Application scenario-oriented molecule generation platform developed for drug discovery.

Design of molecules for candidate compound selection is one of the central challenges in drug discovery due to the complexity of chemical space and requirement of multi-parameter optimization. Here we present an application scenario-oriented platform (ID4Idea) for molecule generation in different scenarios of drug discovery. This platform utilizes both library or rule based and generative based algorithms (VAE, RNN, GAN, etc.), in combination with various AI learning types (pre-training, transfer learning, reinforcement learning, active learning, etc.) and input representations (1D SMILES, 2D graph, 3D shape, binding site, pharmacophore, etc.), to enable customized solutions for a given molecular design scenario. Besides the usual generation followed screening protocol, goal-directed molecule generation can also be conducted towards predefined goals, enhancing the efficiency of hit identification, lead finding, and lead optimization. We demonstrate the effectiveness of ID4Idea platform through case studies, showcasing customized solutions for different design tasks using various input information, such as binding pockets, pharmacophores, and compound representations. In addition, remaining challenges are discussed to unlock the full potential of AI models in drug discovery and pave the way for the development of novel therapeutics.

Hit Identification Driven by Combining Artificial Intelligence and Computational Chemistry Methods: A PI5P4K-ß Case Study.

Computer-aided drug design (CADD), especially artificial intelligence-driven drug design (AIDD), is increasingly used in drug discovery. In this paper, a novel and efficient workflow for hit identification was developed within the ID4Inno drug discovery platform, featuring innovative artificial intelligence, high-accuracy computational chemistry, and high-performance cloud computing. The workflow was validated by discovering a few potent hit compounds (best IC50 is ∼0.80 µM) against PI5P4K-ß, a novel anti-cancer target. Furthermore, by applying the tools implemented in ID4Inno, we managed to optimize these hit compounds and finally obtained five hit series with different scaffolds, all of which showed high activity against PI5P4K-ß. These results demonstrate the effectiveness of ID4inno in driving hit identification based on artificial intelligence, computational chemistry, and cloud computing.

A Cloud Computing Platform for Scalable Relative and Absolute Binding Free Energy Predictions: New Opportunities and Challenges for Drug Discovery.

Free energy perturbation (FEP) has become widely used in drug discovery programs for binding affinity prediction between candidate compounds and their biological targets. However, limitations of FEP applications also exist, including, but not limited to, high cost, long waiting time, limited scalability, and breadth of application scenarios. To overcome these problems, we have developed XFEP, a scalable cloud computing platform for both relative and absolute free energy predictions using optimized simulation protocols. XFEP enables large-scale FEP calculations in a more efficient, scalable, and affordable way, for example, the evaluation of 5000 compounds can be performed in 1 week using 50-100 GPUs with a computing cost roughly equivalent to the cost for the synthesis of only one new compound. By combining these capabilities with artificial intelligence techniques for goal-directed molecule generation and evaluation, new opportunities can be explored for FEP applications in the drug discovery stages of hit identification, hit-to-lead, and lead optimization based not only on structure exploitation within the given chemical series but also including evaluation and comparison of completely unrelated molecules during structure exploration in a larger chemical space. XFEP provides the basis for scalable FEP applications to become more widely used in drug discovery projects and to speed up the drug discovery process from hit identification to preclinical candidate compound nomination.

Scaffold Hopping Transformations Using Auxiliary Restraints for Calculating Accurate Relative Binding Free Energies.

In silico screening of drug-target interactions is a key part of the drug discovery process. Changes in the drug scaffold via contraction or expansion of rings, the breaking of rings, and the introduction of cyclic structures from acyclic structures are commonly applied by medicinal chemists to improve binding affinity and enhance favorable properties of candidate compounds. These processes, commonly referred to as scaffold hopping, are challenging to model computationally. Although relative binding free energy (RBFE) calculations have shown success in predicting binding affinity changes caused by perturbing R-groups attached to a common scaffold, applications of RBFE calculations to modeling scaffold hopping are relatively limited. Scaffold hopping inevitably involves breaking and forming bond interactions of quadratic functional forms, which is highly challenging. A novel method for handling ring opening/closure/contraction/expansion and linker contraction/expansion is presented here. To the best of our knowledge, RBFE calculations on linker contraction/expansion have not been previously reported. The method uses auxiliary restraints to hold the atoms at the ends of a bond in place during the breaking and forming of the bonds. The broad applicability of the method was demonstrated by examining perturbations involving small-molecule macrocycles and mutations of proline in proteins. High accuracy was obtained using the method for most of the perturbations studied. The rigor of the method was isolated from the force field by validating the method using relative and absolute hydration free energy calculations compared to standard simulation results. Unlike other methods that rely on λ-dependent functional forms for bond interactions, the method presented here can be employed using modern molecular dynamics software without modification of codes or force field functions.

Molecular Understanding of Laccase Adsorption on Charged Self-Assembled Monolayers.

Controlling the orientation of laccase on electrodes is crucial for the achievement of fast direct electron transfer. It is important to find a short pathway between the T1 copper site of laccase and a substrate during the laccase immobilization. In this work, we studied the adsorption orientation and conformation of Trametes versicolor laccase (TvL) on two kinds of charged self-assembled monolayers (SAMs), including NH2-SAM and COOH-SAM, by parallel tempering Monte Carlo and all-atom molecular dynamics simulations. TvL adsorbs on positively and negatively charged surface with "end-on" and "lying" orientation, respectively. On the positively charged surface, T1 copper site of TvL is closer to the surface. The orientation of TvL on positively charged surface is narrower than that on negatively charged surface. Thus, the positively charged surface is more conducive to the immobilization of TvL. The conformational changes of TvL on the charged surfaces are analyzed by RMSD, superimposed structures, dipole moment, gyration radius, and eccentricity. Results show that native structures of TvL are well preserved when it adsorbs on the charged surfaces. This work provides atomistic insight into the mechanism of TvL adsorption on charged surface and is helpful for the design and development of laccase-based electrodes.

Molecular Understanding of the Penetration of Functionalized Gold Nanoparticles into Asymmetric Membranes.

In this work, the interactions between surface-functionalized gold nanoparticles (AuNPs) and asymmetric membranes and the associated cytotoxicity were explored by coarse-grained molecular dynamics simulations. Simulation results show that the surface chemistry of AuNPs and the asymmetry of lipid membranes play significant roles. AuNPs with different signs of charges spontaneously adhere to the membrane surface or penetrate the membrane core. Also, the asymmetric distribution of charged lipids in membranes can facilitate the penetration of cationic AuNPs. Increasing the surface charge density (SCD) of AuNPs can not only improve the penetration efficiency but also lead to more disruption of the membrane structure. Moreover, the flip-flop of charged lipids in the inner leaflet can be observed during the translocation of AuNPs with a high SCD. The breakdown of membrane asymmetry may hinder the cellular internalization of AuNPs in a direct penetration mechanism. More importantly, we demonstrate that the hydrophobic contact between protruding solvent-exposed lipid tails and the hydrophobic moieties of ligands can mediate the insertion of AuNPs with a low SCD into cell membranes, which will exhibit less cytotoxicity in most in vivo applications. This may open a new exciting avenue to developing nanocarriers with a higher translocation efficiency and a lower toxicity simultaneously for biomedical applications.

Molecular simulations of cytochrome c adsorption on positively charged surfaces: the influence of anion type and concentration.

Contradictory results have been reported regarding cytochrome c (Cyt-c) adsorption onto the positively charged SAMs, and the role of small anions in the adsorption is still unclear. In this work, the adsorption of Cyt-c on the amino-terminated SAM (NH2-SAM) and the effect of chloride and phosphate ions on the adsorption were studied using molecular dynamics simulations. The results reveal that Cyt-c could not stably adsorb onto the surface even at a relatively high ionic strength when chloride ions were added, while phosphate ions could promote its adsorption. At a low phosphate concentration, Cyt-c can adsorb on the NH2-SAM mainly with two opposite orientations. One is similar to that characterized in the experiments for Cyt-c adsorbed on the NH2-SAMs, in which the heme group points far away from the surface. The other orientation is similar to that for Cyt-c on the carboxyl-terminated SAMs. In the latter case, phosphate ions formed a distinct counterion layer near the surface and overcompensated the positive charge of the surface. Further analysis shows that chloride ions have no significant tendency to aggregate near the NH2-SAM surface and cannot shield the electrostatic repulsion between Cyt-c and the surface, while the phosphate ions can easily adsorb onto the surface and bind specifically to certain lysine residues of Cyt-c, which mediate its adsorption. At a high phosphate concentration, the phosphate and sodium ions will aggregate to form clusters, which results in random adsorption orientation. This work may provide some guidance for the design of Cyt-c-based bioelectronic devices and controlled enzyme immobilization.

Structural properties of polymer-brush-grafted gold nanoparticles at the oil-water interface: insights from coarse-grained simulations.

In this work, the structural properties of amphiphilic polymer-brush-grafted gold nanoparticles (AuNPs) at the oil-water interface were investigated by coarse-grained simulations. The effects of grafting architecture (diblock, mixed and Janus brush-grafted AuNPs) and hydrophilicity of polymer brushes are discussed. Simulation results indicate that functionalized AuNPs present abundant morphologies including typical core-shell, Janus-type, jellyfish-like, etc., in a water or oil-water solvent environment. It is found that hydrophobic/weak hydrophilic polymer-brush-grafted AuNPs have better phase transfer performance, especially for AuNPs modified with hydrophobic chains as outer blocks and weak hydrophilic chains as inner blocks. This kind of AuNP can cross the interface region and move into the oil phase completely. For hydrophobic/strong hydrophilic polymer-brush-grafted AuNPs, they are trapped in the interface region instead of moving into any phase. The mechanism of phase transfer is ascribed to the flexibility and mobility of outer blocks. Besides, we study the desorption energy by PMF analysis. The results demonstrate that Janus brush-grafted AuNPs show the highest interfacial stability and activity, which can be further strengthened by increasing the hydrophilicity of grafted polymer brushes. This work will promote the industrial applications of polymer-brush-grafted NPs such as phase transfer catalysis and Pickering emulsion catalysis.

Molecular simulation study of feruloyl esterase adsorption on charged surfaces: effects of surface charge density and ionic strength.

The surrounding conditions, such as surface charge density and ionic strength, play an important role in enzyme adsorption. The adsorption of a nonmodular type-A feruloyl esterase from Aspergillus niger (AnFaeA) on charged surfaces was investigated by parallel tempering Monte Carlo (PTMC) and all-atom molecular dynamics (AAMD) simulations at different surface charge densities (±0.05 and ±0.16 C·m(-2)) and ionic strengths (0.007 and 0.154 M). The adsorption energy, orientation, and conformational changes were analyzed. Simulation results show that whether AnFaeA can adsorb onto a charged surface is mainly controlled by electrostatic interactions between AnFaeA and the charged surface. The electrostatic interactions between AnFaeA and charged surfaces are weakened when the ionic strength increases. The positively charged surface at low surface charge density and high ionic strength conditions can maximize the utilization of the immobilized AnFaeA. The counterion layer plays a key role in the adsorption of AnFaeA on the negatively charged COOH-SAM. The native conformation of AnFaeA is well preserved under all of these conditions. The results of this work can be used for the controlled immobilization of AnFaeA.

Lipase adsorption on different nanomaterials: a multi-scale simulation study.

Candida antarctica lipase B (CalB) is an efficient biocatalyst for hydrolysis and esterification, which plays an important role in the production of biodiesel in the bioenergy industries. The ordered immobilisation of lipases on different supports would be significant for its enzymatic catalysis in some biodiesel production processes; however, the underlying mechanisms and the preferred lipase orientation are not well understood yet. In this work, a fundamental understanding of the orientation and adsorption mechanism of lipase on four different nanomaterial surfaces with different surface chemistry are explored in detail by a combination of parallel tempering Monte Carlo (PTMC) and molecular dynamics (MD) simulations. Simulation results show that lipase is strongly adsorbed onto the hydrophobic graphite surface, as reflected by the large contact area and interaction energy; while the adsorption onto the hydrophilic TiO2 surface is weak due to two strongly adhered water layers; meanwhile lipase undergoes desorption and reorientation processes. For CalB adsorption on positively and negatively charged surfaces (NH2-SAM and COOH-SAM), the orientation distributions of lipase are narrow, and opposite orientations are obtained. CalB adsorbed on NH2-SAM has its catalytic centre oriented towards the surface, which is not conducive to the substrate binding; while the catalytic centre faces toward the solution when it is adsorbed on the COOH-SAM. Besides, the native structures of CalB adsorbed on different surfaces are preserved, which indicates lipase as a robust enzyme. The simulation results will promote our understanding on how surface properties of nanomaterials, such as charge or hydrophobicity, will affect lipase immobilisation, and help us in the rational design and development of immobilised lipase carriers.

Adsorption of hydrophobin on different self-assembled monolayers: the role of the hydrophobic dipole and the electric dipole.

In this work, the adsorptions of hydrophobin (HFBI) on four different self-assembled monolayers (SAMs) (i.e., CH3-SAM, OH-SAM, COOH-SAM, and NH2-SAM) were investigated by parallel tempering Monte Carlo and molecular dynamics simulations. Simulation results indicate that the orientation of HFBI adsorbed on neutral surfaces is dominated by a hydrophobic dipole. HFBI adsorbs on the hydrophobic CH3-SAM through its hydrophobic patch and adopts a nearly vertical hydrophobic dipole relative to the surface, while it is nearly horizontal when adsorbed on the hydrophilic OH-SAM. For charged SAM surfaces, HFBI adopts a nearly vertical electric dipole relative to the surface. HFBI has the narrowest orientation distribution on the CH3-SAM, and thus can form an ordered monolayer and reverse the wettability of the surface. For HFBI adsorption on charged SAMs, the adsorption strength weakens as the surface charge density increases. Compared with those on other SAMs, a larger area of the hydrophobic patch is exposed to the solution when HFBI adsorbs on the NH2-SAM. This leads to an increase of the hydrophobicity of the surface, which is consistent with the experimental results. The binding of HFBI to the CH3-SAM is mainly through hydrophobic interactions, while it is mediated through a hydration water layer near the surface for the OH-SAM. For the charged SAM surfaces, the adsorption is mainly induced by electrostatic interactions between the charged surfaces and the oppositely charged residues. The effect of a hydrophobic dipole on protein adsorption onto hydrophobic surfaces is similar to that of an electric dipole for charged surfaces. Therefore, the hydrophobic dipole may be applied to predict the probable orientations of protein adsorbed on hydrophobic surfaces.