The bacterial outer membrane serves as a crucial barrier to protect double-membraned bacteria from harmful environmental insults, such as antibiotics, bile acids in the gut, and detergents in your hand soap. In E. coli, an outer leaflet of lipopolysaccharides (LPS) and an inner layer of phospholipids creates a lipid bilayer largely impermeable to both hydrophobic and hydrophilic molecules. Decades of work has lead to a fairly complete model for how LPS is trafficked to the outer membrane, including structures for many components of the LPS transport system. However, we know comparably very little about how newly synthesized phospholipids in the inner membrane are moved to the outer membrane, how asymmetry across the outer membrane is maintained (i.e., LPS outside and phsopholipids inside), and how outer membrane lipids might be imported towards the inner membranes (e.g., for use as a carbon source).
Sometimes, a research project leads you in a direction you didn't initially expect. Our work on outer membrane lipid transport actually began with our interest in understanding host-pathogen interactions in M. tuberculosis (Mtb), an intracellular bacterium that infects approximately one third of the world's population and is the causative agent of tuberculosis (TB). We were intrigued by a mysterious family of proteins in Mtb called MCE, which were originally thought to play in a role in Mammalian Cell Entry, but subsequently have been implicated in the transport of hydrophobic molecules in or out of the cell, including lipids, cholesterol, and organic solvents. Most importantly, MCE proteins are important for virulence in mouse models of TB infection, and have been linked to virulence in other double-membraned bacterial pathogens.
Initially we focused our efforts on Mtb, where MCE proteins have been most extensively studied. But we quickly realized that MCE proteins are ubiquitous among double-membraned bacteria, including the model bacterium E. coli. MCE genes are even conserved in some double-membraned organelles, such as eukaryotic chloroplasts, where they have been implicated in phospholipid import from the ER. In contrast, MCE genes are entirely absent from single-membraned bacteria, suggesting that they play a fundamental role in the double-membraned lifestyle. We believe that understanding the function of MCE genes will have broad implications in many areas of biology: from understanding host-pathogen interaction in Mtb; to fundamental aspects of the physiology of countless bacterial species; to plant biology, organelle maintenance, and domestication.
Lipids are hydrophobic and poorly soluble in water. Yet, the inner and outer membranes are separated by a ~20 nm, water-filled gap called the periplasm, presenting a major obstacle for efficient lipid trafficking. Structures of components of the LPS transporter, particularly LptA and LptD revealed that these proteins can assemble to form a periplasm-spanning bridge with a long hydrophobic groove. These structures have led to a model wherein the lipid tails of LPS insert into the groove and are protected from solvent, and a series of LPS molecules can be pushed along the bridge from the inner membrane to the outer membrane. In contrast, we know much less about how phospholipids cross the periplasm. Enter MCE proteins!
Our crystal structure of MlaD revealed for the first time that MCE domains form hexameric rings surrounding a central, hydrophobic pore. We proposed that this pore serves as part of the phospholipid transport pathway between the inner and outer membranes, though this remains to be demonstrated experimentally. However, MlaD is not large enough to facilitate lipid transport all the way across the periplasm, but likely works together with several other proteins, including an inner membrane ABC transporter complex and an outer membrane complex between MlaA and a porin protein. While the direction of lipid transport remains an open question and needs to be tested directly, one can imagine the system working as an exporter roughly as outlined below. Alternatively, the step may be reversed to drive phospholipid import.
MlaD is one of three MCE proteins in E. coli. Our recent cryo electron microscopy (cryo EM) studies of the other two E. coli MCE proteins, PqiB and YebT, uncovered two additional, distinct architectures. In contrast to the single ring formed by MlaD, YebT stacks together 7 MCE rings to form a long, hollow tube. PqiB combines a stack of 3 MCE rings with a very long 6-helix coiled-coil, resulting in a hollow needle and syringe, reminiscent of several classes of bacterial protein secretion machines. Unlike the relatively small protein, MlaD, both PqiB and YebT are over 200 Å in length, very similar to width of the periplasm that separates the inner and outer membranes. Thus, we hypothesize that PqiB and YebT create a continuous lipid translocation pathway between the inner and outer membranes, in a manner analogous to the current model for LPS transport. But in a remarkable example on convergent evolution, the proteins mediating periplasmic transport of LPS and phospholipids are completely unrelated.
While we have learned a great deal about MCEs and their associated proteins, there is still much that we don't understand. Are lipids really transported through the central channels of MlaD, PqiB, and YebT? How are lipids extracted from the inner membrane and inserted into the outer membrane? Or is transport occurring in the opposite direction? Why does E. coli need three MCE proteins? Why does Mtb need 24, and why does M. smegmatis need 36? How do bacterial MCEs differ from their homologs in plants? These are just a handful of the sorts of questions we are actively pursuing in the lab.