Data from: Harnessing AlphaFold to reveal hERG channel conformational state secrets

Ngo, Khoa 1 ; Yang, Pei-Chi1 ; Yarov-Yarovoy, Vladimir 1 ; Clancy, Colleen1 ; Vorobyov, Igor 1

Published Nov 24, 2024; Updated Apr 02, 2025 on Dryad. https://doi.org/10.5061/dryad.18931zd5x

Data files

Nov 24, 2024 version files 40.19 GB

AlphaFold_Predictions.zip

93.69 MB
Drug_Docking.zip

3.17 GB
Molecular_Dynamics_Simulations.zip

36.93 GB
README.md

12.49 KB

Nov 26, 2024 version files 40.19 GB

AlphaFold_Predictions.zip

93.69 MB
Drug_Docking.zip

3.17 GB
hERG_Models.zip

766.30 KB
Molecular_Dynamics_Simulations.zip

36.93 GB
README.md

13.17 KB

Apr 02, 2025 version files 44.19 GB

AlphaFold_Predictions.zip

93.69 MB
Drug_Docking.zip

4.10 GB
hERG_Models.zip

1.29 MB
Molecular_Dynamics_Simulations_AFic3.zip

3.08 GB
Molecular_Dynamics_Simulations.zip

36.93 GB
README.md

14.39 KB

Abstract

To design safe, selective, and effective new therapies, there must be a deep understanding of the structure and function of the drug target. One of the most difficult problems to solve has been resolution of discrete conformational states of transmembrane ion channel proteins. An example is K_V11.1 (hERG), comprising the primary cardiac repolarizing current, I_Kr. hERG is a notorious drug anti-target against which all promising drugs are screened to determine potential for arrhythmia. Drug interactions with the hERG inactivated state are linked to elevated arrhythmia risk, and drugs may become trapped during channel closure. However, the structural details of multiple conformational states have remained elusive. Here, we guided AlphaFold2 to predict plausible hERG inactivated and closed conformations, obtaining results consistent with multiple available experimental data. Drug docking simulations demonstrated hERG state-specific drug interactions in good agreement with experimental results, revealing that most drugs bind more effectively in the inactivated state and are trapped in the closed state. Molecular dynamics simulations demonstrated ion conduction that aligned with earlier studies. Finally, we identified key molecular determinants of state transitions by analyzing interaction networks across closed, open, and inactivated states in agreement with earlier mutagenesis studies. Here, we demonstrate a readily generalizable application of AlphaFold2 as an effective and robust method to predict discrete protein conformations, reconcile seemingly disparate data and identify novel linkages from structure to function.

https://doi.org/10.1101/2024.01.27.577468

https://doi.org/10.5061/dryad.18931zd5x

Description of the data and file structure

We modeled the hERG channel structures in various states and validated the predictions through extensive computational simulations and comparisons with experimental data. The data were obtained using AlphaFold2 for structural prediction, Rosetta for structural modeling and drug docking, and molecular dynamics simulations.

Data structure

Molecular Dynamics Simulations/

\|- |- i_kk_0mV/  # Inactivated model molecular dynamics simulation with only K+ ions initially in selectivity filter | 0 mV membrane voltage
\|- |- |- i_kk_1000ns.dcd  # Trajectory file for 1000 ns simulation
\|- |- |- step5_input.pdb  # Entire simulation system structure with only K+ ions initially in selectivity filter
\|- |- |- step5_input.psf  # Molecular topology file
\|- |- i_kk_500mV/  # Inactivated model molecular dynamics simulation with only K+ ions initially in selectivity filter | 500 mV membrane voltage
\|- |- |- i_kk_1000ns.dcd  # Trajectory file for 1000 ns simulation
\|- |- |- step5_input.pdb  # Entire simulation system structure with only K+ ions initially in selectivity filter
\|- |- |- step5_input.psf  # Molecular topology file
\|- |- i_kk_750mV/  # Inactivated model molecular dynamics simulation with only K+ ions initially in selectivity filter | 750 mV membrane voltage
\|- |- |- i_kk_1000ns.dcd  # Trajectory file for 1000 ns simulation
\|- |- |- step5_input.pdb  # Entire simulation system structure with only K+ ions initially in selectivity filter
\|- |- |- step5_input.psf  # Molecular topology file
\|- |- i_wk_0mV/  # Inactivated model molecular dynamics simulation with K+ and water molecules initially in selectivity filter | 0 mV membrane voltage
\|- |- |- i_wk_1000ns.dcd  # Trajectory file for 1000 ns simulation
\|- |- |- step5_input.pdb  # Entire simulation system structure with K+ and water initially in selectivity filter
\|- |- |- step5_input.psf  # Molecular topology file
\|- |- i_wk_500mV/  # Inactivated model molecular dynamics simulation with K+ and water molecules initially in selectivity filter | 500 mV membrane voltage
\|- |- |- i_wk_1000ns.dcd  # Trajectory file for 1000 ns simulation
\|- |- |- step5_input.pdb  # Entire simulation system structure with K+ and water initially in selectivity filter
\|- |- |- step5_input.psf  # Molecular topology file
\|- |- i_wk_750mV/  # Inactivated model molecular dynamics simulation with K+ and water molecules initially in selectivity filter | 750 mV membrane voltage
\|- |- |- i_wk_1000ns.dcd  # Trajectory file for 1000 ns simulation
\|- |- |- step5_input.pdb  # Entire simulation system structure with K+ and water initially in selectivity filter
\|- |- |- step5_input.psf  # Molecular topology file
\|- |- o_kk_0mV/  # Open model  (PDB 5VA2-derived)molecular dynamics simulation with only K+ ions initially in selectivity filter | 0 mV membrane voltage
\|- |- |- o_kk_1000ns.dcd  # Trajectory file for 1000 ns simulation
\|- |- |- step5_input.pdb  # Entire simulation system structure with only K+ ions initially in selectivity filter
\|- |- |- step5_input.psf  # Molecular topology file
\|- |- o_kk_500mV/  # Open model (PDB 5VA2-derived) molecular dynamics simulation with only K+ ions initially in selectivity filter | 500 mV membrane voltage
\|- |- |- o_kk_1000ns.dcd  # Trajectory file for 1000 ns simulation
\|- |- |- step5_input.pdb  # Entire simulation system structure with only K+ ions initially in selectivity filter
\|- |- |- step5_input.psf  # Molecular topology file
\|- |- o_kk_750mV/  # Open model (PDB 5VA2-derived) molecular dynamics simulation with only K+ ions initially in selectivity filter | 750 mV membrane voltage
\|- |- |- o_kk_1000ns.dcd  # Trajectory file for 1000 ns simulation
\|- |- |- step5_input.pdb  # Entire simulation system structure with only K+ ions initially in selectivity filter
\|- |- |- step5_input.psf  # Molecular topology file
\|- |- o_wk_0mV/  # Open model (PDB 5VA2-derived) molecular dynamics simulation with K+ and water molecules initially in selectivity filter | 0 mV membrane voltage
\|- |- |- o_wk_1000ns.dcd  # Trajectory file for 1000 ns simulation
\|- |- |- step5_input.pdb  # Entire simulation system structure with K+ and water initially in selectivity filter
\|- |- |- step5_input.psf  # Molecular topology file
\|- |- o_wk_500mV/  # Open model (PDB 5VA2-derived) molecular dynamics simulation with K+ and water molecules initially in selectivity filter | 500 mV membrane voltage
\|- |- |- o_wk_1000ns.dcd  # Trajectory file for 1000 ns simulation
\|- |- |- step5_input.pdb  # Entire simulation system structure with K+ and water initially in selectivity filter
\|- |- |- step5_input.psf  # Molecular topology file
\|- |- o_wk_750mV/  # Open model (PDB 5VA2-derived) molecular dynamics simulation with K+ and water molecules initially in selectivity filter | 750 mV membrane voltage
\|- |- |- o_wk_1000ns.dcd  # Trajectory file for 1000 ns simulation
\|- |- |- step5_input.pdb  # Entire simulation system structure with K+ and water initially in selectivity filter
\|- |- |- step5_input.psf  # Molecular topology file

AlphaFold Predictions/

\|- |- closed/  # Closed model AlphaFold prediction
\|- |- |- co_fb27e.a3m  # Multiple sequence alignment for closed model using AlphaFold2
\|- |- |- co_fb27e.csv  # Prediction metrics for closed model from AlphaFold2
\|- |- |- co_fb27e_coverage.png  # Sequence alignment coverage visualization for closed model
\|- |- |- co_fb27e_pae.png  # Predicted aligned error plot for closed model
\|- |- |- co_fb27e_plddt.png  # Predicted local distance difference test (pLDDT) scores for closed model
\|- |- |- co_fb27e_unrelaxed_rank_*.pdb  # Unrelaxed structural predictions for closed model, various rankings
\|- |- |- log.txt  # Log file capturing AlphaFold model run details for closed model
\|- |- |- |- template/  # Template directory for AlphaFold2 predictions in closed model
\|- |- |- |- |- 5va2.cif  # Template structure file in CIF format
\|- |- |- |- |- 5va2.pdb  # Template structure in PDB format
\|- |- |- |- |- pdb70_a3m.ffdata  # AlphaFold template database file (ffdata format)
\|- |- |- |- |- pdb70_a3m.ffindex  # Index file for pdb70_a3m.ffdata
\|- |- |- |- |- pdb70_cs219.ffdata  # Alternate ffdata format file for template
\|- |- |- |- |- pdb70_cs219.ffindex  # Index for pdb70_cs219.ffdata
\|- |- |- |- co_fb27e_env/  # Environment files for AlphaFold processing
\|- |- |- |- |- bfd.mgnify30.metaeuk30.smag30.a3m  # Combined sequence alignment file
\|- |- |- |- |- msa.sh  # Script for multiple sequence alignment generation
\|- |- |- |- |- out.tar.gz  # Archive of environment data output
\|- |- |- |- |- pdb70.m8  # Alignment database file
\|- |- |- |- |- uniref.a3m  # Uniref sequence alignment data
\|- |- inactivated_step_1/  # Inactivated model AlphaFold prediction (step 1)
\|- |- |- hERG_211f6_3_unrelaxed_rank_*.pdb  # Unrelaxed structure predictions for step 1 inactivated model
\|- |- |- |- selected/
\|- |- |- |- |- inactivated_step_1_chosen.pdb  # Selected structure for further processing
\|- |- |- |- template/  # Template directory for inactivated step 1
\|- |- |- |- |- inactivated_step_1_template.pdb  # Template structure file
\|- |- inactivated_step_2/  # Inactivated model AlphaFold prediction (step 2)
\|- |- |- i_actVSD_SFregion_partialpore_fb27e.a3m  # Sequence alignment file for step 2
\|- |- |- i_actVSD_SFregion_partialpore_fb27e.csv  # Prediction metrics for step 2
\|- |- |- i_actVSD_SFregion_partialpore_fb27e_coverage.png  # Coverage visualization
\|- |- |- i_actVSD_SFregion_partialpore_fb27e_pae.png  # Aligned error plot for step 2
\|- |- |- i_actVSD_SFregion_partialpore_fb27e_plddt.png  # Predicted pLDDT scores for step 2
\|- |- |- i_actVSD_SFregion_partialpore_fb27e_unrelaxed_rank_*.pdb  # Unrelaxed structure predictions for step 2
\|- |- |- |- i_actVSD_SFregion_partialpore_fb27e_env/  # Environment files for AlphaFold step 2
\|- |- |- |- |- bfd.mgnify30.metaeuk30.smag30.a3m  # Sequence alignment file
\|- |- |- |- |- msa.sh  # MSA script
\|- |- |- |- |- out.tar.gz  # Archive of environment data
\|- |- |- |- |- pdb70.m8  # Alignment database file
\|- |- |- |- |- uniref.a3m  # Uniref sequence alignment data
\|- |- |- |- template/  # Template directory for step 2
\|- |- |- |- |- 5va2.cif  # Template structure file
\|- |- |- |- |- 5va2.pdb  # Template in PDB format
\|- |- |- |- |- pdb70_a3m.ffdata  # Database file in ffdata format
\|- |- |- |- |- pdb70_a3m.ffindex  # Index file for pdb70_a3m.ffdata
\|- |- |- |- |- pdb70_cs219.ffdata  # Alternate ffdata format file
\|- |- |- |- |- pdb70_cs219.ffindex  # Index for pdb70_cs219.ffdata
\|- |- open/  # Open model AlphaFold prediction (control case)
\|- |- |- cite.bibtex  # Citation information for open model studies
\|- |- |- config.json  # Configuration file for open model predictions
\|- |- |- log.txt  # Log file for open model AlphaFold run
\|- |- |- o_actVSD_SFregion_partialpore_lowtol_fb27e.a3m  # Sequence alignment for open model, low tolerance
\|- |- |- o_actVSD_SFregion_partialpore_lowtol_fb27e.csv  # Prediction metrics for open model
\|- |- |- o_actVSD_SFregion_partialpore_lowtol_fb27e.done.txt  # Status file for open model prediction completion
\|- |- |- o_actVSD_SFregion_partialpore_lowtol_fb27e_coverage.png  # Sequence coverage plot
\|- |- |- o_actVSD_SFregion_partialpore_lowtol_fb27e_pae.png  # Aligned error plot for open model
\|- |- |- o_actVSD_SFregion_partialpore_lowtol_fb27e_plddt.png  # Predicted pLDDT scores for open model
\|- |- |- o_actVSD_SFregion_partialpore_lowtol_fb27e_unrelaxed_rank_*.pdb  # Unrelaxed structure predictions for open model
\|- |- |- |- o_actVSD_SFregion_partialpore_lowtol_fb27e_env/  # Environment files for open model
\|- |- |- |- |- bfd.mgnify30.metaeuk30.smag30.a3m  # Sequence alignment file
\|- |- |- |- |- msa.sh  # MSA script
\|- |- |- |- |- out.tar.gz  # Archive of environment data
\|- |- |- |- |- pdb70.m8  # Alignment database file
\|- |- |- |- |- uniref.a3m  # Uniref sequence alignment data
\|- |- |- |- template/  # Template directory for open model
\|- |- |- |- |- 5va2.cif  # Template structure file in CIF format
\|- |- |- |- |- 5va2.pdb  # Template structure in PDB format
\|- |- |- |- |- pdb70_a3m.ffdata  # AlphaFold template database file (ffdata format)
\|- |- |- |- |- pdb70_a3m.ffindex  # Index for pdb70_a3m.ffdata
\|- |- |- |- |- pdb70_cs219.ffdata  # Alternate ffdata format file
\|- |- |- |- |- pdb70_cs219.ffindex  # Index for pdb70_cs219.ffdata

Drug Docking/

\|- |- E-4031  # Drug name
\|- |- |- pdb/
\|- |- |- |- co_e4031_1_*.pdb  # Model structures of closed conformation with E-4031
\|- |- |- |- i_e4031_1_*.pdb  # Model structures of inactivated conformation with E-4031
\|- |- |- |- o_e4031_1_*.pdb  # Model structures of open (PDB 5VA2-derived) conformation with E-4031
\|- |- |- |- ic3_e4031_1_*.pdb  # Model structures of open (AF ic3) conformation with E-4031
\|- |- |- scores/
\|- |- |- |- co_e4031.sc  # Score file for closed conformation with E-4031
\|- |- |- |- i_e4031.sc  # Score file for inactivated conformation with E-4031
\|- |- |- |- o_e4031.sc  # Score file for open (PDB 5VA2-derived) conformation with E-4031
\|- |- |- |- ic3_e4031.sc  # Score file for open (AF ic3) conformation with E-4031
...
(Continue similarly for each drug and ionization states in complex with hERG closed, inactivated, and open conformations)

Molecular Dynamics Simulations AFic3/

\|- |- ic3_kk_750mV/  # Open model (AF ic3) molecular dynamics simulation with only K+ ions initially in selectivity filter | 750 mV membrane voltage
\|- |- |- ic3_kk_1000ns.dcd  # Trajectory file for 1000 ns simulation
\|- |- |- step5_input.pdb  # Entire simulation system structure with only K+ ions initially in selectivity filter
\|- |- |- step5_input.psf  # Molecular topology file
\|- |- ic3_wk_750mV/  # Open model (AF ic3) molecular dynamics simulation with K+ and water molecules initially in selectivity filter | 750 mV membrane voltage
\|- |- |- ic3_wk_1000ns.dcd  # Trajectory file for 1000 ns simulation
\|- |- |- step5_input.pdb  # Entire simulation system structure with only K+ ions initially in selectivity filter
\|- |- |- step5_input.psf  # Molecular topology file

hERG Models/

\|- co.pdb  # Structural file of hERG putative closed-state model after Rosetta refinement
\|- i.pdb # Structural file of hERG putative inactivated-state model after Rosetta refinement
\|- o.pdb  # Structural file of hERG open-state model (PDB 5VA2-derived) after Rosetta refinement
\|- ic3.pdb  # Structural file of hERG open-state model (AF ic3) after Rosetta refinement

Code/software

Analysis scripts can be found at: https://github.com/k-ngo/AlphaFold_Analysis

The Rosetta software suite can be downloaded from the following link: https://www.rosettacommons.org/software/license-and-download (opens in new window)

The AMBER software suite can be downloaded from the following link: https://ambermd.org/GetAmber.ph

Structural prediction with AlphaFold2 can be performed on Google Colab through ColabFold: https://colab.research.google.com/github/sokrypton/ColabFold/blob/main/AlphaFold2.ipynb

Change log

04/02/2025: Updated "hERG_Models.zip" containing models of hERG channels in putative closed, inactivated, and open state (PDB 5VA2-derived and from inactivated-state-sampling cluster 3, AF ic3). Added "Molecular_Dynamics_Simulations_AFic3.zip" containing supplemental simulation files for the AF ic3 model. Updated "Drug_Docking.zip" will revised drug docking results, now with data for AF ic3 model.

11/26/2024: Added "hERG_Models.zip" containing models of hERG channels in putative closed, inactivated, and open state. After undergoing structural refinement in Rosetta, these models were used for subsequent drug docking, molecular dynamics simulations, and structural analyses.