Determining Conformational Change within the Escherichia coli Outer Membrane Protein FepA

Abstract

Bacteria have been able to circumvent antibiotic treatment in several ways, e.g., reorganization of the membrane and its permeability, decrease porin content, over expression of efflux pumps, and genetic changes in target sites. Most aerobic and facultative anaerobic microorganisms synthesize at least one siderophore, which chelates iron, making iron transport systems an appealing target in determining new methods of pathogenic preventions.

Ferric enterobactin (FeEnt) is the native siderophore of Escherichia coli (E. coli). The mechanism of FeEnt transport through the outer membrane (OM) receptor protein, FepA, is still unknown. FepA is a ligand-gated porin with a globular N-domain occluding the 22-stranded β-barrel C-domain. The occlusion of the ligand-gated porin makes it necessary for conformational change within the N-domain of FepA during FeEnt transport. Two models have been proposed: the ball-and-chain model and the transient pore model; in the former the N-domain is completely expelled into the periplasmic space and the latter requires a structural rearrangement within the β-barrel of FepA.

TonB is essential for transporting all metal complexes through the OM, including FeEnt. The role of TonB is unknown and the concentration of TonB proteins and TonB-dependent receptor proteins is disproportionate. This suggests that there are two populations of TonB-dependent receptor proteins: active transporters associated with TonB and inactive transporters unassociated with TonB. This discrepancy in protein concentration between the OM transport protein and TonB may explain the slow turnover rate of FeEnt through FepA. The slow turnover rate may also be a result of an intrinsically slow transport mechanism across the OM. Existing radioisotopic assays measure the transport of FeEnt through the passage of the inner membrane (IM), as the accumulation of the iron complex in the cytoplasm. We devised a new assay to observe FeEnt transport through the OM.

The novel post-uptake binding (PUB) determinations provided information on the mechanism of FeEnt transport through FepA. It revealed that all of the FepA proteins were functionally active and could transport FeEnt through the OM. It is possible that the interaction between FepA and TonB is the rate-limiting step, which may explain the low turnover number. PUB determinations from strains lacking FepB or a complete FepCDG inner membrane complex suggested that these proteins were necessary for FeEnt transport through the OM. After determining the FepA proteins were actively transporting FeEnt in the strain lacking FepB, it was determined that FeEnt was not retained in the cell. Accumulation of the ligand in the cytoplasm was impaired in strains lacking the periplasmic binding protein or a complete inner membrane FepCDG complex. Retention experiments and PUB determinations revealed TolC exported FeEnt out of the cell.

Kinetic data suggested that ligand uptake through FepA is triphasic with the initial rate being the most rapid, the second rate has an intermediate rate, and the last rate is the slowest. These results have not been previously observed and it may reflect the mechanistic connections to TonB-ExbBD, FepB and FepCDG-Fes. We determined the activation energy of FeEnt internalization through the OM was approximately 35 kcal/mol. The results indicated that the transport of FeEnt through FepA may involve significant conformational changes. These studies have resulted in a publication, Newton, S. M., Trinh, V., Pi, H., and Klebba, P. E. (2010) J Biol Chem 285, 17488-17497.

After determining that all FepA proteins bound to FeEnt actively transport FeEnt, we tried to determine the mechanism of FeEnt transport through FepA using FRET analyses. We engineered several double Cys substituted mutants in FepA. These FepA derivatives were labeled with the donor dye, fluorescein maleimide (FM), and the acceptor dye, Alexa Fluor 546 maleimide (A546M). The fluorescence intensities from an emission scan with an excitation at 488 nm was used to determine if energy transfer was occurring between the two dyes. Using the energy transfer efficiency, the distance between the two dyes could be calculated. Comparing the distance between the dyes in the absence and presence of FeEnt transport through FepA would help determine the mechanism of ligand transport through FepA.

After determining the double Cys mutant derivatives of FepA transported FeEnt similar to wild-type FepA, we optimized fluorescent labeling conditions with FM and A546M. Then, we ran excitation and emission scans of the single and double Cys substituted mutants in FepA. We were unable to show reproducible results that indicated energy transfer between the dyes in the absence or presence of FeEnt transport through FepA.

It was possible that the dipole orientation of the dyes are more static than anticipated, this would result in poor energy transfer between the fluors. By engineering new double Cys substituted mutants in FepA, it should be possible to remedy this problem. The labeling efficiency of A546M may also contribute to the results that indicated energy transfer did not occur between the dyes. If fractional labeling of A546M occurred, it was possible that FepA could be labeled with two FM dyes rather than just one. If the fractional labeling of A546M was determined, it would be possible to adjust the values obtained in the emission scan and determine the distance between the fluors. If we determined the energy transfer between the dyes in the absence and presence of FeEnt transport through FepA, we would determine the mechanism of FeEnt transport through FepA: ball-and-chain or transient pore model.

After determining FRET analysis could not be used for studies involving the conformational change of FepA during FeEnt transport, disulfide bond formation studies were conducted to determine the conformational change of the N-domain of FepA during FeEnt transport through FepA. We constructed two classes of double Cys mutants in FepA: N-terminus to N-terminus mutants in FepA and N-terminus to C-terminus mutants in FepA.

Siderophore nutrition tests, FeEnt accumulation determinations, and mobility shift assays were conducted under oxidizing and reducing conditions. The first two assays were used to determine if the formation of a disulfide bond hindered the transport of FeEnt through FepA. The N-terminus to N-terminus Cys substituted mutants in FepA were designed so the Cys residues were within cross-linking distance. With the exception of L125C/V141C, all N-terminus to N-terminus mutants in FepA indicated a formation of a disulfide bond within the globular N-domain of FepA. This cross-link hindered FeEnt accumulation through FepA. These results suggested that the N-domain of FepA needed to undergo conformational changes during FeEnt uptake through the OM.

N-terminus to C-terminus mutants were engineered where the Cys residues could not form a disulfide bond in the native FepA or in the absence of FeEnt transport. If the N-domain of FepA was displaced into the periplasmic space, like the ball-and-chain model, it was possible that the Cys residue in the N-domain could come in proximity of the Cys residue in the C-domain and form a disulfide bond. Based on the FeEnt accumulation assays and the siderophore nutrition tests a cross-link did not form between the two Cys substituted residues in FepA. If a disulfide bond formed, it did not hinder FeEnt transport through FepA. The electrophoretic mobility results indicated that all of the double Cys mutants in FepA formed a disulfide cross-link. This result was just qualitative and there were no quantitative results to determine the effects of the disulfide bond during FeEnt binding or transport.