Probing the Structure of Proline-Containing Lipopeptides Aggregates: A Simulation-Driven Approach to SAXS Interpretation

Lipopeptides are amphiphilic molecules composed of short peptides attached to lipid chains. They can self-assemble into peptide-functionalized supramolecular structures with various morphologies. Their structure enables a variety of applications, including catalysis, membrane-disruptive antibiotics, and molecular sensors. However, predicting the final shape of the aggregates as a function of the experimental conditions is extremely challenging. According to experimental results, small changes in concentration or pH can lead to markedly different assembly structures. In this work, we modeled supramolecular aggregates of proline-containing lipopeptides of the proline-arginine-tryptophan-glycine-C18 sequence using a hybrid particle field (hPF) approach. A zero-order reference Hamiltonian was constructed from the radial distribution function obtained from all-atom simulations. Meta-inference molecular simulations combined with SAXS data were used to refine the description of the aggregates, allowing for an improved representation of the structural ensemble present in solution. The addition of the meta-inference bias allows for a much-improved description of the SAXS signal from simulation, revealing a diverse ensemble of structures. The results at different concentrations revealed the presence of spherical and prolate core-shell structures. Furthermore, cluster analysis revealed a complex polymorphic behavior that is concentration dependent and cannot be adequately represented by a typical average structure.