Back

Polymer-extracted structure of the mechanosensitive channel MscS reveals the role of protein-lipid interactions in the gating cycle

Moller, E.; Britt, M.; Zhou, F.; Yang, H.; Anishkin, A.; Ernst, R. K.; Sukharev, S. I.; Matthies, D.

2024-02-05 biophysics
10.1101/2024.01.22.576751 bioRxiv
Show abstract

Interpretations of experimental conformations of mechanosensitive channels gated by force from lipids become more reliable when native lipids are preserved in the structures. Escherichia coli MscS is an adaptive osmolyte release valve that regulates turgor in osmotically stressed cells. MscS promptly opens under abrupt super-threshold tensions in the membrane, but at lower and more gradually applied tensions, it silently inactivates from the closed state. A central question has been whether to assign the commonly observed non-conductive conformation with splayed peripheral helices to a closed or inactivated state. We present a 3-[A] MscS cryo-EM structure obtained in Glyco-DIBMA polymers, which avoid complete lipid removal. Within the complex, we observe densities for endogenous phospholipids intercalating between the peripheral and pore-lining helices. The lipidomic analysis shows a 2-3 fold enrichment of phosphatidylglycerol in Glyco-DIBMA-extracted MscS samples. The computed pressure of these lipids on the protein surface enforces the characteristic kinks in the pore-lining helices, sterically stabilizing the separation of the peripheral helices. Mutations of residues coordinating lipids in the crevices eliminate inactivation, allowing us to classify this group of structures as the inactivated state. Our study reveals a novel inactivation mechanism in which intercalated lipids physically decouple the tension-sensing helices from the gate.

Matching journals

The top 2 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.