Computational Pharmaceutical Chemistry and Molecular Informatics Group

Permanent URI for this collectionhttps://researchdata.hhu.de/handle/entry/76

Our research focusses on understanding, predicting, and modulating biomolecular interactions from an atomistic perspective.

Browse

Now showing 1 - 20 of 27

Chaperone/ETR1 Structural Models for: Molecular Mechanism and Structural Models of Protein-Mediated Copper Transfer to the Arabidopsis thaliana Ethylene Receptor ETR1 at the ER Membrane
(N/A, 2025) Dluhosch, Dominik; Kersten, Lisa Sophie; Minges, Alexander; Schott-Verdugo, Stephan; Gohlke, Holger; Groth, Georg
In plants, the gaseous plant hormone ethylene regulates a wide range of developmental processes and stress responses. The small unsaturated hydrocarbon is detected by a family of receptors (ETRs) located in the membrane of the endoplasmic reticulum, which rely on a monovalent copper cofactor to detect this hydrocarbon. The copper-transporting P-type ATPase RAN1 (HMA7), located in the same membrane, is known to be essential for the biogenesis of ETRs. Still, the precise molecular mechanism by which the receptors acquire their copper cofactor remains unclear. A recent study by our laboratory demonstrated a direct interaction between RAN1 and soluble copper chaperones of the ATX1 family with the model ethylene receptor ETR1, providing initial insights into the mechanism by which copper is transferred from the cytosol to the membrane-bound receptors. In this study, we further investigated these interactions with respect to the function of individual domains in complex formation. To this end, we combined biochemical experiments and computational predictions and unraveled the processes and mechanisms by which copper is transferred to ETR1 at the molecular level.
Data for "Influence of ionic liquids on enzymatic asymmetric carboligations"
(N/A, 2025) El Harrar, Till; Gohlke, Holger
The asymmetric mixed carboligation of aldehydes catalyzed by thiamine diphosphate (ThDP)-dependent enzymes provides a sensitive system for monitoring changes in activity, chemo-, and enantioselectivity. While previous studies have shown that organic cosolvents influence these parameters, we now demonstrate that similar effects occur upon addition of water-miscible ionic liquids (ILs). In this study, six ThDP-dependent enzymes were analyzed in the presence of 14 ILs under comparable conditions to assess their influence on enzymatic carboligation reactions yielding 2-hydroxy ketones. ILs exerted a moderate to strong influence on activity and, more notably, altered enantioselectivity. (R)-selective reactions were generally stable upon IL addition, while (S)-selective reactions frequently showed reduced selectivity or even inversion to the (R)-enantiomer. The most significant change was observed for the ApPDC_E469G variant of pyruvate decarboxylase from Acetobacter pasteurianus, where the enantiomeric excess shifted from 86% (S) to 60% (R) in the presence of 9% (w/v) Ammoeng 102. Control experiments indicated that this shift was primarily due to the Ammoeng cation rather than the anion. To explore the molecular basis of this phenomenon, all-atom molecular dynamics (MD) simulations were performed on wild-type ApPDC and the E469G variant in Ammoeng 101 and Ammoeng 102. The simulations revealed that hydrophobic and hydrophilic regions of the Ammoeng cations interact with the (S)-selective binding pocket, thereby favoring formation of the (R)-product. These results highlight the potential of solvent engineering for modulating enzyme selectivity and demonstrate that MD simulations can capture functionally relevant enzyme–solvent interactions at the atomic level.
Data for "TopEC: Improved classification of enzyme function by a localized 3D protein descriptor and 3D Graph Neural Networks"
(N/A, 2024-08-25) van der Weg, Karel; Merdivan, Erinc; Piraud, Marie; Gohlke, Holger
Accurately annotating molecular function of enzymes remains challenging. Computational methods can aid in this and allow for high-throughput annotation. Tools available for inferring enzyme function from general sequence, fold, or evolutionary information are generally successful. However, they can lead to misclassification if for certain sequences a deviation in local structural features influences the function. Here, we present TopEC, a 3D graph neural network based on a localized 3D descriptor to learn chemical reactions of enzymes from (predicted) enzyme structures and predict Enzyme Commission (EC) classes. Using the message passing frameworks from SchNet and DimeNet++, we include distance and angle information to improve the predictive performance compared to regular 2D graph neural networks. We obtained significantly improved EC classification prediction (F-score: 0.72) to 2D GNNs, without fold bias at residue and atomic resolutions and trained networks that can classify both experimental and computationally generated enzyme structures for a vast functional space (> 800 ECs). Our model is robust to uncertainties in binding site locations and similar functions in distinct binding sites. By investigating the importance of each graph node to the predictive performance, we see that TopEC networks learn from an interplay between biochemical features and local shape-dependent features. TopEC is available as a repository, including accompanying data, on github: https://github.com/IBG4-CBCLab/TopEC. The data in this repository is available under the CC-BY-NC-SA 4.0 license.
Functional and structural characterization of interactions between opposite subunits in HCN pacemaker channels
(N/A, 2021-08-05) Kondapuram, Mahesh; Frieg, Benedikt; Yüksel, Sezin; Schwabe, Tina; Sattler, Christian; Lelle, Marco; Schweinitz, Andrea; Schmauder, Ralf; Benndorf, Klaus; Gohlke, Holger; Kusch, Jana
Hyperpolarization-activated and cyclic nucleotide (HCN) modulated channels are tetrameric cation channels. In each of the four subunits, the intracellular cyclic nucleotide-binding domain (CNBD) is coupled to the transmembrane domain via a helical structure, the C-linker. High-resolution channel structures suggest that the C-linker enables functionally relevant interactions with the opposite subunit, which might be critical for coupling the conformational changes in the CNBD to the channel pore. We combined mutagenesis, patch-clamp technique, confocal patch-clamp fluorometry, and molecular dynamics simulations to show that residue K464 of the C-linker is essential for stabilizing the closed state of the mHCN2 channel by forming interactions with the opposite subunit. MD simulations revealed that both cAMP and K464E induce a rotation of the intracellular domain relative to the channel pore, weakening the autoinhibitory effect of the unoccupied CL-CNBD region. The adopted poses are in excellent agreement with structural results.
MD simulation data for: "Molecular Mechanisms Underlying Medium-Chain Free Fatty Acid-regulated Activity of the Phospholipase PlaF from Pseudomonas aeruginosa"
(N/A, 2023-11) Gentile, Rocco; Schott-Verdugo, Stephan; Gohlke, Holger
PlaF is a membrane-bound phospholipase A1 from P. aeruginosa that is involved in remodeling membrane glycerophospholipids (GPLs) and modulation of virulence-associated signaling and metabolic pathways. Previously, we identified the role of medium-chain free fatty acids (FFA) in inhibiting PlaF activity and promoting homodimerization, yet the underlying molecular mechanism remained elusive. Here, we used unbiased and biased molecular dynamics simulations and free energy computations to assess how PlaF interacts with FFAs localized in the water milieu surrounding the bilayer or within the bilayer, and how these interactions regulate PlaF activity. Medium-chain FFAs localized in the upper bilayer leaflet can stabilize inactive dimeric PlaF, likely through interactions with charged surface residues as experimentally validated. Potential of mean force (PMF) computations indicate that membrane-bound FFAs may facilitate the activation of monomeric PlaF by lowering the activation barrier of changing into a tilted, active configuration. We estimated that the coupled equilibria of PlaF monomerization-dimerization and tilting at the physiological concentration of PlaF lead to the majority of PlaF forming inactive dimers when in a cell membrane loaded with decanoic acid (C10). This is in agreement with a suggested in vivo product feedback loop and GC-MS profiling results indicating that PlaF catalyzes the release of C10 from P. aeruginosa membranes. Additionally, we found that C10 in the water milieu can access the catalytic site of active monomeric PlaF, contributing to the competitive component of C10-mediated PlaF inhibition. Our study provides mechanistic insights into how medium-chain FFA may regulate the activity of PlaF, a potential bacterial drug target.
MD simulation data for: "The cyclophilin A binding loop of the capsid regulates the human TRIM5α sensitivity of nonpandemic HIV-1"
(N/A, 2023-11) Becker, Daniel; Münk, Carsten; Gohlke, Holger
All MD input structures, MD infiles, umbrella sampling files, and scripts that were used to analyze the umbrella sampling results are provided in this supporting repository.
Millisecond-long sampling for a comprehensive energetic evaluation of aqueous ionic liquid effects on amino acid interactions
(N/A, 2022-09-01) El Harrar, Till; Gohlke, Holger
The interactions of amino acid side-chains confer diverse energetic contributions and physical properties to a protein's stability and function. Various computational tools estimate the effect of changing a given amino acid on the protein's stability based on parametrized (free) energy functions. When parametrized for the prediction of protein stability in water, such energy functions can lead to suboptimal results for other solvents, such as ionic liquids (IL), aqueous ionic liquids (aIL), or salt solutions. However, to our knowledge, no comprehensive data is available describing the energetic effect of aIL on intramolecular protein interactions. Here, we present the most comprehensive set of potential of mean force (PMF) profiles of pairwise protein-residue interactions to date, covering 50 relevant interactions in water, the two biotechnologically relevant aIL [BMIM/Cl] and [BMIM/TfO], and [Na/Cl]. These results are based on a cumulated simulation time of > 1 ms. aIL and salt ions can weaken, but also strengthen, specific residue interactions by more than 3 kcal mol 1, depending on the residue pair, residue-residue configuration, participating ions, and concentration, necessitating considering such interactions specifically. These changes originate from a complex interplay of competitive or cooperative noncovalent ion-residue interactions, changes in solvent structural dynamics, or unspecific charge screening effects and occur at the contact distance but also at larger, solvent-separated distances. This data provides explanations at the atomistic and energetic level for complex IL effects on protein stability and should help improve the prediction accuracy of computational tools that estimate protein stability based on (free) energy functions.
Molecular dynamics simulations data and analysis scripts used for Grx5 in the publication "Quantitative assessment of the determinant structural differences between redox-active and inactive glutaredoxins"
(N/A, 2020-02) Wäschenbach, Lucas; Gohlke, Holger
The files in this directory contain the molecular dynamics simulations data and analysis scripts for Grx5 used in the publication "Quantitative assessment of the determinant structural differences between redox-active and inactive glutaredoxins"
Molecular dynamics simulations data and analysis scripts used for Grx7 in the publication "Quantitative assessment of the determinant structural differences between redox-active and inactive glutaredoxins"
(N/A, 2020-02) Wäschenbach, Lucas; Gohlke, Holger
The files in this directory contain the molecular dynamics simulations data and analysis scripts for Grx7 used in the publication "Quantitative assessment of the determinant structural differences between redox-active and inactive glutaredoxins"
Movie for: The architecture of the 10-23 DNAzyme and its implications for DNA-mediated catalysis
(N/A, 2022-09-19) Christoph, Gertzen; Jan, Borggräfe; Aldino, Viegas; Manuel, Etzkorn; Holger, Gohlke
Understanding the molecular features of catalytically active DNA sequences, so-called DNAzymes, is not only essential for our understanding of the fundamental properties of catalytic nucleic acids in general but may well be the key to unraveling their full potential via tailored modifications. Our recent findings contributed to the endeavor to assemble a mechanistic picture of DNA-mediated catalysis by providing high-resolution structural insights into the 10-23 DNAzyme (Dz) and exposing a complex interplay between the Dz´s unique molecular architecture, conformational plasticity, and dynamic modulation by metal ions as central elements of the DNA catalyst. To illustrate the sampled conformational space, the movie depicts one MD trajectory of the Dz:RNA complex.
Regulation of STING activity in DNA sensing by ISG15 modification
(N/A, 2023-09) Gertzen, Christoph; Kaiser, Jesko; Münk, Carsten; Gohlke, Holger
Molecular modeling and simulations suggest that ISGylation of K289 of STING is an important regulator of STING oligomerization
Sequence alignments of alpha and beta spectrin
(N/A, 2020) Wäschenbach, Lucas; Gohlke, Holger
*** Multiple sequence alignment of erythrocytic α-spectrin sequences from 58 different species Sequences of erythrocytic α-spectrin were collected from the UniProtKB database at uniprot.org. 58 sequences with “erythrocytic α-spectrin” annotation were selected with an additional criterium being a comparable sequence length to human erythrocytic α-spectrin (2,419 residues). Multiple sequence alignment was performed with Clustal Omega (webserver version accessed at https://www.ebi.ac.uk/Tools/msa/clustalo/ on 04.30.2020) with default parameters. The result is given in the file with the ending .clustal_num. In addition, the sequence alignment is displayed as a .tif file, with colored residues indicating high identiy according to the ‘identity’ coloring option of Mview; colors vary according to physicochemical properties as specified in the MView documentation (Madeira, Park et al. 2019). *** Multiple sequence alignment of erythrocytic β-spectrin sequences from 11 different species Sequences of erythrocytic β-spectrin were collected from the UniProtKB database at uniprot.org. 11 sequences with “erythrocytic β -spectrin” annotation were selected with an additional criterium being a comparable sequence length to human erythrocytic β-spectrin (2,137 residues). Multiple sequence alignment was performed with Clustal Omega (webserver version accessed at https://www.ebi.ac.uk/Tools/msa/clustalo/ on 04.30.2020) with default parameters. The result is given in the file with the ending .clustal_num. In addition, the sequence alignment is displayed as a .tif file, with colored residues indicating high identiy according to the ‘identity’ coloring option of Mview; colors vary according to physicochemical properties as specified in the MView documentation (Madeira, Park et al. 2019). *** Madeira, F., Park, Y. M., Lee, J., Buso, N., Gur, T., Madhusoodanan, N., Basutkar, P., Tivey, A. R. N., Potter, S. C., Finn, R. D., and Lopez, R. (2019) The EMBL-EBI search and sequence analysis tools APIs in 2019. Nucleic Acids Res. 47, W636-W641
Simulation data for "Substrate access mechanism in a novel membrane-bound phospholipase A of Pseudomonas aeruginosa concordant with specificity and regioselectivity"
(N/A, 2021-07-12) Ahmad, Sabahuddin; Strunk, Christoph; Schott-Verdugo, Stephan; Jaeger, Karl-Erich; Kovacic, Filip; Gohlke, Holger
PlaF is a cytoplasmic membrane-bound phospholipase A1 from Pseudomonas aeruginosa that alters the membrane glycerophospholipid (GPL) composition and fosters the virulence of this human pathogen. PlaF activity is regulated by a dimer-to-monomer transition followed by tilting of the monomer in the membrane. However, how substrates reach the active site and how the characteristics of the active site tunnels determine the activity, specificity, and regioselectivity of PlaF for natural GPL substrates has remained elusive. Here, we combined unbiased and biased all-atom molecular dynamics (MD) simulations and configurational free energy computations to identify access pathways of GPL substrates to the catalytic center of PlaF. Our results map out a distinct tunnel through which substrates access the catalytic center. PlaF variants with bulky tryptophan residues in this tunnel revealed decreased catalysis rates due to tunnel blockage. The MD simulations suggest that GPLs preferably enter the active site with the sn-1 acyl chain first, which agrees with the experimentally demonstrated PLA1 activity of PlaF. We propose that the acyl chain-length specificity of PlaF is determined by the structural features of the access tunnel, which results in favorable free energy of binding of medium-chain GPLs. The suggested egress route conveys fatty acid products to the dimerization interface and, thus, contributes to understanding the product feedback regulation of PlaF by fatty acid-triggered dimerization. These findings open up opportunities for developing potential PlaF inhibitors, which may act as antibiotics against P. aeruginosa.
SQLite file for TopCysteineDB: A Cysteinome-wide Database Integrating Structural and Chemoproteomics Data for Cysteine Ligandability Prediction
(N/A, 2025-04) Bonus, Michele; Greb, Julian; Majmudar, Jaimeen D.; Boehm, Markus; Korczynska, Magdalena; Nazemi, Azadeh; Mathiowetz, Alan M.; Gohlke, Holger
Development of targeted covalent inhibitors and covalent ligand-first approaches have emerged as a powerful strategy in drug design, with cysteines being attractive targets due to their nucleophilicity and relative scarcity. While structural biology and chemoproteomics approaches have generated extensive data on cysteine ligandability, these complementary data types remain largely disconnected. Here, we present TopCysteineDB, a comprehensive resource integrating structural information from the PDB with chemoproteomics data from activity-based protein profiling experiments. Analysis of the complete PDB yielded 264,234 unique cysteines, while the proteomics dataset encompasses 41,898 detectable cysteines across the human proteome. Using TopCovPDB, an automated classification pipeline complemented by manual curation, we identified 787 covalent cysteines and systematically categorized other functional roles, including metal-binding, cofactor-binding, and disulfide bonds. Mapping residue-wise structural information to sequence space enabled cross-referencing between structural and proteomics data, creating a unified view of cysteine ligandability. For TopCySPAL, a machine learning model was developed, integrating structural features and proteomics data, achieving strong predictive performance (AUROC: 0.964, AUPRC: 0.914) and robust generalization to novel cases. TopCysteineDB and TopCySPAL are freely accessible through a webinterface, TopCysteineDBApp (https://topcysteinedb.hhu.de/), designed to facilitate exploration of cysteine sites across the human proteome. The interface provides an interactive visualization featuring a color-coded mapping of chemoproteomics data onto cysteine site structures and the highlighting of identified peptide sequences. It offers customizable dataset downloads and ligandability predictions for user-provided structures. This resource advances targeted covalent inhibitor design by providing integrated access to previously dispersed data types and enabling systematic analysis and prediction of cysteine ligandability.
Structural and mechanistic insights into bacterial phospholipase A involved in membrane phospholipid degradation and virulence
(N/A, 2022) Stephan, Schott-Verdugo; Holger, Gohlke; Renu, Batra-Safferling; Karl-Erich, Jaeger; Filip, Kovacic
Cells steadily adapt their membrane glycerophospholipid (GPL) composition to changing environmental and developmental conditions. While the regulation of membrane homeostasis via GPL synthesis in bacteria has been studied in detail, the mechanisms underlying the controlled degradation of endogenous GPLs remain unknown. Thus far, the function of intracellular phospholipases A (PLAs) in GPL remodeling (Lands cycle) in bacteria is not clearly established. Here, we identified the first cytoplasmic membrane-bound phospholipase A1 (PlaF) from Pseudomonas aeruginosa, which might be involved in the Lands cycle. PlaF is an important virulence factor, as the P. aeruginosa ΔplaF mutant showed strongly attenuated virulence in Galleria mellonella and macrophages. We present a 2.0-Å-resolution crystal structure of PlaF, the first structure that reveals homodimerization of a single-pass transmembrane (TM) full-length protein. PlaF dimerization, mediated solely through the intermolecular interactions of TM and juxtamembrane regions, inhibits its activity. The dimerization site and the catalytic sites are linked by an intricate ligand-mediated interaction network, which might explain the product (fatty acid) feedback inhibition observed with the purified PlaF protein. We used molecular dynamics simulations and configurational free energy computations to suggest a model of PlaF activation through a coupled monomerization and tilting of the monomer in the membrane, which constrains the active site cavity into contact with the GPL substrates. Thus, these data indicate the importance of the PlaF mediated GPL remodeling pathway for virulence and could pave the way for the development of novel therapeutics targeting PlaF.
Structural models of SFV Bet protein as SI for "Foamy Viruses, Bet, and APOBEC3 Restriction"
(N/A, 2021-03-18) Vasudevan, Ananda; Becker, Daniel; Luedde, Tom; Gohlke, Holger; Muenk, Carsten
Domain boundaries within the sequence of the SFV-chimpanzee Bet protein sequence were predicted with TopDomain. The two domains were modeled using the ROBETTA webserver. For each domain, ROBETTA generated five models. The best model was chosen based on the lowest Cα atom RMSD between the generated models and the corresponding ab initio modeled domain of PFV Bet. This consensus approach was also validated by TopScore in that domains with the lowest RMSD also have the lowest TopScore compared to the other four models. For the N-terminal domain, the overall TopScore is 0.47 and for the C-terminal one 0.45, indicating a moderate quality of the models. In the N-terminal domain, the helical part of the structure has low local TopScores, indicating good quality, but the N-terminal loops high local TopScores, indicating low quality. In the C-terminal domain, the β-strands have low local TopScores, indicating good quality, but helices and loops from residues 358–403 have high local TopScores, indicating low quality.
Supporting Information for "Enzyme adaptation to habitat thermal legacy shapes the thermal plasticity of marine microbiomes"
(N/A, 2023-01) Nutschel, Christina; Pfleger, Christopher; Dittrich, Jonas; Gohlke, Holger
The dataset contains: I. 3D structural models of esterases generated by a AlphaFold2-based workflow of ColabFold, II. predicted catalytic triads of esterases and their substrate accessibilities analyzed by CAVER 3.0.3 PyMOL Plugin, III. input and output files for molecular dynamics (MD) simulations performed by Amber21, and, IV. Tp-values of esterases predicted by Constraint Network Analysis (CNA). We added template scripts and a readme file for further explanations.
Supporting Information for "F/G-Region Rigidity is Inversely Correlated to Substrate Promiscuity of Human CYP Isoforms Involved in Metabolism"
(N/A, 2021-07-13) Becker, Daniel; Bharatam, Prasad; Gohlke, Holger
Of 57 human Cytochrome P450 (CYP) enzymes, twelve metabolize 90% of xenobiotics. To our knowledge, no study has addressed the relation between enzyme dynamics and substrate promiscuity for more than three CYPs. Here, we show by constraint dilution simulations with the Constraint Network Analysis for the twelve isoforms that structural rigidity of the F/G-region is significantly inversely correlated to the enzymes’ substrate promiscuity. This highlights the functional importance of structural dynamics of the substrate tunnel.
Supporting Information for "Loading and Co-Solvent-Triggered Release of Okanin, a C4 Plant Key Enzyme Inhibitor, into/from Functional Microgels"
(N/A, 2023-01) Dittrich, Jonas; Gohlke, Holger
The constantly growing world population leads to increasing demands for food, which challenges modern agriculture manifold. Pests, such as weeds, require the application of agrochemicals to increase crop yield. Due to the environmental impact of these potentially hazardous chemicals, the demand for more efficient formulations is increasing. Promising formulations consist of easily adaptable carriers from which controllable stimuli release the agrochemicals. Here, we investigated poly(N vinylcaprolactam) (pVCL)-based microgels as a potential carrier for okanin, an inhibitor of the C4 plant key enzyme phosphoenolpyruvate carboxylase, by combining experiments, molecular simulations, and free energy computations. Dynamic light scattering, scanning transmission electron and atomic force microscopy revealed that pVCL microgels collapse and rigidify upon the loading of okanin. The simulations identified loosely adsorbed okanin and tightly bound okanin mediating inter-chain crosslinks. With increasing okanin concentration, stacking interactions of okanin occur with adsorbed and bound okanin. These findings can explain the experimentally observed collapse and the rigidification of the microgels. Based on the atomistic insights, two poly(N vinylcaprolactam co glycidyl methacrylate) microgels were synthesized, for which a doubled loading capacity of okanin was found. Finally, we investigated the triggered release of okanin using the addition of green solvents as a stimulus. This work establishes a basis for the further optimization of pVCL-based microgels as a carrier for the delivery of polyphenolic agrochemicals.
Supporting Information for "Molecular mechanisms underlying single nucleotide polymorphism-induced reactivity decrease in CYP2D6"
(N/A, 2024-02) Becker, Daniel; Gohlke, Holger
Cytochrome P450 2D6 (CYP2D6) is one of the most important enzymes involved in drug metabolism. Genetic polymorphism can influence drug metabolism by CYP2D6 such that a therapy is seriously affected by under- or overdosing of drugs. However, a general explanation at the atomistic level for poor activity is missing so far. Here we show for the 20 most common single nucleotide polymorphisms (SNPs) of CYP2D6 that poor metabolism is driven by four mechanisms. We found in extensive all-atom molecular dynamics simulations that the rigidity of the I-helix (central helix), distance between central phenylalanines (stabilizing bound substrate), availability of basic residues on the surface of CYP2D6 (binding of Cytochrome P450 reductase), and position of arginine 132 (electron transfer to heme) are essential for an extensive function of the enzyme. These results were applied to SNPs with unknown effects and potential SNPs that may lead to poor drug metabolism were identified. The revealed molecular mechanisms might be important for other drug-metabolizing Cytochrome P450 enzymes.

Browse

Browsing Computational Pharmaceutical Chemistry and Molecular Informatics Group by Title