Structure network analysis, a graph theory-based approach, represents a cutting-edge tool to deeply understand G protein-coupled receptor (GPCR) function, which strongly relies on the communication between the extracellular and intracellular poles of their structure. We have implemented the database psnGPCRdb, which stores the structure networks (i.e., linked nodes, hubs, communities and communication pathways) computed on all updated GPCR structures in the Protein Data Bank (PDB), in their isolated states or in complex with extracellular and/or intracellular molecules. The structure communication signatures of a sub-family or family of GPCRs as well as of their activator or inhibitor molecules are stored as consensus networks. Ligand-centric networks are computed as well, with implications in drug discovery. The analysis of a training set of 480 Class-A GPCR networks from psnGPCRdb was exploited to gain insight into GPCR function. Focus was put on the Principal Component Analysis (PCA) of a number of network features. PCA analyses relied on the strength of links as the most effective variable. The PCA on the orthosteric ligand sub-network could separate the different subfamilies. As for the allosteric communication pathways, expressed in terms of matapaths made of the most recurrent nodes and links in the path pool, independent of the subfamily, for all GPCRs in the ON state and the majority of GPCRs in the OFF state the orthosteric ligand binding-site allosterically communicates with the G protein coupling regions. The PC1 separates ON and OFF states. While in the ON states the structural communication does not differentiate the subfamilies, for the OFF states, segregation into sub-families is operated by the PC2. The metapaths of the ON states are very similar in the cytosolic halves of the receptor, where they involve highly conserved residues, but diverge in the G protein depending on the G protein signalling-competence. By aggregating data according to G protein signaling, a clear separation between Gi/o and Gs is operated by the PC1 by considering the strength of either the metapath or the interface links as a variable. We also predicted the residues critical for receptor-G protein binding (hotspots) and found that the PCA based on their probability contributes as well to differentiate G protein coupling. In summary, network analysis of all non-redundant Class-A GPCR networks in the psnGPCRdb led to the following highlights. GPCR classification in sub-families depends on the ligand-binding site sub-networks. Regardless of the functional state, all Class-A GPCRs exhibit an allosteric communication between the orthosteric ligand-binding site and the G protein-coupling region. The ON and OFF GPCR states show clearly distinct communication features that depend on the presence of the G protein. In the OFF states, the structural communication depends on ligand binding that dictates the subfamily features. The G protein α-subunit homogenizes the metapaths on the receptor side, taking on the task of determining the signaling competence. Indeed, the typology of G protein signaling is dictated by the architecture of the GPCR-G protein interface, which holds the interaction hotspots, and the communication pathways on the G protein. The structure network-based models proved very effective in assigning functional state, sub-family, and G protein-mediated signalling to a test set of 128 Class-A GPCRs. psnGPCRdb represents a very powerful resource to unravel GPCR function with important implications in cell signaling and drug design.
L'analisi delle reti strutturali, un metodo basato sulla teoria dei grafi, costituisce uno strumento all'avanguardia per avanzare nella comprensione del meccanismo di funzionamento dei recettori accoppiati a proteina G (GPCR), che si basa sulla comunicazione allosterica fra i poli extracellulare e intracellulare della loro struttura. Abbiamo creato un database, psnGPCRdb, che contiene le reti strutturali (nodi, links, hubs, communities e percorsi di comunicazione) calcolate su tutte le strutture dei GPCR depositate nel Protein Data Bank (PDB), sia nei loro stati isolati che in complesso con molecole extracellulari e/o intracellulari. Le caratteristiche distintive della comunicazione strutturale di una sottofamiglia o famiglia di GPCR, assieme alle loro molecole attivatrici o inibitrici, sono depositate come reti consenso. Il database contiene anche le sotto-reti di interazione dei vari ligandi, con implicazioni nel drug discovery. L'analisi di un training-set di 480 reti di GPCR di Classe-A presenti nel psnGPCRdb è stata focalizzata sull'Analisi delle Componenti Principali (PCA) utilizzando la forza dei link quale variabile piú efficace. La PCA sulla sotto-rete dei ligandi ortosterici si è rivelata abile nel separare le diverse sottofamiglie. Per quanto riguarda i pathway di comunicazione allosterica, espressi in termini di metapaths composti dai nodi e dai links più ricorrenti nel pool dei paths, indipendentemente dalla sottofamiglia, per tutti i GPCR nello stato ON e per la maggior parte dei GPCR nello stato OFF, il sito di legame del ligando ortosterico comunica allostericamente con quello di legame delle proteine G. La PC1 separa nettamente gli stati ON e OFF. Negli stati ON, la comunicazione strutturale non distingue le sottofamiglie; al contrario, negli stati OFF, la segregazione in sottofamiglie è operata dalla PC2. I metapaths degli stati ON sono molto simili nelle porzioni citosoliche del recettore, coinvolgendo residui molto conservati, ma divergono nella proteina G in linea col signaling da essa attivato. Considerando la forza dei link del metapath o dell'interfaccia come variabile, ed aggregando i dati in base alla signaling della proteina G, si ottiene una chiara separazione tra Gi/o e Gs, operata dalla PC1. Abbiamo, inoltre, previsto i residui critici per l'interazione fra recettore e proteina G (hotspot) e scoperto che la PCA basata sulla loro probabilità contribuisce a differenziare i complessi per signaling della proteina G. In sintesi, l'analisi delle reti di tutti i GPCR non ridondanti di Classe-A nel psnGPCRdb ha condotto ai seguenti risultati. La classificazione dei GPCR in sottofamiglie dipende dalle sotto-reti d’interazione del ligando. Indipendentemente dallo stato funzionale, tutti i GPCR di Classe-A presentano una comunicazione allosterica tra il sito di legame del ligando ortosterico e la regione di accoppiamento della proteina G. Gli stati ON e OFF dei GPCR mostrano differenti caratteristiche di comunicazione che dipendono dalla presenza della proteina G. Negli stati OFF, la comunicazione strutturale dipende dal ligando, che determina le specifiche caratteristiche della sottofamiglia. La subunità-α della proteina G omogeneizza i metapath sul recettore, svolgendo il compito di determinare la tipologia di signaling. Quest’ultimo è dettato dall'architettura dell'interfaccia GPCR-proteina G, che contiene gli hotspots, e dai pathway di comunicazione sulla proteina G. I modelli basati su reti strutturali si sono dimostrati molto efficaci nell'assegnare stato funzionale, sottofamiglia e segnale mediato dalla proteina G ad un test-set di 128 GPCR di Classe-A. Il psnGPCRdb rappresenta una risorsa preziosa per comprendere la funzione dei GPCR, con importanti implicazioni nel signaling cellulare e nel design di farmaci.
Avanzamento nella comprensione del meccanismo di funzionamento dei recettori accoppiati a proteina G mediante l’analisi delle reti strutturali
GENTILE, SARA
2025
Abstract
Structure network analysis, a graph theory-based approach, represents a cutting-edge tool to deeply understand G protein-coupled receptor (GPCR) function, which strongly relies on the communication between the extracellular and intracellular poles of their structure. We have implemented the database psnGPCRdb, which stores the structure networks (i.e., linked nodes, hubs, communities and communication pathways) computed on all updated GPCR structures in the Protein Data Bank (PDB), in their isolated states or in complex with extracellular and/or intracellular molecules. The structure communication signatures of a sub-family or family of GPCRs as well as of their activator or inhibitor molecules are stored as consensus networks. Ligand-centric networks are computed as well, with implications in drug discovery. The analysis of a training set of 480 Class-A GPCR networks from psnGPCRdb was exploited to gain insight into GPCR function. Focus was put on the Principal Component Analysis (PCA) of a number of network features. PCA analyses relied on the strength of links as the most effective variable. The PCA on the orthosteric ligand sub-network could separate the different subfamilies. As for the allosteric communication pathways, expressed in terms of matapaths made of the most recurrent nodes and links in the path pool, independent of the subfamily, for all GPCRs in the ON state and the majority of GPCRs in the OFF state the orthosteric ligand binding-site allosterically communicates with the G protein coupling regions. The PC1 separates ON and OFF states. While in the ON states the structural communication does not differentiate the subfamilies, for the OFF states, segregation into sub-families is operated by the PC2. The metapaths of the ON states are very similar in the cytosolic halves of the receptor, where they involve highly conserved residues, but diverge in the G protein depending on the G protein signalling-competence. By aggregating data according to G protein signaling, a clear separation between Gi/o and Gs is operated by the PC1 by considering the strength of either the metapath or the interface links as a variable. We also predicted the residues critical for receptor-G protein binding (hotspots) and found that the PCA based on their probability contributes as well to differentiate G protein coupling. In summary, network analysis of all non-redundant Class-A GPCR networks in the psnGPCRdb led to the following highlights. GPCR classification in sub-families depends on the ligand-binding site sub-networks. Regardless of the functional state, all Class-A GPCRs exhibit an allosteric communication between the orthosteric ligand-binding site and the G protein-coupling region. The ON and OFF GPCR states show clearly distinct communication features that depend on the presence of the G protein. In the OFF states, the structural communication depends on ligand binding that dictates the subfamily features. The G protein α-subunit homogenizes the metapaths on the receptor side, taking on the task of determining the signaling competence. Indeed, the typology of G protein signaling is dictated by the architecture of the GPCR-G protein interface, which holds the interaction hotspots, and the communication pathways on the G protein. The structure network-based models proved very effective in assigning functional state, sub-family, and G protein-mediated signalling to a test set of 128 Class-A GPCRs. psnGPCRdb represents a very powerful resource to unravel GPCR function with important implications in cell signaling and drug design.File | Dimensione | Formato | |
---|---|---|---|
Gentile.pdf
embargo fino al 09/06/2028
Dimensione
9.48 MB
Formato
Adobe PDF
|
9.48 MB | Adobe PDF |
I documenti in UNITESI sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/20.500.14242/212372
URN:NBN:IT:UNIMORE-212372