1. Structural Biology and Molecular Biophysics
Download icon

Gating and selectivity mechanisms for the lysosomal K+ channel TMEM175

  1. SeCheol Oh
  2. Navid Paknejad
  3. Richard K Hite  Is a corresponding author
  1. Structural Biology Program, Memorial Sloan Kettering Cancer Center, United States
Research Article
  • Cited 0
  • Views 2,568
  • Annotations
Cite this article as: eLife 2020;9:e53430 doi: 10.7554/eLife.53430

Abstract

Transmembrane protein 175 (TMEM175) is a K+-selective ion channel expressed in lysosomal membranes, where it establishes a membrane potential essential for lysosomal function and its dysregulation is associated with the development of Parkinson’s Disease. TMEM175 is evolutionarily distinct from all known channels, predicting novel ion-selectivity and gating mechanisms. Here we present cryo-EM structures of human TMEM175 in open and closed conformations, enabled by resolutions up to 2.6 Å. Human TMEM175 adopts a homodimeric architecture with a central ion-conduction pore lined by the side chains of the pore-lining helices. Conserved isoleucine residues in the center of the pore serve as the gate in the closed conformation. In the widened channel in the open conformation, these same residues establish a constriction essential for K+ selectivity. These studies reveal the mechanisms of permeation, selectivity and gating and lay the groundwork for understanding the role of TMEM175 in lysosomal function.

Introduction

Lysosomes are small, acidic organelles that play essential roles in nutrient sensing, signaling, autophagy and degradation of macromolecules (Ballabio and Bonifacino, 2020; Condon and Sabatini, 2019; Lawrence and Zoncu, 2019; Platt et al., 2018). Many of these processes are intricately linked to ion transport across the membrane via numerous channels and transporters and defects in lysosomal transport proteins lead to a variety of diseases (Grimm et al., 2017; Li et al., 2019; van Veen et al., 2020). Transmembrane protein 175 (TMEM175) was recently identified as a constitutively-active potassium (K+) selective channel expressed in lysosomal membranes responsible for establishing a membrane potential across the lysosomal membrane (Cang et al., 2015). TMEM175 is implicated in cellular proteostasis and its mutation is associated with the development of Parkinson’s disease through a yet unknown mechanism, underscoring the importance of this channel to regulation of cellular homeostasis (Blauwendraat et al., 2019; Cang et al., 2015; Iwaki et al., 2019; Jinn et al., 2019; Jinn et al., 2017; Krohn et al., 2020; Nalls et al., 2014).

TMEM175 is evolutionarily distinct from known K+ channels, displaying a unique membrane topology as well as lacking the conserved TVGYG selectivity filter present in canonical K+ channels such as shaker K+, BK and KcsA (Doyle, 1998; Long et al., 2005; Tao et al., 2017). In accordance with its divergent sequence, TMEM175 differs from canonical K+ channels in its ion-permeation properties and its pharmaceutical sensitivities. Canonical K+ channels are strongly selective for K+ over Na+, but are blocked by Cs+. In contrast, TMEM175 permeates cations according to a lyotropic sequence, with Ca2+ being least permeable, followed by Na+, K+ and Cs+ being most permeable (Cang et al., 2015). Only one of the common K+ channel inhibitors, 4-aminopyridine, can inhibit TMEM175 activity, while others, such as tetraethylammonium, do not alter channel activity (Cang et al., 2015). Together, these features predict unique ion permeation and selectivity mechanisms for TMEM175.

Structures of TMEM175 homologs from the prokaryotes Chamaesiphon minutus (Lee et al., 2017) and Marivirga tractuosa (Brunner et al., 2018) revealed that prokaryotic TMEM175 channels adopt a homotetrameric architecture with each protomer comprising a single 6-TM domain surrounding a central ion-conduction pathway. However, differences in the structures of the ion-conduction pathways led to differing proposals for how TMEM175 channels discriminate between cations, and thus ion selectivity in prokaryotic TMEM175 remains an open question. Moreover, unlike mammalian TMEM175 channels, prokaryotic TMEM175 channels are only minimally selective for K+, permeating only 2 to 4 K+ ions for every Na+ ion compared to 36 for human TMEM175 (Brunner et al., 2018; Cang et al., 2015). Due to the differences in ion selectivity, it remains unknown whether the mechanisms proposed to govern prokaryotic TMEM175 channels are relevant to K+ selectivity in mammalian TMEM175 channels. To elucidate the mechanisms underlying TMEM175 function in mammalian cells, we determined single-particle cryo-electron microscopic (cryo-EM) structures and analyzed the ion-permeation and selectivity properties of human TMEM175 (hTMEM175).

Results

Human TMEM175 is highly selective for K+

To measure the ion selectivity of recombinant hTMEM175 channels, we took advantage of the observation that while hTMEM175 is endogenously expressed in the membranes of endosomes and lysosomes, transient overexpression as a GFP-fusion protein in HEK293T cells leads to expression of hTMEM175 at the plasma membrane (Figure 1—figure supplement 1Lee et al., 2017). The permeation properties of these plasma membrane-localized hTMEM175 channels can be analyzed using whole-cell patch clamp. In a bi-ionic condition, in which the pipette (intracellular) solution contains 150 mM K+ and the bath (extracellular) solution contains 150 mM Na+, hTMEM175 displays a strong preference for K+ (Figure 1A). The reversal potential calculated from voltage families stepping from −100 mV to +100 mV was −55 ± 2.7 mV, corresponding to a K+/Na+ permeation ratio (PK/PNa) of ~9 (Figure 1B). Consistent with previous results (Cang et al., 2015), hTMEM175 is also selective for Cs+ over Na+. In a Cs+/Na+ bi-ionic condition, the mean reversal potential of hTMEM175 is −65 ± 4.1 mV (PCs/PNa of ~13) (Figure 1C–D). We also measured whole-cell currents from non-transfected HEK293T cells, which revealed the presence of non-selective currents whose magnitude varied between 50 and 100 pA at +100 mV (Figure 1—figure supplement 1). Because these endogenous currents are also present in the hTMEM175 transfected cells, the ion-selectivity measurements determined using whole-cell patch clamp underrepresent the selectivity of hTMEM175 and the true values are likely closer to those measured in endolysosomal patch clamp (PK/PNa ~36) (Cang et al., 2015).

Figure 1 with 1 supplement see all
hTMEM175 is a K+ selective channel.

(A, C) Representative whole-cell electrical recordings of hTMEM175-transfected HEK293T cells. In bi-ionic conditions of 150 mM K+ (intracellular) and 150 mM Na+ (extracellular) (A) or 150 mM Cs+ (intracellular) and 150 mM Na+ (extracellular) (C), currents were measured using the following protocol (red): from a holding potential of 0 mV, the voltage was stepped to voltages between −100 and +100 mV, in 20 mV increments, then returned to 0 mV. (B, D) Normalized current-voltage relationships of three independent whole-cell patch clamp recordings of hTMEM175-transfected HEK293T cells in bi-ionic conditions of 150 mM K+ (intracellular) and 150 mM Na+ (extracellular) (B) or 150 mM Cs+ (intracellular) and 150 mM Na+ (extracellular) (D). (E) K+ efflux from purified hTMEM175 reconstituted into liposomes in the presence or absence of 1 mM 4-aminopyridine and from empty liposomes was monitored using a fluorescence-based flux assay. Arrows mark addition of the proton ionophore CCCP to initiate K+ flux and addition of the K+ ionophore valinomycin to measure total flux capacity of the liposomes. All experiments were performed in triplicate and error bars represent SEM.

We next overexpressed hTMEM175 in HEK293S GnTi- cells and purified it to homogeneity (Figure 2—figure supplement 1). To assess the activity of the purified channels, we reconstituted hTMEM175 into proteoliposomes composed of a 3:1 ratio of 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphoethanolamine (POPE) and 1-palmitoyl-2-oleoyl-sn-glycero-3-phospho-(1'-rac-glycerol) (POPG) and measured channel activity. Using a 9-amino-6-chloro-2-methoxyacridine (ACMA)-based flux assay (Su et al., 2016) with 300 mM K+ inside of the vesicles and 300 mM Na+ outside, robust K+ efflux was detected from proteoliposomes containing hTMEM175 following the addition of the ionophore carbonyl cyanide m-chlorophenylhydrazone (CCCP) compared to empty liposomes (Figure 1E). No flux could be detected when the inhibitor 4-aminopyridine (4-AP) was added to the proteoliposomes at a concentration of 1 mM, demonstrating that reconstituted hTMEM175 channels are active and that they retain their K+ selectivity and their pharmacological sensitivity to 4-AP.

Structure of hTMEM175

To investigate the mechanisms that govern hTMEM175 function, we collected cryo-EM images of hTMEM175 purified in 150 mM K+. Three-dimensional classification revealed that two conformations were present among the imaged particles; class 1, which was resolved at a resolution of 2.6 Å, and class 2, which was resolved at a resolution of 3.0 Å (Figure 2A and Figure 2—figure supplement 1, 2 and Table 1). Due to the high degree of similarity between the two classes, we will first describe the higher-resolution class 1 structure (Figure 2B). hTMEM175 is composed of two homologous 6-helix repeat domains that share ~23% sequence identity (Figure 2C and Figure 2—figure supplement 3). Consequently, the density map revealed that while hTMEM175 is homodimeric, similarities between the two 6-helix repeat domains result in a pseudo-four-fold symmetric architecture. Inspection of the map identified features that were sufficiently well resolved to permit us to distinguish repeat I (TM1-TM6) from repeat II (TM7-TM12) and build a de novo structure of hTMEM175, comprising TM1-TM4 and TM7-TM12. The densities corresponding to peripheral helices TM5 and TM6 are too poorly resolved in the map for modelling (Figure 2—figure supplement 2). Viewed from the cytoplasm, hTMEM175 is a diamond-shaped channel measuring ~85 Å along the long axis and ~60 Å along the short axis with the ion-conduction pathway located at the center of the channel along the pseudo-four-fold axis (Figure 2B). The six helices of each repeat form distinct domains and no swapping of domains or helices is evident. Viewed from within the plane of the membrane, most of hTMEM175 is embedded within the membrane with only short loops extending out of the membrane on either side (Figure 2B).

Figure 2 with 4 supplements see all
Structure of hTMEM175.

(A) Cryo-EM density maps of class 1 (cyan) and class 2 (gold) hTMEM175 in KCl depicted from within the membrane. (B) Structure of class 1 hTMEM175 depicted from within the membrane (left) and from the cytoplasm (right). TM1-TM4 (repeat I) and TM7-TM12 (repeat II) of protomer A are shown in blue and red, respectively. Protomer B is shown in grey. Approximate width of the lipid bilayer is shown as grey bars. (C) Topology of hTMEM175. TM1-TM4 (repeat I) and TM7-TM12 (repeat II) are shown in blue and red, respectively. Unmodelled helices TM5 and TM6 are shown in black. (D) Superposition of class 1 hTMEM175 repeat I (blue) with repeat II (red).

Table 1
Cryo-EM data acquisition, reconstruction and model refinement statistics.
hTMEM175hTMEM175hTMEM175hTMEM175
Class 1 K+Class 2 K+Class 1 Cs+Class 2 Cs+
Cryo-EM acquisition and processing
EMDB accession #21603216042160521606
Magnification22,500x22,500x22,500x22,500x
Voltage (kV)300300300300
Total electron exposure (e- / Å2)61616161
Exposure time (s)8888
Defocus range (uM)-1.0 to -2.5-1.0 to -2.5-1.0 to -2.5-1.0 to -2.5
Pixel size (Å)1.0881.0881.0881.088
Symmetry imposedC2C2C2C2
Initial particles4,153,6144,153,6144,275,2194,275,219
Final particles342,34057,15294,65370,132
Resolution (masked FSC = 0.143, Å)2.643.033.173.24
Density modified CC (0.5, Å)2.673.093.123.23
Model Refinement
PDB ID6WC96WCA6WCB6WCC
Model resolution (FSC = 0.50/0.143Å)2.68 / 2.323.07 / 2.673.18 / 2.713.27 / 2.84
Model refinement resolution300-2.6300-3.0300-3.2300-3.2
RMS deviations
 Bond length (Å)0.0050.0020.0040.003
 Bond angle (°)0.5320.5070.4060.506
Ramachandran plot
 Favored (%)96.1396.4299.1799.17
 Allowed (%)3.873.580.830.83
 Disallowed (%)0000
Rotamer Outliers (%)2.271.611.292.26
Validation
 MolProbity score1.711.741.121.49
 Clashscore3.886.282.504.39

Consistent with the high sequence homology between repeat I (TM1-TM6) and repeat II (TM7-TM12), alignment reveals that their structures are nearly identical with an RMSD of 1.7 Å (Figure 2D). The structures of repeat I and repeat II are also homologous with the monomeric structures of TMEM175 channels from the prokaryotes Chamaesiphon minutus (TMEM175Cm) (Lee et al., 2017) and Marivirga tractuosa (TMEM175Mt) (Brunner et al., 2018Figure 2—figure supplement 4). Thus, while prokaryotic TMEM175 channels are homotetramers rather than homodimers, the global architecture of TMEM175 channels is conserved.

Ion-conduction pathway contains ordered ions and waters

The ion-conduction pathway of hTMEM175 is located along the central axis of the channel extending approximately 45 Å from the cytoplasm to the lysosomal lumen (Figure 3A) The pore is lined by the side chains of the kinked pore-lining helices, TM1 and TM7. The side chains of TM1 and TM7 create multiple constrictions whose radii are less than 2.0 Å and would restrict the permeation of hydrated K+ ions. The narrowest of these constrictions is formed by the side chains of Ile46 from TM1 and Ile271 from TM7, which are conserved among all eukaryotic TMEM175 channels, and has minimum radius of 1.7 Å (Figure 3B). Just below the constriction formed by Ile46 and Ile271, the pore contains an expanded vestibule that is also lined by the side chains of Thr49 and Leu53 from TM1 and Thr274, Leu275 and Leu278, making it much more hydrophobic than the rest of the pore (Figure 3A). Despite its hydrophobicity, multiple non-protein density peaks are resolved within the vestibule (Figure 3—figure supplement 1). Non-protein density peaks are also present in the vestibule in a density map calculated without symmetry, indicating that they represent ordered molecules rather than arising from the accumulation of noise along the two-fold symmetry axis during image processing (Figure 3—figure supplement 1).

Figure 3 with 2 supplements see all
hTMEM175 ion conduction pathway.

(A) Ion permeation pathway of class 1 hTMEM175. Pore-lining helices TM1 from protomers A (blue) and B (grey) are shown at left and TM7 from protomers A (red) and B (grey) are shown at right with all other helices removed for clarity. Pore-lining residues are shown as sticks. Surface representation of the ion permeation pathway colored by hydrophobicity calculated using the class 1 structure without ions and water molecules calculated using CHAP (Klesse et al., 2019). (B) Dimensions of the ion conduction pathway in class 1 calculated using CHAP (Klesse et al., 2019). (C) Overlapping non-protein density peaks in the ion permeation pathway of class 1 in the presence of K+ (blue mesh, 12 σ threshold) and Cs+ (gold mesh, 8 σ threshold). hTMEM175 is shown as in A. K+ ions are shown as violet spheres. (D) Density map near the isoleucine constriction displayed as blue mesh and contoured at 12 σ threshold. K+ ions are shown as violet spheres and water molecules are shown as red spheres.

In addition to the non-protein densities resolved in the hydrophobic vestibule, numerous other non-protein densities are resolved in the other, more hydrophilic regions of the pore (Figure 3—figure supplement 1). However, due to the large number of non-protein peaks and lack of obvious protein-coordinated ion-binding sites, it was not possible to unambiguously distinguish ions from water molecules based on the density map alone. To aid in assigning the identity of these peaks, we collected cryo-EM images of hTMEM175 purified in 150 mM Cs+. We chose to determine structures in the presence of Cs+ for two reasons. First, Cs+ scatters electrons approximately three times more strongly than K+ (Peng, 1998) and thus bound Cs+ ions should yield density peaks that can be distinguished from those corresponding to water and other non-protein atoms in the density map. Second, because the permeation of Cs+ is similar to K+, we hypothesized that Cs+ would occupy the same binding sites in the pore as does K+ and thus facilitate identification of the ion-binding sites (Figure 1).

Three-dimensional classification revealed that hTMEM175 adopts the same two conformations in the presence of Cs+ that exist in the presence of K+ (Figure 3—figure supplement 2 and Table 1). A number of non-protein densities are resolved in the pore of class 1 hTMEM175 in the presence of Cs+. Because of the similarities between the class 1 structures determined in Cs+ and K+ (all-atom RMSD = 0.2 Å), we could directly compare the density maps at a threshold of σ = 12 for the class 1 map determined in K+ and of σ = 8 for the class 1 map determined in Cs+, identifying four overlapping non-protein density peaks in both maps (Figure 3C). We therefore modelled these overlapping densities as ions, either K+ or Cs+ depending on the condition. We assigned the remaining non-protein densities as ordered waters. Despite the narrow dimensions of the pore, the protein only minimally participates in the direct coordination of the bound ions (Figure 3D). Instead, ordered water molecules form nearly all of the direct interactions with the bound ions. Near the cytoplasmic entrance to the pore, the ion in the K1 site is directly coordinated by the side chain of Ser38 and indirectly coordinated by the side chain of Asp266 and the backbone oxygens of Ser38 and Glu259 via ordered waters (Figure 4A). The ion in the K2 site is indirectly coordinated by the side chain of Ser45 and the backbone oxygens of Ala263 and Gly267 via ordered waters (Figure 4B). The K3 site, the weakest of the four sites in both density maps, is located near the isoleucine constriction that is flanked by two layers of four water molecules (Figure 4C). The waters on the cytoplasmic side are coordinated by the side chain of Ser45 and the backbone oxygen of Gly267, while the waters on the luminal side are indirectly coordinated by the side chains of Thr49 and Thr274 and by the backbone oxygens of Ile46 and Ile271. Unlike the spherical densities resolved at the other ion-binding sites, the density for the K3 ion is elongated and extends between the two layers of waters on either side of the isoleucine constriction (Figure 3D). The waters on the luminal side of the constriction are located between 3.1 and 3.4 Å away from the center of the luminal portion of the density peak, while the waters on the cytoplasmic side are located between 3.1 and 3.2 Å away from the center of the cytoplasmic portion of the density peak. Because the center of the constriction is hydrophobic, ions are unlikely to stably bind there, and so we speculate that the K3 density corresponds to the sum of two partially-occupied ion-binding sites positioned on either the cytoplasmic or the luminal side of the constriction where the ions can be directly coordinated by the nearby waters. The fourth ion-binding site (K4) is located in the hydrophobic vestibule near the luminal entrance to the pore (Figure 4D). Notably, two symmetry-related copies of sites K1, K2 and K4 are present in the structure because the K1, K2 and K4 sites are located slightly off of the central axis of the pore (Figure 4). However, the distances between the pairs of K1, K2 and K4 ion-binding sites are 3.0 Å, 2.9 Å and 3.3 Å, respectively, which are likely too close for both symmetry-related sites to be simultaneously occupied with ions.

Ion-binding sites in hTMEM175.

Structure of the K1 (A), K2 (B), K3 (C) and K4 (D) binding sites in class 1. K+ ions are shown as violet spheres and water molecules are shown as red spheres. Density for K+ and water molecules shown as blue mesh and contoured at 12 σ threshold.

Structural heterogeneity reveals gating mechanism

To better understand the functional states of the two conformations resolved in our data sets, we next superimposed the class 1 and class 2 structures determined in the presence of K+ (Figure 5A). Overall, the two classes are very similar, with an all-atom RMSD of 0.9 Å. The similarities are especially pronounced in the cytoplasmic side of the channel, which likely arises from the existence of interaction networks at the intra-subunit interfaces between repeats I and II and at the inter-subunit interfaces between protomers (Figure 5B). These interaction networks adopt identical configurations in both conformations and are anchored by the essential RxxxFSD motif on TM1 and TM7 (Cang et al., 2015). In addition to the RxxxFSD motif, the networks involve a conserved histidine and a conserved tryptophan on TM2/TM8 and a conserved asparagine on TM3/TM9 (Figure 2—figure supplement 3). Similar interaction networks were resolved in the prokaryotic structures (Brunner et al., 2018; Lee et al., 2017), suggesting a conserved role for the RxxxFSD motif in maintaining channel quaternary structure during conformational changes.

Figure 5 with 1 supplement see all
Gating in hTMEM175.

(A) Superposition of class 1 (cyan) and class 2 (gold) viewed from within the membrane (left) and from the lysosomal lumen (right). (B) Alignment of the RxxxFSD inter- and intra-subunit interaction networks in class 1 (cyan) and class 2 (gold) depicted as sticks and viewed from the cytosol. Ionic and polar interaction are shown as dashed lines. (C) Ion conduction pathways of class 1 (cyan) and class 2 (gold). TM1 is shown at left and TM7 is shown at right with all other helices removed for clarity. K+ ion binding sites are shown as spheres. Dotted lines correspond to minimum distance between opposing residues at the isoleucine constriction.

In contrast to the rigid cytoplasmic side, differences between the two classes can be readily detected on the luminal side of the channel (Figure 5A and Video 1). When viewed from the luminal side of the channel, the ends of the transmembrane helices in the class 2 structure are rotated in a clockwise manner compared to their positions in class 1 (Figure 5A). The luminal loops between the transmembrane helices also adopt different conformations, with the loop between TM9 and TM10 undergoing the largest change. In class 2, the last turn of TM9 is unwound and moves nearly 12 Å from its position adjacent to the loop between TM11 and TM12 in class 1 to interact with the loop between TM1 and TM2.

Video 1
Morph between open class 1 and closed class 2.

Inspection of the pore-lining helices, TM1 and TM7, reveals that the clockwise rotation of their luminal ends from class 1 to class 2 is accompanied by the adoption of a straighter, α-helical conformation, particularly for TM7 (Figure 5C). In class 1, the kink in TM1 is stabilized by Pro54 and by a water molecule coordinated by the side chain of Thr84 on TM2, the backbone carbonyl oxygen of Ala48 and the backbone amide nitrogen of Met51, while the kink in TM7 is stabilized by a water molecule coordinated by the side chain of Ser316 on TM8, the backbone carbonyl oxygens of Val272 and Ala273 and the backbone amide nitrogen of Leu276 (Figure 5—figure supplement 1). In class 2, the waters stabilizing the kinks are displaced by the side chains of Met51 and Leu276. The resultant straightening of TM1 and TM7 in class 2 alters the shape of the pore, particularly at the isoleucine constriction (Figure 5C). In class 1, the minimum radius of the isoleucine constriction is 1.7 Å with an ion-binding site in the center surrounded by the isoleucine side chains in a nearly four-fold symmetric configuration. In class 2, the four-fold arrangement of the isoleucine side chains is broken by an inward movement of all four isoleucine residues and a rotation of the Ile271 side chains. These changes reduce the minimum pore radius to 0.5 Å, which is too narrow to accommodate dehydrated K+ ions. Accordingly, no density is resolved in the K3 ion-binding site in class 2. Thus, class 2 represents a closed conformation with Ile46 and Ile271 forming the channel gate.

If class 2 represents a closed conformation, what state does class 1 represent? The proteoliposome flux assay demonstrated that purified hTMEM175 can conduct ions in the absence of stimuli, indicating that hTMEM175 can adopt an open state in the similar conditions used for cryo-EM analysis (Figure 1E). However, in order for class 1 to be a conductive state, ions would have to be able to permeate its narrow pore in a partially dehydrated state. In the cytosolic and luminal regions of the pore, the bound ions are coordinated by numerous water molecules and the constrictions are formed by polar and charged side chains, suggesting that partially hydrated ions can readily translocate. In contrast, ions would have to be almost completely dehydrated to penetrate the isoleucine constriction because of its size and hydrophobicity. However, the dehydration need only be transient due to the layers of water molecules on either side of the isoleucine constriction that can rehydrate the ion once it passes through the constriction (Figure 6A).

Figure 6 with 1 supplement see all
Permeation and selectivity through the isoleucine constriction.

(A) The isoleucine constriction is flanked by two layers of ordered water molecules. The cytosolic layer of waters is partially coordinated by Ser45, while the luminal layer is partially coordinated by Thr49 and Thr274. (B) Mean current recorded from HEK293T cells transfected with hTMEM175 (blue), S45A (red dashed), S45T (red), T274V (orange dashed), T274S (orange), I46M (green), I46M/I271M (magenta) and non-transfected (white) at +100 mV in a bi-ionic condition of 150 mM Cs+ (intracellular) and 150 mM Na+ (extracellular). (C) Normalized I-V relationship of whole-cell patch clamp of hTMEM175 transfected (blue), S45T transfected (red), T274S transfected (orange) and I46M transfected (green) HEK293T cells in a bi-ionic condition of 150 mM Cs+ (intracellular) and 150 mM Na+ (extracellular). All experiments were performed at least three times and error bars represent SEM. (D) Model for ion selectivity and gating in hTMEM175. In the open state, ions are transiently dehydrated through the isoleucine constriction, favoring permeation of K+ ions. In the closed state, the isoleucine constriction closes, preventing ion permeation.

To determine if the ordered water molecules facilitate ion permeation through the isoleucine constriction, we mutated Ser45 or Thr274, conserved residues on the cytosolic side and luminal side of the isoleucine constriction, respectively, whose side-chain hydroxyl groups participate in the coordination of ordered waters (Figure 6A). We first analyzed the effects of the mutations on protein stability using fluorescence size-exclusion chromatography (Goehring et al., 2014), which revealed that channels with mutations to Ser45 and Thr274 are properly folded as dimers and express at levels within two-fold of wild-type hTMEM175 (Figure 6—figure supplement 1). We next assessed the effects of the mutations on channel activity using whole-cell patch clamp in a bi-ionic Cs+/Na+ condition. Currents recorded from cells expressing the S45A and T274V mutants, which lack one of the hydroxyl groups involved in water coordination, were indistinguishable from those recorded from non-transfected cells (Figure 6B). In contrast, cells expressing either the S45T or the T274S mutant yielded Cs+-selective currents (Figure 6B–C). These results reveal a critical role for residues that coordinate water molecules in facilitating the permeation of ions through the isoleucine constriction and suggest that the water molecules themselves may participate in ion permeation. We therefore speculate that ions can permeate through the isoleucine constriction of class 1 in a partially hydrated state and that class 1 represents a conductive state.

Mechanisms of ion selectivity

The transient dehydration of ions through the isoleucine constriction implies a mechanism for ion selectivity. Because the enthalpies of dehydration for Cs+ (250 kJ/mol) and K+ (295 kJ/mol) are lower than that of Na+ (365 kJ/mol) (Marcus, 1991), Cs+ and K+ ions can more readily access the partially dehydrated state necessary to permeate through the isoleucine constriction and are thus permeated more efficiently than Na+. Previously, mutation of Ile46 and Ile271 to asparagine was shown to diminish ion selectivity, which led to the proposal that the Ile46 and Ile271 act as a hydrophobic selectivity filter (Lee et al., 2017). However, in the class 1 structure of hTMEM175 the branched side chains of Ile46 and Ile271 form a constriction that we propose is precisely shaped to allow dehydrated ions to permeate, suggesting that the unique shape of isoleucine may also be essential (Figure 3D). To test if hydrophobicity itself is sufficient to impart ion selectivity, we mutated Ile46 to methionine, which is similar to isoleucine in terms of volume occupied and hydrophobicity, and recorded whole-cell currents. In a bi-ionic Cs+/Na+ condition, the I46M mutant displayed less selectivity for Cs+ with a mean reversal potential of −21 ± 1.5 mV compared to −65 ± 4.1 mV for wild-type hTMEM175 (Figure 6C). We also attempted to record whole-cell currents from cells expressing the I46M/I271M double mutant. However, no exogenous currents could be detected, consistent with a previous report that mutations of Ile46 and Ile271 to alanine, valine, leucine and phenylalanine were not functional (Lee et al., 2017). These results point to a unique role for isoleucine side chains in establishing a selectivity filter that cannot be duplicated by other amino acids.

Although the isoleucine constriction participates in the selective permeation of ions by hTMEM175, it is unclear if the energetic differences between K+ and Na+ dehydration are alone sufficient to generate a nearly forty-fold selectivity for K+ over Na+. Moreover, alignment of hTMEM175 with prokaryotic homologs reveal that isoleucine residues are broadly conserved at the position of the isoleucine constrictions (Brunner et al., 2018; Lee et al., 2017Figure 2—figure supplement 3), suggesting that the isoleucine constriction may be a conserved feature among TMEM175 channels. Given the greater selectivity of hTMEM175 compared to its prokaryotic homologs (Brunner et al., 2018; Lee et al., 2017), additional, unknown mechanisms may further contribute to K+ selectivity in hTMEM175. Because the ion-binding sites are largely formed by waters, we hypothesized that ion coordination by waters might represent an additional mechanism of selectivity in hTMEM175. Comparing the structures of the water-mediated ion-binding sites with the structure of the ion-binding sites in KcsA revealed that the distances between water molecules in the layers on either side of the isoleucine constriction in hTMEM175 are 4.8 Å and 5.1 Å, similar to the distances between opposing backbone carbonyl oxygens that form the ion-binding sites in the TVGYG selectivity filter of KcsA (4.5–5.1 Å) (Zhou et al., 2001Figure 6A). The similarity of these structures suggests that the water molecules in hTMEM175 may play an analogous role in hTMEM175 and participate in ion selectivity. To examine the influence of the ordered water molecules near the isoleucine constriction on ion selectivity, we compared the reversal potential of S45T and T274S mutants with wild-type hTMEM175. While the reversal potential of the S45T was similar to wild type (−62 ± 1.9 mV), it was right-shifted to −34 ± 5.1 mV for the T274S mutant, suggesting that the T274S mutant is less selective than wild-type hTMEM175 (Figure 6B–C). These results suggest that minor differences in residues that coordinate water molecules near the isoleucine constriction can influence ion selectivity. Because Thr274 does not directly interact with the permeating ions, these results suggest that the network of ordered molecules in the pore may also contribute to ion selectivity in hTMEM175 channels.

Discussion

To gain insights into the mechanisms underlying TMEM175 channel function, we determined cryo-EM structures of hTMEM175 in two conformations. Numerous non-protein densities are resolved in the ion conduction pathways of both structures. However, unlike anomalous X-ray diffraction experiments, which can specifically localize atoms in protein structures, ligand identification in cryo-EM structures relies on examination of the local environment of the binding sites. While such analyses can be straightforward for well-characterized proteins, densities can be ambiguous in structures of less well-characterized proteins. Because TMEM175 channels are evolutionarily distinct from all other known ion channels, we needed to an alternative approach to distinguish between the ion densities in the pore and ordered waters. Therefore, we determined structures of hTMEM175 in the presence of two ions, K+ and Cs+, which are permeated similarly by hTMEM175, but have different electron scattering properties (Figure 1). The comparison revealed four overlapping peaks in the class 1 maps and three in the class 2 maps (Figure 3). While this approach enabled us to better identify ion densities and may be useful for identifying ion-binding sites in other proteins, the signal is much weaker than that generated by anomalous X-ray diffraction experiments and future efforts will be required to improve identification of ligands in cryo-EM density maps.

Identification of the ion-binding sites enabled us to tentatively assign the two conformations as an open state and a closed state. However, cryo-EM structures represent snapshots of the protein being imaged and it is difficult to explicitly assign functional states to the conformational states resolved. Because the population of imaged particles represents the equilibrium of states at the condition being imaged, it is sometimes possible to infer that the fraction of particles in a particular structural state corresponds to the likelihood of the corresponding functional state (Hite and MacKinnon, 2017). For small proteins like hTMEM175, such correspondence is difficult due to the large number of particles that are removed from the data sets during classification and thus whose conformational state is unknown (Table 1). Moreover, while care is taken during image processing to identify all conformations, it is possible that additional low-abundance conformations remain unidentified in the data sets. Thus, future orthogonal experiments will be necessary to test the functional assignments that we propose for our structures as well as the mechanisms derived from their interpretation.

In contrast to most ion channels, for which gating and selectivity are physically uncoupled, the precise three-dimensional arrangement of Ile46 and Ile271 in the two hTMEM175 structures suggests that they serve as both the gating residues as well as the ion selectivity filter. Notably, what factors influence the likelihood of hTMEM175 adopting open or closed states is an open question. Due to hTMEM175’s association with lysosomal homeostasis and the development of Parkinson’s Disease (Blauwendraat et al., 2019; Cang et al., 2015; Iwaki et al., 2019; Jinn et al., 2019; Jinn et al., 2017; Nalls et al., 2014), future studies will be necessary to identify stimuli that regulate gating and to understand how these stimuli alter the equilibrium between open and closed states to regulate lysosomal K+ flux.

Materials and methods

Key resources table
Reagent type
(species) or resource
DesignationSource or referenceIdentifiersAdditional information
Gene (Homo sapiens)hTMEM175Synbio technologies
Cell line (H. sapiens)HEK-293TATCCCRL-3216
RRID:CVCL_0063
Cell line (H. sapiens)HEK-293S GnTi-ATCCCRL-3022
Chemical compound, drug1-palmitoyl-2-oleoyl-sn-glycero-3-phosphoethanolamineAvanti Polar Lipids850757
Chemical compound, drug1-palmitoyl-2-oleoyl-sn-glycero-3-phospho-(1'-rac-glycerol) (sodium salt)Avanti Polar Lipids840457
Chemical compound, drugcarbonyl cyanide m-chlorophenylhydrazone (CCCP)Thermo Fisher Scientific215911250 MG
Chemical compound, drug9-amino-6-chloro-2-methoxyacridine (ACMA)Thermo Fisher ScientificA1324
Chemical compound, drugvalinomycinSigmaV0627
Chemical compound, drugPolyethylenimine, Linear, MW 25000, Transfection Grade (PEI 25K)Polysciences, Inc23966–1
Chemical compound, drugSodium ButyrateSigma8451440100
Chemical compound, druglauryl maltoside neopentyl glycolAnatraceNG310
Chemical compound, drugn-Octyl-β-D-MaltopyranosideAnatraceO310S
Software, algorithmMotionCor2Zheng et al., 2017RRID:SCR_016499
Software, algorithmCtfFind 4.1.10Rohou and Grigorieff, 2015RRID:SCR_016731
Software, algorithmRELION 3.1Scheres, 2016http://www2.mrc-lmb.cam.ac.uk/relion
RRID:SCR_016274
Software, algorithmSerialEMMastronarde, 2005RRID:SCR_017293
Software, algorithmcryoSPARC v2Structura Biotechnologyhttps://cryosparc.com/
RRID:SCR_016501
Software, algorithmPHENIXLiebschner et al., 2019https://www.phenix-online.org/
RRID:SCR_014224
Software, algorithmCOOTEmsley et al., 2010https://www2.mrc-lmb.cam.ac.uk/personal/pemsley/coot/
RRID:SCR_014222
Software, algorithmPyMOLSchrödinger, 2020https://pymol.org/2/
RRID:SCR_000305
Software, algorithmUCSF ChimeraPettersen et al., 2004https://www.cgl.ucsf.edu/chimera
RRID:SCR_004097
Software, algorithmGraphPad Prism 7GraphPad Software
Software, algorithmSoftMax Pro 6Molecular Devices
Software, algorithmAxon Digidata 1550B digitizerMolecular Devices
Software, algorithmClampex 10.6Molecular Devices
Software, algorithmCHAPKlesse et al., 2019https://www.channotation.org/
Software, algorithmClampfit 10.6Molecular Devices
OthersQUANTIFOIL R1.2/1.3 holey carbon gridsQuantifoil
OthersFEI Vitrobot Mark IVFEI Thermo Fisher

Protein expression and purification

Request a detailed protocol

The gene encoding human TMEM175 was synthesized (SynBio) and subcloned into a BacMam expression vector with a C-terminal EGFP-tag fused via a short linker containing a PreScission protease site (Goehring et al., 2014). The plasmid was mixed with PEI 25 k (Polysciences, Inc) for 30 min and then used to transfect HEK293S GnTi cells (ATCC CRL-3022). After 24 hr incubation at 37 °C, sodium butyrate was added to a final concentration of 10 mM, and cells were allowed to grow at 37 °C for an additional 48–72 hr before harvesting. Cell pellets were washed in phosphate-buffered saline solution and flash frozen in liquid nitrogen. Expressed protein was solubilized in 2% lauryl maltose neopentyl glycol (LMNG, Anatrace), 20 mM HEPES pH 7.5, 150 mM KCl supplemented with protease-inhibitor cocktail (1 mM PMSF, 2.5 μg/mL aprotinin, 2.5 μg/mL leupeptin, 1 μg/mL pepstatin A) and DNase. Solubilized protein was separated by centrifugation 74,766 g for 40 mins, followed by binding to anti-GFP nanobody resin for 2 hr. Anti-GFP nanobody affinity chromatography was performed by 20 column volumes of washing with buffer containing 0.1% LMNG, 20 mM HEPES pH 7.5, 150 mM KCl, 2 mM DTT, followed by overnight PreScission digestion, and elution with wash buffer. Eluted protein was further purified by size exclusion chromatography on a Superdex 200 Increase 10/300 GL (GE healthcare) in SEC buffer (0.1% LMNG, 50 mM Tris pH 8.0, 150 mM KCl, 2 mM DTT). Peak fractions were pooled and concentrated to ~4 mg/mL using CORNING SPIN-X concentrators (100 kDa cutoff). For the CsCl samples, KCl was replaced with CsCl for all steps of the purification.

Proteoliposome reconstitution and flux assay

Request a detailed protocol

1-palmitoyl-2-oleoyl-sn-glycero-3-phosphoethanolamine (POPE) and 1-palmitoyl-2-oleoyl-sn-glycero-3-phospho-(1'-rac-glycerol) (POPG) in chloroform (Avanti) were mixed in a ratio 3:1 (mg:mg) and dried under argon gas. The dried lipid mixture was solubilized in pentane and dried again under argon gas to remove residual chloroform. Dried lipids were then desiccated for 2 hr under vacuum. Lipids were resuspended in 10 mM Hepes pH 7.4, 300 mM KCl to a final concentration of 10 mg/ml. Unilamellar vesicles were formed by sonication and then solubilized using 8% (w/v) octyl maltoside. Full length hTMEM175 purified in LMNG at a concentration of 1 mg/ml was mixed with the octyl maltoside-solubilized lipids and dialyzed using 25 kDa MWC bags (SpectraPor) in 10 mM Hepes pH 7.4, 300 mM KCl, 2 mM dithiothreitol (DTT) for 5 days with daily exchange of dialysis buffer. After dialysis, harvested proteoliposomes were snap frozen in liquid nitrogen and stored at −80 °C until use. Proteoliposomes were rapidly thawed at 37 °C, sonicated for 5 s, incubated at room temperature for 2–4 hr before use, and then diluted 100-fold into a flux assay buffer composed of 10 mM Hepes pH 7.4, 300 mM NaCl, 0.2 µM 9-amino-6-chloro-2-methoxyacridine (ACMA).

Data were collected on a SpectraMax M5 fluorometer (Molecular Devices) using Softmax Pro six software. ACMA excitation/emission wavelengths were 410/490 nm, respectively. Fluorescence intensity measurements were collected every 30 s. The ionophore CCCP (1 µM) and valinomycin (20 nM) were added at 150 s and 600 s, respectively.

Electron microscopy sample preparation and data acquisition

Request a detailed protocol

4–5 μl of purified hTMEM175 at a concentration of 4 mg/ml was applied to glow-discharged Au 400 mesh QUANTIFOIL R1.2/1.3 holey carbon grids (Quantifoil, and then plunged into liquid nitrogen-cooled liquid ethane with an FEI Vitrobot Mark IV (FEI Thermo Fisher). Grids were transferred to a 300 keV FEI Titan Krios microscopy equipped with a K2 summit direct electron detector (Gatan). Images were recorded with SerialEM (Mastronarde, 2005) in super-resolution mode at 22,500x, corresponding to pixel size of 0.544 Å. Dose rate was eight electrons/pixel/s, and defocus range was 1.2–2.5 µm. Images were recorded for 8 s with 0.2 s subframes (total 40 subframes), corresponding to a total dose of 61 electrons/Å2.

Electron microscopy data processing

Request a detailed protocol

40-frame super-resolution movies (0.544 Å/pixel) of TMEM175 in KCl were gain corrected, Fourier cropped by two and aligned using whole-frame and local motion correction algorithms by Motioncor2 (Zheng et al., 2017) (1.088 Å/pixel). Whole-frame CTF parameters were determined using CTFfind 4.1.10 (Rohou and Grigorieff, 2015). Approximately 500 particles were manually selected to generate initial templates for autopicking that were improved by several rounds of two-dimensional classification in Relion 3.0 (Scheres, 2016), resulting in 2,499,425 particles for KCl data set 1 and 1,654,189 particles for KCl data set 2. False-positive selections and contaminants were excluded from the data using multiple rounds of heterogeneous classification in cryoSPARC v2 (Punjani et al., 2017) using models generated from the ab initio algorithm in cryoSPARC v2, resulting in a stack of 571,468 particles. After particle polishing in Relion and local CTF estimation and higher order aberration correction in cryoSPARC v2, a consensus reconstruction was determined at resolution of 2.7 Å. 3D variability analysis in cryoSPARC v2 was then employed to characterize conformational heterogeneity, revealing two states that were subsequently used for iterative rounds of supervised heterogeneous refinement in cryoSPARC v2. The final stack for class 1 contained 342,340 particles and yielded a reconstruction with an estimated resolution of 2.6 Å by non-uniform refinement in cryoSPARC v2 (Punjani et al., 2019). The final stack for class 2 contained 70,132 particles and yielded a reconstruction with an estimated resolution of 3.0 Å by non-uniform refinement in cryoSPARC v2. The final reconstructions of class 1 and class 2 were further improved by employing density modification on the two unfiltered half-maps with a soft mask in Phenix (Terwilliger et al., 2019).

40-frame super-resolution movies (0.544 Å/pixel) of TMEM175 in CsCl were gain corrected, Fourier cropped by two and aligned using whole-frame and local motion correction algorithms by Motioncor2 (1.088 Å/pixel). Approximately 500 particles were manually selected to generate initial templates for autopicking that were improved by several rounds of two-dimensional classification in Relion and autopicking using Relion, resulting in 2,537,436 particles for CsCl data set 1 and 1,737,783 particles for CsCl data set 2. False-positive selections and contaminants were excluded through iterative rounds of heterogeneous classification in cryoSPARC v2 using models generated from the ab initio algorithm in cryoSPARC v2, resulting in a stack of 330,698 particles. After particle polishing in Relion and local CTF estimation and higher order aberration correction in cryoSPARC v2, a consensus reconstruction was determined to 3.1 Å. 3D variability analysis in cryoSPARC v2 was then employed to characterize conformational heterogeneity, revealing two states that were used for iterative rounds of seeded heterogeneous refinement in cryoSPARC v2. The final stack for class 1 contained 104,126 particles and yielded a reconstruction with an estimated resolution of 3.2 Å by non-uniform refinement in cryoSPARC v2. The final stack for class 2 contained 70,132 particles and yielded a reconstruction with an estimated resolution of 3.2 Å by non-uniform refinement in cryoSPARC v2. The final reconstructions of class 1 and class 2 were further improved by employing density modification on the two unfiltered half-maps with a soft mask that includes the detergent micelle in Phenix.

Model building and coordinate refinement

Request a detailed protocol

Poly-alanine helices were manually built into the transmembrane helices of the class 1 K+ density map using coot (Emsley et al., 2010). The helices were manually registered using large side chains and the connecting loops were manually built into the density. Densities corresponding to TM5 and TM6 (residues 174–251) were too poorly ordered and omitted from the model. The final model contains residues 30–173, and 254–476. Four ions were assigned by identifying overlapping non-protein density peaks in the class 1 K+ and Cs+ maps. Atomic coordinates were refined against the density modified map using phenix.real_space_refinement with geometric and Ramachandran restraints maintained throughout (Adams et al., 2010).

The refined class 1 structure was manually docked into the class 2 density map using chimera (Pettersen et al., 2004). The model was manually rebuilt using coot to fit the density. Three ions were assigned by identifying overlapping non-protein density peaks in the K+ and Cs+ maps. Water molecules were placed into the remaining non-protein density peaks. Atomic coordinates were refined against the density modified map using phenix.real_space_refinement with geometric and Ramachandran restraints maintained throughout (Adams et al., 2010).

The Cs+ class 1 and class 2 structures were determined by docking in the K+ class 1 and class 2 structure in Coot and manually rebuilding the protein to best fit the density map. Atomic coordinates were refined against the density modified map using phenix.real_space_refinement with geometric and Ramachandran restraints maintained throughout (Adams et al., 2010).

Fluorescence size exclusion chromatography (FSEC)

Request a detailed protocol

Plasmids encoding GFP-tagged versions of wild-type and mutant hTMEM175 were mixed with PEI 25 k (Polysciences, Inc) for 30 min and then used to transfect HEK293S GnTi cells. After 24 hr incubation at 37 °C, sodium butyrate was added to a final concentration of 10 mM, and cells were allowed to grow at 37 °C for an additional 48–72 hr before harvesting. Cell pellets were washed in phosphate-buffered saline solution and flash frozen in liquid nitrogen. Expressed protein was solubilized in 2% lauryl maltose neopentyl glycol (LMNG), 20 mM HEPES pH 7.5, 150 mM KCl supplemented with protease-inhibitor cocktail (1 mM PMSF, 2.5 μg/mL aprotinin, 2.5 μg/mL leupeptin, 1 μg/mL pepstatin A) and DNase. Solubilized protein was separated by centrifugation 21,130 g for 60 mins. Separated proteins were injected to and monitored by fluorescence size exclusion chromatography on a Superose 6 Increase 10/300 GL (GE healthcare) in SEC buffer (0.1% LMNG, 50 mM Tris pH 8.0, 150 mM KCl, 2 mM DTT). Fluorescence was monitored at 488/509 nm of excitation/emission wavelength, respectively.

Electrophysiology

Request a detailed protocol

Electrophysiological recordings of hTMEM175 constructs were performed in HEK293T cells (ATCC CRL-3216). HEK293T cells cultured in DMEM supplemented with 10% FBS were transfected with 2 μg of hTMEM175 plasmid using 6 μg of PEI 25 k (Polysciences, Inc). 24–48 hr following transfection, cells were detached by trypsin treatment. The detached cells were transferred to poly-Lys-treated 35 mm single dishes (FluoroDish, World Precision Instruments) and incubated overnight at 37 °C in fresh media. Immediately prior to recording, media was replaced with a bath solution containing 145 mM Na- methanesulfonate (MS), 5 mM NaCl or KCl, 1 mM MgCl2, 1 mM CaCl2, 10 mM HEPES/Tris pH 7.4. 10 cm long borosilicate glass were pulled and fire polished (Sutter instrument). The resistance of glass pipette was 5 ~ 8 MΩ were filled with a pipette solution containing 150 mM K-MS or Cs-MS, 5 mM MgCl2, 10 mM EGTA/Tris, 10 mM HEPES/Tris pH 7.4, GΩ seals were formed after gentle suction. The recordings were performed in whole cell patch clamp configuration using the following protocol: from a holding potential of 0 mV, the voltage was stepped to voltages between −100 and +100 mV, in 20 mV increments. The currents were recorded using Axon Digidata 1550B digitizer and Clampex 10.6 (Molecular Devices, LLC) and analyzed using Clampfit 10.6 (Molecular Devices, LLC). Each experiment was performed a unique cell and currents were normalized to the maximum current of each experiment (at +100 mV). Each condition includes cells from at least two independent transfections.

In bi-ionic conditions, the relative permeability between cations are calculated using the following equations.

PXPY= [Y+]ext[X+]inte-(ErevFRT)

Where PX and PY are the permeabilities of intracellular and extracellular cation X and Y, respectively, Erev is the measured reversal potential, F is Faradays’ constant, R is the gas constant, and T is the absolute temperature.

Figures were prepared with UCSF Chimera (Pettersen et al., 2004), PyMol (Schrödinger), CHAP (Klesse et al., 2019) and Prism 7 (GraphPad).

References

  1. 1
  2. 2
  3. 3
  4. 4
  5. 5
  6. 6
  7. 7
  8. 8
    Features and development of coot
    1. P Emsley
    2. B Lohkamp
    3. WG Scott
    4. K Cowtan
    (2010)
    Acta Crystallographica. Section D, Biological Crystallography 66:486–501.
    https://doi.org/10.1107/S0907444910007493
  9. 9
  10. 10
  11. 11
  12. 12
  13. 13
  14. 14
  15. 15
  16. 16
  17. 17
  18. 18
  19. 19
  20. 20
  21. 21
  22. 22
  23. 23
  24. 24
  25. 25
    Electron scattering factors of ions and their parameterization
    1. L-M Peng
    (1998)
    Acta Crystallographica Section a Foundations of Crystallography 54:481–485.
    https://doi.org/10.1107/S0108767398001901
  26. 26
  27. 27
  28. 28
  29. 29
  30. 30
  31. 31
  32. 32
    The PyMOL Molecular Graphics System
    1. L Schrödinger
    (2020)
    Schrödinger, LLC.
  33. 33
  34. 34
  35. 35
  36. 36
  37. 37
  38. 38

Decision letter

  1. Baron Chanda
    Reviewing Editor; University of Wisconsin-Madison, United States
  2. Richard W Aldrich
    Senior Editor; The University of Texas at Austin, United States
  3. Luis G Cuello
    Reviewer; Texas Tech University Health Sciences Center, United States

In the interests of transparency, eLife publishes the most substantive revision requests and the accompanying author responses.

Acceptance summary:

Although the human lysosomal potassium ion channel, TMEM175, lacks a canonical selectivity filter, it is 40-fold more selective to potassium than sodium. In a canonical K+ selectivity filter, K+ ions bind preferentially to the ion binding sites in the filter. This study reveals a potentially alternate mechanism of ion selectivity wherein the cost of desolvation of an ion as it enters a hydrophobic constriction is the most important determinant. K+ can be desolvated more easily than Na+ and therefore these channels are potassium selective.

Decision letter after peer review:

Thank you for submitting your article "Structural basis for permeation and selectivity in the non-canonical human lysosomal K+ channel TMEM175" for consideration by eLife. Your article has been reviewed by Richard Aldrich as the Senior Editor, a Reviewing Editor, and three reviewers.

The reviewers have discussed the reviews with one another and the Reviewing Editor has drafted this decision to help you prepare a revised submission.

Summary:

This study represents a significant contribution to our understanding of ion selectivity in a recently discovered new family of K+-permeable channels, the TMEM175 potassium channel family, which is present in all three kingdoms of life. In prokaryotes, these channels display modest potassium over sodium selectivity but in humans the TMEM175 is ~ 40-folds more selective for K+ than for Na+ ions. Interestingly, these novel channels lack the typical signature sequence (i.e., TTVGYGD) that is characteristic of the highly K+ selective channels, which are exemplified by the shaker and KcsA channels. Although structures of the prokaryotic channels have been solved recently, the ion-binding sites have not been observed in those structures. The authors solved structures of this channel in several conditions: (1) high potassium, (2) low pH, and (3) high Cs. The high K+ structure is at a resolution below 3 Å, and very high quality. There are several interesting aspects to this structure. First, two of the transmembrane segments from one of the two 6-TM repeats in the dimer are disordered and cannot be modeled, while the corresponding two segments from the other repeat are very well ordered. Second, despite the fact that the pore appears very narrow with the narrowest constriction formed by 4 isoleucines where only dehydrated potassium can squeeze by, there are many non-protein densities in the pore that the authors attribute to water and K+. To distinguish between them, they determined the structure in Cs+, under the assumption that since Cs+ scatters electrons stronger than K+, hTMEM175 in Cs+ should display strong peaks at the Cs+ binding sites, which should be in the same locations as K+, since Cs+ permeates the channel as well. They identified 5 K+ binding sites, where 3 are on the symmetry axis and 2 are off it but none of them are fully coordinated by protein atoms. They thus conclude that the selectivity arises from the transient dehydration of the ions as they pass through the narrowest Ile constriction. Overall, this study contributes new knowledge that will ultimately lead to a broader understanding of the mechanisms of ion selectivity. Nevertheless, all the reviewers have expressed concerns that the interpretation of the data is some instances is too narrow and a more nuanced discussion is in order.

Essential revisions:

1) One of the key points here is the role of TMEM175 "hydrophobic constrictions" as sieving barriers for discriminating between the various dehydrated cations. This is difficult to reconcile with our existing knowledge of ion permeation even though a similar mechanism has been proposed previously by Jiang and his colleagues. One possibility is these channels are in a closed conformation and that these hydrophobic constrictions act as dewetted gates (as opposed to steric gates). Although mutations in the hydrophobic residues alter ion conduction, it is not clear whether these mutations modify the structure. I would like the authors to rationalize their point of view and discuss it in the context of dewetting and the possibility that the hydrophobic constriction could be hydrophobic gates (Aryal et al., 2015).

2) Related to the above point. Please add a paragraph in the Discussion section acknowledging the possibility that the channel is in a closed state in the structures presented and that there may be additional determinants for selectivity in this channel family.

3) In subsection “Human TMEM175 is highly selective for K+”, the ACMA assay to test the function of the TMEM175 was done in POPE:POPG liposomes and the phospholipid composition so the lysosome is dramatically different. Can the authors argue why did they do the assay in this lipid composition? Why not to perform the ACMA assay as it has been done before for the Hv1 channel? The Mackinnon group performed the assay in a lipid composition that mimic the membrane of human neutrophil plasma membrane, which is more physiologically relevant.

4) Subsection “Structure of the ion conduction pathway”, the authors concluded too strong that because the T274V was nonconductive, "the water-mediated ion binding site, which is disrupted in the T274V mutant is essential for ion permeation". Although it is indeed very provocative to state this conclusion, we can imagine several possible scenarios in which the T274V halt ion conduction beside disrupting the water-mediated ion binding site. I advise the authors to be more conservative and less conclusive since in the absence of the T274V structure, they cannot categorically conclude this. Other mutations could have been done (as T274S), and the cryo-EM structures of these mutants could have been solved to support the mechanism proposed by the authors.

5) The authors attempted to use a low resolution cryo-EM structure soaked in Cesium to assign the K+-binding sites in the TMTM175 structure. This experimental approach assume that the putative 5 ion binding sites are equivalents and that they have the same ion selectivity. It has been shown that in KcsA, the second binding site although highly selective for potassium ions, at steady state conditions, does not coordinate Cesium or Rubidium ions. So, the authors should explain to the reader the limitation(s) of this experimental approach since it could draw incorrect conclusions. For example, the Cs4 does not directly correspond to densities found on the hTMEM175k density map and that could be explain by a different type of interaction between Cs+ ions and the protein at that specific site.

6) In Subsection “Structure of the ion conduction pathway” the authors concluded that because the cryo-EM structure at pH 5.5, that mimics a late endosomes and lysosomes, seems to be identical to the hTMEM175k (obtained at basic pH), then these results are consistent with a channel no gated by pH. I agree to disagree here, since there are many assumptions and generalization in this sentence. For starters, nobody knows how the hTMEM175 channel is gated. The very simplistic and superficial functional studies of the hTMEM175 solely showed non-inactivating and non-voltage dependent macroscopic currents. But these whole cell currents could arise for channels that are inactivated but at the steady state have a sizable rate of recovery from the inactive state. One could argue that the absence of a lipid bilayer precluded the structural changes elicited by the acidic pH. The lack of structural changes does not categorically prove the point of Oh et al. Hence, I would advise a more conservative narrative that include other reasons of why the structures at different pHs are identical. Finally, it is very intriguing why the protonation state of the acidic residues that coordinate K1 (i.e., D279 or E282), which are very close to the luminal side, did not change the occupancy of the ion at the K1 site. I would argue that in the absence of a lipid bilayer, the local concentration of protons is significantly lower in the cryo-EM structure and therefore the ion occupancy of the K1 site is the same as the one obtained at basic pH.

7) The authors proposed that negative charges on the channel help to select cations over anions. However, there is no experimental data to support this claim. Does the selectivity change if these charges are neutralized? If this model is true, the channel ion selectivity should change as a function of the pHs, as the protonation state of these residues is changed.

8) I would like the authors to discuss the fact that the TMEM175 normally works with a deltapH, which is acidic at the luminal side (pH of ~4.8). How does this fact affect their conclusions about the lack of pH-dependent putative structural changes? In physiologically relevant conditions the channel confront to dramatically different environments in regard to the proton concentration.

9) I am concerned about the reversal potential measurements. If the seals have some degree of leak, the authors could be wrong estimating the reversal potential. For instance, if the cells have some chloride leak, as usually happen, your reversal potential is going to be altered. A leak subtraction protocol (P over N) could solve this problem. As the ion selectivity measurements are an important dataset in the present work, I think the authors should make sure that the reported measurements are fully reliable, hence a very clear explanation how they were performed must be provided.

10) Lack of functional activity in a mutant channel, such as the T274V or the I46M/I271M mutants can arise due to an effect on channel folding/assembly in the membrane. An interpretation of the mutants requires that an effect of the amino acid substitution on folding/assembly is ruled out.

11) A I46N/I274N substitution has been shown to result in a loss of ionic selectivity (in Lee, 2017). This result should be described and discussed in the manuscript.

[Editors' note: further revisions were suggested prior to acceptance, as described below.]

Thank you for resubmitting your work entitled "Gating and selectivity mechanisms for the lysosomal K+ channel TMEM175" for further consideration by eLife. Your revised article has been evaluated by Richard Aldrich (Senior Editor) and a Reviewing Editor.

The manuscript has been improved but there are some remaining issues that need to be addressed before acceptance, as outlined below:

1) The reviewers agree that the manuscript has improved significantly, and the additional data may provide insight into the mechanism of gating in TMEM175. Nonetheless, the reviewers remain concerned that you are over interpreting your data and want you to tone down the claims. While your discovery is certainly exciting, it is important to convey to a general reader the uncertainties associated with interpretations based primarily on the structure. Given that there is no single channel data or additional biochemistry to provide a "functional" framework for interpreting this structural data, it is necessary that you explicitly address these caveats.

2) Even with regards to ion selectivity, one has to be careful because the open structure may not correspond to an open state unless it is validated in another way. If that is the case, then the mechanism of ion selectivity as proposed may not stand. Therefore, please make sure this caveat is taken into consideration.

3) The figure numbers are out registry i.e., in subsection “Ion-conduction pathway contains ordered ions and waters” it reads.…"Glu259 via ordered water (Figure 4A)" it should be (Figure 5A).

4) Can the authors add the distance between the ions on Figure 4A, B and D?

5) The assignment of the density at the Ile constriction as K+ is highly speculative. The density is very weak in the Cs+ map and is very poor when the C2 symmetry is not applied. It is also hard to imagine that K+ ions would be stable in an hydrophobic environment. The authors need to mention that the assignment of the density as K+ is speculative.

6) The FSEC profiles shown in Figure 3—figure supplement 1A are saturating the detector (flat at the top). These curves cannot be used to claim that the mutants were expressed at similar levels to wild type as the levels can be different but not seen due to detector saturation. That claim should be dropped. The curves do indicate that the mutants fold similar to the wild type.

https://doi.org/10.7554/eLife.53430.sa1

Author response

Summary:

This study represents a significant contribution to our understanding of ion selectivity in a recently discovered new family of K+-permeable channels, the TMEM175 potassium channel family, which is present in all three kingdoms of life. In prokaryotes, these channels display modest potassium over sodium selectivity but in humans the TMEM175 is ~ 40-folds more selective for K+ than for Na+ ions. Interestingly, these novel channels lack the typical signature sequence (i.e., TTVGYGD) that is characteristic of the highly K+ selective channels, which are exemplified by the shaker and KcsA channels. Although structures of the prokaryotic channels have been solved recently, the ion-binding sites have not been observed in those structures. The authors solved structures of this channel in several conditions: (1) high potassium, (2) low pH, and (3) high Cs. The high K+ structure is at a resolution below 3 Å, and very high quality. There are several interesting aspects to this structure. First, two of the transmembrane segments from one of the two 6-TM repeats in the dimer are disordered and cannot be modeled, while the corresponding two segments from the other repeat are very well ordered. Second, despite the fact that the pore appears very narrow with the narrowest constriction formed by 4 isoleucines where only dehydrated potassium can squeeze by, there are many non-protein densities in the pore that the authors attribute to water and K+. To distinguish between them, they determined the structure in Cs, under the assumption that since Cs+ scatters electrons stronger than K+, hTMEM175 in Cs+ should display strong peaks at the Cs binding sites, which should be in the same locations as K+, since Cs+ permeates the channel as well. They identified 5 K+ binding sites, where 3 are on the symmetry axis and 2 are off it but none of them are fully coordinated by protein atoms. They thus conclude that the selectivity arises from the transient dehydration of the ions as they pass through the narrowest Ile constriction. Overall, this study contributes new knowledge that will ultimately lead to a broader understanding of the mechanisms of ion selectivity. Nevertheless, all the reviewers have expressed concerns that the interpretation of the data is some instances is too narrow and a more nuanced discussion is in order.

We thank the editors and reviewers for their positive assessment of our manuscript and their detailed and constructive comments. To address some of the reviewers’ concerns, we developed a new image processing workflow to better extract structural information from our cryo-EM images. The re-processing improved the yield of “good” particles and enabled us to resolve a novel conformational state. This new conformation, which we call class 2, exists in the both the KCl and CsCl data sets. By removing the class 2 particles from the KCl data set, we were able to obtain a more homogenous particle population. Refinement of these particles resulted in a 2.6 Å reconstruction, which we call class 1 and is similar to the structure that we described in our initial submission. Comparison of the class 1 and class 2 states reveals significant conformational changes in the pore, including changes to Ile46 and Ile271. In class 2, Ile46 and Ile271 form a constriction whose minimum radius is 0.5 Å that blocks ion permeation through the pore. In class 1, the isoleucine constriction widens to a minimum radius of 1.6 Å and density corresponding to a K+ ion is resolved within the constriction. The ion density is elongated, extending throughout the constriction, consistent with the ion occupying multiple binding sites. Ordered water molecules are resolved on either side of the constriction that can coordinate the ion as it enters the constriction from either side. We propose that class 1 represents a conductive state and that the rigid architecture of isoleucine constriction necessitates the transient dehydration of ions, thereby favoring K+ over Na+. Thus, our work represents a major advance over previous studies of TMEM175 homologs by answering two of the major outstanding questions regarding human TMEM175 function – how it is gated and how it can selectively permeate potassium ions.

Due to the substantial changes in the data, we have extensively rewritten the manuscript and modified several of figures. Our point-by-point responses to the reviewers’ comments are listed below.

Essential revisions:

1) One of the key points here is the role of TMEM175 "hydrophobic constrictions" as sieving barriers for discriminating between the various dehydrated cations. This is difficult to reconcile with our existing knowledge of ion permeation even though a similar mechanism has been proposed previously by Jiang and his colleagues. One possibility is these channels are in a closed conformation and that these hydrophobic constrictions act as dewetted gates (as opposed to steric gates). Although mutations in the hydrophobic residues alter ion conduction, it is not clear whether these mutations modify the structure. I would like the authors to rationalize their point of view and discuss it in the context of dewetting and the possibility that the hydrophobic constriction could be hydrophobic gates (Aryal et al., 2015).

We thank the reviewers for their careful consideration of our proposed selectivity mechanism. To characterize the role of the constriction formed by Ile46 and Ile271 in ion selectivity and channel gating, we re-processed our data and identified a novel conformation, which we call class 2. In this new class 2 state, Ile46 and Ile271 constrict the pore to a minimum radius of 0.5 Å, which is too narrow to accommodate cations, even in a dehydrated state (Figure 5). We therefore conclude that class 2 represents a closed state and that Ile46 and Ile271 form a steric gate to ion permeation.

In the higher-resolution class 1 density map that we describe in the revised manuscript, the isoleucine constriction adopts a wider conformation than in class 2 and density corresponding to a K+ ion is resolved (Figure 3, Figure 4 and Figure 5). The K+ ion density is elongated, rather than spherical, consistent with the ion partially occupying multiple different positions. Density is also resolved for ordered water molecules on either side of the constriction that are positioned to coordinate ions as they approach the constriction. Using whole-cell patch clamp, we assessed the function of channels possessing mutations to several of the residues that coordinate the water molecules near the isoleucine constriction. In both cases, mutations that disrupt the ability of these residues to participate in the coordination of these waters (S45A and T274V) result in non-functional channels, while conservative mutations (S45T and T274S) yield functional channels (Figure 6B). We therefore propose that class 1 represents a conductive state and that Ile46 and Ile271 form a constriction that selectively permeates cations based on their relative enthalpies of dehydration.

Together, these structures reveal two roles for the isoleucine constriction. In the closed structure, the constriction forms a steric gate that prevents ion permeation. In the open structure, the constriction widens are forms a selectivity filter.

2) Related to the above point. Please add a paragraph in the Discussion section acknowledging the possibility that the channel is in a closed state in the structures presented and that there may be additional determinants for selectivity in this channel family.

In the revised version of the manuscript, we describe structures of hTMEM175 in two conformations. In the closed class 2 state, the channel is gated by a closure of the isoleucine constriction (Figure 5C). In the class 1 state, ion density is resolved within the widened isoleucine constriction (Figure 3D). The ion in the constriction is coordinated by water molecules that reside on either side of the constriction that can coordinate ions as the enter the constriction from either side (Figure 6). In the revised manuscript, we discuss in subsection “Structural heterogeneity reveals gating mechanism” these features of the structure and propose that class 1 represents a conductive state of hTMEM175.

In addition to the isoleucine constriction, our electrophysiological analysis of wild-type and mutant TMEM175 channels is consistent with Thr274 also participating in ion selectivity. Mutation of Thr274 to serine right-shifted the reversal potential from ~-65 mV for wild-type hTMEM175 to ~-34 mV.

3) In subsection “Human TMEM175 is highly selective for K+”, the ACMA assay to test the function of the TMEM175 was done in POPE:POPG liposomes and the phospholipid composition so the lysosome is dramatically different. Can the authors argue why did they do the assay in this lipid composition? Why not to perform the ACMA assay as it has been done before for the Hv1 channel? The Mackinnon group performed the assay in a lipid composition that mimic the membrane of human neutrophil plasma membrane, which is more physiologically relevant.

We agree with the reviewers that POPE:POPG liposomes do not reflect the physiological composition of the lysosomal membrane. Instead, we performed the proteolipsome flux assay using channels reconstituted into POPE:POPG liposomes, as has been used for analysis of a number of other ion channels. For the proteolipsome flux assay that we are using, POPE:POPG liposomes are particularly advantageous because they do not allow proton leak. In contrast, protons can readily permeate through protein-free liposomes of several commonly used lipid mixtures, including E. coli polar lipids and those containing PC lipids. Because PC is a major constituent of endolysosomal membranes (Escribá et al., 2015), which would generate background leak in our assay, we performed our experiments with POPE:POPG liposomes. In future studies, it will be essential to characterize the influence of different lipids on the function of hTMEM175.

4) Subsection “Structure of the ion conduction pathway”, the authors concluded too strong that because the T274V was nonconductive, "the water-mediated ion binding site, which is disrupted in the T274V mutant is essential for ion permeation". Although it is indeed very provocative to state this conclusion, we can imagine several possible scenarios in which the T274V halt ion conduction beside disrupting the water-mediated ion binding site. I advise the authors to be more conservative and less conclusive since in the absence of the T274V structure, they cannot categorically conclude this. Other mutations could have been done (as T274S), and the cryo-EM structures of these mutants could have been solved to support the mechanism proposed by the authors.

To further investigate the role of water molecules in the permeation of hTMEM175, we performed additional mutagenesis experiments. In addition to our analysis of the T274V mutant, we examined a more conservative T274S mutant as suggested by the reviewers as well as conservative and non-conservative mutations to Ser45, whose side chain participates in the coordination of water molecules on the cytosolic side of the isoleucine constriction. We first compared the folding of these mutants to wild-type hTMEM175 by FSEC analysis (Goehring et al., 2014), which revealed that all of the mutants eluted single peaks with retention volumes similar to wild type (Figure S8). We next examined their ability to selectively permeate Cs+ ions using whole-cell patch clamp. Cs+-selective currents could be detected from cells expressing the conservative S45T or T274S mutations while cells expressing the non-conservative S45A or T274V mutants displayed current levels that were indistinguishable from background currents measured in non-transfected cells (Figure 6B-C). While it remains possible that the S45A and T274V mutations alter trafficking to the plasma membrane, it is unlikely as the conservative mutants are both localized to the plasma membrane. Thus, we interpret the lack of exogenous currents from cells expressing the S45A and T274V mutants as resulting from defects in channel function.

5) The authors attempted to use a low resolution cryo-EM structure soaked in Cesium to assign the K+-binding sites in the TMTM175 structure. This experimental approach assume that the putative 5 ion binding sites are equivalents and that they have the same ion selectivity. It has been shown that in KcsA, the second binding site although highly selective for potassium ions, at steady state conditions, does not coordinate Cesium or Rubidium ions. So, the authors should explain to the reader the limitation(s) of this experimental approach since it could draw incorrect conclusions. For example, the Cs4 does not directly correspond to densities found on the hTMEM175k density map and that could be explain by a different type of interaction between Cs+ ions and the protein at that specific site.

We agree with the reviewers that the Cs+ is imperfect mimic of K+ and that it may bind to proteins differently. This is a fundamental limitation of our approach and we have expanded on this limitation in the Discussion section.

6) In Subsection “Structure of the ion conduction pathway” the authors concluded that because the cryo-EM structure at pH 5.5, that mimics a late endosomes and lysosomes, seems to be identical to the hTMEM175k (obtained at basic pH), then these results are consistent with a channel no gated by pH. I agree to disagree here, since there are many assumptions and generalization in this sentence. For starters, nobody knows how the hTMEM175 channel is gated. The very simplistic and superficial functional studies of the hTMEM175 solely showed non-inactivating and non-voltage dependent macroscopic currents. But these whole cell currents could arise for channels that are inactivated but at the steady state have a sizable rate of recovery from the inactive state. One could argue that the absence of a lipid bilayer precluded the structural changes elicited by the acidic pH. The lack of structural changes does not categorically prove the point of Oh et al. Hence, I would advise a more conservative narrative that include other reasons of why the structures at different pHs are identical. Finally, it is very intriguing why the protonation state of the acidic residues that coordinate K1 (i.e., D279 or E282), which are very close to the luminal side, did not change the occupancy of the ion at the K1 site. I would argue that in the absence of a lipid bilayer, the local concentration of protons is significantly lower in the cryo-EM structure and therefore the ion occupancy of the K1 site is the same as the one obtained at basic pH.

We agree with the reviewers that the gating of hTMEM175 has yet to be functionally well characterized and it remains unclear how pH will alter channel activity. Our functional analyses of hTMEM175 were performed at pH 7.4 and thus do not inform on channel activity at pH 5.5. Because the focus of our manuscript has changed with the addition of the closed structure and because we have not performed a detailed functional characterization of hTMEM175 in acidic conditions, we have excluded the structure of hTMEM175 in the pH 5.5 condition, which is nearly identical to the pH 8.0 condition from our manuscript.

7) The authors proposed that negative charges on the channel help to select cations over anions. However, there is no experimental data to support this claim. Does the selectivity change if these charges are neutralized? If this model is true, the channel ion selectivity should change as a function of the pHs, as the protonation state of these residues is changed.

We agree with the reviewers that our proposed mechanism for cation over anion selectivity was speculative and have removed this discussion from the manuscript.

8) I would like the authors to discuss the fact that the TMEM175 normally works with a deltapH, which is acidic at the luminal side (pH of ~4.8). How does this fact affect their conclusions about the lack of pH-dependent putative structural changes? In physiologically relevant conditions the channel confront to dramatically different environments in regard to the proton concentration.

As mentioned for point 6, we have removed the pH 5.5 structure from the manuscript. We would therefore prefer not to speculate on the effects of a deltapH on channel gating.

9) I am concerned about the reversal potential measurements. If the seals have some degree of leak, the authors could be wrong estimating the reversal potential. For instance, if the cells have some chloride leak, as usually happen, your reversal potential is going to be altered. A leak subtraction protocol (P over N) could solve this problem. As the ion selectivity measurements are an important dataset in the present work, I think the authors should make sure that the reported measurements are fully reliable, hence a very clear explanation how they were performed must be provided.

We agree with the reviewers that systematic errors exist in our reversal potential measurements due to endogenous background currents. Our experiments were conducted using bath solution containing 145 mM Na-methanesulfonate (MS) 5 mM NaCl or KCl, 1 mM MgCl2, 1 mM CaCl2, 10 mM HEPES/Tris pH 7.4. and a pipette solution containing 150 mM KMS or Cs-MS, 5 mM MgCl2, 10 mM EGTA/Tris, 10 mM HEPES/Tris pH 7.4. The magnitude of the chloride component of these currents is reduced by using methanesulfonate as the predominant anion.

As suggested by the reviewers, we attempted to perform a subtraction experiment using the mean background current of the non-transfected cells. However, as we show in Figure 2—figure supplement 1, the magnitude of the background currents of the non-transfected cells varied substantially and the subtractions resulted in traces with clear artefacts and were not interpretable. Because these systematic errors in our permeability measurement could not be corrected using a background subtraction protocol, we emphasize in subsection “Human TMEM175 is highly selective for K+” in the manuscript that these measurements underestimate the true reversal potential.

10) Lack of functional activity in a mutant channel, such as the T274V or the I46M/I271M mutants can arise due to an effect on channel folding/assembly in the membrane. An interpretation of the mutants requires that an effect of the amino acid substitution on folding/assembly is ruled out.

As described in our response to point 4, we used FSEC to analyze the hydrodynamic radius of GFP-tagged mutant hTMEM175 channels that we characterized by electrophysiology. In all cases, the channels eluted as a single peak with a similar retention volume to wild-type hTMEM175. We interpret this result as the mutants all retaining their dimer architecture (Figure 4—figure supplement 1).

11) A I46N/I274N substitution has been shown to result in a loss of ionic selectivity (in Lee, 2017). This result should be described and discussed in the manuscript.

We agree with the reviewers and have included the I46N/I274N mutant in our discussion of ion selectivity of TMEM175 channels in subsection “Ion binding sites in the pore”.

[Editors' note: further revisions were suggested prior to acceptance, as described below.]

1) The reviewers agree that the manuscript has improved significantly, and the additional data may provide insight into the mechanism of gating in TMEM175. Nonetheless, the reviewers remain concerned that you are over interpreting your data and want you to tone down the claims. While your discovery is certainly exciting, it is important to convey to a general reader the uncertainties associated with interpretations based primarily on the structure. Given that there is no single channel data or additional biochemistry to provide a "functional" framework for interpreting this structural data, it is necessary that you explicitly address these caveats.

We thank the reviewers for the critical assessment of our work. We have added a paragraph to the discussion addressing some of the caveats of structural characterizations of our work on TMEM175 and of ion channels in general.

2) Even with regards to ion selectivity, one has to be careful because the open structure may not correspond to an open state unless it is validated in another way. If that is the case, then the mechanism of ion selectivity as proposed may not stand. Therefore, please make sure this caveat is taken into consideration.

We agree with the reviewers that due to the inherent difficulties in assigning discrete functional states to static cryo-EM structures, it is not possible to be certain that our open structure may correspond to an open state. We have revised the discussion to better address the possibility that our functional assignments are speculative.

3) The figure numbers are out registry i.e., in subsection “Ion-conduction pathway contains ordered ions and waters” it reads.…"Glu 259 via ordered water (Figure 4A)" it should be (Figure 5A).

Thank you very much, we have corrected the figure citations.

4) Can the authors add the distance between the ions on Figure 4A, B and D?

Thank you very much for this suggestion, we have added the distances between the ions in the figures. Furthermore, because the distances between symmetry-related K1, K2 and K4 ion binding sites are ~3 Å, we have noted in the text at the end of the third paragraph in subsection “Ion-conduction pathway contains ordered ions and waters” that it is unlikely that both of the symmetry-related ion binding sites are simultaneously occupied.

5) The assignment of the density at the Ile constriction as K+ is highly speculative. The density is very weak in the Cs+ map and is very poor when the C2 symmetry is not applied. It is also hard to imagine that K+ ions would be stable in a hydrophobic environment. The authors need to mention that the assignment of the density as K+ is speculative.

We thank the reviewers for their comment. We agree that the ion binding site near the Ile constriction is the weakest of the four sites resolved in both density maps and that it is unusual ion binding site being so closely positioned to the hydrophobic Ile side chains. We have revised the text in subsection “Ion-conduction pathway contains ordered ions and waters” to highlight that ions will only transiently access the hydrophobic environment of the isoleucine constriction. The ions will more stably bind at the cytoplasmic or luminal sides of the constriction where they can be partially coordinated by ordered waters.

6) The FSEC profiles shown in Figure 3—figure supplement 1A are saturating the detector (flat at the top). These curves cannot be used to claim that the mutants were expressed at similar levels to wild type as the levels can be different but not seen due to detector saturation. That claim should be dropped. The curves do indicate that the mutants fold similar to the wild type.

We have repeated the FSEC analysis using a lower laser gain power in which the detector is not saturated. The results demonstrate that with the exception of T274V, the mutant constructs express at similar levels as wild-type TMEM175. We have modified the text in subsection “Ion-conduction pathway contains ordered ions and waters” and replaced figure 3—figure supplement 1A with the updated panel.

https://doi.org/10.7554/eLife.53430.sa2

Article and author information

Author details

  1. SeCheol Oh

    Structural Biology Program, Memorial Sloan Kettering Cancer Center, New York, United States
    Contribution
    Conceptualization, Data curation, Formal analysis, Validation, Investigation, Methodology
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-1685-5922
  2. Navid Paknejad

    Structural Biology Program, Memorial Sloan Kettering Cancer Center, New York, United States
    Contribution
    Data curation, Methodology
    Competing interests
    No competing interests declared
  3. Richard K Hite

    Structural Biology Program, Memorial Sloan Kettering Cancer Center, New York, United States
    Contribution
    Conceptualization, Formal analysis, Supervision, Funding acquisition, Investigation, Methodology, Project administration
    For correspondence
    hiter@mskcc.org
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0003-0496-0669

Funding

Searle Scholars Program

  • Richard K Hite

Memorial Sloan Kettering Cancer Center (Josie Robertson Investigators Program)

  • Richard K Hite

National Cancer Institute (P30 CA008748)

  • Richard K Hite

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

Acknowledgements

We thank M de la Cruz at The MSKCC Richard Rifkind Center for cryo-EM for assistance with data collection, the MSKCC HPC group for assistance with data processing and SB Long and T Walz for comments on the manuscript. This work was supported in part by NIH-NCI Cancer Center Support Grant (P30 CA008748), the Josie Robertson Investigators Program (to RKH) and the Searle Scholars Program (to RKH).

Senior Editor

  1. Richard W Aldrich, The University of Texas at Austin, United States

Reviewing Editor

  1. Baron Chanda, University of Wisconsin-Madison, United States

Reviewer

  1. Luis G Cuello, Texas Tech University Health Sciences Center, United States

Publication history

  1. Received: November 7, 2019
  2. Accepted: March 29, 2020
  3. Accepted Manuscript published: March 31, 2020 (version 1)
  4. Version of Record published: April 8, 2020 (version 2)

Copyright

© 2020, Oh et al.

This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.

Metrics

  • 2,568
    Page views
  • 483
    Downloads
  • 0
    Citations

Article citation count generated by polling the highest count across the following sources: Crossref, PubMed Central, Scopus.

Download links

A two-part list of links to download the article, or parts of the article, in various formats.

Downloads (link to download the article as PDF)

Download citations (links to download the citations from this article in formats compatible with various reference manager tools)

Open citations (links to open the citations from this article in various online reference manager services)

Further reading

    1. Cell Biology
    2. Structural Biology and Molecular Biophysics
    Lin Mei et al.
    Research Article

    The actin cytoskeleton mediates mechanical coupling between cells and their tissue microenvironments. The architecture and composition of actin networks are modulated by force, but it is unclear how interactions between actin filaments (F-actin) and associated proteins are mechanically regulated. Here, we employ both optical trapping and biochemical reconstitution with myosin motor proteins to show single piconewton forces applied solely to F-actin enhance binding by the human version of the essential cell-cell adhesion protein αE-catenin, but not its homolog vinculin. Cryo-electron microscopy structures of both proteins bound to F-actin reveal unique rearrangements that facilitate their flexible C-termini refolding to engage distinct interfaces. Truncating α-catenin's C-terminus eliminates force-activated F-actin binding, and addition of this motif to vinculin confers force-activated binding, demonstrating that α-catenin's C-terminus is a modular detector of F-actin tension. Our studies establish that piconewton force on F-actin can enhance partner binding, which we propose mechanically regulates cellular adhesion through a-catenin.

    1. Microbiology and Infectious Disease
    2. Structural Biology and Molecular Biophysics
    Clarissa L Durie et al.
    Research Article Updated

    Legionella pneumophila is an opportunistic pathogen that causes the potentially fatal pneumonia Legionnaires’ Disease. This infection and subsequent pathology require the Dot/Icm Type IV Secretion System (T4SS) to deliver effector proteins into host cells. Compared to prototypical T4SSs, the Dot/Icm assembly is much larger, containing ~27 different components including a core complex reported to be composed of five proteins: DotC, DotD, DotF, DotG, and DotH. Using single particle cryo-electron microscopy (cryo-EM), we report reconstructions of the core complex of the Dot/Icm T4SS that includes a symmetry mismatch between distinct structural features of the outer membrane cap (OMC) and periplasmic ring (PR). We present models of known core complex proteins, DotC, DotD, and DotH, and two structurally similar proteins within the core complex, DotK and Lpg0657. This analysis reveals the stoichiometry and contact interfaces between the key proteins of the Dot/Icm T4SS core complex and provides a framework for understanding a complex molecular machine.