Structures and transport dynamics of a Campylobacter jejuni multidrug efflux pump

Resistance-nodulation-cell division efflux pumps are integral membrane proteins that catalyze the export of substrates across cell membranes. Within the hydrophobe-amphiphile efflux subfamily, these resistance-nodulation-cell division proteins largely form trimeric efflux pumps. The drug efflux process has been proposed to entail a synchronized motion between subunits of the trimer to advance the transport cycle, leading to the extrusion of drug molecules. Here we use X-ray crystallography and single-molecule fluorescence resonance energy transfer imaging to elucidate the structures and functional dynamics of the Campylobacter jejuni CmeB multidrug efflux pump. We find that the CmeB trimer displays a very unique conformation. A direct observation of transport dynamics in individual CmeB trimers embedded in membrane vesicles indicates that each CmeB subunit undergoes conformational transitions uncoordinated and independent of each other. On the basis of our findings and analyses, we propose a model for transport mechanism where CmeB protomers function independently within the trimer.

the absence of ligand. The exponential lifetime can be obtained by fitting the data with a single-exponential function, resulting in (a) the presence of 1 M Tdc. The exponential lifetime can be obtained by fitting the data with a single-exponential function, resulting in (a)  L  I1 = 0.52 s (k L  I1 = 1.92 s -1 ), (b) the presence of 10 M Tdc. The exponential lifetime can be obtained by fitting the data with a single-exponential function, resulting in (a)

In vivo antimicrobial susceptibility assay
We used the C. jejuni 81-176 ΔcmeABC::cat null mutant strain, which lacks the cmeABC genes. We inserted the cmeABC operon that includes cmeABC, cmeR and the intergenic region between cmeR and cmeA into the 16S region of the 81-176 genomic DNA. This approach allowed us to ensure that the expression of cmeB was driven from a single copy of gene in the genomes with native regulator, operator and promoter. The expression level of cmeABC was determined using the anti-CmeB and anti-CmeC antibodies. Western analysis suggested that the expression level of wild-type CmeB, D409A and D410 mutant transporters were more or less the same (Supplementary Figure   8).
We then tested the susceptibility of C. jejuni cells carrying wild-type CmeB or its isogenic mutant (D409A or D410A) to taurocholate, taurodeoxycholate and rifampin.
These three antimicrobials are the known substrates of the CmeABC efflux pump 9 . We found that C. jejuni cells expressing the D409A or D410A mutant were >64-fold less sensitive to taurodeoxycholate when compared with C. jejuni cells carrying the wild-type CmeB pump (Supplementary Table 4). In addition, cells producing the D409A and D410A variants were >32-fold and >16-fold, respectively, less resistance to taurocholate when compared with cells expressing the wild-type pump. We also found that the minimum inhibitory concentrations (MICs) of C. jejuni cells producing D409A and D410A to rifampin were at least 256 and 64 times, respectively, lower than those of C.
jejuni cells carrying wild-type CmeB (Supplementary Table 4). These data show that both D409 and D410 residues are critical for the function of the CmeB pump.

Binding of taurodeoxycholate by CmeB
We used isothermal titration calorimetry (ITC) to determine the binding affinity of taurodeoxycholate (Tdc) for the CmeB multidrug efflux pump. The data indicate that the dissociation constant, K D , for Tdc binding is 3.26 ± 0.21 µM (Supplementary Figure   15 and Supplementary Table 5), confirming the purified CmeB protein is capable of recognizing this bile acid.
As cyclooctatetraene was needed in our FRET experiments to reduce the lifetime of dark states, we also determined the binding affinity of Tdc for CmeB in the presence of 2 mM cyclooctatetraene. ITC data indicated that the K D for Tdc binding is 2.79 ± 0.20 µM, suggesting that the presence of cyclooctatetraene does not affect the binding affinity of Tdc (Supplementary Table 5).

In vitro proton translocation across the CmeB proteoilposomes
We next examined if protons can translocate across the lipid bilayer of the CmeB proteoliposomes. The purified CmeB protein was reconstituted into liposomes containing the fluorescence proton-specific probe pyranine 43 in the intra-vesicular space, where the pH was adjusted to 7.5. When added into buffer solution containing 20 mM Na-HEPES (pH 6.5), we detected a significant quenching of the fluorescence signal in proteoliposomes possessing wild-type CmeB compared with those liposomes without the pump.
We then investigated whether protons can be transferred in the presence of 10 M Tdc. When Tdc was added into the extravesicular medium of the CmeB proteoliposomes,