Atomistic Molecular Dynamics Simulations Research Articles

Exploring the Impact of Physiological C-Terminal Truncation on α-Synuclein Conformations to Unveil Mechanisms Regulating Pathological Aggregation.

Emerging evidence suggests that physiological C-terminal truncation of α-synuclein (αS) plays a critical role in regulating liquid-liquid phase separation and promoting amyloid aggregation, processes implicated in neurodegenerative diseases such as Parkinson's disease (PD). However, the molecular mechanisms through which C-terminal truncation influences αS conformation and modulates its aggregation remain poorly understood. In this study, we investigated the impact of C-terminal truncation on αS conformational dynamics by comparing full-length αS1-140 with truncated αS1-103 monomers using atomistic discrete molecular dynamics simulations. Our findings revealed that both αS1-140 and αS1-103 primarily adopted helical conformations around residues 7-32, while residues 36-95, located in the second half of the N-terminal and NAC domains, predominantly formed a dynamic β-sheet core. The C-terminus of αS1-140 was largely unstructured and dynamically wrapped around the β-sheet core. While residues 1-95 exhibited similar secondary structure propensities in both αS1-140 and αS1-103, the dynamic capping by the C-terminus in αS1-140 slightly enhanced β-sheet formation around residues 36-95. In contrast, key aggregation-driving regions (residues 2-9, 36-42, 45-57, and 68-78) were dynamically shielded by the C-terminus in αS1-140, reducing their exposure and potentially preventing interpeptide interactions that drive aggregation. C-terminal truncation, on the other hand, increased the exposed surface area of these aggregation-prone regions, thereby enhancing interpeptide interactions, phase separation, and amyloid aggregation. Overall, our simulations provide valuable insights into the conformational effects of C-terminal truncation on αS and its role in promoting pathological aggregation.

Heterogeneous Slowdown of Dynamics in the Condensate of an Intrinsically Disordered Protein.

The high concentration of proteins and other biological macromolecules inside biomolecular condensates leads to dense and confined environments, which can affect the dynamic ensembles and the time scales of the conformational transitions. Here, we use atomistic molecular dynamics (MD) simulations of the intrinsically disordered low complexity domain (LCD) of the human fused in sarcoma (FUS) RNA-binding protein to study how self-crowding inside a condensate affects the dynamic motions of the protein. We found a heterogeneous retardation of the protein dynamics in the condensate with respect to the dilute phase, with large-amplitude motions being strongly slowed by up to 2 orders of magnitude, whereas small-scale motions, such as local backbone fluctuations and side-chain rotations, are less affected. The results support the notion of a liquid-like character of the condensates and show that different protein motions respond differently to the environment.

Polar order in a fluid like ferroelectric with a tilted lamellar structure – observation of a polar smectic C (SmCP) phase

The discovery of fluid states of matter with spontaneous bulk polar order is appreciated as a major discovery in the fields of soft matter and liquid crystals. Typically, this manifests as polar order superimposed atop conventional phase structures and is thus far limited to orthogonal phase types. Here we report a family of materials which exhibit a previously unseen state of matter which we conclude is a polar smectic C phase, and so we term it SmCP. The spontaneous polarisation of the SmCP phase is over two orders of magnitude larger than that found in conventional ferroelectric SmC phase of chiral materials used in some LCD devices. Fully atomistic molecular dynamics simulations faithfully and spontaneously reproduce the proposed structure and associated bulk properties; comparison of experimental and simulated X‐ray scattering patterns shows excellent agreement. The materials disclosed here have significantly smaller dipole moments than typical polar liquid crystals such as RM734 which suggests the role of molecular electrical polarity in generating polar order is perhaps overstated, a view supported by consideration of other molecular systems.

Polar order in a fluid like ferroelectric with a tilted lamellar structure - observation of a polar smectic C (SmCP) phase.

The discovery of fluid states of matter with spontaneous bulk polar order is appreciated as a major discovery in the fields of soft matter and liquid crystals. Typically, this manifests as polar order superimposed atop conventional phase structures and is thus far limited to orthogonal phase types. Here we report a family of materials which exhibit a previously unseen state of matter which we conclude is a polar smectic C phase, and so we term it SmCP. The spontaneous polarisation of the SmCP phase is over two orders of magnitude larger than that found in conventional ferroelectric SmC phase of chiral materials used in some LCD devices. Fully atomistic molecular dynamics simulations faithfully and spontaneously reproduce the proposed structure and associated bulk properties; comparison of experimental and simulated X-ray scattering patterns shows excellent agreement. The materials disclosed here have significantly smaller dipole moments than typical polar liquid crystals such as RM734 which suggests the role of molecular electrical polarity in generating polar order is perhaps overstated, a view supported by consideration of other molecular systems.

Atomistic Simulations of Hydration and Antibiofouling Behavior of Amphiphilic Polymer Brush Surfaces Functionalized with TMAO and Short Fluorocarbon.

Developing fouling-resistant materials is of paramount interest in marine industries and biomedical applications. In this work, we studied the interfacial hydration and surface-protein interactions of the amphiphilic brush surface functionalized with hybrid hydrophilic trimethylamine N-oxide (TMAO) and hydrophobic pentafluoroethyl groups using a combination of atomistic molecular dynamics simulations and free-energy computations. Our results show that while the interfacial hydration density of the amphiphilic surface slightly decreases with the introduction of small fluorocarbons compared to that of the pure TMAO-functionalized surface, the amphiphilic surface remains relatively strong in resisting protein adsorption. The nanosized clustering of hydrophobic fluorine atoms on the top of the amphiphilic brush surface introduces weak protein adsorption; however, due to the strong interfacial hydration and weak hydrophobic interaction, the amphiphilic surface exhibits sufficient antibiofouling activities. Our fundamental studies will be critical for the discovery of marine fouling-resistant coating surfaces.

Experimental and simulation study of reverse micelles formed by aerosol-OT and water in non-polar solvents.

The formation of reverse micelles by aerosol-OT [sodium bis(2-ethylhexyl) sulfosuccinate] in hydrocarbon solvents, and in the presence of water, is studied using a combination of atomistic molecular-dynamics simulations and small-angle neutron scattering (SANS). There have been many previous studies of aerosol-OT and its self-assembly in both water and non-aqueous solvents, but this work is focused on a combined experimental and simulation study of reverse-micelle formation. The effects of hydration (with water-to-surfactant molar ratios in the range 0-60) and solvent (cyclohexane and n-dodecane) are investigated. A force field is adapted that results in spontaneous formation of reverse micelles starting from completely randomized configurations. The computed dimensions of the reverse micelles compare very favourably with those determined in SANS experiments, providing validation of the simulation model. The kinetics of reverse-micelle formation are studied with a 50-ns, 1.7-million-atom system which contains, in the steady state, about 50 reverse micelles. The internal structures of reverse micelles are characterized with mass density profiles, and the effects of solvent, and the structural crossover from highly structured water to 'bulk' water in the core, are detailed. The corresponding changes in the molecular reorientation times of sequestered water are also determined. Overall, the combination of experiment and simulation gives a detailed picture of reverse-micelle self-assembly and structure.

Progress in Multiscale Modeling of Silk Materials.

As a result of their hierarchical structure and biological processing, silk fibers rank among nature's most remarkable materials. The biocompatibility of silk-based materials and the exceptional mechanical properties of certain fibers has inspired the use of silk in numerous technical and medical applications. In recent years, computational modeling has clarified the relationship between the molecular architecture and emergent properties of silk fibers and has demonstrated predictive power in studies on novel biomaterials. Here, we review advances in modeling the structure and properties of natural and synthetic silk-based materials, from early structural studies of silkworm cocoon fibers to cutting-edge atomistic simulations of spider silk nanofibrils and the recent use of machine learning models. We explore applications of modeling across length scales: from quantum mechanical studies on model peptides, to atomistic and coarse-grained molecular dynamics simulations of silk proteins, to finite element analysis of spider webs. As computational power and algorithmic efficiency continue to advance, we expect multiscale modeling to become an indispensable tool for understanding nature's most impressive fibers and developing bioinspired functional materials.

Accelerated Prediction of Phase Behavior for Block Copolymer Libraries Using a Molecularly Informed Field Theory.

Solution formulations involving polymers are the basis for a wide range of products spanning consumer care, therapeutics, lubricants, adhesives, and coatings. These multicomponent systems typically show rich self-assembly and phase behavior that are sensitive to even small changes in chemistry and composition. Longstanding computational efforts have sought techniques for predictive modeling of formulation structure and thermodynamics without experimental guidance, but the challenges of addressing the long time scales and large length scales of self-assembly while maintaining chemical specificity have thwarted the emergence of general approaches. As a consequence, current formulation design remains largely Edisonian. Here, we present a multiscale modeling approach that accurately predicts, without any experimental input, the complete temperature-concentration phase diagram of model diblock polymers in solution, as established postprediction through small-angle X-ray scattering. The methodology employs a strategy whereby atomistic molecular dynamics simulations is used to parametrize coarse-grained field-theoretic models; simulations of the latter then easily surmount long equilibration time scales and enable rigorous determination of solution structures and phase behavior. This systematic and predictive approach, accelerated by access to well-defined block copolymers, has the potential to expedite in silico screening of novel formulations to significantly reduce trial-and-error experimental design and to guide selection of components and compositions across a vast range of applications.

Comparative Bindings of Glycosaminoglycans with CXCL8 Monomer and Dimer: Insights from Conformational Dynamics and Kinetics of Hydrogen Bonds.

GAGs bind to both the monomeric and dimeric forms of CXCL8, helping to form a concentration gradient of the chemokine that facilitates the recruitment of neutrophils to an injury site and supports other biological functions. In this study, atomistic molecular dynamics simulations were conducted to investigate the binding behavior of two hexameric GAGs sulfated at two different positions, chondroitin sulfate (CS) and heparan sulfate (HS), with the monomer (SIL8) and dimer (DIL8) forms of the CXCL8 protein. The results support that the conformational diversity of CS and HS appeared to be more when binding with monomer SIL8 than dimer DIL8. CS gained more configurational entropy from glycosidic linkage flexibility when bound to SIL8 than DIL8, with a higher energy barrier, whereas HS exhibited a lower energy barrier for configurational entropy when bound to SIL8 and DIL8. The monomer SIL8 exhibited more favorable and preferential binding with GAGs compared to DIL8. Formation of hydrogen bonds with the basic amino acids of SIL8 and GAG was more rigid and required higher activation energy to break than the other identified hydrogen bondings. Water molecules involved in hydrogen bonding with GAGs, excluding those with basic amino acids of DIL8, showed longer lifetimes and slower relaxation compared to SIL8. This suggests that water-mediated interactions also favor binding of DIL8 with GAGs. Despite having more basic amino acids, DIL8 did not display stronger binding than SIL8, indicating the significant role of basic residues in stabilizing the GAG-protein interactions in the monomers. The reason could be that the greater number of nonbasic amino acids in dimeric CXCL8 stabilizes the complex by forming water-mediated hydrogen bonds, reducing the conformational preferences for binding with GAGs. In contrast, the monomeric form of CXCL8 exhibits a higher conformational preference for protein-GAG interactions.

Assessing the effect of deep eutectic solvents on α-chymotrypsin thermal stability and activity.

Optimizing the liquid reaction phase holds significant potential for enhancing the efficiency of biocatalytic pro- cesses since it determines reaction equilibrium and kinetics. This study investigates the influence of the addition of deep eutectic solvents on the stability and activity of α-chymotrypsin, a proteolytic enzyme with industrial rele- vance. Deep eutectic solvents, composed of choline chloride or betaine mixed with glycerol or sorbitol, were added in the reaction phase at various concentrations. Experimental techniques, including kinetic and fluorometry, were employed to assess the α-chymotrypsin activity, thermal stability, and unfolding reversibility. Atomistic molecular dynamics simulations were also conducted to assess the interactions and provide molecular-level insights between α-chymotrypsin and the solvent. The results showed that among all studied mixtures, adding choline chloride + sorbitol improved thermal stability up to 18 °C and reaction kinetic efficiency up to two-fold upon adding choline chloride + glycerol. Notably, the choline chloride + sorbitol system exhibited the most substantial stabilization effect, attributed to the surface preferential accumulation of sorbitol, as corroborated by the computational anal- yses. This work highlights the potential of tailoring liquid reaction phase of α-chymotrypsin catalyzed reaction using neoteric solventsto enhance α-chymotrypsin performance and stability in industrial applications.

Nanoscale friction and wear of graphite surface in ambient and underwater conditions

The tribological properties of graphite are extremely sensitive to surrounding environment, which affects its longevity and friction reduction performance. This study aims to develop a comprehensive understanding of the effects of a liquid environment on interactions at the sliding interface, thereby influencing the nanoscale tribological properties of graphite surfaces. By combining atomic force microscopy experiments and molecular dynamics simulations, we conducted a comparative analysis of friction and wear behavior of graphite under two different environmental conditions: ambient (air) and underwater condition. Our investigations explored both the step edge and interior (step-free) graphite surfaces. The experimental results revealed a notable contrast in the frictional and wear behavior of graphite at nanoscale in these two environments. The interior surface exhibited a friction coefficient (COF) of approximately 0.003 and 0.006 against a diamond-coated surface in ambient and underwater conditions, respectively. Interestingly, the underwater environment not only increased friction but also significantly compromised the wear resistance of graphene layers near the graphite surface compared to the ambient environment, as evidenced in both step edge and interior step-free regions. The interior region sustained up to ∼7400 nN load in ambient condition but failed at ∼1500 nN under water. Similarly, the step edge failed at ∼375 and ∼187.5 nN in ambient and underwater conditions, respectively. Our simulations revealed that the increased friction in underwater condition is due to resistance of surrounding water molecules during tip sliding. The presence of water at tip-graphite contact interface generated substantial localized stress, leading to the initiation of wear and revealing the pronounced effect of water on the wear characteristics of graphite in underwater condition.

A simplified method for theoretical sum frequency generation spectroscopy calculation and interpretation: The "pop model".

Existing methods to compute theoretical spectra are restricted to the use of time-correlation functions evaluated from accurate atomistic molecular dynamics simulations, often at the abinitio level. The molecular interpretation of the computed spectra requires additional steps to deconvolve the spectroscopic contributions from local water and surface structural populations at the interface. The lack of a standard procedure to do this often hampers rationalization. To overcome these challenges, we rewrite the equations for spectra calculation into a sum of partial contributions from interfacial populations, weighted by their abundance at the interface. We show that SFG signatures from each population can be parameterized into a minimum dataset of reference partial spectra. Accurate spectra can then be predicted by just evaluating the statistics of interfacial populations, which can be done even with force field simulations as well as with analytic models. This approach broadens the range of simulation techniques from which theoretical spectra can be calculated, opening toward non-atomistic and Monte Carlo simulation approaches. Most notably, it allows constructing accurate theoretical spectra for interfacial conditions that cannot even be simulated, as we demonstrate for the pH-dependent SFG spectra of silica/water interfaces.

Structure, orientation, and dynamics of per- and polyfluoroalkyl substance (PFAS) surfactants at the air-water interface: Molecular-level insights

HypothesisUnderstanding the intricate molecular-level details of toxic per- and polyfluoroalkyl substances (PFAS) partitioning to the air–water interface holds paramount importance in evaluating their fate and transport, as well as for finding safer alternatives for various applications, including aqueous film forming foams. The behavior of these substances at interfaces strongly depends on molecular architecture, chemistry, and concentration, which define molecular packing, self-assembly, interfacial diffusion, and the surface tension. SimulationsModeling of three PFAS surfactants, namely, longer-tail (perfluorooctanoate (PFOA)) and shorter-tail (perfluorobutanoate (PFBA) and 2,3,3,3-tetrafluoro-2-(heptafluoropropoxy) propanoate (GenX)) has been conducted using atomistic molecular dynamics simulations. A systematic comparison between these representative PFAS of different sizes and structure reveals factors influencing their association behavior, mechanism of surface tension reduction, and interfacial mobility as a function of surface coverage. FindingsShorter-chain PFAS surfactants (GenX or PFBA) require lower surface coverage compared to longer chain (PFOA) PFAS to achieve the same decrease in surface tension. However, a higher concentration of GenX and PFBA is necessary in the bulk aqueous solution to achieve the same surface coverage as PFOA, due to their higher solubility in water. The PFAS molecular orientation and mobility at the interface are found to be vastly influenced by the length and architecture of the hydrophobic fluorocarbon tail. A significant ordering of the water dipole moment near the anionic headgroup is apparent at high surface concentration. A direct correlation is established between the PFAS interfacial properties and PFAS-PFAS, PFAS-counterion, and PFAS-water interactions.

Simulations reveal unique roles for the FXR hinge in the FXR-RXR nuclear receptor heterodimer.

Nuclear receptors are multidomain transcription factors whose full-length quaternary architecture is poorly described and understood. Most nuclear receptors bind DNA as heterodimers or homodimers, which could encompass a variety of arrangements of the individual domains. Only a handful of experimental structures currently exist describing these architectures. Given that domain interactions and protein-DNA interactions within transcriptional complexes are tightly linked to function, understanding the arrangement of nuclear receptor domains on DNA is of utmost importance. Here, we employ modeling and molecular dynamics (MD) simulations to describe the structure of the full-length farnesoid X receptor (FXR) and retinoid X receptor alpha (RXR) heterodimer bound to DNA. Combining over 100 microseconds of atomistic MD simulations with enhanced sampling simulations, we characterize the dynamic behavior of eight FXR-RXR-DNA complexes, showing that these complexes support a range of quaternary architectures. Our simulations reveal critical roles for the hinge in the DNA-bound dimer in facilitating interdomain allostery, mediating DNA binding and driving the dynamic flexibility of the complex. These roles have been hard to appreciate previously due to experimental limitations in studying the flexible hinge. These studies provide a much-needed framework that will enable the field to obtain a complete understanding of nuclear receptor quaternary architectures.

Deciphering the Morphological Difference of Amyloid-β Fibrils in Familial and Sporadic Alzheimer's Diseases.

The aggregation of amyloid-β (Aβ) into amyloid fibrils is the major pathological hallmark of Alzheimer's disease (AD). Aβ fibrils can adopt a variety of morphologies, the relative populations of which are recently found to be associated with different AD subtypes such as familial and sporadic AD (fAD and sAD, respectively). The two AD subtypes differ in their ages of onset, AD-related genetic predispositions, and dominant Aβ fibril morphologies. We postulate that these disease subtype-dependent fibril morphology differences can be attributed to the intrinsic fibril properties and interacting molecules in the environment. Using atomistic discrete molecular dynamics simulations, we demonstrated that the fAD-dominant morphology exhibited a lower free-energy barrier for fibril growth but also a lower stability compared with the sAD-dominant fibril morphology, resulting in the time-dependent population change consistent with experimental observations. Additionally, we studied the effect of the Bri2 BRICHOS domain, an endogenous protein that has been reported to inhibit Aβ aggregation by preferential binding to fibrils, as one of the possible environmental factors. The Bri2 BRICHOS domain showed stronger binding to the fAD-dominant fibril than the sAD-dominant fibril in silico, suggesting a more effective suppression of fAD-dominant fibril formation. This result explains the high population of the sAD-dominant fibril morphology in sporadic cases with normal Bri2 functions. Genetic predisposition in fAD, on the other hand, might impair or overwhelm Bri2 functions, leading to a high population of fAD-associated fibril morphology. Together, our computational findings provide a theoretical framework for elucidating the AD subtypes entailed by distinct dominant amyloid fibril morphologies.

Sequence-sensitivity in functional synthetic polymer properties.

Recently, a new class of synthetic methyl methacrylate-based random heteropolymers (MMA-based RHPs) has displayed protein-like properties. Their function appears to be insensitive to the precise sequence. Here, through atomistic molecular dynamics simulation, we show that there are universal protein-like features of MMA-based RHPs that are insensitive to the sequence, and mostly depend on the overall composition. In particular, we find that MMA-based RHPs ``fold'' into globules with heterogeneous hydration patterns. However, the insensitivity to sequence identity observed in MMA-based RHPs dramatically changes when we substitute the backbone architecture with acrylate or replace the oxygen atom in the side chain with a nitrogen atom (methacrylamide or acrylamide). Using principal component analysis and an intersection-over-union based index, we demonstrate that different sequences may not overlap in the property space, meaning that their properties are controlled by the sequence rather than fixed composition. We further investigate the sequence-insensitive capability of the MMA-based RHPs as previously reported on bacterial phospholipase OmpLA stabilization through heterodimerization. As experimentally observed, such polymers enhance the stability of OmpLA as reliably as its native bilayer environment. The design of such MMA-based RHPs provides a sequence-insensitive alternative to protein-mimetic biomaterials that is orthogonal to the sequence-structure-function paradigm of proteins.

Sequence‐sensitivity in functional synthetic polymer properties

Recently, a new class of synthetic methyl methacrylate‐based random heteropolymers (MMA‐based RHPs) has displayed protein‐like properties. Their function appears to be insensitive to the precise sequence. Here, through atomistic molecular dynamics simulation, we show that there are universal protein‐like features of MMA‐based RHPs that are insensitive to the sequence, and mostly depend on the overall composition. In particular, we find that MMA‐based RHPs ``fold'' into globules with heterogeneous hydration patterns. However, the insensitivity to sequence identity observed in MMA‐based RHPs dramatically changes when we substitute the backbone architecture with acrylate or replace the oxygen atom in the side chain with a nitrogen atom (methacrylamide or acrylamide). Using principal component analysis and an intersection‐over‐union based index, we demonstrate that different sequences may not overlap in the property space, meaning that their properties are controlled by the sequence rather than fixed composition. We further investigate the sequence‐insensitive capability of the MMA‐based RHPs as previously reported on bacterial phospholipase OmpLA stabilization through heterodimerization. As experimentally observed, such polymers enhance the stability of OmpLA as reliably as its native bilayer environment. The design of such MMA‐based RHPs provides a sequence‐insensitive alternative to protein‐mimetic biomaterials that is orthogonal to the sequence‐structure‐function paradigm of proteins.

Comparative Assessment of Water Models in Protein-Glycan Interaction: Insights from Alchemical Free Energy Calculations and Molecular Dynamics Simulations.

Accurate computational simulations of protein-glycan dynamics are crucial for a comprehensive understanding of critical biological mechanisms, including host-pathogen interactions, immune system defenses, and intercellular communication. The accuracy of these simulations, including molecular dynamics (MD) simulation and alchemical free energy calculations, critically relies on the appropriate parameters, including the water model, because of the extensive hydrogen bonding with glycan hydroxyl groups. However, a systematic evaluation of water models' accuracy in simulating protein-glycan interaction at the molecular level is still lacking. In this study, we used full atomistic MD simulations and alchemical absolute binding free energy (ABFE) calculations to investigate the performance of five distinct water models in six protein-glycan complex systems. We evaluated water models' impact on structural dynamics and binding affinity through over 5.8 μs of simulation time per system. Our results reveal that most protein-glycan complexes are stable in the overall structural dynamics regardless of the water model used, while some show obvious fluctuations with specific water models. More importantly, we discover that the stability of the binding motif's conformation is dependent on the water model chosen when its residues form weak hydrogen bonds with the glycan. The water model also influences the conformational stability of the glycan in its bound state according to density functional theory (DFT) calculations. Using alchemical ABFE calculations, we find that the OPC water model exhibits exceptional consistency with experimental binding affinity data, whereas commonly used models such as TIP3P are less accurate. The findings demonstrate how different water models affect protein-glycan interactions and the accuracy of binding affinity calculations, which is crucial in developing therapeutic strategies targeting these interactions.

Water vapor responsiveness of chitosan: An experimental and simulation analysis.

Stimuli-responsive polymers have gained significant research interest in recent years owing to their potential applications in diverse areas. Here, we present a study on the actuation characteristics of chitosan-based free-standing films that exhibit full reversibility and repeatability in response to water vapor exposure. The effect of pH of the water and the degree of cross-linking of the chitosan films on the actuation performance is studied. In the case of free-standing polymer film-based actuators, the primary driving force behind actuation is understood to be the differential strain induced by the gradient in volume changes across the thickness of the film. To understand it further, we conducted full atomistic molecular dynamics simulation studies to explore water absorption and adsorption into the chitosan matrix. Our simulations revealed an accumulation of water molecules in the surface layer that rapidly desorb when shielded from water vapor. Furthermore, estimates of the energy gain resulting from the adsorption of water on the surface suggest that it is adequate to drive the shape change of the actuator when subjected to asymmetric exposure to water vapor. This finding supports the fact that the adsorbed layer of water on the surface of the chitosan film plays a role in actuation.

One molecule to couple them all: Toward realistic numbers of molecules in multiscale molecular dynamics simulations of exciton-polaritons.

Collective strong coupling of many molecules to the confined light modes of an optical resonator can influence the photochemistry of these molecules, but the origin of this effect is not yet fully understood. To provide atomistic insights, several approaches have been developed based on quantum chemistry or molecular dynamics methods. However, most of these methods rely on coupling a few molecules (or sometimes only one) to a single cavity mode. To reach the strong coupling regime with such a small number of molecules, much larger vacuum field strengths are employed than in experiments. To keep the vacuum field realistic and avoid potential artefacts, the number of coupled molecules should be significantly increased instead, but that is not always possible due to restrictions on computational hardware and software. To overcome this barrier and model the dynamics of an arbitrarily large ensemble of molecules coupled to realistic cavity fields in atomistic molecular dynamics simulations, we propose to coarse-grain subsets of molecules into one or more effective supermolecules with an enhanced dipole moment and concerted dynamics. To verify the validity of the proposed multiscale model, we performed simulations in which we investigated how the number of molecules that are coupled to the cavity affects excited-state intra-molecular proton transfer, polariton relaxation, and exciton transport.