Structural basis of tolvaptan binding to the vasopressin V2 receptor.

The vasopressin V2 receptor (V2R) is a validated therapeutic target for autosomal dominant polycystic kidney disease (ADPKD), with tolvaptan being the first FDA-approved antagonist. Herein, we used Gaussian accelerated molecular dynamics simulations to investigate the spontaneous binding of tolvaptan to both active and inactive V2R conformations at the atomic-level. Overall, the binding process consists of two stages. Tolvaptan binds initially to extracellular loops 2 and 3 (ECL2/3) before overcoming an energy barrier to enter the pocket. Our simulations result highlighted key residues (e.g., R181, Y205, F287, F178) involved in this process, which were experimentally confirmed by site-directed mutagenesis. This work provides structural insights into tolvaptan-V2R interactions, potentially aiding the design of novel antagonists for V2R and other G protein-coupled receptors.

Beware of the generic machine learning-based scoring functions in structure-based virtual screening.

Machine learning-based scoring functions (MLSFs) have attracted extensive attention recently and are expected to be potential rescoring tools for structure-based virtual screening (SBVS). However, a major concern nowadays is whether MLSFs trained for generic uses rather than a given target can consistently be applicable for VS. In this study, a systematic assessment was carried out to re-evaluate the effectiveness of 14 reported MLSFs in VS. Overall, most of these MLSFs could hardly achieve satisfactory results for any dataset, and they could even not outperform the baseline of classical SFs such as Glide SP. An exception was observed for RFscore-VS trained on the Directory of Useful Decoys-Enhanced dataset, which showed its superiority for most targets. However, in most cases, it clearly illustrated rather limited performance on the targets that were dissimilar to the proteins in the corresponding training sets. We also used the top three docking poses rather than the top one for rescoring and retrained the models with the updated versions of the training set, but only minor improvements were observed. Taken together, generic MLSFs may have poor generalization capabilities to be applicable for the real VS campaigns. Therefore, it should be quite cautious to use this type of methods for VS.

Can machine learning consistently improve the scoring power of classical scoring functions? Insights into the role of machine learning in scoring functions.

How to accurately estimate protein-ligand binding affinity remains a key challenge in computer-aided drug design (CADD). In many cases, it has been shown that the binding affinities predicted by classical scoring functions (SFs) cannot correlate well with experimentally measured biological activities. In the past few years, machine learning (ML)-based SFs have gradually emerged as potential alternatives and outperformed classical SFs in a series of studies. In this study, to better recognize the potential of classical SFs, we have conducted a comparative assessment of 25 commonly used SFs. Accordingly, the scoring power was systematically estimated by using the state-of-the-art ML methods that replaced the original multiple linear regression method to refit individual energy terms. The results show that the newly-developed ML-based SFs consistently performed better than classical ones. In particular, gradient boosting decision tree (GBDT) and random forest (RF) achieved the best predictions in most cases. The newly-developed ML-based SFs were also tested on another benchmark modified from PDBbind v2007, and the impacts of structural and sequence similarities were evaluated. The results indicated that the superiority of the ML-based SFs could be fully guaranteed when sufficient similar targets were contained in the training set. Moreover, the effect of the combinations of features from multiple SFs was explored, and the results indicated that combining NNscore2.0 with one to four other classical SFs could yield the best scoring power. However, it was not applicable to derive a generic target-specific SF or SF combination.

Small-Molecule Conformer Generators: Evaluation of Traditional Methods and AI Models on High-Quality Data Sets.

Small-molecule conformer generation (SMCG) is an extremely important task in both ligand- and structure-based computer-aided drug design, especially during the hit discovery phase. Recently, a multitude of artificial intelligence (AI) models tailored for SMCG have emerged. Despite developers typically furnishing performance evaluation data upon releasing their AI models, a comprehensive and equitable performance comparison between AI models and conventional methods is still lacking. In this study, we curated a new benchmarking data set comprising 3354 high-quality ligand bioactive conformations. Subsequently, we conducted a systematic assessment of the performance of four widely adopted traditional methods (i.e., ConfGenX, Conformator, OMEGA, and RDKit ETKDG) and five AI models (i.e., ConfGF, DMCG, GeoDiff, GeoMol, and torsional diffusion) in the tasks of reproducing bioactive and low-energy conformations of small molecules. In the former task, the AI models have no advantage, particularly with a maximum ensemble size of 1. Even the best-performing AI model GeoMol is still worse than any of the tested traditional methods. Conversely, in the latter task, the torsional diffusion model shows obvious advantages, surpassing the best-performing traditional method ConfGenX by 26.09 and 12.97% on the COV-R and COV-P metrics, respectively. Furthermore, the influence of force field-based fine-tuning on the quality of the generated conformers was also discussed. Finally, a user-friendly Web server called fastSMCG was developed to enable researchers to rapidly and flexibly generate small-molecule conformers using both traditional and AI methods. We anticipate that our work will offer valuable practical assistance to the scientific community in this field.

Identification of CBPA as a New Inhibitor of PD-1/PD-L1 Interaction.

Targeting of the PD-1/PD-L1 immunologic checkpoint is believed to have provided a real breakthrough in the field of cancer therapy in recent years. Due to the intrinsic limitations of antibodies, the discovery of small-molecule inhibitors blocking PD-1/PD-L1 interaction has gradually opened valuable new avenues in the past decades. In an effort to discover new PD-L1 small molecular inhibitors, we carried out a structure-based virtual screening strategy to rapidly identify the candidate compounds. Ultimately, CBPA was identified as a PD-L1 inhibitor with a KD value at the micromolar level. It exhibited effective PD-1/PD-L1 blocking activity and T-cell-reinvigoration potency in cell-based assays. CBPA could dose-dependently elevate secretion levels of IFN-γ and TNF-α in primary CD4+ T cells in vitro. Notably, CBPA exhibited significant in vivo antitumor efficacy in two different mouse tumor models (a MC38 colon adenocarcinoma model and a melanoma B16F10 tumor model) without the induction of observable liver or renal toxicity. Moreover, analyses of the CBPA-treated mice further showed remarkably increased levels of tumor-infiltrating CD4+ and CD8+ T cells and cytokine secretion in the tumor microenvironment. A molecular docking study suggested that CBPA embedded relatively well into the hydrophobic cleft formed by dimeric PD-L1, occluding the PD-1 interaction surface of PD-L1. This study suggests that CBPA could work as a hit compound for the further design of potent inhibitors targeting the PD-1/PD-L1 pathway in cancer immunotherapy.

Molecular dynamics simulations reveal the disruption mechanism of a 2,4-thiazolidinedione derivative C30 against tau hexapeptide (PHF6) oligomer.

Derivatives of 2,4-thiazolidinedione have been reported to inhibit the aggregation of tau protein, in which compound 30 (C30) not only inhibit 80% of paired helical filament 6 (PHF6) aggregation, but also inhibit K18 and full-length tau aggregation. However, its inhibitory mechanism is unclear. In this study, to investigate the effect of C30 on tau protein, all-atom molecular dynamics simulation was performed on the PHF6 oligomer with and without C30. The results show that C30 can cause significant conformational changes in the PHF6 oligomer. The nematic order parameter P2 and secondary structure analyses show that C30 destroys the ordered structure of PHF6 oligomer, reduces the content of ß-sheet structure, and transforms ß-sheet into random coil structure. By clustering analysis, it was found that C30 has four possible binding sites on the PFH6 oligomer, and the binding ability order is S1 > S2 > S4 > S3. Following a more in-depth analyses of each site, it was determined that the S1 site is the most possible binding site mainly located between layers of L1 and L3. The hydrophobic interaction is the driving force for the binding of C30 to PHF6 oligomer. In addition, L1P4_Y310, L1P5_Y310, L3P1_V309, and L3P2_V309 are key residues for C30 binding to oligomer. Moreover, π-π interaction formed by L1P4_Y310 and L1P5_Y310 with C30 and the hydrogen bonding interaction formed by C30 with L3P3_Q307 are beneficial to the combination of C30 and oligomer. The fully understanding disrupt the mechanism of 2,4-thiazolidinedione derivative on PHF6 oligomer and the identification of binding sites will help design and discover new AD inhibitors in the future.

Deciphering the Effect of Lysine Acetylation on the Misfolding and Aggregation of Human Tau Fragment 171IPAKTPPAPK180 Using Molecular Dynamic Simulation and the Markov State Model.

The formation of neurofibrillary tangles (NFT) with ß-sheet-rich structure caused by abnormal aggregation of misfolded microtubule-associated protein Tau is a hallmark of tauopathies, including Alzheimer's Disease. It has been reported that acetylation, especially K174 located in the proline-rich region, can largely promote Tau aggregation. So far, the mechanism of the abnormal acetylation of Tau that affects its misfolding and aggregation is still unclear. Therefore, revealing the effect of acetylation on Tau aggregation could help elucidate the pathogenic mechanism of tauopathies. In this study, molecular dynamics simulation combined with multiple computational analytical methods were performed to reveal the effect of K174 acetylation on the spontaneous aggregation of Tau peptide 171IPAKTPPAPK180, and the dimerization mechanism as an early stage of the spontaneous aggregation was further specifically analyzed by Markov state model (MSM) analysis. The results showed that both the actual acetylation and the mutation mimicking the acetylated state at K174 induced the aggregation of the studied Tau fragment; however, the effect of actual acetylation on the aggregation was more pronounced. In addition, acetylated K174 plays a major contributing role in forming and stabilizing the antiparallel ß-sheet dimer by forming several hydrogen bonds and side chain van der Waals interactions with residues I171, P172, A173 and T175 of the corresponding chain. In brief, this study uncovered the underlying mechanism of Tau peptide aggregation in response to the lysine K174 acetylation, which can deepen our understanding on the pathogenesis of tauopathies.

An online target and rapid screening method for α-glucosidase inhibitors based on capillary electrophoresis.

Screening enzymatic active compounds is one of the important fields in drug research. α-Glucosidase can hydrolyze carbohydrates to monosaccharides after meals and lead to the rise of blood glucose levels in human body. Thus, the inhibition of α-glucosidase activity is an effective approach for the diabetes treatment. In this work, we developed a new method to simultaneously screen multiple bioactive compounds within a single CE running. The affect factors on the method performance, including injection, mixing, incubation, separation and detection, were carefully analyzed and discussed. Under the optimum, the mixture consisting of two internal standards (DMSO and 4-nitrophenol) and five compounds (lyoniresinol, hydroxytyrosol, rutin, kaempferol, and quercetin) was simultaneously screened, and kaempferol and quercetin showed stronger activity and this conclusion was also supported by offline assay. Furthermore, molecular docking was employed for investigating its interaction mechanism. Eventually, the established method has been applied to screen potential α-glucosidase inhibitors from an extract of Lycium barbarum and the peak area of rutin, taxifolin, quercetin, and chlorogenic acid in L. barbarum samples changed before and after the enzymatic reaction, confirming that these four compounds had potential inhibitory activities, which was consistent with the literature data. The present work provides a promising method for the target and rapid discovery of bioactive compounds from a plant extract or mixture.

Binding affinity and dissociation pathway predictions for a series of USP7 inhibitors with pyrimidinone scaffold by multiple computational methods.

Ubiquitin specific protease 7 (USP7) has attracted increasing attention because of its multifaceted roles in different tumor types. The crystal structures of USP7-inhibitor complexes resolved recently provide reliable models for computational structure-based drug design (SBDD) towards USP7. How to accurately estimate USP7-ligand binding affinity is quite critical to guarantee the reliability of SBDD. In this study, we assessed the reliability of multiple computational methods to the binding affinity prediction for a series of USP7 inhibitors with the pyrimidinone scaffold, including molecular docking scoring, MM/PB(GB)SA, and umbrella sampling (US). It was found that the accuracy of the evaluated computational methods for binding affinity prediction follows the order: US-based method > MM/PB(GB)SA > Glide XP scoring. The calculation results demonstrate that incorporating protein flexibility through induced-fit docking or ensemble docking cannot improve the performance of the Glide scoring based on rigid-receptor docking. For the MM/PB(GB)SA methods, the choice of the protein structure and the calculation procedure has a marked impact on the predictions. More importantly, we discovered for the first time that there are significant differences in the dissociation pathways of strong-binding inhibitors and weak-binding inhibitors of USP7, which may be used as a new criterion to judge whether an inhibitor is a strong binder or not. It is expected that our work can provide valuable guidance on the design and discovery of potent USP7 inhibitors.

The misfolding mechanism of the key fragment R3 of tau protein: a combined molecular dynamics simulation and Markov state model study.

The formation of neurofibrillary tangles (NFT) by abnormal aggregation of misfolded microtubule-associated protein tau is a hallmark of tauopathies, including Alzheimer's disease. However, it remains unclear how tau monomers undergo conformational changes and further lead to the abnormal aggregation. In this work, molecular dynamics simulation combined with the Markov state model (MSM) analysis was used to uncover the misfolding progress and structural characteristics of the key R3 fragment of tau protein at the atomic level. The simulation results show that R3 exists in disordered structures mainly, which is consistent with the experimental results. The MSM analysis identified multiple ß-sheet conformations of R3. The residues involved in the ß-sheet structure formation are mainly located in three regions: PHF6 at the N-terminal, S324 to N327 at the middle of R3, and K331 to G334 at the C-terminal. In addition, the path analysis of the formation of the ß-sheet structure by transition path theory (TPT) revealed that there are multiple paths to form ß-sheet structures from the disordered state, and the timescales are at the millisecond level, indicating that a large number of structural rearrangements occur during the formation of ß-sheet structures. It is interesting to note that S19 is a critical intermediate state for the formation of two target ß-sheet structures, S23 and S4. In S19, three regions of V306 to K311, C322 to G326, and K331 to G334 form a turn structure, the regions that form the ß-sheet structure in target states S23 and S4, indicating that the formation of a turn structure is necessary to form a ß-sheet structure and then the turn structure will eventually transform into the ß-sheet structure through key hydrogen bonding interactions. These findings can provide insights into the kinetics of tau protein misfolding.

Importance of a crystalline water network in docking-based virtual screening: a case study of BRD4.

As a member of the bromodomain and extra terminal domain (BET) protein family, bromodomain-containing protein 4 (BRD4) is an epigenetic reader and can recognize acetylated lysine residues in histones. BRD4 has been regarded as an essential drug target for cancers, inflammatory diseases and acute heart failure, and therefore the discovery of potent BRD4 inhibitors with novel scaffolds is highly desirable. In this study, the crystalline water molecules in BRD4 involved in ligand binding were analyzed first, and the simulation results suggest that several conserved crystalline water molecules are quite essential to keep the stability of the crystalline water network and therefore they need to be reserved in structure-based drug design. Then, a docking-based virtual screening workflow with the consideration of the conserved crystalline water network in the binding pocket was utilized to identify the potential inhibitors of BRD4. The in vitro fluorescence resonance energy transfer (HTRF) binding assay illustrates that 4 hits have good inhibitory activity against BRD4 in the micromolar regime, including three compounds with IC50 values below 5 µM and one below 1 µM (0.37 µM). The structural analysis demonstrates that three active compounds possess novel scaffolds. Moreover, the interaction patterns between the hits and BRD4 were characterized by molecular dynamics simulations and binding free energy calculations, and then several suggestions for the further optimization of these hits were proposed.

A Molecular Modeling Study of the Hydroxyflutamide Resistance Mechanism Induced by Androgen Receptor Mutations.

Hydroxyflutamide (HF), an active metabolite of the first generation antiandrogen flutamide, was used in clinic to treat prostate cancer targeting androgen receptor (AR). However, a drug resistance problem appears after about one year's treatment. AR T877A is the first mutation that was found to cause a resistance problem. Then W741C_T877A and F876L_T877A mutations were also reported to cause resistance to HF, while W741C and F876L single mutations cannot. In this study, molecular dynamics (MD) simulations combined with the molecular mechanics generalized Born surface area (MM-GBSA) method have been carried out to analyze the interaction mechanism between HF and wild-type (WT)/mutant ARs. The obtained results indicate that AR helix 12 (H12) plays a pivotal role in the resistance of HF. It can affect the coactivator binding site at the activation function 2 domain (AF2, surrounded by H3, H4, and H12). When H12 closes to the AR ligand-binding domain (LBD) like a lid, the coactivator binding site can be formed to promote transcription. However, once H12 is opened to expose LBD, the coactivator binding site will be distorted, leading to invalid transcription. Moreover, per-residue free energy decomposition analyses indicate that N705, T877, and M895 are vital residues in the agonist/antagonist mechanism of HF.

Discovery of Novel Inhibitors of BRD4 for Treating Prostate Cancer: A Comprehensive Case Study for Considering Water Networks in Virtual Screening and Drug Design.

Androgen receptor (AR) is the primary target for treating prostate cancer (PCa), which inevitably progresses due to drug-resistant mutations. Bromodomain-containing protein 4 (BRD4) has been a new potential drug target for PCa treatment. Herein, we report the rational design and discovery of novel BRD4 inhibitors through computer-aided drug design (CADD), and a hit compound SQ-1 (IC50 = 676 nM) was identified by structure-based virtual screening (SBVS) with the conserved water network. To optimize the structure of SQ-1, the free energy landscape was constructed, and the binding mechanism was explored by characterizing the water profile and the dissociation mechanism. Finally, the compound SQ-17 with improved inhibitory activity (IC50 < 100 nM) was discovered, which showed potent antiproliferative activity against LNCaP. These data highlighted a successful attempt to identify and optimize a small molecule by comprehensive CADD application and provided essential clues for developing novel therapeutics for PCa treatment.

Molecular Mechanism of Tau Misfolding and Aggregation: Insights from Molecular Dynamics Simulation.

Tau dysfunction has a close association with many neurodegenerative diseases, which are collectively referred to as tauopathies. Neurofibrillary tangles (NFTs) formed by misfolding and aggregation of tau are the main pathological process of tauopathy. Therefore, uncovering the misfolding and aggregation mechanism of tau protein will help to reveal the pathogenic mechanism of tauopathies. Molecular dynamics (MD) simulation is well suited for studying the dynamic process of protein structure changes. It provides detailed information on protein structure changes over time at the atomic resolution. At the same time, MD simulation can also simulate various conditions conveniently. Based on these advantages, MD simulations are widely used to study conformational transition problems such as protein misfolding and aggregation. Here, we summarized the structural features of tau, the factors affecting its misfolding and aggregation, and the applications of MD simulations in the study of tau misfolding and aggregation.

Donepezil Reduces H2O2-Inflicted Oxidative Stress and Necroptosis in Cardiomyocytes.

OBJECTIVE: Necroptosis, as a form of regulated cell necrosis, could participate in myocardial oxidative damage. We investigated whether donepezil attenuates H2O2-induced oxidative stress injury and necroptosis in rat cardiomyocytes. METHODS: H9c2 cells were incubated with H2O2 (final concentration of 1 mM) and then intervened with donepezil at doses of 2.5 and 10 µM. Subsequently, the necroptosis inhibitor necrostatin-1 (Nec-1) was introduced to treat H9c2 cells. For cell function experiments, cell proliferation; the contents of creatine kinase (CK), lactate dehydrogenase (LDH), superoxide dismutase (SOD), catalase (CAT), glutathione (GSH), and malondialdehyde (MDA); the protein and mRNA levels of the necroptosis-related proteins receptor-interacting serine-threonine kinase 3 (RIP3) and mixed lineage kinase-like (MLKL); and calcium ion fluorescence intensity were detected using Cell Counting Kit-8, enzyme-linked immunosorbent assay (ELISA), Western blotting, quantitative reverse transcription polymerase chain reaction, and flow cytometry, respectively. RESULTS: Cell viability was conspicuously decreased; CK and LDH contents, RIP3 and MLKL expression levels, and MDA production were preeminently elevated; and the production of SOD, CAT, and GSH was prominently reduced under H2O2 stimulation, which were dose-dependently countered by donepezil intervention. Nec-1 decreased the cell necroptosis, oxidative stress, and calcium overload caused by H2O2. However, on the premise of donepezil intervention, the addition of Nec-1 failed to further improve the situation, suggesting that donepezil exerts cardioprotective effects partly by inhibiting RIP3 and MLKL levels. CONCLUSION: Donepezil reduced H2O2-inflicted oxidative stress and necroptosis in cardiomyocytes by suppressing RIP3 and MLKL levels and calcium ion overload.

[White light upconversion luminescence and its mechanism of Yb3+/Ho3+/Tm3+ co-doped tellurite glass].

Yb3+ /Ho3+, Yb3+ /Tm3+ and Yb3+ /Ho3+ /Tm3+ co-doped tellurite glasses were prepared by melt-quenching method. Under the excitation of 980 nm laser, Yb3+ /Ho3+/Tm3+ co-doped glass sample shows strong blue, green and red emissions, corresponding to the transitions 1G4 --> 3H6 of Tm3+, 5F4 (5S2) --> 5 I8 of Ho3+, as well as 5F5 -->5 I8 of Ho3+ and 1G4 --> 3F4 of Tm3+ ions, respectively. It was found that the integrated emission intensity ratio of the red to green in Yb3+/Ho3+ /Tm3+ co-doped sample (3.95) is greater than that in Yb3+/Ho3+ co-doped sample (1.69) due to the cross-relaxation between Ho3+ and Tm3+ ions : 3H4 (Tm3+) + 5 I6 (Ho3+) -->3F4 (Tm3+) + 5F5 (Ho3+), and 3F4 (Tm3+ ) + 5 I8 (Ho3+) --> 3H6 (Tm3+) +5 I7 (Ho3+). When the pump power density is 8.2 W x cm(-2), the calculated color coordinates of Yb3+ /Ho3+ /Tm3+ co-doped sample are x = 0.345, y = 0.338, which is very close to the equal energy white light (x = 0.333, y = 0.333).

Fabrication and luminescent enhancement of Eu3+-doped Y2O3@YOF core-shell nanocrystals.

In this paper, a two-step synthesis method for preparing Eu3+ ion-doped Y2O3@YOF core-shell nanocrystals is introduced. Eu3+ ion-doped Y2O3@YOF core-shell nanocrystals were prepared by combining an autocombustion process with a low temperature solid state reaction. X-ray powder diffraction (XRD), field emission scanning electron microscopy (FE-SEM), transmission electron microscopy (TEM), energy dispersive X-ray spectroscopy (EDS), photoluminescence (PL) and fluorescence decay were employed to characterize the prepared samples. The results of XRD, TEM and EDS indicated that the products prepared by this method were not a mixture of Y2O3:Eu3+ and YOF:Eu3+ nanocrystals, but Eu3+ ion-doped Y2O3@YOF core-shell nanocrystals. Compared with Y2O3:Eu3+ nanocrystals, a 20% increment in luminescence intensity was observed in the Eu3+ ion-doped Y2O3@YOF core-shell nanocrystals, thus suggesting that coating with a YOF:Eu3+-shell can efficiently block the nonradiative relaxation channels that are induced by surface defect states.

Nitrogen doped ZnO thin films prepared by photo-assisted metal-organic chemical vapor deposition.

Nitrogen-doped ZnO (ZnO:N) films were prepared by photo-assisted metal-organic chemical vapor deposition technique using NH3 as N doping source. The effects of in-situ light irradiation on the properties of ZnO:N films were studied by Hall measurements, X-ray diffraction, Raman scattering, and X-ray photoelectron spectroscopy. The results show that stable p-type ZnO:N films with a hole concentration of 3.61 x 10(17) cm(-3) was successfully achieved. Moreover, introducing proper in-situ light irradiation during the growth process can not only effectively improve the crystalline quality of ZnO films, but also enhance the activity of (N)o (N occupies O site) acceptors by removing the undesirable hydrogen atoms from ZnO:N films. Both effects are benefit for the p-type conductivity formation. Our results indicate that photo-assisted MOCVD maybe an effective technology to realize device-quality p-type ZnO:N films.

Uncovering the Effect of pS202/pT205/pS208 Triple Phosphorylations on the Conformational Features of the Key Fragment G192-T212 of Tau Protein.

Microtubule-associated protein tau is abnormally phosphorylated and forms the aggregates of paired helical filaments in Alzheimer's disease (AD) and other tauopathies. So far, the relationship and mechanism between the abnormal phosphorylation of tau and fibril formation is still unclear. Therefore, studying the effect of phosphorylation on the structure of tau protein is helpful to elucidate the pathogenic mechanism of tauopathies. It has been shown that pS202/pT205/pS208 triple phosphorylations located in the proline-rich region can promote tau aggregation. In this work, the effect of triple phosphorylations on tau structure was investigated by molecular dynamics simulations combined with multiple analytical methods of trajectories. The results showed that the conformational diversity of G192-T212 fragments decreased after phosphorylation compared with that of the wild-type. Moreover, the dynamic network and hydrogen bond analyses showed that the addition of pS208 phosphorylation can destroy the key hydrogen bonds and the network structure formed centered on pT205 at the C-terminal of the pS202/pT205 double phosphorylated peptide and then destroy the turn structure formed in the region of G207-R211. The destruction of this turn structure is considered to be the main reason for the aggregation of pS202/pT205/pS208 triple phosphorylations. For the pS202/pT205/pS208 triple phosphorylated system, the G207-R211 region is a coil structure, which is more extended and prone to aggregation. In a word, our results reveal the mechanism that pS202/pT205/pS208 triple phosphorylations promote tau aggregation at the atomic level, which can provide useful theoretical guidance for the rational design of effective therapeutic drugs against AD and other tauopathies.

The impact of cross-docked poses on performance of machine learning classifier for protein-ligand binding pose prediction.

Structure-based drug design depends on the detailed knowledge of the three-dimensional (3D) structures of protein-ligand binding complexes, but accurate prediction of ligand-binding poses is still a major challenge for molecular docking due to deficiency of scoring functions (SFs) and ignorance of protein flexibility upon ligand binding. In this study, based on a cross-docking dataset dedicatedly constructed from the PDBbind database, we developed several XGBoost-trained classifiers to discriminate the near-native binding poses from decoys, and systematically assessed their performance with/without the involvement of the cross-docked poses in the training/test sets. The calculation results illustrate that using Extended Connectivity Interaction Features (ECIF), Vina energy terms and docking pose ranks as the features can achieve the best performance, according to the validation through the random splitting or refined-core splitting and the testing on the re-docked or cross-docked poses. Besides, it is found that, despite the significant decrease of the performance for the threefold clustered cross-validation, the inclusion of the Vina energy terms can effectively ensure the lower limit of the performance of the models and thus improve their generalization capability. Furthermore, our calculation results also highlight the importance of the incorporation of the cross-docked poses into the training of the SFs with wide application domain and high robustness for binding pose prediction. The source code and the newly-developed cross-docking datasets can be freely available at https://github.com/sc8668/ml_pose_prediction and https://zenodo.org/record/5525936 , respectively, under an open-source license. We believe that our study may provide valuable guidance for the development and assessment of new machine learning-based SFs (MLSFs) for the predictions of protein-ligand binding poses.