The simplest description of the gating cycle in the pore domain of a K⁺ channel requires at least four distinct kinetic states (Ostmeyer et al., 2013; Panyi and Deutsch, 2006; Yellen, 1998). At rest, the pore domain’s activation gate formed by the inner helix bundle is closed (C) while the selectivity filter is presumably conductive (O). We define this conformation as the C/O state. When an activating stimulus drives the opening of the channel’s inner helix bundle gate, the selective flow of K⁺ can be sustained for hundredths of milliseconds through the open (O/O) state. However, as a consequence of allosteric coupling between the inner gate and the pore helix near the selectivity filter (Cuello et al., 2010a; Pan et al., 2011), opening of the activation gate in many channels leads to a series of structural changes at the selectivity filter that renders it non-conductive (the O/I state), a process known as C-type inactivation (Hoshi et al., 1991; Liu et al., 1996; López-Barneo et al., 1993). Once the activating stimulus has ceased, the channel’s activation gate returns to the closed conformation and the channel transiently occupies the closed-inactivated (C/I) state with the selectivity filter inactivated. Ultimately, allosteric interactions with the activation gate reset the selectivity filter back to its conductive conformation (C/O), completing the gating cycle (Figure 1a).

Figure 1

Download asset Open asset

The high-resolution structures of KcsA in high and low K⁺ concentrations have been known for a number of years from crystals of KcsA-Fab complexes (Zhou et al., 2001a). These structures have been suggested to correspond to the putative C/O (high K⁺) and C/I (low K⁺) states (Zhou et al., 2001b). Previously, it was determined the structures for KcsA in a putatively partially open-inactivated and in its fully-open and deep-inactivated states at modest resolution by crystallizing a constitutively open mutant (KcsA-OM) displaying different degrees of openings at the activation gate (Cuello et al., 2010b). This strategy allowed to propose a set of structural and ion occupancy changes in the selectivity filter associated with KcsA inactivation gating. The observed changes involve an initial constriction of the permeation pathway at G77, between K⁺ binding sites S2 and S3 at the selectivity filter. This change is accompanied by a concomitant reduction in ion occupancy in S2 and it was proposed represents an early non-conductive state (I¹). As the inner helical bundle gate opens further, there is additional narrowing of the selectivity filter at G77, causing the collapse of the permeation pathway and the loss of K⁺ at both, S2 and S3 binding sites, which likely represents a deep C-type inactivated state of KcsA or I² (Cuello et al., 2010b). The existence of a deep C-type inactivated conformation of KcsA’s selectivity filter (I² state) is supported by solution and solid state Nuclear Magnetic Resonance spectroscopy experiments (Ader et al., 2009; Ader et al., 2008; Bhate et al., 2010; Imai et al., 2010; Wylie et al., 2014) and 2DIR measurements (Kratochvil et al., 2016). However, at ~3.2 Å resolution, the original O/I (Cuello et al., 2010b) structure does not provide information on the position of the carbonyl groups responsible for ion coordination (Zhou et al., 2001b).

Further analyses of these structures revealed that most of the degradation in overall resolution was likely due to conformational heterogeneity at the activation gate formed by the inner helix bundle (Cuello et al., 2010b). In an attempt to increase the resolution of pore domain states where the inner helix bundle gate is open, a set of cysteine-bridges between adjacent subunits was engineered, using the lower resolution structure of KcsA-OM as a template (Cuello et al., 2010b). This experimental approach, combined with targeted mutations known to affect the stability of the C-type inactivated conformation (Cordero-Morales et al., 2006), allowed the structural determination of KcsA in its ‘transient’ open-conductive (O/O, mutant E71A) and deep C-type inactivated (O/I, mutant Y82A) conformations at near atomic resolutions (2.25 Å and 2.4 Å, respectively).

In KcsA, once the activation gate opens, C-type inactivation is initiated by the loss of K⁺ coordination at the S2 binding site due to the narrowing of the permeation pathway at Gly 77 (Cuello et al., 2010b) (Figure 6a). Further constricting the selectivity filter at Gly 77 causes structural changes that eventually leads to a collapsed, deep C-type inactivated conformation of the filter with the loss of S2 and S3 K⁺ coordination sites. This mechanism of C-type inactivation has found support from a variety of spectroscopic and computational studies carried out in KcsA (Ader et al., 2009; Ader et al., 2008; Bhate and McDermott, 2012; Bhate et al., 2010; Imai et al., 2010; Ostmeyer et al., 2013; Wylie et al., 2014; Kratochvil et al., 2016). However, recent studies based on crystal structures of refolded semisynthetic closed-KcsA containing unnatural amino acids, have suggested that the collapsed structure of the selectivity filter might not represent the deep C-type-inactivated conformation in KcsA (Devaraneni et al., 2013).

The atomic-resolution crystal structures of KcsA’s O/O and O/I states strongly suggest otherwise, since the well-established non-inactivating E71A mutant (Cordero-Morales et al., 2006) prevented the collapse of the selectivity filter. This finding indicates that KcsA in the deep C-type inactivated state has a collapsed selectivity filter, as previously suggested (Cuello et al., 2010a). Additionally, the existence of time dependent-conductance changes in a D-Ala semisynthetic construct cannot be used as evidence against the structural correlation between the collapsed conformation of the filter and the process of KcsA’s inactivation. Not only do these studies vastly underestimate the intrinsic conformational dynamics of the selectivity filter, but recent 2DIR data points towards a more complex conformational landscape for the selectivity filter that might include additional non-conductive conformations (Kratochvil et al., 2016). Indeed, recent computational analyses suggest that the D-Alanine ‘rigidized’ selectivity filter could relax to a partially constricted conformation, with the D-Ala side chains at position 77 occluding the permeation pathway (Li et al., 2017). Therefore, it is possible that the 77D-Ala selectivity filter displayed a time-dependent lack of activity due to alternative constricted conformations of KcsA’s filter.

Furthermore, these experimental results, together with a number of recent reports in KcsA and other channels (Ader et al., 2009; Ben-Abu et al., 2009; Imai et al., 2010; Panyi and Deutsch, 2006; Peters et al., 2013) have demonstrated that the key driving force behind KcsA’s inactivation is an allosteric coupling between their activation gate and the selectivity filter. This strongly suggests that to properly study the structural changes associated with C-type inactivation at the selectivity filter of K⁺ channels, a channel must be trapped with its activation gate open, since the structural integrity of a conductive selectivity filter is influenced by a number of parameters, most importantly, the conformation of the activation gate due to their structural coupling. This coupling ultimately will strongly influence the energetics of interaction between permeant ions and the channel’s selectivity filter as well as the pKa of activation and deactivation gating. These observations underlie a novel hysteretic gating behavior in KcsA (Tilegenova et al., 2017). Additionally, the non-inactivating Locked-open-KcsA-E71A mutant structure is consistent with the collapsed filter hypothesis of C-type inactivation, since the open state displayed a non-collapsed selectivity filter in this inactivation-deficient mutant.

Recapitulating existing data with the present high-resolution structures, KcsA’s C-type inactivation can be explained by the following minimal mechanistic model (Figure 8—figure supplement 1). Upon activation gating (C/O-to-O/O transition) KcsA’s inner bundle gate opens, expands its diameter to ~22 Å and allowing the diffusion of hydrated K⁺ ions into the channel central cavity. Once in the cavity, a K⁺ ion sheds off water molecules (Figure 4b) immediately before entering a selectivity filter that has undergone subtle but distinctive structural changes. The O/O is a transient kinetic state that conducts K⁺ ions for hundredths of milliseconds. But a strong allosteric coupling between KcsA’s and the selectivity filter (Phe 103 clashes with I100, Thr 74 and 75 from the signature sequence) deforms the pore helix and as a consequence the selectivity filter undergoes compensatory structural changes (Pan et al., 2011; Cuello et al., 2010a). These changes involve (1) flipping of the selectivity filter backbone C = O, which triggers the consecutive loss of the second and third K⁺ binding sites as a function of the channel’s activation gate opening and (2) in the absence of the second and third K⁺ binding sites the selectivity filter collapses. A KcsA kinetic intermediate that has loss as much as one K⁺ binding site at the selectivity filter is considered an inactivated state (I₁) (Figure 8, Figure 8—figure supplement 1), which can dramatically reduce the unitary conductance of the channel(Morais-Cabral et al., 2001). This sequence of events underlies C-type inactivation gating in KcsA and likely in other Kv channels (O/O to O/I transition). A network of water molecules that resides in a cavity behind KcsA’s selectivity filter stabilizes its collapsed conformation; diffusion of water molecules into the inactivation cavity determines the rate and magnitude of C-type inactivation. Once the stimulus is terminated, the activation gate closes (O/I to C/I transition) although the selectivity filter remains collapsed for a short period of time, likely defined by the exit rate of water molecules from the inactivation cavity (Ostmeyer et al., 2013). Finally, the selectivity filter resets to the conductive conformation (C/I to C/O transition) completing the KcsA kinetic cycle (Figure 8—figure supplement 1.

Figure 8 with 1 supplement see all

Download asset Open asset

Trapping KcsA in its open conformation has allowed the development of an atomic description of its O/O and O/I kinetic states. These two new kinetic intermediate snapshots complement, at high resolution, the existing structures for the C/O (high K⁺-structure) and the C/I (low K⁺-structure) conformations (Zhou et al., 2001a) and highlight a network of water molecules behind the channel’s selectivity filter that stabilizes its collapsed conformation (Figure 6a and b). The structural description of this gating cycle stresses diffusion of water molecules into the inactivation cavity as a key determinant of the rate and magnitude of KcsA’s C-type inactivation.

Finally, a recent Cryo-EM structural study of the human Ether-à-go-go-Related K⁺ channel (hERG) suggested that the structural changes at the selectivity filter associated with C-type inactivation are subtle and might be different from those reported for KcsA (Wang and MacKinnon, 2017). Given hERG distinctive inactivation process, it is not surprising that the structure-function correlation underlying hERG inactivation is unique and different from the one displayed by KcsA and other voltage-gated K⁺ channels (Cuello et al., 2010a).

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. Ader C
   2. Schneider R
   3. Hornig S
   4. Velisetty P
   5. Wilson EM
   6. Lange A
   7. Giller K
   8. Ohmert I
   9. Martin-Eauclaire MF
   10. Trauner D
   11. Becker S
   12. Pongs O
   13. Baldus M

   (2008) A structural link between inactivation and block of a K+ channel

   Nature Structural & Molecular Biology 15:605–612.

   https://doi.org/10.1038/nsmb.1430

   • PubMed
   • Google Scholar
3. 1. Ader C
   2. Schneider R
   3. Hornig S
   4. Velisetty P
   5. Vardanyan V
   6. Giller K
   7. Ohmert I
   8. Becker S
   9. Pongs O
   10. Baldus M

   (2009) Coupling of activation and inactivation gate in a K+-channel: potassium and ligand sensitivity

   The EMBO Journal 28:2825–2834.

   https://doi.org/10.1038/emboj.2009.218

   • PubMed
   • Google Scholar
4. 1. Balendiran GK
   2. Pandian JR
   3. Drake E
   4. Vinayak A
   5. Verma M
   6. Cascio D

   (2014) B-factor Analysis and Conformational Rearrangement of Aldose Reductase

   Current Proteomics 11:151–160.

   https://doi.org/10.2174/157016461103140922163444

   • PubMed
   • Google Scholar
5. 1. Ben-Abu Y
   2. Zhou Y
   3. Zilberberg N
   4. Yifrach O

   (2009) Inverse coupling in leak and voltage-activated K+ channel gates underlies distinct roles in electrical signaling

   Nature Structural & Molecular Biology 16:71–79.

   https://doi.org/10.1038/nsmb.1525

   • PubMed
   • Google Scholar
6. 1. Bhate MP
   2. Wylie BJ
   3. Tian L
   4. McDermott AE

   (2010) Conformational dynamics in the selectivity filter of KcsA in response to potassium ion concentration

   Journal of Molecular Biology 401:155–166.

   https://doi.org/10.1016/j.jmb.2010.06.031

   • PubMed
   • Google Scholar
7. 1. Bhate MP
   2. McDermott AE

   (2012) Protonation state of E71 in KcsA and its role for channel collapse and inactivation

   PNAS 109:15265–15270.

   https://doi.org/10.1073/pnas.1211900109

   • PubMed
   • Google Scholar
8. 1. Chapman ML
   2. VanDongen HM
   3. VanDongen AM

   (1997) Activation-dependent subconductance levels in the drk1 K channel suggest a subunit basis for ion permeation and gating

   Biophysical Journal 72:708–719.

   https://doi.org/10.1016/S0006-3495(97)78707-1

   • PubMed
   • Google Scholar
9. 1. Chapman ML
   2. VanDongen AM

   (2005) K channel subconductance levels result from heteromeric pore conformations

   The Journal of General Physiology 126:87–103.

   https://doi.org/10.1085/jgp.200509253

   • PubMed
   • Google Scholar
10. 1. Cordero-Morales JF
    2. Cuello LG
    3. Zhao Y
    4. Jogini V
    5. Cortes DM
    6. Roux B
    7. Perozo E

    (2006) Molecular determinants of gating at the potassium-channel selectivity filter

    Nature Structural & Molecular Biology 13:311–318.

    https://doi.org/10.1038/nsmb1069

    • PubMed
    • Google Scholar
11. 1. Cuello LG
    2. Jogini V
    3. Cortes DM
    4. Pan AC
    5. Gagnon DG
    6. Dalmas O
    7. Cordero-Morales JF
    8. Chakrapani S
    9. Roux B
    10. Perozo E

    (2010a) Structural basis for the coupling between activation and inactivation gates in K(+) channels

    Nature 466:272–275.

    https://doi.org/10.1038/nature09136

    • PubMed
    • Google Scholar
12. 1. Cuello LG
    2. Jogini V
    3. Cortes DM
    4. Perozo E

    (2010b) Structural mechanism of C-type inactivation in K(+) channels

    Nature 466:203–208.

    https://doi.org/10.1038/nature09153

    • PubMed
    • Google Scholar
13. 1. Devaraneni PK
    2. Komarov AG
    3. Costantino CA
    4. Devereaux JJ
    5. Matulef K
    6. Valiyaveetil FI

    (2013) Semisynthetic K+ channels show that the constricted conformation of the selectivity filter is not the C-type inactivated state

    PNAS 110:15698–15703.

    https://doi.org/10.1073/pnas.1308699110

    • PubMed
    • Google Scholar
14. 1. Emsley P
    2. Cowtan K

    (2004) Coot: model-building tools for molecular graphics

    Acta Crystallographica Section D Biological Crystallography 60:2126–2132.

    https://doi.org/10.1107/S0907444904019158

    • PubMed
    • Google Scholar
15. 1. Hoshi T
    2. Zagotta WN
    3. Aldrich RW

    (1991) Two types of inactivation in Shaker K+ channels: effects of alterations in the carboxy-terminal region

    Neuron 7:547–556.

    https://doi.org/10.1016/0896-6273(91)90367-9

    • PubMed
    • Google Scholar
16. 1. Imai S
    2. Osawa M
    3. Takeuchi K
    4. Shimada I

    (2010) Structural basis underlying the dual gate properties of KcsA

    PNAS 107:6216–6221.

    https://doi.org/10.1073/pnas.0911270107

    • PubMed
    • Google Scholar
17. 1. Kratochvil HT
    2. Carr JK
    3. Matulef K
    4. Annen AW
    5. Li H
    6. Maj M
    7. Ostmeyer J
    8. Serrano AL
    9. Raghuraman H
    10. Moran SD
    11. Skinner JL
    12. Perozo E
    13. Roux B
    14. Valiyaveetil FI
    15. Zanni MT

    (2016) Instantaneous ion configurations in the K+ ion channel selectivity filter revealed by 2D IR spectroscopy

    Science 353:1040–1044.

    https://doi.org/10.1126/science.aag1447

    • PubMed
    • Google Scholar
18. 1. Li J
    2. Ostmeyer J
    3. Boulanger E
    4. Rui H
    5. Perozo E
    6. Roux B

    (2017) Chemical substitutions in the selectivity filter of potassium channels do not rule out constricted-like conformations for C-type inactivation

    PNAS 114:11145–11150.

    https://doi.org/10.1073/pnas.1706983114

    • PubMed
    • Google Scholar
19. 1. Liu Y
    2. Jurman ME
    3. Yellen G

    (1996) Dynamic rearrangement of the outer mouth of a K+ channel during gating

    Neuron 16:859–867.

    https://doi.org/10.1016/S0896-6273(00)80106-3

    • PubMed
    • Google Scholar
20. 1. Lockless SW
    2. Zhou M
    3. MacKinnon R

    (2007) Structural and thermodynamic properties of selective ion binding in a K+ channel

    PLoS Biology 5:e121.

    https://doi.org/10.1371/journal.pbio.0050121

    • PubMed
    • Google Scholar
21. 1. Long SB
    2. Campbell EB
    3. Mackinnon R

    (2005) Crystal structure of a mammalian voltage-dependent Shaker family K+ channel

    Science 309:897–903.

    https://doi.org/10.1126/science.1116269

    • PubMed
    • Google Scholar
22. 1. López-Barneo J
    2. Hoshi T
    3. Heinemann SH
    4. Aldrich RW

    (1993)

    Effects of external cations and mutations in the pore region on C-type inactivation of Shaker potassium channels

    Receptors & Channels 1:61–71.

    • PubMed
    • Google Scholar
23. 1. Morais-Cabral JH
    2. Zhou Y
    3. MacKinnon R

    (2001) Energetic optimization of ion conduction rate by the K+ selectivity filter

    Nature 414:37–42.

    https://doi.org/10.1038/35102000

    • PubMed
    • Google Scholar
24. 1. Ostmeyer J
    2. Chakrapani S
    3. Pan AC
    4. Perozo E
    5. Roux B

    (2013) Recovery from slow inactivation in K+ channels is controlled by water molecules

    Nature 501:121–124.

    https://doi.org/10.1038/nature12395

    • PubMed
    • Google Scholar
25. 1. Otwinowski Z
    2. Minor W

    (1997) Processing of X-ray diffraction data collected in oscillation mode

    Methods in Enzymology 276:307–326.

    https://doi.org/10.1016/S0076-6879(97)76066-X

    • PubMed
    • Google Scholar
26. 1. Pan AC
    2. Cuello LG
    3. Perozo E
    4. Roux B

    (2011) Thermodynamic coupling between activation and inactivation gating in potassium channels revealed by free energy molecular dynamics simulations

    The Journal of General Physiology 138:571–580.

    https://doi.org/10.1085/jgp.201110670

    • PubMed
    • Google Scholar
27. 1. Panyi G
    2. Deutsch C

    (2006) Cross talk between activation and slow inactivation gates of Shaker potassium channels

    The Journal of General Physiology 128:547–559.

    https://doi.org/10.1085/jgp.200609644

    • PubMed
    • Google Scholar
28. 1. Peters CJ
    2. Fedida D
    3. Accili EA

    (2013) Allosteric coupling of the inner activation gate to the outer pore of a potassium channel

    Scientific Reports 3:3025.

    https://doi.org/10.1038/srep03025

    • PubMed
    • Google Scholar
29. 1. Tilegenova C
    2. Cortes DM
    3. Cuello LG

    (2017) Hysteresis of KcsA potassium channel's activation- deactivation gating is caused by structural changes at the channel's selectivity filter

    PNAS 114:3234–3239.

    https://doi.org/10.1073/pnas.1618101114

    • PubMed
    • Google Scholar
30. 1. Uysal S
    2. Cuello LG
    3. Cortes DM
    4. Koide S
    5. Kossiakoff AA
    6. Perozo E

    (2011) Mechanism of activation gating in the full-length KcsA K+ channel

    PNAS 108:11896–11899.

    https://doi.org/10.1073/pnas.1105112108

    • PubMed
    • Google Scholar
31. 1. Wang W
    2. MacKinnon R

    (2017) Cryo-EM Structure of the Open Human Ether-à-go-go -Related K + Channel hERG

    Cell 169:422–430.

    https://doi.org/10.1016/j.cell.2017.03.048

    • PubMed
    • Google Scholar
32. 1. Wylie BJ
    2. Bhate MP
    3. McDermott AE

    (2014) Transmembrane allosteric coupling of the gates in a potassium channel

    PNAS 111:185–190.

    https://doi.org/10.1073/pnas.1319577110

    • PubMed
    • Google Scholar
33. 1. Yellen G

    (1998) The moving parts of voltage-gated ion channels

    Quarterly Reviews of Biophysics 31:239–295.

    https://doi.org/10.1017/S0033583598003448

    • PubMed
    • Google Scholar
34. 1. Zheng J
    2. Sigworth FJ

    (1997) Selectivity changes during activation of mutant Shaker potassium channels

    The Journal of General Physiology 110:101–117.

    https://doi.org/10.1085/jgp.110.2.101

    • PubMed
    • Google Scholar
35. 1. Zheng J
    2. Sigworth FJ

    (1998) Intermediate conductances during deactivation of heteromultimeric Shaker potassium channels

    The Journal of General Physiology 112:457–474.

    https://doi.org/10.1085/jgp.112.4.457

    • PubMed
    • Google Scholar
36. 1. Zhou M
    2. Morais-Cabral JH
    3. Mann S
    4. MacKinnon R

    (2001a) Potassium channel receptor site for the inactivation gate and quaternary amine inhibitors

    Nature 411:657–661.

    https://doi.org/10.1038/35079500

    • PubMed
    • Google Scholar
37. 1. Zhou Y
    2. Morais-Cabral JH
    3. Kaufman A
    4. MacKinnon R

    (2001b) Chemistry of ion coordination and hydration revealed by a K+ channel-Fab complex at 2.0 A resolution

    Nature 414:43–48.

    https://doi.org/10.1038/35102009

    • PubMed
    • Google Scholar
38. 1. Zhuo RG
    2. Peng P
    3. Liu XY
    4. Yan HT
    5. Xu JP
    6. Zheng JQ
    7. Wei XL
    8. Ma XY

    (2016) Allosteric coupling between proximal C-terminus and selectivity filter is facilitated by the movement of transmembrane segment 4 in TREK-2 channel

    Scientific Reports 6:21248.

    https://doi.org/10.1038/srep21248

    • PubMed
    • Google Scholar

Thank you for submitting your article "The gating cycle of a K⁺ channel at atomic resolution" for consideration by eLife. Your article has been reviewed by three peer reviewers, one of whom, Kenton J Swartz (Reviewer #2), is a member of our Board of Reviewing Editors, and the evaluation has been overseen by Richard Aldrich as the Senior Editor. The following individual involved in review of your submission has agreed to reveal their identity: Roderick MacKinnon (Reviewer #1).

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 manuscript addresses conformational states that underlie gating in the KcsA K⁺ channel. This channel opens transiently following steps in pH from 8 to 3 and then inactivates (some would say desensitizes) partly. Activation is associated with opening of an inner helix gate and inactivation as shown here seems associated with conformational changes at the selectivity filter. This conclusion is supported by demonstration that KcsA with a trapped open inner helix gate (using disulfide bridges) still gates. Mutations near the selectivity filter that increase or decrease inactivation favor different conformations in the setting of the trapped open inner helix gate. The same lab determined similar structures at an earlier date but these are at higher resolution. The higher resolution was achieved through the disulfide bridges, which limited heterogeneity. This study makes a compelling case for two distinct gates in KcsA that are both important for gating in KcsA. The following are requested revisions for improving the manuscript.

Essential revisions:

1) Unless we are mistaken, the functional characterization of the Y82A and E71A mutants was not done in the background of the disulfide bonded OM. This is not trivial to do because presumably these constructs would not be responsive to low pH because the lower pH controlled gate is locked open. However, it would be nice to know that they are not responsive to low pH and then to see single channel recording similar to those in Figure 2C, D that are provided for the OM alone. It would be nice to see that the E71A-OM has a high Po and that the Y82A-OM has a low Po, etc. Also, the functional data that are provided for E71A and Y82A in Figure 3B and the OM in Figure 2C, D are all single observations and the authors should provide population data.

2) In subsection “Trapping KcsA with its activation gate in the open conformation” of the manuscript, the authors refer to thermal denaturation experiments in Figure 2B, but no wild-type KcsA control is presented in the ms nor is a result for WT KcsA cited from the literature; this should be addressed. It's probably also worth noting in the text that the 28-118 disulfide bond is formed between residues in adjacent KcsA subunits. The authors should therefore note whether the sample buffer in Figure 2B contained DTT and/or β-mercaptoethanol, or if reducing agents were present during heating of the sample (no experimental details are provided).

3) We are skeptical about the discussion of B factors (subsection “Atomic resolution structure of KcsA open and conductive” paragraph three). How did the B factors compare in the 2 structures outside the selectivity filter? More importantly, one structure has a mutation in a residue surrounding the filter (E71A) and the other does not. Because such a mutation could affect the filter, we would hesitate to give too much interpretation to their B factor differences. Moreover, it's probably not appropriate to directly compare B-factors from at two different resolutions (2.0 Å for the 1K4C structure, 2.25 Å for the 5VK6 structure), in which the structures may have been refined using different parameterization. In addition, the B-factors plotted in Figure 5 appear to include those of both main chain and side chain atoms. Because B-factors for side chain atoms may be less well-defined and more subject to bias than main chain atoms, it would likely be more useful to limit the comparison to main chain atoms only. To further minimize bias in the comparisons, both data sets would need to be truncated to the same resolution (and probably same redundancy) and then re-refined, and then the authors will need to provide a strong rationale for the statistical distribution of the B-factors and provide some error estimates of these values to justify their argument. In some previous work, investigators have justified using a normalization of the B-factors, so this may be a valid approach.

4) The authors state that water exchange is a key determinant in the rate and magnitude of inactivation. This could be true but the evidence is not very strong. Does changing the activity of bulk water alter inactivation, as one would predict on the basis of this assertion? The authors should provide additional experimental support for this conclusion.

5) Subsection “Atomic resolution structure of KcsA open and deep C-type inactivated” and further in the Discussion: historically, the name C-type inactivation was given to the process of inactivation in Shaker K channels following removal of the N-terminal inactivation peptide. How do we know that the process of inactivation in KcsA (or desensitization following opening after a pH step) is the same as inactivation following voltage dependent gate opening in Shaker? We think it would be fine to motivate this work with previous studies on C-type inactivation in Shaker or other K channels, but at the end of the day this body of work is about KcsA and the authors should restrict their implications on mechanism to that protein and be more circumspect throughout.

6) Related to point 3, the discussion contained a large section opposing another study of inactivation based on D-alanine mutations that would seem to prevent the inactivated conformational change observed here. The authors state in the Discussion: "Indeed, recent computational analyses of a D-Alanine "rigidized" selectivity filter demonstrated that the selectivity filter is able to relax to a partially constricted conformation, with the D-Ala side chains at position 77 occluding the permeation pathway (Li et al., 2017). Therefore, the 77D-Ala substitution does not lead to a rigid selectivity filter, and crucially, does not preclude the transition to a constricted conformation for the selectivity filter." This argument supposes the correctness or accuracy of the computational studies. Perhaps more importantly, don't you think it is possible that inactivation in the selectivity filter, or its inability to conduct, might be associated with a number of different filter conformations (i.e. not all inactivated filters/states look the same)?

7) What is the distinction between a C-type inactivated state and a deep C-type inactivated state? Please explain the difference.

8) Subsection “Atomic resolution structure of KcsA open and deep C-type inactivated”: "Crystals of the O/I-state diffracted ~ 2.4 A with good refinement statistics, R_work = 0.1948 and R_free = 0.2267 and its structure was solved by molecular replacement…" Crystals do not diffract with good or bad refinement statistics. You mean the model exhibited those residuals when compared to the data. In addition, the crystallographic data table (table 1) is incomplete. The mean R_merge values for each data set are conspicuously high. Many investigators have switched to R_pim or CC1/2 as an indicator of overall data quality, but these are not provided. In addition, the value of i/sigma appears to only correspond to the high-resolution shell, and the overall i/sigma is not provided. Finally, the redundancy appears to only correspond to overall redundancy, and not high-resolution shell. Without a completed data table, it is difficult for the reader to independently evaluate the X-ray data quality or make comparisons among the data sets analyzed in this manuscript. For the refinement statistics, the authors list the "Wilson B-factor", which is estimated from the experimental data. The authors should instead provide the mean b-factors for atoms in each model; these are usually provided separately for the protein, water, and ligand atoms.

9) In the Discussion, the authors summarize their findings as if they have completed a mathematical proof: "Taking together existing data with the present high resolution structures we can recapitulate the process of C-type inactivation in the framework of a mechanistic model (Figure 8 and Figure 8—figure supplement 1)". Would it not be better to say, “on the basis of data presented we propose the following…” After all, we all know how tricky it can be to figure out what is really happening in these molecules!

Essential revisions:

1) Unless we are mistaken, the functional characterization of the Y82A and E71A mutants was not done in the background of the disulfide bonded OM. This is not trivial to do because presumably these constructs would not be responsive to low pH because the lower pH controlled gate is locked open. However, it would be nice to know that they are not responsive to low pH and then to see single channel recording similar to those in Figure 2C, D that are provided for the OM alone. It would be nice to see that the E71A-OM has a high Po and that the Y82A-OM has a low Po, etc. Also, the functional data that are provided for E71A and Y82A in Figure 3B and the OM in Figure 2C, D are all single observations and the authors should provide population data.

The reviewers are correct. Originally, we considered that sampling the behavior of the pH-activated channel would provide a close sampling of the gating transitions at the selectivity filter in either of the mutants. And while we still stand by that assessment, we do agree that formally, these electrophysiological measurements should be carried out on disulfide-cross-linked OM backgrounds (even if it is more difficult). Single channel traces are now shown in Figure 3D, with Figure 3E depicting overall Open Probability behavior for the population active in the patch (NPo). As expected, when the inner bundle gate is crosslinked, single channel activity becomes pH independent, and solely defined by the mutation at the selectivity filter/pore helix.

2) In subsection “Trapping KcsA with its activation gate in the open conformation” of the manuscript, the authors refer to thermal denaturation experiments in Figure 2B, but no wild-type KcsA control is presented in the ms nor is a result for WT KcsA cited from the literature; this should be addressed. It's probably also worth noting in the text that the 28-118 disulfide bond is formed between residues in adjacent KcsA subunits. The authors should therefore note whether the sample buffer in Figure 2B contained DTT and/or β-mercaptoethanol, or if reducing agents were present during heating of the sample (no experimental details are provided).

Thermal denaturation assays have been performed in our lab to assess overall stability of KcsA mutants for many years (first described in Cortes & Perozo, Biochemistry 36(33):10343-52 (1997). We followed the same protocol here. We now include the control experiment in Figure 2B, state that disulfide bonds are indeed inter-subunit (legend, figure 3), and include the methodological details of the experiment (Materials and methods section).

3) We are skeptical about the discussion of B factors (subsection “Atomic resolution structure of KcsA open and conductive” paragraph three). How did the B factors compare in the 2 structures outside the selectivity filter? More importantly, one structure has a mutation in a residue surrounding the filter (E71A) and the other does not. Because such a mutation could affect the filter, we would hesitate to give too much interpretation to their B factor differences. Moreover, it's probably not appropriate to directly compare B-factors from at two different resolutions (2.0 Å for the 1K4C structure, 2.25 Å for the 5VK6 structure), in which the structures may have been refined using different parameterization. In addition, the B-factors plotted in Figure 5 appear to include those of both main chain and side chain atoms. Because B-factors for side chain atoms may be less well-defined and more subject to bias than main chain atoms, it would likely be more useful to limit the comparison to main chain atoms only. To further minimize bias in the comparisons, both data sets would need to be truncated to the same resolution (and probably same redundancy) and then re-refined, and then the authors will need to provide a strong rationale for the statistical distribution of the B-factors and provide some error estimates of these values to justify their argument. In some previous work, investigators have justified using a normalization of the B-factors, so this may be a valid approach.

This is an important point brought by the reviewers. Formally, an approach to compare Debye-Weller B temperature factors in structures of different resolutions (and hence B factors) has been to artificially increase the overall values of the highest resolution structure. However, these comparisons are always limited by the worst-resolution data of the pair. Alternative, one could subtract the average B factor for a common stretch of residues in each case and compare, not the actual B factor values, but its variance (see Cordero-Morales et al., 2006). We have indeed followed the reviewer’s suggestions and carried out a de novo solution of wild-type KcsA, refined in house with the same parameters used during the refinement of the Locked-open E71A, at equivalent resolutions. This dataset was then used to redo the B-Factor analysis, but now with the normalized B-factors of the α carbons of amino acids of the selectivity filter. This newly generated comparison, between closed and open KcsA is shown in a new Figure 5.

4) The authors state that water exchange is a key determinant in the rate and magnitude of inactivation. This could be true but the evidence is not very strong. Does changing the activity of bulk water alter inactivation, as one would predict on the basis of this assertion? The authors should provide additional experimental support for this conclusion.

We have already carried out an earlier set of experiments to test this issue experimentally (in addition to extensive molecular dynamics simulations) see Ostmeyer et al., 2013. In this publication, we showed that reducing the occupancy of the buried water molecules at the filter (by imposing a high osmotic stress) accelerates the rate of recovery, as measured from the recovery rate of inactivation in the presence of a high concentration osmolyte solution (2M sucrose). The experiment provides strong, and direct experimental evidence supporting the hypothesis that water access to cavities behind the selectivity filter is a key driver of the kinetics on inactivation and inactivation repriming in KcsA. See Figures 4 B and C from Ostmeyer et al.

5) Subsection “Atomic resolution structure of KcsA open and deep C-type inactivated” and further in the Discussion: historically, the name C-type inactivation was given to the process of inactivation in Shaker K channels following removal of the N-terminal inactivation peptide. How do we know that the process of inactivation in KcsA (or desensitization following opening after a pH step) is the same as inactivation following voltage dependent gate opening in Shaker? We think it would be fine to motivate this work with previous studies on C-type inactivation in Shaker or other K channels, but at the end of the day this body of work is about KcsA and the authors should restrict their implications on mechanism to that protein and be more circumspect throughout.

We have to disagree with this assessment. First of all, others and we have demonstrated that the time-dependent reduction in open probability observed in KcsA is not due to a “desensitization” process, which would somehow affect the pH-dependence of the channel but to the existence of a second gate at KcsA’s selectivity filter (an inactivation gate). This is not a controversial point (see Cordero-Morales et al., 2006).

However, regarding nomenclature, perhaps this is an issue of semantics: while it is indeed correct that C-type inactivation was detected (and defined) in voltage dependent K⁺ channels as the time dependent reduction in ion current still present in the absence of fast N-type inactivation, there is ample evidence over the past 20+ years that point to the functional equivalence between the KcsA inactivation gating and that seen in voltage dependent K⁺ channels. The details differ, of course, but there are many common observations that strongly suggest, at the very least, a “kinship” between these evolutionarily related channels:

– Inactivation in KcsA displays the same kinetic relations (transition into and out of C-type inactivated state under a two-pulse protocol);

– Same external dependence of inactivation kinetics on external K⁺ concentrations;

– Same behavior with conductive and non-conductive ions: Rb⁺, plus Tl⁺, NH₄⁺ Cs⁺ (unpublished);

– Equivalent mutations that enhance C-type inactivation in eukaryotes do lock in the pinched filter conformation in KcsA;

– Mutations that inhibit C-type inactivation in eukaryotes lead to conductive filter conformations in KcsA.

We argue that the mechanism of inactivation in KcsA is at least related to those present in more complex eukaryotic channels. We all look for commonalities among evolutionarily related systems, but the jury is definitely out on whether there will be a “standard” C-type inactivation mechanism or different “classes” of it, depending on the pore domain. Our work points to the initiation of inactivation in KcsA as the loss of a potassium ion at the second binding site, accompanied by changes in pore diameter neat Glycine 77. This would lead to increase in the energy barrier for K⁺ translocation between biding sites. Still, there might be other ways to add a (relatively) subtle energy barrier to inhibit K⁺ flow, so that in some instances the Kv channel’s filter might not need to adopt the same pinched conformation to inactivate (as might be the case in the I1 conformation of KcsA). Ultimately, the observation of a similarly pinched filter in a eukaryotic pore domain (something that so far has not been observed) might help settle the score toward a “common mechanism”. Consequently, we have “toned down” the direct equivalence between eukaryotic C-type inactivation and KcsA inactivation and addressed the present results and interpretation within the context of the KcsA results and mechanism.

6) Related to point 3, the discussion contained a large section opposing another study of inactivation based on D-alanine mutations that would seem to prevent the inactivated conformational change observed here. The authors state in the Discussion: "Indeed, recent computational analyses of a D-Alanine "rigidized" selectivity filter demonstrated that the selectivity filter is able to relax to a partially constricted conformation, with the D-Ala side chains at position 77 occluding the permeation pathway (Li et al., 2017). Therefore, the 77D-Ala substitution does not lead to a rigid selectivity filter, and crucially, does not preclude the transition to a constricted conformation for the selectivity filter." This argument supposes the correctness or accuracy of the computational studies. Perhaps more importantly, don't you think it is possible that inactivation in the selectivity filter, or its inability to conduct, might be associated with a number of different filter conformations (i.e. not all inactivated filters/states look the same)?

The argument is clearly supported by the computational calculations, but it is not necessarily dependent on their accuracy. We make the case that simply, the complexity of behavior of the unnatural substitutions at the selectivity filter makes it impossible to univocally conclude that “the constricted conformation of the selectivity filter is not the C-type inactivated state” even when some reduction in open probability is observed in the presumably rigid D-Ala mutation.

We buttress our argument, not only with solid computational work (allowing molecular explorations that are very difficult to evaluate experimentally), but also from recent 2D infrared experiments done by us (and in collaboration with, among others, F Valiyaveetil). In this work, we provide extensive evidence for the complexity of conformations (as coupled vibrational modes) at the selectivity filter and its dependence on the state of the intracellular gate (see, Kratochvil HT et al. JACS 139(26):8837-8845 (2017)). Our findings indicate that changes in the intracellular gate are linked to changes in the selectivity filter conformational heterogeneity that are distinct to those triggered by K⁺ depletion. The existence of previously undetected conformational subensembles appear to underlie changes in ion affinity at the selectivity filter and suggest that this complexity does not disappear (just changes) in the presence of unnatural amino acids. More recent computational work (Li et al., now in press) provides strong evidence for an alternative pinched state (nonconductive) in the D-Ala KcsA mutant. We have reworked this section to provide a more inclusive set of interpretations to the inactivation process when D-Ala is part of the selectivity filter.

7) What is the distinction between a C-type inactivated state and a deep C-type inactivated state? Please explain the difference.

We defined the deep C-type inactivated state as that reached after steady state once the inner bundle gate opens (t →∞). With this definition, we can have a better equivalent to the determined crystal structures, while making a distinction with other (presumably non-conductive) states observed when the inner bundle gate is not fully open (see Cuello et al., 2010).

8) Subsection “Atomic resolution structure of KcsA open and deep C-type inactivated”: "Crystals of the O/I-state diffracted ~ 2.4 A with good refinement statistics, R_work = 0.1948 and R_free = 0.2267 and its structure was solved by molecular replacement…" Crystals do not diffract with good or bad refinement statistics. You mean the model exhibited those residuals when compared to the data. In addition, the crystallographic data table (table 1) is incomplete. The mean R_merge values for each data set are conspicuously high. Many investigators have switched to R_pim or CC1/2 as an indicator of overall data quality, but these are not provided. In addition, the value of i/sigma appears to only correspond to the high-resolution shell, and the overall i/sigma is not provided. Finally, the redundancy appears to only correspond to overall redundancy, and not high-resolution shell. Without a completed data table, it is difficult for the reader to independently evaluate the X-ray data quality or make comparisons among the data sets analyzed in this manuscript. For the refinement statistics, the authors list the "Wilson B-factor", which is estimated from the experimental data. The authors should instead provide the mean b-factors for atoms in each model; these are usually provided separately for the protein, water, and ligand atoms.

Agreed. The crystallographic tables are now fully populated and reformatted according to the reviewer’s suggestions.

9) In the Discussion, the authors summarize their findings as if they have completed a mathematical proof: "Taking together existing data with the present high resolution structures we can recapitulate the process of C-type inactivation in the framework of a mechanistic model (Figure 8 and Figure 8—figure supplement 1)". Would it not be better to say, “on the basis of data presented we propose the following…” After all, we all know how tricky it can be to figure out what is really happening in these molecules!

Fair enough! We have changed the text to be less deterministic and allow for the complexities of a still not fully understood process. Our explanations have been constrained to implications and interpretation relevant to KcsA inactivation while considering the possibility that our findings might apply to other Kv channels.

Author details

1. Luis G Cuello

   Center for Membrane Protein Research, Department of Cell Physiology and Molecular Biophysics, Texas Tech University Health Sciences Center, Lubbock, United States

   Contribution

   Conceptualization, Formal analysis, Supervision, Funding acquisition, Validation, Investigation, Writing—original draft, Project administration, Writing—review and editing

   For correspondence

   luis.cuello@ttuhsc.edu

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0001-5192-7568

2. D Marien Cortes

   Center for Membrane Protein Research, Department of Cell Physiology and Molecular Biophysics, Texas Tech University Health Sciences Center, Lubbock, United States

   Contribution

   Investigation, Methodology

   Competing interests

   No competing interests declared

3. Eduardo Perozo

   Department of Biochemistry and Molecular Biology, The University of Chicago, Chicago, United States

   Contribution

   Conceptualization, Resources, Data curation, Formal analysis, Supervision, Funding acquisition, Investigation, Methodology, Writing—original draft, Project administration, Writing—review and editing

   For correspondence

   eperozo@uchicago.edu

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0001-7132-2793

• 3,446

  Page views

• 692

  Downloads

• 65

  Citations

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