Bollepalli MK et al. (2014), State-dependent network connectivity determines...

XB-ART-50701

Structure 2014 Jul 08;227:1037-46. doi: 10.1016/j.str.2014.04.018.

Show Gene links Show Anatomy links

State-dependent network connectivity determines gating in a K+ channel.

Bollepalli MK , Fowler PW , Rapedius M , Shang L , Sansom MS , Tucker SJ , Baukrowitz T .

???displayArticle.abstract???
X-ray crystallography has provided tremendous insight into the different structural states of membrane proteins and, in particular, of ion channels. However, the molecular forces that determine the thermodynamic stability of a particular state are poorly understood. Here we analyze the different X-ray structures of an inwardly rectifying potassium channel (Kir1.1) in relation to functional data we obtained for over 190 mutants in Kir1.1. This mutagenic perturbation analysis uncovered an extensive, state-dependent network of physically interacting residues that stabilizes the pre-open and open states of the channel, but fragments upon channel closure. We demonstrate that this gating network is an important structural determinant of the thermodynamic stability of these different gating states and determines the impact of individual mutations on channel function. These results have important implications for our understanding of not only K+ channel gating but also the more general nature of conformational transitions that occur in other allosteric proteins.

???displayArticle.pubmedLink??? 24980796
???displayArticle.pmcLink??? PMC4087272
???displayArticle.link??? Structure
???displayArticle.grants??? [+]

Species referenced: Xenopus
Genes referenced: fubp1 kcnj1 kcnj6 ttn
GO keywords: ATP-sensitive potassium channel complex

???displayArticle.disOnts??? Bartter disease
???displayArticle.omims??? BARTTER SYNDROME, TYPE 2, ANTENATAL; BARTS2

???attribute.lit??? ???displayArticles.show???

	Graphical Abstract
	Figure 1. Mutations in Kir1.1 Preferentially Increase pH Sensitivity(A) The TMD and top half of the CTD were systematically mutated at a total of 187 positions (shown in magenta with a single monomer picked out in pink) and their pH0.5 values determined.(B) The F88A mutant substantially increases pH sensitivity compared to WT Kir1.1. Currents were recorded in giant excised patches over a range of pH values, allowing the pH0.5 value and Hill coefficient (n) to be measured. Data points represent mean ± SEM.(C) Mutations at 86 positions had relatively little effect on the pH0.5 value, leading to a shift of less than 0.4 pH units (green). Forty-seven mutants increased the pH0.5 by > 0.4 pH units (red), but only two mutants decreased the pH0.5 value by < 0.4 pH units (blue). No measurement of channel activity was possible for mutants at 52 positions (Table S1).(D) Mutations for which pH0.5 > pH 6.8 (red) and pH0.5 < pH 6.0 (blue) are mapped onto a single monomer (shown in gray) of a Kir1.1 closed-state model. †Indicates the coexpression with WT mRNA in a 1:1 ratio used to rescue the functional expression.See also Figure S1 and Movie S1.
	Figure 2. Gating-Sensitive Residues Interact in a State-Dependent Network(A) The size of the largest cluster formed by the 47 pH0.5 > pH 6.8 residues determined in the closed, pre-open, open-Kir3.2, and open-KirBac-EM models of Kir1.1. Residues interactions were defined as described in the Supplemental Experimental Procedures by using a probe radius of 1.0 Å.(B) Distributions in cluster size obtained for the indicated Kir1.1 models. The largest cluster in each model is indicated by an asterisk. Randomly selected residues exhibit clusters of ≤ 10 in all models (see also Figure S3).(C–F) The size of the largest cluster of pH0.5 > pH 6.8 interacting residues plotted against the probe radius used to define a packing interaction in the indicated Kir1.1 models (colored squares); notice that no large clusters are formed in the closed and open-KirBac-EM models residues, but large clusters appear as the probe radius increases to 0.8–1.0 Å in the pre-open and open-state models (see also Figure S4). Also shown in gray squares is the average size of the largest cluster for an ensemble of 100 models where the 47 positions were chosen randomly out of the 187 investigated positions; notice no larger clusters appear in any of the Kir1.1 models (see also Figure S5).
	Figure 3. Assembly of a Large Gating Network in the Pre-Open and Open States(A) All 47 identified pH0.5 > 6.8 residues are shown on the bottom axis, and those that belong to a network of intermediate or large size are identified by colored squares.(B) The same residues that form the networks are mapped onto the appropriate structural model of Kir1.1 by using the same coloring scheme. Other pH0.5 > 6.8 residues are shown in dark gray. In the closed model there are five networks of intermediate size: one involving the G loop of each monomer (shown in red) and four identical clusters that connect the TM1, TM2, and slide helices (only one is shown for clarity, in pink). As the CTD moves upward, these smaller clusters fuse together in the pre-open model, forming a single large network spanning all four monomers (blue). This connects the TMD to the G loop and the CTD, and almost all these residues remain connected in the open state (green).See also Movie S2 and Figure S6.
	Figure 4. Mutant Cycle Analysis Reveals Long-Range Thermodynamic Coupling(A) Location of network mutations L89A, Y79A (TM1), and S305A (G loop) in the open-Kir3.2 model.(B) The pH sensitivity of WT and indicated mutants. Calculated dose-response curves for double and triple mutants are also shown assuming no thermodynamic coupling (see Experimental Procedures). Data points represent mean ± SEM.(C) Thermodynamic coupling between indicated mutations was determined as the difference between the calculated ΔΔGCalc and experimental ΔΔGExpt values for indicated double and triple mutants. Note that thermodynamic coupling increases with the addition of L89A on Y79A-S305A. The T71A mutant is used as a control because this mutation is not within the network and exhibits WT pH sensitivity. The T71A-S305A double mutant exhibits no coupling.
	Figure 5. Intersubunit Interactions Affect Subunit CooperativityAsp67, Trp69, and Thr70 on the slide helix and Leu220 on the CD loop of the CTD all shift the pH0.5 > 6.8 and also significantly reduce the Hill coefficient (Δn > −1). They are separated by a large distance in the closed model (A), but come together in the open-Kir3.2 model (B) to form a connection between the TMD and the CTD. For context, the large networks are shown as in Figure 4B. Leu220 is indicated by an asterisk to denote this residue belongs to an adjacent subunit and represents part of an intersubunit interaction (see also Figure S1).
	Figure 6. Impact of a Network Mutation on the Gating of Kir1.1Cartoon depicting the assumed free energies of the different conformational states of Kir1.1 at pH 6.5, i.e., the equilibrium point where the open and close states are about equally populated (bottom row). The gating network residues are highlighted in green (open state), in blue (pre-open state), and in red (the closed state). Note that the gating network in the closed state is fragmented into smaller clusters. Mutation of a gating network residue (red dot) will have a local destabilizing effect (local destabilization, orange arrows) on all states. However, it will have a larger effect on the open and pre-open states due to greater connectivity of the gating network (network destabilization, yellow arrows) in these states compared to the closed state. This raises the free energy of the pre-open and open states relative to the closed state and leads to a redistribution in the relative population of states. The closed state now becomes more energetically favorable and therefore more frequently populated, thereby explaining the increased pH sensitivity observed for mutations within this network.

References [+] :

Alam, High-resolution structure of the open NaK channel. 2009, Pubmed