Replica Exchange Molecular Dynamics Simulations Research Articles

Atomistic Insights into the Inhibitory Mechanism of Tyrosine Phosphorylation against the Aggregation of Human Tau Fragment PHF6.

Abnormal aggregation of the microtubule-associated protein tau into intracellular fibrillary inclusions is characterized as the hallmark of tauopathies, including Alzheimer's disease and chronic traumatic encephalopathy. The hexapeptide 306VQIVYK311 (PHF6) of R3 plays an important role in the aggregation of tau. Recent experimental studies reported that phosphorylation of residue tyrosine 310 (Y310) could decrease the propensity of PHF6 to form fibrils and inhibit tau aggregation. However, the underlying inhibitory mechanism is not well understood. In this work, we systematically investigated the influences of phosphorylation on the conformational ensembles and oligomerization dynamics of PHF6 by performing extensive all-atom molecular dynamics (MD) simulations. Our replica exchange MD simulations demonstrate that Y310 phosphorylation could effectively suppress the formation of β-structure and shift PHF6 oligomers toward coil-rich aggregates. The interaction analyses show that hydrogen bonding and hydrophobic interactions among PHF6 peptides, as well as Y310-Y310 π-π stacking and I308-Y310 CH-π interactions, are weakened by phosphorylation. Additional microsecond MD simulations show that Y310 phosphorylation could inhibit the oligomerization of PHF6 by preventing the formation of large β-sheet oligomers and multi-layer β-sheet aggregates. This study provides mechanistic insights into the phosphorylation-inhibited tau aggregation, which may be helpful for the in-depth understanding of the pathogenesis of tauopathies.

Controlled destabilization of caged circularized DNA oligonucleotides predicted by replica exchange molecular dynamics simulations.

Spatiotemporal control is a critical issue in the design of strategies for the photoregulation of oligonucleotide activity. Efficient uncaging, i.e., activation by removal of photolabile protecting groups (PPGs), often necessitates multiple PPGs. An alternative approach is based on circularization strategies, exemplified by intrasequential circularization, also denoted photo-tethering, as introduced in [Seyfried et al., Angew. Chem., Int. Ed., 2017, 56, 359]. Here, we develop a computational protocol, relying on replica exchange molecular dynamics (REMD), in order to characterize the destabilization of a series of circularized, caged DNA oligonucleotides addressed in the aforementioned study. For these medium-sized (32 nt) oligonucleotides, melting temperatures are computed, whose trend is in good agreement with experiment, exhibiting a large destabilization and, hence, reduction of the melting temperature of the order of ΔTm ∼ 30 K as compared with the native species. The analysis of free energy landscapes confirms the destabilization pattern experienced by the circularized oligonucleotides. The present study underscores that computational protocols that capture controlled destabilization and uncaging of oligonucleotides are promising as predictive tools in the tailored photocontrol of nucleic acids.

Site-specific chirality-conferred structural compaction differentially mediates the cytotoxicity of Aβ42.

Growing evidence supports the confident association between distinct amyloid beta (Aβ) isoforms and Alzheimer's Disease (AD) pathogenesis. As such, critical investigations seeking to uncover the translational factors contributing to Aβ toxicity represent a venture of significant value. Herein, we comprehensively assess full-length Aβ42 stereochemistry, with a specific focus on models that consider naturally-occurring isomerization of Asp and Ser residues. We customize various forms of d-isomerized Aβ as natural mimics, ranging from fragments containing a single d residue to full length Aβ42 that includes multiple isomerized residues, systematically evaluating their cytotoxicity against a neuronal cell line. Combining multidimensional ion mobility-mass spectrometry experimental data with replica exchange molecular dynamics simulations, we confirm that co-d-epimerization at Asp and Ser residues within Aβ42 in both N-terminal and core regions effectively reduces its cytotoxicity. We provide evidence that this rescuing effect is associated with the differential and domain-specific compaction and remodeling of Aβ42 secondary structure.

A study of ab initio folding of chignolins using replica-exchange molecular dynamics simulations.

More and more studies have confirmed the importance of polarization effects in hydrogen bonding interactions in protein folding simulations. In this paper, a recently developed charge update scheme termed polarized structure-specific backbone charge (PSBC) model was applied to the folding of 10-residue chignolin. A comparison between simulations performed using PSBC and a nonpolarizable (AMBER99SB) force field demonstrably showed the importance of the electrostatic polarization effect in the folding of the short β-hairpin peptide by a series of analyses such as DSSP, free-energy landscape, hydrogen bond occupancy, and melting curve. The PSBC model was further validated by folding two other chignolin variants.

The effect of histidine behaviors on the structural properties of Aβ(1-42) peptide in protonation stage one, two, and three.

Histidine behaviors (including tautomeric and protonation behaviors, and integration on ε, δ, or p states) have been linked to protein folding and misfolding. However, the histidine behaviors of Aβ(1-42) are unconfirmed, which is the key point to understanding the pathogenesis of Alzheimer's disease. In the current study, 19 replica exchange molecular dynamics (REMD) simulations were performed to investigate the impact of histidine on the structural properties in protonation evolution stages one, two, and three. In contrast to the deprotonated (εεε) state, our current findings demonstrate that any protonated state will promote the formation of the β-sheet structure. The sheet-rich structures of (pεε), (δεp), (pεp), and (ppp) have the same basic characteristics of three-strand structures between the N-terminus, central hydrophobic core (CHC), and C-terminus. We found that the (pεε) (probability of 77.7%) and (δεp) (probability of 60.2%) prefer the abundant conformation over the other systems with higher regularity antiparallel β-sheet structure characteristics. Further H-bonding results indicate that H6 and H14 are more essential than H13. In addition, the Pearson correlation coefficient analysis revealed that the experimental result coincided with our simulated (δpε) system. The current study aids in the better understanding of the mechanisms of histidine behaviors, providing a fresh outlook on protein folding and misfolding.

Molecular dynamics simulations of the calmodulin-induced α-helix in the SK2 calcium-gated potassium ion channel

The family of small-conductance Ca2+-activated potassium ion channels (SK channels) is composed of four members (SK1, SK2, SK3, and SK4) involved in neuron-firing regulation. The gating of these channels depends on the intracellular Ca2+ concentration, and their sensitivity to this ion is provided by calmodulin (CaM). This protein binds to a specific region in SK channels known as the calmodulin-binding domain (CaMBD), an event which is essential for their gating. While CaMBDs are typically disordered in the absence of CaM, the SK2 channel subtype displays a small prefolded α-helical region in its CaMBD even if CaM is not present. This small helix is known to turn into a full α-helix upon CaM binding, although the molecular-level details for this conversion are not fully understood yet. In this work, we offer new insights on this physiologically relevant process by means of enhanced sampling, atomistic Hamiltonian replica exchange molecular dynamics simulations, providing a more detailed understanding of CaM binding to this target. Our results show that CaM is necessary for inducing a full α-helix along the SK2 CaMBD through hydrophobic interactions with V426 and L427. However, it is also necessary that W431 does not compete for these interactions; the role of the small prefolded α-helix in the SK2 CaMBD would be to stabilize W431 so that this is the case. In conclusion, our findings provide further insight into a key interaction between CaM and SK channels that is important for channel sensitivity to Ca2+.

Polyelectrolyte Influence on Beta-Hairpin Peptide Stability: A Simulation Study.

Assemblies of proteins and charged macromolecules (polyelectrolytes) find important applications as pharmaceutical formulations, biocatalysts, and cell-contacting substrates. A key question is how the polymer component influences the structure and function of the protein. The present paper addresses the influence of charged polymers on the thermal stability of two model beta-hairpin-forming peptides through an all-atom, replica exchange molecular dynamics simulation. The (negatively charged) peptides consist of the terminal 16 amino acids of the B1 domain of Protein G (GB1) and a variant with three of the GB1 residues substituted with tryptophan (Tryptophan Zipper 4, or TZ4). A (cationic) lysine polymer is seen to thermally stabilize TZ4 and destabilize GB1, while a (also cationic) chitosan polymer slightly stabilizes GB1 but has essentially no effect on TZ4. Free energy profiles reveal folded and unfolded conformations to be separated by kinetic barriers generally acting in the direction of the thermodynamically favored state. Through application of an Ising-like statistical mechanical model, a mechanism is proposed based on competition between (indirect) entropic stabilization of folded versus unfolded states and (direct) competition for hydrogen-bonding and hydrophobic interactions. These findings have important implications to the design of polyelectrolyte-based materials for biomedical and biotechnological applications.

Elucidating the Protonation State of the γ-Secretase Catalytic Dyad.

γ-Secretase (GS) is an intramembrane aspartyl protease that participates in the sequential cleavage of C99 to generate different isoforms of the amyloid-β (Aβ) peptides that are associated with the development of Alzheimer's disease. Due to its importance in the proteolytic processing of C99 by GS, we performed pH replica exchange molecular dynamics (pH-REMD) simulations of GS in its apo and substrate-bound forms to sample the protonation states of the catalytic dyad. We found that the catalytic dyad is deprotonated at physiological pH in our apo form, but the presence of the substrate at the active site displaces its monoprotonated state toward physiological pH. Our results show that Asp257 acts as the general base and Asp385 as the general acid during the cleavage mechanism. We identified different amino acids such as Lys265, Arg269, and the PAL motif interacting with the catalytic dyad and promoting changes in its acid-base behavior. Finally, we also found a significant pKa shift of Glu280 related to the internalization of TM6-CT in the GS-apo form. Our study provides critical mechanistic insight into the GS mechanism and the basis for future research on the genesis of Aβ peptides and the development of Alzheimer's disease.

Concentration Effect, Structural Properties, and Driving Force on Aβ28 Dimerization with and without Zn2+ Cooperation: Learning from Replica Exchange Sampling.

Zn2+ is a very important factor in promoting the formation of amyloid beta (Aβ) aggregates and amyloid plaques. The Zn2+ -bound Aβ species generate amorphous or low molecular-weight oligomers. However, it is a lack of studies to approach the starting structural features (dimerization) in Aβ nucleation processes with and without Zn2+ , which is the key point in understanding Zn2+ -induced nucleation mechanisms. To better understand the effect of concentration, structural properties, and the driving force, 14 independent replica exchange molecular dynamics simulations were performed in Aβ28 dimerization with and without Zn2+ (zAβ28 ) cooperation. Our scanning results show that the aggregation propensity is easier in Aβ28 -Aβ28 and Aβ28 -zAβ28 systems than zAβ28 -zAβ28 system. In binding property, the Aβ28 -Aβ28 model (-61.5 kcal mol-1 ) is stronger than zAβ28 -zAβ28 (-26.6 kcal mol-1 ) and Aβ28 -zAβ28 (-7.24 kcal mol-1 ) models. Further analysis confirmed that H13 and H14 residues play specific roles in the three systems. The key point is the orientation of N atom of the imidazole ring in histidine residues. Furthermore, we discovered different driving forces for each system. Our current study contributes to the understanding of how the Aβ28 dimer interacts with Zn2+ , which could lead to new insights into Zn2+ -induced nucleation mechanisms.

Insights into the Mixture of Aβ24 and Aβ42 Peptides from Atomistic Simulations

Amyloid-β (Aβ) oligomers play a central role in Alzheimer's disease (AD). Plaques of AD patients consist of Aβ40 and Aβ42 peptides and truncated Aβ peptides. The Aβ24 peptide, identified in human AD brains, was found to impair Aβ42 clearance through the brain-blood barrier. The Aβ24 peptide was also shown to reduce Aβ42 aggregation kinetics in pure buffer, but the underlying mechanism is unknown at atomistic level. In this study, we explored the conformational ensemble of the equimolar mixture of Aβ24 and Aβ42 by replica exchange molecular dynamics simulations and compared it to our previous results on the pure Aβ42 dimer. Our simulations demonstrate that the truncation at residue 24 changes the secondary, tertiary, and quaternary structures of the dimer, offering an explanation of the slower aggregation kinetics of the mixture.

Concerted enhanced-sampling simulations to elucidate the helix-fibril transition pathway of intrinsically disordered α-Synuclein

Fibril formation of α-synuclein is linked with Parkinson's disease. The intrinsically disordered nature of α-syn provides extensive conformational plasticity and becomes difficult to characterize its transition pathway from native monomeric to disease-associated fibril form. We implemented different simulation methods such as steered dynamics-umbrella sampling, and replica-exchange and conventional MD simulations to access various conformational states of α-syn. Nineteen distinct intermediate structures were identified by free energy landscape and cluster analysis. They were then sorted based on secondary structure and solvent exposure of fibril-core residues to illustrate the fibril dissociation pathway. The analysis showed that following the initial dissociation of the polypeptide chain from the fibril, α-syn might form either compact-conformations by long-range interactions or extended-conformations stabilized by local interactions. This leads α-syn to adapt two different pathways. The secondary structure, solvation, contact distance, interaction energies and backbone dihedrals of thirty-two selected residues were analyzed for all the 19 intermediates. The results suggested that formation of β-turns, reorganization of salt bridges, and dihedral changes in the hydrophobic regions are the major driving forces for helix-fibril transition. Structural features of the intermediates also correlated with the earlier experimental and computational studies. The study provides critical information on the fibrillation pathway of α-syn.

Structural properties of Aβ (1–40) peptide in protonation stage of one, two, and three: New insights from the histidine protonation behaviors

Structural properties and aggregation tendency can be significantly influenced by histidine behaviors (histidine on Nδ-H state is defined as δ, likewise, Nε-H: ε and both Nδ-H and Nε-H: p). In current study, we investigated structural properties of Aβ(1–40) peptide during protonation evolution stage of one, two, and three by total 19 independent replica exchange molecular dynamics simulations using implicit solvent. Our results show that any kind of protonated state will promote β-sheet structure formation in comparison with deprotonated (εεε). With increase in number of protonation, the lowest β-sheet content increased. The highest averaged β-sheet structure content was detected in (δpδ) (46.0 %), (εpp) (36.8 %), and (ppp) (16.0 %) in each protonation stage. With three β-strand structures, (δpδ) shows more stable features and high hydrophobic properties. Further analysis confirmed that H13 and H14 are more important than H6. Specifically, H13 and H14 have a synergistic effect for structural formations by controlling H-bond networks in H13(p) with V39/V40 and H14(p/δ) with G37/G38. Finally, the Pearson correlation coefficient results confirmed that experimental result (ref. 44) is corresponding to our (εpp) system. Our current study will be conducive to understanding the effects of the histidine behaviors, it provides new insights for exploration protein folding and misfolding processes.

Enhanced sampling using replica exchange with nonequilibrium switches: A case study on simple models.

Configurational sampling is central to characterize the equilibrium properties of complex molecular systems, but it remains a significant computational challenge. The conventional molecular dynamics (MD) simulations of limited duration often result in inadequate sampling and thus inaccurate equilibrium estimates. Replica exchange with nonequilibrium switches (RENS) is a collective variable-free computational technique to achieve extensive sampling from a sequence of equilibrium and nonequilibrium MD simulations without modifying the underlying potential energy surface of the system. Unlike the conventional replica exchange molecular dynamics (REMD) simulation, which demands a significant number of replicas for better accuracy, RENS employs nonequilibrium heating (forward) and cooling (reverse) work simulations prior to configurational swaps to improve the acceptance probability for replica exchange by using only a few replicas. Here, we have implemented the RENS algorithm on four model systems and examined its performance against the conventional MD and REMD simulations. The desired equilibrium distributions were generated by RENS for all the model systems, whereas REMD and MD simulations could not do so due to inadequate sampling on the same timescales. The calculated work distributions from RENS obeyed the expected nonequilibrium fluctuation theorem. The results indicate that the switching time of the nonequilibrium simulations can be systematically altered to optimize the acceptance probability and the reduced work of switching. The modular implementation of RENS algorithm not only enables us to readily extend it to multiple replicas but also paves the way for extension to larger molecular systems in the future.

On Quality Thresholds for the Clustering of Molecular Structures.

It has been recently suggested that diametral (so-called quality) similarity thresholds are superior to radial ones for the clustering of molecular three-dimensional structures (González-Alemán et al., 2020). The argument has been made for two clustering algorithms available in various software packages for the analysis of molecular structures from ensembles generated by computer simulations, attributed to Daura et al. (1999) (radial threshold) and Heyer et al. (1999) (diametral threshold). Here, we compare these two algorithms using the root-mean-squared difference (rmsd) between the Cartesian coordinates of selected atoms as pairwise similarity metric. We discuss formally the relation between these two methods and illustrate their behavior with two examples, a set of points in two dimensions and the coordinates of the tau polypeptide along a trajectory extracted from a replica-exchange molecular-dynamics simulation (Shea and Levine, 2016). We show that the two methods produce equally sized clusters as long as adequate choices are made for the respective thresholds. The real issue is not whether the threshold is radial or diametral but how to choose in either case a threshold value that is physically meaningful. We will argue that, when clustering molecular structures with the rmsd as a metric, the simplest best guess for a threshold is actually radial in nature.

Molecular interplay of ADAMTS13-MDTCS and von willebrand Factor-A2: deepened insights from extensive atomistic simulations

Thrombotic thrombocytopenic purpura (TTP) is a rare and life-threatening disease. One hallmark is severe ADAMTS13 deficiency, causing ultra-large von Willebrand factor (VWF) multimers to accumulate, leading to microthrombi and lastly to microangiopathic hemolytic anemia and severe thrombocytopenia. Despite great success in recent decades, the molecular picture of the interaction between VWF and ADAMTS13 remains vague. Here, we utilized modern replica-exchange molecular dynamics simulations with the TIGER2h method to sample a vast configurational space of the isolated ADAMTS13-MDTCS domains and the exposure to its substrate and activating cofactor – the unraveled VWF-A2 domain. The sampling of binding sites and conformations was guided and filtered in agreement with available experimental evidence. We provide comprehensive information on exosites for each domain and direct pairs of interacting amino acids, for the first time. The major binding cluster for the active site of the MP domain contrasts the previous mapping of VWF-A2 residues and reciprocal binding pockets. Two major binding modes are revealed and provide access to conformational changes of an extended gatekeeper tetrad upon overcoming local latency during substrate binding and to a dedicated recruitment mechanism. Our work adds the first molecular interaction model that places previous experimental results in perspective to better understand disease-related mutations towards improved therapies. Numerous empirical targets are proposed to verify the given binding modes, to refine the overall picture of MP binding pockets, the role of Dis binding in MP activation and the passage of the Cys-rich domain through VWF-A2, thus deepening the understanding of a highly dynamic interplay. Communicated by Ramaswamy H. Sarma

Compositional Analysis of ssDNA-Coated Single-Wall Carbon Nanotubes through UV Absorption Spectroscopy.

Aqueous suspensions of single-wall carbon nanotubes (SWCNTs) coated by ssDNA are analyzed using UV absorption and total carbon measurements. The results give absolute average concentrations of both components in samples without free ssDNA. From those values, the average mid-UV SWCNT absorptivity is deduced for three different batches of relatively small diameter nanotubes: two HiPco and one CoMoCAT. The absorptivity values enable the use of simple spectrophotometry to measure absolute concentrations of similar SWCNT samples in aqueous SDS. The results also quantify the mass ratio of ssDNA to SWCNT, defining the average number of nanotube carbon atoms suspended by one ssDNA strand of T15GT15 or T30G. Comparing this experimental parameter with results from replica exchange molecular dynamics simulations of one ssDNA strand freely adsorbed on a (6,5) segment shows close agreement between the computed number of SWCNT atoms covered per strand and the measured number of SWCNT atoms suspended per strand.

REMD Simulations of Full-Length Alpha-Synuclein Together with Ligands Reveal Binding Region and Effect on Amyloid Conversion.

Alpha-synuclein is a key protein involved in the development and progression of Parkinson’s disease and other synucleinopathies. The intrinsically disordered nature of alpha-synuclein hinders the computational screening of new drug candidates for the treatment of these neurodegenerative diseases. In the present work, replica exchange molecular dynamics simulations of the full-length alpha-synuclein together with low-molecular ligands were utilized to predict the binding site and effect on the amyloid-like conversion of the protein. This approach enabled an accurate prediction of the binding sites for three tested compounds (fasudil, phthalocyanine tetrasulfonate, and spermine), giving good agreement with data from experiments by other groups. Lots of information about the binding and protein conformational ensemble enabled the suggestion of a putative effect of the ligands on the amyloid-like conversion of alpha-synuclein and the mechanism of anti- and pro-amyloid activity of the tested compounds. Therefore, this approach looks promising for testing new drug candidates for binding with alpha-synuclein or other intrinsically disordered proteins and at the same time the estimation of the effect on protein behavior, including amyloid-like aggregation.

How Cooperatively Folding Are Homopolymer Molecular Knots?

Detailed thermodynamic analysis of complex systems with multiple stable configurational states allows for insight into the cooperativity of each individual transition. In this work we derive a heat capacity decomposition comprising contributions from each individual configurational state, which together sum to a baseline heat capacity, and contributions from each state-to-state transition. We apply this analysis framework to a series of replica exchange molecular dynamics simulations of linear and 1-1 coarse-grained homo-oligomer models which fold into stable, configurationally well-defined molecular knots, in order to better understand the parameters leading to stable and cooperative folding of these knots. We find that a stiff harmonic backbone bending angle potential is key to achieving knots with specific 3D structures. Tuning the backbone equilibrium angle in small increments yields a variety of knot topologies, including $3_1$, $5_1$, $7_1$, and $8_{19}$ types. Populations of different knotted states as functions of temperature can also be manipulated by tuning backbone torsion stiffness or by adding side chain beads. We find that sharp total heat capacity peaks for the homo-oligomer knots are largely due to a coil-to-globule transition, rather than a cooperative knotting step. However, in some cases the cooperativity of globule-to-knot and coil-to-globule transitions are comparable, suggesting that highly cooperative folding to knotted structures can be achieved by refining the model parameters or adding sequence specificity.

Melatonin Inhibits hIAPP Oligomerization by Preventing β-Sheet and Hydrogen Bond Formation of the Amyloidogenic Region Revealed by Replica-Exchange Molecular Dynamics Simulation.

The pathogenesis of type 2 diabetes (T2D) is highly related to the abnormal self-assembly of the human islet amyloid polypeptide (hIAPP) into amyloid aggregates. To inhibit hIAPP aggregation is considered a promising therapeutic strategy for T2D treatment. Melatonin (Mel) was reported to effectively impede the accumulation of hIAPP aggregates and dissolve preformed fibrils. However, the underlying mechanism at the atomic level remains elusive. Here, we performed replica-exchange molecular dynamics (REMD) simulations to investigate the inhibitory effect of Mel on hIAPP oligomerization by using hIAPP20–29 octamer as templates. The conformational ensemble shows that Mel molecules can significantly prevent the β-sheet and backbone hydrogen bond formation of hIAPP20–29 octamer and remodel hIAPP oligomers and transform them into less compact conformations with more disordered contents. The interaction analysis shows that the binding behavior of Mel is dominated by hydrogen bonding with a peptide backbone and strengthened by aromatic stacking and CH–π interactions with peptide sidechains. The strong hIAPP–Mel interaction disrupts the hIAPP20–29 association, which is supposed to inhibit amyloid aggregation and cytotoxicity. We also performed conventional MD simulations to investigate the influence and binding affinity of Mel on the preformed hIAPP1–37 fibrillar octamer. Mel was found to preferentially bind to the amyloidogenic region hIAPP20–29, whereas it has a slight influence on the structural stability of the preformed fibrils. Our findings illustrate a possible pathway by which Mel alleviates diabetes symptoms from the perspective of Mel inhibiting amyloid deposits. This work reveals the inhibitory mechanism of Mel against hIAPP20–29 oligomerization, which provides useful clues for the development of efficient anti-amyloid agents.

Fundamental Redesign of the TIGER2hs Kernel to Address Severe Parameter Sensitivity.

Replica exchange molecular dynamics simulations are one of the most popular approaches to enhance conformational sampling of molecular systems. Applications range from protein folding to protein-protein or other host-guest interactions, as well as binding free energy calculations. While these methods are computationally expensive, highly accurate results can be obtained. We recently developed TIGER2hs, an improved version of the temperature intervals with global exchange of replicas (TIGER2) algorithm. This method combines the replica-based enhanced sampling in an explicit solvent with a hybrid solvent energy evaluation. During the exchange attempts, bulk water is replaced by an implicit solvent model, allowing sampling with significantly less replicas than parallel tempering (REMD). This enables accurate enhanced sampling calculations with only a fraction of computational resources compared to REMD. Our latest results highlight several issues with sampling imbalance and parameter sensitivity within the original TIGER2 exchange algorithms that affect the overall state populations. A high sensitivity on replica number and maximum temperature is eliminated by changing to a pairwise exchange kernel (PE) without additional sorting. Simulations are controlled by adjusting the average temperature change per exchange ⟨ΔT/χ⟩ to below 30 K to mimic a controlled temperature mixing of replicas similar to REMD. Thus, this parameter provides an applicable property for selecting combinations of replica number and maximum temperature to adjust simulations for best accuracy, with flexible resource investment. This increases the robustness of the method and ensures results in excellent agreement with REMD, as demonstrated for three different peptides.