Better Together: Lipopeptide Micelle Formation Enhances Antimicrobial Selectivity

Main Text

Antimicrobial resistance stands as a growing threat to public health (1). Antibiotics, designed to specifically target bacterial cells, have inadvertently accelerated the development of resistant microorganisms, making any infection potentially dangerous. Antimicrobial lipopeptides (AMLPs—part lipid, part peptide) have recently gained increasing interest (2), because they combine two attractive features: just like antimicrobial peptides, their mode of action is very generic and thus evolutionarily conserved; but unlike antimicrobial peptides, they involve much smaller molecular-weight peptides and are thus more in line with both practical and economical considerations of drug development.

In a previous coarse-grained simulation study, Lin and Grossfield (3) had investigated the mechanisms of action of single AMLPs on both types of model membranes: anionic (bacterial-like) and zwitterionic (mammal-like). The results showed that the lipid-tail-like chain dominates the binding while the peptide part provides selectivity to anionic membranes. The simulations indicated significant binding to both types of membranes, with a slight preference for the anionic composition. Given that host mammalian membranes significantly outnumber bacterial membranes, how can a small difference in free energy ensure selectivity toward the latter?

In a surprising but significant twist presented in this issue of the Biophysical Journal, Lin and Grossfield (4) extended their analysis to finite concentrations of AMLPs. The binding of a 48-mer AMLP micelle yields somewhat similar conclusions as to the thermodynamics of the monomer systems: favorable insertion in both membranes and preference for anionic membranes. Scaling the results per lipopeptide, the authors observe weaker binding to both membranes, due to the added stability of the micelle relative to the monomer in solution. On the other hand, a notable difference arises when considering not only the overall free-energy difference of the binding process, but also the paths to get there: interactions of the micelle with the model mammalian membrane exhibited significant free-energy barriers (79 kcal/mol), while, in comparison, the anionic mixture yielded almost-downhill behavior (barrier of 1 kcal/mol).

The authors discuss the significance of these findings: the consideration of the biologically relevant micelle in solution, rather than the monomer alone, is key to rationalize the selectivity of AMLPs toward bacterial membranes. The lead optimization of such constructs will thus need to rely on a proper control of the oligomerization in solution. More specifically, their analysis leads to a number of rational design rules: enhancing the kinetic-barrier difference and potentially hedging selectivity issues for small micelles, which display much smaller free-energy barriers.

As in any simulation-based oligomerization study, finite-size aggregation effects creep in, although the authors go to some length to understand the errors associated with them. The many biophysical roles oligomerization can take emphasize the ever-increasing need for the simulation community to better address finite-size effects in a robust and systematic way.

In the end, the work of Lin and Grossfield is an exemplary application of coarse-grained simulations and enhanced-sampling methodologies done carefully. It provides detailed molecular and thermodynamic insight into a biomolecular process that is way beyond contemporary capabilities of atomistic simulations, for which exhaustive phase-space sampling still requires enormous efforts even for ostensibly small systems (5). One striking aspect of this work is the combination of several advanced methodologies, including the string method (6) and the numerical optimization of a likelihood estimator (7) for the weighted histogram analysis method (8), allowing convergence for an impressive dynamical free-energy range. Efficiently sampling the important degrees of freedom remains of utmost importance as the biophysics community is gradually closing the gap between experiments and simulations.

Editor: Markus Deserno.