TL;DR
This paper introduces a novel computational and experimental approach to stabilize membrane proteins, specifically G protein-coupled receptors (GPCRs), which are notoriously difficult to study due to their inherent flexibility. By identifying and reinforcing "metastable" regions within these receptors, the researchers successfully engineered significantly more stable variants of the beta1-adrenergic receptor, paving the way for better structural studies and drug development.
What Problem Does This Paper Address?
Membrane proteins, which are crucial for cellular signaling and are targets for over 60% of current drugs, are severely underrepresented in structural databases. This is largely due to their inherent flexibility and instability when extracted from their native membrane environment for structural analysis. G protein-coupled receptors (GPCRs), a major class of membrane proteins, signal by undergoing conformational changes. This flexibility, while essential for function, makes them challenging to crystallize and study. Current methods for stabilizing membrane proteins are often time-consuming, labor-intensive, and can disrupt crucial functional regions. The paper aims to address this by developing a more efficient and rational approach to identify and stabilize these flexible regions.
Methods
The researchers developed an integrated computational and experimental strategy.
- Bioinformatics and Sequence Analysis: They began by analyzing multiple sequence alignments of class A GPCRs to identify residues with low sequence conservation. This is based on the hypothesis that naturally evolved receptors might be "metastable," meaning they exist in multiple, closely related energy states, and that stabilizing these states could be achieved by reinforcing suboptimal interactions.
- Computational Structure Analysis: Using existing crystal structures of GPCRs (specifically the beta1-adrenergic receptor, B1AR), they identified two types of "metastable" regions:
- Nonconserved unsatisfied polar residues: These are polar amino acids that are not well-stabilized by hydrogen bonds or electrostatic interactions, suggesting they could be points of flexibility.
- Regions exhibiting packing defects: These are areas within the protein's core where amino acids don't fit together optimally, leading to instability.
- Computational Protein Design (RosettaMembrane): The RosettaMembrane software was employed to computationally redesign these identified metastable regions. The goal was to predict amino acid substitutions that would minimize the free energy of the receptor in its inactive state, thereby stabilizing it. The design process involved exploring multiple substitutions simultaneously.
- Experimental Validation: Promising computational designs were then experimentally tested. The primary assay was measuring the apparent melting temperature (Tm) of the receptor variants using radioligand binding assays. Higher Tm values indicate greater conformational stability.
- Iterative Design: The feedback from experimental results was used to refine subsequent computational design rounds, allowing for the combination of locally stabilized regions into more hyperstable variants.
Datasets
- Class A GPCR sequences: Used for multiple sequence alignments to identify conserved and non-conserved residues.
- Crystal structures of GPCRs: Specifically, the beta1-adrenergic receptor (B1AR) structure (PDB ID: 18) was used as the primary model for computational design. Other class A GPCR structures were used for broader sequence and structural analysis.
- Experimental data on B1AR-M23: A previously stabilized variant of B1AR was used as a benchmark for comparison.
- Experimental data on A2AR Rant receptor variants: Used to validate the predictive power of the method for experimentally identified stabilizing mutations.
Key Insights and Ideas
- Natural GPCRs are inherently metastable: The findings strongly suggest that naturally evolved GPCR sequences are not optimized for maximal stability but rather for a degree of conformational flexibility, allowing them to exist in multiple functional states. This metastability is achieved through suboptimal polar and van der Waals interactions.
- Targeting nonconserved, unsatisfied polar residues and packing defects is effective: These specific types of regions are key determinants of metastability and can be rationally targeted for stabilization.
- Computational design can identify novel stabilizing interaction networks: The approach can uncover combinations of mutations that work synergistically to enhance stability, going beyond single-point mutations.
- Stabilization can be achieved at both the core and surface: The method proved effective in stabilizing regions buried within the receptor's core as well as surface-exposed residues.
- The method can modulate pharmacological properties: Beyond just stabilizing the receptor, the designed mutations also altered agonist binding affinity, demonstrating the potential to reprogram receptor function.
Key Results
- 13 novel B1AR variants stabilized: The approach successfully generated 13 variants of the beta1-adrenergic receptor stabilized in its inactive conformation.
- Hyperstabilized variants: Two variants, M23-5m and M23-7m, showed significantly increased apparent thermostability. M23-7m exhibited a Tm increase of over 30°C compared to wild-type B1AR and over 11°C compared to the previously stabilized M23 variant.
- Predictive power: The method could predict the stabilizing effects of 70% of substitutions previously identified through experimental scanning mutagenesis for B1AR-M23 and A2AR Rant variants.
- Modulation of agonist affinity: The designed variants showed a significant loss in agonist affinity, with EC50 values shifting by up to 396-fold compared to wild-type B1AR.
Limitations and Future Work
- Accuracy of structural models: The authors acknowledge that the success of structure-based design relies on the accuracy of the starting structural models.
- Generalizability to other membrane proteins: While applied to GPCRs, the method's applicability to other classes of membrane proteins would require further validation.
- Future Work: The authors suggest that this approach can be extended to design receptors with reprogrammed signaling properties, to create thermostabilized molecular sensors, and to engineer membrane-embedded scaffolds. They also propose developing a multistate design framework to predict and engineer conformational regulation of receptor function in diverse states.
Summary updated