Some proteins are pleased to remain in the cytosol, where they are synthesized. But for many, their fate lies elsewhere, and whether they seek to be inside an organelle, integrated into a membrane, or released from their natal cell altogether, these wayfaring polypeptides all soon face a common barrier: a lipid bilayer. Getting proteins across membranes is one of the most fundamental challenges of cellular life, and the solution is fittingly similar in all kingdoms. Yet to human onlookers, this process remains, largely, a mystery.
Proteins to be secreted from a cell must cross the endoplasmic reticulum membrane (in eukaryotes) or the plasma membrane (in prokaryotes). In either case, their way is paved by a heterotrimeric membrane channel—the Sec61 complex in eukaryotes and the SecY complex in bacteria and archaea—and driven by a variety of different partners. The ribosome is one; it can dock and send nascent peptides directly into the channel as they are being synthesized. Proteins can also be secreted post-translationally—in eukaryotes, with the help of the chaperone BiP or, in bacteria, with the ATP-dependent motor SecA.
A structure of the SecY complex and a handful of SecA structures have been solved in several labs, including Tom Rapoport’s. Now, Rapoport and his team have solved a structure of the two together; it appears in the Oct. 16 Nature. “This is the first time that we can actually tell how SecA might work,” said Rapoport, a Howard Hughes investigator and HMS professor of cell biology. In a second study published in the same issue, the lab used biochemical techniques to test and extend their predictions.
SecA–SecY seems to share fundamental mechanistic features with protein-degrading and unfolding enzymes. So beyond describing central events in protein secretion, this work may inform studies on a wide range of protein-handling processes.
After four years of trials, Jochen Zimmer, a postdoctoral fellow and the lead author of the structure paper, arrived at a 4.5 ångström structure of a SecA–SecY complex from the bacterium Thermotoga maritima. A lower-resolution structure of components from other bacterial species was also solved by postdoctoral fellow Yunsun Nam. SecA appears to be preparing SecY to receive a polypeptide: transmembrane helices are spread apart, and a short helical plug is partially displaced. In a separate study reported in the same issue, a group led by researchers in Japan solved the structure of a SecY complex held in another “pre-open” state by an antibody that binds at a site similar to SecA.
Based on the structure, Rapoport and colleagues propose that a clamp in SecA holds the polypeptide substrate and then opens upon ATP binding. The release allows a two-helix finger in SecA to push the polypeptide through a pore in the SecY channel. Upon ATP hydrolysis, the clamp would retighten and the finger would move back to get ready for another stroke.
Importantly, the channel takes proteins not only across the membrane but, for transmembrane proteins, sideways, into the membrane. The structure provides ready possibilities for how this might happen. The center of the SecA clamp sits directly above the pore in SecY, and further, the clamp’s opening lines up with a gap in the channel, generating a long, continuous seam. This seam may open briefly to give hydrophobic segments their exit strategy into the lipid bilayer.
The lab initially thought that the seam might open and close constantly, allowing segments to move into the lipid environment if they were hydrophobic enough. There was a problem, however. If the seam opened completely, lipids would be able to flood into the channel. Yet polypeptides need a way to sample the lipid environment to see if that’s where they belong. The group then found that the structure was suggesting a solution: the seam is open, but not along its full length.
What we see is that we just have a window pretty much at the center of the lipid bilayer, and this window is closed toward the cytoplasmic- and periplasmic-exposed ends. So you could speculate that this would basically allow the alkyl chains of the phospholipids to move inwards, but would prevent the head groups from moving into the channel,” Zimmer explained. The lipids, then, could give polypeptides the “taste” they need just by dangling their tails into the channel.
Marveling at this mechanism, Jon Beckwith, American Cancer Society professor of microbiology and molecular genetics at HMS, said, “I think the great thing is that it just seems like as [Rapoport] gets deeper and deeper, it isn’t like, OK, now we know everything…. All these new questions get raised.” Beckwith and his lab identified the sec genes in E. coli in the 1980s and have studied their functions, including collaborating recently with the Rapoport lab, although Beckwith was not involved in the present work.
As part of the second study, Stephanie Miller, a visiting student from Germany, found that a tyrosine in the finger’s tip is critical for translocating polypeptides in vitro. Evolution seems to agree, since that tyrosine is conserved in 80 percent of SecAs. Intriguingly, all but one of the 18 SecAs that instead have a small hydrophilic residue appear to be part of “accessory” SecA–SecY pairs that specialize in exporting specific substrates—in particular, heavily glycosylated proteins in pathogenic gram-positive bacteria. The researchers are left asking how these peculiarities enable the complexes to export their substrates.
By doubly cross-linking a stalled polypeptide to the SecA finger and the pore of SecY, Karl Erlandson, a graduate student and the lead author of the second paper, found direct evidence that the fingertip does contact the polypeptide close to the pore. His experiments indicate that the polypeptide makes a straight path from the finger to the pore.
Rapoport was quick to point out, “This is only a first step. The real issue would be to delineate the entire path of the polypeptide chain…. And then, of course, the whole movement that we’re implying with the model is speculative…. We don’t know whether the clamp actually opens and closes; we don’t know whether the finger moves up and down.”
Like SecA, AAA+ ATPases such as the 19S proteasome, p97, and Clp proteins thread polypeptides through a narrow pore using loops tipped by bulky hydrophobics, especially tyrosine. Rather than delivering their substrates across a membrane, these enzymes unfold proteins for degradation or refolding. The burden they all share is how to handle the full lengths of a vast array of substrates, noted Tania Baker, a Howard Hughes investigator and professor at MIT who, largely in collaboration with fellow MIT professor Bob Sauer, has done extensive work on the operation of ClpX. All of these enzymes solve the problem of “how to build a generic nanomachine that can translocate anything,” she said.
For the SecA–SecY machine, the new work “delimits very significantly how the whole thing works,” said Art Horwich, a professor at Yale who has studied ClpA. “The notion of being able to do this structural work with SecY is really impressive—membrane proteins terrify me.”
The universal hope of researchers who study this far-flung collection of cellular machines is to get a glimpse of a polypeptide in the act of being translocated. “The big next step would be to get a structure with the polypeptide chain in the channel,” Rapoport said. “How to do that, we’re not really sure.
For Students: Contact Tom Rapoport at tom_rapoport@hms.harvard.edu for more information on this and other lab projects.
Conflict Disclosure: The researchers report no conflicts.
Funding Sources: The National Institutes of Health, Howard Hughes Medical Institute, Damon Runyon Cancer Research Foundation
