Dr. Xavier Deupi i Corral

Communication with the environment is a key aspect of life. At the cellular level, this task can be carried out by proteins embedded in the cell membrane, which are exposed both to the extracellular environment and to the cytoplasm, and can thus act as 'cellular intercoms'. Due to their diversity and prevalence, G protein-coupled receptors (GPCRs) are one of the most important groups of such signalling membrane receptors. They transduce information carried by external sensory signals, such as photons and odours, and by endogenous signals, such as biogenic amines, neuropeptides, and hormones, into cellular responses. This versatility makes them essential to cell physiology and prime pharmaceutical targets, with around 30% of prescribed drugs acting through them.

While thousands of GPCR structures are now available, we still understand poorly how a receptor selects among its intracellular partners, and how differences in sequence translate into differences in function. These are the questions my group addresses, using sequence and structural analysis, molecular modelling, and molecular dynamics simulations, in close collaboration with experimental laboratories at PSI and elsewhere. Much of our work concerns opsins, the light-activated GPCRs.

Opsins are light-activated GPCRs. Their ligand is retinal, a derivative of vitamin A, covalently bound to the receptor through a protonated Schiff base. Absorption of a photon isomerises the retinal and triggers activation of the protein. The protein environment around the chromophore determines which wavelength of light each opsin absorbs. Thanks to this arrangement, evolution has been able to tune the colour sensitivity of opsins repeatedly and independently in different lineages. This makes opsins an unusually good system to study how differences in sequence translate into differences in function.

A protonated Schiff base requires a nearby negative charge, the counterion, which in most opsins is provided by conserved residues in the binding pocket, one in invertebrates and two in vertebrates. We have shown that there are other possibilities. For instance, in box jellyfish opsins the counterion has moved to a different position in the transmembrane bundle, and in coral opsins, which lack a counterion residue at any of the known sites, a chloride ion takes its place.

At present we are working on the evolution of opsins and on the fundamental mechanisms of rhodopsin activation. We are also developing machine-learning models that predict absorption maxima directly from sequence, for both animal and microbial opsins. In addition, we contribute structural modelling to the ERC Synergy project SOL, which aims to engineer bistable opsins into light-controlled tools.

• Sakai Y, Sen S, Sugihara T, et al. Coral anthozoan-specific opsins employ a novel chloride counterion for spectral tuning. eLife 2025;14:RP105451. https://doi.org/10.7554/eLife.105451
• Gruhl T, Weinert T, Rodrigues MJ, et al. Ultrafast structural changes direct the first molecular events of vision. Nature 2023;615:939–944. https://doi.org/10.1038/s41586-023-05863-6
• Tejero O, Pamula F, Koyanagi M, et al. Active state structures of a bistable visual opsin bound to G proteins. Nature Communications 2024;15:8928. https://doi.org/10.1038/s41467-024-53208-2
• Deupi X, Edwards P, Singhal A, et al. Stabilized G protein binding site in the structure of constitutively active metarhodopsin-II. PNAS 2012;109:119–124. https://doi.org/10.1073/pnas.1114089108

Beyond opsins, we also work on receptors activated by diffusible ligands. Upon activation, every GPCR recruits intracellular partners such as G proteins, arrestins, and kinases. Which partners it recruits, and how efficiently, determines the cellular response, and partly explains why two drugs acting at the same receptor can produce different effects. This selectivity is encoded in the sequence of the receptor and in its conformational behaviour, but it remains difficult to predict.

For instance, we have studied the role of the phosphorylated C-terminus in the recruitment of G proteins and arrestins, the activation of the chemokine receptor CCR5 by its agonist, the signalling profiles of natural variants of the melatonin receptors MT1 and MT2, and cyclic peptides that stabilise Gq/11 heterotrimers.

• Mühle J, Alenfelder J, Rodrigues MJ, et al. Cyclic peptide inhibitors function as molecular glues to stabilize Gq/11 heterotrimers. PNAS 2025;122:e2418398122. https://doi.org/10.1073/pnas.2418398122
• Venkatakrishnan AJ, Deupi X, Lebon G, et al. Diverse activation pathways in class A GPCRs converge near the G-protein-coupling region. Nature 2016;536:484–487. https://doi.org/10.1038/nature19107
• Isogai S, Deupi X, Opitz C, et al. Backbone NMR reveals allosteric signal transduction networks in the β1-adrenergic receptor. Nature 2016;530:237–241. https://doi.org/10.1038/nature16577
• Venkatakrishnan AJ, Deupi X, Lebon G, Tate CG, Schertler GF, Babu MM. Molecular signatures of G-protein-coupled receptors. Nature 2013;494:185–194. https://doi.org/10.1038/nature11896
• Deupi X, Kobilka BK. Energy landscapes as a tool to integrate GPCR structure, dynamics, and function. Physiology 2010;25:293–303. https://doi.org/10.1152/physiol.00002.2010

Together with PSI's Center for Radiopharmaceutical Sciences, we apply a similar approach to model the binding mode of radioligands, molecules that carry a radionuclide to a chosen target for cancer imaging or therapy. A useful radioligand has to bind its target with high affinity and selectivity, and at the same time reach the target and clear from the rest of the body at appropriate rates. Structural modelling allows us to understand how these molecules are recognised by their targets, and to suggest modifications that improve them. For instance, we have contributed to the development of radiopeptides directed against the urokinase receptor uPAR and against ACE2.

• Benoit RM, Wang J, Beyer D, et al. Development and structure-guided characterization of a novel ACE2-binding macrocyclic peptide. Journal of Structural Biology: X 2026;13:100145. https://doi.org/10.1016/j.yjsbx.2026.100145
• Vaccarin C, Beyer D, Schmid JV, et al. Optimizing uPAR-targeting radiopeptides for improved tissue distribution: progress towards radionuclide therapy. European Journal of Nuclear Medicine and Molecular Imaging 2026;53:2502–2517. https://doi.org/10.1007/s00259-025-07602-7
• Beyer D, Vaccarin C, Deupi X, et al. A tool for nuclear imaging of the SARS-CoV-2 entry receptor: molecular model and preclinical development of ACE2-selective radiopeptides. EJNMMI Research 2023;13:32. https://doi.org/10.1186/s13550-023-00979-2

I am a senior scientist at the Paul Scherrer Institute and a group leader at the Swiss Institute of Bioinformatics. I have a degree in Organic Chemistry from the Institut Químic de Sarrià in Barcelona. I moved to computational biology for my PhD at the Universitat Autònoma de Barcelona, where I modelled GPCRs at a time when bovine rhodopsin was the only available structure. A postdoctoral position in Brian Kobilka's laboratory at Stanford put me in daily contact with the experiments I had been modelling, which has shaped how I work since. I then returned to Barcelona as a research scientist, and in 2010 I left that position for PSI, to work alongside structural biologists at an institute with a synchrotron and a free-electron laser.

Curriculum vitae (PDF)