Predicting Pose Distribution of Protein Domains Connected by Flexible Linkers Is an Unsolved Problem

Area of Science:

• Structural Biology
• Computational Biology
• Biophysics

Background:

• Protein domain-linker-domain (D-L-D) protein conformational ensembles influence function, allostery, and binding thermodynamics.
• Predicting these conformational distributions is crucial for understanding protein behavior.
• The CASP16 Conformational Ensembles Experiment specifically addressed this challenge.

Purpose of the Study:

• To assess the accuracy of computational methods in predicting the distribution of relative orientations of protein domains connected by flexible linkers.
• To evaluate predictions for a Staphylococcal protein A (SpA) construct (ZLBT-C) with wild-type (WT) and all-glycine (Gly6) linkers.
• To compare computational predictions against experimental Nuclear Magnetic Resonance (NMR) residual dipolar coupling (RDC) and Small Angle X-ray Scattering (SAXS) data.

Main Methods:

• Twenty-five groups submitted predicted conformational distributions (ensembles of structures) for the ZLBT-C construct.
• Prediction accuracy was evaluated by back-calculating NMR RDCs and SAXS curves from predicted ensembles and comparing them to experimental data.
• Kernelization was used to compare predicted ensembles with continuous orientational distributions derived from experimental data.

Main Results:

• Predicted conformational distributions varied in accuracy but none closely matched the combined NMR and SAXS experimental data.
• No computational methods successfully reproduced the distinct conformational differences observed between the WT and Gly6 linker constructs in SAXS data.
• Analysis revealed both strengths and weaknesses in the prediction methods, emphasizing the complementary nature of NMR RDC and SAXS data.

Conclusions:

• Current computational methods are insufficient for accurately predicting protein domain orientation distributions and conformational ensembles.
• The specific sequence and flexibility of linkers significantly impact protein conformation, a factor not adequately captured by current prediction models.
• Further development of computational approaches is needed, integrating insights from biophysical techniques like NMR and SAXS for improved accuracy.