Search for: All records

The activation of transcription factor Max-Like Protein x (MLX) is modulated by competition between active dimerization and inactive association with cytosolic lipid droplets (LDs). However, LD association has been shown to depend on the neutral lipid composition. This work explores the mechanism by which MLX specifically targets LDs rich in triacylglycerol (TG) over those with abundant sterol esters (SE). We compare the association ensembles for a potential minimal targeting sequence, an amphipathic helix-loop-helix hairpin, and the full dimerization and cytoplasmic localization domain (DCD), finding the latter requires larger packing defects and quantifiably alters LD membrane properties. Surprisingly, direct interactions with TG neutral lipids are not observed for either sequence. Instead, targeting to SE-rich LDs is blocked for both sequences by insufficient packing defects. We additionally explore the full mechanism of hairpin association, aiming to understand sequence-specific features that enable strong membrane association. We find that there are multiple association pathways, but that each involves a catch, dive, snorkeling, and embedding phase. The combination of multiple catch and dive residues placed on opposing ends of amphipathic helices lengthens the catch phase, greatly enhancing association in a manner that resembles kinetic selection. Once bound, locking interhelical interactions block dissociation. Collectively, our findings suggest that in addition to relative binding affinities, both kinetics and altered surface properties due to protein association could influence competition within the LD proteome. SignificanceThe transcription factor Max-Like Protein x (MLX plays) a central role in metabolic regulation by responding to nutrient status and, simultaneously, neutral lipid composition. This work reveals how MLX selectively targets triacylglycerol-rich lipid droplets (LDs) through sequence-specific interactions with packing defects. We show that LD surface modulates MLX binding and that MLX in turn alters monolayer properties, highlighting a dynamic interplay between protein association and membrane properties. These findings provide new insight into how protein localization and function may be regulated at LD surfaces, with implications for nutrient sensing and, more broadly, transcriptional control relevant to metabolic and disease states. 

Triacylglycerols (TG) are the primary neutral lipids in lipid droplets (LDs), organelles responsible for lipid storage, metabolism, and signaling. Molecular dynamics (MD) simulations have provided valuable insight into LD structure, but fixed-charge force fields struggle to capture TG behavior across both hydrophobic cores and polar interfaces. Here, we develop and evaluate a polarizable TG model using the Drude2023 lipid force field and benchmark its performance against experimental measurements of bulk density, TG−water interfacial tension, core hydration, and monolayer expansion. The Drude model accurately reproduces the experimental properties and captures key monolayer features such as surface-oriented TGs (SURF-TGs) and chemically distinct membrane packing defects. Compared to fixed-charge models such as C36-standard and C36-cutoff, the Drude polarizable model is the only force field able to capture the dual nature of TG at polar−nonpolar interfaces like the LD monolayer and more homogeneous hydrophobic environments, like the LD core. However, C36-standard is consistent with the Drude results for the LD monolayer, while C36-cutoff is consistent with the decreased hydration in the LD core. Even with large applied surface tensions, C36-cutoff does not produce Drude-like LD monolayer properties. These results highlight the importance of dynamic polarizability and establish Drude2023 as a more reliable framework for simulating TG in heterogeneous systems like LDs.