Modeling the Biophysics of Cellular Membranes

Cholesterol is the major unique lipid in the outer plasma membrane of cells, where it is necessary for many basic cellular functions. A major question in lipid biology and biophysics is why we have evolved to have specific sterol species, e.g., cholesterol and ergosterol. In yeast, the vacuole membrane phase-separates, forming domains with both distinct physical properties as well as affinities for particular membrane-associated proteins. In a collaborative study with Itay Budin's lab, enzymes of the ergosterol biosynthetic pathway were systematically knocked out, yielding distinct morphology and fluidity of the vacuole membrane. Ergosterol was not the strongest ordering lipid, perhaps contradicting intuition that the evolutionarily targeted sterol would be a maximum of some simple property. Instead, ergosterol led to both phase separation as well as to an ordered domain that was cleanly fluid, suggesting that ergosterol is optimal for the functional properties of the vacuole beyond phase separation. Our molecular simulations show the structural basis for this fluidity, with a distinct sub-population of ordered lipids at the nanometer scale induced by the sterol. To be fluid at the scale of the domain in the vacuole, the ordered lipid complexes must be of a finite scale, that is, the sterol must both induce order and limit it to small scales, similarly to that observed by cholesterol.

Myomaker and myomerger are proteins in muscle cells necessary for fusion (a critical process in building multi-nucleated muscle cells), but whose molecular mechanisms for inducing fusion are unknown. Inducing myomaker and myomerger expression in normally fusion-incompetent cells induces fusion, suggesting that these proteins have a direct role. Yet unlike typical fusion machinery, myomaker does not extend significantly above the plasma membrane, suggesting it does not have a receptor-targeting domain or a stalk that would reconfigure to mechanically influence cell-cell fusion.

Data instead suggest a lipid-centered effect. Myomaker is homologous to a ceramidase, yet in our work we did not detect an ability to regulate ceramides. Instead, myomaker was associated with a number of lipid-related observations. First, treatment by a ceramide synthase inhibitor increased both fusion and the expression of ether lipids. Myomaker and ether lipids were correlated in pseudoviral budding experiments, suggesting that the two lipids are tightly spatially correlated on the plasma membrane. Increasing ether lipid content also increased plasma-membrane expression of myomaker and fusion, also correlating with the exterior presentation of phosphatidylethanolamine and phosphatidylserine lipids, lipids that are normally present on the inner plasma membrane leaflet. While molecular simulations found no clear correlation of myomaker with ether lipids, we detected a substantial membrane-thinning effect near residues, with functional implications inferred from mutational analysis. This membrane deformation is suggestive of a mechanism in which myomaker is able to scramble nearby lipids, exposing fusogenic phosphatidylethanolamine lipids to the outer leaflet.

Bacteria must control the amount of intracellular magnesium, an important cofactor for numerous enzymes and nucleic acid structures. In this collaboration including two other intramural labs, our simulations support the Matthies' lab's structure of MgtA, a magnesium-transporting P-Type ATPase. Unexpectedly, the transporter appears as a stable dimer, with cryo-EM providing a detailed molecular interface between the two units with multiple hydrophobic residues and salt bridges. Molecular simulations show the dynamics and stability of this interface. Additionally, simulations indicate the detailed coordination and hydration of a critical magnesium ion binding site, and predict the binding stability of magnesium compared with sodium. Simulations predict the hydration of the transporter's interior beyond structural studies, which are typically unable to resolve dynamic structures.