Conformational mapping of GPCR activation: dynamic allosteric site discovery in V2R through MD-MSM and mutual information analysis

Conformational mapping of GPCR activation: dynamic allosteric site discovery in V2R through MD-MSM and mutual information analysis | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Conformational mapping of GPCR activation: dynamic allosteric site discovery in V2R through MD-MSM and mutual information analysis Shaoyong Lu, Xin Qiao, Chunhao Zhu, Xiaobing Lan, Mingyu Li, Nuan Li, and 3 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-6427090/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Allostery governs‌ the functional dynamics of proteins by regulating their conformational transitions. ‌The development of allosteric modulators has emerged as a promising therapeutic strategy‌, leveraging their superior target specificity ‌and reduced off-target effects compared to orthosteric compounds‌. ‌A critical barrier in this field remains‌ the identification of dynamic allosteric sites, ‌which are often undetectable in conventional structural analyses due to their transient nature‌. ‌To address this challenge,‌ we established ‌an integrative computational framework‌ combining molecular dynamics (MD), Markov state modeling (MSM), and mutual information (MI) analysis ‌to probe‌ dynamic allosteric sites ‌in the class A G protein-coupled receptor (GPCR) prototype, vasopressin V2 receptor (V2R)‌. ‌Through‌ multi-replica MD simulations, ‌we reconstructed‌ the receptor's conformational landscape, ‌which was statistically refined‌ via MSM ‌to resolve‌ equilibrium populations ‌and transition kinetics‌. ‌Key mechanistic features‌ of activation-related structural motifs ‌were quantitatively characterized‌. ‌Candidate allosteric sites were systematically ranked‌ through MI-driven residue interaction network analysis, ‌prioritizing‌ pharmacologically targetable regions. ‌This strategy revealed‌ a novel dynamic allosteric site ‌on the V2R intracellular interface‌, ‌whose functional relevance was confirmed through‌ structure-guided mutagenesis ‌and‌ BRET-based signaling assays. ‌Our findings‌ not only ‌elucidate the allosteric activation mechanism of V2R at atomic resolution‌ but also ‌establish a conformation-aware platform‌ for ‌rational discovery of dynamic binding pockets‌, ‌providing a transformative approach for‌ GPCR-targeted drug discovery. Biological sciences/Biophysics/Computational biophysics Biological sciences/Computational biology and bioinformatics/Protein function predictions Physical sciences/Chemistry/Theoretical chemistry/Molecular dynamics Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 Figure 9 Figure 10 Introduction Allostery, often referred to as ‘the second secret of life’, occurs in various biological macromolecules, including dynamic proteins, DNA, and RNA polymers 1 – 3 . It plays a critical role in regulating conformational changes and functional modulation of biomolecules 4 – 7 . Typically, allosteric perturbations occur at sites distant from orthosteric sites on proteins. Consequently, allosteric modulators do not compete with endogenous substrates that bind to orthosteric sites but instead finely tune receptor affinity in a highly predictable manner, thereby enhancing safety 8 – 10 . Moreover, allosteric modulators can achieve better selective targeting of different receptor subfamilies because allosteric sites are less conserved compared to orthosteric sites, thereby reducing the risk of side effects 11 – 13 . Despite these advantages, many clinically important targets lack suitable allosteric sites for the design of therapeutic molecules 14 – 16 . Currently, certain transcription factors, small GTPases, and phosphatases are considered as ‘undruggable proteins’ due to the absence of appropriate small molecule binding sites 17 , 18 . For example, the transcription factor p53, a tumor suppressor protein, exerts its effects by interacting with mouse double minute 2 homolog (MDM2). Drug design targeting p53 remains challenging because its protein-protein interaction (PPI) interface is broad and flat, lacking druggable binding sites 19 . However, allosteric modulators could inhibit PPIs by targeting allosteric sites distant from their interfaces. Notably, dynamic allosteric sites, which are only visible in low-populated intermediate states, enrich the diversity of allosteric sites and expand the range of potential drug targets 20 – 22 . Dynamic allosteric sites have addressed another previously undruggable challenge, as exemplified by the small GTPase KRAS, one of the most frequently mutated oncogenes 23 . The two currently approved drugs, Sotorasib (AMG-510) and Adagrasib (MRTX849), are covalent inhibitors targeting the residue Cys12 of the KRAS G12C mutant. Both drugs bind to a cryptic allosteric site, which is not clearly visible in other published KRAS structures, although in some cases, a groove can be observed. However, this site was discovered accidentally during the experimental process. Generally, most allosteric sites are identified based on available bound molecules and involve various exhaustive and tedious methods, such as crystallography, mutagenesis experiments, and high-throughput screening experiments 24 , 25 . Identification of dynamic allosteric sites in the absence of bound molecules would significantly accelerate the drug development process. Furthermore, since cryptic allosteric sites cannot be observed in static structures, their identification remains a significant challenge 26 – 28 . To date, there is a lack of effective methods for locating dynamic allosteric sites. While structural approaches provide valuable insights, their static nature offers limited information about dynamics underling receptor activation 22 . In contrast, computational techniques, particularly those based on molecular dynamics (MD) simulations, have advanced the identification of novel dynamic allosteric sites without bound molecules. MD simulations effectively offer atomic-level insights into the spatial and temporal organization of biomolecular conformations, enabling the uncovering of interactions between distant functional regions 29 – 32 . Additionally, Markov state models (MSMs), constructed from extensive MD simulations, provide a simplified representation of dynamic biomolecular ensembles. A key advantage of MSMs is their ability to identify dynamic allosteric sites that often exist in subtle conformational states 33 . Notably, the residue-based mutual information (MI) matrix is an approach for examining the interdependencies between residues in a protein. It quantifies statistical correlations between pairs of residues and proves valuable in identifying cooperative interactions and uncovering cryptic allosteric sites 34 , 35 . As a typical class A G protein-coupled receptor (GPCR), the vasopressin V2 receptor (V2R) regulates fluid homeostasis by binding to the endogenous substrate arginine vasopressin (AVP), which stimulates the activation of downstream Gs proteins 36 – 38 . In the cryo-EM structure, the initial six residues in AVP form a cyclic insertion deep within the orthosteric site. Specifically, V2R exhibits a kink in the L312 7.40 −A314 7.42 region (superscripts indicate the Ballesteros–Weinstein numbering for GPCR residues), a polar network among S127 3.40 , T218 5.51 , and Y280 6.44 , and weak interactions between the intracellular ends of TM3 and TM6 39 (Fig. 1 ). Current research has extensively explored the activation mechanism of GPCRs, identifying a common activation pathway that directly connects the ligand-binding pocket to the G protein-coupling region. This pathway integrates key conserved motifs such as CWxP, PIF, Na + pocket, DRY, and NPxxY, elucidating how residue contact rearrangements drive conformational changes in transmembrane helices 40 . Although some studies have resolved the cryo-EM structures of the V2R receptor bound to the agonist AVP, the lack of the apo or inactive state structures has constrained the scope of these investigations. Consequently, these studies relied on comparisons between the active structure of V2R and the inactive conformation of the highly homologous oxytocin receptor (OTR), leaving the activation mechanism of the V2R receptor insufficiently explored 39 , 41 . Additionally, suitable allosteric sites for drug design remain elusive for V2R 42 . Here, we developed an integrated computational and experimental framework that comprehensively delineates the activation mechanism of the receptor and identifies a previously unexplored dynamic allosteric site on V2R. Initially, we performed multiple replica MD simulations coupled with MSMs to explore the activation transition pathway of the receptor 43 – 46 . Subsequent analyses compared representative conformations to evaluate changes in key residues and motifs during activation. Following this, we devised an algorithmic framework based on residue side-chain fluctuations to analyze the allosteric network among the predicted dynamic sites. Ultimately, a promising dynamic allosteric site was identified within the intracellular region of the receptor through site-directed mutagenesis experiments. Results The integrated computational and experimental framework for identifying dynamic allosteric sites To initiate the computational workflow (Fig. 2 ), we established three states of V2R for five independent runs of each 1 us MD simulations: the apo inactive state, the AVP-bound intermediate active state, and both AVP- and Gs-bound fully active state (PDB ID: 7KH0). This timescale has been demonstrated to effectively capture the ensemble of dynamic conformations of receptors 47 , 48 . Subsequently, the simulation trajectories were dimensionally reduced using characteristic features of receptor activation. MSMs were then constructed, dividing the system into several metastates and analyzing the ensemble’s dynamic properties and equilibrium distribution 49 . Next, representative conformations of each metastate were extracted from the simulated trajectories, and the conformational changes of key residues and motifs during receptor activation were analyzed 48 , 50 . After gaining a comprehensive understanding of the ensemble, the fpocket program was used to predict potential allosteric sites in both the intermediate active and fully active states of the receptor. Thereafter, based on the MI matrix of residue side-chain fluctuations, the potential sites were ranked according to eigenvector centrality, a method used to assess node importance within networks, particularly suited for those with complex weight distributions 51 . Finally, high-potential allosteric sites were selected for experimental validation via site-directed mutagenesis and bioluminescence resonance energy transfer (BRET) assays 22 , 28 , 47 . Exploring the activation transition pathway of the receptor among different metastates We conducted 1 µs × 5 independent runs for three distinct states of V2R using random initial velocities, yielding a cumulative trajectory length of 15 µs. The apo inactive state’s initial structure, generated via Swiss-Model, exhibited near-identical topology to the AlphaFold2-predicted structure with the highest pLDDT score, as evidenced by a low root-mean-square deviation (RMSD) of 0.93 Å between their atomic coordinates (Figure S1 ). In contrast, the receptor portion of the fully active state’s initial structure was differed from the apo inactive state’s receptor structure in four key respects (Figure S2): bending at the extracellular part of TM7, absence of pronounced outward rotation in the CWxP motif, inward displacement at the intracellular part of TM5, and outward displacement at the intracellular part of TM6. Previous investigations into the activation mechanism of V2R compared its active structure with the inactive structure of the oxytocin receptor (OTR), which shares high homology. These studies identified a bend at the extracellular end of TM7, a relatively small degree of outward rotation at the intracellular end of TM6, and a mutation in the PIF motif 39 , 41 . However, their analysis of the entire activation pathway lacked comprehensiveness, failing to detect the subtle outward rotation of the CWxP motif during activation, the inward movement of the intracellular end of TM5, and the nonpolar mutation of the residue V266 6.30 . Following MD simulations, we calculated the RMSD for Cα atoms in the receptor portions across three distinct states relative to their initial conformations (Figure S3A). The receptor system reached equilibrium after 100 ns of the simulation time. Notably, the AVP-bound intermediate active state exhibited the highest RMSD values (5.90 ± 0.89 Å) (Figure S3A), suggesting substantial ligand-induced structural reorganization. Comparative analysis revealed significantly lower deviations in the apo inactive state (4.48 ± 0.63 Å) and the fully active ternary complex containing both AVP and Gs protein (3.94 ± 0.37 Å) (Figure S3A), indicating more constrained conformational flexibility in these states. This differential RMSD pattern suggested that ligand binding selectively enhanced conformational plasticity in the intermediate active state while maintaining relative stability in other conformations during simulations. To characterize regional mobility variations, we calculated per-residue root-mean-square fluctuation of backbone heavy atoms across receptor segments in all three states 52 . Significant fluctuation differences were observed in intracellular loop (ICL) regions 2 and 3 (Figure S3B). These loops likely participate in Gs protein interactions, potentially stabilizing their conformation. The inherent flexibility of ICL2 and ICL3 in GPCRs, typically results in structural ambiguity, as evidenced by poorly resolved electron density maps that complicate molecular modeling. This technical challenge has historically limited investigation of ICL3-G protein functional interactions 39 , 41 . Our study addresses this gap through computational reconstruction of missing ICL segments and systematic characterization of loop-Gαs protein interplay, providing mechanistic insights for future research. Detailed interaction analysis identified key polar and hydrophobic contacts between both intracellular loops and the Gs protein (Figures S4-S5), with specific residue participation documented. Class A GPCRs feature seven transmembrane α-helices (TM1-TM7), with agonist-induced conformational changes primarily involving outward displacement of TM6 and inward movement of TM7. These structural rearrangements facilitate the formation of a G protein-binding pocket 53 – 55 . To characterize the receptor activation dynamics, we employed two distinct order parameters: 1) the Cα-Cα distance between V270 6.34 (TM6) and R137 3.50 (TM3) quantifying TM6 outward displacement 50 , and 2) the Cα-Cα separation between P322 7.50 (TM7) and L81 2.46 (TM2) measuring TM7 inward movement 56 . Using MD trajectories, we constructed a two-dimensional free energy landscape by projecting these structural parameters (Fig. 3 A). The apo inactive state, characterized by initial distances of 7.1 Å (V270 6.34 -R137 3.50 ) and 10.1 Å (P322 7.50 -L81 2.46 ), occupied a distinct energy basin (5.5–6.7 Å and 8.6–10.3 Å, respectively). Activation involved sequential energy barrier crossings: first transitioning to an intermediate active state (7.4–12.1 Å and 7.6–9.9 Å), then progressing to the fully active state (13.7–15.6 Å and 7.4–8.4 Å). Notably, TM6 displacement showed a progressive increase in V270 6.34 -R137 3.50 distance, while TM7 movement reduced the P322 7.50 -L81 2.46 separation. This trajectory analysis delineates the complete activation pathway of V2R from the apo inactive through an intermediate active to the fully active states. Convergence validation through subsampling and multiple MD replicates 47 , 48 (Figures S6-S7) confirmed sufficient sampling of the activation landscape. The observed energy barriers and metastable states align with established GPCR activation mechanisms, demonstrating the system's capacity to capture essential conformational transitions. To investigate the kinetic properties and equilibrium distribution of the ensemble, we established MSMs based on key activation parameters. From a statistical mechanics perspective, MSMs enabled systematic characterization of conformational distributions within the equilibrium ensemble. This methodology not only permitted precise identification of metastable states but also enabled quantitative determination of kinetic properties, including transition timescales between states 57 . The modeling workflow proceeded as follows: first, the conformational ensemble was discretized into 300 microstates, with Markovian behavior verified through implied timescale analysis (Figure S8A). Subsequently, the Robust Perron Cluster Analysis (PCCA+) algorithm was implemented to cluster these microstates into three distinct metastates. Transition path theory (TPT) was then employed to calculate inter-metastate transition timescales. To validate model consistency, we conducted Chapman-Kolmogorov tests comparing predicted and observed transition probabilities between metastates (Figure S8B). The conformational space was partitioned into three regions, with blue, yellow, and pink designating the apo inactive, intermediate active, and fully active states, respectively. MSMs-derived conformational distribution showed strong agreement with the free energy landscape (Fig. 3 B), with equilibrium populations of approximately 30% for each metastate due to minimal energy barrier differences. Notably, the fully active state demonstrated the highest population (39.4%), consistent with previous reports of receptor basal activity 58 , 59 , thereby validating the MSM construction. Transition kinetics analysis revealed differential timescales between activation and deactivation pathways: The apo inactive→fully active transition (96.72 µs) occurred faster than the reverse process (108.24 µs) (Fig. 3 C), aligning with the observed population bias toward the fully active state. Furthermore, transitions from both apo inactive (29.23 µs) and fully active (22.21 µs) states to the intermediate active state exhibited accelerated kinetics compared to their reverse transitions (intermediate active→apo inactive: 70.01 µs; intermediate active→fully active: 40.77 µs). This kinetic hierarchy identifies the intermediate active→fully active transition as the rate-limiting step for receptor activation, while the intermediate active→apo inactive transition governs deactivation kinetics. Molecular insights into key steps of the specific activation pathway To investigate conformational changes during receptor activation, we extracted representative conformations of three metastates from MSMs-derived conformational landscapes and systematically analyzed dynamic alterations in key residues and structural motifs. Following AVP ligand binding, activation signals propagated through conserved motifs to the receptor intracellular end, facilitating Gs protein interaction 60 . Notably, V2R exhibited distinct activation-related conformational changes compared to other class A GPCRs. Structural analysis revealed that the extracellular ends of TM6 and TM7 cooperatively formed the AVP binding site. Specifically, Tyr2 of AVP established a hydrogen bond with L312 7.40 (Figure S9), inducing characteristic bending in the L312 7.40 −A314 7.42 segment of TM7 41 (Fig. 4 A). Conformational sampling demonstrated increased prevalence of the smaller angle conformations among P306 7.34 , A314 7.42 , and Y325 7.53 in the fully active state (Fig. 4 B), corroborating TM7 bending observed in cryo-EM structural comparisons. This contrasts with other class A GPCRs (e.g., NTSR1 and β2AR) 61 , 62 where agonist binding fails to induce significant TM7 bending. Concurrent with these changes, residues F287 6.51 , F288 6.52 , and Q291 6.55 underwent inward rotation (Fig. 4 C), facilitating binding site formation. Signal transduction to the adjacent CWxP motif revealed divergent mechanisms: while β2AR and α2AR receptors exhibit pronounced outward rotation of W 6 . 48 followed by F 6 . 44 movement 63 , 64 . V2R demonstrated minimal dihedral variation among S127 3.40 , Y280 6.44 , W284 6.48 , and V281 6.45 across the three states (Fig. 4 D). This indicates attenuated outward rotation of Y280 6.44 /W284 6.48 during activation 39 (Fig. 4 E), differing from conventional GPCR activation paradigms. Comparative analysis with OTR structures revealed partial outward rotation of W 6.48 /Y 6.44 in activated V2R 39 , 41 . This discrepancy may arise from evolutionary substitutions at positions 280 6.44 (F→Y) and 127 3.40 (I→S), establishing a stabilizing polar network between S127 3.40 , T218 5.51 , and Y280 6.44 (Fig. 5 A). This network likely restricts TM6 mobility, as evidenced by limited angular changes among S127 3.40 , P217 5.50 , and Y280 6.44 during activation (Fig. 5 B). Consistently, V2R exhibits reduced intracellular TM6 displacement (V 6 . 34 -R 3 . 50 distance: 14.5 Å, PDB ID: 7KH0) compared to α2AR (15.9 Å, PDB ID: 6K41) 64 . Upon transmission of the activation signal to the central V2R region, the hydrophobic interaction triad comprising I130 3.43 , I276 6.40 , and V277 6.41 became destabilized 47 , facilitating outward displacement of TM6 (Figure S10A, B). Concurrently, TM3 and TM7 underwent mutual approach 40 , leading to structural collapse of the Na + binding pocket formed by D85 2.50 , S126 3.39 , N317 7.45 and N321 7.49 (Figure S10C, D). These conformational rearrangements align with canonical activation patterns observed across class A GPCRs. Progressive signal propagation induced distinct helical motions: inward displacement of TM3, outward movement of TM6, and inward repositioning of TM7 (Figure S11). Notably, a novel discovery emerged with TM5 exhibiting inward movement (Fig. 5 C), narrowing its spatial relationship with TM6 and TM7. This was quantitatively corroborated by reduced interatomic distance between Q225 5.58 and Y325 7.53 in both the intermediate active and fully active states (Fig. 5 D). Comparative analysis revealed that TM5 intracellular domain dynamics exhibit receptor-specific variability within class A GPCRs: AT1R systems display outward movement, while CB1R systems show inward displacement 65 – 68 . A hallmark of V2R architecture lies in the differential regulation of TM6 dynamics (Figs. 5 E, F). Specifically, the polar network at the middle interface of TM6 constrained its outward displacement amplitude, while the intracellular mutation at V266 6.30 enhanced TM6 mobility, correlating with elevated basal receptor activity 58 , 59 . This contrasts with conserved class A GPCR features where the acidic residue (typically D 6 . 30 ) forms stabilizing ionic interactions with R 3 . 50 in inactive states. Integration of these findings enabled comprehensive mapping of activation-associated signal transduction pathways, highlighting critical structural motifs including CWxP, PSY, NPxxY, and DRH (Fig. 6 , red highlights). The collective data elucidate unique V2R activation mechanics characterized by: (1) constrained TM6 middle movement, (2) facilitated TM6 intracellular displacement, and (3) TM5 inward trajectory enhancing proximity to TM6/TM7. MI analysis uncovers allosteric communication networks among potential sites To investigate dynamic correlations between residues, we first computed the absolute deviation of each residue's side chain conformation across the simulation trajectory. Subsequent construction of MSMs employed Euclidean distance-based clustering of side chain deviation vectors, with model validation performed through implied timescale analysis (Figure S12). Each microstate in the MSMs was represented by a quintenary vector encoding side chain fluctuation patterns, where vector elements were visually encoded using a gradient scale (darker shades indicating larger deviations, lighter tones representing smaller variations) (Fig. 7 A). We then quantified residue-residue correlations through MI analysis, generating a comprehensive MI matrix (Fig. 7 B). MI provides a robust measure of statistical dependence between random variables, quantifying the information shared between two systems. In information-theoretic terms, higher MI values indicate stronger associations, while lower values suggest weaker relationships. This method is particularly advantageous for detecting nonlinear correlations and automatically excludes invariant residues (zero entropy) from contributing to MI calculations 69 – 71 . Our implementation extends MI's proven utility in bioinformatics and machine learning to protein dynamics analysis. Potential allosteric sites were identified using fpocket 72 across representative conformations of both the intermediate active and fully active states, yielding 13 candidate sites. The residue-level MI matrix was subsequently transformed into a site-specific MI matrix (Fig. 7 C), enabling network construction where MI values served as edge weights. To identify key regulatory hubs, we computed eigenvector centrality measures from the MI-derived weighted adjacency matrix. This network metric reveals globally influential nodes that maintain strong connections with other high-MI sites 73 , 74 , effectively highlighting sites with maximal impact on allosteric communication pathways during protein conformational changes. Identification of a dynamic allosteric site on the activation pathway We systematically ranked predicted potential sites using eigenvector centrality metrics, with the top five sites visualized as spherical structures (Fig. 8 A, B). MI calculations required careful consideration of continuous variable discretization, as the selection of interval ranges directly impacts probability distribution estimations and subsequent computational outcomes 75 . To ensure robustness, we performed comprehensive analyses by discretizing side-chain deviation values into varying interval categories (6–10 bins). While interval selection induced fluctuations in site rankings, the five highest-ranked sites consistently maintained their positions (1–5) across different discretization schemes (Figure S13). Based on the representative five-interval discretization results, we constructed an allosteric network through MI analysis of the top-ranked sites, with each site represented as distinct spherical elements. Notably, the fifth site (Site 5) corresponded to the orthosteric binding site, while the fourth site (Site 4) exhibited eigenvector centrality values comparable to those of Site 5. Comprehensive analysis of the allosteric network revealed that the top five sites, encompassing the orthosteric site, maintain intricate allosteric interconnectivity through dynamic coupling (Fig. 8 C). Structural characterization demonstrated that Site 1 and Site 3 occupy adjacent positions within the interhelical crevice formed by mid-regions of TM6 and TM7 (Fig. 8 D, F; 9 A, C; S14A). While Site 1 faces the lipid bilayer interface, Site 3 resides within the receptor’s transmembrane core. Computational characterization using fpocket quantified distinct physicochemical properties (Table S1 ): Site 1 (volume = 193 Å 3 , SASA = 59 Å 2 , α-sphere density = 1.9 Å −3 ) and Site 3 (volume = 115 Å 3 , SASA = 24 Å 2 , α-sphere density = 1.7 Å −3 ). These metrics collectively indicate suboptimal druggability for both sites, characterized by limited surface accessibility and loose packing geometry. Notably, Site 3 demonstrates reduced hydrophobic character, evidenced by an apolar α-sphere proportion of 0.16. Site 2 occupies a lipid-exposed niche at mid-regions of TM3 and TM5 (Fig. 8 E, 9 B, S14B). Its formation correlates with intracellular TM5 inward displacement during activation, which reduces the interhelical distance between TM5 and TM3 side chains. This site emerges exclusively in intermediate conformational states of the ensemble and remains undetected in the initial active-state structure. Despite moderate dimensions (SASA = 47 Å 2 , volume = 201 Å 3 ) and packing density (α-sphere = 2.0 Å −3 ), its structural plasticity and limited accessibility compromise therapeutic targeting potential. In contrast, Site 4 demonstrates unique structural and dynamic features at the intracellular TM5-TM6 interface (Fig. 8 G, 9 D, S14C, D). This pocket forms through coordinated helical movements during activation: TM6 intracellular shift combined with TM5 inward displacement toward TM6. Comparative conformational analysis revealed V2R-specific activation mechanics: unlike canonical class A GPCRs, which exhibit pronounced TM6 outward movement, V2R activation features attenuated TM6 displacement coupled with TM5 inward migration. This distinctive motion pattern creates a receptor-specific dynamic pocket absent in the initial active structure, where excessive TM6 outward displacement (T269 6.33 , T273 6.37 , V277 6.41 side chain repositioning) obstructs cavity formation. Notably, Site 4 presents superior druggability metrics: substantial volume (515 Å 3 ), enhanced surface accessibility (SASA = 139 Å 2 ), and favorable hydrophobicity (apolar α-sphere proportion = 0.75) (Table S1 ). Its compact architecture (α-sphere density = 4.1 Å −3 ) and high fpocket druggability score (0.855) suggest strong therapeutic potential. Trajectory analysis confirmed dynamic persistence of this pocket throughout conformational sampling (Figure S15A), supporting its identification as a viable allosteric site. Notably, among the 13 potential sites identified, two particularly significant sites (Site 7 and Site 10) emerged, underscoring the efficacy of our methodology in detecting allosteric sites. However, given our primary objective to discover novel allosteric sites, these two well-characterized locations were excluded from further investigation. Site 7 occupies the intracellular cavity formed by TM2, TM3, and TM5–TM7, spatially overlapping with the canonical G protein binding domain (Figure S16A, C). This evolutionarily conserved site across class A GPCRs has been extensively studied, with numerous allosteric modulators already developed and characterized. Due to the substantial existing research foundation, no additional exploration of this site was pursued. Site 10 localizes to the intracellular interface between TM3, TM4, and ICL2 (Figure S16B, D; S17A, B). Its structural formation relies on ICL2 conformational rearrangement coupled with intracellular TM3 displacement-features absent in the initial active-state structure due to ICL2 disorder and suboptimal TM3 positioning. Mechanistic studies suggest that allosteric modulation at this site could regulate ICL2-G protein interactions, enabling dual Gq/Gs signaling crosstalk as previously demonstrated in β2AR and GPR40 76, 77 (Figure S17C, D). Given the comprehensive prior investigations of this site, it was similarly excluded. Through systematic integration of empirical knowledge and MD simulations, we ultimately selected the dynamic, system-specific allosteric Site 4 as our primary candidate for experimental validation. Experimental validation of the predicted dynamic allosteric site As the activated V2R predominantly exerts its physiological effects through Gs protein-mediated downstream signaling, we conducted alanine-scanning mutagenesis coupled with BRET assays to validate the functional role of Site 4 as a potential allosteric site of Gs protein signaling. The TRUPATH biosensor system 78 , 79 , utilizing BRET2 technology, was implemented to monitor receptor-G protein coupling dynamics. This methodology incorporates a NanoLuc energy donor fused to the Gα subunit and a GFP2 acceptor tagged to the Gγ subunit's N-terminus. Upon V2R activation, the heterotrimeric G protein complex (Gαβγ) undergoes conformational dissociation, releasing the Gα subunit to interact with the activated receptor. This structural reorganization reduces energy transfer efficiency between the donor-acceptor pair, reflecting a transition from high-affinity to low-affinity interaction states. The resultant BRET signal attenuation provides quantitative measurement of receptor-G protein coupling efficiency and dynamic interactions. Site 4 comprises 10 residues, including Q225 5.58 , V226 5.59 , I228 5.61 , F229 5.62 , I232 5.65 , V266 6.30 , T269 6.33 , V270 6.34 , T273 6.37 , and L274 6.38 (Figure S15B). Systematic alanine substitution of each residue generated 10 distinct mutants, enabling comprehensive evaluation of their functional impact on V2R-mediated Gs protein signaling efficacy. Gene expression analysis (Figure S18) demonstrated comparable expression levels between all mutants and wild-type receptors, indicating the mutations did not affect receptor biosynthesis or stability. Subsequent BRET assays under ligand-free conditions (Fig. 10 A-C) revealed that only the Q225A mutant exhibited substantial impairment (~ 50% reduction) in basal Gs protein activity, suggesting its critical role in maintaining constitutive receptor activity. The remaining nine mutants showed no significant alterations in this unliganded state. Ligand-dependent activation assays with AVP yielded distinct patterns (Fig. 10 D-F): While V226A and I228A mutants maintained substantial Gs activation capacity (∼15% reduction only), four mutants (F229A, I232A, V266A, T269A) showed severe activation deficits (~ 85% reduction). Notably, Q225A and V270A mutants displayed impaired Gs activation (~ 60% reduction), whereas T273A and L274A mutants were nearly incapable of Gs activation. These mutation-induced functional perturbations systematically mapped the structural determinants of orthosteric signaling, revealing key residues mediating allosteric transitions during receptor activation. The intracellular Site 4 shares striking topological similarity with the validated allosteric pocket in GPR88 80 (Figure S19), reinforcing its potential as a druggable domain. The hydrophobic character and membrane-proximal location of Site 4 align with established principles for intracellular allosteric modulator design 53 . Particularly compelling is its spatial correspondence to successful drug targets in class A GPCRs, suggesting conserved mechanisms of allosteric regulation exploitable for V2R-targeted therapeutic development. This integrated experimental and structural evidence positions Site 4 as a promising candidate for developing allosteric modulators capable of fine-tuning V2R signaling through selective stabilization of specific receptor conformations. Discussion Compared to conventional orthosteric drugs, allosteric modulators offer superior subtype selectivity and safety profiles, yet their development has been hindered by the challenges in identifying suitable allosteric sites 81 , 82 . Traditional approaches face significant limitations: experimental methods depend on serendipitous ligand discovery 68 , 83 , while computational predictions suffer from dataset biases (data-driven methods) or limited resolution (normal mode analyses) 84 – 86 . Our framework addresses these limitations through three key aspects: (1) MD simulations capturing conformational ensembles across three distinct inactivation and activation states, (2) MSM construction enabling analysis of thermodynamic and kinetic properties while mapping free energy landscapes, and (3) integrated pocket prediction algorithms (fpocket) combined with MI-based residue importance ranking. The integration of MD simulations with MSMs establishes an advanced methodological framework that significantly enhances protein dynamics investigation and accelerates drug discovery processes. By systematically aggregating data from multiple short-timescale simulations 87 , 88 , MSMs achieve comprehensive reconstruction of free energy landscapes that resolve three critical aspects: metastable states, rare transition events, and long-term conformational evolution. This methodology provides unique dual perspectives on both thermodynamic stability and kinetic accessibility 89 , 90 , enabling: (1) identification of transient functional states such as cryptic allosteric sites, and (2) quantitative characterization of transition pathways between conformational states. Our integrated computational-experimental framework has successfully identified a novel high-potential dynamic allosteric site in V2R, a critical therapeutic target for modulating fluid homeostasis. While current drug development efforts targeting V2R have predominantly focused on optimizing orthosteric ligands through structural modifications, there remains a critical gap in identifying viable allosteric sites for molecular intervention 91 , 92 . The discovery of this dynamic allosteric site presents new opportunities for developing selective allosteric modulators with potentially improved therapeutic profiles. This methodological framework demonstrates broad applicability beyond V2R, offering a systematic approach for identifying dynamic allosteric sites across various drug targets. However, the translation of these computational predictions into therapeutic agents faces significant challenges, particularly in the technical complexity of developing appropriate molecular probes for experimental validation. Notably, our previous research established methodological validity through mutagenesis studies that experimentally confirmed a predicted dynamic allosteric site 47 , 48 . Subsequent structure-based drug discovery efforts targeting this validated site yielded a novel allosteric modulator 47 , 93 , providing empirical validation of our predictive framework. These findings collectively demonstrate that our approach not only accelerates the identification of dynamic allosteric sites but also offers a viable pathway for translating computational predictions into bona fide therapeutic candidates. In summary, we established an integrated computational-experimental framework for identifying dynamic allosteric sites in protein receptors, employing V2R as a prototypical model system. Through synergistic application of MD simulations and MSMs, we systematically analyzed receptor conformational landscapes to reveal a novel allosteric site formed through V2R-specific activation-induced structural reconfiguration between TM5 and TM6. This dynamic allosteric site, identified as a central hub in the allosteric communication network through quantitative analysis of residue side-chain dynamics, was subsequently validated through experimental interrogation. Our methodology uniquely enables detection of dynamic allosteric sites independent of ligand binding constraints, thereby providing an accelerated path for developing allosteric modulators targeting conventionally challenging therapeutic targets. The presented framework - combining conformational ensemble analysis with MI-driven site prioritization - possesses inherent generalizability for application to diverse drug targets, offering transformative potential for expanding the repertoire of actionable allosteric sites in pharmacological discovery. Methods Preparation of simulation systems The fully active V2R structure complexed with the ligand AVP, Gs, nanobody35, and single Fab chain (PDB ID: 7KH0) was downloaded from the Protein Data Bank (PDB). The nanobody35 and single Fab chain were removed and the truncated N-terminal, Gαs subunit, ECL2, ICL2, and ICL3 were recovered by a loop building program, according to the sequence of the WT V2R. The active V2R structure was derived from the extracted coordinates of the receptor and ligand in the fully active state. For the apo V2R structure, we utilized Swiss Model to perform homology modeling and it can be found in the alphafold2 database. The modeled structure exhibited a RMSD of 0.93 Å when compared to the highest pLDDT-scoring structure in the AlphaFold2 database. Referring to the common sequence in the solved structures, the 32–342 residues were extracted as the input structures for the following processes. MD simulation settings All structures were oriented using the Orientations of Proteins in Membranes server 94 . Subsequently, the structures were incorporated into a POPC membrane utilizing the CHARMM additive force field through the CHARMM-GUI platform 95 . TIP3P water molecules were introduced at both the top and bottom of the system, and counterions such as K⁺ or Cl⁻ were also included in the solvation process. The bilayer components have frequently been utilized in various other simulation studies 96 . Using the input generator from CHARMM-GUI, we generated the coordinate and topology files in Amber format. Firstly, to prevent unrealistic collisions, the 2500 steepest descent cycles followed by 5000 conjugate gradient cycles was performed with the restraint of 10.0 kcal mol − 1 Å −2 on the V2R and 2.5 kcal mol − 1 Å −2 on the lipids. Secondly, the system underwent gradual heating from 0 to 300 K within the canonical ensemble (NVT) for 125 ps, with the restraint of 5.0 kcal mol − 1 Å −2 on the V2R and 2.5 kcal mol − 1 Å −2 on the lipids. Thirdly, four steps of equilibration molecular dynamics were performed in the isothermal isobaric (NPT) ensemble. The restrained force applied to all solute atoms was gradually decreased to 0 kcal· mol − 1 Å −2 to release all restraints. Then, the three states underwent 5 rounds of 1 µs MD simulations, with an integration step of 2.0 fs. In the end, we gathered 15 independent trajectories, each initiated with random velocities. The total simulation timescale was 15 µs. During the simulations, the Particle Mesh Ewald method was employed to compute long-range electrostatic interactions, while a cutoff of 9 Å was set for short-range electrostatic and van der Waals interactions. The SHAKE algorithm was used to handle covalent bonds involving hydrogen. A temperature of 310 K was maintained using a Langevin thermostat, with a collision frequency of 1.0 ps⁻¹. Snapshots were recorded every 100 ps. MSM construction According to the activation parameters, an MSM was built using the PyEMMA protocol ( http://www.emma-project.org/latest/ ) 90 . By validating the implied timescales (Figure S4), we verified that the V2R systems were Markovian and dependable, utilizing a model of 300 microstates with a lag time of 0.07 ps (35 steps) and a maximum of 100 iterations for k-means clustering. Subsequently, the microstates were grouped into three metastates using the PCCA + algorithm, a process that was validated through a Chapman–Kolmogorov test. Employing Transition Path Theory (TPT), we calculated the transition probability matrix for the MSM and determined the mean first passage time for each activation and inactivation event 97 . To obtain the most representative structure of each metastable state, we initially selected structures near the center of the metastates and compiled them into a condensed trajectory using the MDTraj package. Based on the new trajectories, we selected the representative snapshot of each metastate according to the pairwise similarity score S ij : $$\:{S}_{ij}={e}^{{-d}_{ij}/{d}_{scale}}$$ 1 Here, d ij is the RMSD between snapshots i and j, and d scale indicates the standard deviation of d. The snapshot with the highest similarity score was chosen as the most representative structure of each metastate. Residue fluctuation featurization and MSM construction To create a residue fluctuation feature set for each dataset, we computed the absolute deviation of each residue's side chain in every frame compared to the first frame. This resulted in a collection of t vectors, each with a length of n, where n represents the number of residues and t denotes the total number of frames. We established MSM microstates by clustering the side-chain fluctuation feature representation. The clustering process utilized k-centers, which continuously added new cluster centers until the maximum within-cluster distance fell below the threshold of 3.6 nm. This threshold was determined based on the results of the implied timescales test (Figure S8). Next, we conducted ten rounds of k-medoids updates, accepting updates when the largest distance to the nearest medoid decreased. To estimate transition probabilities based on frame assignments to clusters, we initially created a transition count matrix, where the element C ij represents the number of observed transitions from state i to state j. Subsequently, we incorporated a pseudocount of 1/n (where n is the number of states) into each element of the transition count matrix and performed row normalization to derive the transition probability matrix 98 , 99 . The lag times was 0.08 ps (40 steps), which was chosen by the implied timescales test (Figure S8). The optimal flux pathways between two sets of states were subsequently identified using transition path theory 100 , 101 . Computation of the MI matrix Beginning with the feature representation of the representative conformation for each MSM state, we established four thresholds to classify each side chain within each state into five distinct fluctuation levels with equal frequency. This process resulted in a feature set for each MSM state, where each snapshot is depicted by a quinary vector containing one entry per residue, representing one of five potential values that correspond to different fluctuation levels. We then proceeded to compute the MI between each pair of residues, which serves as an indicator of the statistical dependence between two random variables. It is given by the equation: $$\:MI\left(X,\:Y\right)=\sum\:_{y\in\:Y}\sum\:_{x\in\:X}p(x,y)\text{l}\text{o}\text{g}\left(\frac{p(x,y)}{p\left(x\right)p\left(y\right)}\right)$$ 2 where X and Y denote any pair of residues, while x and y indicate the fluctuation states of the respective residues. The probability p (x) represents the likelihood of observing a residue in state x, and p (x, y) refers to the joint probability of x and y. These probabilities are derived from the equilibrium probabilities calculated during the MSM fitting process. Although other approaches could be employed to identify a set of representative structures and their equilibrium probabilities, MSM is particularly beneficial as they tend to provide more accurate estimates of true equilibrium probabilities in sets of finite-length trajectories. To compute MI matrices for V2R, we utilized its dimer symmetry to enhance the sampling of fluctuation states. If A i and B i ​ are the random variables representing the fluctuation states of residue i from chains A and B, respectively, then due to the chemical identity of the two chains, at equilibrium, P (A i , B j ) = P (A j , B i ). To leverage this relationship for improved sampling of the state space and to enhance the robustness of our predictions against sampling errors, we calculate the mean of these two probabilities when determining MI. Visualization of the allosteric site network We used fpocket to identify thirteen potential allosteric sites from the representative conformations of the active intermediate state and the fully active state. The residue composition of each site was calculated using a distance cutoff of 4 Å. Based on the residues assigned to each site, we calculate the MI between residues for each pair of sites and take the average to obtain a symmetric MI matrix between sites. In this matrix, each element represents the average MI value between site pairs. Then, a weighted network graph is constructed based on the site MI matrix, where each node represents a site, and the edge weights between nodes reflect the average MI between sites. Next, we calculate the eigenvector centrality of each node to assess the importance of sites within the network. Finally, sites are ranked by eigenvector centrality to identify those with higher centrality. Eigenvector centrality is a method used to assess node importance within a network, particularly suited for networks with complex weight distributions 102 . Its fundamental concept is that a node's importance depends not only on its own connections but also on the centrality of its neighboring nodes. A node linked to many high-centrality nodes will, in turn, have a higher centrality score. For a given network, let A represent its adjacency matrix (or weighted matrix), where A ij denotes the connection weight between node i and node j. The eigenvector centrality c i of node i is defined as the vector that satisfies the following relationship: $$\:{c}_{i}=\frac{1}{\lambda\:}\sum\:_{j=1}^{n}{A}_{ij}{c}_{j}$$ 3 where n is the total number of nodes in the network and λ is the corresponding eigenvalue. λ is the largest eigenvalue, which is commonly used for normalization during the computation process. Cell culture and site-directed mutagenesis HEK293T cells were obtained from ATCC. Cells were maintained, passaged and transfected in DMEM medium containing 10% FBS, 100 U/ml penicillin and 100 µg/ml streptomycin (Gibco-ThermoFisher) in a humidified atmosphere at 37°C and 5% CO 2 . After transfection, cells were plated in DMEM containing 1% dialyzed FBS, 100 U/ml penicillin, and 100 µg/ml streptomycin for cell surface expression assays and G protein dissociation. All V2R mutants used in the present study were generated by site-directed mutagenesis. The successful introduction of the mutations in the polymerase chain reaction products was verified by DNA sequencing. V2R cell surface expression HEK293 cells were transiently transfected with 100 ng of HiBiT-tagged WT or mutated V2R, which contained a HiBiT sequence and a linker at the N terminus (MVSGWRLFKKISGSSGGSSGGNSGGGS; gene synthesized with codon optimization), After 24 h, the cells were seeded on 96-well microplates at a density of 15000 cells per well, and incubated for 12 h at 37°C. Bring the microplate back to room temperature, cells were mixed with 50 µL of assay buffer consisting of 1:50 of a LgBiT stock solution (Promega) and 1:25 extracellular substrate stock solution (Promega). Cells were incubated for 8 min at room temperature, then the luminescence were recorded using a Synergy Neo microplate reader (BioTek). G protein dissociation assay Gs (GαsS-RLuc8, Gβ3, Gγ9-GFP2) BRET probes were from the TRUPATH kit, which was a gift from Bryan Roth (Addgene kit #1000000163). HEK293 cells were transiently co-transfected with WT or mutated V2R along with specific G protein BRET probes according to the experimental setting. After 24 h, the cells were seeded on 96-well microplates at a density of 30,000–50,000 cells per well, and incubated for an additional 24 h. For the constitutive activity measurement, cells transfected with varying amounts of WT or mutated V2R (200ng, 400ng, 600ng, 800ng and 1000ng/well) were washed once with assay buffer (1× Hank’s balanced salt solution (HBSS) + 20 mM HEPES, pH 7.4) and the BRET signal was directly recorded after the addition of 5 µM RLuc8 substrate coelenterazine-400a using a a Synergy Neo microplate reader (BioTek). For the AVP-stimulated G protein activation, the cells were washed once with assay buffer and stimulated with AVP at different concentrations. BRET signal was subsequently measured after the addition of the coelenterazine-400a and was calculated as the ratio of the GFP2 emission to RLuc8 emission. Experimental data analysis All concentration-response curves were fit to a three-parameter logistic equation in Prism (Graphpad Software). BRET concentration-response curves were analyzed as either raw Net BRET (fit Emax-fit Baseline). Declarations Conflicts of interest The authors declare no conflicts of interest regarding this manuscript. Author Contributions J.Z. and S.L. conceived and supervised the project. J.Z., S.L., X.Q., C.Z. and X.L. designed the experiments. X.Q. performed the computational experiments and analyzed the data. C.Z. and X.L. performed the biological experiments. X.Q. and C.Z. drafted the manuscript, and all authors contributed to specific parts of the manuscript. J.Z. and S.L. assumed responsibility for the manuscript in its entirety. M.L., N.L., N.Li, J.H. and N.Liu discussed the results and revised the manuscript. Acknowledgements This study was supported by grants from the National Key R&D Program of China (No. 2023YFC3404700), the Noncommunicable Chronic Diseases-National Science and Technology Major Project (No. 2024ZD0531200), and the Innovative Research Team of High-Level Local Universities in Shanghai. Data availability The data that support this study are available from the corresponding authors upon reasonable request. Initial structures for MD simulation are obtained from the RCSB PDB database (PDB ID: 7KH0) ( https://www.rcsb.org/ ) and Swiss model ( https://swissmodel.expasy.org/ ). The analysis protocol for Markov State Model refers to http://www.emma-project.org/latest/ . Pocket prediction is accomplished by fpocket, see http://fpocket.sourceforge.net/ . The calculation of mutual information for residue side chains is based on modules from https://github.com/bowman-lab/enspara/tree/master/enspara . Other simulation analyses were based on AMBER suite, according to http://ambermd.org/ . The code for calculating the allosteric network of sites can be referenced from https://github.com/MaienLii/allosteric-site-network . References Fenton AW (2008) Allostery: an illustrated definition for the 'second secret of life'. Trends Biochem Sci 33:420–425 Goodey NM, Benkovic SJ (2008) Allosteric regulation and catalysis emerge via a common route. Nat Chem Biol 4:474–482 Zhang M et al (2024) G protein-coupled receptors (GPCRs): advances in structures, mechanisms, and drug discovery. Signal Transduct Target Ther 9:88 Nussinov R, Tsai CJ (2013) Allostery in disease and in drug discovery. Cell 153:293–305 Jiang H et al (2017) Peptidomimetic inhibitors of APC-Asef interaction block colorectal cancer migration. Nat Chem Biol 13:994–1001 Changeux JP, Christopoulos A (2016) Allosteric Modulation as a Unifying Mechanism for Receptor Function and Regulation. Cell 166:1084–1102 Lu S et al (2021) Mechanism of allosteric activation of SIRT6 revealed by the action of rationally designed activators. Acta Pharm Sin B 11:1355–1361 Ni D et al (2022) Computational elucidation of allosteric communication in proteins for allosteric drug design. Drug Discov Today 27:2226–2234 Zha J et al (2023) AlloReverse: multiscale understanding among hierarchical allosteric regulations. Nucleic Acids Res 51:W33–W38 Zhou Z et al (2012) Total glucosides of paeony attenuated functional maturation of dendritic cells via blocking TLR4/5 signaling in vivo. Int Immunopharmacol 14:275–282 De Smet F, Christopoulos A, Carmeliet P (2014) Allosteric targeting of receptor tyrosine kinases. Nat Biotechnol 32:1113–1120 Wootten D, Christopoulos A, Sexton PM (2013) Emerging paradigms in GPCR allostery: implications for drug discovery. Nat Rev Drug Discov 12:630–644 Wu P, Clausen MH, Nielsen TE (2015) Allosteric small-molecule kinase inhibitors. Pharmacol Ther 156:59–68 Dang CV, Reddy EP, Shokat KM, Soucek L (2017) Drugging the 'undruggable' cancer targets. Nat Rev Cancer 17:502–508 Zorn JA, Wells JA (2010) Turning enzymes ON with small molecules. Nat Chem Biol 6:179–188 Sun Z et al (2020) Covalent Inhibitors Allosterically Block the Activation of Rho Family Proteins and Suppress Cancer Cell Invasion. Adv Sci (Weinh) 7:2000098 Lazo JS, Sharlow ER (2016) Drugging Undruggable Molecular Cancer Targets. Annu Rev Pharmacol Toxicol 56:23–40 Xie X et al (2023) Recent advances in targeting the undruggable proteins: from drug discovery to clinical trials. Signal Transduct Target Ther 8:335 Huang L, Guo Z, Wang F, Fu L (2021) KRAS mutation: from undruggable to druggable in cancer. Signal Transduct Target Ther 6:386 Cimermancic P et al (2016) CryptoSite: Expanding the Druggable Proteome by Characterization and Prediction of Cryptic Binding Sites. J Mol Biol 428:709–719 Ni D et al (2021) Along the allostery stream: Recent advances in computational methods for allosteric drug discovery. WIREs Comput Mol Sci 12 Lu S, Ji M, Ni D, Zhang J (2018) Discovery of hidden allosteric sites as novel targets for allosteric drug design. Drug Discov Today 23:359–365 Ostrem JM, Peters U, Sos ML, Wells JA, Shokat KM (2013) K-Ras(G12C) inhibitors allosterically control GTP affinity and effector interactions. Nature 503:548–551 Lu S et al (2020) Emergence of allosteric drug-resistance mutations: new challenges for allosteric drug discovery. Drug Discov Today 25:177–184 Latallo MJ, Cortina GA, Faham S, Nakamoto RK, Kasson PM (2017) Predicting allosteric mutants that increase activity of a major antibiotic resistance enzyme. Chem Sci 8:6484–6492 Zhang Q et al (2022) Targeting a cryptic allosteric site of SIRT6 with small-molecule inhibitors that inhibit the migration of pancreatic cancer cells. Acta Pharm Sin B 12:876–889 Lu S et al (2019) Deactivation Pathway of Ras GTPase Underlies Conformational Substates as Targets for Drug Design. ACS Catal 9:7188–7196 Ni D et al (2020) Discovery of cryptic allosteric sites using reversed allosteric communication by a combined computational and experimental strategy. Chem Sci 12:464–476 Zhang M, Jang H, Nussinov R (2019) The mechanism of PI3Kalpha activation at the atomic level. Chem Sci 10:3671–3680 Evenseth LSM et al (2020) Exploring Conformational Dynamics of the Extracellular Venus flytrap Domain of the GABA(B) Receptor: A Path-Metadynamics Study. J Chem Inf Model 60:2294–2303 Powers AS et al (2024) A non-canonical mechanism of GPCR activation. Nat Commun 15:9938 Otun O et al (2024) Conformational dynamics underlying atypical chemokine receptor 3 activation. Proc Natl Acad Sci U S A 121:e2404000121 Husic BE, Pande VS (2018) Markov State Models: From an Art to a Science. J Am Chem Soc 140:2386–2396 McClendon CL, Friedland G, Mobley DL, Amirkhani H, Jacobson MP (2009) Quantifying Correlations Between Allosteric Sites in Thermodynamic Ensembles. J Chem Theory Comput 5:2486–2502 Stokely AM et al (2024) NetSci: A Library for High Performance Biomolecular Simulation Network Analysis Computation. J Chem Inf Model 64:7966–7976 Juul KV, Bichet DG, Nielsen S, Norgaard JP (2014) The physiological and pathophysiological functions of renal and extrarenal vasopressin V2 receptors. Am J Physiol Ren Physiol 306:F931–940 Makita N, Manaka K, Sato J, Iiri T (2020) V2 vasopressin receptor mutations. Vitam Horm 113:79–99 Stoof SC, Cnossen MH, de Maat MP, Leebeek FW, Kruip MJ (2016) Side effects of desmopressin in patients with bleeding disorders. Haemophilia 22:39–45 Wang L et al (2022) Author Correction: Cryo-EM structure of the AVP-vasopressin receptor 2-Gs signaling complex. Cell Res 32:957 Zhou Q et al (2019) Common activation mechanism of class A GPCRs. Elife 8 Zhou F et al (2021) Molecular basis of ligand recognition and activation of human V2 vasopressin receptor. Cell Res 31:929–931 Natochin YV, Golosova DV (2020) Vasopressin receptor subtypes and renal sodium transport. Vitam Horm 113:239–258 Mancinelli CD et al (2024) Control of G protein-coupled receptor function via membrane-interacting intrinsically disordered C-terminal domains. Proc Natl Acad Sci U S A 121:e2407744121 Shenol A et al (2024) Molecular dynamics-based identification of binding pathways and two distinct high-affinity sites for succinate in succinate receptor 1/GPR91. Mol Cell 84:955–966e954 Heng J et al (2023) Function and dynamics of the intrinsically disordered carboxyl terminus of beta2 adrenergic receptor. Nat Commun 14:2005 Li Y, Sun J, Li D, Lin J (2022) The full activation mechanism of the adenosine A(1) receptor revealed by GaMD and Su-GaMD simulations. Proc Natl Acad Sci U S A 119:e2203702119 Lu S et al (2021) Activation pathway of a G protein-coupled receptor uncovers conformational intermediates as targets for allosteric drug design. Nat Commun 12:4721 Li M et al (2024) Delineating the stepwise millisecond allosteric activation mechanism of the class C GPCR dimer mGlu5. Nat Commun 15:7519 Harrigan MP et al (2017) MSMBuilder: Statistical Models for Biomolecular Dynamics. Biophys J 112:10–15 Dror RO et al (2011) Activation mechanism of the beta2-adrenergic receptor. Proc Natl Acad Sci U S A 108:18684–18689 Medeiros Almeida V et al (2021) Role of a high centrality residue in protein dynamics and thermal stability. J Struct Biol 213:107773 Li M et al (2022) Mechanistic Insights into the Long-range Allosteric Regulation of KRAS Via Neurofibromatosis Type 1 (NF1) Scaffold Upon SPRED1 Loading. J Mol Biol 434:167730 Zhang M, Lan X, Li X, Lu S (2023) Pharmacologically targeting intracellular allosteric sites of GPCRs for drug discovery. Drug Discov Today 28:103803 Lu S, Zhang J (2019) Small Molecule Allosteric Modulators of G-Protein-Coupled Receptors: Drug-Target Interactions. J Med Chem 62:24–45 Ignatowska-Jankowska BM et al (2015) A Cannabinoid CB1 Receptor-Positive Allosteric Modulator Reduces Neuropathic Pain in the Mouse with No Psychoactive Effects. Neuropsychopharmacology 40:2948–2959 Qu Q et al (2023) Insights into distinct signaling profiles of the microOR activated by diverse agonists. Nat Chem Biol 19:423–430 Noe F, Clementi C (2015) Kinetic distance and kinetic maps from molecular dynamics simulation. J Chem Theory Comput 11:5002–5011 Ranieri M et al (2019) Gain-of-function mutations of the V2 vasopressin receptor in nephrogenic syndrome of inappropriate antidiuresis (NSIAD): a cell-based assay to assess constitutive water reabsorption. Pflugers Arch 471:1291–1304 Gilles N (2023) Natural Peptide Toxins as an Option for Renewed Treatment of Type 2 Vasopressin Receptor-Related Diseases. Biology (Basel) 12 Liu HL et al (2024) Structural basis of tolvaptan binding to the vasopressin V(2) receptor. Acta Pharmacol Sin 45:2441–2449 Krumm BE et al (2023) Neurotensin Receptor Allosterism Revealed in Complex with a Biased Allosteric Modulator. Biochemistry 62:1233–1248 Ring AM et al (2013) Adrenaline-activated structure of beta2-adrenoceptor stabilized by an engineered nanobody. Nature 502:575–579 Rasmussen SG et al (2011) Structure of a nanobody-stabilized active state of the beta(2) adrenoceptor. Nature 469:175–180 Yuan D et al (2020) Activation of the alpha(2B) adrenoceptor by the sedative sympatholytic dexmedetomidine. Nat Chem Biol 16:507–512 Zhang H et al (2015) Structure of the Angiotensin receptor revealed by serial femtosecond crystallography. Cell 161:833–844 Wingler LM, McMahon C, Staus DP, Lefkowitz RJ, Kruse AC (2019) Distinctive Activation Mechanism for Angiotensin Receptor Revealed by a Synthetic Nanobody. Cell 176, 479–490 e412 Shao Z et al (2016) High-resolution crystal structure of the human CB1 cannabinoid receptor. Nature 540:602–606 Yang X et al (2022) Molecular mechanism of allosteric modulation for the cannabinoid receptor CB1. Nat Chem Biol 18:831–840 Feinstein AG, Cole JL, May ER (2024) Dimerization Promotes PKR Activation by Modulating Energetics of alphaC Helix Conversion between Active and Inactive Conformations. J Phys Chem B 128:9305–9314 Demirtas K, Erman B, Haliloglu T (2024) Dynamic correlations: exact and approximate methods for mutual information. Bioinformatics 40 Hacisuleyman A, Erman B (2024) Synergy and anti-cooperativity in allostery: Molecular dynamics study of WT and oncogenic KRAS-RGL1. Proteins 92:665–678 Schmidtke P, Le Guilloux V, Maupetit J, Tuffery P (2010) fpocket: online tools for protein ensemble pocket detection and tracking. Nucleic Acids Res 38:W582–589 Negre CFA et al (2018) Eigenvector centrality for characterization of protein allosteric pathways. Proc Natl Acad Sci U S A 115:E12201–E12208 Tekpinar M, Yildirim A (2022) Impact of dimerization and N3 binding on molecular dynamics of SARS-CoV and SARS-CoV-2 main proteases. J Biomol Struct Dyn 40:6243–6254 Schneider M, Antes I (2023) Comparison of allosteric signaling in DnaK and BiP using mutual information between simulated residue conformations. Proteins 91:237–255 Ho JD et al (2018) Structural basis for GPR40 allosteric agonism and incretin stimulation. Nat Commun 9:1645 Liu X et al (2019) Mechanism of beta(2)AR regulation by an intracellular positive allosteric modulator. Science 364:1283–1287 DiBerto JF, Olsen RHJ, Roth BL (2022) TRUPATH: An Open-Source Biosensor Platform for Interrogating the GPCR Transducerome. Methods Mol Biol 2525:185–195 Olsen RHJ et al (2020) TRUPATH, an open-source biosensor platform for interrogating the GPCR transducerome. Nat Chem Biol 16:841–849 Chen G et al (2022) Activation and allosteric regulation of the orphan GPR88-Gi1 signaling complex. Nat Commun 13:2375 Oleinikovas V, Saladino G, Cossins BP, Gervasio FL (2016) Understanding Cryptic Pocket Formation in Protein Targets by Enhanced Sampling Simulations. J Am Chem Soc 138:14257–14263 Lu S, Shen Q, Zhang JA (2019) Methods and Their Applications: Facilitating the Discovery of Allosteric Drugs and the Investigation of Allosteric Mechanisms. Acc Chem Res 52:492–500 Lees JA et al (2023) An inverse agonist of orphan receptor GPR61 acts by a G protein-competitive allosteric mechanism. Nat Commun 14:5938 Panjkovich A, Daura X (2014) PARS: a web server for the prediction of Protein Allosteric and Regulatory Sites. Bioinformatics 30:1314–1315 Huang M et al (2018) AlloFinder: a strategy for allosteric modulator discovery and allosterome analyses. Nucleic Acids Res 46:W451–W458 Liu X et al (2020) Unraveling allosteric landscapes of allosterome with ASD. Nucleic Acids Res 48:D394–D401 Bowman GR, Huang X, Pande VS (2010) Network models for molecular kinetics and their initial applications to human health. Cell Res 20:622–630 Noe F, Fischer S (2008) Transition networks for modeling the kinetics of conformational change in macromolecules. Curr Opin Struct Biol 18:154–162 Beauchamp KA et al (2011) MSMBuilder2: Modeling Conformational Dynamics at the Picosecond to Millisecond Scale. J Chem Theory Comput 7:3412–3419 Scherer MK et al (2015) PyEMMA 2: A Software Package for Estimation, Validation, and Analysis of Markov Models. J Chem Theory Comput 11:5525–5542 Xiong X et al (2024) Long-Residence Time Peptide Antagonist for the Vasopressin V(2) Receptor to Treat Autosomal Dominant Polycystic Kidney Disease. J Med Chem 67:5935–5944 Cao X et al (2022) Benzodiazepine Derivatives as Potent Vasopressin V(2) Receptor Antagonists for the Treatment of Autosomal Dominant Kidney Disease. J Med Chem 65:9295–9311 Zhang D et al (2023) Structural insights into angiotensin receptor signaling modulation by balanced and biased agonists. EMBO J 42:e112940 Lomize MA, Pogozheva ID, Joo H, Mosberg HI, Lomize AL (2012) OPM database and PPM web server: resources for positioning of proteins in membranes. Nucleic Acids Res 40:D370–376 Jo S et al (2017) CHARMM-GUI 10 years for biomolecular modeling and simulation. J Comput Chem 38:1114–1124 Mattedi G, Acosta-Gutierrez S, Clark T, Gervasio FL (2020) A combined activation mechanism for the glucagon receptor. Proc Natl Acad Sci U S A 117:15414–15422 E W, Vanden-Eijnden E (2006) Towards a Theory of Transition Paths. J Stat Phys 123:503–523 Zimmerman MI, Porter JR, Sun X, Silva RR, Bowman GR (2018) Choice of Adaptive Sampling Strategy Impacts State Discovery, Transition Probabilities, and the Apparent Mechanism of Conformational Changes. J Chem Theory Comput 14:5459–5475 Prinz JH et al (2011) Markov models of molecular kinetics: generation and validation. J Chem Phys 134:174105 Metzner P, Schütte C, Vanden-Eijnden E (2009) Transition Path Theory for Markov Jump Processes. Multiscale Model Simul 7:1192–1219 E W, Vanden-Eijnden E (2010) Transition-path theory and path-finding algorithms for the study of rare events. Annu Rev Phys Chem 61:391–420 Foutch D, Pham B, Shen T (2021) Protein conformational switch discerned via network centrality properties. Comput Struct Biotechnol J 19:3599–3608 Additional Declarations There is NO Competing Interest. Supplementary Files SupplementaryInformation.docx Supplementary Information Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-6427090","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":442839330,"identity":"6685dd8b-d41e-4b0f-941b-3778aa4db890","order_by":0,"name":"Shaoyong Lu","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA40lEQVRIiWNgGAWjYBACPgYGNiiT+QCQOEBYCxtCC1sCVAsz0Vp4DIjUIpH+7MGPisNy5vxrvkl8qLnDYM7ej991bBI55oY9Zw4bW854u01yxrFnDJY9hwnZksMmwdt2OHHDjbPbpHnYDjMY3Egm7DDJv2AtZ55J//kH1HL/MSEtCWbSYFvO97BJM7aBbCHkfZ43ZtIyZ9KNDW6wGVv29h3mMTiTbIBXCz870GFvKqzlDM4ffnjjx7fDcgbHDz7Abw0ENDMwSCSwSABZPMQoB4E6oH0HmD8Qq3wUjIJRMApGFgAAsRRKH3LIG4kAAAAASUVORK5CYII=","orcid":"https://orcid.org/0000-0002-1334-6292","institution":"Shanghai Jiao Tong University School of Medicine","correspondingAuthor":true,"prefix":"","firstName":"Shaoyong","middleName":"","lastName":"Lu","suffix":""},{"id":442839331,"identity":"36b3d030-878d-4372-949b-b724f447ec85","order_by":1,"name":"Xin Qiao","email":"","orcid":"","institution":"Shanghai Jiao Tong University School of Medicine","correspondingAuthor":false,"prefix":"","firstName":"Xin","middleName":"","lastName":"Qiao","suffix":""},{"id":442839332,"identity":"5d9547cd-24fb-4520-8ba3-3a2a618055a2","order_by":2,"name":"Chunhao Zhu","email":"","orcid":"","institution":"Ningxia Medical University","correspondingAuthor":false,"prefix":"","firstName":"Chunhao","middleName":"","lastName":"Zhu","suffix":""},{"id":442839333,"identity":"fcdeace6-30c8-4c63-a56b-06fc37a0d3ab","order_by":3,"name":"Xiaobing Lan","email":"","orcid":"","institution":"College of Pharmacy, Ningxia Medical University, Yinchuan, Ningxia Hui Autonomous Region 750004, China","correspondingAuthor":false,"prefix":"","firstName":"Xiaobing","middleName":"","lastName":"Lan","suffix":""},{"id":442839334,"identity":"ae8abaf0-5bd6-46ee-8487-36e9577559a8","order_by":4,"name":"Mingyu Li","email":"","orcid":"","institution":"Shanghai Jiao Tong University School of Medicine","correspondingAuthor":false,"prefix":"","firstName":"Mingyu","middleName":"","lastName":"Li","suffix":""},{"id":442839335,"identity":"21d70b41-dae2-47ad-8a86-6b5e56bec742","order_by":5,"name":"Nuan Li","email":"","orcid":"","institution":"Shanghai Jiao Tong University School of Medicine","correspondingAuthor":false,"prefix":"","firstName":"Nuan","middleName":"","lastName":"Li","suffix":""},{"id":442839336,"identity":"bd555a7f-e1a6-4cf0-a926-810844d7a78e","order_by":6,"name":"Jianxiang Huang","email":"","orcid":"","institution":"Shanghai Jiao Tong University School of Medicine","correspondingAuthor":false,"prefix":"","firstName":"Jianxiang","middleName":"","lastName":"Huang","suffix":""},{"id":442839337,"identity":"c21dc674-7839-4662-b2d8-f2361587c53f","order_by":7,"name":"Ning Liu","email":"","orcid":"","institution":"Ningxia Medical University","correspondingAuthor":false,"prefix":"","firstName":"Ning","middleName":"","lastName":"Liu","suffix":""},{"id":442839338,"identity":"d7999dad-4dfa-416e-9920-388ed23c3f46","order_by":8,"name":"Jian Zhang","email":"","orcid":"https://orcid.org/0000-0002-6558-791X","institution":"Shanghai Jiao Tong University School of Medicine","correspondingAuthor":false,"prefix":"","firstName":"Jian","middleName":"","lastName":"Zhang","suffix":""}],"badges":[],"createdAt":"2025-04-11 09:50:10","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-6427090/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-6427090/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":81797760,"identity":"63b2896f-2c40-4600-9f4e-189cc29427dd","added_by":"auto","created_at":"2025-05-02 04:24:01","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":301442,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eCryo-EM structure of the fully active V2R−Gs protein complex reveals ligand binding and activation mechanism.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e(\u003cstrong\u003eA\u003c/strong\u003e) The overall cryo-EM structure of the fully active vasopressin V2 receptor (V2R) in complex with the agonist arginine vasopressin (AVP) (cyan), Gαs (palegreen), Gβs (orange) and Gγs (lightblue).\u003c/p\u003e\n\u003cp\u003e(\u003cstrong\u003eB\u003c/strong\u003e) The structure of the active V2R in complex with the agonist AVP. The zoom-in views represent the binding mode of AVP, the kink in the L312\u003csup\u003e7.40\u003c/sup\u003e−A314\u003csup\u003e7.42\u003c/sup\u003e region of TM7, the polar network among S127\u003csup\u003e3.40\u003c/sup\u003e, T218\u003csup\u003e5.51\u003c/sup\u003e, and Y280\u003csup\u003e6.44\u003c/sup\u003e, and the weak hydrophobic interactions between R137\u003csup\u003e3.50\u003c/sup\u003e and V266\u003csup\u003e6.30\u003c/sup\u003e.\u003c/p\u003e","description":"","filename":"image1.png","url":"https://assets-eu.researchsquare.com/files/rs-6427090/v1/203042794ef6af7e404b14c5.png"},{"id":81797113,"identity":"a0248686-30fd-4b95-937a-beef8c570fad","added_by":"auto","created_at":"2025-05-02 04:08:01","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":262954,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eSchematic workflow for identifying dynamic allosteric sites. Conformational ensemble sampling:\u003c/strong\u003e initiate molecular dynamics (MD) simulations using receptor structures representing distinct functional states to comprehensively sampling the conformational ensemble. \u003cstrong\u003eActivation pathway analysis:\u003c/strong\u003e employ Markov state models (MSMs) to elucidate activation transition pathways and characterize state-specific conformational rearrangements within the receptor system. \u003cstrong\u003eAllosteric site prediction:\u003c/strong\u003e Quantify residue-level dynamic coupling through mutual information (MI) analysis of side chain fluctuations. Prioritize potential allosteric sites based on their MI network connectivity scores. \u003cstrong\u003eExperimental validation: \u003c/strong\u003everify high-potential allosteric sites through site-directed mutagenesis coupled with bioluminescence resonance energy transfer (BRET) functional assays.\u003c/p\u003e","description":"","filename":"image2.png","url":"https://assets-eu.researchsquare.com/files/rs-6427090/v1/4ca35ddbf526711f350efbdd.png"},{"id":81797109,"identity":"6239a0d0-958a-4c93-b062-346f3039dbdb","added_by":"auto","created_at":"2025-05-02 04:08:01","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":188939,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eIntegrated profiling of V2R activation via free energy landscape and Markov state models (MSMs)\u003c/strong\u003e.\u003c/p\u003e\n\u003cp\u003e(\u003cstrong\u003eA\u003c/strong\u003e) The activation pathway free energy landscape of V2R was constructed using two key order parameters: the Cα-Cα distance between residues V270\u003csup\u003e6.34\u003c/sup\u003e and R137\u003csup\u003e3.50\u003c/sup\u003e, and the Cα-Cα distance between P322\u003csup\u003e7.50\u003c/sup\u003e and L81\u003csup\u003e2.46\u003c/sup\u003e. Distinct conformational states are indicated by arrows within the landscape, with corresponding color scaling shown in the adjacent bar.\u003c/p\u003e\n\u003cp\u003e(\u003cstrong\u003eB\u003c/strong\u003e) MSM analysis reveals three metastates of V2R. Spatial distributions of the apo inactive (blue), intermediate active (yellow), and fully active (pink) states are superimposed on the free energy landscape, accompanied by metastate assignment probabilities shown in the adjacent panel.\u003c/p\u003e\n\u003cp\u003e(\u003cstrong\u003eC\u003c/strong\u003e) State transition kinetics quantified through mean first passage time (MFPT) analysis, demonstrating temporal progression between the apo inactive (blue), intermediate active (yellow), and fully active (pink) states.\u003c/p\u003e","description":"","filename":"image3.png","url":"https://assets-eu.researchsquare.com/files/rs-6427090/v1/90e0839be5dfb997de0ff33e.png"},{"id":81797115,"identity":"dfac8c49-8fb7-4161-bd56-6b2c2a1c73e4","added_by":"auto","created_at":"2025-05-02 04:08:01","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":308229,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eComparative analysis of structural variations among V2R states: the apo inactive (blue), intermediate active (yellow), and fully active (pink) states.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e(\u003cstrong\u003eA\u003c/strong\u003e) Structural variations in the TM7 kink geometry within the AVP binding pocket (L312\u003csup\u003e7.40\u003c/sup\u003e−A314\u003csup\u003e7.42\u003c/sup\u003e region).\u003c/p\u003e\n\u003cp\u003e(\u003cstrong\u003eB\u003c/strong\u003e) State-dependent angular distributions of the Cα-triangulated angle (P306\u003csup\u003e7.34\u003c/sup\u003e-A314\u003csup\u003e7.42\u003c/sup\u003e-Y325\u003csup\u003e7.53\u003c/sup\u003e) derived from MD trajectories.\u003c/p\u003e\n\u003cp\u003e(\u003cstrong\u003eC\u003c/strong\u003e) Differential inward rotation patterns of F287\u003csup\u003e6.51\u003c/sup\u003e, F288\u003csup\u003e6.52\u003c/sup\u003e and Q291\u003csup\u003e6.55\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003e(\u003cstrong\u003eD\u003c/strong\u003e) Comparative analysis of dihedral angle probability profiles calculated between Cβ atoms (S127\u003csup\u003e3.40\u003c/sup\u003eand V281\u003csup\u003e6.45\u003c/sup\u003e) and aromatic ring centroids (Y280\u003csup\u003e6.44\u003c/sup\u003e and W284\u003csup\u003e6.48\u003c/sup\u003e).\u003c/p\u003e\n\u003cp\u003e(\u003cstrong\u003eE\u003c/strong\u003e) Differential outward rotational displacement of Y280\u003csup\u003e6.44\u003c/sup\u003e and W284\u003csup\u003e6.48\u003c/sup\u003e within the CWxP motif.\u003c/p\u003e","description":"","filename":"image4.png","url":"https://assets-eu.researchsquare.com/files/rs-6427090/v1/e5fa7568804c158ea7e65d73.png"},{"id":81797123,"identity":"bbc8efa2-6794-4c8a-acab-2b9196761530","added_by":"auto","created_at":"2025-05-02 04:08:02","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":1741399,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eComparative analysis of structural variations among V2R states: the apo inactive (blue), intermediate active (yellow), and fully active (pink) states.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e(\u003cstrong\u003eA\u003c/strong\u003e) The polar network among S127\u003csup\u003e3.40\u003c/sup\u003e, T218\u003csup\u003e5.51\u003c/sup\u003e, and Y280\u003csup\u003e6.44\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003e(\u003cstrong\u003eB\u003c/strong\u003e) Probability distribution of the Cβ-P217\u003csup\u003e5.50\u003c/sup\u003e(pyrrolidine)-Y280\u003csup\u003e6.44\u003c/sup\u003e(benzene) centroid angle across all simulation trajectories.\u003c/p\u003e\n\u003cp\u003e(\u003cstrong\u003eC\u003c/strong\u003e) Comparative analysis of Q225\u003csup\u003e5.58\u003c/sup\u003e-Y325\u003csup\u003e7.53\u003c/sup\u003e spatial proximity.\u003c/p\u003e\n\u003cp\u003e(\u003cstrong\u003eD\u003c/strong\u003e) Probability distribution of the Q225\u003csup\u003e5.58\u003c/sup\u003e (carbonyl carbon)-Y325\u003csup\u003e7.53\u003c/sup\u003e (benzene centroid) distance.\u003c/p\u003e\n\u003cp\u003e(\u003cstrong\u003eE)\u003c/strong\u003e The comparison of the upward bending of R137\u003csup\u003e3.50\u003c/sup\u003e and the outward movement of V266\u003csup\u003e6.30\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003e(\u003cstrong\u003eF\u003c/strong\u003e) Probability distribution of the R137\u003csup\u003e3.50\u003c/sup\u003e (guanidine carbon)-V266\u003csup\u003e6.30\u003c/sup\u003e (Cβ) distance.\u003c/p\u003e","description":"","filename":"image5.png","url":"https://assets-eu.researchsquare.com/files/rs-6427090/v1/17592a2b85af7ebc4d3d79f5.png"},{"id":81797125,"identity":"f8d4bb7d-bc1e-4324-92e1-378bf361b369","added_by":"auto","created_at":"2025-05-02 04:08:02","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":233457,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eThe schematic diagram illustrates V2R activation pathway through four functionally distinct layers.\u003c/strong\u003e The layers are organized by structural topology and mechanistic roles in signal transduction. Node represents structurally equivalent residues along the activation pathway, with critical residues and functional motifs highlighted in red. The hierarchical organization comprises: Layer 1 (signal initiation) containing primary activation triggers, Layers 2-3 (signal propagation) mediating sequential conformational changes through the transmembrane helices, and Layer 4 (G protein coupling interface) facilitating interactions with downstream effector proteins. This multi-layered architecture demonstrates the spatial-temporal progression of activation signals from the AVP ligand binding site to the Gs protein coupling interface.\u003c/p\u003e","description":"","filename":"image6.png","url":"https://assets-eu.researchsquare.com/files/rs-6427090/v1/0d7e5c86d86959e765fe5742.png"},{"id":81797112,"identity":"5c1c4784-b3f3-477d-8234-8cbb72e63d37","added_by":"auto","created_at":"2025-05-02 04:08:01","extension":"png","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":165298,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eA schematic outline of calculating the allosteric network among sites based on mutual information.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e(\u003cstrong\u003eA\u003c/strong\u003e) Markov state model (MSMs) illustrate a dynamic system comprising multiple conformational states (represented as circles with diameters proportional to their population sizes) and inter-state transitions (depicted by single-headed arrows whose lengths inversely correspond to transition probabilities). Each state features a five-level classification system (quinary scale) for individual residues, where color intensity within circular elements reflects the magnitude of side chain conformational fluctuations - darker hues indicate greater structural flexibility.\u003c/p\u003e\n\u003cp\u003e(\u003cstrong\u003eB\u003c/strong\u003e) A comprehensive pairwise mutual information (MI) matrix, derived from the MSMs in panel A, displays all-against-all residue correlations. Matrix elements are color-coded such that intensified darkness corresponds to stronger mutual information between residue pairs, quantitatively representing their dynamic coupling.\u003c/p\u003e\n\u003cp\u003e(\u003cstrong\u003eC\u003c/strong\u003e) This panel presents a residue-resolution allosteric communication network, where double-edged circles (color-coded by site affiliation) represent individual residues. Solid-line enclosures demarcate functional sites, while inter-residue connections (straight lines) visualize MI-derived correlations. Line thickness quantitatively encodes interaction strength.\u003c/p\u003e","description":"","filename":"image7.png","url":"https://assets-eu.researchsquare.com/files/rs-6427090/v1/1f9584097c046dd256d8056c.png"},{"id":81797116,"identity":"72cb9585-d799-4bb9-b7d9-f1b26197c143","added_by":"auto","created_at":"2025-05-02 04:08:01","extension":"png","order_by":8,"title":"Figure 8","display":"","copyAsset":false,"role":"figure","size":2025345,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eEigenvector centrality-driven ranking and allosteric network integration of potential sites.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e(\u003cstrong\u003eA\u003c/strong\u003e) The predicted potential sites were ranked according to their eigenvector centrality scores. The top five sites exhibited eigenvector centrality values exceeding or approaching 0.3, with Site 5 notably identified as an orthosteric site.\u003c/p\u003e\n\u003cp\u003e(\u003cstrong\u003eB\u003c/strong\u003e) Spherical representations of the top five sites were generated based on their eigenvector centrality values, visualized at a 180° clockwise rotation angle.\u003c/p\u003e\n\u003cp\u003e(\u003cstrong\u003eC\u003c/strong\u003e) The allosteric network connecting the top five sites was mapped, where edge weights reflect the pairwise total correlation between sites, and node sizes correspond to their eigenvector centrality. For clarity, self-edges and edges with negligible correlations were excluded.\u003c/p\u003e\n\u003cp\u003e(\u003cstrong\u003eD-H\u003c/strong\u003e) Each of the top five sites is represented as spherical structures within the fully active conformation, with color-coding consistent across both the network diagram (\u003cstrong\u003eC\u003c/strong\u003e) and structural visualizations (\u003cstrong\u003eD-H\u003c/strong\u003e).\u003c/p\u003e","description":"","filename":"image8.png","url":"https://assets-eu.researchsquare.com/files/rs-6427090/v1/850c76a2359660c0004171f4.png"},{"id":81797473,"identity":"0381eef8-bd52-4eab-aa64-0b2daeaeb071","added_by":"auto","created_at":"2025-05-02 04:16:01","extension":"png","order_by":9,"title":"Figure 9","display":"","copyAsset":false,"role":"figure","size":2026570,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eStructural topology and spatial mapping of top four allosteric sites.\u003c/strong\u003e The four distinct binding sites (\u003cstrong\u003eA-D\u003c/strong\u003e) are visualized through stick representations with color-coded differentiation: Site 1 (yellow) occupies the interhelical cleft formed between the central transmembrane segments of TM6 and TM7. Site 2 (blue) is positioned along the lipid-exposed interface created by the mid-regions of TM3 and TM5. Site 3 (red) localizes within the solvent-accessible cavity framed by the central portions of TM3, TM6, and TM7. Site 4 (orange) resides in the intracellular groove formed between the cytoplasmic ends of TM5 and TM6.\u003c/p\u003e","description":"","filename":"image9.png","url":"https://assets-eu.researchsquare.com/files/rs-6427090/v1/69365d166a9246db44f380a9.png"},{"id":81797477,"identity":"0a0f0886-9606-4854-adb4-9d2bc9e5a346","added_by":"auto","created_at":"2025-05-02 04:16:02","extension":"png","order_by":10,"title":"Figure 10","display":"","copyAsset":false,"role":"figure","size":722131,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eConstitutive and AVP-induced activity of Site 4 mutants in BRET assays\u003c/strong\u003e. (\u003cstrong\u003eA-C\u003c/strong\u003e) To systematically evaluate constitutive activity, we established graded cell surface expression levels of V2R and its Site 4 mutants in HEK293T cells through precise titration of receptor-encoding plasmid transfection quantities.\u003c/p\u003e\n\u003cp\u003e(\u003cstrong\u003eD-F\u003c/strong\u003e) Gαs dissociation assays revealed that mutations at Site 4 residues substantially impaired AVP-induced signaling relative to the wild-type (WT) receptor, with mutant membrane expression levels quantified in Figure S18. All measurements were normalized to equivalent cell surface expression to ensure comparable receptor densities.\u003c/p\u003e","description":"","filename":"image10.png","url":"https://assets-eu.researchsquare.com/files/rs-6427090/v1/f70e4ec31e00e9f5844bc0cc.png"},{"id":83617867,"identity":"a96d6b6d-debe-4ec4-8180-455a88d3e225","added_by":"auto","created_at":"2025-05-29 14:19:48","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":8726649,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6427090/v1/e447d42a-0d35-44a7-a328-d3308ca03518.pdf"},{"id":81797118,"identity":"d8fd16f4-8af2-43d7-a7c5-5f7587ac8f0a","added_by":"auto","created_at":"2025-05-02 04:08:01","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":7684208,"visible":true,"origin":"","legend":"Supplementary Information","description":"","filename":"SupplementaryInformation.docx","url":"https://assets-eu.researchsquare.com/files/rs-6427090/v1/69124ba8ac74188feacc6846.docx"}],"financialInterests":"There is \u003cb\u003eNO\u003c/b\u003e Competing Interest.","formattedTitle":"Conformational mapping of GPCR activation: dynamic allosteric site discovery in V2R through MD-MSM and mutual information analysis","fulltext":[{"header":"Introduction","content":"\u003cp\u003eAllostery, often referred to as \u0026lsquo;the second secret of life\u0026rsquo;, occurs in various biological macromolecules, including dynamic proteins, DNA, and RNA polymers\u003csup\u003e\u003cspan additionalcitationids=\"CR2\" citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e\u003c/sup\u003e. It plays a critical role in regulating conformational changes and functional modulation of biomolecules\u003csup\u003e\u003cspan additionalcitationids=\"CR5 CR6\" citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e\u003c/sup\u003e. Typically, allosteric perturbations occur at sites distant from orthosteric sites on proteins. Consequently, allosteric modulators do not compete with endogenous substrates that bind to orthosteric sites but instead finely tune receptor affinity in a highly predictable manner, thereby enhancing safety\u003csup\u003e\u003cspan additionalcitationids=\"CR9\" citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e\u003c/sup\u003e. Moreover, allosteric modulators can achieve better selective targeting of different receptor subfamilies because allosteric sites are less conserved compared to orthosteric sites, thereby reducing the risk of side effects\u003csup\u003e\u003cspan additionalcitationids=\"CR12\" citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eDespite these advantages, many clinically important targets lack suitable allosteric sites for the design of therapeutic molecules\u003csup\u003e\u003cspan additionalcitationids=\"CR15\" citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e\u003c/sup\u003e. Currently, certain transcription factors, small GTPases, and phosphatases are considered as \u0026lsquo;undruggable proteins\u0026rsquo; due to the absence of appropriate small molecule binding sites\u003csup\u003e\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e, \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e\u003c/sup\u003e. For example, the transcription factor p53, a tumor suppressor protein, exerts its effects by interacting with mouse double minute 2 homolog (MDM2). Drug design targeting p53 remains challenging because its protein-protein interaction (PPI) interface is broad and flat, lacking druggable binding sites\u003csup\u003e\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e\u003c/sup\u003e. However, allosteric modulators could inhibit PPIs by targeting allosteric sites distant from their interfaces. Notably, dynamic allosteric sites, which are only visible in low-populated intermediate states, enrich the diversity of allosteric sites and expand the range of potential drug targets\u003csup\u003e\u003cspan additionalcitationids=\"CR21\" citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eDynamic allosteric sites have addressed another previously undruggable challenge, as exemplified by the small GTPase KRAS, one of the most frequently mutated oncogenes\u003csup\u003e\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e\u003c/sup\u003e. The two currently approved drugs, Sotorasib (AMG-510) and Adagrasib (MRTX849), are covalent inhibitors targeting the residue Cys12 of the KRAS G12C mutant. Both drugs bind to a cryptic allosteric site, which is not clearly visible in other published KRAS structures, although in some cases, a groove can be observed. However, this site was discovered accidentally during the experimental process. Generally, most allosteric sites are identified based on available bound molecules and involve various exhaustive and tedious methods, such as crystallography, mutagenesis experiments, and high-throughput screening experiments\u003csup\u003e\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e, \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e\u003c/sup\u003e. Identification of dynamic allosteric sites in the absence of bound molecules would significantly accelerate the drug development process. Furthermore, since cryptic allosteric sites cannot be observed in static structures, their identification remains a significant challenge\u003csup\u003e\u003cspan additionalcitationids=\"CR27\" citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e\u003c/sup\u003e. To date, there is a lack of effective methods for locating dynamic allosteric sites.\u003c/p\u003e \u003cp\u003eWhile structural approaches provide valuable insights, their static nature offers limited information about dynamics underling receptor activation\u003csup\u003e\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e\u003c/sup\u003e. In contrast, computational techniques, particularly those based on molecular dynamics (MD) simulations, have advanced the identification of novel dynamic allosteric sites without bound molecules. MD simulations effectively offer atomic-level insights into the spatial and temporal organization of biomolecular conformations, enabling the uncovering of interactions between distant functional regions\u003csup\u003e\u003cspan additionalcitationids=\"CR30 CR31\" citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e\u003c/sup\u003e. Additionally, Markov state models (MSMs), constructed from extensive MD simulations, provide a simplified representation of dynamic biomolecular ensembles. A key advantage of MSMs is their ability to identify dynamic allosteric sites that often exist in subtle conformational states\u003csup\u003e\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e\u003c/sup\u003e. Notably, the residue-based mutual information (MI) matrix is an approach for examining the interdependencies between residues in a protein. It quantifies statistical correlations between pairs of residues and proves valuable in identifying cooperative interactions and uncovering cryptic allosteric sites\u003csup\u003e\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e, \u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eAs a typical class A G protein-coupled receptor (GPCR), the vasopressin V2 receptor (V2R) regulates fluid homeostasis by binding to the endogenous substrate arginine vasopressin (AVP), which stimulates the activation of downstream Gs proteins\u003csup\u003e\u003cspan additionalcitationids=\"CR37\" citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e\u003c/sup\u003e. In the cryo-EM structure, the initial six residues in AVP form a cyclic insertion deep within the orthosteric site. Specifically, V2R exhibits a kink in the L312\u003csup\u003e7.40\u003c/sup\u003e\u0026minus;A314\u003csup\u003e7.42\u003c/sup\u003e region (superscripts indicate the Ballesteros\u0026ndash;Weinstein numbering for GPCR residues), a polar network among S127\u003csup\u003e3.40\u003c/sup\u003e, T218\u003csup\u003e5.51\u003c/sup\u003e, and Y280\u003csup\u003e6.44\u003c/sup\u003e, and weak interactions between the intracellular ends of TM3 and TM6\u003csup\u003e39\u003c/sup\u003e (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). Current research has extensively explored the activation mechanism of GPCRs, identifying a common activation pathway that directly connects the ligand-binding pocket to the G protein-coupling region. This pathway integrates key conserved motifs such as CWxP, PIF, Na\u003csup\u003e+\u003c/sup\u003e pocket, DRY, and NPxxY, elucidating how residue contact rearrangements drive conformational changes in transmembrane helices\u003csup\u003e\u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e\u003c/sup\u003e. Although some studies have resolved the cryo-EM structures of the V2R receptor bound to the agonist AVP, the lack of the apo or inactive state structures has constrained the scope of these investigations. Consequently, these studies relied on comparisons between the active structure of V2R and the inactive conformation of the highly homologous oxytocin receptor (OTR), leaving the activation mechanism of the V2R receptor insufficiently explored\u003csup\u003e\u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e, \u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e41\u003c/span\u003e\u003c/sup\u003e. Additionally, suitable allosteric sites for drug design remain elusive for V2R\u003csup\u003e\u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e42\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eHere, we developed an integrated computational and experimental framework that comprehensively delineates the activation mechanism of the receptor and identifies a previously unexplored dynamic allosteric site on V2R. Initially, we performed multiple replica MD simulations coupled with MSMs to explore the activation transition pathway of the receptor\u003csup\u003e\u003cspan additionalcitationids=\"CR44 CR45\" citationid=\"CR43\" class=\"CitationRef\"\u003e43\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e46\u003c/span\u003e\u003c/sup\u003e. Subsequent analyses compared representative conformations to evaluate changes in key residues and motifs during activation. Following this, we devised an algorithmic framework based on residue side-chain fluctuations to analyze the allosteric network among the predicted dynamic sites. Ultimately, a promising dynamic allosteric site was identified within the intracellular region of the receptor through site-directed mutagenesis experiments.\u003c/p\u003e"},{"header":"Results","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003eThe integrated computational and experimental framework for identifying dynamic allosteric sites\u003c/h2\u003e \u003cp\u003eTo initiate the computational workflow (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e), we established three states of V2R for five independent runs of each 1 us MD simulations: the apo inactive state, the AVP-bound intermediate active state, and both AVP- and Gs-bound fully active state (PDB ID: 7KH0). This timescale has been demonstrated to effectively capture the ensemble of dynamic conformations of receptors\u003csup\u003e\u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e47\u003c/span\u003e, \u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e48\u003c/span\u003e\u003c/sup\u003e. Subsequently, the simulation trajectories were dimensionally reduced using characteristic features of receptor activation. MSMs were then constructed, dividing the system into several metastates and analyzing the ensemble\u0026rsquo;s dynamic properties and equilibrium distribution\u003csup\u003e\u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e49\u003c/span\u003e\u003c/sup\u003e. Next, representative conformations of each metastate were extracted from the simulated trajectories, and the conformational changes of key residues and motifs during receptor activation were analyzed\u003csup\u003e\u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e48\u003c/span\u003e, \u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e50\u003c/span\u003e\u003c/sup\u003e. After gaining a comprehensive understanding of the ensemble, the fpocket program was used to predict potential allosteric sites in both the intermediate active and fully active states of the receptor. Thereafter, based on the MI matrix of residue side-chain fluctuations, the potential sites were ranked according to eigenvector centrality, a method used to assess node importance within networks, particularly suited for those with complex weight distributions\u003csup\u003e\u003cspan citationid=\"CR51\" class=\"CitationRef\"\u003e51\u003c/span\u003e\u003c/sup\u003e. Finally, high-potential allosteric sites were selected for experimental validation via site-directed mutagenesis and bioluminescence resonance energy transfer (BRET) assays\u003csup\u003e\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e, \u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e, \u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e47\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003e\u003c/h3\u003e\n\u003cdiv class=\"Heading\"\u003e\u003cb\u003eExploring the activation transition pathway of the receptor among different metastates\u003c/b\u003e\u003c/div\u003e \u003cp\u003eWe conducted 1 \u0026micro;s \u0026times; 5 independent runs for three distinct states of V2R using random initial velocities, yielding a cumulative trajectory length of 15 \u0026micro;s. The apo inactive state\u0026rsquo;s initial structure, generated via Swiss-Model, exhibited near-identical topology to the AlphaFold2-predicted structure with the highest pLDDT score, as evidenced by a low root-mean-square deviation (RMSD) of 0.93 \u0026Aring; between their atomic coordinates (Figure \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e). In contrast, the receptor portion of the fully active state\u0026rsquo;s initial structure was differed from the apo inactive state\u0026rsquo;s receptor structure in four key respects (Figure S2): bending at the extracellular part of TM7, absence of pronounced outward rotation in the CWxP motif, inward displacement at the intracellular part of TM5, and outward displacement at the intracellular part of TM6. Previous investigations into the activation mechanism of V2R compared its active structure with the inactive structure of the oxytocin receptor (OTR), which shares high homology. These studies identified a bend at the extracellular end of TM7, a relatively small degree of outward rotation at the intracellular end of TM6, and a mutation in the PIF motif\u003csup\u003e\u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e, \u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e41\u003c/span\u003e\u003c/sup\u003e. However, their analysis of the entire activation pathway lacked comprehensiveness, failing to detect the subtle outward rotation of the CWxP motif during activation, the inward movement of the intracellular end of TM5, and the nonpolar mutation of the residue V266\u003csup\u003e6.30\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eFollowing MD simulations, we calculated the RMSD for Cα atoms in the receptor portions across three distinct states relative to their initial conformations (Figure S3A). The receptor system reached equilibrium after 100 ns of the simulation time. Notably, the AVP-bound intermediate active state exhibited the highest RMSD values (5.90\u0026thinsp;\u0026plusmn;\u0026thinsp;0.89 \u0026Aring;) (Figure S3A), suggesting substantial ligand-induced structural reorganization. Comparative analysis revealed significantly lower deviations in the apo inactive state (4.48\u0026thinsp;\u0026plusmn;\u0026thinsp;0.63 \u0026Aring;) and the fully active ternary complex containing both AVP and Gs protein (3.94\u0026thinsp;\u0026plusmn;\u0026thinsp;0.37 \u0026Aring;) (Figure S3A), indicating more constrained conformational flexibility in these states. This differential RMSD pattern suggested that ligand binding selectively enhanced conformational plasticity in the intermediate active state while maintaining relative stability in other conformations during simulations.\u003c/p\u003e \u003cp\u003eTo characterize regional mobility variations, we calculated per-residue root-mean-square fluctuation of backbone heavy atoms across receptor segments in all three states\u003csup\u003e\u003cspan citationid=\"CR52\" class=\"CitationRef\"\u003e52\u003c/span\u003e\u003c/sup\u003e. Significant fluctuation differences were observed in intracellular loop (ICL) regions 2 and 3 (Figure S3B). These loops likely participate in Gs protein interactions, potentially stabilizing their conformation. The inherent flexibility of ICL2 and ICL3 in GPCRs, typically results in structural ambiguity, as evidenced by poorly resolved electron density maps that complicate molecular modeling. This technical challenge has historically limited investigation of ICL3-G protein functional interactions\u003csup\u003e\u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e, \u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e41\u003c/span\u003e\u003c/sup\u003e. Our study addresses this gap through computational reconstruction of missing ICL segments and systematic characterization of loop-Gαs protein interplay, providing mechanistic insights for future research. Detailed interaction analysis identified key polar and hydrophobic contacts between both intracellular loops and the Gs protein (Figures S4-S5), with specific residue participation documented.\u003c/p\u003e \u003cp\u003eClass A GPCRs feature seven transmembrane α-helices (TM1-TM7), with agonist-induced conformational changes primarily involving outward displacement of TM6 and inward movement of TM7. These structural rearrangements facilitate the formation of a G protein-binding pocket\u003csup\u003e\u003cspan additionalcitationids=\"CR54\" citationid=\"CR53\" class=\"CitationRef\"\u003e53\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR55\" class=\"CitationRef\"\u003e55\u003c/span\u003e\u003c/sup\u003e. To characterize the receptor activation dynamics, we employed two distinct order parameters: 1) the Cα-Cα distance between V270\u003csup\u003e6.34\u003c/sup\u003e (TM6) and R137\u003csup\u003e3.50\u003c/sup\u003e (TM3) quantifying TM6 outward displacement\u003csup\u003e\u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e50\u003c/span\u003e\u003c/sup\u003e, and 2) the Cα-Cα separation between P322\u003csup\u003e7.50\u003c/sup\u003e (TM7) and L81\u003csup\u003e2.46\u003c/sup\u003e (TM2) measuring TM7 inward movement\u003csup\u003e\u003cspan citationid=\"CR56\" class=\"CitationRef\"\u003e56\u003c/span\u003e\u003c/sup\u003e. Using MD trajectories, we constructed a two-dimensional free energy landscape by projecting these structural parameters (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eA). The apo inactive state, characterized by initial distances of 7.1 \u0026Aring; (V270\u003csup\u003e6.34\u003c/sup\u003e-R137\u003csup\u003e3.50\u003c/sup\u003e) and 10.1 \u0026Aring; (P322\u003csup\u003e7.50\u003c/sup\u003e-L81\u003csup\u003e2.46\u003c/sup\u003e), occupied a distinct energy basin (5.5\u0026ndash;6.7 \u0026Aring; and 8.6\u0026ndash;10.3 \u0026Aring;, respectively). Activation involved sequential energy barrier crossings: first transitioning to an intermediate active state (7.4\u0026ndash;12.1 \u0026Aring; and 7.6\u0026ndash;9.9 \u0026Aring;), then progressing to the fully active state (13.7\u0026ndash;15.6 \u0026Aring; and 7.4\u0026ndash;8.4 \u0026Aring;). Notably, TM6 displacement showed a progressive increase in V270\u003csup\u003e6.34\u003c/sup\u003e-R137\u003csup\u003e3.50\u003c/sup\u003e distance, while TM7 movement reduced the P322\u003csup\u003e7.50\u003c/sup\u003e-L81\u003csup\u003e2.46\u003c/sup\u003e separation. This trajectory analysis delineates the complete activation pathway of V2R from the apo inactive through an intermediate active to the fully active states. Convergence validation through subsampling and multiple MD replicates\u003csup\u003e\u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e47\u003c/span\u003e, \u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e48\u003c/span\u003e\u003c/sup\u003e (Figures S6-S7) confirmed sufficient sampling of the activation landscape. The observed energy barriers and metastable states align with established GPCR activation mechanisms, demonstrating the system's capacity to capture essential conformational transitions.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eTo investigate the kinetic properties and equilibrium distribution of the ensemble, we established MSMs based on key activation parameters. From a statistical mechanics perspective, MSMs enabled systematic characterization of conformational distributions within the equilibrium ensemble. This methodology not only permitted precise identification of metastable states but also enabled quantitative determination of kinetic properties, including transition timescales between states\u003csup\u003e\u003cspan citationid=\"CR57\" class=\"CitationRef\"\u003e57\u003c/span\u003e\u003c/sup\u003e. The modeling workflow proceeded as follows: first, the conformational ensemble was discretized into 300 microstates, with Markovian behavior verified through implied timescale analysis (Figure S8A). Subsequently, the Robust Perron Cluster Analysis (PCCA+) algorithm was implemented to cluster these microstates into three distinct metastates. Transition path theory (TPT) was then employed to calculate inter-metastate transition timescales. To validate model consistency, we conducted Chapman-Kolmogorov tests comparing predicted and observed transition probabilities between metastates (Figure S8B).\u003c/p\u003e \u003cp\u003eThe conformational space was partitioned into three regions, with blue, yellow, and pink designating the apo inactive, intermediate active, and fully active states, respectively. MSMs-derived conformational distribution showed strong agreement with the free energy landscape (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eB), with equilibrium populations of approximately 30% for each metastate due to minimal energy barrier differences. Notably, the fully active state demonstrated the highest population (39.4%), consistent with previous reports of receptor basal activity\u003csup\u003e\u003cspan citationid=\"CR58\" class=\"CitationRef\"\u003e58\u003c/span\u003e, \u003cspan citationid=\"CR59\" class=\"CitationRef\"\u003e59\u003c/span\u003e\u003c/sup\u003e, thereby validating the MSM construction. Transition kinetics analysis revealed differential timescales between activation and deactivation pathways: The apo inactive\u0026rarr;fully active transition (96.72 \u0026micro;s) occurred faster than the reverse process (108.24 \u0026micro;s) (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eC), aligning with the observed population bias toward the fully active state. Furthermore, transitions from both apo inactive (29.23 \u0026micro;s) and fully active (22.21 \u0026micro;s) states to the intermediate active state exhibited accelerated kinetics compared to their reverse transitions (intermediate active\u0026rarr;apo inactive: 70.01 \u0026micro;s; intermediate active\u0026rarr;fully active: 40.77 \u0026micro;s). This kinetic hierarchy identifies the intermediate active\u0026rarr;fully active transition as the rate-limiting step for receptor activation, while the intermediate active\u0026rarr;apo inactive transition governs deactivation kinetics.\u003c/p\u003e\n\u003ch3\u003eMolecular insights into key steps of the specific activation pathway\u003c/h3\u003e\n\u003cp\u003eTo investigate conformational changes during receptor activation, we extracted representative conformations of three metastates from MSMs-derived conformational landscapes and systematically analyzed dynamic alterations in key residues and structural motifs. Following AVP ligand binding, activation signals propagated through conserved motifs to the receptor intracellular end, facilitating Gs protein interaction\u003csup\u003e\u003cspan citationid=\"CR60\" class=\"CitationRef\"\u003e60\u003c/span\u003e\u003c/sup\u003e. Notably, V2R exhibited distinct activation-related conformational changes compared to other class A GPCRs.\u003c/p\u003e \u003cp\u003eStructural analysis revealed that the extracellular ends of TM6 and TM7 cooperatively formed the AVP binding site. Specifically, Tyr2 of AVP established a hydrogen bond with L312\u003csup\u003e7.40\u003c/sup\u003e (Figure S9), inducing characteristic bending in the L312\u003csup\u003e7.40\u003c/sup\u003e\u0026minus;A314\u003csup\u003e7.42\u003c/sup\u003e segment of TM7\u003csup\u003e41\u003c/sup\u003e (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eA). Conformational sampling demonstrated increased prevalence of the smaller angle conformations among P306\u003csup\u003e7.34\u003c/sup\u003e, A314\u003csup\u003e7.42\u003c/sup\u003e, and Y325\u003csup\u003e7.53\u003c/sup\u003e in the fully active state (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eB), corroborating TM7 bending observed in cryo-EM structural comparisons. This contrasts with other class A GPCRs (e.g., NTSR1 and β2AR)\u003csup\u003e\u003cspan citationid=\"CR61\" class=\"CitationRef\"\u003e61\u003c/span\u003e, \u003cspan citationid=\"CR62\" class=\"CitationRef\"\u003e62\u003c/span\u003e\u003c/sup\u003e where agonist binding fails to induce significant TM7 bending. Concurrent with these changes, residues F287\u003csup\u003e6.51\u003c/sup\u003e, F288\u003csup\u003e6.52\u003c/sup\u003e, and Q291\u003csup\u003e6.55\u003c/sup\u003e underwent inward rotation (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eC), facilitating binding site formation. Signal transduction to the adjacent CWxP motif revealed divergent mechanisms: while β2AR and α2AR receptors exhibit pronounced outward rotation of W\u003csup\u003e\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e.\u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e48\u003c/span\u003e\u003c/sup\u003e followed by F\u003csup\u003e\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e.\u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e44\u003c/span\u003e\u003c/sup\u003e movement\u003csup\u003e\u003cspan citationid=\"CR63\" class=\"CitationRef\"\u003e63\u003c/span\u003e, \u003cspan citationid=\"CR64\" class=\"CitationRef\"\u003e64\u003c/span\u003e\u003c/sup\u003e. V2R demonstrated minimal dihedral variation among S127\u003csup\u003e3.40\u003c/sup\u003e, Y280\u003csup\u003e6.44\u003c/sup\u003e, W284\u003csup\u003e6.48\u003c/sup\u003e, and V281\u003csup\u003e6.45\u003c/sup\u003e across the three states (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eD). This indicates attenuated outward rotation of Y280\u003csup\u003e6.44\u003c/sup\u003e/W284\u003csup\u003e6.48\u003c/sup\u003e during activation\u003csup\u003e\u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e\u003c/sup\u003e (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eE), differing from conventional GPCR activation paradigms. Comparative analysis with OTR structures revealed partial outward rotation of W\u003csup\u003e6.48\u003c/sup\u003e/Y\u003csup\u003e6.44\u003c/sup\u003e in activated V2R\u003csup\u003e\u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e, \u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e41\u003c/span\u003e\u003c/sup\u003e. This discrepancy may arise from evolutionary substitutions at positions 280\u003csup\u003e6.44\u003c/sup\u003e (F\u0026rarr;Y) and 127\u003csup\u003e3.40\u003c/sup\u003e (I\u0026rarr;S), establishing a stabilizing polar network between S127\u003csup\u003e3.40\u003c/sup\u003e, T218\u003csup\u003e5.51\u003c/sup\u003e, and Y280\u003csup\u003e6.44\u003c/sup\u003e (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eA). This network likely restricts TM6 mobility, as evidenced by limited angular changes among S127\u003csup\u003e3.40\u003c/sup\u003e, P217\u003csup\u003e5.50\u003c/sup\u003e, and Y280\u003csup\u003e6.44\u003c/sup\u003e during activation (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eB). Consistently, V2R exhibits reduced intracellular TM6 displacement (V\u003csup\u003e\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e.\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e\u003c/sup\u003e-R\u003csup\u003e\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e.\u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e50\u003c/span\u003e\u003c/sup\u003e distance: 14.5 \u0026Aring;, PDB ID: 7KH0) compared to α2AR (15.9 \u0026Aring;, PDB ID: 6K41)\u003csup\u003e\u003cspan citationid=\"CR64\" class=\"CitationRef\"\u003e64\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eUpon transmission of the activation signal to the central V2R region, the hydrophobic interaction triad comprising I130\u003csup\u003e3.43\u003c/sup\u003e, I276\u003csup\u003e6.40\u003c/sup\u003e, and V277\u003csup\u003e6.41\u003c/sup\u003e became destabilized\u003csup\u003e\u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e47\u003c/span\u003e\u003c/sup\u003e, facilitating outward displacement of TM6 (Figure S10A, B). Concurrently, TM3 and TM7 underwent mutual approach\u003csup\u003e\u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e\u003c/sup\u003e, leading to structural collapse of the Na\u003csup\u003e+\u003c/sup\u003e binding pocket formed by D85\u003csup\u003e2.50\u003c/sup\u003e, S126\u003csup\u003e3.39\u003c/sup\u003e, N317\u003csup\u003e7.45\u003c/sup\u003e and N321\u003csup\u003e7.49\u003c/sup\u003e (Figure S10C, D). These conformational rearrangements align with canonical activation patterns observed across class A GPCRs.\u003c/p\u003e \u003cp\u003eProgressive signal propagation induced distinct helical motions: inward displacement of TM3, outward movement of TM6, and inward repositioning of TM7 (Figure S11). Notably, a novel discovery emerged with TM5 exhibiting inward movement (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eC), narrowing its spatial relationship with TM6 and TM7. This was quantitatively corroborated by reduced interatomic distance between Q225\u003csup\u003e5.58\u003c/sup\u003e and Y325\u003csup\u003e7.53\u003c/sup\u003e in both the intermediate active and fully active states (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eD). Comparative analysis revealed that TM5 intracellular domain dynamics exhibit receptor-specific variability within class A GPCRs: AT1R systems display outward movement, while CB1R systems show inward displacement\u003csup\u003e\u003cspan additionalcitationids=\"CR66 CR67\" citationid=\"CR65\" class=\"CitationRef\"\u003e65\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR68\" class=\"CitationRef\"\u003e68\u003c/span\u003e\u003c/sup\u003e. A hallmark of V2R architecture lies in the differential regulation of TM6 dynamics (Figs.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eE, F). Specifically, the polar network at the middle interface of TM6 constrained its outward displacement amplitude, while the intracellular mutation at V266\u003csup\u003e6.30\u003c/sup\u003e enhanced TM6 mobility, correlating with elevated basal receptor activity\u003csup\u003e\u003cspan citationid=\"CR58\" class=\"CitationRef\"\u003e58\u003c/span\u003e, \u003cspan citationid=\"CR59\" class=\"CitationRef\"\u003e59\u003c/span\u003e\u003c/sup\u003e. This contrasts with conserved class A GPCR features where the acidic residue (typically D\u003csup\u003e\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e.\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e\u003c/sup\u003e) forms stabilizing ionic interactions with R\u003csup\u003e\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e.\u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e50\u003c/span\u003e\u003c/sup\u003e in inactive states.\u003c/p\u003e \u003cp\u003eIntegration of these findings enabled comprehensive mapping of activation-associated signal transduction pathways, highlighting critical structural motifs including CWxP, PSY, NPxxY, and DRH (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003e, red highlights). The collective data elucidate unique V2R activation mechanics characterized by: (1) constrained TM6 middle movement, (2) facilitated TM6 intracellular displacement, and (3) TM5 inward trajectory enhancing proximity to TM6/TM7.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e\n\u003ch3\u003eMI analysis uncovers allosteric communication networks among potential sites\u003c/h3\u003e\n\u003cp\u003eTo investigate dynamic correlations between residues, we first computed the absolute deviation of each residue's side chain conformation across the simulation trajectory. Subsequent construction of MSMs employed Euclidean distance-based clustering of side chain deviation vectors, with model validation performed through implied timescale analysis (Figure S12). Each microstate in the MSMs was represented by a quintenary vector encoding side chain fluctuation patterns, where vector elements were visually encoded using a gradient scale (darker shades indicating larger deviations, lighter tones representing smaller variations) (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eA).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eWe then quantified residue-residue correlations through MI analysis, generating a comprehensive MI matrix (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eB). MI provides a robust measure of statistical dependence between random variables, quantifying the information shared between two systems. In information-theoretic terms, higher MI values indicate stronger associations, while lower values suggest weaker relationships. This method is particularly advantageous for detecting nonlinear correlations and automatically excludes invariant residues (zero entropy) from contributing to MI calculations\u003csup\u003e\u003cspan additionalcitationids=\"CR70\" citationid=\"CR69\" class=\"CitationRef\"\u003e69\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR71\" class=\"CitationRef\"\u003e71\u003c/span\u003e\u003c/sup\u003e. Our implementation extends MI's proven utility in bioinformatics and machine learning to protein dynamics analysis.\u003c/p\u003e \u003cp\u003ePotential allosteric sites were identified using fpocket\u003csup\u003e\u003cspan citationid=\"CR72\" class=\"CitationRef\"\u003e72\u003c/span\u003e\u003c/sup\u003e across representative conformations of both the intermediate active and fully active states, yielding 13 candidate sites. The residue-level MI matrix was subsequently transformed into a site-specific MI matrix (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eC), enabling network construction where MI values served as edge weights. To identify key regulatory hubs, we computed eigenvector centrality measures from the MI-derived weighted adjacency matrix. This network metric reveals globally influential nodes that maintain strong connections with other high-MI sites\u003csup\u003e\u003cspan citationid=\"CR73\" class=\"CitationRef\"\u003e73\u003c/span\u003e, \u003cspan citationid=\"CR74\" class=\"CitationRef\"\u003e74\u003c/span\u003e\u003c/sup\u003e, effectively highlighting sites with maximal impact on allosteric communication pathways during protein conformational changes.\u003c/p\u003e\n\u003ch3\u003eIdentification of a dynamic allosteric site on the activation pathway\u003c/h3\u003e\n\u003cp\u003eWe systematically ranked predicted potential sites using eigenvector centrality metrics, with the top five sites visualized as spherical structures (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003eA, B). MI calculations required careful consideration of continuous variable discretization, as the selection of interval ranges directly impacts probability distribution estimations and subsequent computational outcomes\u003csup\u003e\u003cspan citationid=\"CR75\" class=\"CitationRef\"\u003e75\u003c/span\u003e\u003c/sup\u003e. To ensure robustness, we performed comprehensive analyses by discretizing side-chain deviation values into varying interval categories (6\u0026ndash;10 bins). While interval selection induced fluctuations in site rankings, the five highest-ranked sites consistently maintained their positions (1\u0026ndash;5) across different discretization schemes (Figure S13). Based on the representative five-interval discretization results, we constructed an allosteric network through MI analysis of the top-ranked sites, with each site represented as distinct spherical elements. Notably, the fifth site (Site 5) corresponded to the orthosteric binding site, while the fourth site (Site 4) exhibited eigenvector centrality values comparable to those of Site 5.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eComprehensive analysis of the allosteric network revealed that the top five sites, encompassing the orthosteric site, maintain intricate allosteric interconnectivity through dynamic coupling (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003eC). Structural characterization demonstrated that Site 1 and Site 3 occupy adjacent positions within the interhelical crevice formed by mid-regions of TM6 and TM7 (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003eD, F; \u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e9\u003c/span\u003eA, C; S14A). While Site 1 faces the lipid bilayer interface, Site 3 resides within the receptor\u0026rsquo;s transmembrane core. Computational characterization using fpocket quantified distinct physicochemical properties (Table \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e): Site 1 (volume\u0026thinsp;=\u0026thinsp;193 \u0026Aring;\u003csup\u003e\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e\u003c/sup\u003e, SASA\u0026thinsp;=\u0026thinsp;59 \u0026Aring;\u003csup\u003e\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e\u003c/sup\u003e, α-sphere density\u0026thinsp;=\u0026thinsp;1.9 \u0026Aring;\u003csup\u003e\u0026minus;3\u003c/sup\u003e) and Site 3 (volume\u0026thinsp;=\u0026thinsp;115 \u0026Aring;\u003csup\u003e\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e\u003c/sup\u003e, SASA\u0026thinsp;=\u0026thinsp;24 \u0026Aring;\u003csup\u003e\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e\u003c/sup\u003e, α-sphere density\u0026thinsp;=\u0026thinsp;1.7 \u0026Aring;\u003csup\u003e\u0026minus;3\u003c/sup\u003e). These metrics collectively indicate suboptimal druggability for both sites, characterized by limited surface accessibility and loose packing geometry. Notably, Site 3 demonstrates reduced hydrophobic character, evidenced by an apolar α-sphere proportion of 0.16.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eSite 2 occupies a lipid-exposed niche at mid-regions of TM3 and TM5 (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003eE, \u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e9\u003c/span\u003eB, S14B). Its formation correlates with intracellular TM5 inward displacement during activation, which reduces the interhelical distance between TM5 and TM3 side chains. This site emerges exclusively in intermediate conformational states of the ensemble and remains undetected in the initial active-state structure. Despite moderate dimensions (SASA\u0026thinsp;=\u0026thinsp;47 \u0026Aring;\u003csup\u003e\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e\u003c/sup\u003e, volume\u0026thinsp;=\u0026thinsp;201 \u0026Aring;\u003csup\u003e\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e\u003c/sup\u003e) and packing density (α-sphere\u0026thinsp;=\u0026thinsp;2.0 \u0026Aring;\u003csup\u003e\u0026minus;3\u003c/sup\u003e), its structural plasticity and limited accessibility compromise therapeutic targeting potential.\u003c/p\u003e \u003cp\u003eIn contrast, Site 4 demonstrates unique structural and dynamic features at the intracellular TM5-TM6 interface (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003eG, \u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e9\u003c/span\u003eD, S14C, D). This pocket forms through coordinated helical movements during activation: TM6 intracellular shift combined with TM5 inward displacement toward TM6. Comparative conformational analysis revealed V2R-specific activation mechanics: unlike canonical class A GPCRs, which exhibit pronounced TM6 outward movement, V2R activation features attenuated TM6 displacement coupled with TM5 inward migration. This distinctive motion pattern creates a receptor-specific dynamic pocket absent in the initial active structure, where excessive TM6 outward displacement (T269\u003csup\u003e6.33\u003c/sup\u003e, T273\u003csup\u003e6.37\u003c/sup\u003e, V277\u003csup\u003e6.41\u003c/sup\u003e side chain repositioning) obstructs cavity formation. Notably, Site 4 presents superior druggability metrics: substantial volume (515 \u0026Aring;\u003csup\u003e\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e\u003c/sup\u003e), enhanced surface accessibility (SASA\u0026thinsp;=\u0026thinsp;139 \u0026Aring;\u003csup\u003e\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e\u003c/sup\u003e), and favorable hydrophobicity (apolar α-sphere proportion\u0026thinsp;=\u0026thinsp;0.75) (Table \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e). Its compact architecture (α-sphere density\u0026thinsp;=\u0026thinsp;4.1 \u0026Aring;\u003csup\u003e\u0026minus;3\u003c/sup\u003e) and high fpocket druggability score (0.855) suggest strong therapeutic potential. Trajectory analysis confirmed dynamic persistence of this pocket throughout conformational sampling (Figure S15A), supporting its identification as a viable allosteric site.\u003c/p\u003e \u003cp\u003eNotably, among the 13 potential sites identified, two particularly significant sites (Site 7 and Site 10) emerged, underscoring the efficacy of our methodology in detecting allosteric sites. However, given our primary objective to discover novel allosteric sites, these two well-characterized locations were excluded from further investigation. Site 7 occupies the intracellular cavity formed by TM2, TM3, and TM5\u0026ndash;TM7, spatially overlapping with the canonical G protein binding domain (Figure S16A, C). This evolutionarily conserved site across class A GPCRs has been extensively studied, with numerous allosteric modulators already developed and characterized. Due to the substantial existing research foundation, no additional exploration of this site was pursued. Site 10 localizes to the intracellular interface between TM3, TM4, and ICL2 (Figure S16B, D; S17A, B). Its structural formation relies on ICL2 conformational rearrangement coupled with intracellular TM3 displacement-features absent in the initial active-state structure due to ICL2 disorder and suboptimal TM3 positioning. Mechanistic studies suggest that allosteric modulation at this site could regulate ICL2-G protein interactions, enabling dual Gq/Gs signaling crosstalk as previously demonstrated in β2AR and GPR40\u003csup\u003e76, 77\u003c/sup\u003e (Figure S17C, D). Given the comprehensive prior investigations of this site, it was similarly excluded.\u003c/p\u003e \u003cp\u003eThrough systematic integration of empirical knowledge and MD simulations, we ultimately selected the dynamic, system-specific allosteric Site 4 as our primary candidate for experimental validation.\u003c/p\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003eExperimental validation of the predicted dynamic allosteric site\u003c/h2\u003e \u003cp\u003eAs the activated V2R predominantly exerts its physiological effects through Gs protein-mediated downstream signaling, we conducted alanine-scanning mutagenesis coupled with BRET assays to validate the functional role of Site 4 as a potential allosteric site of Gs protein signaling.\u003c/p\u003e \u003cp\u003eThe TRUPATH biosensor system\u003csup\u003e\u003cspan citationid=\"CR78\" class=\"CitationRef\"\u003e78\u003c/span\u003e, \u003cspan citationid=\"CR79\" class=\"CitationRef\"\u003e79\u003c/span\u003e\u003c/sup\u003e, utilizing BRET2 technology, was implemented to monitor receptor-G protein coupling dynamics. This methodology incorporates a NanoLuc energy donor fused to the Gα subunit and a GFP2 acceptor tagged to the Gγ subunit's N-terminus. Upon V2R activation, the heterotrimeric G protein complex (Gαβγ) undergoes conformational dissociation, releasing the Gα subunit to interact with the activated receptor. This structural reorganization reduces energy transfer efficiency between the donor-acceptor pair, reflecting a transition from high-affinity to low-affinity interaction states. The resultant BRET signal attenuation provides quantitative measurement of receptor-G protein coupling efficiency and dynamic interactions.\u003c/p\u003e \u003cp\u003eSite 4 comprises 10 residues, including Q225\u003csup\u003e5.58\u003c/sup\u003e, V226\u003csup\u003e5.59\u003c/sup\u003e, I228\u003csup\u003e5.61\u003c/sup\u003e, F229\u003csup\u003e5.62\u003c/sup\u003e, I232\u003csup\u003e5.65\u003c/sup\u003e, V266\u003csup\u003e6.30\u003c/sup\u003e, T269\u003csup\u003e6.33\u003c/sup\u003e, V270\u003csup\u003e6.34\u003c/sup\u003e, T273\u003csup\u003e6.37\u003c/sup\u003e, and L274\u003csup\u003e6.38\u003c/sup\u003e (Figure S15B). Systematic alanine substitution of each residue generated 10 distinct mutants, enabling comprehensive evaluation of their functional impact on V2R-mediated Gs protein signaling efficacy. Gene expression analysis (Figure S18) demonstrated comparable expression levels between all mutants and wild-type receptors, indicating the mutations did not affect receptor biosynthesis or stability. Subsequent BRET assays under ligand-free conditions (Fig.\u0026nbsp;\u003cspan refid=\"Fig10\" class=\"InternalRef\"\u003e10\u003c/span\u003eA-C) revealed that only the Q225A mutant exhibited substantial impairment (~\u0026thinsp;50% reduction) in basal Gs protein activity, suggesting its critical role in maintaining constitutive receptor activity. The remaining nine mutants showed no significant alterations in this unliganded state.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eLigand-dependent activation assays with AVP yielded distinct patterns (Fig.\u0026nbsp;\u003cspan refid=\"Fig10\" class=\"InternalRef\"\u003e10\u003c/span\u003eD-F): While V226A and I228A mutants maintained substantial Gs activation capacity (\u0026sim;15% reduction only), four mutants (F229A, I232A, V266A, T269A) showed severe activation deficits (~\u0026thinsp;85% reduction). Notably, Q225A and V270A mutants displayed impaired Gs activation (~\u0026thinsp;60% reduction), whereas T273A and L274A mutants were nearly incapable of Gs activation. These mutation-induced functional perturbations systematically mapped the structural determinants of orthosteric signaling, revealing key residues mediating allosteric transitions during receptor activation.\u003c/p\u003e \u003cp\u003eThe intracellular Site 4 shares striking topological similarity with the validated allosteric pocket in GPR88\u003csup\u003e80\u003c/sup\u003e (Figure S19), reinforcing its potential as a druggable domain. The hydrophobic character and membrane-proximal location of Site 4 align with established principles for intracellular allosteric modulator design\u003csup\u003e\u003cspan citationid=\"CR53\" class=\"CitationRef\"\u003e53\u003c/span\u003e\u003c/sup\u003e. Particularly compelling is its spatial correspondence to successful drug targets in class A GPCRs, suggesting conserved mechanisms of allosteric regulation exploitable for V2R-targeted therapeutic development. This integrated experimental and structural evidence positions Site 4 as a promising candidate for developing allosteric modulators capable of fine-tuning V2R signaling through selective stabilization of specific receptor conformations.\u003c/p\u003e \u003c/div\u003e"},{"header":"Discussion","content":"\u003cp\u003eCompared to conventional orthosteric drugs, allosteric modulators offer superior subtype selectivity and safety profiles, yet their development has been hindered by the challenges in identifying suitable allosteric sites\u003csup\u003e\u003cspan citationid=\"CR81\" class=\"CitationRef\"\u003e81\u003c/span\u003e, \u003cspan citationid=\"CR82\" class=\"CitationRef\"\u003e82\u003c/span\u003e\u003c/sup\u003e. Traditional approaches face significant limitations: experimental methods depend on serendipitous ligand discovery\u003csup\u003e\u003cspan citationid=\"CR68\" class=\"CitationRef\"\u003e68\u003c/span\u003e, \u003cspan citationid=\"CR83\" class=\"CitationRef\"\u003e83\u003c/span\u003e\u003c/sup\u003e, while computational predictions suffer from dataset biases (data-driven methods) or limited resolution (normal mode analyses)\u003csup\u003e\u003cspan additionalcitationids=\"CR85\" citationid=\"CR84\" class=\"CitationRef\"\u003e84\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR86\" class=\"CitationRef\"\u003e86\u003c/span\u003e\u003c/sup\u003e. Our framework addresses these limitations through three key aspects: (1) MD simulations capturing conformational ensembles across three distinct inactivation and activation states, (2) MSM construction enabling analysis of thermodynamic and kinetic properties while mapping free energy landscapes, and (3) integrated pocket prediction algorithms (fpocket) combined with MI-based residue importance ranking.\u003c/p\u003e \u003cp\u003eThe integration of MD simulations with MSMs establishes an advanced methodological framework that significantly enhances protein dynamics investigation and accelerates drug discovery processes. By systematically aggregating data from multiple short-timescale simulations\u003csup\u003e\u003cspan citationid=\"CR87\" class=\"CitationRef\"\u003e87\u003c/span\u003e, \u003cspan citationid=\"CR88\" class=\"CitationRef\"\u003e88\u003c/span\u003e\u003c/sup\u003e, MSMs achieve comprehensive reconstruction of free energy landscapes that resolve three critical aspects: metastable states, rare transition events, and long-term conformational evolution. This methodology provides unique dual perspectives on both thermodynamic stability and kinetic accessibility\u003csup\u003e\u003cspan citationid=\"CR89\" class=\"CitationRef\"\u003e89\u003c/span\u003e, \u003cspan citationid=\"CR90\" class=\"CitationRef\"\u003e90\u003c/span\u003e\u003c/sup\u003e, enabling: (1) identification of transient functional states such as cryptic allosteric sites, and (2) quantitative characterization of transition pathways between conformational states.\u003c/p\u003e \u003cp\u003eOur integrated computational-experimental framework has successfully identified a novel high-potential dynamic allosteric site in V2R, a critical therapeutic target for modulating fluid homeostasis. While current drug development efforts targeting V2R have predominantly focused on optimizing orthosteric ligands through structural modifications, there remains a critical gap in identifying viable allosteric sites for molecular intervention\u003csup\u003e\u003cspan citationid=\"CR91\" class=\"CitationRef\"\u003e91\u003c/span\u003e, \u003cspan citationid=\"CR92\" class=\"CitationRef\"\u003e92\u003c/span\u003e\u003c/sup\u003e. The discovery of this dynamic allosteric site presents new opportunities for developing selective allosteric modulators with potentially improved therapeutic profiles.\u003c/p\u003e \u003cp\u003eThis methodological framework demonstrates broad applicability beyond V2R, offering a systematic approach for identifying dynamic allosteric sites across various drug targets. However, the translation of these computational predictions into therapeutic agents faces significant challenges, particularly in the technical complexity of developing appropriate molecular probes for experimental validation. Notably, our previous research established methodological validity through mutagenesis studies that experimentally confirmed a predicted dynamic allosteric site\u003csup\u003e\u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e47\u003c/span\u003e, \u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e48\u003c/span\u003e\u003c/sup\u003e. Subsequent structure-based drug discovery efforts targeting this validated site yielded a novel allosteric modulator\u003csup\u003e\u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e47\u003c/span\u003e, \u003cspan citationid=\"CR93\" class=\"CitationRef\"\u003e93\u003c/span\u003e\u003c/sup\u003e, providing empirical validation of our predictive framework. These findings collectively demonstrate that our approach not only accelerates the identification of dynamic allosteric sites but also offers a viable pathway for translating computational predictions into bona fide therapeutic candidates.\u003c/p\u003e \u003cp\u003eIn summary, we established an integrated computational-experimental framework for identifying dynamic allosteric sites in protein receptors, employing V2R as a prototypical model system. Through synergistic application of MD simulations and MSMs, we systematically analyzed receptor conformational landscapes to reveal a novel allosteric site formed through V2R-specific activation-induced structural reconfiguration between TM5 and TM6. This dynamic allosteric site, identified as a central hub in the allosteric communication network through quantitative analysis of residue side-chain dynamics, was subsequently validated through experimental interrogation. Our methodology uniquely enables detection of dynamic allosteric sites independent of ligand binding constraints, thereby providing an accelerated path for developing allosteric modulators targeting conventionally challenging therapeutic targets. The presented framework - combining conformational ensemble analysis with MI-driven site prioritization - possesses inherent generalizability for application to diverse drug targets, offering transformative potential for expanding the repertoire of actionable allosteric sites in pharmacological discovery.\u003c/p\u003e"},{"header":"Methods","content":"\u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003ePreparation of simulation systems\u003c/h2\u003e \u003cp\u003eThe fully active V2R structure complexed with the ligand AVP, Gs, nanobody35, and single Fab chain (PDB ID: 7KH0) was downloaded from the Protein Data Bank (PDB). The nanobody35 and single Fab chain were removed and the truncated N-terminal, Gαs subunit, ECL2, ICL2, and ICL3 were recovered by a loop building program, according to the sequence of the WT V2R. The active V2R structure was derived from the extracted coordinates of the receptor and ligand in the fully active state. For the apo V2R structure, we utilized Swiss Model to perform homology modeling and it can be found in the alphafold2 database. The modeled structure exhibited a RMSD of 0.93 \u0026Aring; when compared to the highest pLDDT-scoring structure in the AlphaFold2 database. Referring to the common sequence in the solved structures, the 32\u0026ndash;342 residues were extracted as the input structures for the following processes.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec12\" class=\"Section2\"\u003e \u003ch2\u003eMD simulation settings\u003c/h2\u003e \u003cp\u003eAll structures were oriented using the Orientations of Proteins in Membranes server\u003csup\u003e\u003cspan citationid=\"CR94\" class=\"CitationRef\"\u003e94\u003c/span\u003e\u003c/sup\u003e. Subsequently, the structures were incorporated into a POPC membrane utilizing the CHARMM additive force field through the CHARMM-GUI platform\u003csup\u003e\u003cspan citationid=\"CR95\" class=\"CitationRef\"\u003e95\u003c/span\u003e\u003c/sup\u003e. TIP3P water molecules were introduced at both the top and bottom of the system, and counterions such as K⁺ or Cl⁻ were also included in the solvation process. The bilayer components have frequently been utilized in various other simulation studies\u003csup\u003e\u003cspan citationid=\"CR96\" class=\"CitationRef\"\u003e96\u003c/span\u003e\u003c/sup\u003e. Using the input generator from CHARMM-GUI, we generated the coordinate and topology files in Amber format.\u003c/p\u003e \u003cp\u003eFirstly, to prevent unrealistic collisions, the 2500 steepest descent cycles followed by 5000 conjugate gradient cycles was performed with the restraint of 10.0 kcal mol\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e \u0026Aring;\u003csup\u003e\u0026minus;2\u003c/sup\u003e on the V2R and 2.5 kcal mol\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e \u0026Aring;\u003csup\u003e\u0026minus;2\u003c/sup\u003e on the lipids. Secondly, the system underwent gradual heating from 0 to 300 K within the canonical ensemble (NVT) for 125 ps, with the restraint of 5.0 kcal mol\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e \u0026Aring;\u003csup\u003e\u0026minus;2\u003c/sup\u003e on the V2R and 2.5 kcal mol\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e \u0026Aring;\u003csup\u003e\u0026minus;2\u003c/sup\u003e on the lipids. Thirdly, four steps of equilibration molecular dynamics were performed in the isothermal isobaric (NPT) ensemble. The restrained force applied to all solute atoms was gradually decreased to 0 kcal\u0026middot; mol\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e \u0026Aring;\u003csup\u003e\u0026minus;2\u003c/sup\u003e to release all restraints. Then, the three states underwent 5 rounds of 1 \u0026micro;s MD simulations, with an integration step of 2.0 fs. In the end, we gathered 15 independent trajectories, each initiated with random velocities. The total simulation timescale was 15 \u0026micro;s. During the simulations, the Particle Mesh Ewald method was employed to compute long-range electrostatic interactions, while a cutoff of 9 \u0026Aring; was set for short-range electrostatic and van der Waals interactions. The SHAKE algorithm was used to handle covalent bonds involving hydrogen. A temperature of 310 K was maintained using a Langevin thermostat, with a collision frequency of 1.0 ps⁻\u0026sup1;. Snapshots were recorded every 100 ps.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec13\" class=\"Section2\"\u003e \u003ch2\u003eMSM construction\u003c/h2\u003e \u003cp\u003eAccording to the activation parameters, an MSM was built using the PyEMMA protocol (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://www.emma-project.org/latest/\u003c/span\u003e\u003cspan address=\"http://www.emma-project.org/latest/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e)\u003csup\u003e90\u003c/sup\u003e. By validating the implied timescales (Figure S4), we verified that the V2R systems were Markovian and dependable, utilizing a model of 300 microstates with a lag time of 0.07 ps (35 steps) and a maximum of 100 iterations for k-means clustering. Subsequently, the microstates were grouped into three metastates using the PCCA\u0026thinsp;+\u0026thinsp;algorithm, a process that was validated through a Chapman\u0026ndash;Kolmogorov test. Employing Transition Path Theory (TPT), we calculated the transition probability matrix for the MSM and determined the mean first passage time for each activation and inactivation event\u003csup\u003e\u003cspan citationid=\"CR97\" class=\"CitationRef\"\u003e97\u003c/span\u003e\u003c/sup\u003e. To obtain the most representative structure of each metastable state, we initially selected structures near the center of the metastates and compiled them into a condensed trajectory using the MDTraj package. Based on the new trajectories, we selected the representative snapshot of each metastate according to the pairwise similarity score S\u003csub\u003eij\u003c/sub\u003e:\u003cdiv id=\"Equ1\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equ1\" name=\"EquationSource\"\u003e\n$$\\:{S}_{ij}={e}^{{-d}_{ij}/{d}_{scale}}$$\u003c/div\u003e\u003cdiv class=\"EquationNumber\"\u003e1\u003c/div\u003e\u003c/div\u003e\u003c/p\u003e \u003cp\u003eHere, d\u003csub\u003eij\u003c/sub\u003e is the RMSD between snapshots i and j, and d\u003csub\u003escale\u003c/sub\u003e indicates the standard deviation of d. The snapshot with the highest similarity score was chosen as the most representative structure of each metastate.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec14\" class=\"Section2\"\u003e \u003ch2\u003eResidue fluctuation featurization and MSM construction\u003c/h2\u003e \u003cp\u003eTo create a residue fluctuation feature set for each dataset, we computed the absolute deviation of each residue's side chain in every frame compared to the first frame. This resulted in a collection of t vectors, each with a length of n, where n represents the number of residues and t denotes the total number of frames.\u003c/p\u003e \u003cp\u003eWe established MSM microstates by clustering the side-chain fluctuation feature representation. The clustering process utilized k-centers, which continuously added new cluster centers until the maximum within-cluster distance fell below the threshold of 3.6 nm. This threshold was determined based on the results of the implied timescales test (Figure S8). Next, we conducted ten rounds of k-medoids updates, accepting updates when the largest distance to the nearest medoid decreased. To estimate transition probabilities based on frame assignments to clusters, we initially created a transition count matrix, where the element C\u003csub\u003eij\u003c/sub\u003e represents the number of observed transitions from state i to state j. Subsequently, we incorporated a pseudocount of 1/n (where n is the number of states) into each element of the transition count matrix and performed row normalization to derive the transition probability matrix\u003csup\u003e\u003cspan citationid=\"CR98\" class=\"CitationRef\"\u003e98\u003c/span\u003e, \u003cspan citationid=\"CR99\" class=\"CitationRef\"\u003e99\u003c/span\u003e\u003c/sup\u003e. The lag times was 0.08 ps (40 steps), which was chosen by the implied timescales test (Figure S8). The optimal flux pathways between two sets of states were subsequently identified using transition path theory\u003csup\u003e\u003cspan citationid=\"CR100\" class=\"CitationRef\"\u003e100\u003c/span\u003e, \u003cspan citationid=\"CR101\" class=\"CitationRef\"\u003e101\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec15\" class=\"Section2\"\u003e \u003ch2\u003eComputation of the MI matrix\u003c/h2\u003e \u003cp\u003eBeginning with the feature representation of the representative conformation for each MSM state, we established four thresholds to classify each side chain within each state into five distinct fluctuation levels with equal frequency. This process resulted in a feature set for each MSM state, where each snapshot is depicted by a quinary vector containing one entry per residue, representing one of five potential values that correspond to different fluctuation levels. We then proceeded to compute the MI between each pair of residues, which serves as an indicator of the statistical dependence between two random variables. It is given by the equation:\u003cdiv id=\"Equ2\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equ2\" name=\"EquationSource\"\u003e\n$$\\:MI\\left(X,\\:Y\\right)=\\sum\\:_{y\\in\\:Y}\\sum\\:_{x\\in\\:X}p(x,y)\\text{l}\\text{o}\\text{g}\\left(\\frac{p(x,y)}{p\\left(x\\right)p\\left(y\\right)}\\right)$$\u003c/div\u003e\u003cdiv class=\"EquationNumber\"\u003e2\u003c/div\u003e\u003c/div\u003e\u003c/p\u003e \u003cp\u003ewhere X and Y denote any pair of residues, while x and y indicate the fluctuation states of the respective residues. The probability p (x) represents the likelihood of observing a residue in state x, and p (x, y) refers to the joint probability of x and y. These probabilities are derived from the equilibrium probabilities calculated during the MSM fitting process. Although other approaches could be employed to identify a set of representative structures and their equilibrium probabilities, MSM is particularly beneficial as they tend to provide more accurate estimates of true equilibrium probabilities in sets of finite-length trajectories. To compute MI matrices for V2R, we utilized its dimer symmetry to enhance the sampling of fluctuation states. If A\u003csub\u003ei\u003c/sub\u003e and B\u003csub\u003ei\u003c/sub\u003e​ are the random variables representing the fluctuation states of residue i from chains A and B, respectively, then due to the chemical identity of the two chains, at equilibrium, P (A\u003csub\u003ei\u003c/sub\u003e, B\u003csub\u003ej\u003c/sub\u003e)\u0026thinsp;=\u0026thinsp;P (A\u003csub\u003ej\u003c/sub\u003e, B\u003csub\u003ei\u003c/sub\u003e). To leverage this relationship for improved sampling of the state space and to enhance the robustness of our predictions against sampling errors, we calculate the mean of these two probabilities when determining MI.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec16\" class=\"Section2\"\u003e \u003ch2\u003eVisualization of the allosteric site network\u003c/h2\u003e \u003cp\u003eWe used fpocket to identify thirteen potential allosteric sites from the representative conformations of the active intermediate state and the fully active state. The residue composition of each site was calculated using a distance cutoff of 4 \u0026Aring;. Based on the residues assigned to each site, we calculate the MI between residues for each pair of sites and take the average to obtain a symmetric MI matrix between sites. In this matrix, each element represents the average MI value between site pairs. Then, a weighted network graph is constructed based on the site MI matrix, where each node represents a site, and the edge weights between nodes reflect the average MI between sites. Next, we calculate the eigenvector centrality of each node to assess the importance of sites within the network. Finally, sites are ranked by eigenvector centrality to identify those with higher centrality. Eigenvector centrality is a method used to assess node importance within a network, particularly suited for networks with complex weight distributions\u003csup\u003e\u003cspan citationid=\"CR102\" class=\"CitationRef\"\u003e102\u003c/span\u003e\u003c/sup\u003e. Its fundamental concept is that a node's importance depends not only on its own connections but also on the centrality of its neighboring nodes. A node linked to many high-centrality nodes will, in turn, have a higher centrality score. For a given network, let A represent its adjacency matrix (or weighted matrix), where A\u003csub\u003eij\u003c/sub\u003e denotes the connection weight between node i and node j. The eigenvector centrality c\u003csub\u003ei\u003c/sub\u003e of node i is defined as the vector that satisfies the following relationship:\u003cdiv id=\"Equ3\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equ3\" name=\"EquationSource\"\u003e\n$$\\:{c}_{i}=\\frac{1}{\\lambda\\:}\\sum\\:_{j=1}^{n}{A}_{ij}{c}_{j}$$\u003c/div\u003e\u003cdiv class=\"EquationNumber\"\u003e3\u003c/div\u003e\u003c/div\u003e\u003c/p\u003e \u003cp\u003ewhere n is the total number of nodes in the network and λ is the corresponding eigenvalue. λ is the largest eigenvalue, which is commonly used for normalization during the computation process.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec17\" class=\"Section2\"\u003e \u003ch2\u003eCell culture and site-directed mutagenesis\u003c/h2\u003e \u003cp\u003eHEK293T cells were obtained from ATCC. Cells were maintained, passaged and transfected in DMEM medium containing 10% FBS, 100 U/ml penicillin and 100 \u0026micro;g/ml streptomycin (Gibco-ThermoFisher) in a humidified atmosphere at 37\u0026deg;C and 5% CO\u003csub\u003e2\u003c/sub\u003e. After transfection, cells were plated in DMEM containing 1% dialyzed FBS, 100 U/ml penicillin, and 100 \u0026micro;g/ml streptomycin for cell surface expression assays and G protein dissociation.\u003c/p\u003e \u003cp\u003eAll V2R mutants used in the present study were generated by site-directed mutagenesis. The successful introduction of the mutations in the polymerase chain reaction products was verified by DNA sequencing.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec18\" class=\"Section2\"\u003e \u003ch2\u003eV2R cell surface expression\u003c/h2\u003e \u003cp\u003eHEK293 cells were transiently transfected with 100 ng of HiBiT-tagged WT or mutated V2R, which contained a HiBiT sequence and a linker at the N terminus (MVSGWRLFKKISGSSGGSSGGNSGGGS; gene synthesized with codon optimization), After 24 h, the cells were seeded on 96-well microplates at a density of 15000 cells per well, and incubated for 12 h at 37\u0026deg;C. Bring the microplate back to room temperature, cells were mixed with 50 \u0026micro;L of assay buffer consisting of 1:50 of a LgBiT stock solution (Promega) and 1:25 extracellular substrate stock solution (Promega). Cells were incubated for 8 min at room temperature, then the luminescence were recorded using a Synergy Neo microplate reader (BioTek).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec19\" class=\"Section2\"\u003e \u003ch2\u003eG protein dissociation assay\u003c/h2\u003e \u003cp\u003eGs (GαsS-RLuc8, Gβ3, Gγ9-GFP2) BRET probes were from the TRUPATH kit, which was a gift from Bryan Roth (Addgene kit #1000000163). HEK293 cells were transiently co-transfected with WT or mutated V2R along with specific G protein BRET probes according to the experimental setting. After 24 h, the cells were seeded on 96-well microplates at a density of 30,000\u0026ndash;50,000 cells per well, and incubated for an additional 24 h. For the constitutive activity measurement, cells transfected with varying amounts of WT or mutated V2R (200ng, 400ng, 600ng, 800ng and 1000ng/well) were washed once with assay buffer (1\u0026times; Hank\u0026rsquo;s balanced salt solution (HBSS)\u0026thinsp;+\u0026thinsp;20 mM HEPES, pH 7.4) and the BRET signal was directly recorded after the addition of 5 \u0026micro;M RLuc8 substrate coelenterazine-400a using a a Synergy Neo microplate reader (BioTek). For the AVP-stimulated G protein activation, the cells were washed once with assay buffer and stimulated with AVP at different concentrations. BRET signal was subsequently measured after the addition of the coelenterazine-400a and was calculated as the ratio of the GFP2 emission to RLuc8 emission.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec20\" class=\"Section2\"\u003e \u003ch2\u003eExperimental data analysis\u003c/h2\u003e \u003cp\u003eAll concentration-response curves were fit to a three-parameter logistic equation in Prism (Graphpad Software). BRET concentration-response curves were analyzed as either raw Net BRET (fit Emax-fit Baseline).\u003c/p\u003e \u003c/div\u003e"},{"header":"Declarations","content":"\u003cp\u003e \u003ch2\u003eConflicts of interest\u003c/h2\u003e \u003cp\u003eThe authors declare no conflicts of interest regarding this manuscript.\u003c/p\u003e \u003c/p\u003e\u003ch2\u003eAuthor Contributions\u003c/h2\u003e \u003cp\u003eJ.Z. and S.L. conceived and supervised the project. J.Z., S.L., X.Q., C.Z. and X.L. designed the experiments. X.Q. performed the computational experiments and analyzed the data. C.Z. and X.L. performed the biological experiments. X.Q. and C.Z. drafted the manuscript, and all authors contributed to specific parts of the manuscript. J.Z. and S.L. assumed responsibility for the manuscript in its entirety. M.L., N.L., N.Li, J.H. and N.Liu discussed the results and revised the manuscript.\u003c/p\u003e\u003ch2\u003eAcknowledgements\u003c/h2\u003e \u003cp\u003eThis study was supported by grants from the National Key R\u0026amp;D Program of China (No. 2023YFC3404700), the Noncommunicable Chronic Diseases-National Science and Technology Major Project (No. 2024ZD0531200), and the Innovative Research Team of High-Level Local Universities in Shanghai.\u003c/p\u003e\u003ch2\u003eData availability\u003c/h2\u003e \u003cp\u003eThe data that support this study are available from the corresponding authors upon reasonable request. Initial structures for MD simulation are obtained from the RCSB PDB database (PDB ID: 7KH0) (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.rcsb.org/\u003c/span\u003e\u003cspan address=\"https://www.rcsb.org/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) and Swiss model (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://swissmodel.expasy.org/\u003c/span\u003e\u003cspan address=\"https://swissmodel.expasy.org/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e). The analysis protocol for Markov State Model refers to \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://www.emma-project.org/latest/\u003c/span\u003e\u003cspan address=\"http://www.emma-project.org/latest/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. Pocket prediction is accomplished by fpocket, see \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://fpocket.sourceforge.net/\u003c/span\u003e\u003cspan address=\"http://fpocket.sourceforge.net/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. The calculation of mutual information for residue side chains is based on modules from \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://github.com/bowman-lab/enspara/tree/master/enspara\u003c/span\u003e\u003cspan address=\"https://github.com/bowman-lab/enspara/tree/master/enspara\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. Other simulation analyses were based on AMBER suite, according to \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://ambermd.org/\u003c/span\u003e\u003cspan address=\"http://ambermd.org/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. The code for calculating the allosteric network of sites can be referenced from \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://github.com/MaienLii/allosteric-site-network\u003c/span\u003e\u003cspan address=\"https://github.com/MaienLii/allosteric-site-network\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eFenton AW (2008) Allostery: an illustrated definition for the 'second secret of life'. Trends Biochem Sci 33:420\u0026ndash;425\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGoodey NM, Benkovic SJ (2008) Allosteric regulation and catalysis emerge via a common route. Nat Chem Biol 4:474\u0026ndash;482\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhang M et al (2024) G protein-coupled receptors (GPCRs): advances in structures, mechanisms, and drug discovery. Signal Transduct Target Ther 9:88\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNussinov R, Tsai CJ (2013) Allostery in disease and in drug discovery. Cell 153:293\u0026ndash;305\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJiang H et al (2017) Peptidomimetic inhibitors of APC-Asef interaction block colorectal cancer migration. Nat Chem Biol 13:994\u0026ndash;1001\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChangeux JP, Christopoulos A (2016) Allosteric Modulation as a Unifying Mechanism for Receptor Function and Regulation. Cell 166:1084\u0026ndash;1102\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLu S et al (2021) Mechanism of allosteric activation of SIRT6 revealed by the action of rationally designed activators. Acta Pharm Sin B 11:1355\u0026ndash;1361\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNi D et al (2022) Computational elucidation of allosteric communication in proteins for allosteric drug design. Drug Discov Today 27:2226\u0026ndash;2234\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZha J et al (2023) AlloReverse: multiscale understanding among hierarchical allosteric regulations. Nucleic Acids Res 51:W33\u0026ndash;W38\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhou Z et al (2012) Total glucosides of paeony attenuated functional maturation of dendritic cells via blocking TLR4/5 signaling in vivo. Int Immunopharmacol 14:275\u0026ndash;282\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDe Smet F, Christopoulos A, Carmeliet P (2014) Allosteric targeting of receptor tyrosine kinases. Nat Biotechnol 32:1113\u0026ndash;1120\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWootten D, Christopoulos A, Sexton PM (2013) Emerging paradigms in GPCR allostery: implications for drug discovery. Nat Rev Drug Discov 12:630\u0026ndash;644\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWu P, Clausen MH, Nielsen TE (2015) Allosteric small-molecule kinase inhibitors. Pharmacol Ther 156:59\u0026ndash;68\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDang CV, Reddy EP, Shokat KM, Soucek L (2017) Drugging the 'undruggable' cancer targets. Nat Rev Cancer 17:502\u0026ndash;508\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZorn JA, Wells JA (2010) Turning enzymes ON with small molecules. Nat Chem Biol 6:179\u0026ndash;188\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSun Z et al (2020) Covalent Inhibitors Allosterically Block the Activation of Rho Family Proteins and Suppress Cancer Cell Invasion. Adv Sci (Weinh) 7:2000098\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLazo JS, Sharlow ER (2016) Drugging Undruggable Molecular Cancer Targets. Annu Rev Pharmacol Toxicol 56:23\u0026ndash;40\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eXie X et al (2023) Recent advances in targeting the undruggable proteins: from drug discovery to clinical trials. Signal Transduct Target Ther 8:335\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHuang L, Guo Z, Wang F, Fu L (2021) KRAS mutation: from undruggable to druggable in cancer. Signal Transduct Target Ther 6:386\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCimermancic P et al (2016) CryptoSite: Expanding the Druggable Proteome by Characterization and Prediction of Cryptic Binding Sites. J Mol Biol 428:709\u0026ndash;719\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNi D et al (2021) Along the allostery stream: Recent advances in computational methods for allosteric drug discovery. WIREs Comput Mol Sci 12\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLu S, Ji M, Ni D, Zhang J (2018) Discovery of hidden allosteric sites as novel targets for allosteric drug design. Drug Discov Today 23:359\u0026ndash;365\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eOstrem JM, Peters U, Sos ML, Wells JA, Shokat KM (2013) K-Ras(G12C) inhibitors allosterically control GTP affinity and effector interactions. Nature 503:548\u0026ndash;551\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLu S et al (2020) Emergence of allosteric drug-resistance mutations: new challenges for allosteric drug discovery. Drug Discov Today 25:177\u0026ndash;184\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLatallo MJ, Cortina GA, Faham S, Nakamoto RK, Kasson PM (2017) Predicting allosteric mutants that increase activity of a major antibiotic resistance enzyme. Chem Sci 8:6484\u0026ndash;6492\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhang Q et al (2022) Targeting a cryptic allosteric site of SIRT6 with small-molecule inhibitors that inhibit the migration of pancreatic cancer cells. Acta Pharm Sin B 12:876\u0026ndash;889\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLu S et al (2019) Deactivation Pathway of Ras GTPase Underlies Conformational Substates as Targets for Drug Design. ACS Catal 9:7188\u0026ndash;7196\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNi D et al (2020) Discovery of cryptic allosteric sites using reversed allosteric communication by a combined computational and experimental strategy. Chem Sci 12:464\u0026ndash;476\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhang M, Jang H, Nussinov R (2019) The mechanism of PI3Kalpha activation at the atomic level. Chem Sci 10:3671\u0026ndash;3680\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eEvenseth LSM et al (2020) Exploring Conformational Dynamics of the Extracellular Venus flytrap Domain of the GABA(B) Receptor: A Path-Metadynamics Study. J Chem Inf Model 60:2294\u0026ndash;2303\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePowers AS et al (2024) A non-canonical mechanism of GPCR activation. Nat Commun 15:9938\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eOtun O et al (2024) Conformational dynamics underlying atypical chemokine receptor 3 activation. Proc Natl Acad Sci U S A 121:e2404000121\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHusic BE, Pande VS (2018) Markov State Models: From an Art to a Science. J Am Chem Soc 140:2386\u0026ndash;2396\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMcClendon CL, Friedland G, Mobley DL, Amirkhani H, Jacobson MP (2009) Quantifying Correlations Between Allosteric Sites in Thermodynamic Ensembles. J Chem Theory Comput 5:2486\u0026ndash;2502\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eStokely AM et al (2024) NetSci: A Library for High Performance Biomolecular Simulation Network Analysis Computation. J Chem Inf Model 64:7966\u0026ndash;7976\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJuul KV, Bichet DG, Nielsen S, Norgaard JP (2014) The physiological and pathophysiological functions of renal and extrarenal vasopressin V2 receptors. Am J Physiol Ren Physiol 306:F931\u0026ndash;940\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMakita N, Manaka K, Sato J, Iiri T (2020) V2 vasopressin receptor mutations. Vitam Horm 113:79\u0026ndash;99\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eStoof SC, Cnossen MH, de Maat MP, Leebeek FW, Kruip MJ (2016) Side effects of desmopressin in patients with bleeding disorders. Haemophilia 22:39\u0026ndash;45\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWang L et al (2022) Author Correction: Cryo-EM structure of the AVP-vasopressin receptor 2-Gs signaling complex. Cell Res 32:957\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhou Q et al (2019) Common activation mechanism of class A GPCRs. Elife 8\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhou F et al (2021) Molecular basis of ligand recognition and activation of human V2 vasopressin receptor. Cell Res 31:929\u0026ndash;931\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNatochin YV, Golosova DV (2020) Vasopressin receptor subtypes and renal sodium transport. Vitam Horm 113:239\u0026ndash;258\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMancinelli CD et al (2024) Control of G protein-coupled receptor function via membrane-interacting intrinsically disordered C-terminal domains. Proc Natl Acad Sci U S A 121:e2407744121\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eShenol A et al (2024) Molecular dynamics-based identification of binding pathways and two distinct high-affinity sites for succinate in succinate receptor 1/GPR91. Mol Cell 84:955\u0026ndash;966e954\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHeng J et al (2023) Function and dynamics of the intrinsically disordered carboxyl terminus of beta2 adrenergic receptor. Nat Commun 14:2005\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLi Y, Sun J, Li D, Lin J (2022) The full activation mechanism of the adenosine A(1) receptor revealed by GaMD and Su-GaMD simulations. Proc Natl Acad Sci U S A 119:e2203702119\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLu S et al (2021) Activation pathway of a G protein-coupled receptor uncovers conformational intermediates as targets for allosteric drug design. Nat Commun 12:4721\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLi M et al (2024) Delineating the stepwise millisecond allosteric activation mechanism of the class C GPCR dimer mGlu5. Nat Commun 15:7519\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHarrigan MP et al (2017) MSMBuilder: Statistical Models for Biomolecular Dynamics. Biophys J 112:10\u0026ndash;15\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDror RO et al (2011) Activation mechanism of the beta2-adrenergic receptor. Proc Natl Acad Sci U S A 108:18684\u0026ndash;18689\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMedeiros Almeida V et al (2021) Role of a high centrality residue in protein dynamics and thermal stability. J Struct Biol 213:107773\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLi M et al (2022) Mechanistic Insights into the Long-range Allosteric Regulation of KRAS Via Neurofibromatosis Type 1 (NF1) Scaffold Upon SPRED1 Loading. J Mol Biol 434:167730\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhang M, Lan X, Li X, Lu S (2023) Pharmacologically targeting intracellular allosteric sites of GPCRs for drug discovery. Drug Discov Today 28:103803\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLu S, Zhang J (2019) Small Molecule Allosteric Modulators of G-Protein-Coupled Receptors: Drug-Target Interactions. J Med Chem 62:24\u0026ndash;45\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eIgnatowska-Jankowska BM et al (2015) A Cannabinoid CB1 Receptor-Positive Allosteric Modulator Reduces Neuropathic Pain in the Mouse with No Psychoactive Effects. Neuropsychopharmacology 40:2948\u0026ndash;2959\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eQu Q et al (2023) Insights into distinct signaling profiles of the microOR activated by diverse agonists. Nat Chem Biol 19:423\u0026ndash;430\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNoe F, Clementi C (2015) Kinetic distance and kinetic maps from molecular dynamics simulation. J Chem Theory Comput 11:5002\u0026ndash;5011\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRanieri M et al (2019) Gain-of-function mutations of the V2 vasopressin receptor in nephrogenic syndrome of inappropriate antidiuresis (NSIAD): a cell-based assay to assess constitutive water reabsorption. Pflugers Arch 471:1291\u0026ndash;1304\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGilles N (2023) Natural Peptide Toxins as an Option for Renewed Treatment of Type 2 Vasopressin Receptor-Related Diseases. Biology (Basel) 12\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLiu HL et al (2024) Structural basis of tolvaptan binding to the vasopressin V(2) receptor. Acta Pharmacol Sin 45:2441\u0026ndash;2449\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKrumm BE et al (2023) Neurotensin Receptor Allosterism Revealed in Complex with a Biased Allosteric Modulator. Biochemistry 62:1233\u0026ndash;1248\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRing AM et al (2013) Adrenaline-activated structure of beta2-adrenoceptor stabilized by an engineered nanobody. Nature 502:575\u0026ndash;579\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRasmussen SG et al (2011) Structure of a nanobody-stabilized active state of the beta(2) adrenoceptor. Nature 469:175\u0026ndash;180\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYuan D et al (2020) Activation of the alpha(2B) adrenoceptor by the sedative sympatholytic dexmedetomidine. Nat Chem Biol 16:507\u0026ndash;512\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhang H et al (2015) Structure of the Angiotensin receptor revealed by serial femtosecond crystallography. Cell 161:833\u0026ndash;844\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWingler LM, McMahon C, Staus DP, Lefkowitz RJ, Kruse AC (2019) Distinctive Activation Mechanism for Angiotensin Receptor Revealed by a Synthetic Nanobody. \u003cem\u003eCell\u003c/em\u003e 176, 479\u0026ndash;490 e412\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eShao Z et al (2016) High-resolution crystal structure of the human CB1 cannabinoid receptor. Nature 540:602\u0026ndash;606\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYang X et al (2022) Molecular mechanism of allosteric modulation for the cannabinoid receptor CB1. Nat Chem Biol 18:831\u0026ndash;840\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFeinstein AG, Cole JL, May ER (2024) Dimerization Promotes PKR Activation by Modulating Energetics of alphaC Helix Conversion between Active and Inactive Conformations. J Phys Chem B 128:9305\u0026ndash;9314\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDemirtas K, Erman B, Haliloglu T (2024) Dynamic correlations: exact and approximate methods for mutual information. Bioinformatics 40\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHacisuleyman A, Erman B (2024) Synergy and anti-cooperativity in allostery: Molecular dynamics study of WT and oncogenic KRAS-RGL1. Proteins 92:665\u0026ndash;678\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSchmidtke P, Le Guilloux V, Maupetit J, Tuffery P (2010) fpocket: online tools for protein ensemble pocket detection and tracking. Nucleic Acids Res 38:W582\u0026ndash;589\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNegre CFA et al (2018) Eigenvector centrality for characterization of protein allosteric pathways. Proc Natl Acad Sci U S A 115:E12201\u0026ndash;E12208\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTekpinar M, Yildirim A (2022) Impact of dimerization and N3 binding on molecular dynamics of SARS-CoV and SARS-CoV-2 main proteases. J Biomol Struct Dyn 40:6243\u0026ndash;6254\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSchneider M, Antes I (2023) Comparison of allosteric signaling in DnaK and BiP using mutual information between simulated residue conformations. Proteins 91:237\u0026ndash;255\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHo JD et al (2018) Structural basis for GPR40 allosteric agonism and incretin stimulation. Nat Commun 9:1645\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLiu X et al (2019) Mechanism of beta(2)AR regulation by an intracellular positive allosteric modulator. Science 364:1283\u0026ndash;1287\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDiBerto JF, Olsen RHJ, Roth BL (2022) TRUPATH: An Open-Source Biosensor Platform for Interrogating the GPCR Transducerome. Methods Mol Biol 2525:185\u0026ndash;195\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eOlsen RHJ et al (2020) TRUPATH, an open-source biosensor platform for interrogating the GPCR transducerome. Nat Chem Biol 16:841\u0026ndash;849\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChen G et al (2022) Activation and allosteric regulation of the orphan GPR88-Gi1 signaling complex. Nat Commun 13:2375\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eOleinikovas V, Saladino G, Cossins BP, Gervasio FL (2016) Understanding Cryptic Pocket Formation in Protein Targets by Enhanced Sampling Simulations. J Am Chem Soc 138:14257\u0026ndash;14263\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLu S, Shen Q, Zhang JA (2019) Methods and Their Applications: Facilitating the Discovery of Allosteric Drugs and the Investigation of Allosteric Mechanisms. Acc Chem Res 52:492\u0026ndash;500\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLees JA et al (2023) An inverse agonist of orphan receptor GPR61 acts by a G protein-competitive allosteric mechanism. Nat Commun 14:5938\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePanjkovich A, Daura X (2014) PARS: a web server for the prediction of Protein Allosteric and Regulatory Sites. Bioinformatics 30:1314\u0026ndash;1315\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHuang M et al (2018) AlloFinder: a strategy for allosteric modulator discovery and allosterome analyses. Nucleic Acids Res 46:W451\u0026ndash;W458\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLiu X et al (2020) Unraveling allosteric landscapes of allosterome with ASD. Nucleic Acids Res 48:D394\u0026ndash;D401\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBowman GR, Huang X, Pande VS (2010) Network models for molecular kinetics and their initial applications to human health. Cell Res 20:622\u0026ndash;630\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNoe F, Fischer S (2008) Transition networks for modeling the kinetics of conformational change in macromolecules. Curr Opin Struct Biol 18:154\u0026ndash;162\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBeauchamp KA et al (2011) MSMBuilder2: Modeling Conformational Dynamics at the Picosecond to Millisecond Scale. J Chem Theory Comput 7:3412\u0026ndash;3419\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eScherer MK et al (2015) PyEMMA 2: A Software Package for Estimation, Validation, and Analysis of Markov Models. J Chem Theory Comput 11:5525\u0026ndash;5542\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eXiong X et al (2024) Long-Residence Time Peptide Antagonist for the Vasopressin V(2) Receptor to Treat Autosomal Dominant Polycystic Kidney Disease. J Med Chem 67:5935\u0026ndash;5944\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCao X et al (2022) Benzodiazepine Derivatives as Potent Vasopressin V(2) Receptor Antagonists for the Treatment of Autosomal Dominant Kidney Disease. J Med Chem 65:9295\u0026ndash;9311\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhang D et al (2023) Structural insights into angiotensin receptor signaling modulation by balanced and biased agonists. EMBO J 42:e112940\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLomize MA, Pogozheva ID, Joo H, Mosberg HI, Lomize AL (2012) OPM database and PPM web server: resources for positioning of proteins in membranes. Nucleic Acids Res 40:D370\u0026ndash;376\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJo S et al (2017) CHARMM-GUI 10 years for biomolecular modeling and simulation. J Comput Chem 38:1114\u0026ndash;1124\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMattedi G, Acosta-Gutierrez S, Clark T, Gervasio FL (2020) A combined activation mechanism for the glucagon receptor. Proc Natl Acad Sci U S A 117:15414\u0026ndash;15422\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eE W, Vanden-Eijnden E (2006) Towards a Theory of Transition Paths. J Stat Phys 123:503\u0026ndash;523\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZimmerman MI, Porter JR, Sun X, Silva RR, Bowman GR (2018) Choice of Adaptive Sampling Strategy Impacts State Discovery, Transition Probabilities, and the Apparent Mechanism of Conformational Changes. J Chem Theory Comput 14:5459\u0026ndash;5475\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePrinz JH et al (2011) Markov models of molecular kinetics: generation and validation. J Chem Phys 134:174105\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMetzner P, Sch\u0026uuml;tte C, Vanden-Eijnden E (2009) Transition Path Theory for Markov Jump Processes. Multiscale Model Simul 7:1192\u0026ndash;1219\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eE W, Vanden-Eijnden E (2010) Transition-path theory and path-finding algorithms for the study of rare events. Annu Rev Phys Chem 61:391\u0026ndash;420\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFoutch D, Pham B, Shen T (2021) Protein conformational switch discerned via network centrality properties. Comput Struct Biotechnol J 19:3599\u0026ndash;3608\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-6427090/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-6427090/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eAllostery governs\u0026zwnj; the functional dynamics of proteins by regulating their conformational transitions. \u0026zwnj;The development of allosteric modulators has emerged as a promising therapeutic strategy\u0026zwnj;, leveraging their superior target specificity \u0026zwnj;and reduced off-target effects compared to orthosteric compounds\u0026zwnj;. \u0026zwnj;A critical barrier in this field remains\u0026zwnj; the identification of dynamic allosteric sites, \u0026zwnj;which are often undetectable in conventional structural analyses due to their transient nature\u0026zwnj;. \u0026zwnj;To address this challenge,\u0026zwnj; we established \u0026zwnj;an integrative computational framework\u0026zwnj; combining molecular dynamics (MD), Markov state modeling (MSM), and mutual information (MI) analysis \u0026zwnj;to probe\u0026zwnj; dynamic allosteric sites \u0026zwnj;in the class A G protein-coupled receptor (GPCR) prototype, vasopressin V2 receptor (V2R)\u0026zwnj;. \u0026zwnj;Through\u0026zwnj; multi-replica MD simulations, \u0026zwnj;we reconstructed\u0026zwnj; the receptor's conformational landscape, \u0026zwnj;which was statistically refined\u0026zwnj; via MSM \u0026zwnj;to resolve\u0026zwnj; equilibrium populations \u0026zwnj;and transition kinetics\u0026zwnj;. \u0026zwnj;Key mechanistic features\u0026zwnj; of activation-related structural motifs \u0026zwnj;were quantitatively characterized\u0026zwnj;. \u0026zwnj;Candidate allosteric sites were systematically ranked\u0026zwnj; through MI-driven residue interaction network analysis, \u0026zwnj;prioritizing\u0026zwnj; pharmacologically targetable regions. \u0026zwnj;This strategy revealed\u0026zwnj; a novel dynamic allosteric site \u0026zwnj;on the V2R intracellular interface\u0026zwnj;, \u0026zwnj;whose functional relevance was confirmed through\u0026zwnj; structure-guided mutagenesis \u0026zwnj;and\u0026zwnj; BRET-based signaling assays. \u0026zwnj;Our findings\u0026zwnj; not only \u0026zwnj;elucidate the allosteric activation mechanism of V2R at atomic resolution\u0026zwnj; but also \u0026zwnj;establish a conformation-aware platform\u0026zwnj; for \u0026zwnj;rational discovery of dynamic binding pockets\u0026zwnj;, \u0026zwnj;providing a transformative approach for\u0026zwnj; GPCR-targeted drug discovery.\u003c/p\u003e","manuscriptTitle":"Conformational mapping of GPCR activation: dynamic allosteric site discovery in V2R through MD-MSM and mutual information analysis","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-05-02 04:07:56","doi":"10.21203/rs.3.rs-6427090/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"a573a06f-2ae7-45a1-b19b-333c715531d1","owner":[],"postedDate":"May 2nd, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":47145974,"name":"Biological sciences/Biophysics/Computational biophysics"},{"id":47145975,"name":"Biological sciences/Computational biology and bioinformatics/Protein function predictions"},{"id":47145976,"name":"Physical sciences/Chemistry/Theoretical chemistry/Molecular dynamics"}],"tags":[],"updatedAt":"2025-05-29T14:11:36+00:00","versionOfRecord":[],"versionCreatedAt":"2025-05-02 04:07:56","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-6427090","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-6427090","identity":"rs-6427090","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}