Biology subjects

Mukhaleva, E.

Publications and source records attributed to Mukhaleva, E..

11 recordsLinked to original sources

bioRxiv2026

Interpretable Machine Learning Model of Receptor Dynamics Reveals AT1R Allostery and a Negative Allosteric Modulator

Allosteric modulation of G protein-coupled receptors (GPCRs) offers major advantages in receptor selectivity and signaling control; yet systematic approaches to identify allosteric modulators, define their binding sites, and map the underlying allosteric networks remain limited. Current molecular dynamics (MD) and machine learning (ML)-based methods often rely on correlation-driven or black-box models that provide limited mechanistic insight. We developed an interpretable probabilistic framework that extracts residue-level dependencies from MD ensembles using Bayesian network modeling (BNM). By representing each residue through its local interaction energy, BNM identifies both local and long-range energetic couplings and maps the allosteric communication pathways linking the AngII binding site to the G-protein interface in the angiotensin II type 1 receptor (AT1R). To functionally prioritize these pathways, we integrated BNM with comprehensive mutational analysis, combining whole-receptor alanine mutagenesis data with exhaustive in silico deep mutational scanning to validate BNM-predicted hotspots. This approach recovered state-dependent allosteric communities, revealed residues in noncanonical regions that regulate Gq coupling and identified positions whose functional importance emerged only with specific, predicted substitutions, as well as highlighted a cryptic intracellular pocket enriched in communication hubs. Guided by these network-derived residues and pocket geometries, structure-based virtual screening identified a small, fragment-like molecule negative allosteric modulator (NAM) named Q2 that attenuates AngII-mediated Gq signaling. Mutational mapping supports Q2 binding adjacent to the G-protein interface, consistent with its mechanism of action. Together, these results establish a generalizable and interpretable framework for uncovering GPCR allosteric communication networks and discovering modulators that exploit these networks.

molecular biology↗

bioRxiv2025

Allostery between distant structural regions dictates selectivity in GPCR:G protein coupling.

Selectivity in GPCR-G protein coupling is widely regarded as a solved problem that derives from a combination of structural hotspot pairings and dynamics at the protein-protein interface. Here, using interpretable machine learning Bayesian Network model with Molecular Dynamics simulations and experiments, we reveal the influence of distant residue communities within the G protein core on coupling selectivity. We observed distinct cooperative hotspot residues across different G protein subtypes, including key regions such as the N-terminus, h4s6 loop, and H5 helix. These results demonstrate the intricate allosteric dependencies between the core and the H5 helix in stabilizing selective interactions. The functional significance of these cooperative regions is validated through subtype swapping mutations. By introducing targeted Gq like mutations in the Gs core, we successfully altered receptor coupling profile to signal through Gq. Our findings emphasize that cooperative interactions in the G core are not only crucial for selectivity but can also be leveraged to engineer G proteins with tailored coupling preferences. TeaserAllosteric interactions in the G protein core shape GPCR selectivity and can be engineered to rewire receptor signaling preferences.

biophysics↗

bioRxiv2025

DRUMBEAT: Temporally resolved interpretable machine learning model for characterizing state transitions in protein dynamics

Conformational transitions are central to protein function, yet their mechanistic analysis remains challenging due to the multi-dimensionality and timescales underlying the molecular motions. While interpretable network models such as Bayesian networks have advanced the identification of key residue interactions in molecular dynamics (MD) data, they lack temporal resolution and cannot capture the sequence of events during transitions. Here, we introduce Dynamically Resolved Universal Model for BayEsiAn network Tracking or DRUMBEAT, a machine learning approach that combines a universal graph topology with sliding-window rescoring to generate interpretable, time-resolved maps of cooperative events in MD trajectories. Applying DRUMBEAT to the benchmark Fip35 WW domain folding trajectories from DE Shaw Research Group, we recover both major folding pathways and critical residues previously highlighted by experiment. Importantly, DRUMBEAT provides new insight in two ways: (1) uncover unknown protein features important for transition, and (2) dissect the order and timing of conformational changes, revealing the precise sequence of residue contact closures during individual folding events. Robustness analysis demonstrates that both the universal graph and time-resolved results are highly consistent across multiple sampling replicates. These findings establish DRUMBEAT as a scalable and interpretable machine learning framework for dissecting the dynamics of protein folding and other conformational transitions, offering a generalizable tool for the mechanistic study of biomolecular dynamics.

biochemistry↗

bioRxiv2025

Protein Frustration Reveals Active Sites in Co-Evolved GPCR:G Protein Complexes and in Engineered Targeted Degrader Complexes

The folded structure of a protein is understood to be an optimal energy state. However, previous studies have shown that certain amino acid residue positions that play a critical role in protein function are often in a suboptimal energy state or "frustrated". Here, we leverage over 1200 three-dimensional structures of G protein-coupled receptors (GPCRs) to demonstrate that residues at the interface between GPCR and its ligand or G protein contain a higher density of frustrated residues compared to other structural regions in the receptor. Likewise, the G subunit of the trimeric G proteins shows multiple clusters of highly frustrated residues on its surface that overlap with their effector protein (G{beta}{gamma}, RGS, Adenyl cyclase, Ric8) binding interfaces. Compared to the co-evolved GPCR:G protein complexes, engineered protein complexes, such as those facilitated by molecular degraders, show a much greater density of highly frustrated residues in the degrader interface. Our study highlights the use of protein frustration as an invaluable tool to evaluate both native protein-protein interfaces and design strategies to facilitate engineered protein complexes.

biophysics↗

bioRxiv2025

Temporally Resolved and Interpretable Machine Learning Model of GPCR Conformational Transition

Identifying target-specific drugs remains a challenge in pharmacology, especially for highly homologous proteins such as dopamine receptors D2R and D3R. Differences in target-specific cryptic druggable sites for such receptors arise from the distinct conformational ensembles underlying their dynamic behavior. While Molecular Dynamics (MD) simulations has emerged as a powerful tool for dissecting protein dynamics, the sheer volume of MD data requires scalable and unbiased data analysis strategies to pinpoint residue communities regulating conformational state ensembles. We have developed the Dynamically Resolved Universal Model for BayEsiAn network Tracking (DRUMBEAT) interpretable machine learning algorithm and validated it by identifying residue communities that enable the deactivation of the {beta}2-adrenergic receptor. Further, upon analyzing dopamine receptor dynamics we identified distinct and non-conserved residue communities around the contacts F1704.62_F172ECL2 and S1464.38_G14134.56 that are specific to D3R conformational transitions compared to D2R. This information can be tapped to design subtype-specific drugs for neuropsychiatric and substance use disorders.

biochemistry↗

bioRxiv2025

Insights From Protein Frustration Analysis Of BRD4-Cereblon Degrader Ternary Complexes Show Separation of Strong from Weak Degraders

PROteolysis TArgeting Chimeras (PROTACs), also known as Ligand-Directed Degraders (LDDs), are an innovative class of small molecules that leverage the ubiquitin-proteasome system to induce the degradation of target proteins. Structure based design methods are not readily applicable for designing LDDs due to the dynamic nature of the ternary complexes. This study investigates the dynamic properties of five LDD-mediated BRD4-Cereblon complexes, focusing on the challenges of evaluating linker efficiency due to the difficulty in identifying suitable computational metrics that correlate well with the cooperativity or degradation propensity of LDDs. We uncovered that protein frustration, a concept originally developed to understand protein folding, calculated for the residues in the protein-protein interface of the LDD-mediated ternary complexes recapitulate the strength of degradation of the LDDs. Our findings indicated that hydrophobic residues in the interface are among the highly frustrated residues pairs, and they are crucial in distinguishing strong degraders from weak ones. By analyzing frustration patterns, we identified key residues and interactions critical to the effectiveness of the ternary complex. These insights provide practical guidelines for designing and prioritizing more efficient degraders, paving the way for the development of next-generation LDDs with improved therapeutic potential.

biophysics↗

bioRxiv2024

BaNDyT: Bayesian Network modeling of molecular Dynamics Trajectories

Bayesian network modeling (BN modeling, or BNM) is an interpretable machine learning method for constructing probabilistic graphical models from the data. In recent years, it has been extensively applied to diverse types of biomedical datasets. Concurrently, our ability to perform long-timescale molecular dynamics (MD) simulations on proteins and other materials has increased exponentially. However, the analysis of MD simulation trajectories has not been data-driven but rather dependent on the users prior knowledge of the systems, thus limiting the scope and utility of the MD simulations. Recently, we pioneered using BNM for analyzing the MD trajectories of protein complexes. The resulting BN models yield novel fully data-driven insights into the functional importance of the amino acid residues that modulate proteins function. In this report, we describe the BaNDyT software package that implements the BNM specifically attuned to the MD simulation trajectories data. We believe that BaNDyT is the first software package to include specialized and advanced features for analyzing MD simulation trajectories using a probabilistic graphical network model. We describe here the softwares uses, the methods associated with it, and a comprehensive Python interface to the underlying generalist BNM code. This provides a powerful and versatile mechanism for users to control the workflow. As an application example, we have utilized this methodology and associated software to study how membrane proteins, specifically the G protein-coupled receptors, selectively couple to G proteins. The software can be used for analyzing MD trajectories of any protein as well as polymeric materials.

bioinformatics↗

bioRxiv2023

Cellular Lipids Regulate the Conformational Ensembles of the Disordered Intracellular Loop 3 in beta-2 Adrenergic Receptor

The structurally disordered intracellular loops (ICLs) of G protein-coupled receptors (GPCRs) play a critical role in G protein coupling. In our previous work, we used a combination of FRET-based and computational methodologies to show that the third intracellular loop (ICL3) modulates the activity and G protein coupling selectivity in GPCRs. In the current study, we have uncovered the role of several lipid components in modulating the conformational ensemble of ICL3 of the {beta}2-adrenergic receptor ({beta}2AR). Our findings indicate that phosphatidylinositol 4,5-bisphosphate (PIP2) in the inner leaflet of the membrane bilayer acts as a stabilizing anchor for ICL3, opening the intracellular cavity to facilitate G protein coupling. This interaction between PIP2 and ICL3 causes tilting of {beta}2AR within the cellular membrane. Notably, this tilting of the receptor is supported by ganglioside GM3 stabilizing the extracellular loops on the outer leaflet of the bilayer, thereby exerting an allosteric effect on the orthosteric ligand binding pocket. Our results underscore the significance of lipids in modulating GPCR activity, proposing an allosteric mechanism that occurs through the receptors orientation within the membrane.

biophysics↗

bioRxiv2023

Bayesian network models identify co-operative GPCR:G protein interactions that contribute to G protein coupling

Cooperative interactions in protein-protein interfaces demonstrate the interdependency or the linked network-like behavior of interface interactions and their effect on the coupling of proteins. Cooperative interactions also could cause ripple or allosteric effects at a distance in protein-protein interfaces. Although they are critically important in protein-protein interfaces it is challenging to determine which amino acid pair interactions are cooperative. In this work we have used Bayesian network modeling, an interpretable machine learning method, combined with molecular dynamics trajectories to identify the residue pairs that show high cooperativity and their allosteric effect in the interface of G protein-coupled receptor (GPCR) complexes with G proteins. Our results reveal a strong co-dependency in the formation of interface GPCR:G protein contacts. This observation indicates that cooperativity of GPCR:G protein interactions is necessary for the coupling and selectivity of G proteins and is thus critical for receptor function. We have identified subnetworks containing polar and hydrophobic interactions that are common among multiple GPCRs coupling to different G protein subtypes (Gs, Gi and Gq). These common subnetworks along with G protein-specific subnetworks together confer selectivity to the G protein coupling. This work underscores the potential of data-driven Bayesian network modeling in elucidating the intricate dependencies and selectivity determinants in GPCR:G protein complexes, offering valuable insights into the dynamic nature of these essential cellular signaling components.

biophysics↗

bioRxiv2023

Molecular recognition of an aversive odorant by the murine trace amine-associated receptor TAAR7f

There are two main families of G protein-coupled receptors that detect odours in humans, the odorant receptors (ORs) and the trace amine-associated receptors (TAARs). Their amino acid sequences are distinct, with the TAARs being most similar to the aminergic receptors such as those activated by adrenaline, serotonin and histamine. To elucidate the structural determinants of ligand recognition by TAARs, we have determined the cryo-EM structure of a murine receptor, mTAAR7f, coupled to the heterotrimeric G protein Gs and bound to the odorant N,N-dimethylcyclohexylamine (DMCH) to an overall resolution of 2.9 [A]. DMCH is bound in a hydrophobic orthosteric binding site primarily through van der Waals interactions and a strong charge-charge interaction between the tertiary amine of the ligand and an aspartic acid residue. This site is distinct and non-overlapping with the binding site for the odorant propionate in the odorant receptor OR51E2. The structure, in combination with mutagenesis data and molecular dynamics simulations suggests that the activation of the receptor follows a similar pathway to that of the {beta}-adrenoceptors, with the significant difference that DMCH interacts directly with one of the main activation microswitch residues.

molecular biology↗

bioRxiv2023

An Ionic Lock Mediates Interdomain Interactions in G protein αi subunits and Determines Gαi Subtype Signaling Specificity

Highly homologous members of the Gi family, Gi1-3, have distinct tissue distributions and physiological functions, yet the functional properties of these proteins with respect to GDP/GTP binding and regulation of adenylate cyclase are very similar. We recently identified PDZ-RhoGEF (PRG) as a novel Gi1 effector, however, it is poorly activated by Gi2. Here, in a proteomic proximity labeling screen we observed a strong preference for Gi1 relative to Gi2 with respect to engagement of a broad range of potential targets. We investigated the mechanistic basis for this selectivity using PRG as a representative target. Substitution of either the helical domain (HD) from Gi1 into Gi2 or substitution of a single amino acid, A230 in Gi2 to the corresponding D in Gi1, largely rescues PRG activation and interactions with other Gi targets. Molecular dynamics simulations combined with Bayesian network models revealed that in the GTP bound state, dynamic separation at the HD-Ras-like domain (RLD) interface is prevalent in Gi2 relative to Gi1 and that mutation of A230s4h3.3 to D in Gi2 stabilizes HD-RLD interactions through formation of an ionic interaction with R145HD.11 in the HD. These interactions in turn modify the conformation of Switch III. These data support a model where D229s4h3.3 in Gi1 interacts with R144HD.11 stabilizes a network of interactions between HD and RLD to promote protein target recognition. The corresponding A230 in Gi2 is unable to form the "ionic lock" to stabilize this network leading to an overall lower efficacy with respect to target interactions. This study reveals distinct mechanistic properties that could underly differential biological and physiological consequences of activation of Gi1 or Gi2 by GPCRs.

pharmacology and toxicology↗