ATP-sensitive potassium (KATP) channels are widely expressed and play key roles in many tissues by coupling metabolic state to membrane excitability. The SUR subunits confer drug and enhanced nucleotide sensitivity to the pore-forming Kir6 subunit, and so information transfer between the subunits must occur. In our previous study, we identified an electrostatic interaction between Kir6 and SUR2 subunits that was key for allosteric information transfer between the regulatory and pore-forming subunit. In this study, we demonstrate a second putative interaction between Kir6.2-D323 and SUR2A-Q1336 using patch clamp electrophysiological recording, where charge swap mutation of the residues on either side of the potential interaction compromise normal channel function. The Kir6.2-D323K mutation gave rise to a constitutively active, glibenclamide and ATP-insensitive KATP complex, further confirming the importance of information transfer between the Kir6 and SUR2 subunits. Sensitivity to modulators was restored when Kir6.2-D323K was co-expressed with a reciprocal charge swap mutant, SUR-Q1336E. Importantly, equivalent interactions have been identified in both Kir6.1 and Kir6.2 suggesting this is a second important interaction between Kir6 and the proximal C terminus of SUR.

Introduction

The ATP-sensitive potassium (KATP) channel forms a sub-group of the inward-rectifying potassium ion channel family that are voltage-insensitive and ubiquitously expressed. The key roles for KATP channels include energy-sparing in cardiac and skeletal muscle [1,2], neuroprotection [3,4], regulation of insulin secretion from pancreatic β-cells [5], and regulation of vascular tone [6]. KATP channels were once front-line targets used in the treatment for type-II diabetes [5], hypertension or angina [6–8], indicating that they represent clinically important and ‘druggable’ proteins.

KATP channels are unique amongst the inwardly rectifying K+ channel family due to the requirement of a large accessory protein for its surface expression. The channel is a complex formed from a Kir6 tetramer co-assembled with a tetramer of sulphonylurea receptor (SUR) proteins that form a heteroctameric functional complex [9,10]. Kir6 subunits have intracellular N- and C-termini and two transmembrane segments M1 and M2, separated by a P-loop that forms the outer mouth of the weakly rectifying K+ selective pore [5]. The SUR subunits have extracellular N- and intracellular C-termini and 17 transmembrane segments.

Due to a C-terminal endoplasmic reticulum retention sequence (RKR), the Kir6 cannot express at the membrane surface without the SUR interaction, which presumably masks this motif. There are two mammalian Kir6 family members, Kir6.1 and Kir6.2. The Kir6.2 pore-forming subunit has an ATP binding site on the C terminus that inhibits the current through the weakly rectifying K+ selective pore [11]. Kir6.2 can functionally express at the cell surface if the subunit is truncated at the C-terminal 26 or 36 residues (Kir6.2ΔC26, Kir6.2ΔC36) and shows current that is ATP sensitive, albeit with a 10-fold right-shifted IC50 of ∼120 µM [12]. Similar Kir6.1 truncations still allow the protein to traffic to the membrane, however, the recorded current is very small, not metabolically sensitive and can only be inhibited by the Kir6.1 pore blocker, PNU37883A (Kir6.1ΔC48 and Kir6.1ΔN13/ΔC48) [13].

Co-assembly of Kir6.2 with an SUR enhances the ATP sensitivity of the complex, with reports in the literature showing an ATP-sensitivity in Kir6.2/SUR2A (the cardiac and skeletal muscle isoform) of ∼20 µM [14–17]. In addition to the enhanced ATP sensitivity, the SUR imparts ADP and nucleotide diphosphate (NDP) sensitivity to the current [18–20]. Furthermore, the accessory subunit imparts sulphonylurea drug sensitivity to the channel which includes activators, such as pinacidil [14,21], and inhibitors, such as glibenclamide [14,22,23]. This, and given the ubiquity of expression, led to an interest in KATP as a pharmacologically useful therapeutic target that continues to this day.

The Kir6.1 containing channel complex shows markedly different properties to the Kir6.2. This isoform, most often associated with the vasculature, is reported to be relatively ATP-insensitive and dependent on NDP's for its activity [13,24]. The Kir6.1/SUR2B channel complex is also highly regulated by protein kinase activity in the vasculature, which is attributed to phosphorylation of either the SUR or Kir6 component of the channel complex [25–27]. The fact that not all modulators of the full complex act via the pore-forming subunit led to the hypothesis that there must be regions of interaction between the pore forming and β-subunits that allow the exchange of information between the two subunits, i.e. binding of a drug to the SUR causing an allosteric modification to the gating of the Kir6 pore.

Our first publication in this area identified a 64-amino acid sequence, between residues 1294 and 1358 on the SUR2 subunit, that was a key region of interaction between the Kir6 and SUR2 family, as identified in co-immunoprecipitation experiments [15]. When co-expressed with full length Kir6.2/SUR2A subunits, this small protein fragment was able to act like a dominant negative on channel function by disrupting functional expression [15,28]. Our subsequent manuscript demonstrated that within this region of SUR2 there was a minimally interacting fragment of the C terminus (1318–1337), and a corresponding region of interaction within the C-terminal 75 residues of Kir6.2-(316–390) [14]. Two putative residues of interaction on the Kir6.2 C terminus were identified (D323 and K338), and 3 on the SUR C-terminal region (E1318, K1322 and Q1336) (Figure 1). Co-immunoprecipitation experiments, and electrophysiological studies, demonstrated that the K338 and E1318 residues were able to interact, demonstrated by performing charge swap experiments where a Kir6.2-K338E mutation, when co-expressed with SUR2A-E1318R, fully restored function to the channel, including pinacidil, glibenclamide, ATP and ADP sensitivity [14]. Corresponding charge swap mutations in Kir6.1 (R347E), also fully restored pinacidil and glibenclamide sensitivity. These data demonstrated that the Kir6.2-K338 (Kir6.1-R374) formed a salt bridge with the SUR2A-E1318 residue to allow the transfer of information between the subunits. Disruption of this salt bridge with a mutation in either the pore-forming subunit, or the accessory subunit, disrupted the normal channel function [14]. The recent publication of cryo-electron microscopy structures for the Kir6.2/SUR1 complex have given us new information regarding the structure and putative interactions between the subunits [29–32].

Co-immunoprecipitation experiments in our previous study demonstrated that there was a second potential salt bridge between Kir6.2-D323 and either SUR2A-K1332 or Q1336 [14]. In this study, we used electrophysiological measurements of membrane currents to test the hypothesis that there was a salt bridge between putative interacting residues. To investigate this, a charge swap at residue Kir6.2-D323 was used to disrupt the putative interaction, and the salt bridge ‘restored' by co-expressing a reciprocal charge swap on the SUR2A-K1332 or -Q1336 to investigate the transmission of regulatory information between subunits.

Methods

Cell culture

Human embryonic kidney (HEK)-293 cells were transiently transfected with both Kir6.2 and SUR2A subunits, and mutant subunits as indicated, 48 h before patch-clamp investigations. Cells were released from the plate by a 2 min trypsin digestion (0.5 g/l trypsin, 0.2 g/l EDTA; Sigma), followed by repeated washing with sterile 2 mM Ca2+ Tyrode's solution (2CaT) (5 mM KCl, 135 mM NaCl, 0.33 mM NaH2PO4, 5 mM sodium pyruvate, 5 mM glucose, 10 mM HEPES, 2 mM CaCl2 and 1 mM MgCl2), adjusted to pH 7.4 with NaOH. Cells were maintained in 2CaT at room temperature until use. For excised inside-out patch recording, cells were cultured and transfected on coverslips 48 h before use. Patches were excised from the cells already adhered to the coverslip.

Molecular biology

Point mutants were produced by overlap PCR with inserts confirmed by DNA sequencing. For electrophysiology, Kir6.0 wild-type (WT) and mutants were expressed in pcDNA3.1/myc/hisA (Invitrogen) and SUR2A WT and mutants expressed in pIRES2-EGFP-F [14,15].

Patch clamp

For conventional whole-cell recording from HEK293 cells, electrodes were pulled from thick-walled filamented borosilicate glass to a resistance of 3–6 MΩ. Cells were continuously perfused at 32 ± 2°C at a rate of 5 ml min−1 with 2CaT solution. Pinacidil and glibenclamide were added as indicated in each figure. Currents were recoded using an Axopatch 200B amplifier, digitised using a Digidata 1440 interface and recorded to computer using pCLAMP10.7. Transfected cells were identified by EGFP-F fluorescence at 488 nm under mercury lamp illumination. Cells were voltage-clamped at 0 mV (approximate EK for these solutions would be ∼−89 mV and so a large outward current was recorded). Membrane potentials were recorded for each cell by switching to current clamp prior to adding pinacidil. For excised inside-out patch recording, experiments were carried out at room temperature with the excised patch placed directly into the perfusion flow [14,15]. Pipette and perfusing solutions both contained 140 mM K+ so the pipette potential was held at 0 mV and currents recorded with a hyperpolarising step to −80 mV to avoid rectification seen with outward KATP currents.

Data analysis

All patch clamp data was recorded and analysed in pCLAMP10.7, and further analysed in Microsoft Excel 2016. Figures were prepared, and statistical analysis was carried out, in Graphpad Prism 7. Data is presented in bar charts as mean ± S.D. with the individual data points superimposed; n is reported as a number of cells.

Modelling putative interactions

Figures 11(C–G) and 12(A–C) were created using Pymol software. The cryo-EM structure of human pancreatic KATP channel (Kir6.2/SUR1) in the propeller conformation (PDB code:6C3P [32]) has been used for Figure 12.

The distance between D323 and R248 was measured using the measurement tool in Pymol. This measurement was done before (Figure 12B) and after modelling the D323K point mutation (Figure 12C). Virtual D323K point mutation was carried out using the mutagenesis tool in Pymol. In this mutagenesis, the most common rotamer of Lysine was chosen, indicated as a percentage associated with each potential conformation of Lysine in the protein, calculated by Pymol. These percentage values are based on the commonality of all the possible conformations of residues in proteins.

Results

Kir6.2-D323K co-expressed with WT-SUR2A results in a constitutively active current

In our previous study, the point mutation Kir6.2-K338E caused a 10-fold leftwards shift in the EC50 for pinacidil, a 100-fold rightwards shift in the IC50 for glibenclamide and small rightwards shift in ATP sensitivity when co-expressed with WT SUR2A [14]. The Kir6.2-D323K mutation was, however, somewhat more severe in terms of its phenotype when co-expressed with WT SUR2A, showing significant constitutive activity (Figure 2B,C) that was unresponsive to glibenclamide in whole-cell recording (Figure 2B,C), and unresponsive to ATP or ADP in excised inside-out patch recording (Figure 2F). The confirmation of a functional potassium current expression was demonstrated by measuring the resting membrane potential of the cells co-expressing Kir6.2-D323K/SUR2A-WT and showing a mean resting membrane potential of −73.4 ± 4.1 mV (n = 6) (Figure 2D,E).

Biochemical evidence [14] indicates that the Kir6.2-D323K mutation could co-immunoprecipitate with either the SUR2A-K1322D or -Q1336E charge swaps, so suggesting an interaction. It was therefore hypothesised that expression of the SUR2A mutants with a WT Kir6.2 would also impair normal current modulation by sulphonylurea drugs and ATP. Co-expression of Kir6.2-WT and SUR2A-K1322D resulted in a very limited current that was not constitutively active, was still modulated by sulphonylureas and could be potentiated by metabolic inhibition of the HEK293 cells (2 mM cyanide and 1 mM iodoacetic acid) (Figure 3B). The resting membrane potential of the Kir6.2-WT/SUR2A-K1322D expressing cells was not different from that of KATP-WT (Figure 3C). These data show that the Kir6.2-WT/SUR2A-K1322D complex is able to form functional channels, however there was sufficient disruption of the channel that function was very limited.

Co-expression of Kir6.2-WT with the SUR2A-Q1336E mutant yielded a more robust current density in the HEK293 cells (Figure 4B) with no significant change in basal constitutive current or resting membrane potential in cells compared with those expressing KATP-WT (Figure 4B,C). Kir6.2-WT/SUR2A-Q1336E channels showed a reduced sensitivity to glibenclamide compared with the KATP-WT complex (Figure 4D, 433 nM vs 4.3 nM, respectively). Additionally, the ATP sensitivity was right-shifted from 23.4 µM to 90 µM in Kir6.2-WT/SUR2A-Q1336E compared with control (Figure 4E). These data show that the Kir6.2-WT/SUR2A-Q1336E complex forms a functional channel, however some of the key modulatory characteristics have been lost.

Kir6.2-D323K co-expressed with SUR2A-Q1336E, but not SUR2A-K1322D, partly restores the functional characteristics of the KATP complex

Co-immunoprecipitation data from our previous publication shows that the charge swap mutation in Kir6.2-D323K restores interaction with SUR2A-K1322D or SUR2A-Q1336E [14]. To confirm whether this charge swap could restore the normal regulation of channel gating by the SUR2A subunit to the constitutively active Kir6.2-D323K mutant, this mutant was co-expressed with either SUR2A-K1322D or SUR2A-Q1336E. The Kir6.2-D323K/SUR2A-K1322D combination showed constitutive activity and a hyperpolarised membrane potential with no effect of sulphonylurea drug or metabolic inhibition of the current (Figure 5B,C). Furthermore, this current was not ATP sensitive (Figure 5D). The Kir6.2-D323K/SUR2A-Q1336E subunit combination still showed some limited constitutive activity as evident by the slightly hyperpolarised membrane potential (Figure 6B), however also showed responsiveness to pinacidil, metabolic inhibition and to glibenclamide (Figure 6C). The Kir6.2-D323K/SUR2A-Q1336E showed a full restoration of glibenclamide sensitivity, being indistinguishable from KATP-WT control (IC50 of 3.9 nM and 3.7 nM for KATP-WT and Kir6.2-D323K/SUR2A-Q1336E, respectively) (Figure 6D). Unlike the Kir6.2-D323K mutant expressed with SUR2A or SUR2A-K1322D mutant, the Kir6.2-D323K/SUR2A-Q1336E was ATP sensitive, however it was significantly right shifted (IC50 23 µM and 123 µM control and charge swap, respectively) (Figure 6E).

Co-expression of Kir6.2-D323K with a double mutant SUR2A-K1322D/Q1336E does not restore functional parameters of the WT KATP complex

It was hypothesised that as both SUR2A-K1322D and -Q1336E mutations co-immunoprecipitated with Kir6.2-D323K in our previous study [14], that perhaps Kir6.2-D323 interacted with both residues on the SUR2A subunit in the fully formed channel complex. To investigate this, a double SUR2A mutant was expressed in HEK293 cells with the Kir6.2-D323K mutant. Expression of Kir6.2-WT with the SUR2A-K1322D/Q1336E double mutant yielded a current that, although pinacidil and glibenclamide sensitive, was too small to record meaningful data from (Figure 7B,C). Co-expression of the Kir6.2-D323K/SUR2A-K1322D/Q1336E complex yielded a channel that was constitutively active, sulphonylurea drug insensitive and ATP insensitive (Figure 7B–E).

Co-expression of the Kir6.2-D323K equivalent mutation in Kir6.1 (E332K) with SUR2A-K1322D, but not Q1336E, restores glibenclamide sensitivity to a constitutively active Kir6.1

Kir6.1 is a markedly different pore forming subunit to Kir6.2 in that it shows little ATP-dependence, but that is highly regulated by kinases and ADP. These modulators are suggested to act via the SUR subunit and so a similar transfer of information must occur between these conserved interacting residues. This has been previously reported for a residue in Kir6.1-E347 where a mutant Kir6.1-E347R assembled with SUR2A-D1318E and restored sulphonylurea drug sensitivity [14]. In this current study, we identified that the Kir6.1-E332K mutant expressed with SUR2A-WT also yielded a constitutively active channel, like its Kir6.2-D323K counterpart, however showed right-shifted glibenclamide sensitivity rather than completely abolishing it Figure 8B–D). Co-expression of Kir6.1-E332K with SUR2A-K1322D restored glibenclamide sensitivity to that indistinguishable from Kir6.1-WT/SUR2A-WT, however co-expression of Kir6.1-E332K/SUR2A-Q1336E did not restore the glibenclamide sensitivity (Figure 8E–G). These data suggest that Kir6.1-E332K interacts with SUR2A-K1332 whereas Kir6.2-D323 interacts with SUR2A-Q1336.

Mutation of Kir6.2-D323 does not cause the pore-forming subunit to become constitutively active

To determine whether mutation of Kir6.2-D323 is the cause of the switch from ATP-sensitive to constitutively active, a modified Kir6.2ΔC26 truncation mutant was used. The deletion of the C-terminal 26 amino acid removes the RKR-endoplasmic reticulum retention sequence that is usually masked by co-assembly with the SUR accessory subunit. This Kir6.2ΔC26 truncation mutant can express without the SUR, and retains ATP sensitivity [12], albeit right shifted 10-fold as the presence of the SUR enhances the ATP sensitivity. Kir6.2ΔC26 truncation, or Kir6.2ΔC26-D323K, was expressed in HEK293 cells. Both were expressed, and both showed activation in response to ATP depletion with metabolic inhibition (CN and IAA). Both Kir6.2ΔC26 and Kir6.2ΔC26-D323K were insensitive to pinacidil or glibenclamide, as would be expected with no SUR co-expression, but neither channel showed constitutive activity in whole-cell recording (Figure 9B,C). Furthermore, both channels were ATP-sensitive with an IC50 of ∼123 µM (Figure 9D), comparable with previous literature for the Kir6.2 truncation, and right shifted ∼10-fold from the Kir6.2/SUR2A complex [12]. These findings suggest that the Kir6.2-D323K mutation does not itself cause the constitutive activity in the channel.

Co-expression of Kir6.2ΔC26-D323K with SUR2A-WT results in a channel that is constitutively active, as seen in the full length D323K mutant in Figure 2, however that was sensitive to glibenclamide (Figure 10). Given the instability of the current, and the lack of sensitivity to pinacidil to try to stabilise the opening, it was not possible to record a concentration response curve.

Data to this point has suggested that the SUR2A-Q1336 residue gives the closest to WT phenotype when co-expressed with the Kir6.2-D323K mutant. With this Kir6.2 mutant, even when co-expressed with the SUR2A-Q1336E, there was still constitutive activity. Co-expression of Kir6.2ΔC26-D323K/SUR2A-Q1336E resulted in channels that showed constitutive activity but retained pinacidil, metabolic inhibition and glibenclamide sensitivity (Figure 10B–D). The Kir6.2ΔC26-D323K/SUR2A-Q1336E complex also retained ATP sensitivity (Figure 10E), however it remained 10-fold right shifted suggesting that, despite the presence of the SUR accessory subunit, and this mutation did not restore the full ATP-sensitivity seen with the KATP-WT.

Discussion

The data from this study suggests that there is a physical interaction between residues D323 on Kir6.2 and Q1336 of SUR2A (Figure 11A). This interaction allows the transfer of information between the SUR and Kir6 subunits with respect to channel gating in the presence of sulphonylurea drugs, such as glibenclamide. Furthermore, we show that a conserved charge at the analogous position in Kir6.1 (E332K) interacts with SUR2A, however at an alternative residue than Kir6.2, SUR2A-K1322 (Figure 11B). These data presented here functionally characterise a further electrostatic interaction between the C-terminal domain of Kir6.2 and the nucleotide binding domain 2 in the proximal C-terminal region of the SUR2A subunit. As with our previously identified electrostatic interaction (Kir6.2-K338/SUR2A-E1318) [14], disruption of the pairing between Kir6.2-D323/SUR2A-Q1336 alters the regulation of the pore by ATP, glibenclamide and pinacidil, with the near complete restoration of function on repair of the electrostatic interaction by opposite charge swap mutations. This highlights the importance of these regions of the proteins for the interaction between the Kir6 and SUR subunits. The main functional data from the Figures 1 to 10 are summarised in Table 1.

This newly identified region of interaction between the C-terminal regions of Kir6 and SUR2 subunits is distinct from reported interactions, with alternative regions previously identified in other studies. Interaction between TMD0 and the M2 TMD of Kir6 subunits has previously been deemed important for the assembly of the channel [33]. Additionally, isoform-specific regulation by nucleotides of channel opening has also been linked to the L0 linking sequence between TMD0 and TMD1 in SUR subunits [33–35]. Here we report a second electrostatic interaction between a single residue (D323) on Kir6.2 at the C-terminal region and a corresponding charge partner (Q1336) on the C-terminal region of SUR2A.

Our earlier studies [14,15,36] focussed on this C-terminal interacting region as a point of allosteric information transfer between the heterologous subunits. There were changes in the functional parameters of the channel when the Kir6.2-K338/SUR2A-E1318 electrostatic interaction was broken, which were restored with a direct charge swap. Our data has been further strengthened with the recent Cryo-electron microscopy structures of Kir6.2 co-expressed with SUR1 [37]. In this study, the authors suggest that our identified Kir6.2-K338 residue comes into close proximity with the proposed charge partner, the equivalent of SUR2A-E1318, in one of the three conformations identified [37].

In this present study, it was hypothesised that Kir6.2-D323K/SUR2A-Q1336E charge swap would restore the function of the channel complex as the reciprocal charge swap for our previous pairing had shown [14]. Single charge mutations, such as Kir6.2-D323K co-expressed with SUR2A-WT, had the surprising effect of causing a constitutive activity in the normally electrically silent channel complex. The apparent lack of inhibition by ATP for this Kir6.2-D323K with the SUR2A-WT suggests that there has been either a disruption of the ATP-binding site, or a disruption of the interaction between Kir6.2 and SUR2A that modulates normal gating. Evidence for a disruption of gating rather than an alteration of the Kir6.2 ATP binding site comes from the Kir6.2ΔC26-D323K mutant where expression of this truncated and mutant pore-forming subunit yields a current that is not constitutively active and is ATP sensitive. Furthermore, co-expression with the charge swap SUR2A-Q1336E mutant fully restores channel function with the exception of the ATP-sensitivity, which remains at ∼120 µM consistent with the truncation of Kir6.2 expressed alone [12]. Perhaps somewhat unexpectedly, the Kir6.2ΔC26-D323K mutant when expressed with SUR2A-WT did form a channel showing substantial constitutive activity that was insensitive to pinacidil, however was fully inhibited by 10 µM glibenclamide. These findings suggest that this single point mutation triggers something deleterious to the formation of a normally functioning channel complex with the full length Kir6.2-D323K mutant compared with the truncated mutant, Kir6.2ΔC26-D323K, co-expressed with the SUR2A-WT. Perhaps the interaction with the RKR motif of the full length Kir6.2 protein has a constraining effect on the correct interaction between Kir6.2-D323 and SUR2A-Q1336.

Constitutive activity of Kir6.1-E332K, the analogous residue to Kir6.2-D323, was also seen when co-expressed with SUR2A-WT. In this case, the co-expression of Kir6.1-E332K with the SUR2A-K1322D, but not the SUR2A-Q1336E, fully restored glibenclamide sensitivity to the complex. These findings suggest that Kir6.1 and Kir6.2 interact at alternative SUR2A residues, K1332 and Q1336, respectively. The similarities in pharmacological response suggest that there is a common zone of interaction between SUR2A-E1318 and Q1336 and the proximal C terminus of the Kir6 pore forming subunits despite the regulatory differences in these proteins. It has been suggested that the relative insensitivity of Kir6.1 to ATP, that largely fail to open spontaneously in the absence of ATP, results from an enhanced response to stimulatory ATP binding to SUR rather than a reduced sensitivity to the inhibitory binding of ATP to the Kir6.1 subunit [24].

Dupuis et al. [38] used a similar approach to ours to identify three residues in SUR2A which, when changed to the equivalent residues in the non-interacting maltose binding protein (MBP1), were sufficient to attenuate the activation of the channel complex by ADP and potassium channel openers. Expanding this, Prinicpalli et al. demonstrated further evidence that the region identified by our previous studies in the SUR [14,15], included additional residues in SUR1 (Q1342, I1347, and L1350) that confer allosteric regulation on channel gating [39]. These studies add to the understanding of the complexity of the mechanisms behind allosteric information transfer between SUR and Kir6 subunits.

With the recent cryo-EM structure of the Kir6.2/SUR1, we have new insight into the folding and alignment of the subunits within the KATP channel complex [37]. In the three main structural patterns published, it is unlikely that our identified residues of interaction could form any permanent salt bridge interaction. The D323 residue is located on an outside face of the cytoplasmic domain that brings it within close proximity of a highly-conserved region within SUR subunit (the ‘lasso region' residues 193–261) [32,37], whilst the SUR2A-Q1336 equivalent in the SUR1 model is located on a region located in the lipid environment externally facing the channel complex (Figure 11C–E, identified as green). The SUR subunit in particular shows a significant degree of flexibility between the five published structures shown in Figure 11C–G, with an overlay of the five models shown in Figure 11H for comparison. This suggests that the interactions that we have described in this manuscript may be important for the correct folding between orientations rather than a key functional structure. Interestingly, the lasso region (193–261), which is in close proximity to D323 (Figure 12A), has been previously shown to be important in gain-of-function mutations in some diabetic states [32,40]. Through these mutations, the open-state of the KATP channel is stabilised and the ATP-inhibition of the channel is overcome, both of which lead to over-activity of the channel [32]. Examples of residues involved are R248 deletion (located on the Lasso extension) which led to immature SUR1 subunit formation and ultimately to hyperinsulinism of infancy and Y330 (located on the ATP binding pocket and therefore adjacent to the lasso extension) which led to permanent neonatal diabetes [40].

Given that our electrophysiology and previous co-immunoprecipitation data suggest some form of interaction is possible between these residues, we suggest that the degree of flexibility possible within the SUR2A subunit could allow the formation of a temporary interaction that perhaps stabilises a transition state between the activated and inactive conformations. Alternatively, this interaction may occur between adjacent Kir6.2-SUR heteroctoamers or an additional protein, as yet unidentified, may interact with the KATP complex and these residues form part of this interaction. It is also possible that this interaction represents conformations of Kir6.2-SUR2A that have yet to be visualised using cryo-EM as the current models are based on SUR1.

Interestingly, using the current cryo-EM structures, modelling of the D323K mutation changes the distance between the Kir6.2-D323 and the SUR-R248 residues from 7.7 Å to 3 Å (Figure 12B,C). We suggest that this may cause the change from a physiologically silent to constitutively active channel in the D323K mutant. Whilst D323 is not located immediately at the ATP-binding pocket on the Kir6.2 subunit, it is in close proximity to the lasso extension which plays important role in regulating the function and gating of KATP channels. Disruption of the lasso extension has been shown to cause significant conformational changes in KATP channel (propeller and quatrefoil forms) [32], and point mutations along the lasso extension as well as to residues adjacent to it have been observed to associate with gain-of-function of KATP channel [40,41]. Therefore, we suggest that the Kir6.2-D323K mutation elicits its effect via its disruptive influence on the lasso extension, so modulating ATP sensitivity, but perhaps fails to allow the full appropriate interaction with the SUR2A-Q1336E ‘equivalent’ charge swap during the transition between states.

In summary, an interface between heterologous KATP subunits was identified containing an electrostatic interaction between Kir6.2-D323 and SUR-Q1336 residues, and the equivalent residue on Kir6.1, Kir6.1-E332, and alternate SUR2A residue K1332. This interaction, together with our previously identified interaction between Kir6.2-K338 and SUR2A-E1318, are important sites of information transfer between heterologous KATP subunits determining sensitivity to allosteric regulators.

Abbreviations

     
  • HEK

    human embryonic kidney

  •  
  • MBP

    maltose binding protein

  •  
  • NDP

    nucleotide diphosphate

  •  
  • SUR

    sulphonylurea receptor

  •  
  • WT

    wild-type

Author Contribution

R.I.N., D.L. and R.D.R. designed the experiments which were carried out by R.D.R., S.B. and H.N.R. All authors contributed to data analysis and preparation of the figures. RDR and SB wrote the manuscript, R.I.N., D.L., H.N.R., R.R. and S.I. critically appraised and edited the text.

Funding

This work was supported by the British Heart Foundation [PG/16/14/32039] (S.B., R.D.R.), and studentship to and H.N.R. (T.H. Wathes Fund, University of Leicester). S.B. was also supported by the University of Liverpool.

Acknowledgements

We thank Sonja Khemiri, Barbara Horley and Dr Lorenza Francescut for expert technical assistance.

Open Access

Open access for this article was enabled by the participation of University of Liverpool in an all-inclusive Read & Publish pilot with Portland Press and the Biochemical Society under a transformative agreement with JISC.

Competing Interests

The authors declare that there are no competing interests associated with the manuscript.

References

References
1
Zingman
,
L.V.
,
Alekseev
,
A.E.
,
Hodgson-Zingman
,
D.M.
and
Terzic
,
A.
(
2007
)
ATP-sensitive potassium channels: metabolic sensing and cardioprotection
.
J. Appl. Physiol.
103
,
1888
1893
2
Jovanović
,
A.
and
Jovanović
,
S.
(
2009
)
SURA2 targeting for cardioprotection?
Curr. Opin. Pharmacol.
9
,
189
193
3
Sun
,
H.
and
Feng
,
Z.
(
2013
)
Neuroprotective role of ATP-sensitive potassium channels in cerebral ischemia
.
Acta Pharmacol. Sin.
34
,
24
32
4
Yamada
,
K.
and
Inagaki
,
N.
(
2005
)
Neuroprotection by KATP channels
.
J. Mol. Cell. Cardiol.
38
,
945
949
5
Bonfanti
,
D.H.
,
Alcazar
,
L.P.
,
Arakaki
,
P.A.
,
Martins
,
L.T.
,
Agustini
,
B.C.
,
de Moraes Rego
,
F.G.
et al (
2015
)
ATP-dependent potassium channels and type 2 diabetes mellitus
.
Clin. Biochem.
48
,
476
482
6
Aziz
,
Q.
,
Thomas
,
A.M.
,
Gomes
,
J.
,
Ang
,
R.
,
Sones
,
W.R.
,
Li
,
Y.
et al (
2014
)
The ATP-sensitive potassium channel subunit, Kir6.1, in vascular smooth muscle plays a major role in blood pressure control
.
Hypertension
64
,
523
529
7
Pan
,
Z.
,
Huang
,
J.
,
Cui
,
W.
,
Long
,
C.
,
Zhang
,
Y.
and
Wang
,
H.
(
2010
)
Targeting hypertension with a new adenosine triphosphate-sensitive potassium channel opener iptakalim
.
J. Cardiovasc. Pharmacol.
56
,
215
228
8
Ito
,
N.
,
Nanto
,
S.
,
Doi
,
Y.
,
Kurozumi
,
Y.
,
Natsukawa
,
T.
,
Shibata
,
H.
, et al (
2013
)
Beneficial effects of intracoronary nicorandil on microvascular dysfunction after primary percutaneous coronary intervention: demonstration of its superiority to nitroglycerin in a cross-over study
.
Cardiovasc. drugs Ther.
27
,
279
287
9
Foster
,
M.N.
and
Coetzee
,
W.A.
(
2016
)
KATP channels in the cardiovascular system
.
Physiol. Rev.
96
,
177
252
10
Aguilar-Bryan
,
L.
,
Clement
,
J.P.
,
Gonzalez
,
G.
,
Kunjilwar
,
K.
,
Babenko
,
A.
and
Bryan
,
J.
(
1998
)
Toward understanding the assembly and structure of KATP channels
.
Physiol. Rev.
78
,
227
245
11
Vanoye
,
C.G.
,
MacGregor
,
G.G.
,
Dong
,
K.
,
Tang
,
L.
,
Buschmann
,
A.S.
,
Hall
,
A.E.
et al (
2002
)
The carboxyl termini of K(ATP) channels bind nucleotides
.
J. Biol. Chem.
277
,
23260
23270
12
Tucker
,
S.J.
,
Gribble
,
F.M.
,
Zhao
,
C.
,
Trapp
,
S.
and
Ashcroft
,
F.M.
(
1997
)
Truncation of Kir6.2 produces ATP-sensitive K+ channels in the absence of the sulphonylurea receptor
.
Nature
387
,
179
183
13
Farzaneh
,
T.
and
Tinker
,
A.
(
2008
)
Differences in the mechanism of metabolic regulation of ATP-sensitive K+ channels containing Kir6.1 and Kir6.2 subunits
.
Cardiovasc. Res.
79
,
621
631
14
Lodwick
,
D.
,
Rainbow
,
R.D.
,
Rubaiy
,
H.N.
,
Al Johi
,
M.
,
Vuister
,
G.W.
and
Norman
,
R.I.
(
2014
)
Sulfonylurea receptors regulate the channel pore in ATP-sensitive potassium channels via an intersubunit salt bridge
.
Biochem. J.
464
,
343
354
15
Rainbow
,
R.D.
,
James
,
M.
,
Hudman
,
D.
,
Al Johi
,
M.
,
Singh
,
H.
,
Watson
,
P.J.
et al (
2004
)
Proximal C-terminal domain of sulphonylurea receptor 2A interacts with pore-forming Kir6 subunits in KATP channels
.
Biochem. J.
379
,
173
181
16
Babenko
,
A.P.
,
Gonzalez
,
G.
,
Aguilar-Bryan
,
L.
and
Bryan
,
J.
(
1999
)
Sulfonylurea receptors set the maximal open probability, ATP sensitivity and plasma membrane density of KATP channels
.
FEBS Lett.
445
,
131
136
17
Takano
,
M.
,
Xie
,
L.H.
,
Otani
,
H.
and
Horie
,
M.
(
1998
)
Cytoplasmic terminus domains of Kir6.x confer different nucleotide-dependent gating on the ATP-sensitive K+ channel
.
J. Physiol.
512
,
395
406
18
Matsuo
,
M.
,
Tanabe
,
K.
,
Kioka
,
N.
,
Amachi
,
T.
and
Ueda
,
K.
(
2000
)
Different binding properties and affinities for ATP and ADP among sulfonylurea receptor subtypes, SUR1, SUR2A, and SUR2B
.
J. Biol. Chem.
275
,
28757
28763
19
Babenko
,
A.P.
,
Aguilar-Bryan
,
L.
and
Bryan
,
J.
(
1998
)
A view of sur/KIR6.X, KATP channels
.
Annu. Rev. Physiol.
60
,
667
687
20
Ueda
,
K.
,
Inagaki
,
N.
and
Seino
,
S.
(
1997
)
MgADP antagonism to Mg2+-independent ATP binding of the sulfonylurea receptor SUR1
.
J. Biol. Chem.
272
,
22983
6
21
Shindo
,
T.
,
Yamada
,
M.
,
Isomoto
,
S.
,
Horio
,
Y.
and
Kurachi
,
Y.
(
1998
)
SUR2 subtype (A and B)-dependent differential activation of the cloned ATP-sensitive K+ channels by pinacidil and nicorandil
.
Br. J. Pharmacol.
124
,
985
991
22
Gribble
,
F.M.
and
Reimann
,
F.
(
2003
)
Sulphonylurea action revisited: the post-cloning era
.
Diabetologia
46
,
875
891
23
Findlay
,
I.
(
1993
)
Sulphonylurea drugs no longer inhibit ATP-sensitive K+ channels during metabolic stress in cardiac muscle
.
J. Pharmacol. Exp. Ther.
266
,
456
467
PMID:
[PubMed]
24
Babenko
,
A.P.
and
Bryan
,
J.
(
2001
)
A conserved inhibitory and differential stimulatory action of nucleotides on K(IR)6.0/SUR complexes is essential for excitation-metabolism coupling by K(ATP) channels
.
J. Biol. Chem.
276
,
49083
49092
25
Light
,
P.E.
,
Bladen
,
C.
,
Winkfein
,
R.J.
,
Walsh
,
M.P.
and
French
,
R.J.
(
2000
)
Molecular basis of protein kinase C-induced activation of ATP-sensitive potassium channels
.
Proc. Natl Acad. Sci. U.S.A.
97
,
9058
9063
26
Jun
,
J.Y.
,
Kong
,
I.D.
,
Koh
,
S.D.
,
Wang
,
X.Y.
,
Perrino
,
B.A.
,
Ward
,
S.M.
et al (
2001
)
Regulation of ATP-sensitive K(+) channels by protein kinase C in murine colonic myocytes
.
Am. J. Physiol. Cell Physiol.
281
,
C857
C864
27
Thorneloe
,
K.S.
,
Maruyama
,
Y.
,
Malcolm
,
A.T.
,
Light
,
P.E.
,
Walsh
,
M.P.
and
Cole
,
W.C.
(
2002
)
Protein kinase C modulation of recombinant ATP-sensitive K(+) channels composed of Kir6.1 and/or Kir6.2 expressed with SUR2B
.
J. Physiol.
541
,
65
80
28
Rainbow
,
R.D.
,
Norman
,
R.I.
,
Hudman
,
D.
,
Davies
,
N.W.
and
Standen
,
N.B.
(
2005
)
Reduced effectiveness of HMR 1098 in blocking cardiac sarcolemmal K(ATP) channels during metabolic stress
.
J. Mol. Cell. Cardiol.
39
,
637
646
29
Li
,
N.
,
Wu
,
J.-X.
,
Ding
,
D.
,
Cheng
,
J.
,
Gao
,
N.
and
Chen
,
L.
(
2017
)
Structure of a pancreatic ATP-sensitive potassium channel
.
Cell
168
,
101
110.e10
30
Martin
,
G.M.
,
Kandasamy
,
B.
,
DiMaio
,
F.
,
Yoshioka
,
C.
and
Shyng
,
S.-L.
(
2017
)
Anti-diabetic drug binding site in a mammalian KATP channel revealed by cryo-EM
.
eLife
6
,
e31054
31
Martin
,
G.M.
,
Yoshioka
,
C.
,
Rex
,
E.A.
,
Fay
,
J.F.
,
Xie
,
Q.
,
Whorton
,
M.R.
et al (
2017
)
Cryo-EM structure of the ATP-sensitive potassium channel illuminates mechanisms of assembly and gating
.
eLife
6
,
e24149
32
Lee
,
K.P.K.
,
Chen
,
J.
and
MacKinnon
,
R.
(
2017
)
Molecular structure of human KATP in complex with ATP and ADP
.
eLife
6
,
e32481
33
Babenko
,
A.P.
and
Bryan
,
J.
(
2003
)
Sur domains that associate with and gate KATP pores define a novel gatekeeper
.
J. Biol. Chem.
278
,
41577
41580
34
Chan
,
K.W.
,
Zhang
,
H.
and
Logothetis
,
D.E.
(
2003
)
N-terminal transmembrane domain of the SUR controls trafficking and gating of Kir6 channel subunits
.
EMBO J.
22
,
3833
3843
35
Babenko
,
A.P.
and
Vaxillaire
,
M.
(
2011
)
Mechanism of KATP hyperactivity and sulfonylurea tolerance due to a diabetogenic mutation in L0 helix of sulfonylurea receptor 1 (ABCC8)
.
FEBS Lett.
585
,
3555
3559
36
Rainbow
,
R.D.
,
Lodwick
,
D.
,
Hudman
,
D.
,
Davies
,
N.W.
,
Norman
,
R.I.
and
Standen
,
N.B.
(
2004
)
SUR2A C-terminal fragments reduce KATP currents and ischaemic tolerance of rat cardiac myocytes
.
J. Physiol.
557
,
785
794
37
Puljung
,
M.C.
(
2018
)
Cryo-electron microscopy structures and progress toward a dynamic understanding of KATP channels
.
J. Gen. Physiol.
150
,
653
669
38
Dupuis
,
J.P.
,
Revilloud
,
J.
,
Moreau
,
C.J.
and
Vivaudou
,
M.
(
2008
)
Three C-terminal residues from the sulphonylurea receptor contribute to the functional coupling between the K(ATP) channel subunits SUR2A and Kir6.2
.
J. Physiol.
586
,
3075
3085
39
Principalli
,
M.A.
,
Dupuis
,
J.P.
,
Moreau
,
C.J.
,
Vivaudou
,
M.
and
Revilloud
,
J.
(
2015
)
Kir6.2 activation by sulfonylurea receptors: a different mechanism of action for SUR1 and SUR2A subunits via the same residues
.
Physiol. Rep.
3
,
e12533
40
Lang
,
V.
and
Light
,
P.E.
(
2010
)
The molecular mechanisms and pharmacotherapy of ATP-sensitive potassium channel gene mutations underlying neonatal diabetes
.
Pharmgenomics. Pers. Med.
3
,
145
161
41
Edghill
,
E.L.
,
Flanagan
,
S.E.
and
Ellard
,
S.
(
2010
)
Permanent neonatal diabetes due to activating mutations in ABCC8 and KCNJ11
.
Rev. Endocr. Metab. Disord.
11
,
193
198

Author notes

*

These authors contributed equally to this work.

This is an open access article published by Portland Press Limited on behalf of the Biochemical Society and distributed under the Creative Commons Attribution License 4.0 (CC BY). Open access for this article was enabled by the participation of University of Liverpool in an all-inclusive Read & Publish pilot with Portland Press and the Biochemical Society under a transformative agreement with JISC.