Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2015 Jan 15;517(7534):396-400.
doi: 10.1038/nature13872. Epub 2014 Nov 2.

Subnanometre-resolution electron cryomicroscopy structure of a heterodimeric ABC exporter

Affiliations

Subnanometre-resolution electron cryomicroscopy structure of a heterodimeric ABC exporter

JungMin Kim et al. Nature. .

Abstract

ATP-binding cassette (ABC) transporters translocate substrates across cell membranes, using energy harnessed from ATP binding and hydrolysis at their nucleotide-binding domains. ABC exporters are present both in prokaryotes and eukaryotes, with examples implicated in multidrug resistance of pathogens and cancer cells, as well as in many human diseases. TmrAB is a heterodimeric ABC exporter from the thermophilic Gram-negative eubacterium Thermus thermophilus; it is homologous to various multidrug transporters and contains one degenerate site with a non-catalytic residue next to the Walker B motif. Here we report a subnanometre-resolution structure of detergent-solubilized TmrAB in a nucleotide-free, inward-facing conformation by single-particle electron cryomicroscopy. The reconstructions clearly resolve characteristic features of ABC transporters, including helices in the transmembrane domain and nucleotide-binding domains. A cavity in the transmembrane domain is accessible laterally from the cytoplasmic side of the membrane as well as from the cytoplasm, indicating that the transporter lies in an inward-facing open conformation. The two nucleotide-binding domains remain in contact via their carboxy-terminal helices. Furthermore, comparison between our structure and the crystal structures of other ABC transporters suggests a possible trajectory of conformational changes that involves a sliding and rotating motion between the two nucleotide-binding domains during the transition from the inward-facing to outward-facing conformations.

PubMed Disclaimer

Figures

Extended Data Figure 1
Extended Data Figure 1. Binding characterization of Fabs against TmrAB
a, Expression levels of Fabs used in ELISA in Fig 1a. Expression levels were assessed by immunoblotting with anti-c-myc antibody and normalized against the highest expression level such that undiluted AD12 equals normalized relative concentration of 1. Binding was monitored by anti-c-myc antibody. b, ELISA with purified Fabs against TmrAB. Purified flag tagged Fabs (Fab-flag) were used in binding reactions. Binding was monitored by anti-flag® M2-peroxidase. Experiments were repeated twice. c, Relative affinities of the Fabs. The ELISA signal data from the Extended Data Fig 1b were fit to a bimolecular binding equation to produce binding curves and apparent Kd values. AH5 showed the highest affinity, followed by BA6, AD12 and AH11. DH5 and BG12 did not show significant binding. Apparent Kd values do not yield meaningful biophysical properties because the concentration of TmrAB participating in the binding reaction is unknown and the transporter is not free in solution. However, comparison among the Fabs should be sufficient enough to determine their relative affinities. Despite unpurified DH5 showed significant binding (Extended Data Fig 1a), purified DH5 did not show significant binding. n.d. not determined. d, Negative stain EM analysis of Fab +TmrAB mixes. Representative 2-D class averages include complex images that show clear Fab densities and images that do not. The two typical Fab views, the dumbbell and donut shaped views, are indicated by yellow and red arrows, respectively. Fab images indicate that Fabs are rigid and form rigid complexes. Percentages of complex particles were assessed by fractions of the numbers of images that clearly show Fab densities (# complex particles) to the total numbers of images that were included (# total particles) in 2-D class averages. The percentages correspond to relative affinity ranking determined by ELISA (Extended Data Fig 1c).
Extended Data Figure 2
Extended Data Figure 2. Purification of a-DDM solubilized and Single particle cryo-EM of TmrAB-AH5 complex
a, Elution profiles of TmrAB alone and TmrAB-AH5 from Superdex 200 are shown in blue and red curves respectively, showing a clear shift of the elution peak of the TmrAB-AH5 complex to a higher molecular weight position. The shifted peak corresponding to fractions A6-A8 contained TmrAB and AH5, confirmed by SDS-PAGE. Fractions B3-B5 correspond to unbound AH5 and the loading material was run for comparison. b, Raw micrograph of TmrAB-AH5 (∼185kDa) embedded in a thin layer of vitreous ice. c, Fourier power spectrum calculated from micrograph shown in a. d, 2D class averages of TmrAB-AH5 complex. Fab AH5 is clearly visible in many class averages. e, Initial 3D reconstruction calculated from 2D class averages using common lines method implemented in SPIDER. f, Fourier shell correlation (FSC) curves of TmrAB-AH5 (red), TmrAB-BA6 (purple) and TmrAB alone (blue). g, Euler angle distribution of all particles used in the final reconstruction. h, Final 3D reconstruction colored with local resolution. i, Voxel histogram corresponding to local resolution. The majority of voxels is at 6 ∼ 7Å resolution. Estimation of local resolution that is too close to the Nyquist (3.9Å) may not be accurate.
Extended Data Figure 3
Extended Data Figure 3. Selected slice views of the 3D reconstruction of TmrAB-AH5
The views are oriented in parallel with the membrane plane. The numbers of slices are marked. a, All transmembrane helices of both TmrA and TmrB are labeled. The arrow points to the extra density in the cavity. b, Two NBDs are in contact with each other. c, The C-terminal helices of TmrA and TmrB are in close contact.
Extended Data Figure 4
Extended Data Figure 4. Single particle cryo-EM of TmrAB-BA6 complex
a, Raw micrograph of TmrAB-BA6 (∼185kDa) embedded in a thin layer of vitreous ice. Images were collected on a Tecnai TF20 microscope using scintillator based TVIPS 8K × 8K CMOS camera. b, Fourier power spectrum calculated from micrograph shown in a. c, 2D class averages of TmrAB-BA6 complex. Fab BA6 is clearly visible in many class averages. d, Initial 3D reconstruction of TmrAB-BA6 determined using common lines method implemented in SPIDER. e, Two different views of the final 3D reconstruction of TmrAB-BA6 filtered to a resolution of 9.4Å. Same as in the 3D reconstruction of TmrAB-AH5, density of micelle is split into two halves and tilted with each other. The orientation of micelle is marked with a pair of black solid lines and the gap in the micelle density generated by the helix H4 from TmrB is marked with a pair of red dashed lines. f, Densities of TmrAB in the 3D reconstructions of TmrAB-AH5 (khaki) and TmrAB-BA6 (grey mesh) overlap. Fabs AH5 and BA6 are pointed with arrows. g, An enlarged view to show the interface between TmrAB and BA6, which has a linear epitope in the NBD of TmrB.
Extended Data Figure 5
Extended Data Figure 5. Single particle cryo-EM of TmrAB alone without Fab
a, Raw micrograph of TmrAB alone (∼135kDa) embedded in a thin layer of vitreous ice. Images were collected on a Tecnai TF20 microscope using scintillator based TVIPS 8K × 8K CMOS camera. b, Fourier power spectrum calculated from micrograph shown in a. c, 2D class averages of TmrAB. d-f, Three different views of TmrAB 3D reconstruction shown in different (low: grey, high: gold) isosurface thresholds. Model of TmrAB (in ribbon diagram) was docked into the density map. The orientation of micelle is indicated with pairs of solid black lines in f and the gap of micelle is indicated with a pair of red dashed lines. g-h, Densities of TmrAB in the 3D reconstructions of TmrAB alone (transparent khaki) and in complex with AH5 (grey mesh) overlap with each other.
Extended Data Figure 6
Extended Data Figure 6. Cross correlation between TmrAB-AH5, TmrAB-BA6 and TmrAB
Left: density map of TmrAB-AH5 is colored according to the value of local cross correlation values of TmrAB-AH5 with TmrAB-BA6 (upper), with TmrAB (lower). Middle: density map of TrmAB-BA6 is colored according to the value of local cross correlation values of TmrAB-BA6 with TmrAB-AH5 (upper), and with TmrAB (lower). Right: density map of TmrAB is colored according to the local cross correlation value of TmrAB with TmrAB-AH5 (upper) and with TmrAB-BA6 (lower).
Extended Data Figure 7
Extended Data Figure 7. Atomic model of TmrAB
a, b, Two different views of the atomic model of TmrAB, generated by flexible fitting of the sequence homology model of TmrAB into the density map of TmrAB-AH5 complex. TmrA is colored in salmon, and TmrB is colored in blue. Intracellular loop 4 is colored in green. c, Two subunits are arranged with a pseudo-two fold symmetry.
Extended Data Figure 8
Extended Data Figure 8. AH5 and BA6 inhibit the ATPase activity of the TmrAB
a, ATP standard for panels b to e. b - e, ATP hydrolysis assay at 37°C. Reactions were carried out at 37 °C for 20 min with 6.25 μM of TmrAB, 250 μM ATP and 2 mM MgCl2 in the presence of 25 μM of AH5 (c), BA6 (d) or a negative control Fab, U33 (e). ATP hydrolysis by TmrAB was reduced in the presence of AH5 or BA6 when compared to the equivalent reaction in the absence of Fabs (b). ATP hydrolysis was not affected by the presence of U33, which does not bind to TmrAB (e). f, ATP and ADP standards (250 μM each) for g. Two peaks were resolved corresponding to ATP and ADP (black and red curves respectively). g, ATP hydrolysis by TmrAB was carried out with 70 nM of TmrAB, 250 μM ATP and 2 mM MgCl2 at 60 °C for 30 min. h, Identification of the TmrAB nucleotide binding state. ATP was not detected from the protein-extracted aqueous phase (red curve). ATP at an equivalent concentration (blue curve) is shown as a control to demonstrate sufficient sensitivity for nucleotide detection.
Extended Data Figure 9
Extended Data Figure 9. Cysteine cross-linking validating interaction between the C-terminal helices of TmrAB in the nucleotide free state
a, Three samples (marked with * in panel d of Fig. 3) were visualized by negative stain EM, showing that TmrAB with the double cysteine mutation has the native dimeric shape of TmrAB. TmrAB contains an exposed native cysteine residue (TmrA-C416) that could not be removed. It causes some inter-dimer cross-linking (marked by arrows) under the oxidative condition. b, Analytical HPLC demonstrating that purified TmrAB containing the A591C/A567C mutation is nucleotide free. c, ATP hydrolysis assay indicating that disulfide cross-linking inhibits the ATPase activity of TmrAB containing the double cysteine mutation. Assays were performed in triplicate for 1 hour at 60°C with 70nM reduced or oxidized TmrAB, 250μM ATP, and 2mM MgCl2 prior to analysis by analytical HPLC.
Figure 1
Figure 1. TmrAB Fab characterization
a, Qualitative ELISA to assess Fab binding to TmrAB. All Fabs showed binding to TmrAB except BG12 in both experiments A and B where independently prepared Fab samples were used. Expression levels were assessed by immunoblotting and normalized (Extended Data Fig 1a). b, Representative competitive ELISA between AD12-myc and Fab-flag against TmrAB. AD12-myc binding was not affected by the presence of AH5-flag or DH5-flag (top left). AH5-flag and DH5-flag maintained near maximum binding (*) at all AD12-myc concentrations (top right). AD12-myc binding was almost abolished in the presence of AD12-flag or BA6-flag (middle left). AD12-flag and BA6-flag maintained near maximum binding (*) at all AD12-myc concentrations (middle right). AD12-myc binding was not significantly affected in the presence of AH11-flag (bottom left). AH11-flag binding decreased as AD12-myc concentrations increased shown (bottom right). Uninhibited binding of AD12-myc and Fab-flag suggests independent binding between AD12 and AH5 or DH5. Inhibited binding of either AD12-myc or Fab-flag suggests overlapping epitopes between AD12 and BA6 or AH11. c, Representative negative stain 2D class averages of TmrAB-Fab complexes. d, Immunoblotting of TmrAB, using Fab-flag. Class A Fabs recognized the denatured form of TmrB, and AH5 did not recognize the denatured form of either strand. BG12 was used as a non-binder control, which did not detect either strand significantly.
Figure 2
Figure 2. 3D reconstruction of TmrAB-AH5 at subnanometer resolution
a, Cryo-EM density map of the TmrAB-AH5 complex filtered to a resolution of 8.2Å. The atomic model of TmrAB and an atomic structure of a Fab (pdb code: 1M71) are docked into the density map. The map shows two NBDs, a bi-lobed DDM micelle, which is separated by TM4 (marked by two dashed lines in the top view), and well-defined AH5 density. In the bottom view, two parallel solid lines and dashed lines indicate orientations of the front and back halves of the micelle respectively. The two halves are tilted by ∼30° with each other. b, The density map at a higher isosurface threshold shows clearly resolved TM helices. c, A cross section view through the TMDs shows well resolved TM helices labeled red and blue for TmrA and TmrB respectively. d, NBDs of TmrAB in two different views. C-terminal helices (*) of the two NBDs are in close proximity depicted in both of the density map and the docked atomic model.
Figure 3
Figure 3. Atomic model of TmrAB showing the laterally open inward-facing conformation
a, Representation of internal volume and opening to the external surface of the transporter. TmrA is colored salmon and TmrB is colored cyan. b, Two different views (tilted around the axis perpendicular to the membrane plan) of the substrate-binding cavity in the TMDs. Density bound to helix H4 (Tyr187) and H5 (His246) of TmrB was observed at a threshold of 5σ. It has the size to accommodate a Hoechst 33342 molecule, which is a known cargo molecule of TmrAB, but was not added during the protein purification. The cargo-like density is inside the cavity but near the inner leaflet of membrane. This position suggests a possible substrate pathway. c, Ribbon diagram of the TmrAB NBDs in the nucleotide free state. Predicted locations of two pairs of cysteine mutations are marked. d, Non-reducing SDS-PAGE gel demonstrating disulfide cross-linking of both double cysteine mutants in the apo-state, and showed a clear difference in cross-linking behavior between nucleotide free and bound TmrAB.
Figure 4
Figure 4. Comparison of the inward facing conformation of TmrAB with the intermediate inward facing structure of TM287/288 and the outward facing structure of Sav1866
a, Conformational changes in NBDs viewed from the membrane towards the cytosol. The Walker A motifs (blue) do not face the ABC signature motifs (green) in the atomic model of TmrAB (top) while they do in the TM287/288 structure (middle), but a single AMPPNP molecule (shown as sticks) is bound predominantly in one NBD. In the Sav1866 structure (bottom), the two motifs from the opposite domains come close to sandwich two ADP molecules (shown as sticks). b, Changes in lateral opening to the cavity. The lateral opening between TM4 and TM6 shown by dashed lines (left) narrows down from the TmrAB model (top) to the TM287/288 structure (middle), and completely closes in the Sav1866 structure (bottom), resulting in an opening to the extracellular space.

Similar articles

Cited by

References

    1. Rees DC, Johnson E, Lewinson O. ABC transporters: the power to change. Nature review Molecular cell biology. 2009;10:218–227. - PMC - PubMed
    1. Schmitt L, Tampe R. Structure and mechanism of ABC transporters. Current opinion in structural biology. 2002;12:754–760. - PubMed
    1. Gottesman MM, Ambudkar SV. Overview: ABC transporters and human disease. Journal of bioenergetics and biomembranes. 2001;33:453–458. - PubMed
    1. Parcej D, Tampe R. ABC proteins in antigen translocation and viral inhibition. Nature chemical biology. 2010;6:572–580. - PubMed
    1. Zutz A, et al. Asymmetric ATP hydrolysis cycle of the heterodimeric multidrug ABC transport complex TmrAB from Thermus thermophilus. J Biol Chem. 2010;286:7104–7115. - PMC - PubMed

Publication types

MeSH terms

-