Martini 3 as a Transferable Force Field for Lipopolysaccharide Parametrization

Area of Science:

• Biophysics
• Computational Biology
• Microbiology

Background:

• Lipopolysaccharides (LPS) are crucial outer membrane components of Gram-negative bacteria.
• All-atom molecular dynamics (AA-MD) simulations offer high resolution but are computationally expensive for long timescales.
• Coarse-grained (CG) models like Martini 3 provide computational efficiency but require accurate LPS parametrization.

Purpose of the Study:

• To develop and validate transferable coarse-grained (CG) parameters for lipopolysaccharides (LPS).
• To enable efficient simulation of Gram-negative bacterial outer membranes (OMs) at biologically relevant scales.
• To address the limited availability of CG models for LPS in computational studies.

Main Methods:

• Parametrization of 57 unique disaccharide units using the automated Bartender tool for CG-to-AA mapping.
• Combination and manual refinement of parameters to model complete LPS molecules.
• Extensive AA and CG simulations of asymmetric bilayers mimicking bacterial OMs.

Main Results:

• Developed and validated CG parameters for LPS from *Escherichia coli*, *Salmonella enterica*, *Klebsiella pneumoniae*, and *Pseudomonas aeruginosa*.
• CG simulations closely agreed with AA simulations for key structural and dynamic properties.
• Demonstrated the accuracy and robustness of the transferable parameter set.

Conclusions:

• The developed transferable CG parameter set is a valuable tool for simulating Gram-negative bacterial OMs.
• Enables larger-scale and longer-timescale simulations compared to AA-MD.
• Facilitates deeper understanding of bacterial outer membrane structure and dynamics.