Title: The structural basis of functional selectivity in the neuropeptide Y receptor Y2R

Project duration: 09/2023 – 08/2026

Research Area: Computational Biochemistry and Biophysics

The neuropeptide Y receptor subtype 2 (Y2R) is a promising target for the treatment of obesity and cancer. The specific activation of cellular signaling pathways poses a significant challenge, necessitating the development of drugs with minimal side effects.
Our project aims to investigate the phenomenon of functional selectivity, where biased ligands induce receptor interaction with specific intracellular transducers, initiating certain physiological responses. Using techniques from biophysics, computational biochemistry, and deep learning, we explore functional selectivity in Y2R.
DEER spectroscopy provides insights into the structure and population of protein conformations. We integrate experimental restraints into atomically precise protein structure modeling using novel AI approaches. The resulting conformations will undergo refinement and validation using physics-based computational methods, such as molecular dynamics (MD) simulation.

Aims

• Obtaining DEER results for Y2R conformations associated with functional selectivity
• Developing and validating an AI-based computational approach that integrates DEER spectroscopy data into the directed modeling process of protein conformation structure
• Constructing structural models of Y2R receptor conformations associated with functional selectivity using the developed approach
• Preparing a description of our integrative computational approach for the research community

Problems

Development of drugs with minimal side effects requires knowledge about the structure of protein conformation associated with functional selectivity. DEER spectroscopy provides insights into the structure and populations of such conformations. However, there is currently no established protocol for converting DEER distance distributions into a structural model of protein conformation. Therefore, it is necessary to develop and validate a computationally efficient approach for this task.

Practical Example

One of the main results of the work is an atomistically precise model of the Y2R conformation coupled with functional selectivity. Subsequently, this model can be used as a target for the design of therapeutics with minimal side effects. In addition, we will develop a computational approach using deep learning and computational biochemistry methods to construct and validate structural models of protein conformations derived from DEER spectroscopy experiments.

Technology

For the structural modeling of protein conformations guided by experimental DEER distance restraints, we will employ the AlphaLink approach. This will generate several thousand models, from which a few representative variants will be chosen via cluster analysis. The stability assessment and refinement of these models will be carried out through MD simulation replicas. Analysing the simulation results will include calculating RMSD, simulating DEER data, and comparing them with the experiment.

Outlook

Our project aims to establish a robust methodological framework by integrating experimental data with computationally efficient AI methods and physics-based computational biochemistry techniques.Additionally, we will construct an atomistically precise structural model of Y2R conformations linked to functional selectivity. These models can be utilised by other research groups in developing drugs targeting cancer or obesity with minimal adverse effects.
In addition to our current objectives, we are proposing a future direction for the project. This new focus will address a key challenge encountered in interpreting DEER experimental results. The challenge arises from the potential blurring of protein conformational states due to spin label motion, particularly notable with flexible spin labels. To tackle this issue, we plan to develop a novel approach that integrates deep learning methods and enhanced sampling techniques in MD simulations.

Team

Lead

• Prof. Dr. Jens Meiler (University Leipzig)
• Prof. Dr. Sayan Mukherjee (University Leipzig, Duke University)

Team Members

• Julia Belyaeva
• Dr. Matthias Elgeti