High-resolution mining of the SARS-CoV-2 main protease conformational space: supercomputer-driven unsupervised adaptive sampling

We provide an unsupervised adaptive sampling strategy capable of producing microseconds-timescale molecular dynamics (MD) simulations of large biosystems using many-body polarizable force fields (PFF). The global exploration problem is decomposed into a set of separate MD trajectories that can be restarted within a selective process to achieve sufficient phase-space sampling. Accurate statistical properties can be obtained through reweighting. Within this highly parallel setup, the Tinker--HP package can be powered by an arbitrary large number of GPUs on supercomputers, reducing exploration time from years to days. This approach is used to tackle the urgent modeling problem of the SARS--CoV--2 Main Protease (Mpro) producing more than 38 microseconds of all-atom simulations of its apo, ligand-free, dimer using the high-resolution AMOEBA PFF. A first 15.14 microseconds simulation (physiological pH) is compared to available non--PFF long-timescale simulation data. A detailed clustering analysis exhibits striking differences between FFs, AMOEBA showing a richer conformational space. Focusing on key structural markers related to the oxyanion hole stability, we observe an asymmetry between protomers. One of them appears less structured resembling the experimentally inactive monomer for which a 6 microseconds simulation was performed as a basis of comparison. Results highlight the plasticity of Mpro active site. The C--terminal end of its less structured protomer is shown to oscillate between several states, being able to interact with the other protomer, potentially modulating its activity. Active and distal sites volumes are found to be larger in the most active protomer within our AMOEBA simulations compared to non-PFFs as additional cryptic pockets are uncovered. A second 17 microseconds AMOEBA simulation is performed with protonated His172 residues mimicking lower pH. Data show the protonation impact on the destructuring of the oxyanion loop. We finally analyze the solvation patterns around key histidine residues. The confined AMOEBA polarizable water molecules are able to explore a wide range of dipole moments, going beyond bulk values, leading to a water molecule counts consistent with experiment. Results suggest that the use of PFFs could be critical in drug discovery to accurately model the complexity of the molecular interactions structuring Mpro

Download Full-text

An Efficient GaMD Multi-Level Enhanced Sampling Strategy: Application to Polarizable Force Fields Simulations of Large Biological Systems

10.33774/chemrxiv-2021-ggjfx-v4 ◽

2021 ◽

Author(s):

Fréderic Célerse ◽

Theo Jaffrelot-Inizan ◽

Louis Lagardère ◽

Olivier Adjoua ◽

Pierre Monmarché ◽

...

Keyword(s):

Molecular Dynamics ◽

Adaptive Sampling ◽

Force Fields ◽

Biological Systems ◽

Sampling Strategy ◽

Umbrella Sampling ◽

Potential Of Mean Force ◽

Enhanced Sampling ◽

Polarizable Force Fields ◽

Multi Level

We introduce a novel multi-level enhanced sampling strategy grounded on Gaussian accelerated Molecular Dynamics (GaMD). First, we propose a GaMD multi-GPUs-accelerated implementation within the Tinker-HP molecular dynamics package. We introduce the new "dual-water" mode and its use with the flexible AMOEBA polarizable force field.By adding harmonic boosts to the water stretching and bonding terms, it accelerates the solvent-solute interactions while enabling speedups thanks to the use of fast multiple--timestep integrators. To further reduce time-to-solution, we couple GaMD to Umbrella Sampling (US). The GaMD—US/dual--water approach is tested on the 1D Potential of Mean Force (PMF) of the solvated CD2--CD58 system (168000 atoms) allowing the AMOEBA PMF to converge within 1 kcal/mol of the experimental value. Finally, Adaptive Sampling (AS) is added enabling AS-GaMD capabilities but also the introduction of the new Adaptive Sampling--US--GaMD (ASUS-GaMD) scheme. The highly parallel ASUS--GaMD setup decreases time to convergence by respectively 10 and 20 times compared to GaMD-US and US. Overall, beside the acceleration of PMF computations, Tinker-HP now allows for the simultaneous use of Adaptive Sampling and GaMD-"dual water" enhanced sampling approaches increasing the applicability of polarizable force fields to large scale simulations of biological systems.

Download Full-text

High-Resolution Mining of SARS-CoV-2 Main Protease Conformational Space: Supercomputer-Driven Unsupervised Adaptive Sampling

10.26434/chemrxiv.13003166.v3 ◽

2020 ◽

Author(s):

Theo Jaffrelot Inizan ◽

Frédéric Célerse ◽

Olivier Adjoua ◽

Dina El Ahdab ◽

Luc-Henry Jolly ◽

...

Keyword(s):

High Resolution ◽

Adaptive Sampling ◽

Sampling Strategy ◽

The Other ◽

Conformational Space ◽

Processing Unit ◽

X Rays ◽

Main Protease ◽

The Stability ◽

Apo State

We provide a new unsupervised adaptive sampling strategy capable of producing microsecondtimescale molecular dynamics (MD) simulations using many-body polarizable force fields (PFF) on modern supercomputers. The global exploration problem is decomposed into a set of separate MD trajectories that can be restarted within an iterative/selective process to achieve sufficient phase-space sampling within large biosystems, while accurate statistical properties can be obtained through debiasing. With this pleasingly parallel setup, the Tinker-HP package can be powered by an arbitrary large number of GPUs (Graphics Processing Unit) cards available on pre-exascale supercomputers, reducing to days explorations that would have taken years. We applied the approach to the urgent problem of the modeling of the SARS–CoV–2 Main protease (Mpro) dimer. A 15.14 microsecond high-resolution all-atom simulation (AMOEBA PFF) of its apo state is provided and compared to other available long-timescale non-PFF data. Noticeable differences are found between clustering analysis of the simulations, the AMOEBA adaptive results exhibiting a richer conformational space. Overall, our high-resolution AMOEBA structural analysis captures key experimental observations concerning the stability of the oxyanion hole, a marker of activity through the stability of different stacking and salt bridge interactions. A dissymmetry is found between the enzyme protomers that exhibit different volumes. One of them appears fully inactive while the other is "activable", exhibiting some partial activity features. This activity evaluation can be further traced back to the large flexibility of the C terminal domain, fully captured by AMOEBA but not seen in X-rays due to insufficient electron densities related to the domain high mobility. The C–terminal region of the fully inactive protomer is shown to oscillate between several states, one of them interacting with the other protomer active site, therefore potentially modulating down its activity. Overall, these results reinforce the experimental hypothesis of a full inactivation of the apo state and clearly capture the asymmetric nature of protomers. Additional analysis show that the cavities volumes of the active and distal sites are found to be larger in the most active protomer with AMOEBA. To a larger extend, the PFF finds significantly larger cavities than those obtained with classical, non-polarizable simulations. The consequences on druggability are discussed as additional potential druggable cryptic pockets are found. All data produced within this research are fully accessible to the community for further analysis.

Download Full-text

High-Resolution Mining of SARS-CoV-2 Main Protease Conformational Space: Supercomputer-Driven Unsupervised Adaptive Sampling

10.26434/chemrxiv.13003166.v5 ◽

2020 ◽

Author(s):

Theo Jaffrelot Inizan ◽

Frédéric Célerse ◽

Olivier Adjoua ◽

Dina El Ahdab ◽

Luc-Henri Jolly ◽

...

Keyword(s):

High Resolution ◽

Adaptive Sampling ◽

Sampling Strategy ◽

The Other ◽

Conformational Space ◽

Processing Unit ◽

X Rays ◽

Main Protease ◽

The Stability ◽

Apo State

We provide a new unsupervised adaptive sampling strategy capable of producing microsecondtimescale molecular dynamics (MD) simulations using many-body polarizable force fields (PFF) on modern supercomputers. The global exploration problem is decomposed into a set of separate MD trajectories that can be restarted within an iterative/selective process to achieve sufficient phase-space sampling within large biosystems, while accurate statistical properties can be obtained through debiasing. With this pleasingly parallel setup, the Tinker-HP package can be powered by an arbitrary large number of GPUs (Graphics Processing Unit) cards available on pre-exascale supercomputers, reducing to days explorations that would have taken years. We applied the approach to the urgent problem of the modeling of the SARS–CoV–2 Main protease (Mpro) dimer. A 15.14 microsecond high-resolution all-atom simulation (AMOEBA PFF) of its apo state is provided and compared to other available long-timescale non-PFF data. Noticeable differences are found between clustering analysis of the simulations, the AMOEBA adaptive results exhibiting a richer conformational space. Overall, our high-resolution AMOEBA structural analysis captures key experimental observations concerning the stability of the oxyanion hole, a marker of activity through the stability of different stacking and salt bridge interactions. A dissymmetry is found between the enzyme protomers that exhibit different volumes. One of them appears fully inactive while the other is "activable", exhibiting some partial activity features. This activity evaluation can be further traced back to the large flexibility of the C terminal domain, fully captured by AMOEBA but not seen in X-rays due to insufficient electron densities related to the domain high mobility. The C–terminal region of the fully inactive protomer is shown to oscillate between several states, one of them interacting with the other protomer active site, therefore potentially modulating down its activity. Overall, these results reinforce the experimental hypothesis of a full inactivation of the apo state and clearly capture the asymmetric nature of protomers. Additional analysis show that the cavities volumes of the active and distal sites are found to be larger in the most active protomer with AMOEBA. To a larger extend, the PFF finds significantly larger cavities than those obtained with classical, non-polarizable simulations. The consequences on druggability are discussed as additional potential druggable cryptic pockets are found. All data produced within this research are fully accessible to the community for further analysis.

Download Full-text

High-Resolution Mining of SARS-CoV-2 Main Protease Conformational Space: Supercomputer-Driven Unsupervised Adaptive Sampling

10.26434/chemrxiv.13003166.v2 ◽

2020 ◽

Author(s):

Theo Jaffrelot Inizan ◽

Frédéric Célerse ◽

Olivier Adjoua ◽

Dina El Ahdab ◽

Luc-Henry Jolly ◽

...

Keyword(s):

High Resolution ◽

Adaptive Sampling ◽

Sampling Strategy ◽

The Other ◽

Conformational Space ◽

Processing Unit ◽

X Rays ◽

Main Protease ◽

The Stability ◽

Apo State

We provide a new unsupervised adaptive sampling strategy capable of producing microsecondtimescale molecular dynamics (MD) simulations using many-body polarizable force fields (PFF) on modern supercomputers. The global exploration problem is decomposed into a set of separate MD trajectories that can be restarted within an iterative/selective process to achieve sufficient phase-space sampling within large biosystems, while accurate statistical properties can be obtained through debiasing. With this pleasingly parallel setup, the Tinker-HP package can be powered by an arbitrary large number of GPUs (Graphics Processing Unit) cards available on pre-exascale supercomputers, reducing to days explorations that would have taken years. We applied the approach to the urgent problem of the modeling of the SARS–CoV–2 Main protease (Mpro) dimer. A 15.14 microsecond high-resolution all-atom simulation (AMOEBA PFF) of its apo state is provided and compared to other available long-timescale non-PFF data. Noticeable differences are found between clustering analysis of the simulations, the AMOEBA adaptive results exhibiting a richer conformational space. Overall, our high-resolution AMOEBA structural analysis captures key experimental observations concerning the stability of the oxyanion hole, a marker of activity through the stability of different stacking and salt bridge interactions. A dissymmetry is found between the enzyme protomers that exhibit different volumes. One of them appears fully inactive while the other is "activable", exhibiting some partial activity features. This activity evaluation can be further traced back to the large flexibility of the C terminal domain, fully captured by AMOEBA but not seen in X-rays due to insufficient electron densities related to the domain high mobility. The C–terminal region of the fully inactive protomer is shown to oscillate between several states, one of them interacting with the other protomer active site, therefore potentially modulating down its activity. Overall, these results reinforce the experimental hypothesis of a full inactivation of the apo state and clearly capture the asymmetric nature of protomers. Additional analysis show that the cavities volumes of the active and distal sites are found to be larger in the most active protomer with AMOEBA. To a larger extend, the PFF finds significantly larger cavities than those obtained with classical, non-polarizable simulations. The consequences on druggability are discussed as additional potential druggable cryptic pockets are found. All data produced within this research are fully accessible to the community for further analysis.

Download Full-text

High-Resolution Mining of SARS-CoV-2 Main Protease Conformational Space: Supercomputer-Driven Unsupervised Adaptive Sampling

10.26434/chemrxiv.13003166.v4 ◽

2020 ◽

Author(s):

Theo Jaffrelot Inizan ◽

Frédéric Célerse ◽

Olivier Adjoua ◽

Dina El Ahdab ◽

Luc-Henri Jolly ◽

...

Keyword(s):

High Resolution ◽

Adaptive Sampling ◽

Sampling Strategy ◽

The Other ◽

Conformational Space ◽

Processing Unit ◽

X Rays ◽

Main Protease ◽

The Stability ◽

Apo State

We provide a new unsupervised adaptive sampling strategy capable of producing microsecondtimescale molecular dynamics (MD) simulations using many-body polarizable force fields (PFF) on modern supercomputers. The global exploration problem is decomposed into a set of separate MD trajectories that can be restarted within an iterative/selective process to achieve sufficient phase-space sampling within large biosystems, while accurate statistical properties can be obtained through debiasing. With this pleasingly parallel setup, the Tinker-HP package can be powered by an arbitrary large number of GPUs (Graphics Processing Unit) cards available on pre-exascale supercomputers, reducing to days explorations that would have taken years. We applied the approach to the urgent problem of the modeling of the SARS–CoV–2 Main protease (Mpro) dimer. A 15.14 microsecond high-resolution all-atom simulation (AMOEBA PFF) of its apo state is provided and compared to other available long-timescale non-PFF data. Noticeable differences are found between clustering analysis of the simulations, the AMOEBA adaptive results exhibiting a richer conformational space. Overall, our high-resolution AMOEBA structural analysis captures key experimental observations concerning the stability of the oxyanion hole, a marker of activity through the stability of different stacking and salt bridge interactions. A dissymmetry is found between the enzyme protomers that exhibit different volumes. One of them appears fully inactive while the other is "activable", exhibiting some partial activity features. This activity evaluation can be further traced back to the large flexibility of the C terminal domain, fully captured by AMOEBA but not seen in X-rays due to insufficient electron densities related to the domain high mobility. The C–terminal region of the fully inactive protomer is shown to oscillate between several states, one of them interacting with the other protomer active site, therefore potentially modulating down its activity. Overall, these results reinforce the experimental hypothesis of a full inactivation of the apo state and clearly capture the asymmetric nature of protomers. Additional analysis show that the cavities volumes of the active and distal sites are found to be larger in the most active protomer with AMOEBA. To a larger extend, the PFF finds significantly larger cavities than those obtained with classical, non-polarizable simulations. The consequences on druggability are discussed as additional potential druggable cryptic pockets are found. All data produced within this research are fully accessible to the community for further analysis.

Download Full-text

High-Resolution Mining of SARS-CoV-2 Main Protease Conformational Space: Supercomputer-Driven Unsupervised Adaptive Sampling

10.26434/chemrxiv.13003166.v1 ◽

2020 ◽

Author(s):

Theo Jaffrelot Inizan ◽

Frédéric Célerse ◽

Olivier Adjoua ◽

Dina El Ahdab ◽

Luc-Henry Jolly ◽

...

Keyword(s):

High Resolution ◽

Adaptive Sampling ◽

Sampling Strategy ◽

The Other ◽

Conformational Space ◽

Processing Unit ◽

X Rays ◽

Main Protease ◽

The Stability ◽

Apo State

We provide a new unsupervised adaptive sampling strategy capable of producing microsecondtimescale molecular dynamics (MD) simulations using many-body polarizable force fields (PFF) on modern supercomputers. The global exploration problem is decomposed into a set of separate MD trajectories that can be restarted within an iterative/selective process to achieve sufficient phase-space sampling within large biosystems, while accurate statistical properties can be obtained through debiasing. With this pleasingly parallel setup, the Tinker-HP package can be powered by an arbitrary large number of GPUs (Graphics Processing Unit) cards available on pre-exascale supercomputers, reducing to days explorations that would have taken years. We applied the approach to the urgent problem of the modeling of the SARS–CoV–2 Main protease (Mpro) dimer. A 15.14 microsecond high-resolution all-atom simulation (AMOEBA PFF) of its apo state is provided and compared to other available long-timescale non-PFF data. Noticeable differences are found between clustering analysis of the simulations, the AMOEBA adaptive results exhibiting a richer conformational space. Overall, our high-resolution AMOEBA structural analysis captures key experimental observations concerning the stability of the oxyanion hole, a marker of activity through the stability of different stacking and salt bridge interactions. A dissymmetry is found between the enzyme protomers that exhibit different volumes. One of them appears fully inactive while the other is "activable", exhibiting some partial activity features. This activity evaluation can be further traced back to the large flexibility of the C terminal domain, fully captured by AMOEBA but not seen in X-rays due to insufficient electron densities related to the domain high mobility. The C–terminal region of the fully inactive protomer is shown to oscillate between several states, one of them interacting with the other protomer active site, therefore potentially modulating down its activity. Overall, these results reinforce the experimental hypothesis of a full inactivation of the apo state and clearly capture the asymmetric nature of protomers. Additional analysis show that the cavities volumes of the active and distal sites are found to be larger in the most active protomer with AMOEBA. To a larger extend, the PFF finds significantly larger cavities than those obtained with classical, non-polarizable simulations. The consequences on druggability are discussed as additional potential druggable cryptic pockets are found. All data produced within this research are fully accessible to the community for further analysis.

Download Full-text

High-Resolution Mining of SARS-CoV-2 Main Protease Conformational Space: Supercomputer-Driven Unsupervised Adaptive Sampling

10.26434/chemrxiv.13003166 ◽

2020 ◽

Author(s):

Theo Jaffrelot Inizan ◽

Frédéric Célerse ◽

Olivier Adjoua ◽

Dina El Ahdab ◽

Luc-Henri Jolly ◽

...

Keyword(s):

High Resolution ◽

Adaptive Sampling ◽

Sampling Strategy ◽

The Other ◽

Conformational Space ◽

Processing Unit ◽

X Rays ◽

Main Protease ◽

The Stability ◽

Apo State

We provide a new unsupervised adaptive sampling strategy capable of producing microsecondtimescale molecular dynamics (MD) simulations using many-body polarizable force fields (PFF) on modern supercomputers. The global exploration problem is decomposed into a set of separate MD trajectories that can be restarted within an iterative/selective process to achieve sufficient phase-space sampling within large biosystems, while accurate statistical properties can be obtained through debiasing. With this pleasingly parallel setup, the Tinker-HP package can be powered by an arbitrary large number of GPUs (Graphics Processing Unit) cards available on pre-exascale supercomputers, reducing to days explorations that would have taken years. We applied the approach to the urgent problem of the modeling of the SARS–CoV–2 Main protease (Mpro) dimer. A 15.14 microsecond high-resolution all-atom simulation (AMOEBA PFF) of its apo state is provided and compared to other available long-timescale non-PFF data. Noticeable differences are found between clustering analysis of the simulations, the AMOEBA adaptive results exhibiting a richer conformational space. Overall, our high-resolution AMOEBA structural analysis captures key experimental observations concerning the stability of the oxyanion hole, a marker of activity through the stability of different stacking and salt bridge interactions. A dissymmetry is found between the enzyme protomers that exhibit different volumes. One of them appears fully inactive while the other is "activable", exhibiting some partial activity features. This activity evaluation can be further traced back to the large flexibility of the C terminal domain, fully captured by AMOEBA but not seen in X-rays due to insufficient electron densities related to the domain high mobility. The C–terminal region of the fully inactive protomer is shown to oscillate between several states, one of them interacting with the other protomer active site, therefore potentially modulating down its activity. Overall, these results reinforce the experimental hypothesis of a full inactivation of the apo state and clearly capture the asymmetric nature of protomers. Additional analysis show that the cavities volumes of the active and distal sites are found to be larger in the most active protomer with AMOEBA. To a larger extend, the PFF finds significantly larger cavities than those obtained with classical, non-polarizable simulations. The consequences on druggability are discussed as additional potential druggable cryptic pockets are found. All data produced within this research are fully accessible to the community for further analysis.

Download Full-text

An Efficient GaMD Multi-Level Enhanced Sampling Strategy: Application to Polarizable Force Fields Simulations of Large Biological Systems

10.33774/chemrxiv-2021-ggjfx-v3 ◽

2021 ◽

Author(s):

Fréderic Célerse ◽

Theo Jaffrelot-Inizan ◽

Louis Lagardère ◽

Olivier Adjoua ◽

Pierre Monmarché ◽

...

Keyword(s):

Molecular Dynamics ◽

Adaptive Sampling ◽

Sampling Strategy ◽

Umbrella Sampling ◽

Potential Of Mean Force ◽

Enhanced Sampling ◽

Polarizable Force Field ◽

Accelerated Molecular Dynamics ◽

Polarizable Force Fields ◽

Multi Level

We detail a novel multi-level enhanced sampling strategy grounded on Gaussian accelerated Molecular Dynamics (GaMD). First, we propose a GaMD multi-GPUs-accelerated implementation within the Tinker-HP molecular dynamics package. We then introduce the new "dual-water" mode and its use with the flexible AMOEBA polarizable force field. By adding harmonic boosts to the water stretching and bonding terms, it accelerates the solvent-solute interactions while enabling speedups thanks to the use of fast multiple--timestep integrators. To further reduce time-to-solution, we couple GaMD to Umbrella Sampling (US). The GaMD—US/dual-water approach is tested on the 1D Potential of Mean Force (PMF) of the CD2-CD58 system (168000 atoms) allowing the AMOEBA PMF to converge within 1 kcal/mol of the experimental value. Finally, Adaptive Sampling (AS) is added enabling AS-GaMD capabilities but also the introduction of the new Adaptive Sampling--US--GaMD (ASUS--GaMD) scheme. The highly parallel ASUS--GaMD setup decreases time to convergence by respectively 10 and 20 compared to GaMD--US and US.

Download Full-text

Effects of Force Fields on the Conformational and Dynamic Properties of Amyloid β(1-40) Dimer Explored by Replica Exchange Molecular Dynamics Simulations

10.1101/210286 ◽

2017 ◽

Author(s):

Charles R. Watts ◽

Andrew Gregory ◽

Cole Frisbie ◽

Sándor Lovas

Keyword(s):

Molecular Dynamics ◽

Force Field ◽

Force Fields ◽

Dynamic Properties ◽

Amyloid Β ◽

Md Simulations ◽

Conformational Space ◽

Random Coil ◽

Replica Exchange ◽

Β Sheet

AbstractAlzheimer’s disease is histologically marked by fibrils of Amyloid beta (Aβ) peptide within the extracellular matrix. Fibrils themselves are benign compared to the cytotoxicity of the oligomers and pre-fibrillary aggregates. The conformational space and structural ensembles of Aβ peptides and their oligomers in solution are inherently disordered and proven to be challenging to study. Optimum force field selection for molecular dynamics (MD) simulations and the biophysical relevance of results are still unknown. We compared the conformational space of the Aβ(1–40) dimers by 300 ns replica exchange MD simulations at physiological temperature (310 K) using: the AMBER-ff99sb-ILDN, AMBER-ff99sb*-ILDN, AMBER-ff99sb-NMR, and CHARMM22* force fields. Statistical comparisons of simulation results to experimental data and previously published simulations utilizing the CHARMM22* and CHARMM36 force fields were performed. All force fields yield sampled ensembles of conformations with collision cross sectional areas for the dimer that are statistically significantly larger than experimental results. All force fields, with the exception of AMBER-ff99sb-ILDN (8.8±6.4%) and CHARMM36 (2.7±4.2%), tend to overestimate the α-helical content compared to experimental CD (5.3±5.2%). Using the AMBER-ff99sb-NMR force field resulted in the greatest degree of variance (41.3±12.9%). Except for the AMBER-ff99sb-NMR force field, the others tended to under estimate the expected amount of β-sheet and over estimate the amount of turn/bend/random coil conformations. All force fields, with the exception AMBER-ff99sb-NMR, reproduce a theoretically expected β-sheet-turn-β-sheet conformational motif, however, only the CHARMM22* and CHARMM36 force fields yield results compatible with collapse of the central and C-terminal hydrophobic cores from residues 17-21 and 30-36. Although analyses of essential subspace sampling showed only minor variations between force fields, secondary structures of lowest energy conformers are different.

Download Full-text