Structural Biochemistry/Membrane Proteins/Topology

Introduction
It has been well established that membrane proteins must be inserted in a specific topology, an orientation in the lipid bilayer, in order to properly function. However, recent evidence from experiments done by Shimon Schuldiner has challenged this conventional thought through studying EmrE, a small homodimeric multidrug transporter from Escherichia coli that transports positively charged aromatic drugs out and takes in two protons.

It should be noted that dimeric membrane proteins can exist in a parallel and/or antiparallel orientation. In the specific case of EmrE, the parallel orientation can be visualized as having the N-terminus and C-terminus on the same side; the antiparallel orientation can be visualized as having the N-terminus on one side and the C-terminus on the other.

While it is well known that proteins such as receptors must be oriented in one way and one way only, this may not necessarily be the case for EmrE. Past experiments on EmrE have indicated that even though EmrE can be found in a parallel topology, only the antiparallel topology can induce normal function. However, each of these experiments may have certain conditions that deserve a counterargument.

Antiparallel Topology: Experiments and Counterarguments
In one experiment, scientists created a 3D artificial model of an antiparallel topology of EmrE. They then used an actual antiparallel 3D model of a native EmrE and compared the two, concluding that both were similar and thus proved that EmrE fits the antiparallel topology. While the evidence is convincing, the fact that the resolution of the structures was very low and so similarities were much easier to “identify.” Furthermore, the crystals used for the antiparallel model came from proteins in detergents that inhibit the protein function. This implies that the way the 3D model was created was by the least energy required for a crystal formation, not the way EmrE naturally exists in.

In another experiment, EmrE was fused to a green fluorescent protein that did not affect the function of EmrE. However, upon manipulation of the positive charges in the protein, parallel mutant EmrE proteins were created. These mutations were shown to be resistant to the green fluorescent protein, leading to the conclusion that this was because of the changed topology. To further support this, the mutant EmrE proteins were coexpressed with the native EmrE and normal function was restored within one generation. A problem with this experiment was that the researchers assumed that the parallel mutants were inactive. In fact, further experiments done by Schuldiner showed that parallel mutant EmrE proteins had continuous growth. In addition, the inactivity of the mutant parallel proteins can be explained by impaired dimerization—if the protein cannot be formed properly due to the mutation, its function will not be normal. This idea is independent of topology.

Parallel Topology: Experiments
Two experiments were done to show the ability of EmrE to function properly even with a parallel topology. In one experiment, unique cysteines were created in the termini of the protein. Crosslinkers, or molecules that can form covalent bonds with other similar molecules, were formed within 9 to 11 angstroms apart. Previous studies had established that in antiparallel topology, cysteine residues had to be at least 35 to 40 angstroms apart. To solidify this evidence, repeated experiments on proteins crosslinked in other loops with similar results showed that they were fully functional EmrE proteins.

In another experiment, the C-terminus of one protomer was artificially fused to the N-terminus of another by a very short, hydrophilic linker, ensuring that the terminus of both protomers were on the same side of the membrane (therefore parallel). These artificially created proteins were also shown to be fully functional.

Another study is that parallel homodimers like EmrE and antiparallel heterodimers like EbrAB from Bacillus subtilis both perform identical functions regardless of their topology. In fact, EbrB was observed to be able to form parallel homodimers and function in the same way. This has very strong implications for EmrE. Because EbrAB and EmrE are closely related proteins, EmrE may also have a similar “promiscuity” in that it can be functional in both the parallel and antiparallel topology.

Conclusion
It has been shown that EmrE exists in a parallel topology and can function normally. Furthermore, it can exist in an antiparallel topology and function just as well. A conceivable mechanism into how this can work is that the active sites are conserved in both topologies. This active site is occupied by either an H+ or a substrate, depending on the two glutamate residues that vary the pKa. This in turn, is strong evidence that the topology is determined by the direction of the driving force, mainly the H+ or the substrate.

Although one should note that these evidence came from one protein, the significance of these findings is clear. It should be noted that EmrE is a small membrane protein capable of evolving into much more complex membrane proteins. If a protein can exist in different forms directed by their function, the conventional idea of structure determining function may be challenged. Instead, function may determine the structure.