The Molecular Mechanism of Cooperative Activation and Control of G-Protein Coupled Receptors
Friedrich-Alexander-Universität Erlangen-Nürnberg, Computer-Chemie-Centrum and Interdisciplinary Center for Molecular Materials
Local Project ID:
HPC Platform used:
G-protein coupled receptors (GPCRs) are membrane proteins that transmit the effects of extracellular ligands to effect changes in the intracellular G-protein signaling system. Approximately 800 GPCRs are encoded in the human genome and approximately half of marketed drugs target GPCRs. It is therefore not surprising that GPCRs research was recognized by the award of the 2012 Nobel Prize in Chemistry to Robert Lefkowitz and Brian Kobilka.
Crystal structures of only 64 different GPCRs are currently available. Importantly, GPCRs can exist in active or inactive conformations and in binary complexes with ligands or intracellular binding partners (IBPs, G-proteins or β-arrestin) or in ternary complexes with both a ligand and an IBP. Crystal structures often deviate from the natural system: Proteins, especially membrane-bound ones, do not necessarily crystallize in their biologically active structures and the measures needed to obtain suitable GPCR crystals tend to increase the diversity between the natural environment and the crystal. It is within this context that molecular-dynamics simulations play a special role in GPCR research as a full-value complement to experimental studies.
The ternary complex model and experimental findings suggest that both an agonist ligand and a bound G-protein are necessary in order to activate GPCRs. It is therefore significant that the first molecular dynamics (MD) simulations of a ternary GPCR complex were reported only eight years ago. Such simulations are now commonplace and the comparison between binary ligand-receptor and ternary complexes has become a valuable tool in GPCR research.
Very long timescale MD simulations can be performed on specialized hardware but are less effective on more conventional massively parallel supercomputers because the simulations only scale up to a relatively limited number of CPUs or GPUs. Luckily, modern variations of metadynamics can make very effective use of massively parallel CPU-based hardware, as has been shown in this project.
One goal of the project is to characterize and simulate receptor activation. An initial study of variants of the histamine 4 receptor (H4R), for which a wide range of basal activities is observed, showed that the observed activity correlates well with stabilities calculated using Constraint Network Analysis (Wifling et al., 2019).
A general activation index was trained using sixteen multi-microsecond simulations of class A GPCRs in defined activation states. This index, denoted A100, can also be calculated for X-ray crystal structures. A100 was calculated for 275 X-ray crystal structures for 114 different receptors as validation. The “active” and “inactive” classes are separated extremely well, whereas the histogram for the “intermediate” structures shows no indication of stable partially active structures, which oscillate between stable active and inactive states (Ibrahim et al., 2019).
Importantly, A100 can be used as a single collective variable for metadynamics simulations of the activation process. Metadynamics simulations using A100 as CV have successfully calculated free-energy profiles for the activation/deactivation of the b2-adrenergic receptor.
Simulations have also addressed spontaneous binding of peptides derived from the Gsα C-terminus, the key interaction site of the Gs protein to the β2-adrenergic receptor. The aim of the analysis was to find the intermediate states of binding to investigate the initial steps of complex formation. X-ray structures from the Kobilka lab indicate that Gs may bind to the β2-adrenergic receptor in at least two different binding modes. MD simulations identified both binding sites and potential paths for G protein binding (Liu et al., 2019). The ultimate goal is to establish a computational protocol that makes it possible to identify binding sites, intermediates and determine free energies of Gα-binding using a single standard CV that is generally applicable.
The Figure illustrates our proposed process of GPCR–G protein complex formation.
A third sub-project has studied ligand binding to class A GPCRs and how this process is affected by GPCR mutations. Unbiased MD simulations on a msecond time scale were used to deduce the histamine-binding site in the histamine H1 receptor (Söldner et al., 2018). A complementary investigation of a D107A mutant, which has been shown experimentally to abolish ligand binding, revealed that the mutation results in a significantly weaker interaction and increased ligand dynamics.
Finally, a protocol for metadynamics simulations to predict ligand binding modes accurately has been established (Söldner et al., 2019). This strategy was tested for the histamine H1 receptor in combination with its physiological ligand histamine, the β2 adrenoceptor with its agonist adrenaline and its antagonist alprenolol. We have recently used the protocol to predict the binding site of histamine in the H2 receptor.
Most simulations currently use Gromacs 2019.4 and Plumed 2.5.3. Unbiased MD simulations achieved on average 200ns/day using 16 nodes (48 cores each) corresponding to ≈ 100 CPUh/ns on SuperMUC-ng. Metadynamics simulations perform similarly and converge within 2 μs ≈ 200,000 CPUh/run. Performance was optimized via the compute protocol, rather than by modifying software.
Timothy Clark1, Heinrich Sticht2, Peter Hildebrand3
1Computer-Chemie-Centrum and Interdisciplinary Center for Molecular Materials, Friedrich-Alexander-Universität Erlangen-Nürnberg, Nägelsbachstraße 25, 91052 Erlangen (Germany)
2Professur für Bioinformatik, Institut für Biochemie, Friedrich-Alexander-Universität Erlangen-Nürnberg, Fahrstraße 17, 91054 Erlangen (Germany)
3Institut für Medizinische Physik und Biophysik, Universität Leipzig, Härtelstraße 16-18, 04107 Leipzig (Germany)
Wifling D, Pfleger C, Kaindl J, Ibrahim P, Kling RC, Buschauer A, Gohlke H, Clark T (2019) Basal Histamine H4 Receptor Activation: Agonist Mimicry by the Diphenylalanine Motif, Chem. Eur. J.25:14613-14624. https://doi.org/10.1002/chem.201902801.
Ibrahim P, Wifling D, Clark T (2019) A universal activation index for class A GPCRs. J. Chem. Inf. Model. 59:3938-3945. https://doi.org/10.1021/acs.jcim.9b00604.
Ibrahim P, Clark T (2019) Metadynamics Simulations of Ligand Binding to GPCRs. Opin. Struct. Biol. 55:129-137. https://doi.org/10.1016/j.sbi.2019.04.002.
Söldner CA, Horn AHC, Sticht H. (2018) Binding of histamine to the H1 receptor - a molecular dynamics study. J Mol Model. 24:346. https://doi.org/10.1007/s00894-018-3873-7.
Söldner CA, Horn AHC, Sticht H. (2019) A metadynamics-based protocol for the determination of GPCR-ligand binding modes. Int. J. Mol. Sci. 20. pii: E1970. https://doi.org/10.3390/ijms20081970.
Liu X, Xu X, Hilger D, Aschauer P, Tiemann J.K., Du Y., Liu H., Hirata K., Sun X., Guixà-González R., Mathiesen J.M., Hildebrand P.W., Kobilka B.K. (2019) Structural insights into the process of GPCR-G protein complex formation. Cell, 177:1243-1251, https://doi.org/10.1016/j.cell.2019.04.021.
Prof. Dr. Timothy Clark
Computer-Chemie-Centrum and Interdisciplinary Center for Molecular Materials
Nägelsbachstraße 25, D-91052 Erlangen (Germany)
e-mail: tim.clark [@] fau.de