Increasing the Realism of in Silico pHLIP Peptide Models with a Novel pH Gradient CpHMD Method.

The pH-low insertion peptides (pHLIP) are pH-dependent membrane inserting peptides, whose function depends on the cell microenvironment acidity. Several peptide variants have been designed to improve upon the wt-sequence, particularly the state transition kinetics and the selectivity for tumor pH. The variant 3 (Var3) peptide is a 27 residue long peptide, with a key titrating residue (Asp-13) that, despite showing a modest performance in liposomes (pKins ∼ 5.0), excelled in tumor cell experiments. To help rationalize these results, we focused on the pH gradient in the cell membrane, which is one of the crucial properties that are not present in liposomes. We extended our CpHMD-L method and its pH replica-exchange (pHRE) implementation to include a pH gradient and mimic the pHLIP-membrane microenvironment in a cell where the internal pH is fixed (pH 7.2) and the external pH is allowed to change. We showed that, by properly modeling the pH-gradient, we can correctly predict the experimentally observed loss and gain of performance in tumor cells experiments by the wt and Var3 sequences, respectively. In sum, the pH gradient implementation allowed for more accurate and realistic pKa estimations and was a pivotal step in bridging the in silico data and the in vivo cell experiments.

Optimization of an in Silico Protocol Using Probe Permeabilities to Identify Membrane Pan-Assay Interference Compounds.

Membrane pan-assay interference compounds (PAINS) are a class of molecules that interact nonspecifically with lipid bilayers and alter their physicochemical properties. An early identification of these compounds avoids chasing false leads and the needless waste of time and resources in drug discovery campaigns. In this work, we optimized an in silico protocol on the basis of umbrella sampling (US)/molecular dynamics (MD) simulations to discriminate between compounds with different membrane PAINS behavior. We showed that the method is quite sensitive to membrane thickness fluctuations, which was mitigated by changing the US reference position to the phosphate atoms of the closest interacting monolayer. The computational efficiency was improved further by decreasing the number of umbrellas and adjusting their strength and position in our US scheme. The inhomogeneous solubility-diffusion model (ISDM) used to calculate the membrane permeability coefficients confirmed that resveratrol and curcumin have distinct membrane PAINS characteristics and indicated a misclassification of nothofagin in a previous work. Overall, we have presented here a promising in silico protocol that can be adopted as a future reference method to identify membrane PAINS.

Smartphone-Based Body Location-Independent Functional Mobility Analysis in Patients with Parkinson's Disease: A Step towards Precise Medicine.

Ecological evaluation of gait using mobile technologies provides crucial information regarding the evolution of symptoms in Parkinson's disease (PD). However, the reliability and validity of such information may be influenced by the smartphone's location on the body. This study analyzed how the smartphone location affects the assessment of PD patients' gait in a free-living environment. Twenty PD patients (mean ± SD age, 64.3 ± 10.6 years; 9 women (45%) performed 3 trials of a 250 m outdoor walk using smartphones in 5 different body locations (pants pocket, belt, hand, shirt pocket, and a shoulder bag). A method to derive gait-related metrics from smartphone sensors is presented, and its reliability is evaluated between different trials as well as its concurrent validity against optoelectronic and smartphone criteria. Excellent relative reliability was found with all intraclass correlation coefficient values above or equal to 0.85. High absolute reliability was observed in 21 out of 30 comparisons. Bland-Altman analysis revealed a high level of agreement (LoA between 4.4 and 17.5%), supporting the use of the presented method. This study advances the use of mobile technology to accurately and reliably quantify gait-related metrics from PD patients in free-living walking regardless of the smartphone's location on the body.

Feasibility of a Mobile-Based System for Unsupervised Monitoring in Parkinson's Disease.

Mobile health (mHealth) has emerged as a potential solution to providing valuable ecological information about the severity and burden of Parkinson's disease (PD) symptoms in real-life conditions. Objective: The objective of our study was to explore the feasibility and usability of an mHealth system for continuous and objective real-life measures of patients' health and functional mobility, in unsupervised settings. Methods: Patients with a clinical diagnosis of PD, who were able to walk unassisted, and had an Android smartphone were included. Patients were asked to answer a daily survey, to perform three weekly active tests, and to perform a monthly in-person clinical assessment. Feasibility and usability were explored as primary and secondary outcomes. An exploratory analysis was performed to investigate the correlation between data from the mKinetikos app and clinical assessments. Results: Seventeen participants (85%) completed the study. Sixteen participants (94.1%) showed a medium-to-high level of compliance with the mKinetikos system. A 6-point drop in the total score of the Post-Study System Usability Questionnaire was observed. Conclusions: Our results support the feasibility of the mKinetikos system for continuous and objective real-life measures of a patient's health and functional mobility. The observed correlations of mKinetikos metrics with clinical data seem to suggest that this mHealth solution is a promising tool to support clinical decisions.

Identification of Pan-Assay INterference compoundS (PAINS) Using an MD-Based Protocol.

Pan-assay interference compounds (PAINS) are promiscuous molecules with multiple behaviors that interfere with assay readouts. Membrane PAINS are a subset of these compounds that influence the function of membrane proteins by nonspecifically perturbing the lipid membranes that surround them. Here, we describe a computational protocol to identify potential membrane PAINS molecules by calculating the effect that a given compound has on the bilayer deformation propensity.

Improved Protocol to Tackle the pH Effects on Membrane-Inserting Peptides.

Many important biological pathways rely on membrane-interacting peptides or proteins, which can alter the biophysical properties of the cell membrane by simply adsorbing to its surface to undergo a full insertion process. To study these phenomena with atomistic detail, model peptides have been used to refine the current computational methodologies. Improvements have been made with force-field parameters, enhanced sampling techniques to obtain faster sampling, and the addition of chemical-physical properties, such as pH, whose influence dramatically increases at the water/membrane interface. The pH (low) insertion peptide (pHLIP) is a peptide that inserts across a membrane bilayer depending on the pH due to the presence of a key residue (Asp14) whose acidity-induced protonation triggers the whole process. The complex nature of these peptide/membrane interactions resulted in sampling limitations of the protonation and configurational space albeit using state-of-the-art methods such as the constant-pH molecular dynamics. To address this issue and circumvent those limitations, new simulations were performed with our newly developed pH-replica exchange method using wild-type (wt)-pHLIP in different 2-oleoyl-1-palmitoyl-sn-glycero-3-phosphocholine membrane sizes. This technique provided enhanced sampling and allowed for the calculation of more complete Asp14 pKa profiles. The conformational heterogeneity derived from strong electrostatic interactions between Asp14 and the lipid phosphate groups was identified as the source of most pKa variability. In spite of these persistent and harder-to-equilibrate phosphate interactions, the pKa values at deeper regions (6.0-6.2) still predicted the experimental pK of insertion (6.0) since the electrostatic perturbation decays as the residue inserts further into the membrane. We also observed that reducing the system size leads to membrane deformations where it increasingly loses the ability to accommodate the pHLIP-induced perturbations. This indicates that large membrane patches, such as 256 or even 352 lipids, are needed to obtain stable and more realistic pHLIP/membrane systems. These results strengthen our method pKa predictive and analytical capabilities to study the intricate play of electrostatic effects of the peptide/membrane interface, granting confidence for future applications in similar systems.

Halogen Bonding: An Underestimated Player in Membrane-Ligand Interactions.

Halogen bonds (XBs) are noncovalent interactions where halogen atoms act as electrophilic species interacting with Lewis bases. These interactions are relevant in biochemical systems being increasingly explored in drug discovery, mainly to modulate protein-ligand interactions, but are also found in engineered protein or nucleic acid systems. In this work, we report direct evidence for the existence of XBs in the context of biological membrane systems, thus expanding the scope of application of these interactions. Indeed, our molecular dynamics simulations show the presence of favorable interactions between halobenzene derivatives and both phosphate or ester oxygen acceptors from a model phospholipid bilayer, thus supporting the existence of XB-mediated phospholipid-halogen recognition phenomena influencing the membrane insertion profile of the ligands and their orientational preferences. This represents a relevant interaction, previously overlooked, eventually determining the pharmacological or toxicological activity of halogenated compounds and hence with potential implications in drug discovery and development, a place where such species account for a significant part of the chemical space. We also provide insights into a potential role for XBs in the water-to-membrane insertion of halogenated ligands as XBs are systematically observed during this process. Therefore, our data strongly suggest that, as the ubiquitous hydrogen bond, XBs should be accounted for in the development of membrane partition models.

PypKa: A Flexible Python Module for Poisson-Boltzmann-Based pKa Calculations.

The protonation of titratable residues has a significant impact on the structure and function of biomolecules, influencing many physicochemical and ADME properties. Thus, the importance of the estimation of protonation free energies (pKa values) is paramount in different scientific communities, including bioinformatics, structural biology, or medicinal chemistry. Here, we introduce PypKa, a flexible tool to predict Poisson-Boltzmann/Monte Carlo-based pKa values of titratable sites in proteins. This application was benchmarked using a large data set of experimental values to show that our single structure-based method is fast and has a competitive performance. This is a free and open-source tool that provides a simple, reusable, and extensible Python API and CLI for pKa calculations with a valuable trade-off between fast and accurate predictions. PypKa allows pKa calculations in existing protocols with the addition of a few extra lines of code. PypKa supports CPU parallel computing on solvated proteins obtained from the PDB repository but also from MD simulations using three common naming schemes: GROMOS, AMBER, and CHARMM. The code and documentation to this open-source project is publicly available at https://github.com/mms-fcul/PypKa.

The Early Phase of ß2m Aggregation: An Integrative Computational Study Framed on the D76N Mutant and the ΔN6 Variant.

Human ß2-microglobulin (b2m) protein is classically associated with dialysis-related amyloidosis (DRA). Recently, the single point mutant D76N was identified as the causative agent of a hereditary systemic amyloidosis affecting visceral organs. To get insight into the early stage of the ß2m aggregation mechanism, we used molecular simulations to perform an in depth comparative analysis of the dimerization phase of the D76N mutant and the ΔN6 variant, a cleaved form lacking the first six N-terminal residues, which is a major component of ex vivo amyloid plaques from DRA patients. We also provide first glimpses into the tetramerization phase of D76N at physiological pH. Results from extensive protein-protein docking simulations predict an essential role of the C- and N-terminal regions (both variants), as well as of the BC-loop (ΔN6 variant), DE-loop (both variants) and EF-loop (D76N mutant) in dimerization. The terminal regions are more relevant under acidic conditions while the BC-, DE- and EF-loops gain importance at physiological pH. Our results recapitulate experimental evidence according to which Tyr10 (A-strand), Phe30 and His31 (BC-loop), Trp60 and Phe62 (DE-loop) and Arg97 (C-terminus) act as dimerization hot-spots, and further predict the occurrence of novel residues with the ability to nucleate dimerization, namely Lys-75 (EF-loop) and Trp-95 (C-terminus). We propose that D76N tetramerization is mainly driven by the self-association of dimers via the N-terminus and DE-loop, and identify Arg3 (N-terminus), Tyr10, Phe56 (D-strand) and Trp60 as potential tetramerization hot-spots.

Halogen bonding in halocarbon-protein complexes and computational tools for rational drug design.

Introduction: Halogens have a prominent role in drug design. Often used as a mean to improve ADME properties, they are also becoming a tool in protein-ligand recognition given their ability to form a non-covalent interaction, termed halogen bond, where halogens act as electrophilic species interacting with electron-rich partners. Rational drug design of halogen-bonding lead molecules requires an accurate description of halocarbon-protein complexes by computational tools though not all methods are able to tackle this non-covalent interaction. Areas covered: The authors present a review of computational methodologies that can be used to properly describe halogen bonds in the context of protein-ligand complexes, providing also insights on how these methods can be used in the context of computer-aided drug design. Expert opinion: Although in the last few years many computational tools, ranging from fast screening methods to the more expensive QM calculations, have been developed to tackle the halogen bonding phenomenon, they are not yet standard in the literature. This will eventually change as official software distributions are including support for halogen bonding in their methods. Tackling desolvation of halogenated species seems to be a good strategy to improve the accuracy of computational methods, that will be more commonly used prior to laboratory work in the future.

Tackling Halogenated Species with PBSA: Effect of Emulating the σ-Hole.

To model halogen-bond phenomena using classical force fields, an extra point (EP) of charge is frequently introduced at a given distance from the halogen (X) to emulate the σ-hole. The resulting molecular dynamics (MD) trajectories can be used in subsequent molecular mechanics (MM) combined with Poisson-Boltzmann and surface area calculations (PBSA) to estimate protein-ligand binding free energies (Δ Gbind). While EP addition improves the MM/MD description of halogen-containing systems, its effect on the calculation of solvation free energies (Δ Gsolv) using the PBSA approach is yet to be assessed. As the PBSA calculations depend, among other parameters, on the empirical assignment of radii (PB radii), a problematic issue arises, since standard halogen radii are smaller than the typical X···EP distances, thus placing the EP within the solvent dielectric. Herein, we took a common literature EP parametrization scheme, which uses X···EP = Rmin and RESP charges in the context of GAFF, and performed a comprehensive study on the performance of PBSA (using three different setups) in the calculation of Δ Gsolv values for 142 halogenated compounds (bearing Cl, Br, or I) for which the experimental values are known. By conducting an optimization (minimizing the error against experimental values), we provide a new optimized set of halogen PB radii, for each PBSA setup, that should be used in the context of the aforementioned scenario. A simultaneous optimization of PB radii and X···EP distances shows that a wide range of distance/radius pairs can be used without significant loss of accuracy, therefore laying the basis for expanding this halogen radii optimization strategy to other force fields and EP implementations. As ligand Δ Gsolv estimation is an important term in the determination of protein-ligand Δ Gbind, this work is particularly relevant in the framework of structure-based virtual screening and related computer-aided drug design routines.

A pH Replica Exchange Scheme in the Stochastic Titration Constant-pH MD Method.

Solution pH is a physicochemical property that has a key role in cellular regulation, and its impact at the molecular level is often difficult to study by experimental methods. In this context, several theoretical methods were developed to study pH effects in macromolecules. The stochastic titration constant-pH molecular dynamics method (CpHMD) was developed by coupling molecular sampling methods, which are appropriate to study the conformational ensemble of biomolecules, with continuum electrostatics approaches, which properly describe pH-dependent protonation states. However, in difficult cases, the protonation sampling can be too slow for the commonly accessible computational times. In this work, we combined a pH replica exchange scheme with this CpHMD method and explored several optimization strategies and possible limitations.

Sugar-based bactericides targeting phosphatidylethanolamine-enriched membranes.

Anthrax is an infectious disease caused by Bacillus anthracis, a bioterrorism agent that develops resistance to clinically used antibiotics. Therefore, alternative mechanisms of action remain a challenge. Herein, we disclose deoxy glycosides responsible for specific carbohydrate-phospholipid interactions, causing phosphatidylethanolamine lamellar-to-inverted hexagonal phase transition and acting over B. anthracis and Bacillus cereus as potent and selective bactericides. Biological studies of the synthesized compound series differing in the anomeric atom, glycone configuration and deoxygenation pattern show that the latter is indeed a key modulator of efficacy and selectivity. Biomolecular simulations show no tendency to pore formation, whereas differential metabolomics and genomics rule out proteins as targets. Complete bacteria cell death in 10 min and cellular envelope disruption corroborate an effect over lipid polymorphism. Biophysical approaches show monolayer and bilayer reorganization with fast and high permeabilizing activity toward phosphatidylethanolamine membranes. Absence of bacterial resistance further supports this mechanism, triggering innovation on membrane-targeting antimicrobials.

The Impact of Using Single Atomistic Long-Range Cutoff Schemes with the GROMOS 54A7 Force Field.

With the recent increase in computing power, the molecular modeling community is now more focused on improving the accuracy and overall quality of biomolecular simulations. For the available simulation packages, force fields, and all other associated methods used, this relates to how well they describe the conformational space and thermodynamic properties of a biomolecular system. The parameter sets of GROMOS force fields have been parametrized and validated with the reaction field (RF) method using charge groups and a twin-range cutoff scheme (0.8/1.4 nm). However, the most recent versions of GROMACS (since v.2016) discontinued the support for charge groups. To take full advantage of the newer and faster versions of this software package with GROMOS 54A7 and RF, we need to evaluate the impact of using a single cutoff scheme (vs twin-range) and of using the Verlet list update method (which is atomistic) compared to the group-based cutoff scheme. Our results show that the GROMOS 54A7 force field seems consistent with a single cutoff, since the resulting conformation and protonation ensembles were indistinguishable. The GROMOS parametrization procedure was also reproduced using an atomistic cutoff scheme, and we have observed that the hydration free energy values of small amino acid side-chain analogues were similar to the ones obtained with the group-based protocol. We do observe a small impact of the atomistic cutoff scheme in the conformational space of the model systems studied (G1-PAMAM and DMPC). However, since the structural properties of these systems are well converged for the cutoff range used (1.4-2.0 nm), unlike with the group-based cutoff schemes, we are confident that the atomistic cutoff can be adopted with RF for MD and constant-pH MD biomolecular simulations using the GROMOS 54A7 force field.

Biomolecular Simulations of Halogen Bonds with a GROMOS Force Field.

Halogen bonds (XBs) are non-covalent interactions in which halogens (X), acting as electrophiles, interact with Lewis bases. XBs are able to mediate protein-ligand recognition and therefore play an important role in rational drug design. In this context, the development of molecular modeling tools that can tackle XBs is paramount. XBs are predominantly explained by the existence of a positive region on the electrostatic potential of X named the σ-hole. Typically, with molecular mechanics force fields, this region is modeled using a charged extra point (EP) linked to X along the R-X covalent bond axis. In this work, we developed the first EP-based strategy for GROMOS force fields (specifically GROMOS 54A7) using bacteriophage T4 lysozyme in complex with both iodobenzene and iodopentafluorobenzene as a prototype system. Several EP parametrization schemes were tested by adding a virtual interaction site to ligand topologies retrieved from the Automated Topology Builder (ATB) and Repository. Contrary to previous approaches using other force fields, our analysis is based on the capability of each parametrization scheme to sample XBs during MD simulations. Our results indicate that the implementation of an EP at a distance from iodine corresponding to Rmin provides a good qualitative description of XBs in MD simulations, supporting the compatibility of our approach with the GROMOS 54A7 force field.

Role of Counterions in Constant-pH Molecular Dynamics Simulations of PAMAM Dendrimers.

Electrostatic interactions play a pivotal role in the structure and mechanism of action of most biomolecules. There are several conceptually different methods to deal with electrostatics in molecular dynamics simulations. Ionic strength effects are usually introduced using such methodologies and can have a significant impact on the quality of the final conformation space obtained. We have previously shown that full system neutralization can lead to wrong lipidic phases in the 25% PA/PC bilayer (J. Chem. Theory Comput. 2014,10, 5483-5492). In this work, we investigate how two limit approaches to the ionic strength treatment (implicitly with GRF or using full system neutralization with either GRF or PME) can influence the conformational space of the second-generation PAMAM dendrimer. Constant-pH MD simulations were used to map PAMAM's conformational space at its full pH range (from 2.5 to 12.5). Our simulations clearly captured the coupling between protonation and conformation in PAMAM. Interestingly, the dendrimer conformational distribution was almost independent of the ionic strength treatment methods, which is in contrast to what we have observed in charged lipid bilayers. Overall, our results confirm that both GRF with implicit ionic strength and a fully neutralized system with PME are valid approaches to model charged globular systems, using the GROMOS 54A7 force field.

Membrane-Induced p Ka Shifts in wt-pHLIP and Its L16H Variant.

The pH (low) insertion peptides (pHLIPs) is a family of peptides that are able to insert into a lipid bilayer at acidic pH. The molecular mechanism of pHLIPs insertion, folding, and stability in the membrane at low pH is based on multiple protonation events, which are challenging to study at the molecular level. More specifically, the relation between the experimental p K of insertion (p Kexp) of pHLIPs and the p Ka of the key residues is yet to be clarified. We carried out a computational study, complemented with new experimental data, and established the influence of (de)protonation of titrable residues on the stability of the peptide membrane-inserted state. Constant-pH molecular dynamics simulations were employed to calculate the p Ka values of these residues along the membrane normal. In the wt-pHLIP, we identified Asp14 as the key residue for the stability of the membrane-inserted state, and its p Ka value is strongly correlated with the experimental p Kexp measured in thermodynamics studies. Also, in order to narrow down the pH range at which pHLIP is stable in the membrane, we designed a new pHLIP variant, L16H, where Leu in the 16th position was replaced by a titrable His residue. Our results showed that the L16H variant undergoes two transitions. The calculated p Ka and experimentally observed p Kexp values are in good agreement. Two distinct p Kexp values delimit a pH range where the L16H peptide is stably inserted in the membrane, while, outside this range, the membrane-inserted state is destabilized and the peptide exits from the bilayer. pHLIP peptides have been successfully used to target cancer cells for the delivery of diagnostics and therapeutic agents to acidic tumors. The fine-tuning of the stability of the pHLIP inserted state and its restriction to a narrow well-defined pH range might allow the design of new peptides, able to discriminate between tissues with different extracellular pH values.

Differential targeting of membrane lipid domains by caffeic acid and its ester derivatives.

Phenolic acids have been associated to a wide range of important health benefits underlain by a common molecular mechanism of action. Considering that significant membrane permeation is prevented by their hydrophilic character, we hypothesize that their main effects result from the interplay with cell membrane surface. This hypothesis was tested using the paradigmatic caffeic acid (CA) and two of its ester derivatives, rosmarinic (RA) and chlorogenic (CGA) acids, for which we predict, based on molecular dynamics simulations, a shallow location in phospholipid bilayers dependent on the protonation-state. Using complementary experimental approaches, an interaction with the membrane was definitely revealed for the three compounds, with RA exhibiting the highest lipid bilayer partition, and the redox signals of membrane-bound RA and CA being clearly detected. Cholesterol decreased the compounds bilayer partition, but not their ability to lower membrane dipole potential. In more complex membrane models containing also sphingomyelin, with liquid disordered (ld)/ liquid ordered (lo) phases coexistence, mimicking domains in the external leaflet of human plasma membrane, all compounds were able to affect nanodomains lateral organization. RA, and to a lesser extent CGA, decreased the size of lo domains. The most significant effect of CA was the possible formation of a rigid gel-like phase, enriched in sphingomyelin. In addition, all phenolic acids decreased the order of lo domains. In sum, phenolic acid effects on the membrane are enhanced in cholesterol-rich lo phases, which predominate in the outer leaflet of human cell membranes and are involved in many key cellular processes.

Insights on the Mechanism of Action of INH-C10 as an Antitubercular Prodrug.

Tuberculosis remains one of the top causes of death worldwide, and combating its spread has been severely complicated by the emergence of drug-resistance mutations, highlighting the need for more effective drugs. Despite the resistance to isoniazid (INH) arising from mutations in the katG gene encoding the catalase-peroxidase KatG, most notably the S315T mutation, this compound is still one of the most powerful first-line antitubercular drugs, suggesting further pursuit of the development of tailored INH derivatives. The N'-acylated INH derivative with a long alkyl chain (INH-C10) has been shown to be more effective than INH against the S315T variant of Mycobacterium tuberculosis, but the molecular details of this activity enhancement are still unknown. In this work, we show that INH N'-acylation significantly reduces the rate of production of both isonicotinoyl radical and isonicotinyl-NAD by wild type KatG, but not by the S315T variant of KatG mirroring the in vivo effectiveness of the compound. Restrained and unrestrained MD simulations of INH and its derivatives at the water/membrane interface were performed and showed a higher preference of INH-C10 for the lipidic phase combined with a significantly higher membrane permeability rate (27.9 cm s-1), compared with INH-C2 or INH (3.8 and 1.3 cm s-1, respectively). Thus, we propose that INH-C10 is able to exhibit better minimum inhibitory concentration (MIC) values against certain variants because of its better ability to permeate through the lipid membrane, enhancing its availability inside the cell. MIC values of INH and INH-C10 against two additional KatG mutations (S315N and D735A) revealed that some KatG variants are able to process INH faster than INH-C10 into an effective antitubercular form (wt and S315N), while others show similar reaction rates (S315T and D735A). Altogether, our results highlight the potential of increased INH lipophilicity for treating INH-resistant strains.