Predictive modelling of GPCR druggable allosteric sites

Información del proyecto

MIDAS

Identificador del acuerdo de subvención: 748830

DOI

10.3030/748830

Proyecto finalizado el 19 Noviembre 2019

Fecha de la firma de la CE 22 Marzo 2017

Fecha de inicio 1 Noviembre 2017

Fecha de finalización 31 Octubre 2019

Financiado con arreglo a

• EXCELLENT SCIENCE - Marie Skłodowska-Curie Actions

Coste total

€ 195 454,80

Aportación de la UE

€ 195 454,80

Coordinado por

THE QUEEN'S UNIVERSITY OF BELFAST
United Kingdom

Periodic Reporting for period 1 - MIDAS (Predictive modelling of GPCR druggable allosteric sites)

Período documentado: 2017-11-01 hasta 2019-10-31

Resumen del contexto y de los objetivos generales del proyecto

G protein-coupled receptors (GPCRs) are the most successful druggable targets in the human genome. Up to 60 % of prescription drugs in the European Union act by modulating biological pathways involving GPCRs. GPCRs can be targeted by small molecules acting on allosteric binding sites (ABSs), i.e. pockets spatially distinct from highly conserved orthosteric sites that host endogenous modulators. The design of allosteric modulators (AMs), therefore, represents a viable strategy to achieve more selective drugs with fewer side effects. The discovery of ABSs targeted by AMs in GPCRs is a challenging task that is largely achieved through cost-intensive and time-consuming high-throughput screening campaigns. The recent release of experimental three-dimensional structures of GPCRs in complex with AMs, mainly solved by X-ray crystallography, has opened new opportunities to develop structure-based computer-aided strategies to identify ABSs and screen for novel AMs. MIDAS aimed at developing the first cost-effective and rapid computational methodology based on probe mapping MD simulations for the identification of ABSs in GPCRs by pursuing four objectives:
1) Identification of suitable probes for MD-based co-solvent mapping
2) Development of a computational procedure to identify ABSs within the helical bundle
3) Development of a computational procedure to identify ABSs between helices and lipids
4) Prospective validation of the computational procedures via site-directed mutagenesis experiments

Trabajo realizado desde el comienzo del proyecto hasta el final del período abarcado por el informe y los principales resultados hasta la fecha

The Experienced Researcher (ER) has developed a novel computational methodology to locate ABSs in GPCRs based on molecular dynamics (MD) simulations. The methodology, combining enhanced sampling MD simulations with fragment-based drug design (FBDD), has been successfully validated in four retrospective case studies representing all known scenarios for GPCR AM binding:
a) the M2 muscarinic receptor in ternary complex with the agonist iperoxo and the extra-cellular positive allosteric modulator (PAM) LY2119620;
b) the Beta2 adrenergic receptor in complex with the intra-cellular peptide-like negative allosteric modulator (NAM) Cmp-15;
c) the CCR2 chemokine receptor in complex with the intra-cellular small molecule NAM CCR2-RA-[R];
d) the purinergic P2Y1 receptor in complex with the interface NAM BPTU.

MIDAS technology has been applied in conjunction with the conventional co-solvent mapping approach and its performance compared with a novel ad hoc developed strategy that implements in silico FBDD concepts. The novel, alternative approach applies chemo-informatics analyses to identify “privileged” fragments to use in probe mapping MD simulations. After MIDAS successful retrospective validation in the four selected case studies, the ER has applied the methodology prospectively in two case studies by predicting the location of the ABSs in class A bio-aminergic GPCRs – namely the Dopamine D2 and serotonin 5HT2C receptors – with available X-ray structures and known AMs. The ER has tested the computational predictions experimentally through secondment in Dr. Peter McCormick’s lab at Queen Mary University London by performing site-directed mutagenesis experiments and cell-based fluorescent binding assays.

Avances que van más allá del estado de la técnica e impacto potencial esperado (incluida la repercusión socioeconómica y las implicaciones sociales más amplias del proyecto hasta la fecha)

MIDAS represents the first computational methodology for the identification of ABSs in GPCRs. Notably, the approach proved to be successful in all known scenarios, including the most challenging case where the AM binds at the interface between the receptor and the cellular membrane. MIDAS relies upon a fully automatized pipeline that spans from fragment identification to result analysis. To this aim, ad hoc protocols for membrane-protein systems probe-mapping MD simulation have been developed and thoroughly tested. In particular, the combination of such protocols and the use of fragments instead of standard co-solvents enables to overcome current limitations of co-solvent MD-mapping such as probe non-specific binding and/or low sampling, protein denaturation, and membrane distortion. MIDAS technology outperforms any other existing strategy based on co-solvent mapping and is able to identify ABSs in simulation times as short as 20 ns. FBDD is nowadays being intensively applied in drug discovery through cost-intensive NMR campaigns, thus being suitable mostly for soluble drug targets and not yet routinely applicable to membrane proteins like GPCRs. As the first approach computationally enabling FBDD for GPCRs, MIDAS has the potential to foster the development of novel health technologies for the discovery of safer drugs targeting membrane proteins. To this aim, MIDAS has already achieved a significant breakthrough by identifying a previously unknown ABS in the 5HT2CR, thus paving the way for the development of novel safer anti-obesity drugs acting via a different mechanism with respect to the clinically approved drug locanserin.

Predictive modeling of GPCR druggable allosteric sites