| Continued from page 1 | ||
|
The first x-ray crystal structure of the membrane protein MsbA, a bacterial ABC transporter superfamily member, will no doubt be viewed as a milestone in cystic fibrosis research, even though some important details of this lipid pump will not apply directly to CFTR. Importantly, the MsbA crystal structure constitutes the first tangible proof that the transmembrane helices of ABC proteins are indeed alpha-helices, as was originally predicted in 1989 for CFTR. And these lipid-embedded helices are all tilted with respect to the plane of the membrane by between 30 and 40 degrees, forming a "teepee" like structure encompassing a chamber in the center of the membrane having two openings into the inner bilayer (probably for lipid diffusion into the pore). And from the MsbA structure, it is also evident that the NBD domains do not constitute part of the pore pathway. Another interesting finding is that the intracellular loops are distinct alpha-helical intracellular domains, one of which is sandwiched between the alpha-helices in the membrane and the NBD domain, serving as a bridge, or "conduit", for the transfer of energy from the NBD domains to the transmembrane domains. Reassuringly, what is resolvable of the NBD domains agrees well, structurally, with the 4 previously known crystal structures of NBD domains (HisP, MalK, Rad50, MJ1267). The chamber of MsbA is closed to the extracellular side, implying that there is no unobstructed pathway for substrate during transport. And there are two side entrances within the membrane's inner leaflet, which probably allow lipid and substrate to diffuse into the lower part of the chamber. This lower half of the chamber contains an area of distinct positive charge on the interior, while the upper extracellular side is more hydrophopic (on the interior). The authors do sound some cautionary alarms, however, for those interested in applying the new MsbA structure to proteins like CFTR. For example, no lipid co-crystallized along with MsbA, including lipid substrate (i.e. Lipid A) or lipid from the membrane As MsbA is known to transport a hydrophobic substance, lack of lipid in the structure makes it difficult to infer mechanisms for this transport. In addition, the resolution of MsbA is given as 4.5 Å, making it difficult to distinguish individual amino acid side chains, even though this resolution allows the tracing of the protein's polypeptide backbone as well as observing secondary structures including alpha-helices and beta-sheets. Also, CFTR being a chloride channel, has an aqueous, rather than a lipid-filled pore. No doubt for this reason, the authors point out that MsbA will probably end up being a better model for the multi-drug resistance protein (also called P-glycoprotein, 36% identical to MsbA in sequence), than it will for CFTR. Still another drawback was the necessity of inferring exactly where the cell membrane should be in the MsbA crystal structure in relation to the MsbA protein itself. And perhaps significantly, MsbA has never been shown to be active in purified form. Another piece lacking from the MsbA structure for the time being is ATP or an ATP analog. And lastly, a portion of the NBD domain was found to be largely unstructured. It was the acid-terminal part containing the Walker A motif. But in spite of these difficulties, there is still valuable information to be obtained from the MsbA structure. |
We are now provided with a possible mechanism for Lipid A transport thru MsbA, which may involve the tilting (or perhaps twisting) of the transmembrane helices during ATP binding and hydrolysis at the NBD domains. This predicted conformational change in the transmembrane helices may be the way in which substrate is "pumped" thru the membrane. What implications do the MsbA structure hold out for providing a better understanding of CFTR? At the very least, MsbA will provide a new "molecular scaffold" for future structural studies of CFTR to be tested on. And the MsbA structure serves to graphically remind one of the dynamics ABC family members are capable of, however, the R-domain of CFTR, which does not exist in MsbA, will have to be modeled in at some point in the future. Chang and Roth write: "...MsbA may not be able to serve a close structural representative of the ABC transporters that transport hydrophilic substances (i.e. CFTR)...", and "..the crystal structure reveals MsbA is not a pore thru the cell membrane, but is a molecular machine scanning the lower bilayer leaflet for substrates (i.e. Lipid A), accepting them laterally, and flipping them to the outer membrane leaflet". Geoffrey Chang and Christopher Roth Science Vol 293 9/7/01 pgs 1793-1800
|
|
