Mining for Potent Inhibitors through Artificial Intelligence and Physics: A Unified Methodology for Ligand Based and Structure Based Drug Design.

Determining the viability of a new drug molecule is a time- and resource-intensive task that makes computer-aided assessments a vital approach to rapid drug discovery. Here we develop a machine learning algorithm, iMiner, that generates novel inhibitor molecules for target proteins by combining deep reinforcement learning with real-time 3D molecular docking using AutoDock Vina, thereby simultaneously creating chemical novelty while constraining molecules for shape and molecular compatibility with target active sites. Moreover, through the use of various types of reward functions, we have introduced novelty in generative tasks for new molecules such as chemical similarity to a target ligand, molecules grown from known protein bound fragments, and creation of molecules that enforce interactions with target residues in the protein active site. The iMiner algorithm is embedded in a composite workflow that filters out Pan-assay interference compounds, Lipinski rule violations, uncommon structures in medicinal chemistry, and poor synthetic accessibility with options for cross-validation against other docking scoring functions and automation of a molecular dynamics simulation to measure pose stability. We also allow users to define a set of rules for the structures they would like to exclude during the training process and postfiltering steps. Because our approach relies only on the structure of the target protein, iMiner can be easily adapted for the future development of other inhibitors or small molecule therapeutics of any target protein.

Glycoproteomic landscape and structural dynamics of TIM family immune checkpoints enabled by mucinase SmE.

Mucin-domain glycoproteins are densely O-glycosylated and play critical roles in a host of biological functions. In particular, the T cell immunoglobulin and mucin-domain containing family of proteins (TIM-1, -3, -4) decorate immune cells and act as key regulators in cellular immunity. However, their dense O-glycosylation remains enigmatic, primarily due to the challenges associated with studying mucin domains. Here, we demonstrate that the mucinase SmE has a unique ability to cleave at residues bearing very complex glycans. SmE enables improved mass spectrometric analysis of several mucins, including the entire TIM family. With this information in-hand, we perform molecular dynamics (MD) simulations of TIM-3 and -4 to understand how glycosylation affects structural features of these proteins. Finally, we use these models to investigate the functional relevance of glycosylation for TIM-3 function and ligand binding. Overall, we present a powerful workflow to better understand the detailed molecular structures and functions of the mucinome.

SARS-CoV-2 evolved variants optimize binding to cellular glycocalyx.

Viral variants of concern continue to arise for SARS-CoV-2, potentially impacting both methods for detection and mechanisms of action. Here, we investigate the effect of an evolving spike positive charge in SARS-CoV-2 variants and subsequent interactions with heparan sulfate and the angiotensin converting enzyme 2 (ACE2) in the glycocalyx. We show that the positively charged Omicron variant evolved enhanced binding rates to the negatively charged glycocalyx. Moreover, we discover that while the Omicron spike-ACE2 affinity is comparable to that of the Delta variant, the Omicron spike interactions with heparan sulfate are significantly enhanced, giving rise to a ternary complex of spike-heparan sulfate-ACE2 with a large proportion of double-bound and triple-bound ACE2. Our findings suggest that SARS-CoV-2 variants evolve to be more dependent on heparan sulfate in viral attachment and infection. This discovery enables us to engineer a second-generation lateral-flow test strip that harnesses both heparin and ACE2 to reliably detect all variants of concern, including Omicron.

Transmissible SARS-CoV-2 variants with resistance to clinical protease inhibitors.

Vaccines and drugs have helped reduce disease severity and blunt the spread of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). However, ongoing virus transmission, continuous evolution, and increasing selective pressures have the potential to yield viral variants capable of resisting these interventions. Here, we investigate the susceptibility of natural variants of the main protease [Mpro; 3C-like protease (3CLpro)] of SARS-CoV-2 to protease inhibitors. Multiple single amino acid changes in Mpro confer resistance to nirmatrelvir (the active component of Paxlovid). An additional clinical-stage inhibitor, ensitrelvir (Xocova), shows a different resistance mutation profile. Importantly, phylogenetic analyses indicate that several of these resistant variants have pre-existed the introduction of these drugs into the human population and are capable of spreading. These results encourage the monitoring of resistance variants and the development of additional protease inhibitors and other antiviral drugs with different mechanisms of action and resistance profiles for combinatorial therapy.

Glycoproteomic landscape and structural dynamics of TIM family immune checkpoints enabled by mucinase SmE.

Mucin-domain glycoproteins are densely O-glycosylated and play critical roles in a host of biological functions. In particular, the T cell immunoglobulin and mucin-domain containing family of proteins (TIM-1, -3, -4) decorate immune cells and act as key checkpoint inhibitors in cancer. However, their dense O-glycosylation remains enigmatic both in terms of glycoproteomic landscape and structural dynamics, primarily due to the challenges associated with studying mucin domains. Here, we present a mucinase (SmE) and demonstrate its ability to selectively cleave along the mucin glycoprotein backbone, similar to others of its kind. Unlike other mucinases, though, SmE harbors the unique ability to cleave at residues bearing extremely complex glycans which enabled improved mass spectrometric analysis of several mucins, including the entire TIM family. With this information in-hand, we performed molecular dynamics (MD) simulations of TIM-3 and -4 to demonstrate how glycosylation affects structural features of these proteins. Overall, we present a powerful workflow to better understand the detailed molecular structures of the mucinome.

Targeted protein S-nitrosylation of ACE2 inhibits SARS-CoV-2 infection.

Prevention of infection and propagation of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) is a high priority in the Coronavirus Disease 2019 (COVID-19) pandemic. Here we describe S-nitrosylation of multiple proteins involved in SARS-CoV-2 infection, including angiotensin-converting enzyme 2 (ACE2), the receptor for viral entry. This reaction prevents binding of ACE2 to the SARS-CoV-2 spike protein, thereby inhibiting viral entry, infectivity and cytotoxicity. Aminoadamantane compounds also inhibit coronavirus ion channels formed by envelope (E) protein. Accordingly, we developed dual-mechanism aminoadamantane nitrate compounds that inhibit viral entry and, thus, the spread of infection by S-nitrosylating ACE2 via targeted delivery of the drug after E protein channel blockade. These non-toxic compounds are active in vitro and in vivo in the Syrian hamster COVID-19 model and, thus, provide a novel avenue to pursue therapy.

Transmissible SARS-CoV-2 variants with resistance to clinical protease inhibitors.

Vaccines and drugs have helped reduce disease severity and blunt the spread of SARS-CoV-2. However, ongoing virus transmission, continuous evolution, and increasing selective pressures have the potential to yield viral variants capable of resisting these interventions. Here, we investigate the susceptibility of natural variants of the main protease (Mpro/3CLpro) of SARS-CoV-2 to protease inhibitors. Multiple single amino acid changes in Mpro confer resistance to nirmatrelvir (the active component of Paxlovid). An additional clinical-stage inhibitor, ensitrelvir (Xocova), shows a different resistance mutation profile. Importantly, phylogenetic analyses indicate that several of these resistant variants have pre-existed the introduction of these drugs into the human population and are capable of spreading. These results encourage the monitoring of resistance variants and the development of additional protease inhibitors and other antiviral drugs with different mechanisms of action and resistance profiles for combinatorial therapy.

Spike-heparan sulfate interactions in SARS-CoV-2 infection.

Recent biochemical, biophysical, and genetic studies have shown that heparan sulfate, a major component of the cellular glycocalyx, participates in infection of SARS-CoV-2 by facilitating the so-called open conformation of the spike protein, which is required for binding to ACE2. This review highlights the involvement of heparan sulfate in the SARS-CoV-2 infection cycle and argues that there is a high degree of coordination between host cell heparan sulfate and asparagine-linked glycans on the spike in enabling ACE2 binding and subsequent infection. The discovery that spike protein binding and infection depends on both viral and host glycans provides insights into the evolution, spread and potential therapies for SARS-CoV-2 and its variants.

Tuaimenal A, a Meroterpene from the Irish Deep-Sea Soft Coral Duva florida, Displays Inhibition of the SARS-CoV-2 3CLpro Enzyme.

Cold water benthic environments are a prolific source of structurally diverse molecules with a range of bioactivities against human disease. Specimens of a previously chemically unexplored soft coral, Duva florida, were collected during a deep-sea cruise that sampled marine invertebrates along the Irish continental margin in 2018. Tuaimenal A (1), a cyclized merosesquiterpenoid representing a new carbon scaffold with a highly substituted chromene core, was discovered through exploration of the soft coral secondary metabolome via NMR-guided fractionation. The absolute configuration was determined through vibrational circular dichroism. Functional biochemical assays and in silico docking experiments found tuaimenal A selectively inhibits the viral main protease (3CLpro) of SARS-CoV-2.

Targeted protein S-nitrosylation of ACE2 as potential treatment to prevent spread of SARS-CoV-2 infection.

Prevention of infection and propagation of SARS-CoV-2 is of high priority in the COVID-19 pandemic. Here, we describe S-nitrosylation of multiple proteins involved in SARS-CoV-2 infection, including angiotensin converting enzyme 2 (ACE2), the receptor for viral entry. This reaction prevents binding of ACE2 to the SARS-CoV-2 Spike protein, thereby inhibiting viral entry, infectivity, and cytotoxicity. Aminoadamantane compounds also inhibit coronavirus ion channels formed by envelope (E) protein. Accordingly, we developed dual-mechanism aminoadamantane nitrate compounds that inhibit viral entry and thus spread of infection by S-nitrosylating ACE2 via targeted delivery of the drug after E-protein channel blockade. These non-toxic compounds are active in vitro and in vivo in the Syrian hamster COVID-19 model, and thus provide a novel avenue for therapy.

GlycoGrip: Cell Surface-Inspired Universal Sensor for Betacoronaviruses.

Inspired by the role of cell-surface glycoproteins as coreceptors for pathogens, we report the development of GlycoGrip: a glycopolymer-based lateral flow assay for detecting SARS-CoV-2 and its variants. GlycoGrip utilizes glycopolymers for primary capture and antispike antibodies labeled with gold nanoparticles for signal-generating detection. A lock-step integration between experiment and computation has enabled efficient optimization of GlycoGrip test strips which can selectively, sensitively, and rapidly detect SARS-CoV-2 and its variants in biofluids. Employing the power of the glycocalyx in a diagnostic assay has distinct advantages over conventional immunoassays as glycopolymers can bind to antigens in a multivalent capacity and are highly adaptable for mutated strains. As new variants of SARS-CoV-2 are identified, GlycoGrip will serve as a highly reconfigurable biosensor for their detection. Additionally, via extensive ensemble-based docking simulations which incorporate protein and glycan motion, we have elucidated important clues as to how heparan sulfate and other glycocalyx components may bind the spike glycoprotein during SARS-CoV-2 host-cell infection. GlycoGrip is a promising and generalizable alternative to costly, labor-intensive RT-PCR, and we envision it will be broadly useful, including for rural or low-income populations that are historically undertested and under-reported in infection statistics.

CIFDock: A novel CHARMM-based flexible receptor-flexible ligand docking protocol.

Docking studies play a critical role in the current workflow of drug discovery. However, limitations may often arise through factors including inadequate ligand sampling, a lack of protein flexibility, scoring function inadequacies (e.g., due to metals, co-factors, etc.), and difficulty in retaining explicit water molecules. Herein, we present a novel CHARMM-based induced fit docking (CIFDock) workflow that can circumvent these limitations by employing all-atom force fields coupled to enhanced sampling molecular dynamics procedures. Self-guided Langevin dynamics simulations are used to effectively sample relevant ligand conformations, side chain orientations, crystal water positions, and active site residue motion. Protein flexibility is further enhanced by dynamic sampling of side chain orientations using an expandable rotamer library. Steps in the procedure consisting of fixing individual components (e.g., the ligand) while sampling the other components (e.g., the residues in the active site of the protein) allow for the complex to adapt to conformational changes. Ultimately, all components of the complex-the protein, ligand, and waters-are sampled simultaneously and unrestrained with SGLD to capture any induced fit effects. This modular flexible docking procedure is automated using CHARMM scripting, interfaced with SLURM array processing, and parallelized to use the desired number of processors. We validated the CIFDock procedure by performing cross-docking studies using a data set comprised of 21 pharmaceutically relevant proteins. Five variants of the CHARMM-based SWISSDOCK scoring functions were created to quantify the results of the final generated poses. Results obtained were comparable to, or in some cases improved upon, commercial docking program data.

A glycan gate controls opening of the SARS-CoV-2 spike protein.

SARS-CoV-2 infection is controlled by the opening of the spike protein receptor binding domain (RBD), which transitions from a glycan-shielded 'down' to an exposed 'up' state to bind the human angiotensin-converting enzyme 2 receptor and infect cells. While snapshots of the 'up' and 'down' states have been obtained by cryo-electron microscopy and cryo-electron tomagraphy, details of the RBD-opening transition evade experimental characterization. Here over 130 µs of weighted ensemble simulations of the fully glycosylated spike ectodomain allow us to characterize more than 300 continuous, kinetically unbiased RBD-opening pathways. Together with ManifoldEM analysis of cryo-electron microscopy data and biolayer interferometry experiments, we reveal a gating role for the N-glycan at position N343, which facilitates RBD opening. Residues D405, R408 and D427 also participate. The atomic-level characterization of the glycosylated spike activation mechanism provided herein represents a landmark study for ensemble pathway simulations and offers a foundation for understanding the fundamental mechanisms of SARS-CoV-2 viral entry and infection.

A glycan gate controls opening of the SARS-CoV-2 spike protein.

SARS-CoV-2 infection is controlled by the opening of the spike protein receptor binding domain (RBD), which transitions from a glycan-shielded "down" to an exposed "up" state in order to bind the human ACE2 receptor and infect cells. While snapshots of the "up" and "down" states have been obtained by cryoEM and cryoET, details of the RBD opening transition evade experimental characterization. Here, over 130 µs of weighted ensemble (WE) simulations of the fully glycosylated spike ectodomain allow us to characterize more than 300 continuous, kinetically unbiased RBD opening pathways. Together with ManifoldEM analysis of cryo-EM data and biolayer interferometry experiments, we reveal a gating role for the N-glycan at position N343, which facilitates RBD opening. Residues D405, R408, and D427 also participate. The atomic-level characterization of the glycosylated spike activation mechanism provided herein achieves a new high-water mark for ensemble pathway simulations and offers a foundation for understanding the fundamental mechanisms of SARS-CoV-2 viral entry and infection.

Characterization and engineering of a two-enzyme system for plastics depolymerization.

Plastics pollution represents a global environmental crisis. In response, microbes are evolving the capacity to utilize synthetic polymers as carbon and energy sources. Recently, Ideonella sakaiensis was reported to secrete a two-enzyme system to deconstruct polyethylene terephthalate (PET) to its constituent monomers. Specifically, the I. sakaiensis PETase depolymerizes PET, liberating soluble products, including mono(2-hydroxyethyl) terephthalate (MHET), which is cleaved to terephthalic acid and ethylene glycol by MHETase. Here, we report a 1.6 Å resolution MHETase structure, illustrating that the MHETase core domain is similar to PETase, capped by a lid domain. Simulations of the catalytic itinerary predict that MHETase follows the canonical two-step serine hydrolase mechanism. Bioinformatics analysis suggests that MHETase evolved from ferulic acid esterases, and two homologous enzymes are shown to exhibit MHET turnover. Analysis of the two homologous enzymes and the MHETase S131G mutant demonstrates the importance of this residue for accommodation of MHET in the active site. We also demonstrate that the MHETase lid is crucial for hydrolysis of MHET and, furthermore, that MHETase does not turnover mono(2-hydroxyethyl)-furanoate or mono(2-hydroxyethyl)-isophthalate. A highly synergistic relationship between PETase and MHETase was observed for the conversion of amorphous PET film to monomers across all nonzero MHETase concentrations tested. Finally, we compare the performance of MHETase:PETase chimeric proteins of varying linker lengths, which all exhibit improved PET and MHET turnover relative to the free enzymes. Together, these results offer insights into the two-enzyme PET depolymerization system and will inform future efforts in the biological deconstruction and upcycling of mixed plastics.

Mechanism of proton transfer in class A ß-lactamase catalysis and inhibition by avibactam.

Gram-negative bacteria expressing class A ß-lactamases pose a serious health threat due to their ability to inactivate all ß-lactam antibiotics. The acyl-enzyme intermediate is a central milestone in the hydrolysis reaction catalyzed by these enzymes. However, the protonation states of the catalytic residues in this complex have never been fully analyzed experimentally due to inherent difficulties. To help unravel the ambiguity surrounding class A ß-lactamase catalysis, we have used ultrahigh-resolution X-ray crystallography and the recently approved ß-lactamase inhibitor avibactam to trap the acyl-enzyme complex of class A ß-lactamase CTX-M-14 at varying pHs. A 0.83-Å-resolution CTX-M-14 complex structure at pH 7.9 revealed a neutral state for both Lys73 and Glu166. Furthermore, the avibactam hydroxylamine-O-sulfonate group conformation varied according to pH, and this conformational switch appeared to correspond to a change in the Lys73 protonation state at low pH. In conjunction with computational analyses, our structures suggest that Lys73 has a perturbed acid dissociation constant (pKa) compared with acyl-enzyme complexes with ß-lactams, hindering its function to deprotonate Glu166 and the initiation of the deacylation reaction. Further NMR analysis demonstrated Lys73 pKa to be ∼5.2 to 5.6. Together with previous ultrahigh-resolution crystal structures, these findings enable us to follow the proton transfer process of the entire acylation reaction and reveal the critical role of Lys73. They also shed light on the stability and reversibility of the avibactam carbamoyl acyl-enzyme complex, highlighting the effect of substrate functional groups in influencing the protonation states of catalytic residues and subsequently the progression of the reaction.

Modeling Boronic Acid Based Fluorescent Saccharide Sensors: Computational Investigation of d-Fructose Binding to Dimethylaminomethylphenylboronic Acid.

Designing organic saccharide sensors for use in aqueous solution is a nontrivial endeavor. Incorporation of hydrogen bonding groups on a sensor's receptor unit to target saccharides is an obvious strategy but not one that is likely to ensure analyte-receptor interactions over analyte-solvent or receptor-solvent interactions. Phenylboronic acids are known to reversibly and covalently bind saccharides (diols in general) with highly selective affinity in aqueous solution. Therefore, recent work has sought to design such sensors and understand their mechanism for allowing fluorescence with bound saccharides. In past work, binding orientations of several saccharides were determined to dimethylaminomethylphenylboronic acid (DMPBA) receptors with an anthracene fluorophore; however, the binding orientation of d-fructose to such a sensor could not be determined. In this work, we investigate the potential binding modes by generating 20 possible bidentate and six possible tridentate modes between fructose and DMPBA, a simplified receptor model. Gas phase and implicit solvent geometry optimizations, with a myriad functional/basis set pairs, were carried out to identify the lowest energy bidentate and tridentate binding modes of d-fructose to DMPBA. An interesting hydrogen transfer was observed during selected bidentate gas phase optimizations; this transfer suggests a strong sharing of the hydrogen atom between the boronate hydroxyl and amine nitrogen.

The Good, the Bad, and the Ugly: "HiPen", a New Dataset for Validating (S)QM/MM Free Energy Simulations.

Indirect (S)QM/MM free energy simulations (FES) are vital to efficiently incorporating sufficient sampling and accurate (QM) energetic evaluations when estimating free energies of practical/experimental interest. Connecting between levels of theory, i.e., calculating Δ A l o w → h i g h , remains to be the most challenging step within an indirect FES protocol. To improve calculations of Δ A l o w → h i g h , we must: (1) compare the performance of all FES methods currently available; and (2) compile and maintain datasets of Δ A l o w → h i g h calculated for a wide-variety of molecules so that future practitioners may replicate or improve upon the current state-of-the-art. Towards these two aims, we introduce a new dataset, "HiPen", which tabulates Δ A g a s M M → 3 o b (the free energy associated with switching from an M M to an S C C - D F T B molecular description using the 3ob parameter set in gas phase), calculated for 22 drug-like small molecules. We compare the calculation of this value using free energy perturbation, Bennett's acceptance ratio, Jarzynski's equation, and Crooks' equation. We also predict the reliability of each calculated Δ A g a s M M → 3 o b by evaluating several convergence criteria including sample size hysteresis, overlap statistics, and bias metric ( Π ). Within the total dataset, three distinct categories of molecules emerge: the "good" molecules, for which we can obtain converged Δ A g a s M M → 3 o b using Jarzynski's equation; "bad" molecules which require Crooks' equation to obtain a converged Δ A g a s M M → 3 o b ; and "ugly" molecules for which we cannot obtain reliably converged Δ A g a s M M → 3 o b with either Jarzynski's or Crooks' equations. We discuss, in depth, results from several example molecules in each of these categories and describe how dihedral discrepancies between levels of theory cause convergence failures even for these gas phase free energy simulations.

Characterization and engineering of a plastic-degrading aromatic polyesterase.

Poly(ethylene terephthalate) (PET) is one of the most abundantly produced synthetic polymers and is accumulating in the environment at a staggering rate as discarded packaging and textiles. The properties that make PET so useful also endow it with an alarming resistance to biodegradation, likely lasting centuries in the environment. Our collective reliance on PET and other plastics means that this buildup will continue unless solutions are found. Recently, a newly discovered bacterium, Ideonella sakaiensis 201-F6, was shown to exhibit the rare ability to grow on PET as a major carbon and energy source. Central to its PET biodegradation capability is a secreted PETase (PET-digesting enzyme). Here, we present a 0.92 Å resolution X-ray crystal structure of PETase, which reveals features common to both cutinases and lipases. PETase retains the ancestral α/ß-hydrolase fold but exhibits a more open active-site cleft than homologous cutinases. By narrowing the binding cleft via mutation of two active-site residues to conserved amino acids in cutinases, we surprisingly observe improved PET degradation, suggesting that PETase is not fully optimized for crystalline PET degradation, despite presumably evolving in a PET-rich environment. Additionally, we show that PETase degrades another semiaromatic polyester, polyethylene-2,5-furandicarboxylate (PEF), which is an emerging, bioderived PET replacement with improved barrier properties. In contrast, PETase does not degrade aliphatic polyesters, suggesting that it is generally an aromatic polyesterase. These findings suggest that additional protein engineering to increase PETase performance is realistic and highlight the need for further developments of structure/activity relationships for biodegradation of synthetic polyesters.

Computing converged free energy differences between levels of theory via nonequilibrium work methods: Challenges and opportunities.

We demonstrate that Jarzynski's equation can be used to reliably compute free energy differences between low and high level representations of systems. The need for such a calculation arises when employing the so-called "indirect" approach to free energy simulations with mixed quantum mechanical/molecular mechanical (QM/MM) Hamiltonians; a popular technique for circumventing extensive simulations involving quantum chemical computations. We have applied this methodology to several small and medium sized organic molecules, both in the gas phase and explicit solvent. Test cases include several systems for which the standard approach; that is, free energy perturbation between low and high level description, fails to converge. Finally, we identify three major areas in which the difference between low and high level representations make the calculation of ΔAlow→high difficult: bond stretching and angle bending, different preferred conformations, and the response of the MM region to the charge distribution of the QM region. © 2016 Wiley Periodicals, Inc.