The family of bestrophin proteins (BEST1-4) was identified by linkage analysis to hereditary macular degenerations caused by mutations in BEST1 (Petrukhin et al., 1998; Marquardt et al., 1998). To date more than 200 mutations in BEST1 are linked with eye disease (Johnson et al., 2017; Xiao et al., 2010). BEST1-4 proteins are expressed in the plasma membrane and form Ca²⁺-activated Cl^- channels by assembling as pentamers (Hartzell et al., 2008; Kane Dickson et al., 2014; Vaisey et al., 2016; Sun et al., 2002; Tsunenari et al., 2006). Data indicate that BEST1 mediates a Ca²⁺-activated Cl^- current that is integral to human retinal pigment epithelial function (Li et al., 2017). The broad tissue distribution of the family suggests additional physiological functions that are not fully realized (Hartzell et al., 2008), and these may include processes as diverse as cell volume regulation (Fischmeister and Hartzell, 2005; Milenkovic et al., 2015), pH homeostasis (Yu et al., 2010), and neurotransmitter release (Lee et al., 2010).

An X-ray structure of chicken BEST1, which shares 74% sequence identity with human BEST1 and possesses analogous Ca²⁺-activation and anion-selectivity properties, revealed an architecture that is distinct from other ion channel families (Kane Dickson et al., 2014; Vaisey et al., 2016). The channel comprises five BEST1 subunits arranged symmetrically around a central ion pore that is ~95 Å long. The diameter of the pore varies along its length and contains two constrictions, the ‘neck’ and the ‘aperture’, both of which are lined by hydrophobic amino acids. The X-ray structure of a bestrophin channel from the prokaryotic organism Klebsiella pneumoniae (KpBest) revealed a similar architecture, including analogous neck and aperture regions, in spite of minimal (14%) sequence identity to human BEST1 (Yang et al., 2014). Surprisingly, however, whereas metazoan bestrophin channels are anion-selective and activated by Ca²⁺, KpBest is cation-selective and does not contain the Ca²⁺ sensor region (the ‘Ca²⁺ clasp’) that is conserved amongst metazoan organisms.

From electrophysiological studies of purified chicken BEST1 and of mutants inspired by the structure, it was deduced that Ca²⁺-activation of the channel is due to the binding of Ca²⁺ ions to the cytosolic Ca²⁺ clasps, of which there are five owing to the pentameric architecture, and to subsequent conformational changes in the neck (Vaisey et al., 2016). These experiments made it clear that the neck forms the activation gate that opens in response to the binding of cytosolic Ca²⁺ ions. Perplexingly, however, even though the Ca²⁺ clasps were occupied by Ca²⁺ ions in the X-ray structure, the conformation of the neck was too narrow to allow hydrated ions to pass. We initially postulated that phenylalanine residues within the neck (F80 and F84) might coordinate permeating anions via anion-quadropole interactions (also referred to as anion-π interactions) and thereby stabilize dehydrated anions as they passed though a narrow neck (Kane Dickson et al., 2014). However, we found that mutation of these residues to alanine had no effect on the ion selectivity properties of the channel, which made it seem unlikely that permeation through the neck was facilitated by anion-quadropole interactions in the wild type channel (Vaisey et al., 2016). Further, molecular simulation studies suggested that the conformation of the neck observed in the X-ray structure would form a significant energetic barrier that would likely prevent ion permeation (Rao et al., 2017). Thus, it was not clear what conformation the X-ray structure represented or how anions might be able to pass through the neck.

Our recent electrophysiological studies using purified chicken BEST1 in synthetic bilayers show that, like many ion channels, BEST1 undergoes inactivation wherein ionic currents through BEST1 decrease over time (Vaisey and Long, 2018). We now realize that the ‘current rundown’ previously observed for mammalian bestrophin channels is due to channel inactivation and that channel inactivation is an inherent property of bestrophin channels and not due to other cellular factors. Inactivation is stimulated by high (>500 nM) concentrations of Ca²⁺ and is caused by the binding of a C-terminal ‘inactivation peptide’, contained within amino acids 346–367, to a receptor on the channel’s cytosolic surface that is located near the Ca²⁺ clasp (Vaisey and Long, 2018). Inactivation can be substantially diminished by mutation of the receptor site or mutation of the inactivation peptide, and it can be prevented altogether by removal of the inactivation peptide (Vaisey and Long, 2018). From these and other mutagenesis studies it was additionally concluded that the neck functions as the inactivation gate that prevents ion flow in the inactivated state (Vaisey and Long, 2018). These results began to suggest that the X-ray structure might represent an inactivated conformation of the channel.

The X-ray structure of BEST1 was determined in complex with an antibody (denoted 10D10) to facilitate crystallization. Electrophysiological analyses of the antibody’s effect on BEST1 currents recorded using purified channels that were reconstituted into lipid bilayers indicate that it stimulates inactivation (Vaisey and Long, 2018), providing further evidence that the X-ray structure represents an inactivated state of the channel. Thus, because the inactivation peptide is bound to its receptor in the X-ray structure and because the antibody used for co-crystallization stimulates inactivation, we suspect that the X-ray structure represents an inactivated state in which the neck is too narrow to allow ions to flow through it (Kane Dickson et al., 2014; Vaisey and Long, 2018). The aforementioned studies suggest that the neck must widen to allow ion flow, but the extent of this widening and the conformational changes that underlie it are not clear. For example, minor movements of the side chain residues lining the neck might be sufficient for ion conduction, or opening might involve more dramatic rearrangements. Another unanswered question is how Ca²⁺ binding controls the neck. It is also unclear what conformation the channel might adopt when the Ca²⁺ clasps are emptied of their Ca²⁺ ions. To address these questions we determined a series of cryo-electron microscopy (cryo-EM) structures of chicken BEST1 with and without Ca²⁺ in activated, deactivated, and inactivated states that represent the major gating transitions of the channel. We find that ion conduction is permitted by dramatic widening of the neck and by correspondingly dramatic conformational changes in the protein. The conformational changes are more akin to localized protein refolding rather than the twisting or domain movement changes that typically underlie gating in ion channels and define the molecular bases of activation and inactivation gating in bestrophin channels.

To begin studying the structural basis of gating in BEST1, we determined a 3.1 Å resolution cryo-EM structure of Ca²⁺-bound BEST1 without an antibody (the same construct used for X-ray studies, comprising residues 1–405 of chicken BEST1, and termed BEST1₄₀₅). From single particle analysis we obtained only a single conformation, which indicates substantial conformational homogeneity among the ~300,000 channel molecules in the dataset. The conformation of the channel is indistinguishable from the previously published X-ray structure (Kane Dickson et al., 2014) (Figure 1—figure supplements 1,2, RMSD for Cα atoms = 0.5 Å). As in the X-ray structure, each of the inactivation peptides, of which there are five owing to the symmetry of the channel, is bound to its cytosolic receptor and well defined in the density. Therefore, this cryo-EM structure also represents a Ca²⁺-bound inactivated state. This result indicates that antibody binding does not distort the channel from a native conformation and reflects a satisfying correspondence of structures determined using X-ray crystallography and cryo-EM methodologies.

An open channel conformation

In an aim to obtain structural information on an open conformation of the channel, we used a construct spanning amino acids 1–345 (termed BEST1₃₄₅) for cryo-EM studies. BEST1₃₄₅ does not inactivate because it is missing the C-terminal inactivation peptide (Vaisey and Long, 2018). Importantly, whilst it does not inactivate, we have previously shown that BEST1₃₄₅ possesses normal Cl^- selectivity and Ca²⁺-dependent activation (Vaisey and Long, 2018). Single particle cryo-EM analysis of Ca²⁺-bound BEST1₃₄₅ (Figure 1—figure supplements 2,3) revealed two distinct conformations of the channel (Figure 1, Figure 1—figure supplement 4). The first, determined at 3.0 Å resolution, represents 86% of the particles and is essentially indistinguishable from the structure of BEST1₄₀₅, except for the absence of the inactivation peptide (Figure 1A,C, Figure 1—figure supplement 1C, RMSD for Cα atoms = 0.5 Å). Because BEST1₃₄₅ does not inactivate, the structure presumably represents a Ca²⁺-bound closed conformation. The conformation, in which the activation gate (the neck) is closed even though Ca²⁺ is bound, is consistent with single channel recordings of drosophila BEST1, which indicate that the channel has a low probability of being open (Chien et al., 2006). The neck adopts the same conformation in both the Ca²⁺-bound closed structure and the inactivated structure, (Figure 2C). The second conformation present within the cryo-EM dataset, which represents 14% of the particles and is determined to 2.9 Å resolution, contains a dramatically widened pore within the neck (Figure 1B,D). Based on discussions presented herein we conclude that this represents the open conformation of the channel. The relative abundance of the closed conformation suggests that it is energetically favorable.

Figure 1 with 5 supplements see all

Download asset Open asset

Figure 2 with 2 supplements see all

Download asset Open asset

In the closed conformation, the neck is less than 3.5 Å in diameter and approximately 15 Å long; three hydrophobic amino acids on the neck helix (S2b) from each of the channel’s five subunits, I76, F80 and F84, form its walls (Figure 1C,E,F) (Kane Dickson et al., 2014). Its narrowness, length and hydrophobicity would create an energetic barrier to ion permeation, consistent with simulation studies of wild-type and neck mutant BEST1 channel structures (Rao et al., 2017). In the open conformation, the neck has dilated to approximately 13 Å in diameter, which is more than sufficient to allow permeation of hydrated Cl^- ions (Figure 1E). No appreciable conformational difference is present in the cytosolic region of the channel, and in particular, the aperture constriction of the pore retains its dimensions. In accord with conclusions drawn from previous electrophysiological studies of mutations made within the neck and aperture (Vaisey et al., 2016; Vaisey and Long, 2018), we conclude that the neck functions as a gate. It permits ion flow in the open conformation and prevents it in the closed conformation.

Ca²⁺-free conformation

To address how Ca²⁺ binding contributes to BEST1 channel gating, we determined the cryo-EM structure of BEST1₃₄₅ in the absence of Ca²⁺ (Figure 3, Figure 3—figure supplements 1 and 2). The structure is determined to 3.0 Å resolution and is indistinguishable from a cryo-EM structure of Ca²⁺-free BEST1₄₀₅ at 3.6 Å resolution. As was expected from biochemical and electrophysiological analyses (Vaisey and Long, 2018), the inactivation peptide is not bound to its receptor in the structure of BEST1₄₀₅ without Ca²⁺ (Figure 3—figure supplement 3). The neck is closed in the absence of Ca²⁺ (Figure 3A), and notably, it shares the same closed conformation as in the inactivated and Ca²⁺-bound closed structures (Figure 2C). The overall structure of the channel is very similar to the inactivated and Ca²⁺-bound closed conformations with the only notable differences near the Ca²⁺ clasp. In structures with Ca²⁺ bound, the five Ca²⁺ clasps, one from each subunit, resemble a belt that wraps around the midsection of BEST1 (Figure 1A,B). Without Ca²⁺, the majority of each Ca²⁺ clasp becomes disordered (Figure 3B). 3D classification of the Ca²⁺-free BEST1₃₄₅ dataset yielded only closed conformations of the neck but did indicate a degree of flexibility in the channel between the transmembrane and cytosolic regions that was manifested as a ~ 5° rotation along the symmetry axis (Figure 3C). This conformational flexibility was not observed in the Ca²⁺-bound datasets, which suggests that Ca²⁺ binding rigidifies the channel. We hypothesize that this Ca²⁺-dependent rigidification may be necessary for opening of the neck.

Figure 3 with 3 supplements see all

Download asset Open asset

Coupling between the Ca²⁺ clasp and the gate

When considering the amino acid side chain movements that are involved in the transition between the open and closed conformations of the neck and how Ca²⁺ binding to the Ca²⁺ clasp might bias these, our attention was drawn to the S4a helix, which is located behind the neck helix, and to W287 in particular. W287 is perfectly conserved among metazoan bestrophin channels and even conserved in the prokaryotic KpBEST channel (Yang et al., 2014) (Figure 4A); this sequence conservation would be consistent with an important role in channel gating. W287 is also positioned approximately at the midpoint between the Ca²⁺ clasp and the neck (Figure 2A,B). W287 adopts one side chain rotamer to pack with F80 and F84 in the open conformation and a different side chain rotamer in the closed conformation that buttresses the space between adjacent neck helices (Figure 2A,B, Figure 4E,F). Because of its location and the conformational changes it undergoes, we hypothesized that W287 might propagate conformational changes in the Ca²⁺ clasp to the neck and thereby govern the conformation of the neck. We find that the W287F mutation (BEST1₃₄₅ W287F) produces channels with dramatically altered gating. Whilst BEST1₃₄₅ W287F retained normal Cl^- versus K⁺ selectivity (Figure 4B), the W287F mutation constitutively activates the channel, making it nearly insensitive to the presence or absence of Ca²⁺. Approximately, 80% of the Cl^- current level through BEST1₃₄₅ W287F is maintained when Ca²⁺ was chelated with EGTA, whereas currents through wild type BEST1 drop to zero (Figure 4B,C). To understand the molecular basis of this behavior, we determined cryo-EM structures of BEST1₃₄₅ W287F in the presence and absence of Ca²⁺ at 2.7 Å and 3.0 Å resolutions, respectively (Figure 4—figure supplements 1 and 2). In the presence of Ca²⁺, BEST1₃₄₅ W287F adopts an open conformation that is essentially indistinguishable from the open conformation observed for BEST1₃₄₅ (Figure 4—figure supplement 3, Cα RMSD = 0.2 Å). Unlike BEST1₃₄₅, 3D classification of the BEST1₃₄₅ W287F dataset did not reveal a closed conformation, which indicates that essentially all of the particles are in an open conformation and that the open state is preferential for this mutant. Remarkably, in the absence of Ca²⁺, and in spite of missing density for Ca²⁺ and a disordered Ca²⁺-clasp region, the neck also adopts an open conformation (Figure 4D, Figure 4—figure supplement 3). Thus, in accord with the electrophysiological recordings, the W287F mutation decouples Ca²⁺ binding from the conformational changes in the neck that are observed for the wild type channel. Modeling of the W287F mutation on a closed conformation of the channel introduces a void behind the neck (Figure 4—figure supplement 3E), which we hypothesize energetically disfavors the closed conformation. The effects of the relatively conservative mutation of this tryptophan to a phenylalanine give context to a myriad of disease-causing missense mutations in and around the neck (Figure 2F).

Figure 4 with 3 supplements see all

Download asset Open asset

The aperture

The structures reveal that the open pore of BEST1 comprises a 90 Å-long water-filled vestibule and a short constriction at the cytosolic aperture (Figure 1D). The aperture, which is located approximately 55 Å from the neck, is only ~3 Å long (measured where the pore diameter is <4 Å); its walls are formed solely by the V205 side chains of the five subunits (Figure 5A). Retinitis pigmentosa can be caused by mutation of the corresponding I205 residue of human BEST1 to threonine (Davidson et al., 2009). The aperture adopts the same conformation in all of the structures of BEST1 determined to date, regardless of whether the neck is open or closed or whether or not Ca²⁺ is bound (Figure 1C,D, Figure 3, Figure 4—figure supplement 3). We have concluded previously from electrophysiological studies that the aperture does not function as either the activation or the inactivation gate (Vaisey and Long, 2018; Vaisey et al., 2016), and the structural data support this conclusion. Accordingly, the V205A mutation of chicken BEST1, which would be expected to widen the aperture markedly, has no effect on Ca²⁺-dependent activation or inactivation (Vaisey and Long, 2018; Vaisey et al., 2016). Mutations of V205 do, however, have dramatic effects on ion permeability (Vaisey et al., 2016) (Figure 5C,D, Figure 5—figure supplement 1). Both human and chicken BEST1 exhibit a lyotropic relative permeability sequence in which permeable anions that are more easily dehydrated than Cl^-, such as Br^-, I^- and SCN^-, are more permeable than Cl^- (Vaisey et al., 2016; Qu and Hartzell, 2000; Hartzell et al., 2005; Hille, 2001) (Figure 5C, P_X/P_Cl ~ 2.7 when X = Br, ~5.1 when X = I and ~14.4 when X = SCN). The lyotropic permeability sequence is an indication that at least partial dehydration occurs for ions as they permeate through the channel. Given the structural context of the open channel, the aperture is the only constriction where dehydration would occur (Figure 1D,E); the dimensions of the aperture are too narrow for a hydrated ion to pass and therefore the aperture would necessitate that anions shed at least some water molecules as they pass through it. The structures support the conclusion that the aperture is responsible for the lyotropic permeability sequence (Vaisey et al., 2016). Consistently, we find that mutation of V205 to a smaller residue (e.g. glycine, alanine, or serine) abolishes the lyotropic sequence among Cl^-, Br^-, I^- and SCN^- (Figure 5C), which suggests that the aperture is made sufficiently wide by these mutations to allow these ions to pass without dehydration. (It is worthwhile to note that our data are moot on the relative conductances among wild type and mutant channels but we hypothesize that wider apertures would yield larger conductances.) In contrast to the effects of small amino acids, mutation of V205 to isoleucine, a bulkier hydrophobic amino acid that would presumably reduce the diameter of the aperture, makes the lyotropic permeability differences between Cl^-, Br^-, I^- and SCN^- more dramatic (Figure 5C). These effects suggest that ions would shed more water molecules when passing through the V205I aperture than through the wild type one. Thus, the aperture controls permeability among halide anions by causing ion dehydration, and it thereby engenders the characteristic lyotropic permeability sequence of BEST1.

Figure 5 with 1 supplement see all

Download asset Open asset

We also investigated the permeability of the larger organic anions acetate, propionate and butyrate relative to Cl^- for wild-type and aperture mutant channels (Figure 5D). These organic anions were less permeable than Cl^- (and less permeable than Br^-, I^- and SCN^-), which was expected due to their significantly larger size and the narrow dimensions of the aperture. We found that the relative permeability of the channel to acetate, propionate and butyrate increases when V205 is substituted with the smaller alanine or glycine residues (Figure 5D). Thus, when the aperture is made wider by mutation, larger organic anions are more permeable. These data support the previous inference (Vaisey et al., 2016) that the aperture functions in the wild type channel as a size-selective filter. The aperture would tend to prevent the permeation of large cellular constituents such as proteins or nucleic acids but would allow smaller solutes such as Cl^- ions to pass. However, the data indicate that the aperture is not perfectly discriminatory for the smallest solutes, as indicated by the measureable permeability of the wild type channel to acetate (Figure 5D). Notably, the amino acid sequence at and around amino acid 205 varies among human BEST1-4 channels and among bestrophin channels in general (Figure 5B). This sequence variability at the aperture may endow BEST1-4 channels with distinct permeabilities related to their specific physiological functions and their particular biological locations.

The structures presented herein represent the major gating transitions in the metazoan bestrophin channel. The work provides structural context to previous conclusions drawn from electrophysiological studies that the neck functions as both the activation and the inactivation gate (Vaisey et al., 2016; Vaisey and Long, 2018). Figure 6 outlines our current understanding of the gating cycle. When unoccupied by Ca²⁺, the Ca²⁺ clasps are disordered, the neck is closed, and ions are prevented from permeating through the channel by the neck. When Ca²⁺ ions bind, the Ca²⁺ clasps become ordered and the open and closed conformations of the channel have free energies that are similar enough that the neck can sample both conformations. The open conformation of the neck is dramatically wider than the closed conformation and opening (activation) exposes small hydrophilic amino acids to the pore where large hydrophobic ones had been. The wide neck explains why the neck has essentially no influence on ion selectivity (Vaisey et al., 2016), in spite of speculation that it might be involved in ion selectivity from the initial X-ray structure of BEST1 (Kane Dickson et al., 2014) and the structure of KpBest (Yang et al., 2014). The channel inactivates at higher concentrations of Ca²⁺ through a mechanism that involves the binding of the C-terminal inactivation peptide to its receptor. The receptor is on the cytosolic surface of the channel and the binding of the inactivation peptide to it allosterically controls the conformation of the neck from a distance. The inactivated conformation of the neck is indistinguishable from the deactivated (Ca²⁺-free) one.

Figure 6

Download asset Open asset

Unlike voltage-dependent K⁺ and Na⁺ channels, in which ions are prevented from flowing by different mechanisms in the inactivated and deactivated states (Hille, 2001; Armstrong and Bezanilla, 1977), the same closed conformation of the neck is responsible for the deactivated and inactivated states of BEST1. Although this property in BEST1 is unusual among ion channels from a functional perspective, it seems less unusual from a structural perspective. Unlike many ion channels that have separate self-contained domains that are responsible for controlling channel gating, such as extracellular or intracellular ligand-binding domains, or voltage sensor domains within the membrane, which can often be studied in isolation using structural or functional approaches, BEST1 does not contain separable domains. The BEST1 channel comprises a single structural unit that resembles a barrel, with a portion of the barrel embedded in the membrane and a portion exposed to the cytosol. Additionally, the primary sequence of BEST1 does not have separate domains; rather the polypeptide portions that form the transmembrane and cytosolic regions of the channel are dispersed throughout the primary sequence (Kane Dickson et al., 2014). The midsection of the barrel contains the region that functions as a gate (the neck) and the cytosolic surface of the barrel is decorated with sensors that influence the gate – namely the Ca²⁺ clasps and the inactivation peptide receptors. These sensors, which function as integral parts of the channel as a whole and could not be studied in isolation, modulate the energetics of opening and closing of the neck through allosteric means. The sensors do so by relaying information received at the surface of the channel, such as the binding of Ca²⁺ or the inactivation peptide, through the core of the protein to the neck. The design of the neck itself appears well suited to receive this subtle energetic information in that the neck comprises five helices (the S2b neck helices) that are tethered on each end by linker regions that allow the helices to toggle between two preferred orientations that form the closed and open conformations of the neck. Thus, while localized twisting or domain motions that are instigated by separate sensor domains often constitute the activation mechanism of ion channels, in bestrophin cytosolic sensors on the surface of the channel control a dramatic molecular choreography within the core of the protein that underlies gating and reveals a fascinating new gating mechanism among ion channels.

Our electrophysiological and structural data suggest that the aperture remains essentially fixed throughout the Ca²⁺-dependent activation and inactivation processes. Rather than functioning as a gate, as has been proposed (Yang et al., 2014), our data indicate that the aperture functions as a size-selective filter that discriminates permeant species on the basis of size and dehydration energy. Its dimensions and its hydrophobicity are such that any ion or hydrophilic entity passing through it must be at least partially dehydrated to fit. The aperture is thereby responsible for the lyotropic permeability sequence among small anions (e.g. Cl^-, Br^-, I^-, SCN^-) in which ions that are more easily dehydrated are more permeable through the channel. The aperture also limits the flow of larger anions such as acetate and mutations that increase the width of the aperture allow such ions to permeate more readily. Our data does not exclude the possibility that the aperture does change in dimensions somewhat due to thermal motion or due to some other factor. It has been observed, for example, that single channel currents obtained from purified KpBest are stimulated by ATP (Zhang et al., 2018). The mechanism of this regulation, and if it occurs in metazoan bestrophin channels, is not yet clear, but activation of bestrophins by ATP, which has been suggested to be mediated through changes at the aperture (Zhang et al., 2018), remains an intriguing possibility.

It has been proposed that the inhibitory neurotransmitter GABA permeates through BEST1 and underlies a tonic form of synaptic inhibition in glia (Lee et al., 2010). Previously, this possibility seemed incongruous with the narrowness of the neck observed in the X-ray structure (Kane Dickson et al., 2014) but the widened neck of the open channel conformation makes the possibility of slow conductance of GABA and/or other larger solutes more plausible from a structural perspective. While the dimensions of the aperture in the atomic models of BEST1, based on the cryo-EM and X-ray studies, would appear to preclude permeation of molecules the size of GABA, thermal motions that transiently widen the aperture may allow slow permeation of such solutes at physiologically relevant rates. Further, it is conceivable that regulatory mechanisms that influence the aperture exist and would modulate such permeation. Given the importance of the aperature in regulating permeation, the sequence diversity in the aperture region is particularly intriguing. This diversity may give rise to distinct permeation properties among bestrophins and hints that the physiological functions of the channels are broader and more nuanced than currently appreciated.

Atomic coordinates and cryo-EM density maps of have been deposited with the PDB and Electron Microscopy Data Bank with the accession numbers: 6N23 (BEST1405, inactivated; EMD-9321), 6N24 (BEST1345 W287F mutant, Ca2+-free; EMD-9322), 6N25 (BEST1345 W287F mutant, Ca2+-bound; EMD-9323), 6N26 ( BEST1345 Ca2+-free closed state; EMD-9324), 6N27 (BEST1345 Ca2+-bound closed state; EMD-9325), and 6N28 ( BEST1345 Ca2+-bound open state; EMD-9326).

1. 1. Miller AN
   2. Vaisey G
   3. Long SB

   (2018) Protein Databank

   ID 6N23. BEST1 in a calcium-bound inactivated state.

   https://www.rcsb.org/structure/6N23
2. 1. Miller AN
   2. Vaisey G
   3. Long SB

   (2018) Protein Databank

   ID 6N24. BEST1 open state W287F mutant, calcium-free.

   https://www.rcsb.org/structure/6N24
3. 1. Miller AN
   2. Vaisey G
   3. Long SB

   (2018) Protein Databank

   ID 6N25. BEST1 open state W287F mutant, calcium-bound.

   https://www.rcsb.org/structure/6N25
4. 1. Miller AN
   2. Vaisey G
   3. Long SB

   (2018) Protein Databank

   ID 6N26. BEST1 calcium-free closed state (deactivated).

   https://www.rcsb.org/structure/6N26
5. 1. Miller AN
   2. Vaisey G
   3. Long SB

   (2018) Protein Databank

   ID 6N27. BEST1 calcium-bound closed state.

   https://www.rcsb.org/structure/6N27
6. 1. Miller AN
   2. Vaisey G
   3. Long SB

   (2018) Protein Databank

   ID 6N28. BEST1 calcium-bound open state.

   https://www.rcsb.org/structure/6N28
7. 1. Miller AN
   2. Vaisey G
   3. Long SB

   (2018) Electron Microscopy Data Bank

   ID EMD-9321. BEST1 in a calcium-bound inactivated state.

   https://www.ebi.ac.uk/pdbe/entry/emdb/EMD-9321
8. 1. Miller AN
   2. Vaisey G
   3. Long SB

   (2018) Electron Microscopy Data Bank

   ID EMD-9322. BEST1 open state W287F mutant, calcium-free.

   https://www.ebi.ac.uk/pdbe/entry/emdb/EMD-9322
9. 1. Miller AN
   2. Vaisey G
   3. Long SB

   (2018) Electron Microscopy Data Bank

   ID EMD-9323. BEST1 open state W287F mutant, calcium-bound.

   https://www.ebi.ac.uk/pdbe/entry/emdb/EMD-9323
10. 1. Miller AN
    2. Vaisey G
    3. Long SB

    (2018) Electron Microscopy Data Bank

    ID EMD-9324. BEST1 calcium-free closed state (deactivated).

    https://www.ebi.ac.uk/pdbe/entry/emdb/EMD-9324
11. 1. Miller AN
    2. Vaisey G

    (2018) Electron Microscopy Data Bank

    ID EMD-9325. BEST1 calcium-bound closed state.

    https://www.ebi.ac.uk/pdbe/entry/emdb/EMD-9325
12. 1. Miller AN
    2. Vaisey G
    3. Long SB

    (2018) Electron Microscopy Data Bank

    ID EMD-9326. BEST1 calcium-bound open state.

    https://www.ebi.ac.uk/pdbe/entry/emdb/EMD-9326

1. 1. Adams PD
   2. Afonine PV
   3. Bunkóczi G
   4. Chen VB
   5. Davis IW
   6. Echols N
   7. Headd JJ
   8. Hung LW
   9. Kapral GJ
   10. Grosse-Kunstleve RW
   11. McCoy AJ
   12. Moriarty NW
   13. Oeffner R
   14. Read RJ
   15. Richardson DC
   16. Richardson JS
   17. Terwilliger TC
   18. Zwart PH

   (2010) PHENIX: a comprehensive Python-based system for macromolecular structure solution

   Acta Crystallographica Section D Biological Crystallography 66:213–221.

   https://doi.org/10.1107/S0907444909052925

   • PubMed
   • Google Scholar
2. 1. Armstrong CM
   2. Bezanilla F

   (1977) Inactivation of the sodium channel. II. gating current experiments

   The Journal of General Physiology 70:567–590.

   https://doi.org/10.1085/jgp.70.5.567

   • PubMed
   • Google Scholar
3. 1. Chien LT
   2. Zhang ZR
   3. Hartzell HC

   (2006) Single cl- channels activated by Ca2+ in Drosophila S2 cells are mediated by bestrophins

   The Journal of General Physiology 128:247–259.

   https://doi.org/10.1085/jgp.200609581

   • PubMed
   • Google Scholar
4. 1. Davidson AE
   2. Millar ID
   3. Urquhart JE
   4. Burgess-Mullan R
   5. Shweikh Y
   6. Parry N
   7. O'Sullivan J
   8. Maher GJ
   9. McKibbin M
   10. Downes SM
   11. Lotery AJ
   12. Jacobson SG
   13. Brown PD
   14. Black GC
   15. Manson FD

   (2009) Missense mutations in a retinal pigment epithelium protein, bestrophin-1, cause retinitis pigmentosa

   The American Journal of Human Genetics 85:581–592.

   https://doi.org/10.1016/j.ajhg.2009.09.015

   • PubMed
   • Google Scholar
5. 1. Emsley P
   2. Lohkamp B
   3. Scott WG
   4. Cowtan K

   (2010) Features and development of coot

   Acta Crystallographica. Section D, Biological Crystallography 66:486–501.

   https://doi.org/10.1107/S0907444910007493

   • PubMed
   • Google Scholar
6. 1. Fernandez-Leiro R
   2. Scheres SHW

   (2017) A pipeline approach to single-particle processing in RELION

   Acta Crystallographica Section D Structural Biology 73:496–502.

   https://doi.org/10.1107/S2059798316019276

   • PubMed
   • Google Scholar
7. 1. Fischmeister R
   2. Hartzell HC

   (2005) Volume sensitivity of the bestrophin family of chloride channels

   The Journal of Physiology 562:477–491.

   https://doi.org/10.1113/jphysiol.2004.075622

   • PubMed
   • Google Scholar
8. 1. Grant T
   2. Rohou A
   3. Grigorieff N

   (2018) cisTEM, user-friendly software for single-particle image processing

   eLife 7:e14874.

   https://doi.org/10.7554/eLife.35383

   • PubMed
   • Google Scholar
9. 1. Hartzell C
   2. Putzier I
   3. Arreola J

   (2005) Calcium-activated chloride channels

   Annual Review of Physiology 67:719–758.

   https://doi.org/10.1146/annurev.physiol.67.032003.154341

   • PubMed
   • Google Scholar
10. 1. Hartzell HC
    2. Qu Z
    3. Yu K
    4. Xiao Q
    5. Chien LT

    (2008) Molecular physiology of bestrophins: multifunctional membrane proteins linked to best disease and other retinopathies

    Physiological Reviews 88:639–672.

    https://doi.org/10.1152/physrev.00022.2007

    • PubMed
    • Google Scholar
11. Book

    1. Hille B

    (2001)

    Ionic Channels of Excitable Membranes

    Sunderland, MA: Sinauer Associates.

    • Google Scholar
12. 1. Johnson AA
    2. Guziewicz KE
    3. Lee CJ
    4. Kalathur RC
    5. Pulido JS
    6. Marmorstein LY
    7. Marmorstein AD

    (2017) Bestrophin 1 and retinal disease

    Progress in Retinal and Eye Research 58:45–69.

    https://doi.org/10.1016/j.preteyeres.2017.01.006

    • PubMed
    • Google Scholar
13. 1. Kane Dickson V
    2. Pedi L
    3. Long SB

    (2014) Structure and insights into the function of a Ca(2+)-activated Cl(-) channel

    Nature 516:213–218.

    https://doi.org/10.1038/nature13913

    • PubMed
    • Google Scholar
14. 1. Lee S
    2. Yoon BE
    3. Berglund K
    4. Oh SJ
    5. Park H
    6. Shin HS
    7. Augustine GJ
    8. Lee CJ

    (2010) Channel-mediated tonic GABA release from glia

    Science 330:790–796.

    https://doi.org/10.1126/science.1184334

    • PubMed
    • Google Scholar
15. 1. Li Y
    2. Zhang Y
    3. Xu Y
    4. Kittredge A
    5. Ward N
    6. Chen S
    7. Tsang SH
    8. Yang T

    (2017) Patient^-specific mutations impair BESTROPHIN1's essential role in mediating Ca^2+-dependent Cl^- currents in human RPE

    eLife 6:213.

    https://doi.org/10.7554/eLife.29914

    • PubMed
    • Google Scholar
16. 1. Lovell SC
    2. Word JM
    3. Richardson JS
    4. Richardson DC

    (2000) The penultimate rotamer library

    Proteins: Structure, Function, and Genetics 40:389–408.

    https://doi.org/10.1002/1097-0134(20000815)40:3<389::AID-PROT50>3.0.CO;2-2

    • PubMed
    • Google Scholar
17. 1. Marquardt A
    2. Stöhr H
    3. Passmore LA
    4. Krämer F
    5. Rivera A
    6. Weber BH

    (1998) Mutations in a novel gene, VMD2, encoding a protein of unknown properties cause juvenile-onset vitelliform macular dystrophy (Best's disease)

    Human Molecular Genetics 7:1517–1525.

    https://doi.org/10.1093/hmg/7.9.1517

    • PubMed
    • Google Scholar
18. 1. Mastronarde DN

    (2005) Automated electron microscope tomography using robust prediction of specimen movements

    Journal of Structural Biology 152:36–51.

    https://doi.org/10.1016/j.jsb.2005.07.007

    • PubMed
    • Google Scholar
19. 1. Milenkovic A
    2. Brandl C
    3. Milenkovic VM
    4. Jendryke T
    5. Sirianant L
    6. Wanitchakool P
    7. Zimmermann S
    8. Reiff CM
    9. Horling F
    10. Schrewe H
    11. Schreiber R
    12. Kunzelmann K
    13. Wetzel CH
    14. Weber BH

    (2015) Bestrophin 1 is indispensable for volume regulation in human retinal pigment epithelium cells

    PNAS 112:E2630–E2639.

    https://doi.org/10.1073/pnas.1418840112

    • PubMed
    • Google Scholar
20. 1. Petrukhin K
    2. Koisti MJ
    3. Bakall B
    4. Li W
    5. Xie G
    6. Marknell T
    7. Sandgren O
    8. Forsman K
    9. Holmgren G
    10. Andreasson S
    11. Vujic M
    12. Bergen AA
    13. McGarty-Dugan V
    14. Figueroa D
    15. Austin CP
    16. Metzker ML
    17. Caskey CT
    18. Wadelius C

    (1998) Identification of the gene responsible for Best macular dystrophy

    Nature Genetics 19:241–247.

    https://doi.org/10.1038/915

    • PubMed
    • Google Scholar
21. 1. Pettersen EF
    2. Goddard TD
    3. Huang CC
    4. Couch GS
    5. Greenblatt DM
    6. Meng EC
    7. Ferrin TE

    (2004) UCSF Chimera--a visualization system for exploratory research and analysis

    Journal of Computational Chemistry 25:1605–1612.

    https://doi.org/10.1002/jcc.20084

    • PubMed
    • Google Scholar
22. 1. Qu Z
    2. Hartzell HC

    (2000)

    Anion permeation in ca(2+)-activated cl(-) channels

    The Journal of General Physiology 116:825–844.

    • PubMed
    • Google Scholar
23. 1. Rao S
    2. Klesse G
    3. Stansfeld PJ
    4. Tucker SJ
    5. Sansom MSP

    (2017) A BEST example of channel structure annotation by molecular simulation

    Channels 11:347–353.

    https://doi.org/10.1080/19336950.2017.1306163

    • PubMed
    • Google Scholar
24. 1. Rohou A
    2. Grigorieff N

    (2015) CTFFIND4: Fast and accurate defocus estimation from electron micrographs

    Journal of Structural Biology 192:216–221.

    https://doi.org/10.1016/j.jsb.2015.08.008

    • PubMed
    • Google Scholar
25. 1. Smart OS
    2. Neduvelil JG
    3. Wang X
    4. Wallace BA
    5. Sansom MS

    (1996) HOLE: a program for the analysis of the pore dimensions of ion channel structural models

    Journal of Molecular Graphics 14:354–360.

    https://doi.org/10.1016/S0263-7855(97)00009-X

    • PubMed
    • Google Scholar
26. 1. Sun H
    2. Tsunenari T
    3. Yau KW
    4. Nathans J

    (2002) The vitelliform macular dystrophy protein defines a new family of chloride channels

    PNAS 99:4008–4013.

    https://doi.org/10.1073/pnas.052692999

    • PubMed
    • Google Scholar
27. 1. Tang G
    2. Peng L
    3. Baldwin PR
    4. Mann DS
    5. Jiang W
    6. Rees I
    7. Ludtke SJ

    (2007) EMAN2: an extensible image processing suite for electron microscopy

    Journal of Structural Biology 157:38–46.

    https://doi.org/10.1016/j.jsb.2006.05.009

    • PubMed
    • Google Scholar
28. 1. Tsunenari T
    2. Nathans J
    3. Yau KW

    (2006) Ca2+-activated cl- current from human bestrophin-4 in excised membrane patches

    The Journal of General Physiology 127:749–754.

    https://doi.org/10.1085/jgp.200609527

    • PubMed
    • Google Scholar
29. 1. Vaisey G
    2. Miller AN
    3. Long SB

    (2016) Distinct regions that control ion selectivity and calcium-dependent activation in the bestrophin ion channel

    PNAS 113:E7399–E7408.

    https://doi.org/10.1073/pnas.1614688113

    • PubMed
    • Google Scholar
30. 1. Vaisey G
    2. Long SB

    (2018) An allosteric mechanism of inactivation in the calcium-dependent chloride channel BEST1

    The Journal of General Physiology 150:1484–1497.

    https://doi.org/10.1085/jgp.201812190

    • PubMed
    • Google Scholar
31. 1. Xiao Q
    2. Hartzell HC
    3. Yu K

    (2010) Bestrophins and retinopathies

    Pflügers Archiv - European Journal of Physiology 460:559–569.

    https://doi.org/10.1007/s00424-010-0821-5

    • PubMed
    • Google Scholar
32. 1. Yang T
    2. Liu Q
    3. Kloss B
    4. Bruni R
    5. Kalathur RC
    6. Guo Y
    7. Kloppmann E
    8. Rost B
    9. Colecraft HM
    10. Hendrickson WA

    (2014) Structure and selectivity in bestrophin ion channels

    Science 346:355–359.

    https://doi.org/10.1126/science.1259723

    • PubMed
    • Google Scholar
33. 1. Yu K
    2. Lujan R
    3. Marmorstein A
    4. Gabriel S
    5. Hartzell HC

    (2010) Bestrophin-2 mediates bicarbonate transport by goblet cells in mouse colon

    Journal of Clinical Investigation 120:1722–1735.

    https://doi.org/10.1172/JCI41129

    • PubMed
    • Google Scholar
34. 1. Zhang Y
    2. Kittredge A
    3. Ward N
    4. Ji C
    5. Chen S
    6. Yang T

    (2018) ATP activates bestrophin ion channels through direct interaction

    Nature Communications 9:12758.

    https://doi.org/10.1038/s41467-018-05616-4

    • PubMed
    • Google Scholar
35. 1. Zheng SQ
    2. Palovcak E
    3. Armache JP
    4. Verba KA
    5. Cheng Y
    6. Agard DA

    (2017) MotionCor2: anisotropic correction of beam-induced motion for improved cryo-electron microscopy

    Nature Methods 14:331–332.

    https://doi.org/10.1038/nmeth.4193

    • PubMed
    • Google Scholar

Author details

1. Alexandria N Miller

   Structural Biology Program, Memorial Sloan Kettering Cancer Center, New York, United States

   Contribution

   Conceptualization, Data curation, Formal analysis, Validation, Visualization, Methodology, Writing—original draft, Writing—review and editing, Determined structures of BEST1405, the W287F mutant, and structures in the absence of Ca2+, Performed electrophysiology experiments

   Contributed equally with

   George Vaisey

   Competing interests

   No competing interests declared

2. George Vaisey

   1. Structural Biology Program, Memorial Sloan Kettering Cancer Center, New York, United States
   2. Louis V. Gerstner Jr. Graduate School of Biomedical Sciences, Memorial Sloan Kettering Cancer Center, New York, United States

   Contribution

   Conceptualization, Data curation, Formal analysis, Validation, Visualization, Methodology, Writing—original draft, Writing—review and editing, Determined structures of the Ca2+-bound closed state and the Ca2+-bound open state, Performed electrophysiology experiments

   Contributed equally with

   Alexandria N Miller

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0002-8359-1314

3. Stephen B Long

   Structural Biology Program, Memorial Sloan Kettering Cancer Center, New York, United States

   Contribution

   Conceptualization, Supervision, Funding acquisition, Validation, Visualization, Methodology, Project administration, Writing—review and editing

   For correspondence

   longs@mskcc.org

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0002-8144-1398

• 5,067

  views

• 725

  downloads

• 38

  citations

Views, downloads and citations are aggregated across all versions of this paper published by eLife.