissociating TNF mutant.•R1antTNF produces shortened nuclear duration of NF-κB.•R1antTNF induces biased gene expression of early-response genes.•R1antTNF selectively activates apoptosis at low concentration.Tumor necrosis factor (TNF) is a pluripotent inflammatory cytokine that can induce both the pro-survival nuclear factor kappa B (NF-κB) pathway and the pro-apoptotic caspase pathway. Selectively activating only one of the two pathways remains challenging. We used TNF mutants with different receptor binding kinetics to study their effects on NF-κB signaling dynamics and cell apoptosis. A TNF mutant, R1antTNF, which binds to TNFR1 with increased association and dissociation rates, induced NF-κB signaling with shorter response time and first peak duration. The short nuclear stay of NF-κB led to biased activation of downstream genes, favoring the fast response ones. At the same time, R1antTNF retains pro-apoptotic activity. At 10 ng/ml, R1antTNF selectively activated the pro-apoptotic pathway rather than the pro-survival NF-κB pathway. Our study provides a new example for the emerging evidence that ligand-receptor binding kinetics play a key role in the selective activation of downstream pathways, which deserves more attention in future drug discovery and disease studies.

Polo-like kinase 1(Plk1) is vital for cell mitosis and has been identified as anticancer target. Its polo-box domain (PBD) mediates substrate binding, blocking of which may offer selective Plk1 inhibition compared to kinase domain inhibitors. Although several PBD inhibitors were reported, most of them suffer from low cell activity. Here, we report the discovery of novel inhibitors to induce mitotic arrest in HeLa cells by virtual screening with Plk1 PBD and cellular activity testing. Of the 81 compounds tested in the cell assay, 10 molecules with diverse chemical scaffolds are potent to induce mitotic arrest of HeLa at low micromolar concentrations. The best compound induces mitotic arrest of HeLa cells with an EC50 of 4.4 μM. The cellular active inhibitors showed binding to Plk1 PBD and compete with PBD substrate in microscale thermophoresis analysis.

For disease network intervention, up-regulating enzyme activities is equally as important as down-regulating activities. However, the design of enzyme activators presents a challenging route for drug discovery. Previous studies have suggested that activating 15-lipoxygenase (15-LOX) is a promising strategy to intervene the arachidonic acid (AA) metabolite network and control inflammation. To prove this concept, we used a computational approach to discover a previously unknown allosteric site on 15-LOX. Both allosteric inhibitors and novel activators were discovered using this site. The influence of activating 15-LOX on the AA metabolite network was then investigated experimentally. The activator was found to increase levels of 15-LOX products and reduce production of pro-inflammatory mediators in human whole blood assays. These results demonstrate the promising therapeutic value of enzyme activators and aid in further development of activators of other proteins.

Allostery is the phenomenon in which a ligand binding at one site affects other sites in the same macromolecule. Allostery has important roles in many biological processes. Theoretically, all nonfibrous proteins are potentially allosteric. However, few allosteric proteins have been validated, and the identification of novel allosteric sites remains a challenge. The motion of residues and subunits underlies protein function; therefore, we hypothesized that the motions of allosteric and orthosteric sites are correlated. We utilized a data set of 24 known allosteric sites from 23 monomer proteins to calculate the correlations between potential ligand-binding sites and corresponding orthosteric sites using a Gaussian network model (GNM). Most of the known allosteric site motions showed high correlations with corresponding orthosteric site motions, whereas other surface cavities did not. These high correlations were robust when using different structural data for the same protein, such as structures for the apo state and the orthosteric effector-binding state, whereas the contributions of different frequency modes to motion correlations depend on the given protein. The high correlations between allosteric and orthosteric site motions were also observed in oligomeric allosteric proteins. We applied motion correlation analysis to predict potential allosteric sites in the 23 monomer proteins, and some of these predictions were in good agreement with published experimental data. We also performed motion correlation analysis to identify a novel allosteric site in 15-lipoxygenase (an enzyme in the arachidonic acid metabolic network) using recently reported activating compounds. Our analysis correctly identified this novel allosteric site along with two other sites that are currently under experimental investigation. Our study demonstrates that the motions of allosteric sites are highly correlated with the motions of orthosteric sites. Our correlation analysis method provides new tools for predicting potential allosteric sites.

•A series of acylthioureas were designed and synthesized as Plk1 inhibitors.•Compounds with halogen in sulfamoylphenyl group showed better binding affinities.•The best compound 3v binds to Plk1 PBD with a Kd of 2.3 ± 0.1 μM.•Compound 3v also inhibits the kinase activity of full-length Plk1.Thiourea derivatives have drawn much attention for their latent capacities of biological activities. In this study, we designed acylthiourea compounds as polo-like kinase 1 (Plk1) polo-box domain (PBD) inhibitors. A series of acylthiourea derivatives without pan assay interference structure (PAINS) were synthesized. Four compounds with halogen substituents exhibited binding affinities to Plk1 PBD in low micromole range. The most potent compound (3v) showed selectivity over other subtypes of Plk PBDs and inhibited the kinase activity of full-length Plk1.

Persisters are a small fraction of drug-tolerant bacteria without any genotype variations. Their existence in many life-threatening infectious diseases presents a major challenge to antibiotic therapy. Persistence is highly related to toxin–antitoxin modules. HipA (high persistence A) was the first toxin found to contribute to Escherichia coli persistence. In this study, we used structure-based virtual screening for HipA inhibitors discovery and identified several novel inhibitors of HipA that remarkably reduced E. coli persistence. The most potent one decreased the persister fraction by more than five-fold with an in vitro KD of 270 ± 90 nM and an ex vivo EC50 of 46 ± 2 and 28 ± 1 μM for ampicillin and kanamycin screening, respectively. These findings demonstrated that inhibition of toxin can reduce bacterial persistence independent of the antibiotics used and provided a framework for persistence treatment by interfering with the toxin–antitoxin modules.Keywords: drug discovery; HipA (high persistence A); Persistence; toxin-antitoxin (TA) module

A computational strategy was used to design helical peptides that can bind to tumor necrosis factor-α (TNFα) dimers to prevent active trimer formation. Three designed peptides showed TNFα inhibition at the cellular level. Chemical crosslinking and mass spectrometry studies verified that these peptides function by breaking TNFα trimers.

Inflammation and other common disorders including diabetes, cardiovascular disease, and cancer are often the result of several molecular abnormalities and are not likely to be resolved by a traditional single-target drug discovery approach. Though inflammation is a normal bodily reaction, uncontrolled and misdirected inflammation can cause inflammatory diseases such as rheumatoid arthritis and asthma. Nonsteroidal anti-inflammatory drugs including aspirin, ibuprofen, naproxen, or celecoxib are commonly used to relieve aches and pains, but often these drugs have undesirable and sometimes even fatal side effects. To facilitate safer and more effective anti-inflammatory drug discovery, a balanced treatment strategy should be developed at the biological network level.In this Account, we focus on our recent progress in modeling the inflammation-related arachidonic acid (AA) metabolic network and subsequent multiple drug design. We first constructed a mathematical model of inflammation based on experimental data and then applied the model to simulate the effects of commonly used anti-inflammatory drugs. Our results indicated that the model correctly reproduced the established bleeding and cardiovascular side effects. Multitarget optimal intervention (MTOI), a Monte Carlo simulated annealing based computational scheme, was then developed to identify key targets and optimal solutions for controlling inflammation. A number of optimal multitarget strategies were discovered that were both effective and safe and had minimal associated side effects. Experimental studies were performed to evaluate these multitarget control solutions further using different combinations of inhibitors to perturb the network. Consequently, simultaneous control of cyclooxygenase-1 and -2 and leukotriene A4 hydrolase, as well as 5-lipoxygenase and prostaglandin E2 synthase were found to be among the best solutions.A single compound that can bind multiple targets presents advantages including low risk of drug–drug interactions and robustness regarding concentration fluctuations. Thus, we developed strategies for multiple-target drug design and successfully discovered several series of multiple-target inhibitors. Optimal solutions for a disease network often involve mild but simultaneous interventions of multiple targets, which is in accord with the philosophy of traditional Chinese medicine (TCM). To this end, our AA network model can aptly explain TCM anti-inflammatory herbs and formulas at the molecular level. We also aimed to identify activators for several enzymes that appeared to have increased activity based on MTOI outcomes. Strategies were then developed to predict potential allosteric sites and to discover enzyme activators based on our hypothesis that combined treatment with the projected activators and inhibitors could balance different AA network pathways, control inflammation, and reduce associated adverse effects.Our work demonstrates that the integration of network modeling and drug discovery can provide novel solutions for disease control, which also calls for new developments in drug design concepts and methodologies. With the rapid accumulation of quantitative data and knowledge of the molecular networks of disease, we can expect an increase in the development and use of quantitative disease models to facilitate efficient and safe drug discovery.

Drug-induced liver injury (DILI) has been the single most frequent cause of safety-related drug marketing withdrawals for the past 50 years. Recently, deep learning (DL) has been successfully applied in many fields due to its exceptional and automatic learning ability. In this study, DILI prediction models were developed using DL architectures, and the best model trained on 475 drugs predicted an external validation set of 198 drugs with an accuracy of 86.9%, sensitivity of 82.5%, specificity of 92.9%, and area under the curve of 0.955, which is better than the performance of previously described DILI prediction models. Furthermore, with deep analysis, we also identified important molecular features that are related to DILI. Such DL models could improve the prediction of DILI risk in humans. The DL DILI prediction models are freely available at http://www.repharma.cn/DILIserver/DILI_home.php.

Targeting mitotic regulation is recognized as an important strategy for cancer therapy. Aurora A/B kinase and polo-like kinase 1 (PLK1) are the key mitotic regulators, and many inhibitors have been developed. Combinations of these inhibitors are anticipated to be more effective therapeutics compared with single-inhibitor treatments; however, a systematic analysis of the combined effects is lacking. Here, we constructed the first mammalian cell mitotic regulation network model, which spans from mitotic entry to anaphase initiation, and contains all key mitotic kinase targets. The combined effects of different kinase inhibitors and microtubule inhibitors were systematically explored. Simultaneous inhibition of Aurora B and PLK1 strongly induces polyploidy. Microtubule inhibitor dosage can be significantly reduced when combined with a PLK1 inhibitor. The efficacy of these inhibitor combinations was validated by our experimental results. The mitotic regulatory network model provides a platform to study the complex interactions during mitosis, enables identification of mitotic regulators, and determines targets for drug discovery research. The suggested use of combining microtubule inhibitors with PLK1 inhibitors is anticipated to enhance microtubule-inhibitor tolerance in a wide range of patients.

Bacteria use chemotaxis signaling pathways to sense environmental changes. Escherichia coli chemotaxis system represents an ideal model that illustrates fundamental principles of biological signaling processes. Chemoreceptors are crucial signaling proteins that mediate taxis toward a wide range of chemoeffectors. Recently, in deep study of the biochemical and structural features of chemoreceptors, the organization of higher-order clusters in native cells, and the signal transduction mechanisms related to the on–off signal output provides us with general insights to understand how chemotaxis performs high sensitivity, precise adaptation, signal amplification, and wide dynamic range. Along with the increasing knowledge, bacterial chemoreceptors can be engineered to sense novel chemoeffectors, which has extensive applications in therapeutics and industry. Here we mainly review recent advances in the E. coli chemotaxis system involving structure and organization of chemoreceptors, discovery, design, and characterization of chemoeffectors, and signal recognition and transduction mechanisms. Possible strategies for changing the specificity of bacterial chemoreceptors to sense novel chemoeffectors are also discussed.

In this Perspective, we focus on new, systems-centric views of structure-based drug design (SBDD) that we believe will impact future drug discovery research and development. We will first discuss new ways to identify drug targets based on systems intervention analysis, and then we will introduce emerging SBDD methods driven by advancements in systems biology.

•Virtual screen targeting TNFα dimer was performed.•Both SPR assay and cell-based study were used to confirm active compounds.•Compound 11 was among the most potent TNFα small molecule inhibitors.•All active compounds represent novel scaffolds as TNFα inhibitors.Tumor necrosis factor-α (TNFα) is a validated therapeutic target for various autoimmune disorders such as rheumatoid arthritis and asthma. All TNFα inhibitors currently on the market are biologics, making the development of small molecule alternatives in urgent need. However, only a few successful cases of direct TNFα antagonization in vitro have been reported. Here, we present the identification of several small molecule candidates able to effectively reduce TNFα activity in vitro and in cell assays. Virtual screen targeting TNFα dimer was performed on the SPECS database and 101 compounds were selected for experimental testing. Two compounds, 1 and 2, displayed considerable inhibitory activity. Follow-up structure–activity relationship analysis of compound 1 identified 3 molecules with low micromolar cell-level inhibitory activity. Compound 11 showed an IC50 value of 14 μM, making it among the most potent TNFα small molecule inhibitors reported. These compounds provide new scaffolds for future development of small molecule drugs against TNFα.We used virtual screen and discovered potent TNFα small molecule inhibitors with novel scaffolds. The best compound, 11, is among the strongest direct TNFα inhibitors reported to date.

The discovery of multitarget drugs has recently attracted much attention. Most of the reported multitarget ligands have been serendipitous discoveries. Although a few methods have been developed for rational multitarget drug discovery, there is a lack of elegant methods for de novo multitarget drug design and optimization, especially for multiple targets with large differences in their binding sites. In this paper, we report the first de novo multitarget ligand design method, with an iterative fragment-growing strategy. Using this method, dual-target inhibitors for COX-2 and LTA4H were designed, with the most potent one inhibiting PGE2 and LTB4 production in the human whole blood assay with IC50 values of 7.0 and 7.1 μM, respectively. Our strategy is generally applicable in rational and efficient multitarget drug design, especially for the design of highly integrated inhibitors for proteins with dissimilar binding pockets.

Ordinary differential equations (ODEs) are widely used to model the dynamic properties of biological networks. Due to the complexity of biological networks and limited quantitative experimental data available, estimating kinetic parameters for these models remains challenging. We present a novel global optimization algorithm, differential simulated annealing (DSA), for estimating kinetic parameters for biological network models robustly and efficiently. DSA was tested on 95 models sizing from a few to several hundreds of parameters from the BioModels database and compared with other five widely used algorithms for parameter estimation, including both deterministic and stochastic optimization algorithms. Our study showed that DSA gave the highest success rate in the whole dataset and performed especially well for large models. Further analysis revealed that DSA outperformed the five algorithms compared in both accuracy and efficiency.

Microsomal prostaglandin E2 synthase 1 (mPGES-1) has been identified as a promising drug target due to its key role in prostaglandin biosynthesis. However, the lack of a well-characterized structure constitutes a great challenge for the development of inhibitors. Recently, we have built a model for the active conformation of mPGES-1. In the present study, the model was used for structure-based virtual screen of novel mPGES-1 inhibitors. Of the 142 compounds tested in the cell-free assay, 10 molecules are highly potent with IC50 values of single digit nanomolar and the strongest inhibition of 1.1 nM. Moreover, nine compounds showed strong activity in the human whole blood (HWB) assay with IC50 values of less than 10 μM. The lead compounds 1 and 2 showed HWB IC50 values of 0.3 and 0.7 μM which are among the most potent mPGES-1 inhibitors reported. These compounds represent new scaffolds for future development of drugs against mPGES-1.

Through history, traditional Chinese medicine (TCM) has adopted oriental philosophical practices of drug combination and interaction to address human diseases. To investigate this from a systems biology point of view, we analysed 28 TCM herbs for their anti-inflammatory function, using molecular docking and arachidonic acid (AA) metabolic network simulation. The inhibition potential of each herb toward five essential enzymes as well as their possible side effects were examined. Three commonly prescribed anti-inflammatory formulae were simulated to discover the combinatorial properties of each contained herb in regulating the whole metabolic network. We discovered that different ingredients of a formula tend to inhibit different targets, which almost covered all the targets in the whole network. We also found that herbal combinations could achieve the same therapeutic effect at lower doses compared with individual usage. New herbal combinations were also predicted based on the inhibition potentials and two types of synergistic drug combinations of TCM theory were discussed from the perspective of systems biology. Using this combined approach of molecular docking and network simulation, we were able to computationally elucidate the combinatorial effects of TCM to intervene disease networks. We expect novel TCM formulae or modern drug combinations to be developed based on this research.

Binding kinetics is closely related to the efficacy of drugs. Several aspects of binding kinetics, such as long residence or frequent dissociation, have been proposed to affect drug properties such as efficacy, selectivity, and multi-target potency. However, a comprehensive and balanced study of binding kinetics in various scenarios is still needed. We performed a comprehensive computational analysis of the role of drug binding kinetics in various situations such as enzyme inhibition, receptor binding, multi-target drug targeting, signal transduction pathways, and metabolic networks. Molecular studies of enzyme inhibition, receptor binding, and multi-target drugs have shown that at constant binding affinity, fast associating drugs show better enzyme inhibitory effects, earlier and higher receptor occupancy peaks, and better multi-target performances, while slow dissociating drugs show prolonged receptor occupancy, as suggested by others. Different situations exemplify slightly different kinetic–efficacy relationships, and each must be considered separately. At the systems level, binding kinetics can not only change the overall effect of drugs, but can also affect signaling dynamics. For example, in the tumor necrosis factor α-induced nuclear factor-κB pathway, inhibitor addition can delay the onset of oscillations and decrease their frequencies, with these changes varying with the binding kinetics of the inhibitor. The effects of drug binding kinetics also depend on network topology and where the target is located in the network. For successful drug discovery, both molecular binding kinetics and systems level requirements need to be considered.

Human non-pancreatic secretory phospholipase A2 (hnpsPLA2) catalyzes the sn-2 acyl hydrolysis of phospholipids. It was reported that hnpsPLA2 is involved in various diseases like inflammation, cancer, and so on. This enzyme also exhibits anti-bacterial and anti-virus activities. It is active over a broad pH range, with higher activity at alkaline conditions. In order to make it suitable as a potential bactericide, high activity at neutral pH is preferable. We have tried to tailor the pH dependence of hnpsPLA2 activity by replacing its surface charged residues. Three surface charge replacements, Arg42Glu, Arg100Glu, and Glu89Lys, showed increased activities at neutral pH, which are 2.3, 2.8, and 2.3 times that of the wild-type enzyme at pH 7. Both the positive-to-negative and negative-to-positive mutations lowered the optimum enzymatic reaction pH of hnpsPLA2, indicating that the enzyme pH profile depends on a delicate balance of charged residues. The activity changes are in good agreement with the recently proposed calcium-coordinated catalytic triad mechanism. This study also provides a general means of enhancing hnpsPLA2 activity at low pH.

Bacterial chemoreceptors mediate chemotactic responses to diverse stimuli. Here, by using an integrated in silico, in vitro, and in vivo approach, we screened a large compound library and found eight novel chemoeffectors for the Escherichia coli chemoreceptor Tar. Six of the eight new Tar binding compounds induce attractant responses, and two of them function as antagonists that can bind Tar without inducing downstream signaling. Comparison between the antagonist and attractant binding patterns suggests that the key interactions for chemotaxis signaling are mediated by the hydrogen bonds formed between a donor group in the attractant and the main-chain carbonyls (Y149 and/or Q152) on the α4 helix of Tar. This molecular insight for signaling is verified by converting an antagonist to an attractant when introducing an N-H group into the antagonist to restore the hydrogen bond. Similar signal triggering effect by an O-H group is also confirmed. Our study suggests that the Tar chemoeffector binding pocket may be separated into two functional regions: region I mainly contributes to binding and region II contributes to both binding and signaling. This scenario of binding and signaling suggests that Tar may be rationally designed to respond to a nonnative ligand by altering key residues in region I to strengthen binding with the novel ligand while maintaining the key interactions in region II for signaling. Following this strategy, we have successfully redesigned Tar to respond to l-arginine, a basic amino acid that does not have chemotactic effect for WT Tar, by two site-specific mutations (R69′E and R73′E).

Human 5-lipoxygenase (5-LOX) is one of the key anti-inflammatory drug targets due to its key role in leukotrienes biosynthesis. We have built a model for the active conformation of human 5-LOX using comparative modeling, docking of known inhibitors, and molecular dynamics simulation. Using this model, novel 5-LOX inhibitors were identified by virtual screen. Of the 105 compounds tested in a cell-free assay, 30 have IC50 values less than 100 μM and 11 less than 10 μM with the strongest inhibition of 620 nM. Compounds 4, 7, and 11 showed strong inhibition activity in the human whole blood (HWB) assay with IC50 values of 8.6, 9.7, 8.1 μM, respectively. Moreover, compounds 4 and 7 were also found to inhibit microsomal prostaglandin E synthase (mPGES)-1 with micromolar IC50 values, similar to licofelone, a dual functional inhibitor of 5-LOX/mPGES-1. The compounds reported here provide new scaffolds for anti-inflammatory drug design.

Allostery is a common mechanism of controlling many biological processes such as enzyme catalysis, signal transduction, and metabolic regulation. The use of allostery to regulate protein activity is an important and promising strategy in drug discovery and biological network regulation. In order to modulate protein activity by allostery, predictive methods need to be developed to discover allosteric binding sites. In the present study, we developed a new approach to identify allosteric sites in proteins based on the coarse-grained two-state Go̅ model. Starting from the concept that allostery is a conformation population shift process, we first constructed an ensemble of two functional states of a protein and tuned the energy landscape to bias one state. We then added perturbations to a binding site and monitored the population distribution of the new ensemble. If population redistribution occurred, then the binding perturbed site was predicted as a potential allosteric site. Our approach successfully identified all the known allosteric sites in a set of test proteins. Several new allosteric sites in the test proteins were also predicted. By use of one of the new allosteric sites predicted from Escherichia coli phosphoglycerate dehydrogenase (PGDH), novel allosteric regulating molecules were screened by molecular docking and enzymatic assay. Three novel allosteric inhibitors were discovered and their binding modes were confirmed by mutation experiments and competitive assay. The IC50 of the strongest inhibitor discovered was 21 μM, which is comparable to that of the native allosteric inhibitor l-serine. The novel allosteric site discovered in PGDH is l-serine-independent, and inhibitors targeting this site can be used as novel regulators of the E. coli serine synthesis pathway. Our approach for allosteric site prediction is generally applicable and the predicted sites can be used in discovering novel allosteric regulating molecules.

The arachidonic acid (AA) metabolic network produces key inflammatory mediators which have been considered as hallmark contributors in various inflammatory related diseases. Enzymes in this network, such as 5-lipoxygenase (5-LOX), cyclooxygenase (COX), leukotriene A4 hydrolase (LTA4H) and prostaglandin E synthase (PGES), have been used as targets for anti-inflammatory drug discovery. Multi-target drugs and drug combinations have also been developed for this network. However, how the inhibitors alter the dynamics of metabolite production and which combinatorial target intervention solutions are better needs further exploration. We did a system based intervention analysis on the AA metabolic network. Using an LC-MS/MS method, we quantitatively studied the eicosanoid metabolites responses of AA metabolic network during stimulation of Sprague Dawley rat blood samples with the calcium ionophore. Our results indicate that inhibiting the upstream rather than the downstream target of 5-LOX pathway will simultaneously alter the AA metabolism to the COX pathway (and vice versa). Therefore, single-target inhibitors cannot control all the inflammatory mediators at the same time. We also suggest that in the case of multiple-target anti-inflammatory solutions, the combination of inhibitors of the downstream enzymes may have stronger inhibition efficiency and cause less side-effects compared to the other solutions. One therapeutic strategy, LTA4H/COX inhibition solution, was found promising for the intervention of inflammatory mediator biosynthesis and at the same time stimulating the production of anti-inflammatory agents.

Dual target inhibitors against COX-2 and LTA4H were designed by adding functional groups from a marketed COX-2 inhibitor, Nimesulide, to an existing LTA4H inhibitor 1-(2-(4-phenoxyphenoxy) ethyl) pyrrolidine. A series of phenoxyphenyl pyrrolidine compounds were synthesized and tested for their inhibition activities using enzyme assays and human whole blood assay. Introduction of small electron withdrawing groups like NO2 and CF3 in the ortho-position of the terminal phenyl ring was found to change the original single target LTA4H inhibitor to dual target LTA4H and COX-2 inhibitors. Compound 5a and 5m showed dual LTA4H and COX-2 inhibition activities in the enzyme assays and the HWB assay with IC50 values in the micromolar to submicromolar range. As their activities in HWB assay were comparable to the two starting single target inhibitors, the two compounds are promising for further studies. The strategy used in the current study may be generally applicable to other dual target drug designs.

We have developed a new version (2.0) of the de novo drug design program LigBuilder. With LigBuilder 2.0, the synthesis accessibility of designed compounds can be analyzed, and a cavity detection procedure is implemented to detect the positions and shapes of the binding sites on the surface of a given protein structure and to quantitatively estimate drugability. Ligands are designed to best fit the detected cavities using a set of rules for evaluation. Drug-like and privileged fragments are used to construct the ligands with the aid of internal and external absorption, distribution, metabolism, excretion, and toxicity (ADME/T) and drug-like filters.

Microsomal prostaglandin E synthase-1 (mPGES-1) is a newly recognized therapeutic target for the treatment of inflammation, pain, cancer, atherosclerosis, and stroke. Many mPGES-1 inhibitors have been discovered. However, as the structure of the binding site is not well-characterized, none of these inhibitors was designed based on the mPGES-1 structure, and their inhibition mechanism remains to be fully disclosed. Recently, we built a new structural model of mPGES-1 which was well supported by experimental data. Based on this model, molecular docking and competition experiments were used to investigate the binding modes of four representive mPGES-1 inhibitors. As the inhibitor binding sites predicted by docking overlapped with both the substrate and the cofactor binding sites, mPGES-1 inhibitors might act as dual-site inhibitors. This inhibitory mechanism was further verified by inhibitor-cofactor and inhibitor-substrate competition experiments. To investigate the potency-binding site relationships of mPGES-1 inhibitors, we also carried out molecular docking studies for another series of compounds. The docking results correlated well with the different inhibitory effects observed experimentally. Our data revealed that mPGES-1 inhibitors could bind to the substrate and the cofactor binding sites simultaneously, and this dual-site binding mode improved their potency. Future rational design and optimization of mPGES-1 inhibitors can be carried out based on this binding mechanism.

A series of novel fused heterocycle methyl esters were designed and synthesized as human nonpancreatic secretory phospholipase A2 (hnps-PLA2) competitive inhibitors. Among the 22 synthesized compounds, 17 quinoline-4-methyl esters displayed hnps-PLA2 inhibition activity in the in vitro bioassay. The IC50 value for the best compound 3o was 1.5 μM. The structure-inhibition–activity relationships of the compounds were studied using molecular docking.

Functional proteins designed de novo have potential application in chemical engineering, agriculture and healthcare. Metal binding sites are commonly used to incorporate functions. Based on a de novo designed protein DS119 with a βαβ structure, we have computationally engineered zinc binding sites into it using a home-made searching program. Seven out of the eight designed sequences tested were shown to bind Zn2+ with micromolar affinity, and one of them bound Zn2+ with 1:1 stoichiometry. This is the first time that metalloproteins with an α, β mixed structure have been designed from scratch.

Although virtual screening through molecular docking has been widely applied in lead discovery, it is still challenging to distinguish true hits from high-scoring decoys because of the difficulty in accurately predicting protein−ligand binding affinities. Following the successful application of energy landscape analysis to both protein folding and biomolecular binding studies, we attempted to use protein−ligand binding energy landscape analysis to recognize true binders from high-scoring decoys. Two parameters describing the binding energy landscape were used for this purpose. The energy gap, defined as the difference between the binding energy of the native binding mode and the average binding energy of other binding modes in the “denatured binding phase”, was used to describe the thermodynamic stability of binding, and the number of local binding wells in the landscapes was used to account for the kinetic accessibility. These parameters, together with the docking score, were combined using logistic regression to investigate their capability to discriminate true ligands from high-scoring decoys. Inhibitors and the noninhibitors of two enzyme systems, neuraminidase and cyclooxygenase-2, were used to test their discrimination capability. Using a five-fold cross-validation, the areas under the receiver operator characteristic curves (AUCs) from the best linear combinations of parameters reached 0.878 for neuraminidase and 0.776 for cyclooxygenase-2. To make a more independent test, inhibitors and high-scoring decoys in a directory of useful decoys (DUD), the largest and most comprehensive public data set for benchmarking virtual screen programs by far, were used as independent test sets to test the discrimination capability of these parameters. The AUCs of the best linear combinations of parameters for the independent test sets were 0.750 for neuraminidase and 0.855 for cyclooxygenase-2. Furthermore, combining these two parameters with the docking scoring function improved the enrichment ratio to 200−300% compared to that using the scoring function alone. This study suggests that incorporating information from binding energy landscape analysis can significantly increase the success rate of virtual screening.

The viral infectivity factor (Vif) was found to be essential for controlling HIV-1 virus infectivity. It targets cellular antiviral proteins in APOBEC family (APO) to trigger its fast degradation and inhibits APO packaging into nascent virion. In the present study, we propose a mathematical model to quantitatively study the intracellular dynamics of these typical virus-host interactions. Four sets of published experimental data were compared with simulation results to justify the model. Systematic parameter sensitivity and perturbation analysis showed that parameters related to APO are crucial to the infectivity of newly synthesized HIV-1 virus. Interestingly, we discovered that the synthesis rate of the viral structure protein Gag and the required number per nascent virion are optimized to achieve high virion production with minimal level of packaged APO, and large portion of model parameters are beneficial to virus only within a relatively small range. Furthermore, minor variations in several parameters related to viral protein Tat, the activator of viral RNA synthesis, were found to induce switch-like behaviors on both incorporated Vif and APO. These findings may provide new insights for understanding the high mutation rate of HIV-1 virus and its latency, as well as help identify key targets for therapeutic design.

Small heat shock proteins (sHSPs) exist ubiquitously among all organisms, with a variety of functions. All small heat shock proteins assemble into a native large oligomeric state containing 9–40 monomers. The sHSPs show chaperone-like activity to prevent the aggregation of nonnative proteins under stressful cellular conditions such as non-optimal temperatures, pH changes, osmotic pressure, and exposure to toxic chemicals. It was found that a common dimeric subunit of sHSPs might be the major active species, but whether the native large oligomeric state is only a storage state or a state crucial to its molecular chaperone activity is still under debate. The native large oligomeric state of the small heat shock protein from a hyperthermophilic methanarchaeon, Methanococcus jannaschii (Mj HSP 16.5), is a stable icositetramer, which is a symmetric hollow sphere that is very stable even at 85°C, and no small active subunit has been detected till now. Our results show that Mj sHSP 16.5 changes into small and active oligomeric state at pH 3, likely as octamers (average result) at 25°C, and dimers at 65°C. The dimer of Mj HSP 16.5 at pH 3.0 and 65°C is very active and efficient, even 7-fold more efficient than the high-temperature-activated icositetramer at neutral pH. Monomer exchange can be observed between dimers of Mj HSP 16.5 at pH 3.0 and 65°C. These results not only demonstrate that the icositetramer structure of Mj sHSP16.5 is not necessary for its molecular chaperone activity, but also suggest that Mj sHSP16.5 is a very efficient chaperone acting at high temperature and under the acidic condition. Even though it is not clear whether the native environment of Methanococcus jannaschii is acidic or not, given its ability to excrete acidic compounds, it is likely that Methanococcus jannaschii will encounter acidic internal or external environments at high temperature. Our results demonstrate that Mj HSP 16.5 may help Methanococcus jannaschii to survive better under those extreme environmental conditions.

A series of novel bis-indole compounds, 1,ω-bis(((3-acetamino-5-methoxy-2-methylindole)-2-methylene)phenoxy)alkane, have been designed and synthesized on the basis of the enzyme structure of human nonpancreatic secretory phospholipase A2 (hnps PLA2). Their inhibition activities against hnps PLA2 were improved compared to that of the monofunctional protocompound. These bivalent ligands not only inhibited hnps PLA2 but also drove the dimerization of hnps PLA2. Their dimerization ability correlated with the linker length and position. Further study on the potent compound 5 (1,5-bis(((3-acetamino-5-methoxy-2-methylindole)-2-methylene)phenoxy)pentane, IC50 = 24 nM) revealed that cooperative binding interactions between the two enzyme molecules also contributed to the stability of the ternary complex. The combination of bivalent ligands and hnps PLA2 can be used as a novel chemically induced dimerization (CID) system for designing regulatory inhibitors.

Multitarget drugs have been to be found effective in controlling complex diseases. However, how to design multitarget drugs presents a great challenge. We have developed a computer-assisted strategy to screen for multitarget inhibitors using a combination of molecular docking and common pharmacophore matching. This strategy was successfully applied to screen for dual-target inhibitors against both the human leukotriene A4 hydrolase (LTA4H-h) and the human nonpancreatic secretory phospholipase A2 (hnps-PLA2). Three compounds screened from the chemical database MDL Available Chemical Directory were found to inhibit these two enzymes at the 10 μM level. Moreover, one synthetic compound matching the common pharmacophores inhibits LTA4H-h and hnps-PLA2 with IC50 values of 35 nM and 7.3 μM, respectively. The common pharmacophore model can also be used to search small molecule databases or collections of existing inhibitors, as well as to guide combinatorial library design to search for ideal multitarget inhibitors.

An unusual phenomenon, the specific interaction between tris(hydroxymethyl)aminomethane (Tris) and lysozyme (LZM), was demonstrated for the first time by rapid screen analysis of interactions using a quartz crystal microbalance (QCM) biosensor. This phenomenon was also observed in a surface plasmon resonance (SPR) system. Further study using high-performance affinity chromatography (HPAC) confirmed this specific interaction between LZM and immobilized Tris with an apparent dissociation constant (KD) of 6.7 × 10−5 M. Molecular docking was carried out to identify possible modes of binding between LZM and Tris linked to a binding arm. The estimated binding free energy was −6.34 kcal mol−1, corresponding to a KD of 2.3 × 10−5 M, which correlated well with the experimental value. Based on the docking model, the three hydroxyl groups of Tris form intermolecular H bonds with Asp52, Glu35, and Ala107 in LZM. This study reinforces the importance of buffer selection in quantitative biochemical investigations. For a lysozyme ligand binding study, it is better to avoid using Tris when the ligands under study are weak binders.

The synthesis and biological evaluation of a series of diphenyl ether derivatives were described. The compounds can either activate or inhibit the aminopeptidase activity of leukotriene A4 hydrolase, while at the same time do not influence the hydrolase activity. Further enzyme kinetics and molecular modeling investigation on these novel chemical activators revealed their possible activation mechanism. These compounds can be used as probes to regulate the aminopeptidase activity of leukotriene A4 hydrolase.Diphenyl ether and derivatives can either activate or inhibit the aminopeptidase activity of leukotriene A4 hydrolase, by binding at the hydrophobic pocket of LTA4H.

Mj HSP16.5 is a small heat shock protein (sHSP) from the hyperthermophilic methanoarchaeon, Methanococcus jannaschii (Mj), which lives at the environment of high temperature up to 94 °C. The structural data showed that Mj HSP16.5 was a 24-mer that formed a hollow sphere with octahedral symmetry. Mj HSP16.5 was very stable at pH 7 that it maintained the 24-mer structure even at 85 °C. In the present study, we investigated the unfolding process of Mj HSP16.5 in the presence of denaturants using several techniques, including circular dichroism (CD), dynamic light scattering (DLS), fluorescence spectroscopy, and size exclusive chromatography (SEC). We found that 8 mol·L−1 urea had no obvious effect on the structure of Mj HSP16.5 at pH 7. The unfolding of Mj HSP16.5 at pH 7 in the presence of guanidine hydrochloride (GdHCl) showed hierarchical behavior. Three significant transitions were observed around 2.0, 3.0, and 6.0 mol·L−1 GdHCl at pH 7. ANS (8-anilino-1- naphthalenesulfonic acid) titration results showed that the binding ability of Mj HSP16.5 to ANS decreased gradually as the concentration of GdHCl increased until around 2.0 mol·L−1 GdHCl, indicating surface hydrophobic area change, and this first transition was companioned with precipitation of Mj HSP16.5. Acrylamide quenching of fluorescence showed that the Stern-Volmer constant changed at about 3.0 mol·L−1 GdHCl, indicating changes of the dimeric interface, and this phase transition was companioned with oligomeric state change from 24-mer to small oligomers (4-mer to 8-mer). The last unfolding phase started around 5.0 mol·L−1 GdHCl, with a midpoint of 6.1 mol·L−1 GdHCl, and Mj HSP16.5 was completely unfolded at 7.0 mol·L−1 GdHCl. We also found that Mj HSP16.5 could be quite easily unfolded at pH 3, where it could be completely unfolded in 4.0 mol·L−1 GdHCl.

Protein–protein interface design is one of the most exciting fields in protein science; however, designing nonnatural protein–protein interaction pairs remains difficult. In this article we report a de novo design of a nonnatural protein–protein interaction pair by scanning the Protein Data Bank for suitable scaffold proteins that can be used for grafting key interaction residues and can form stable complexes with the target protein after additional mutations. Using our design algorithm, an unrelated protein, rat PLCδ1-PH (pleckstrin homology domain of phospholipase C-δ1), was successfully designed to bind the human erythropoietin receptor (EPOR) after grafting the key interaction residues of human erythropoietin binding to EPOR. The designed mutants of rat PLCδ1-PH were expressed and purified to test their binding affinities with EPOR. A designed triple mutation of PLCδ1-PH (ERPH1) was found to bind EPOR with high affinity (K D of 24 nM and an IC50 of 5.7 μM) both in vitro and in a cell-based assay, respectively, although the WT PLCδ1-PH did not show any detectable binding under the assay conditions. The in vitro binding affinities of the PLCδ1-PH mutants correlate qualitatively to the computational binding affinities, validating the design and the protein–protein interaction model. The successful practice of finding a proper protein scaffold and making it bind with EPOR demonstrates a prospective application in protein engineering targeting protein–protein interfaces.

The ubiquitin–proteasome pathway plays a crucial role in the regulation of many physiological processes and in the development of a number of major human diseases, such as cancer, Alzheimer’s, Parkinson’s, diabetes, etc. As a new target, the study on the proteasome inhibitors has received much attention recently. Three-dimensional quantitative structure–activity relationship (3D-QSAR) studies using comparative molecule field analysis (CoMFA) and comparative molecule similarity indices analysis (CoMSIA) techniques were applied to analyze the binding affinity of a set of tripeptide aldehyde inhibitors of 20S proteasome. The optimal CoMFA and CoMSIA models obtained for the training set were all statistically significant with cross-validated coefficients (q2) of 0.615, 0.591 and conventional coefficients (r2) of 0.901, 0.894, respectively. These models were validated by a test set of eight molecules that were not included in the training set. The predicted correlation coefficients (r2) of CoMFA and CoMSIA are 0.944 and 0.861, respectively. The CoMFA and CoMSIA field contour maps agree well with the structural characteristics of the binding pocket of β5 subunit of 20S proteasome, which suggests that the 3D-QSAR models built in this paper can be used to guide the development of novel inhibitors of 20S proteasome.

Human nonpancreatic secreted phospholipase A2 (hnps PLA2) is considered to be an important drug target for antiinflammation therapy. We have established a new fluorescence assay by using 1-anilinonaphthalene-8-sulfonate (ANS) as an interfacial probe for hydrophobic environment detection. The fitted apparent kcat/Km of hnps PLA2 is 0.0181 ± 0.0005 RFU/μM s. Tests on known synthesized inhibitor gave IC50 values similar to those from isotope-labeled assay. Because ANS is a commonly used probe for hydrophobic environment detection that needs no modification in the current assay, this strategy may be widely applicable for interfacial catalytic reactions.

3C-like proteinase of severe acute respiratory syndrome (SARS) coronavirus has been demonstrated to be a key target for drug design against SARS. The interaction between SARS coronavirus 3C-like (3CL) proteinase and an octapeptide interface inhibitor was studied by affinity capillary electrophoresis (ACE). The binding constants were estimated by the change of migration time of the analytes in the buffer solution containing different concentrations of SARS 3CL proteinase. The results showed that SARS 3CL proteinase was able to complex with the octapeptide competitively, with binding constants of 2.44 × 104 M−1 at 20 °C and 2.11 × 104 M−1 at 37 °C. In addition, the thermodynamic parameters deduced reveal that hydrophobic interaction might play major roles, along with electrostatic force, in the binding process. The ACE method used here could be developed to be an effective and simple way of applying large-scale drug screening and evaluation.

De novo protein design offers a unique means to test and advance our understanding of how proteins fold. However, most current design methods are native structure eccentric and folding kinetics has rarely been considered in the design process. Here, we show that a de novo designed mini-protein DS119, which folds into a βαβ structure, exhibits unusually slow and concentration-dependent folding kinetics. For example, the folding time for 50 μM of DS119 was estimated to be ∼2 s. Stopped-flow fluorescence resonance energy transfer experiments further suggested that its folding was likely facilitated by a transient dimerization process. Taken together, these results highlight the need for consideration of the entire folding energy landscape in de novo protein design and provide evidence suggesting nonnative interactions can play a key role in protein folding.

βαβ structural motifs are commonly used building blocks in protein structures containing parallel β-sheets. However, to our knowledge, no stand-alone βαβ structure has been observed in nature to date. Recently, for the first time that we know of, a small protein with an independent βαβ structure (DS119) was successfully designed in our laboratory. To understand the folding mechanism of DS119, in the study described here, we carried out all-atom molecular dynamics and coarse-grained simulations to investigate its folding pathways and energy landscape. From all-atom simulations, we successfully observed the folding event and got a stable folded structure with a minimal root mean-square deviation of 2.6 Å with respect to the NMR structure. The folding process can be described as a fast collapse phase followed by rapid formation of the central helix, and then slow formation of a parallel β-sheet. By using a native-centric Gō-like model, the cooperativity of the system was characterized in terms of the calorimetric criterion, sigmoidal transitions, conformation distribution shifts, and free-energy profiles. DS119 was found to be an incipient downhill folder that folds more cooperatively than a downhill folder, but less cooperatively than a two-state folder. This may reflect the balance between the two structural elements of DS119: the rapidly formed α-helix and the slowly formed parallel β-sheet. Folding times estimated from both the all-atom simulations and the coarse-grained model were at microsecond level, making DS119 another fast folder. Compared to fast folders reported previously, DS119 is, to the best of our knowledge, the first that exhibits a parallel β-sheet.

Statistical potentials have been widely used in protein studies despite the much-debated theoretical basis. In this work, we have applied two physical reference states for deriving the statistical potentials based on protein structure features to achieve zero interaction and orthogonalization. The free-rotating chain-based potential applies a local free-rotating chain reference state, which could theoretically be described by the Gaussian distribution. The self-avoiding chain-based potential applies a reference state derived from a database of artificial self-avoiding backbones generated by Monte Carlo simulation. These physical reference states are independent of known protein structures and are based solely on the analytical formulation or simulation method. The new potentials performed better and yielded higher Z-scores and success rates compared to other statistical potentials. The end-to-end distance distribution produced by the self-avoiding chain model was similar to the distance distribution of protein atoms in structure database. This fact may partly explain the basis of the reference states that depend on the atom pair frequency observed in the protein database. The current study showed that a more physical reference model improved the performance of statistical potentials in protein fold recognition, which could also be extended to other types of applications.