Intermolecular structure formation in cell membranes is central to many biological processes, including signaling and viral infection. In vitro models of the cell membrane are unable to reproduce this structure formation on physiologically relevant spatial and temporal scales. We are building systems for fabricating lipid bilayers that properly mimic both the cytoskeletal attachment and leaflet asymmetry of the plasma membrane. As well as serving as tools for studying lipid biophysics, these systems are an important step towards completely synthetic cells.

In the past decade, there has been a significant research effort towards understanding the role that lipid rafts—phase-segregated domains of lipids present in the plasma membrane—play in a variety of cellular processes. Various lines of evidence from cell biology-based assays suggested that rafts associate with specific membrane proteins and thereby help to regulate signaling processes. Lipid rafts have also been implicated in viral docking and budding and have been suggested as potential apoptosis triggers for cancer therapies.

Much of our biophysical understanding of the nature of rafts comes from in vitro investigations of synthetic lipid bilayers. These systems have elucidated the temperature-dependant phase behavior of several physiologically relevant phospholipid-cholesterol mixtures. Briefly, many lipid bilayers containing both saturated and unsaturated phospholipids as well as cholesterol will separate, below a miscibility transition temperature, into a liquid ordered (l_o) and a liquid disordered (l_d) phase. Both of these phases are fluid, with no long-range crystalline packing of the lipids themselves. The critical difference is that the tail groups of the lipids in the l_o phase, which is concentrated in saturated lipids, sphingolipids, and cholesterol, are in extended configurations any unsaturated lipids in an all-trans state.

There has been, however, considerable difficulty in reconciling in vitro studies of synthetic model bilayers with observations of actual cells. The core of this difficulty lies in the scale of phase separation. In synthetic model systems, separated phases will condense and grow to diameters of 5-10 µm or larger. No such large phase-separated domains have been seen in cells, however. Rather, the best evidence shows that any phase separated domains in cells are present on the nanoscale: likely 200 nm or smaller. Understanding the source of this scale discrepancy is key to understanding fully the biological role of lipids rafts.

Several studies have suggested that the raft growth is limited in cells because of mechanical connectivity between proteins in the plasma membrane and the cytoskeleton. These connections limit the long-range diffusivity of lipids and prevent the growth and condensation of l_o domains beyond the nanoscale. This connectivity is an aspect of the plasma membrane that previous synthetic model systems have failed to capture; building a truly biomimetic cell membrane requires mimicking cytoskeleton-membrane attachment. Most model lipid membranes also differ from the physiological plasma membrane in that the two leaflets of the bilayer are compositionally identical. In real cells, the bilayer is asymmetric, with the outer leaflet concentrated in sphingolipids while the inner leaflet is concentrated in charged lipids such as phosphatidylserine.

We are fabricating synthetic lipid membranes that mimic both the cytoskeletal connectivity and compositional asymmetry of the plasma membrane. The cytoskeletal anchoring is based on a system we developed earlier for supporting a planar lipid bilayer with a polymer hydrogel. In this system, the bilayer contains lipids that have been modified to display vinyl headgroups. When a hydrogel precursor is polymerized around this bilayer, the vinyl groups are covalently incorporated into the growing polymer chain. We are adapting this approach to a giant unilamellar vesicle (GUV) system. GUVs have been widely used to study lipid phase separation processes. By synthesizing GUVs that contain a hydrogel precursor and polymerizing the hydrogel in situ, we can construct models membranes anchored to an underlying polymer matrix that will behave like the cytoskeleton.

Unfortunately, conventional systems for fabricating GUVs are very limited in terms of what solution environments they can successfully operate in. This restricts our ability to include the proper reaction components for a hydrogel polymerization on the interior of a GUV. We are developing a microfluidic layer-by-layer GUV synthesis system that allows for the inclusion of arbitrary components within a GUV. This system is based on a two-phase droplet flow configuration: water droplets, with arbitrary contents, are dispensed into a continuous oil phase. The oil contains lipids, which form a monolayer at the phase interface. Droplets are then dispensed into a microcentrifuge tube containing a layer of oil floating over water. The oil superphase contains dissolved lipids, which form a monolayer at the oil-water interface. When the droplets fall through this interface, they pick up a second lipid monolayer, creating a water droplet surrounded by a lipid bilayer: a GUV, with the contents determined completely by the aqueous stream in the microfluidic device.

This system has allowed us to construct GUVs containing a poly(ethylene glycol) hydrogel.

This same system allows us to mimic the asymmetry of the plasma membrane. Simply by using different lipids in the two different oil phases (microfluidic flow stream and microcentrifuge tube), we can control the contents of the two monolayers independently. Lipids with fluorescently labeled headgroups can be included in either leaflet. We have demonstrated asymmetry by showing that upon addition of a hydrophilic fluorescent quencher, GUVs labeled externally are quenched significantly more than GUVs labeled internally. We are proceeding with the construction of vesicles the mimic the phosphatidylserine asymmetry of the plasma membrane.

Once biomimetic GUVs are fabricated, they will be investigated using high resolution microscopy techniques based on Förster resonance energy transfer in total internal reflection fluorescence. These techniques will allow us to detect and track lipid raft formation on the nanoscale.

Extended Abstract Status: Not Uploaded