Rapid Formation of Peptide/Lipid Coaggregates by the Amyloidogenic Seminal Peptide PAP248-286.

Protein/lipid coassembly is an understudied phenomenon that is important to the function of antimicrobial peptides as well as the pathological effects of amyloid. Here, we study the coassembly process of PAP248-286, a seminal peptide that displays both amyloid-forming and antimicrobial activity. PAP248-286 is a peptide fragment of prostatic acid phosphatase and has been reported to form amyloid fibrils, known as semen-derived enhancer of viral infection (SEVI), that enhance the viral infectivity of human immunodeficiency virus. We find that in addition to forming amyloid, PAP248-286 much more readily assembles with lipid vesicles into peptide/lipid coaggregates that resemble amyloid fibrils in some important ways but are a distinct species. The formation of these PAP248-286/lipid coaggregates, which we term "messicles," is controlled by the peptide:lipid (P:L) ratio and by the lipid composition. The optimal P:L ratio is around 1:10, and at least 70% anionic lipid is required for coaggregate formation. Once formed, messicles are not disrupted by subsequent changes in P:L ratio. We propose that messicles form through a polyvalent assembly mechanism, in which a critical surface density of PAP248-286 on liposomes enables peptide-mediated particle bridging into larger species. Even at ∼50-fold lower PAP248-286 concentrations, messicles form at least 10-fold faster than amyloid fibrils. It is therefore possible that some or all of the biological activities assigned to SEVI, the amyloid form of PAP248-286, could instead be attributed to a PAP248-286/lipid coaggregate. More broadly speaking, this work could provide a potential framework for the discovery and characterization of nonamyloid peptide/lipid coaggregates by other amyloid-forming proteins and antimicrobial peptides.

Tuning Length Scales of Small Domains in Cell-Derived Membranes and Synthetic Model Membranes.

Micron-scale, coexisting liquid-ordered (Lo) and liquid-disordered (Ld) phases are straightforward to observe in giant unilamellar vesicles (GUVs) composed of ternary lipid mixtures. Experimentally, uniform membranes undergo demixing when temperature is decreased: domains subsequently nucleate, diffuse, collide, and coalesce until only one domain of each phase remains. The sizes of these two domains are limited only by the size of the system. Under different conditions, vesicles exhibit smaller-scale domains of fixed sizes, leading to the question of what sets the length scale. In membranes with excess area, small domains are expected when coarsening is hindered or when a microemulsion or modulated phase arises. Here, we test predictions of how the size, morphology, and fluorescence levels of small domains vary with the membrane's temperature, tension, and composition. Using GUVs and cell-derived giant plasma membrane vesicles, we find that 1) the characteristic size of domains decreases when temperature is increased or membrane tension is decreased, 2) stripes are favored over circular domains for lipid compositions with low energy per unit interface, 3) fluorescence levels are consistent with domain registration across both monolayer leaflets of the bilayer, and 4) small domains form in GUVs composed of lipids both with and without ester-linked lipids. Our experimental results are consistent with several elements of current theories for microemulsions and modulated phases and inconsistent with others, suggesting a motivation to modify or enhance current theories.

Identifying the Onset of Phase Separation in Quaternary Lipid Bilayer Systems from Coarse-Grained Simulations.

Understanding the (de)mixing behavior of multicomponent lipid bilayers is an important step toward unraveling the nature of spatial composition heterogeneities in cellular membranes and their role in biological function. We use coarse-grained molecular dynamics simulations to study the composition phase diagram of a quaternary mixture of phospholipids and cholesterol. This mixture is known to exhibit both uniform and coexisting phases. We compare and combine different statistical measures of membrane structure to identify the onset of phase coexistence in composition space. An important element in our approach is the dependence of composition heterogeneities on the size of the system. While homogeneous phases can be structured and display long correlation lengths, the hallmark behavior of phase coexistence is the scaling of the apparent correlation length with system size. Because the latter cannot be easily varied in simulations, our method instead uses information obtained from observation windows of different sizes to accurately distinguish phase coexistence from structured homogeneous phases. This approach is built on very general physical principles, and will be beneficial to future studies of the phase behavior of multicomponent lipid bilayers.