Insights into the bilayer-mediated toppling mechanism of a folate-specific ECF transporter by cryo-EM

Edited by Wolfgang Baumeister, Max Planck Institute of Biochemistry, Martinsried, Germany, and approved July 3, 2021 (received for review March 15, 2021)
August 18, 2021
118 (34) e2105014118

Significance

Gram-positive bacteria that lack biosynthesis pathways for micronutrients such as vitamins use a class of membrane proteins to acquire them from the environment. These membrane proteins are called energy-coupling factor (ECF) transporters and function in an unusual way. A membrane-embedded substrate-binding protein rotates within the membrane to bring a molecule from the outside to the inside of the cell. It was, however, not clear how this motion can occur within a bilayer environment. We use cryogenic electron microscopy at 200 kV to visualize an ECF transporter within a lipid bilayer. The transporter causes deformations to the surrounding lipid environment. These insights offer an explanation for how changes of the lipid environment enable such motion in the transport process.

Abstract

Energy-coupling factor (ECF)–type transporters are small, asymmetric membrane protein complexes (∼115 kDa) that consist of a membrane-embedded, substrate-binding protein (S component) and a tripartite ATP-hydrolyzing module (ECF module). They import micronutrients into bacterial cells and have been proposed to use a highly unusual transport mechanism, in which the substrate is dragged across the membrane by a toppling motion of the S component. However, it remains unclear how the lipid bilayer could accommodate such a movement. Here, we used cryogenic electron microscopy at 200 kV to determine structures of a folate-specific ECF transporter in lipid nanodiscs and detergent micelles at 2.7- and 3.4-Å resolution, respectively. The structures reveal an irregularly shaped bilayer environment around the membrane-embedded complex and suggest that toppling of the S component is facilitated by protein-induced membrane deformations. In this way, structural remodeling of the lipid bilayer environment is exploited to guide the transport process.
Energy-coupling factor (ECF)-type transporters are membrane protein complexes that belong to the superfamily of ATP binding cassette transporters (ABC), where they form the phylogenetically, mechanistically, and structurally distinct type III clade (1). A full ECF transporter complex is formed by a minimum of four polypeptide chains or domains: two cytosolic nucleotide-binding proteins (EcfA and EcfA’), a membrane-embedded scaffold protein (EcfT), and a membrane-embedded substrate-binding protein (S component). The former three proteins form a tripartite module of ∼93 kDa (named the energising or ECF module), while the S component of ∼19 kDa can be either associated with the ECF module (“full complex”) or be present in the membrane as a solitary protein. ECF transporter complexes catalyze the uptake of micronutrients such as vitamins and transition metals in bacteria, and their substrate specificity is mediated by a different S component for each substrate (2, 3). X-ray crystal structures suggest that the S component must topple within the membrane to bring the substrate from the outside to the inside of the cell (35). Coarse-grained molecular dynamics simulations have indicated that the membrane bends and thins near the ECF module, providing a favorable environment for the S component to topple within the lipid bilayer (5). However, experimental evidence for membrane deformations around ECF transporter complexes, required to substantiate this idea, is lacking. Most X-ray crystallization methods use detergent-solubilized proteins, rendering a direct visualizing of the bilayer environment and, consequently, the correct positioning of the protein bilayer, impossible (3, 4, 69). Therefore, structures of ECF transporter complexes embedded in a more native environment are required to understand their interplay with the lipid bilayer. In this context, single-particle analysis by cryogenic electron microscopy (cryo-EM) has shown to be a suitable method to study the role of membrane remodeling with respect to the function of membrane proteins (1017).
Here, we present two cryo-EM structures of the folate-specific ECF transporter complex (ECF-FolT2) in the inward-facing apo conformation at an overall resolution of 2.7 and 3.4 Å, with better-resolved local regions allowing to model water molecules. The cryo-EM structures obtained in lipid nanodiscs and detergent micelles reveal protein–lipid interactions and deformations of the surrounding environment, providing experimental evidence for a bilayer-mediated toppling mechanism. With a total molecular weight of around 115 kDa for the full complex and the absence of symmetry, ECF transporter complexes are challenging targets for cryo-EM. We demonstrate that cryo-EM reconstructions at better than 3-Å resolution of such small asymmetric membrane proteins from data obtained with a 200-kV cryogenic transmission electron microscope (cryo-TEM) is conceivable.

Results

Cryo-EM Structure of ECF-FolT2 in Lipid Nanodiscs.

We chose the folate-specific transporter complex ECF-FolT2 from Lactobacillus delbrueckii for analysis by cryo-EM because this transporter has been characterized previously by X-ray crystallography, liposome-based uptake assays, and coarse-grained molecular dynamics simulations, which provides a context for interpretation of the data (35). ECF-FolT2 was purified in n-dodecyl-β-D-maltopyranoside (DDM) micelles as described previously (4) (SI Appendix, Fig. S1A) and subsequently reconstituted into lipid nanodiscs. Escherichia coli polar lipid extracts were chosen for the reconstitution, since the transporter displayed high transport activities in liposomes composed of these lipids (4, 5). For the formation of lipid nanodiscs, we selected the membrane scaffold protein 2N2 (MSP2N2), as it can form lipid nanodiscs with diameters of ∼16 nm (18). This size would be large enough to ensure that the membrane-embedded domains of the complex are accommodated without size constraints, which might otherwise hinder the ability to make conformational changes.
To find suitable reconstitution conditions, we varied the amount of lipids while keeping the ratio of the transport complex to membrane scaffold protein constant and assessed the quality of the nanodisc reconstitutions by size-exclusion chromatography (SEC) (SI Appendix, Fig. S1B). The elution volumes of the complexes indicate that higher lipid content resulted in lipid nanodiscs of increasing diameter, while peak broadening suggests that higher lipid-to-protein ratios led to more heterogeneous size distribution of the protein–lipid nanodisc complexes. This was indeed observed in a cryo-EM dataset collected with the transporter reconstituted at a molar ratio of 1:5:500 protein: MSP2N2: lipids (SI Appendix, Fig. S1C). Already at a two-dimensional (2D) classification level, heterogeneity in the size of the lipid nanodiscs surrounding the transporter was evident. The three-dimensional (3D) classification results showed that the transporter is enclosed by elliptical lipid nanodiscs, ranging from ∼98 to 123 Å in the longest dimension. The refinement of the two dominant classes, however, only yielded reconstructions between 9- and 10-Å resolution.
We reasoned that higher resolution reconstruction might be obtained in single-particle analysis using ECF-FolT2 reconstituted at a molar ratio of 1:5:250 (protein:MSP2N2:lipids) because of the more-homogenous sample behavior by SEC (SI Appendix, Fig. S1 B and D). Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) analysis and ATPase activity assays confirmed that a full-length functional protein was reconstituted into lipid nanodiscs (SI Appendix, Fig. S1 D and E). Cryo-EM data acquisition was optimized by selection of regions with an even particle distribution based on sample thickness using an in-house–written script (SI Appendix, Fig. S2A) (19). While we generally target regions with a sample thickness of 20 to 40 nm for data collection, the dataset for the ECF transporter complex in nanodiscs was acquired in the 30- to 50-nm range. This however did not affect the quality of the data, with 90% of the cryo-EM images revealing contrast transfer function (CTF) resolution estimations of 4 Å or better (SI Appendix, Fig. S2B). The 2D class averages displayed relatively uniform lipid nanodisc dimensions along with visible secondary structure elements and distinct views corresponding to ECF-FolT2 (SI Appendix, Fig. S2C).
The lipid nanodisc is elliptical in shape and ∼88 × 76 Å in size. The relative small size is likely the result of the lipid-to-protein ratio used during the reconstitution and is in agreement with the observation that the membrane scaffold protein MSP2N2 forms lipid nanodiscs with a diameter of ∼16 nm at higher lipid-to-protein ratios, while smaller lipid nanodiscs are formed at lower ratios (18). Notably, smaller than 16-nm lipid nanodisc diameters with MSP2N2 were also observed in other structures where low amounts of lipids were used, such as the archaeal glutamate transporter homolog from Thermococcus kodakarensis (12). This elliptical (rather than circular) shape of the nanodiscs is most likely dictated by the reconstituted protein, which is elliptical with dimensions of ∼60 × 30 Å within the plane of membrane.
In the end, we obtained a reconstruction of ECF-FolT2 in lipid nanodiscs at 2.7-Å resolution (SI Appendix, Fig. S3 AD and Table S1), with better-resolved local regions allowing an unambiguous fit of the atomic coordinates (Fig. 1A) and several water molecules (SI Appendix, Fig. S4A). At present, this is the highest-resolution structure for an ECF transporter complex, and in fact the highest reported resolution for a small asymmetric membrane protein with cryo-EM data acquired with a 200-kV cryo-TEM.
Fig. 1.
Cryo-EM structure of the folate-specific ECF transporter complex in the inward-facing apo conformation embedded in lipid nanodiscs. (A) Atomic model of ECF-FolT2 in lipid nanodiscs as viewed from the membrane plane shown in cartoon representation built into the cryo-EM density map (transparent contours at 9.5 σ) with an overlay of the density corresponding to the lipid nanodisc drawn transparent (unsharpened map lowpass filtered to 6 Å and contoured at 0.325 σ). The two nucleotide-binding proteins (EcfA in red and EcfA’ in orange) and the two membrane-embedded proteins (EcfT in blue and FolT2 in yellow) form together ECF-FolT2. Water molecules shown as green spheres. Approximate boundaries of the membrane are indicated by black dotted lines. (B–E) Interactions of the S component with EcfT and EcfA’. Interactions formed between FolT2 to EcfT in the cryo-EM structure (B) and in the AMP-PNP–bound X-ray crystal structure of ECF-FolT2 (PDB: 5d3m) (C). Interactions formed between FolT2 to EcfA’ in the cryo-EM structure (D) and in the AMP-PNP–bound structure (E). Individual polypeptides are colored as in (A). Highlighted residues are represented in the ball-and-stick style. Hydrogen bonds and π-interactions are shown as black and green dotted lines. Selected structural elements and residues are labels. Schematic representations Above panels (B and D) indicate the zoomed in regions for panels Below.

The Inward-Facing apo State Is a Structurally Stable Conformation.

ECF-FolT2 in lipid nanodiscs adopts an inward-facing apo conformation, where the S component FolT2 assumes a substrate-free “toppled’” state with both nucleotide-binding proteins in the open and nucleotide-free conformation. This conformation resembles the postsubstrate-release state (3, 4, 69) obtained in previous crystallographic studies on ECF transporter complexes (Fig. 1A), showing that the inward-facing apo conformation represents a structurally stable conformation in near-native and nonnative environments.
Despite the overall similarity, there are pronounced differences in the interactions between the subunits of the complex. While the core of both nucleotide-binding proteins, as well as the coupling helices of the scaffold protein EcfT align overall well between the cryo-EM and X-ray crystal structures of ECF-FolT2 (4) (rmsd: 1.5 Å, SI Appendix, Fig. S4 B and C), deviations are seen in the C-terminal part of both nucleotide-binding proteins and in the membrane protein components of the complex (SI Appendix, Fig. S4 B and C).
Compared to the two X-ray crystal structures of ECF-FolT2, the transmembrane helices (TMH) of EcfT have hinged toward FolT2, with the pivot point corresponding to a conserved Pro71 (SI Appendix, Fig. S4D). As a consequence, FolT2 is packed more tightly against the coupling helices in the cryo-EM structures compared to both X-ray crystal structures (SI Appendix, Fig. S4E). This difference in domain conformation is translated into altered interactions. A π-interaction between FolT2 and EcfT is formed between Tyr134 in FolT2 and Phe116 in EcfT (Fig. 1B), which was not observed previously in the crystal structures (Fig. 1C). Also, interactions of loops L3 and L5 in FolT2 with TMH4 and TMH3 in EcfT, respectively, are lost when compared to the AMP-PNP–bound structure (Fig. 1 B and C).
Additional interactions are formed between the carboxyl-terminal tail of FolT2 and the nucleotide-binding protein EcfA’, which were also not observed in the X-ray crystal structures (Fig. 1 D and E). Here, two positively charged residues Arg174 and Lys175 in FolT2 are rotated toward and extend into the space between both nucleotide-binding proteins (Fig. 1D and SI Appendix, Fig. S4E). The arginine residue forms interactions with the backbone of residues Asp145 and Leu146 in EcfA’, while the sidechain of the same aspartate residue in EcfA’ forms an interaction with the backbone of Arg174 in FolT2 (Fig. 1D). Such interactions are not observed in the AMP-PNP–bound structure as the two positively charged residues in FolT2 are rotated away from EcfA’ (Fig. 1E). Notably, the surfaces of FolT2 and EcfA’ in this region are highly complementary in terms of shape and surface charges (SI Appendix, Fig. S4F).
The modular nature of ECF transporter complexes requires a degree of flexibility to facilitate the docking and toppling of the S components. The differences in the interactions observed between EcfT and the S component are in agreement with the dynamic nature of the transmembrane region of EcfT (3, 4, 6, 9), which is an important feature to accommodate different S components and to facilitate the toppling of the substrate-bound S component during the transport cycle (36). Although unrelated in sequence and in substrate specificities, different S components have a common structural fold and can dock onto the same ECF module (3, 4, 6, 20, 21). Interestingly, a previously unnoticed common feature is the presence of positively charged residues at the carboxyl-terminal end of different S components (SI Appendix, Fig. S5A). Analysis of a recent crystal structure of the pantothenate-specific ECF transporter complex (ECF-PanT) (3), which uses the same ECF module as FolT2, indeed revealed two positively charged residues, Arg203 and Lys205, that point into the space between both nucleotide-binding proteins (SI Appendix, Fig. S5B). Although not previously described, they form a similar interaction as observed in the here reported cryo-EM structure of ECF-FolT2. In ECF-PanT, the arginine residue in PanT interacts with the backbone of residue Phe144 in EcfA’, while Asp145 in EcfA’ forms an interaction with Lys202 in PanT (SI Appendix, Fig. S5C). It is likely that other S components form similar interactions, contributing to their ability to use of the same ECF module.

Protein-Mediated Deformation of the Surrounding Environment.

One of the advantages of cryo-EM is that apart from the target protein, also its surrounding environment is resolved. Visualization of the lipid nanodisc surrounding ECF-FolT2 makes it possible to define the lipid bilayer boundaries and potential bilayer deformations (Fig. 2). The membrane components EcfT and FolT2 form together a highly asymmetrical structure that is inserted into the lipid nanodisc. The two nucleotide-binding proteins EcfA and EcfA’ protrude from the lipid nanodisc on the side corresponding to the cytoplasmic side. On the opposite side corresponding to the extracellular side, neither EcfT nor FolT2 have regions that protrude from the lipid nanodisc. Notably, the long and predominantly hydrophobic loop connecting TMH3 and TMH4 in EcfT is embedded within the boundary of the lipid nanodisc, in contrast to the more-exposed conformation modeled in the crystal structure (Fig. 2 and SI Appendix, Fig. S4C).
Fig. 2.
Membrane deformations mediated by ECF-FolT2. Densities for cryo-EM maps displayed as described in Fig. 1. (A) Viewed from the membrane plane on the side of the exit pathway for folate from the cavity in FolT2. The exit pathway is indicated by a green arrow. The pivot point of the membrane and coupling domains of EcfT at conserved residue Pro71 in THM3 is indicated by an orange sphere. (B) Viewed from the membrane plane as in (A) but rotated by 180° highlighting the deformation of the lipid nanodisc at the cytoplasmic base of FolT2. The approximate thickness of the lipid nanodiscs was determined using Chimera X (22) and indicated by black arrows. Selected structural elements are labeled. Approximate boundaries of the membrane are indicated by black dotted lines.
The lipid nanodisc is deformed, nonuniform in thickness, and not shaped symmetrically. While the bulk of the nanodisc displays a thickness of ∼32 to 33 Å, it is substantially thinner near the interface between EcfT and FolT2 (Fig. 2). At the level of the exit path for folate to the cytosol, which is lined by the tip of loop L1 of FolT2 and a conserved Pro71 in TMH3 of EcfT, a thinning to ∼28 Å is observed (Fig. 2A). An even-stronger thinning of the lipid nanodisc to ∼23 Å is observed on the opposite side, where the cytoplasmic face of FolT2 rests on the coupling helices, as well as around the terminal regions of coupling helices 1 to 3 and the amino-terminal part of the loop connecting coupling helix 3 with TMH5 of EcfT (Fig. 2B).
It could be argued that the observed bilayer deformations might be caused by a poor fit of the transporter in the nanodisc or by direct interactions with the scaffold protein. However, the surface area of the bilayer in the nanodiscs is much larger than the footprint of the membrane components of the transporter (∼3,800 Å2 versus ∼3,100 Å2). In addition, an argument against the deformation due to interactions between the transporter and the scaffold protein is the lack of defined density for the membrane scaffold protein. Yet, to test whether nanodisc constraints might contribute to the observed deformation, we analyzed data obtained with larger nanodisc sizes by using higher amounts of lipids than described in section Cryo-EM Structure of ECF-FolT2 in Lipid Nanodiscs. Although these larger nanodiscs were heterogeneous in size and cryo-EM analysis only yielded low-resolution reconstructions, the ECF transporter complex appears to adopt a similar conformation to the one observed with tighter nanodiscs (SI Appendix, Fig. S1B).
This result shows that the size of the nanodisc does not dictate the observed structural features. To further rule out a potential artifact, we decided to collect data with ECF-FolT2 in detergent micelles, which are unconstraint by scaffold proteins. We obtained a good distribution of particles (SI Appendix, Fig. S6 A and B), and 2D class averages of particles extracted from around 700 movies also showed visible secondary structure elements, details within the detergent micelles, and distinct views corresponding to ECF-FolT2 (SI Appendix, Fig. S6C). Despite the relatively small dataset, we managed to obtain a reconstruction of ECF-FolT2 in detergent micelles at 3.4-Å resolution (SI Appendix, Fig. S7 AD and Table S1). The structure of ECF-FolT2 in detergent micelles (SI Appendix, Fig. S8A) is virtually identical to that of ECF-FolT2 in lipid nanodiscs (rmsd: 0.77 Å, SI Appendix, Fig. S8B), adopting the same inward-facing apo conformation. Altogether, this suggests that the surrounding environment has little influence on the structural conformation that ECF-FolT2 adopts. Analogously to the lipid bilayer in the nanodiscs, the detergent micelle is not uniform in thickness either. Even though the detergent micelle is overall thicker than the nanodisc, it thins around to comparable thickness as the nanodiscs in the same locations where the bilayer of the nanodisc reaches its thinnest parts (SI Appendix, Fig. S8C). Since the detergent micelle is not constrained by lateral forces of a scaffold protein, this result strongly indicates that the deformation of the surrounding environment is mediated by the transporter itself. Consistently, the detergent molecules of the micelle adopt similar orientations as the lipid molecules of the lipid nanodisc in the vicinity of the protein.

Mechanistic and Structural Importance of Protein–Lipid Interactions.

The reconstitution of ECF-FolT2 into lipid nanodiscs also made it possible to analyze the protein–lipid interactions. Multiple nonprotein densities were found around the membrane protein components of the complex (Fig. 3). The overall shape of the membrane-embedded part of the complex is a highly asymmetrical, with a much larger footprint on the cytoplasmic side compared to the extracellular side. Most of the observed lipid densities are confined to the respective inner leaflet of the membrane (Fig. 3 A–C). We interpreted the densities as annular phospholipids, whereby the precise identity (head group and acyl chains) remains ambiguous. Computational lipid fingerprint analysis indicated a preferred interaction of anionic lipids with the ECF transporter complex (5) and hence lipid densities were modeled where possible as phosphatidylglycerols. Based on their location, modeled lipids were grouped into three regions (Fig. 3 A–C).
Fig. 3.
Protein–lipid interactions. (A) Lipids at the cytoplasmic base of FolT2. The Inset is a zoomed view of the large boxed panel rotated by 90°, highlighting the density (blue) for another acyl chain (labeled I*), which either belongs to a second lipid molecule in this region or forms with the remaining density of lipid I (purple density) a cardiolipin molecule. (B) Lipids near the exit pathway for the substrate. (C) Lipids around EcfT viewed from the membrane plane. Global views of the interpreted lipids (green spheres) around the membrane components EcfT and FolT2 (surface representation) within the lipid nanodisc (contoured as in Fig. 2) as viewed from the membrane plane are shown next to schematic representations of the full complex. Large boxed panels represent zoomed views of regions as highlighted in the schematic representations. Proteins are shown in cartoon and surface representation. Modeled lipids are shown in stick representation (green) with the corresponding cryo-EM density contoured at 6.5 σ (purple). Roman numbers in the panel represent individual lipids or lipid groups referred to in the main text. Selected structural elements are labeled. The contours of the lipid nanodisc (unsharpened map lowpass filtered to 6 Å and contoured at 0.325 σ) are indicated in black. The exit pathway in (B) is indicated by a green arrow, and the inner and outer leaflet of the membrane are indicated by double-headed arrows (A–C). In all representations, proteins are colored according to Fig. 1.
The first region corresponds to the cytoplasmic base of FolT2 that is facing the lipid environment (Fig. 3A). A lipid molecule (labeled I in Fig. 3A) was found sandwiched between the interface shared by FolT2 and EcfT, from which one of the acyl chains protrudes into the interface almost parallel to the lipid nanodisc, while the other acyl chain wraps around TMH3 of FolT2. In this region, an additional density is present (labeled I* in Fig. 3A, Inset). It could be indicative for a second lipid molecule in this region or represent a third acyl chain forming a cardiolipin molecule, which was suggested previously based on computational lipid finger printing analysis (5). Another lipid molecule could be resolved between TMH4 and TMH5, where one of the acyl chains lies within a narrow groove between TMH4 and TMH5, while the other is wrapped around the amino-terminal region of TMH5 (labeled II in Fig. 3A). The third lipid molecule was observed at the same height as the amino-terminal end of TMH5 between TMH6 and the amino-terminal region of coupling helix 3 in EcfT (labeled III in Fig. 3A). Notably, all three resolved lipids demark the boundary of the thinnest region of the lipid nanodisc and lie almost horizontal to the membrane plane, indicating that the observed thinning is caused by tilting of the lipid molecules. Additionally, the lipids are located near positively charged residues (SI Appendix, Figs. S5A and S9A). Similarly, a detergent molecule (labeled as α in SI Appendix, Fig. S9B) and two potentially copurified lipid molecules (labeled as β and γ in SI Appendix, Fig. S9B) were observed in the same region with similar tilted orientations in the structure of ECF-FolT2 in detergent micelles. While the preferential occurrence of positively charged residues agrees with the positive-inside rule (23), this region was also suggested to provide binding sites for phosphatidylglycerols or cardiolipins (5).
The second region corresponds to the exit pathway for the substrate (Fig. 3B). A density corresponding to an acyl chain of a lipid is present at the amino-terminal part of coupling helix 2 in EcfT and the loop L1 of FolT2 (labeled IV in Fig. 3B), with the latter likely shielding the exit pathway from the hydrophobic milieu. A second density for a phospholipid acyl chain is sandwiched between two tryptophan residues near the pivot point of the domains of EcfT (labeled V in Fig. 3B). It might help to stabilize the hydrophobic cleft that is formed by the kink in TMH3 of EcfT induced by the conserved Pro71.
The third region corresponds to the TMH of EcfT (Fig. 3C). Densities corresponding to acyl chains of lipids could be interpreted near a positively charged belt around the TMH of EcfT on the inner leaflet (labeled VI in Fig. 3C and SI Appendix, Fig. S9C), while the only densities on the outer leaflet are found by the amino-terminal end of TMH4 and between TMH2 and TMH3 (labeled VII in Fig. 3C). Notably, all lipids in this region have a canonical orientation of the acyl chain (roughly perpendicular to the membrane plane) rather than the tilted orientation as in the region by the cytoplasmic base of FolT2 (Fig. 3A).

Discussion

ECF transporter complexes mediate the uptake of micronutrients such as vitamins and transition metals in bacteria. They have been proposed to operate by a bilayer-mediated toppling mechanism (5). Our work provides experimental evidence that ECF transporter complexes are likely to remodel the lipid bilayer in order to facilitate transport.
The cryo-EM reconstructions determined in lipid nanodiscs and detergent micelles demonstrate that the S component assumes a toppled orientation in the inward-facing apo conformation, irrespective of the surrounding environment. The structures also reveal a deformation and thinning of the surrounding environment and hint at the mechanistic and structural role of lipids. In the inward-open apo conformation, a thinning of the local environment occurs near the cytoplasmic base of the S component, which is solvent exposed and overall positively charged. Lipid molecules are resolved in this region, highlighting the importance of specific interactions to maintain the boundary of the deformation. When replacing the model of ECF-FolT2 with the model of the pantothenate-specific ECF transporter complex (ECF-PanT, Protein Data Bank [PDB]: 6zg3), the positively charged cytoplasmic base of PanT also remains solvent exposed (SI Appendix, Fig. S9D). It is likely that the deformation of the local environment provides the means to keep the highly positively charged cytoplasmic base of S components solvent exposed when toppling within the membrane. Furthermore, it could also be important for the stabilization of the inward-facing conformation together with the interaction formed with EcfA’ (Fig. 1 D and E). Notably, this region is resolved to the highest resolution in the cryo-EM reconstructions (SI Appendix, Figs. S3C and S7C). The thinning of the lipid nanodisc observed near the exit pathway for the transported substrate is likely a long-range effect caused by the deformation on the opposite side and in terms of biological membranes can be explained by the formation of an indentation in the local environment of the transporter. The remodeling of the local environment by thinning of the lipid bilayer and the protein–lipid interactions could provide a mechanism to reduce the energy barrier for the S component to topple within the membrane in the transport cycle and to create a pathway leading to the correct relative orientation of the ECF module and the S component for docking.
We therefore extend the bilayer-mediated toppling mechanism by also considering the synergetic properties of the components in the ECF transporter complex (Fig. 4). The unengaged S component in the outward-facing open conformation captures the substrate from the external environment. A freely available ECF module thins and bends the membrane locally, where it adopts a slightly tilted conformation (5). This conformation and local membrane thinning attract the substrate-bound S component to dock and drive the toppling of the S component within the membrane. (State 1 in Fig. 4). In the toppled state, the bound substrate is released into the cell. In this state, the S component is tightly engaged with the ECF module stabilizing the inward-facing conformation. At the same time, the local environment is deformed to ensure that the charged cytoplasmic base remains solvent exposed, resulting in a further stabilization of the inward-facing conformation (State 2 in Fig. 4). While the inward-facing apo conformation represents a conformation that is stabilized by protein–protein interactions, the bilayer deformations observed in our structure may cause some strain. Binding and hydrolysis of ATP releases the empty S component and possibly relieves the strain, allowing the rotation of the S component back to the outward-facing state so that in can engage in binding of a new substrate (State 3 in Fig. 4).
Fig. 4.
Substrate transport by a bilayer-mediated toppling mechanism of ECF transporter complexes. Schematic representation of the different states of the transport cycle by ECF transporter complexes with exchangeable S components (so-called group II ECF transporters) (24). State 1: The substrate-bound S component docks to a freely available ECF module, which is in a slightly tiled conformation and thins and bends the membrane. The S component topples within the membrane. State 2: The ECF transporter complex assumes the inward-facing apo conformation, in which the bound substrate is released into the cell. The S component locally deforms the membrane by which the positively charged base of the FolT2 remains solvent exposed. State 3: Binding and hydrolysis of Mg-ATP results in the rotation of the apo S component to the outward-facing conformation and the release from the ECF module for renewed substrate binding. The ECF module tilts and bends the membrane for binding a substrate-bound S component. State 3 is based on structural evidence by cryo-EM (this work) and X-ray crystallographic studies. The remaining states are based on molecular dynamics simulations (5).
Small asymmetric membrane proteins with a molecular weight of 150 kDa or lower pose challenges for single-particle cryo-EM because of their inherently low signal-to-noise ratio and lack of contrast in vitreous ice. Imaging with 200-kV cryo-TEMs or lower voltages is gaining increasing prominence, in particular with membrane proteins. Especially small (membrane) proteins benefit from the increased image contrast originated from the higher elastic cross-section at lower acceleration voltages (25, 26). However, adequate ice thickness ranges are required to maximize the quality and efficacy of the output, especially at lower acceleration voltages, for which our improved data collection scheme based on prior sample thickness measurements has proven to be instrumental (19). The capabilities of 200-kV TEMs of reaching better than 3-Å resolution have been demonstrated by various groups with soluble proteins (2733). Our work is a further demonstration that this can be also accomplished with a small asymmetric membrane protein. It is likely that a combination of the biochemistry, technical aspects, and image processing contributed to achieve such a resolution in the end.
The influence of lipids and the deformation of the lipid bilayer on the function of membrane proteins is becoming increasingly acknowledged and has been highlighted in single-particle cryo-EM studies on the proteins belonging to the TMEM16 family (10, 11), the archaeal glutamate transporter homologs (12, 13), the human endoplasmic reticulum (ER) membrane protein complex (14), the β-barrel assembly machinery (15), the bacterial mechanosensitive channel of small conductance (16), and the human signal peptidase complex (17), to name a few. Our work overall provides experimentally a glimpse on the bilayer-mediated toppling mechanism by single-particle cryo-EM, which allows the dynamic interaction of different S components with the same ECF module. Moreover, it lays the foundations to further explore the influence of membrane deformations on the transport cycle in ECF transporter complexes within a near-native environment. ECF transporter complexes have recently gained renewed attention with the identification of a novel haem transporter belonging to the family of ECF-type ABC transporters (3436). As various ECF transporters complexes are encoded in the genome of various human pathogens but not in humans (37, 38), it makes them attractive targets for newly designed antimicrobial compounds, for which this work represents a steppingstone toward studying their means of action on a structural level.

Materials and Methods

Cloning, Expression, and Membrane Vesicle Preparation.

The cloning of ECF-FolT2 into the p2BAD expression vector was described previously (4). His10-ECF-FolT2-StrepII was expressed in E. coli MC1061 cells as described previously with minor changes (4). Briefly, cells were grown in 5-L baffled flasks containing 2 L 2xYT liquid medium supplemented with 2.5 mM potassium phosphate (KPi), pH 7.5, 0.5% [vol/vol] glycerol, 100 µg/mL ampicillin, and 200 µL/L Antifoam 204 (Sigma) at 37 °C with shaking at 200 rpm for 2 h. The temperature was subsequently reduced to 25 °C, and protein expression was induced at an optical density of 600 nm of ∼0.8 with 0.01% (wt/vol) L-arabinose. Cells were harvested 3 h postinduction (7,500 × g, 15 min, 4 °C), washed in storage buffer (50 KPi, pH 7.5, 10% [vol/vol] glycerol), and membrane vesicles prepared as previously described (4). Membrane vesicles were stored in aliquots containing 30 mg total protein as determined by a bicinchoninic acid assay kit (Thermo Fisher Scientific) at −80 °C after flash-freezing in liquid nitrogen.

Purification of ECF-FolT2.

ECF-FolT2 was purified as described previously with minor changes to the protocol (4). Briefly, membrane vesicles containing 60 mg total protein were buffered in 50 mM KPi, pH 7.5, 300 mM NaCl, 1% (wt/vol) DDM, Glycon), and solubilized for 1 h at 4 °C at the lowest speed on a rotator. The solubilized sample was clarified by ultracentrifugation at 286,625 × g for 35 min, and the supernatant was incubated with 0.5 mL Nickel-Sepharose 6 Fast Flow beads (Cytiva) for 1 h at 4 °C at the lowest speed on a rotator, which were previously equilibrated with wash buffer (50 mM KPi, pH 7.5, 300 mM NaCl, 50 mM imidazole, pH 8.0, 0.05% (wt/vol) DDM). The suspension was packed into an empty polypropylene column (Bio-Rad), and the beads were washed with 30 column volumes wash buffer. Bound proteins were eluted with elution buffer (50 mM KPi, pH 7.5, 300 mM NaCl, 500 mM imidazole, pH 8.0, 0.04% [wt/vol] DDM) in 250-, 850-, and 250-µL fractions. The second elution fraction was used for further purification by SEC using a Superose 6 Increase 10/300 GL column (Cytiva) equilibrated with SEC buffer (20 mM Tris, pH 8.0, 150 mM NaCl, 0.0261% [wt/vol] DDM). Peak fractions containing ECF-FolT2 were pooled and used either for grid preparation or for reconstitution into lipid nanodiscs.

Reconstitution of ECF-FolT2 into Lipid Nanodiscs.

The reconstitution procedure was performed according to established protocols (18, 39). Briefly, E. coli polar lipid extracts (Avanti Polar Lipids) in chloroform were vacuum dried and rehydrated in 20 mM Tris, pH 8.0, 150 mM NaCl, 1.75% (wt/vol) DDM to a concentration of 20 mg/mL Freshly purified ECF-FolT2 from the SEC step was mixed with purified MSP2N2 and lipids at a 1:5:250 molar ratio of ECF-FolT2: MSP2N2: lipids and kept on ice for 30 min, after which 200 mg/mL washed Bio-Beads SM-2 (Bio-Rad) were added and left to incubate at 4 °C on a rotator at the lowest speed. Another portion of 200 mg/mL Bio-Beads were added and left to incubate for 16 h overnight. The reconstitution mixture was separated from the Bio-Beads and incubated with 0.5 mL Nickel-Sepharose 6 Fast Flow beads for 1 h at 4 °C at the lowest speed on a rotator, which were previously equilibrated with wash buffer (20 mM Tris, pH 8.0, 150 mM NaCl). The suspension was packed into an empty polypropylene column, and the beads were washed with 40 column volumes wash buffer. Bound proteins were eluted with elution buffer (20 mM Tris, pH 8.0, 150 mM NaCl, 500 mM imidazole, pH 8.0) in 250-, 850-, and 250-µL fractions. The second elution fraction was used for further purification by SEC using a Superose 6 Increase 10/300 GL column (Cytiva) equilibrated with SEC buffer (20 mM Tris, pH 8.0, 150 mM NaCl). Peak fractions containing the ECF-FolT2-MSP2N2-lipid complex were pooled and used for grid preparation.

Malachite Green Phosphate Assay.

The ATPase activity was determined using a commercially available Malachite Green Phosphate Assay Kit (Sigma). The reactions were prepared according to the manufacturer’s instructions in volume of 100 µL. Briefly, the ATPase activity was allowed to proceed for 15 min at 30 °C with 0.5-µg nanodisc–embedded protein in SEC buffer (20 mM Tris, pH 8.0, 150 mM NaCl) supplemented either 200 µM Mg-ATP or 200 µM Mg-ADP or no adenine nucleotides. The reactions were stopped with 1.25 mM ethylenediaminetetraacetic acid (EDTA), pH 8.0, and 20 µL working solution from the Malachite Green Phosphate Assay Kit was added to a reaction volume of 80 µL in 96-well plates. The color was allowed to develop for 30 min at room temperature before measuring the ATPase activity at 640 nm with a TECAN Sparx plate reader. The measurement was performed as biological duplicate containing each technical triplicate. Phosphate concentrations were calculated from standard curves obtained with the supplied phosphate standard of the kit.

Cryo-EM Sample Preparation of ECF-FolT2 in DDM Micelles and Data Acquisition.

Prior to grid preparation, the sample was concentrated to 7.9 mg/mL with a 100-kDa molecular weight cut-off (MWCO) centrifugal concentrator (Millipore) and used for immediate for grid preparation. Quantifoil Au R1.2/1.3 300-mesh and UltrAuFoil R1.2/1.3 300-mesh (Quantifoil) grids were glow-discharged (Edwards Scancoat 6) at 5 mA for 30 and 60 s, respectively. The sample was applied at a volume of 2.8 µL onto grids, which were blotted for 4 s after a wait time of 2.5 s using a Vitrobot Mark IV (Thermo Fisher Scientific) at 15 °C and 100% humidity, subsequently plunge frozen in a liquid ethane-propane mixture and stored in liquid nitrogen until use. Cryo-EM data were collected on a Talos Arctica microscope (Thermo Fisher Scientific) operating at 200-kV with a BioQuantum postcolumn energy filter (Gatan) in zero-loss mode, a 20-eV slit, and a 100-µm objective aperture, in an automated fashion with EPU version 1.12 (Thermo Fisher Scientific) on a K2 Summit detector (Gatan) in counting mode. Movies were recorded at a physical pixel size of 1.012 Å (calibrated magnification of 49,407×), a defocus range from −0.8 to −1.8 µm, an exposure time of 9 s with a subframe exposure time of 150 ms (60 frames), and a total electron exposure on the specimen level of 53.3 electrons per Å2. The best regions on the grid for data acquisition were screened and selected with an in-house–written script to calculate the ice thickness (19), which was set between 20 and 40 nm, and the data quality was monitored on-the-fly using the software FOCUS version 1.0.0 (40).

Image Processing and Model Building of ECF-FolT2 in DDM Micelles.

A total of 464 and 357 dose-fractioned cryo-EM images were recorded back to back from the Quantifoil (dataset 1) and UltrAuFoil (dataset 2) grids (frozen at the same time with protein from the same purification), respectively, and subjected to motion correction and dose weighting of frames in MotionCor2 version 1.2.1 (41). The CTF parameters were estimated on the movie frames with CTFFIND4.1.13 (42). Low-quality images were removed on the basis of visual inspection of the images showing contamination or poor CTF estimation, and a total of 443 and 265 images remained for further processing from dataset 1 and 2, respectively. MotionCor2 and CTFFIND4 were used within the FOCUS software. Particles were picked with crYOLO version 1.2.2 using the PhosaurusNet architecture with anchors set to 160 (43). A model was created and trained on eight and three micrographs from dataset 1 and 2, respectively, which was subsequently used for picking on all images. A total of 75,688 and 85,775 particles were picked from dataset 1 and 2, respectively, and imported into RELION version 3.0.7 (44). Particles from each set were extracted separately using a box size of 256 pixels, which were subjected to three rounds of 2D classification with a particle mask diameter of 200 Å and the expectation step limited to 10 Å resolution, after which particles from both data sets were combined, and three additional rounds of 2D classification was performed. A robust classification of the extracted particles was achieved by ignoring the effects of the CTF until the first zero transition in all 2D classification rounds. A final 2D classification step was performed without ignoring the effects of the CTF until the first zero transition. 2D classes that showed high-resolution features were selected (73,900 particles) for 3D auto-refinement into a single consensus map using an initial model generated de novo by Stochastic Gradient Descent (SGD) method in RELION and lowpass filtered to 60 Å. The overall resolution of the consensus map after postprocessing was 3.81 Å resolution. The overall resolution was improved to 3.76 Å of the best 3D class (52,109 particles) after 3D classification into four classes. The resolution was further improved to 3.70 Å after CTF refinement. Particles were subjected to Bayesian polishing (45) after obtaining optimal parameters (s_vel: 0.7995, s_div 5190, s_acc 5.46) yielding a map at 3.55-Å resolution. Another round of CTF refinement improved the resolution to 3.45-Å resolution. The resolution dropped even further after removing particles with a CtfMaxResolution below 4-Å resolution using the subset selection option, yielding a final map containing 41,963 particles at 3.41-Å resolution. All reported resolutions were estimated with the 0.143 cutoff criterion (46) with gold-standard Fourier shell correlation (FSC) between two independently refined half maps (47). During postprocessing, high-resolution noise substitution was used to correct for convolution effects of real-space masking on the FSC curve (48). The directional resolution anisotropy of density maps was quantitatively evaluated using the remote 3DFSC processing server (accessible at https://3dfsc.salk.edu) (49). Local resolution variations were estimated with RELION. The data collection and processing parameters are summarized in SI Appendix, Table S1. The final map was sharpened with LocalDeblur (50) using the local resolution map estimated with MonoRes (51) through Scipion version 2.0 (52). Initially, individual chains from the crystal structure of ECF-FolT2 in the apo state (PDB: 5jsz) was fitted into density of the sharpened map in UCSF Chimera version 1.14 (53). The entire model was subsequently adjusted and improved in COOT version 0.9-pre (54). Residues for amino acids 1,15 to 16, and 282 in EcfA, amino acids 1, 13 to 21, and 283 to 287 in EcfA’, amino acids 1 to 9 in FolT2, and amino acids 1 to 6 in EcfT could not be place due to poor or absent densities. In EcfA, amino acid 1 was modeled as a glycine, which is part of the tobacco etch virus protease (TEV)-cleavable decahistidine affinity tag and replaces the starting methionine at that position. Two residues of the STREPII affinity tag at the carboxyl-terminal end of FolT2 could be also modeled. The model was subjected to an iterative process of real-space refinement against the postprocessed map in PHENIX version 1.18-3861 (55). All programs used for image processing were accessed through SBGrid (56).

Cryo-EM Sample Preparation of ECF-FolT2 Embedded in Lipid Nanodiscs and Data Acquisition.

Prior to grid preparation, nanodisc-embedded ECF-FolT2 used for the inward-facing apo conformation was concentrated to 7.9 mg/mL with a 100-kDa MWCO centrifugal concentrator and then used immediately for grid preparation. Quantifoil Au R1.2/1.3 300-mesh grids were glow-discharged at 5 mA for 30 s. Prior to sample application, fluorinated Fos-choline 8 (Anatrace) was added to a final concentration of 2.9 mM to obtain an even distribution of well-separated particles. The sample was immediately applied at a volume of 2.8 µL onto grids, blotted for 3 to 4 s using the Vitrobot at 15 °C and 100% humidity, and subsequently flash frozen in a liquid ethane-propane mixture and stored in liquid nitrogen until use. Cryo-EM data were collected as described in section Cryo-EM Sample Preparation of ECF-FolT2 in DDM Micelles and Data Acquisition with minor changes as follows. Data were collected in an automated fashion with EPU version 2.4 (Thermo Fisher Scientific). The detector was operating in superresolution mode, with movies recorded at a physical pixel size of 0.771 Å (calibrated magnification of 64,851×), a defocus range from −0.8 to −1.8 µm, an exposure time of 9 s with a subframe exposure time of 150 ms (60 frames), and a total electron exposure on the specimen level of 51.1 electrons per Å2. As a result of adding fluorinated Fos-choline 8, the ice thickness script was adjusted to 30 and 50 nm to obtain the best regions on the grid for data acquisition.

Image Processing and Model Building of ECF-FolT2 Embedded in Lipid Nanodiscs.

A total of 4,722 movies were collected from four grids (frozen at the same time with protein from the same purification) in three sessions over the course of 3 mo and subjected to gain correction, Fourier cropping by a factor of two, motion correction, and dose weighting of frames in MotionCor2. The CTF parameters were estimated on the movie frames with CTFFIND4. Low-quality images were removed on the basis of visual inspection of the images showing contamination or poor CTF estimation, and a total of 3,974 movies remained for further processing. Particles were picked with crYOLO version 1.5.0 using the PhosaurusNet architecture with anchors set to 180. A picking model was created and trained on 10 micrographs, which was subsequently used for picking on all images. A total of 1,362,547 particles were picked and imported into the beta-release of RELION version 3.1 (57). Particles were extracted with a box size of 320 px and downscaled initially to 3.084 Å/px, which were then subjected to four rounds of 2D classification with a particle mask diameter of 180 Å, with the expectation step gradually increased from 10 to 6 Å in each step. A robust classification of the extracted particles was achieved by ignoring the effects of the CTF until the first zero transition in all 2D classification rounds. Resulting 2D classes displaying high-resolution features were selected (658,218 particles) for 3D classification into three classes using an initial model generated de novo by the SGD method in RELION and lowpass filtered to 60 Å. The particles from the best class (392,698 particles) were re-extracted and recentered with a box size of 320 pixels without downscaling, which were subjected to 3D auto-refinement resulting in a reconstruction at 3.16-Å resolution after postprocessing with the resolution not improving after a round of CTF refinement (beam-tilt estimation, per-particle defocus, and per-micrograph astigmatism). The resolution, however, improved beyond 3 Å after Bayesian polishing with default parameters and CTF refinement (beam-tilt estimation, per-particle defocus, and per-micrograph astigmatism) to 2.94 Å. Particles from the best class (282,236 particles) after 3D classification (tau_fudge = 8 and E-step limit = 6 Å) using a soft mask (4-pixel extension and 8-pixel soft cosine edge) without image alignment into three classes, and CTF refinement (beam-tilt estimation, per-particle defocus, and per-micrograph astigmatism) resulted in a reconstruction at 2.87-Å resolution. Another round of Bayesian polishing with particles re-extracted with a box size of 386 pixels resulted in a small improvement to 2.85-Å resolution. CTF refinement (trefoil estimation, anisotropic magnification estimation, beam-tilt estimation, per-particle defocus, and per-micrograph astigmatism) yielded a reconstruction at 2.74-Å resolution. Particles from the best class (166,524 particles) of a final round of 3D classification (tau_fudge = 12 and E-step limit = 6 Å) using a soft mask around the protein density (4-pixel extension and 8-pixel soft cosine edge) without image alignment into three classes and two rounds of CTF refinement trefoil estimation, anisotropic magnification estimation, beam-tilt estimation, per-particle defocus, and per-micrograph astigmatism) resulted in a final reconstruction at 2.69-Å resolution. All reported resolutions were estimated as described in section Image Processing and Model Building of ECF-FolT2 in DDM Micelles, and local resolution variations for both maps were estimated with RELION. Directional anisotropy was evaluated as described in section Image Processing and Model Building of ECF-FolT2 in DDM Micelles. The data collection and processing parameters are summarized in SI Appendix, Table S1. The final map was sharpened as described in section Image Processing and Model Building of ECF-FolT2 in DDM Micelles. The model from the cryo-EM structure of ECF-FolT2 in DDM micelles was fitted into the density in Chimera after which the model was subsequently adjusted and improved in COOT. Residues for amino acids 1, 15 to 16, and 282 in EcfA, amino acids 1, 12 to 20, and 283 to 287 in EcfA’, amino acids 1 to 9 in FolT2, and amino acids 1 to 4 in EcfT could not be placed due to poor or absent densities. Water molecules were added to the model using phenix.douse as part of the PHENIX program, which were manually curated and retained based on logical interaction geometries and clear and visible densities. The model was subjected to an iterative process of real-space refinement against the postprocessed map in PHENIX. All programs used for image processing were accessed through SBGrid.

Structure Validation.

The quality of the fit was validated by a Fourier shell cross correlation (FSCsum) between the refined model and the final map. To monitor the effects of potential overfitting, random shifts (up to 0.5 Å) were introduced into the coordinates of the final model, followed by refinement against the first unfiltered half map. The FSC between this shaken-refined model and the first half map used during validation refinement is termed FSCwork, and the FSC against the second half map, which was not used at any point during refinement, is termed FSCfree. A marginal gap between the curves describing FSCwork and FSCfree indicates no overfitting of the model. The chemophysical properties of the final model were validated by MolProbity (58).

Figure Preparation.

Figures were prepared with PyMOL version 2.4.1 (The PyMOL Molecular Graphics System, Schrödinger, LLC), Chimera and UCSF ChimeraX version 1.1 (22). Final illustrations were prepared with Affinity Designer version 1.8.6 (Serif Ltd).

Data Availability

Data supporting the findings of this manuscript are available from the corresponding authors upon request. Cryo-EM density maps, half maps, and masks of ECF-FolT2 in DDM micelles and MSP2N2 lipid nanodiscs have been deposited in the Electron Microscopy Data Bank (EMDB) under Accession Nos. EMD-12483 and EMD-12484, respectively. The respective atomic models are available through the PDB under the accession codes 7NNT (ECF-FolT2 in DDM micelles) and 7NNU (ECF-FolT2 in MSP2N2 lipid nanodiscs).

Acknowledgments

We thank A. A. Garaeva for freezing grids of ECF-FolT2 in detergent micelles, G. T. Oostergetel for help with initial cryo-EM data acquisition, and M. Punter for maintenance of the image processing cluster. As part of the COFUND project oLife, C.T. acknowledges funding from the European Union’s Horizon 2020 research and innovation program under the Grant Agreement 847675. C.P. acknowledges funding from the Dutch Research Council: Nederlandse Organisatie voor Wetenschappelijk Onderzoek (NWO) Veni Grant 722.017.001 and NWO Start-Up Grant 40.018.016. D.J.S. acknowledges funding from the Dutch Research Council: NWO TOP Grant 714.018.003.

Supporting Information

Appendix (PDF)

References

1
C. Thomas, R. Tampé, Structural and mechanistic principles of ABC transporters. Annu. Rev. Biochem. 89, 605–636 (2020).
2
S. Rempel, W. K. Stanek, D. J. Slotboom, ECF-type ATP-binding cassette transporters. Annu. Rev. Biochem. 88, 551–576 (2019).
3
I. Setyawati et al., In vitro reconstitution of dynamically interacting integral membrane subunits of energy-coupling factor transporters. eLife 9, e64389 (2020).
4
L. J. Y. M. Swier, A. Guskov, D. J. Slotboom, Structural insight in the toppling mechanism of an energy-coupling factor transporter. Nat. Commun. 7, 11072 (2016).
5
I. Faustino et al., Membrane mediated toppling mechanism of the folate energy coupling factor transporter. Nat. Commun. 11, 1763 (2020).
6
J. A. Santos et al., Functional and structural characterization of an ECF-type ABC transporter for vitamin B12. eLife 7, e35828 (2018).
7
T. Wang et al., Structure of a bacterial energy-coupling factor transporter. Nature 497, 272–276 (2013).
8
K. Xu et al., Crystal structure of a folate energy-coupling factor transporter from Lactobacillus brevis. Nature 497, 268–271 (2013).
9
M. Zhang et al., Structure of a pantothenate transporter and implications for ECF module sharing and energy coupling of group II ECF transporters. Proc. Natl. Acad. Sci. U.S.A. 111, 18560–18565 (2014).
10
V. Kalienkova et al., Stepwise activation mechanism of the scramblase nhTMEM16 revealed by cryo-EM. eLife 8, e44364 (2019).
11
M. E. Falzone et al., Structural basis of Ca2+-dependent activation and lipid transport by a TMEM16 scramblase. eLife 8, e43229 (2019).
12
V. Arkhipova, A. Guskov, D. J. Slotboom, Structural ensemble of a glutamate transporter homologue in lipid nanodisc environment. Nat. Commun. 11, 998 (2020).
13
X. Wang, O. Boudker, Large domain movements through the lipid bilayer mediate substrate release and inhibition of glutamate transporters. eLife 9, e58417 (2020).
14
T. Pleiner et al., Structural basis for membrane insertion by the human ER membrane protein complex. Science 369, 433–436 (2020).
15
M. G. Iadanza et al., Distortion of the bilayer and dynamics of the BAM complex in lipid nanodiscs. Commun. Biol. 3, 766 (2020).
16
Y. Zhang et al., Visualization of the mechanosensitive ion channel MscS under membrane tension. Nature 590, 509–514 (2021).
17
A. M. Liaci, et al., “Structure of the Human Signal Peptidase Complex Reveals the Determinants for Signal Peptide Cleavage” (2021). Available at SSRN: https://ssrn.com/abstract=3778304. Accessed 1 August 2021.
18
Y. V. Grinkova, I. G. Denisov, S. G. Sligar, Engineering extended membrane scaffold proteins for self-assembly of soluble nanoscale lipid bilayers. Protein Eng. Des. Sel. 23, 843–848 (2010).
19
J. Rheinberger, G. Oostergetel, G. P. Resch, C. Paulino, Optimized cryo-EM data-acquisition workflow by sample-thickness determination. Acta Crystallogr. D Struct. Biol. 77, 565–571 (2021).
20
G. B. Erkens et al., The structural basis of modularity in ECF-type ABC transporters. Nat. Struct. Mol. Biol. 18, 755–760 (2011).
21
R. P.-A. Berntsson et al., Structural divergence of paralogous S components from ECF-type ABC transporters. Proc. Natl. Acad. Sci. U.S.A. 109, 13990–13995 (2012).
22
T. D. Goddard et al., UCSF ChimeraX: Meeting modern challenges in visualization and analysis. Protein Sci. 27, 14–25 (2018).
23
G. von Heijne, Y. Gavel, Topogenic signals in integral membrane proteins. Eur. J. Biochem. 174, 671–678 (1988).
24
D. A. Rodionov, et al., A novel class of modular transporters for vitamins in prokaryotes. J. Bacteriol 191, 42–51 (2009).
25
M. J. Peet, R. Henderson, C. J. Russo, The energy dependence of contrast and damage in electron cryomicroscopy of biological molecules. Ultramicroscopy 203, 125–131 (2019).
26
R. Henderson, The potential and limitations of neutrons, electrons and X-rays for atomic resolution microscopy of unstained biological molecules. Q. Rev. Biophys. 28, 171–193 (1995).
27
M. A. Herzik Jr, M. Wu, G. C. Lander, Achieving better-than-3-Å resolution by single-particle cryo-EM at 200 keV. Nat. Methods 14, 1075–1078 (2017).
28
M. A. Herzik Jr, M. Wu, G. C. Lander, High-resolution structure determination of sub-100 kDa complexes using conventional cryo-EM. Nat. Commun. 10, 1032 (2019).
29
M. Wu, G. C. Lander, M. A. Herzik Jr, Sub-2 Angstrom resolution structure determination using single-particle cryo-EM at 200 keV. J Struct Biol X 4, 100020 (2020).
30
F. Hamdi et al., 2.7 Å cryo-EM structure of vitrified M. musculus H-chain apoferritin from a compact 200 keV cryo-microscope. PLoS One 15, e0232540 (2020).
31
A. Merk et al., 1.8 Å resolution structure of β-galactosidase with a 200 kV CRYO ARM electron microscope. IUCrJ 7, 639–643 (2020).
32
B. J. Greber et al., The cryoelectron microscopy structure of the human CDK-activating kinase. Proc. Natl. Acad. Sci. U.S.A. 117, 22849–22857 (2020).
33
B. J. Greber, J. Remis, S. Ali, E. Nogales, 2.5 Å-resolution structure of human CDK-activating kinase bound to the clinical inhibitor ICEC0942. Biophys. J. 120, 677–686 (2021).
34
A. Jochim et al., An ECF-type transporter scavenges heme to overcome iron-limitation in Staphylococcus lugdunensis. eLife 9, e57322 (2020).
35
N. Chatterjee et al., A novel heme transporter from the energy coupling factor family is vital for Group A Streptococcus colonization and infections. J. Bacteriol. 202, e00205–e00220 (2020).
36
E. Verplaetse et al., Heme uptake in Lactobacillus sakei evidenced by a new energy coupling factor (ECF)-like transport system. Appl. Environ. Microbiol. 86, e02847-19 (2020).
37
S. Bousis, I. Setyawati, E. Diamanti, D. J. Slotboom, A. K. H. Hirsch, Energy-coupling factor transporters as novel antimicrobial targets. Adv. Ther. (Weinh.) 2, 1800066 (2019).
38
B. Zhu, et al., Genome-wide identification of Streptococcus sanguinis fitness genes in human serum and discovery of potential selective drug targets. Mol. Microbiol. 115, 658–671 (2021).
39
T. K. Ritchie et al., “Reconstitution of membrane proteins in phospholipid bilayer nanodiscs” in Methods in Enzymology, N. Düzgünes, Ed. (Elsevier, 2009), pp. 211–231.
40
N. Biyani et al., Focus: The interface between data collection and data processing in cryo-EM. J. Struct. Biol. 198, 124–133 (2017).
41
S. Q. Zheng et al., MotionCor2: Anisotropic correction of beam-induced motion for improved cryo-electron microscopy. Nat. Methods 14, 331–332 (2017).
42
A. Rohou, N. Grigorieff, CTFFIND4: Fast and accurate defocus estimation from electron micrographs. J. Struct. Biol. 192, 216–221 (2015).
43
T. Wagner et al., SPHIRE-crYOLO is a fast and accurate fully automated particle picker for cryo-EM. Commun. Biol. 2, 218 (2019).
44
J. Zivanov et al., New tools for automated high-resolution cryo-EM structure determination in RELION-3. eLife 7, e42166 (2018).
45
J. Zivanov, T. Nakane, S. H. W. Scheres, A Bayesian approach to beam-induced motion correction in cryo-EM single-particle analysis. IUCrJ 6, 5–17 (2019).
46
P. B. Rosenthal, R. Henderson, Optimal determination of particle orientation, absolute hand, and contrast loss in single-particle electron cryomicroscopy. J. Mol. Biol. 333, 721–745 (2003).
47
S. H. Scheres, S. Chen, Prevention of overfitting in cryo-EM structure determination. Nat. Methods 9, 853–854 (2012).
48
S. Chen et al., High-resolution noise substitution to measure overfitting and validate resolution in 3D structure determination by single particle electron cryomicroscopy. Ultramicroscopy 135, 24–35 (2013).
49
Y. Z. Tan et al., Addressing preferred specimen orientation in single-particle cryo-EM through tilting. Nat. Methods 14, 793–796 (2017).
50
E. Ramírez-Aportela et al., Automatic local resolution-based sharpening of cryo-EM maps. Bioinformatics 36, 765–772 (2020).
51
J. L. Vilas et al., MonoRes: Automatic and accurate estimation of local resolution for electron microscopy maps. Structure 26, 337–344.e4 (2018).
52
J. M. de la Rosa-Trevín et al., Scipion: A software framework toward integration, reproducibility and validation in 3D electron microscopy. J. Struct. Biol. 195, 93–99 (2016).
53
E. F. Pettersen et al., UCSF Chimera – a visualization system for exploratory research and analysis. J. Comput. Chem. 25, 1605–1612 (2004).
54
P. Emsley, K. Cowtan, Coot: Model-building tools for molecular graphics. Acta Crystallogr. D Biol. Crystallogr. 60, 2126–2132 (2004).
55
D. Liebschner et al., Macromolecular structure determination using X-rays, neutrons and electrons: Recent developments in Phenix. Acta Crystallogr. D Struct. Biol. 75, 861–877 (2019).
56
A. Morin et al., Collaboration gets the most out of software. eLife 2, e01456 (2013).
57
J. Zivanov, T. Nakane, S. H. W. Scheres, Estimation of high-order aberrations and anisotropic magnification from cryo-EM data sets in RELION-3.1. IUCrJ 7, 253–267 (2020).
58
C. J. Williams et al., MolProbity: More and better reference data for improved all-atom structure validation. Protein Sci. 27, 293–315 (2018).

Information & Authors

Information

Published in

Go to Proceedings of the National Academy of Sciences
Go to Proceedings of the National Academy of Sciences
Proceedings of the National Academy of Sciences
Vol. 118 | No. 34
August 24, 2021
PubMed: 34408021

Classifications

Data Availability

Data supporting the findings of this manuscript are available from the corresponding authors upon request. Cryo-EM density maps, half maps, and masks of ECF-FolT2 in DDM micelles and MSP2N2 lipid nanodiscs have been deposited in the Electron Microscopy Data Bank (EMDB) under Accession Nos. EMD-12483 and EMD-12484, respectively. The respective atomic models are available through the PDB under the accession codes 7NNT (ECF-FolT2 in DDM micelles) and 7NNU (ECF-FolT2 in MSP2N2 lipid nanodiscs).

Submission history

Published online: August 18, 2021
Published in issue: August 24, 2021

Keywords

  1. membrane transport
  2. ABC transporter
  3. lipid bilayer deformation
  4. cryo-EM

Acknowledgments

We thank A. A. Garaeva for freezing grids of ECF-FolT2 in detergent micelles, G. T. Oostergetel for help with initial cryo-EM data acquisition, and M. Punter for maintenance of the image processing cluster. As part of the COFUND project oLife, C.T. acknowledges funding from the European Union’s Horizon 2020 research and innovation program under the Grant Agreement 847675. C.P. acknowledges funding from the Dutch Research Council: Nederlandse Organisatie voor Wetenschappelijk Onderzoek (NWO) Veni Grant 722.017.001 and NWO Start-Up Grant 40.018.016. D.J.S. acknowledges funding from the Dutch Research Council: NWO TOP Grant 714.018.003.

Notes

This article is a PNAS Direct Submission.

Authors

Affiliations

Chancievan Thangaratnarajah https://orcid.org/0000-0002-0279-6642
Faculty of Science and Engineering, Groningen Biomolecular Sciences and Biotechnology, Membrane Enzymology Group, University of Groningen, 9747 AG, Groningen, Netherlands;
Faculty of Science and Engineering, Groningen Biomolecular Sciences and Biotechnology, Electron Microscopy Group, University of Groningen, 9747 AG, Groningen, Netherlands
Faculty of Science and Engineering, Groningen Biomolecular Sciences and Biotechnology, Membrane Enzymology Group, University of Groningen, 9747 AG, Groningen, Netherlands;
Faculty of Science and Engineering, Groningen Biomolecular Sciences and Biotechnology, Electron Microscopy Group, University of Groningen, 9747 AG, Groningen, Netherlands
Faculty of Science and Engineering, Groningen Biomolecular Sciences and Biotechnology, Membrane Enzymology Group, University of Groningen, 9747 AG, Groningen, Netherlands;
Faculty of Science and Engineering, Groningen Biomolecular Sciences and Biotechnology, Electron Microscopy Group, University of Groningen, 9747 AG, Groningen, Netherlands
Faculty of Science and Engineering, Groningen Biomolecular Sciences and Biotechnology, Membrane Enzymology Group, University of Groningen, 9747 AG, Groningen, Netherlands;

Notes

1
To whom correspondence may be addressed. Email: [email protected] or [email protected].
Author contributions: C.T., C.P., and D.J.S. designed research; C.T. and J.R. performed research; C.T., J.R., C.P., and D.J.S. analyzed data; and C.T., C.P., and D.J.S. wrote the paper.

Competing Interests

The authors declare no competing interest.

Metrics & Citations

Metrics

Note: The article usage is presented with a three- to four-day delay and will update daily once available. Due to ths delay, usage data will not appear immediately following publication. Citation information is sourced from Crossref Cited-by service.


Citation statements




Altmetrics

Citations

If you have the appropriate software installed, you can download article citation data to the citation manager of your choice. Simply select your manager software from the list below and click Download.

Cited by

    Loading...

    View Options

    View options

    PDF format

    Download this article as a PDF file

    DOWNLOAD PDF

    Get Access

    Login options

    Check if you have access through your login credentials or your institution to get full access on this article.

    Personal login Institutional Login

    Recommend to a librarian

    Recommend PNAS to a Librarian

    Purchase options

    Purchase this article to get full access to it.

    Single Article Purchase

    Insights into the bilayer-mediated toppling mechanism of a folate-specific ECF transporter by cryo-EM
    Proceedings of the National Academy of Sciences
    • Vol. 118
    • No. 34

    Media

    Figures

    Tables

    Other

    Share

    Share

    Share article link

    Share on social media