|Home | About | Journals | Submit | Contact Us | Français|
Transporter proteins from the multidrug and toxic compound extrusion (MATE)1 family play vital roles in metabolite transport in plants2-3, directly affecting crop yields worldwide4. MATE transporters also mediate multidrug resistance (MDR) in bacteria and mammals5, modulating the efficacy of many pharmaceutical drugs used in the treatment of a variety of diseases6-9. MATE transporters couple substrate transport to electrochemical gradients and are the only remaining class of MDR transporters whose structure has not been determined10. Here we report the x-ray structure of the MATE transporter NorM from Vibrio cholerae determined to 3.65 Å, revealing an outward-facing conformation with two portals open to the outer leaflet of the membrane and a unique topology of the predicted 12 transmembrane helices distinct from any other known MDR transporter. We also report a cation-binding site in close proximity to residues previously deemed critical for transport11. This conformation likely represents a stage of the transport cycle with high-affinity to monovalent cations and low-affinity to substrates.
Cellular export of toxins and substrates is a fundamental life process, and members of the MATE family represent the last class of multidrug resistance (MDR) transporters to be structurally characterized. MATE transporters are involved in a variety of important biological functions across all kingdoms of life. In plants, MATE transporters are highly prevalent with 58 paralogues found in Arabidopsis thaliana12 and secrete a diverse range of secondary metabolites as a defense against herbivores and microbial pathogens3,5. Additionally, plant MATE transporters play an important role in tolerance towards phytotoxic aluminum in acidic soils, a major limitation of crop production in 50% of the world's arable land4. In mammals, they export a structurally diverse array of xenobiotic cations in the liver and kidney, influencing the plasma concentrations of many drugs including metformin, a widely prescribed Type 2 diabetes medication, thereby mitigating therapeutic efficacy8-9. Bacterial MATE transporters function primarily as xenobiotic efflux pumps and can confer resistance to tigecycline, a new glycylcycline class antibiotic developed to overcome methicillian-resistant and vancomycin-resistant Staphylococcus aureus6-7.
MATE transporters use either H+ or Na+ gradients across the membrane to drive substrate export, although the coupling mechanism is not well understood. All MATE proteins share ~40% protein sequence similarity5 (Fig. S1a), suggesting an overall conserved structure and transport function. To provide a basis for understanding the function of MATE transporters, we present two x-ray structures (Fig. 1) of NorM from V. cholerae (NorM-VC) with and without rubidium (Rb+) using protein purified in n-dodecyl-β-D-maltopyranoside, and crystallized with n-nonyl-β-glucopyranoside (β-NG) and a novel facial amphiphile13, 3α-hydroxy-7α,12α-bis[(β-D-maltopyranosyl)ethyloxy] cholane (FA-231) as additives by methods described previously14. Although nearly identical, the NorM-VC apo-structure was solved to a higher resolution at 3.65 Å. The structure reveals a novel topology distinct from all other MDR transporter families15-19. This outward-facing conformation also presents a monovalent cation-binding site within the internal cavity. NorM-VC retains transport activity in Escherichia coli (Fig. S2) and purified NorM-VC protein binds MATE substrates5 (Fig. S3; Table S1).
The structure of NorM-VC spans ~50 Å in the plane of the lipid bilayer and is arranged as two bundles of six transmembrane helices (TMs 1-6 and TMs 7-12) forming a large internal cavity open to the extracellular space (Fig. 1). The two halves are related by an intramolecular 2-fold, likely a result of gene duplication as reflected within its primary sequence. A cytoplasmic loop (residues 218-232 between TMs 6 and 7) connects the two halves, as also seen in transporters of the Major Facilitator Superfamily (MFS)19-21, while the initial helix of each half (TM 1 and TM7) are preceded by a helical extension (residues 2-18, and 233-247) from the inner membrane leaflet side. An additional helix (residues 450-461) after TM12 is nestled under the cytoplasmic side of TM11. The model was obtained as described in Supplementary Information (Fig. S4), refined (Table S2), and verified by multiple Fobs – Fcalc simulated annealing omit maps (Figs. S5-S6). The topology and structural correctness of the model was validated using a series of 16 mercury-labeled single cysteine mutant crystals (Figs. 2a, b, S1a, S7–S8; Table S3). Two structurally similar NorM-VC molecules (NorM1 and NorM2) constitute the asymmetric unit allowing independent verification of the position and identity for each mercury-labeled cysteine residue (Fig. S8).
The topology of NorM-VC is unique amongst all known transporters and unlike other 12 TM transporters such as those of the MFS, no TM-helices intersect the space defined by any cluster of three consecutive helices within each bundle (Fig. 2c, d). For example, in the MFS, TM1 intersects TMs 4-6. Likewise, TM7 intersects TMs 10-12. No such helical arrangement is seen in NorM. Two non-equivalent portals, formed by TMs 1 and 8 on one side and TMs 2 and 7 on the other, open within the lipid bilayer allowing the expulsion of substrates directly to the outer membrane leaflet or the extracellular space (Figs. 3 and S9). Within the lipid bilayer, the portals are maximally ~12 Å wide, sufficient to accommodate substrate passage. The volume of the internal cavity embedded within the lipid bilayer is ~4,300 Å3 (Fig. S10), approximately four times larger than the binding pockets of the transcription MDR regulators xPXR and BmrR22-23, and nearly two thirds that of ABC exporters Pgp or MsbA15,17. Residues facing the internal cavity are both relatively conserved and contributed by all TM helices except for TMs 3 and 9, which are located at the periphery in this conformation (Figs. 1 and S11). Of the 91 residues facing the cavity, 57 are hydrophobic, 14 are aromatic, 20 are polar, and 5 are charged. Nine aromatic residues are clustered in the c-terminal half of the internal cavity (Table S4).
In the internal cavity, we observed the binding of Rb+ and also cesium (Cs+) ions, which are heavier alkali metal analogs of sodium more easily visualized by x-ray crystallography (Figs. 4a, b)14. The cation binding site of NorM-VC is composed of residues from TMs 7, 8, and 10-12 (Table S4), a fundamentally different topology then found in members of the MFS family (Fig. 2c, d). Previous studies have identified three conserved acidic residues that are critical for transport function in NorM from V. parahaemolyticus (NorM-VP)11, a closely related ortholog of NorM-VC with 76% sequence identity and 86% similarity24-25. These residues correspond to D36, E255, and D371 in NorM-VC (D32, E251, and D367 in NorM-VP). In our Rb+-bound NorM-VC structure, residues E255 and D371 face the internal cavity on the C-terminal half of the molecule and are located near the cation (Fig. 4). For NorM-VP, both of these negatively charged residues are important for transport function as certain mutations at these positions decouple substrate/Na+ antiport11. Like NorM-VP, mutations at residue 255 in NorM-VC affect protein folding and do not express well11. Substitution mutations at D371 to alanine or asparagine abolish Rb+ and Cs+-binding in the outward-facing conformation (Fig. S12). These findings suggest that these and possibly other surrounding residues are crucial for the recognition and binding of cations from the extracellular space. Mutations of D367 in NorM-VP (D371 in NorM-VC) disrupt transport activity, further suggesting that cation binding at this site plays an important role in the transport cycle11. In our outward-facing NorM-VC structure, residue D36 (D32 in NorM-VP) is located in the N-terminal half, far from the C-terminal cation-binding site. We did not observe the presence of either Rb+ or Cs+ near D36 and speculate that this residue may play an important role in other conformations.
We propose that the outward-facing conformation of NorM-VC, like other efflux pumps26-27, represents a state in the transport cycle that has high affinity for monovalent cations (Fig. 4c; step 1), and a lower affinity for substrates. Although the internal cavity is accessible to mercury compounds (Figs. 2a, b and S7), this conformation is not amenable to co-crystallization or soaking with several well-established drugs or transport substrates. Upon cation binding, we propose that NorM-VC undergoes an outward- to inward-facing conformational change more favorable to substrate binding (Fig. 4c; step 2). Herein, cation release/substrate binding could induce a structural change back to the outward-facing conformation26-27, where the substrate is released into the outer leaflet of the lipid bilayer and/or extracellular space (Fig. 4c; step 3).
The structure of NorM reveals the last known MDR transporter family elucidated by x-ray crystallography. Unlike the Resistance, Nodulation, and Cell-Division (RND) transporter family where an extracellular domain mediates substrate binding, MATE transporters alongside MDR transporters from the ABC, MFS, and SMR families have poly-specific drug binding sites composed of TM helices within the lipid bilayer15-17,19. Despite wide primary sequence divergence, extraction and expulsion of substrates directly from the inner to the outer membrane leaflet through portals facing the lipid bilayer (Fig. 4c) may be a common theme for these transporter families. The shared V-shaped conformations suggests this may be the case, providing a molecular basis for hydrophobic/amphipathic substrate transport.
NorM-VC protein was recombinantly expressed in E. coli BL21 DE3 using pET19b, solubilized with n-dodecyl-β-D-maltopyranoside, and purified by Ni-affinity chromatography. Crystals were grown by the sitting drop method in buffer containing D2O with Tris-HCL, (NH4)2SO4, PEG250DME, 0.23% n-nonyl-β-glucopyranoside (β-NG), and 0.02% FA-231. Wild-type and mutant crystals were derivatized by mercury compounds dissolved in mother liquor. Diffraction data was processed with the programs HKL2000 and MOSFLM. Protein phasing, model building, and crystallographic refinement were achieved using the programs PHASES, SOLVE/RESOLVE, and CNS v1.2. To verify the identity and position of certain residues as well as characterizing the cation-binding site, NorM-VC wild-type and mutant crystals were soaked with the appropriate compounds and diffraction data collected. Isomorphous difference Fourier maps, calculated using suitable pairs of data sets, were calculated using CNS v1.2 and confirmed the mercury labeling of cysteine residues and the binding of Rb or Cs to the cation-binding site.
We thank the contribution of Dr. Y. Yin for initial clones of NorM. We thank CLS, SSRL, ALS, and APS. This work was supported by grants from the NIH GM70480 (GC) and GM73197 (QZ), the Beckman Foundation, the Skaggs Chemical Biology Foundation.
Full Methods and any associated references are available in the Supplementary Information section of this manuscript.
Author Contributions. X.H., P.S., A.K., and G.C. designed the experiments and wrote the manuscript. X.H., P.S., A.K., and R.E. performed experiments. W.H. and Q.Z. did chemical synthesis.
Coordinates and structure factors for the structures reported here are available from the Protein Data Bank with accession codes 3MKT and 3MKU.
Author Information Reprints and permissions information is available at www.nature.com/reprints.
The authors declare no competing financial interests.
Readers are welcome to comment on the online version of this article at www.nature.com/nature.