Polarity of the ATP binding site of the Na+,K+-ATPase, gastric H+,K+-ATPase and sarcoplasmic reticulum Ca2+-ATPase

BBA - Biomembranes 1862 (2020) 183138

Contents lists available at ScienceDirect

BBA - Biomembranes journal homepage: www.elsevier.com/locate/bbamem

Polarity of the ATP binding site of the Na+,K+-ATPase, gastric H+,K+ATPase and sarcoplasmic reticulum Ca2+-ATPase K.R. Hossaina, X. Lia, T. Zhangc, S. Paulac,1, F. Corneliusd, R.J. Clarkea,b,

T

⁎

a

School of Chemistry, University of Sydney, Sydney, NSW 2006, Australia The University of Sydney Nano Institute, Sydney, NSW 2006, Australia c Department of Chemistry, Purdue University, West Lafayette, IN 47907, USA d Department of Biomedicine, University of Aarhus, DK-8000 Aarhus C, Denmark b

A R T I C LE I N FO

A B S T R A C T

Keywords: Na+,K+-ATPase Ca2+-ATPase H+,K+-ATPase Eosin N-terminus Docking calculations

A fluorescence ratiometric method utilizing the probe eosin Y is presented for estimating the ATP binding site polarity of P-type ATPases in different conformational states. The method has been calibrated by measurements in a series of alcohols and tested using complexation of eosin Y with methyl-β-cyclodextrin. The results obtained with the Na+,K+-, H+,K+- and sarcoplasmic reticulum Ca2+-ATPases indicate that the ATP binding site, to which eosin is known to bind, is significantly more polar in the case of the Na+,K+- and H+,K+-ATPases compared to the Ca2+-ATPase. This result was found to be consistent with docking calculations of eosin with the E2 conformational state of the Na+,K+-ATPase and the Ca2+-ATPase. Fluorescence experiments showed that eosin binds significantly more strongly to the E1 conformation of the Na+,K+-ATPase than the E2 conformation, but in the case of the Ca2+-ATPase both fluorescence experiments and docking calculations showed no significant difference in binding affinity between the two conformations. This result could be due to the fact that, in contrast to the Na+,K+- and H+,K+-ATPases, the E2-E1 transition of the Ca2+-ATPase does not involve the movement of a lysine-rich N-terminal tail which may affect the overall enzyme conformation. Consistent with this hypothesis, the eosin affinity of the E1 conformation of the Na+,K+-ATPase was significantly reduced after N-terminal truncation. It is suggested that changes in conformational entropy of the N-terminal tail of the Na+, K+- and the H+,K+-ATPases during the E2-E1 transition could affect the thermodynamic stability of the E1 conformation and hence its ATP binding affinity.

1. Introduction

or Post-Albers cycle. In all cases the energy required for ion pumping is derived from the hydrolysis of ATP. A key feature of the cycle is the existence of two major conformational forms of the enzyme, which can exist in both unphosphorylated and phosphorylated states, E1 and E2 (or E1P and E2P in the case of the phosphorylated states). These two forms differ significantly in their ion binding affinities. In the case of the Na+,K+-ATPase, the E1 form binds preferentially Na+ ions to its transport sites, whereas the E2 form is stabilized by the binding of K+ ions. In the case of the H+,K+-ATPase, the E1 form binds H+ ions instead of Na+. In the SR Ca2+-ATPase, the E1 form binds Ca2+ ions, and the E2 form is thought to bind H+ ions [4]. Because the transition from the E2 to the E1 form is an essential step in the mechanism of all P-type ATPases to allow ion pumping to occur, experimental techniques which are able to detect this transition and to determine its rate are highly desirable. A technique which has proved to be very useful is the monitoring of the fluorescence of the probe eosin Y

P-type ATPases play crucial roles in cell physiology. The Na+ electrochemical potential gradient created across the plasma membrane of animal cells by the Na+,K+-ATPase (or sodium pump) is used to drive nutrient reabsorption in kidney, and both the Na+ and K+ gradients it generates are essential for nerve and muscle function [1]. The closely related enzyme, the H+,K+-ATPase (or proton pump) of gastric parietal cells creates the low pH of the stomach necessary for digestion [2]. Another P-type ATPase, the sarco(endo)plasmic reticulum Ca2+ATPase (SR calcium pump or SERCA) transports Ca2+ ions out of the cytoplasm into the sarcoplasmic reticulum, which allows muscle relaxation [3]. All of these ion pumps undergo a cyclic series of reaction steps involving conformational changes, phosphorylation by ATP, dephosphorylation and ion binding and release, which is often referred to as the E1-E2 cycle or, in the case of the Na+,K+-ATPase, the Albers-Post ⁎

Corresponding author at: School of Chemistry, University of Sydney, Sydney, NSW 2006, Australia. E-mail address: [email protected] (R.J. Clarke). 1 Present address: Department of Chemistry, California State University at Sacramento, Sacramento CA 95819, USA. https://doi.org/10.1016/j.bbamem.2019.183138 Received 24 September 2019; Received in revised form 15 November 2019; Accepted 27 November 2019 Available online 29 November 2019 0005-2736/ © 2019 Elsevier B.V. All rights reserved.

BBA - Biomembranes 1862 (2020) 183138

K.R. Hossain, et al.

0.9 mM EDTA. The protein concentration was 4 mg ml−1, determined by the Peterson modification [19] of the Lowry method [20] using bovine serum albumin as a standard. Preparation of pig gastric H+,K+-ATPase-enriched membrane vesicles was performed as described by Sachs et al. [21] with some modifications [22]. The protein concentration was 12.8 mg ml−1, determined by the Peterson modification [19] of the Lowry method [20], using bovine serum albumin as a standard. The specific ATPase activity of the preparation used was 23 μmol ATP hydrolysed h−1 (mg of protein)−1, which was measured by the method of Baginski et al. [23] as described previously [24]. Preparation of rabbit skeletal muscle sarco/endoplasmic reticulum Ca2+-ATPase microsomes was performed as described previously [25,26]. Briefly, 120 g of thawed rabbit muscle tissue was cut into small pieces, removing any connective tissue. The muscle tissue was then mixed with 340 ml of buffer containing 10 mM MOPS, 0.3 M sucrose, 0.1 mM EDTA (pH 7.0). The mixture was homogenized in a blender for 15 s in 5 min intervals over 1 h during which the pH of the solution was readjusted to 7.0 using 0.1 M NaOH solution between homogenisations. The homogenized solution was then centrifuged at 15,000g for 20 min in a JLA-16.250 rotor (Beckman Coulter, Lane Cove, Australia) using an Avanti J-E high-speed centrifuge (Beckman Coulter). After this the supernatant was filtered through several layers of cheesecloth to remove any remaining fat and centrifuged at 40,000g for 90 min using a JA-20 rotor (Beckman Coulter). The pellet was collected and resuspended using a glass homogenizer in 40 ml of buffer containing 10 mM MOPS, 0.6 M KCl (pH 7.0) and incubated for 40 min at 4 °C. The suspension was centrifuged at 15,000g for 20 min using a JA-20 rotor (Beckman Coulter); after which the top 10% of the supernatant and the pellet were discarded and the rest of the supernatant again centrifuged at 40,000g for 90 min. The pellet was collected and the microsomes were suspended with a glass homogenizer in 5 ml of a solution containing 10 mM MOPS, 0.9 M sucrose (pH 7.0), and stored in 1 ml aliquots at −80 °C. The protein concentration was 10.5 mg ml−1, determined by the Peterson modification [19] of the Lowry method [20], using bovine serum albumin as a standard. The specific ATPase activity of the preparation used was 305 μmol ATP hydrolysed h−1 (mg of protein)−1, measured by the pyruvate kinase/lactate dehydrogenase coupled assay at 37 °C [27] as described previously [28].

Fig. 1. Chemical structure of eosin Y in its doubly deprotonated form.

(hereafter simply referred to as eosin) (see Fig. 1). This probe was first introduced to the field by Skou and Esmann [5,6], but it has since been used in many studies on both the Na+,K+-ATPase [7–11] and the H+,K+-ATPase [12] by several research groups and has also been applied in studies of the plasma membrane Ca2+-ATPase (PMCA) [13] and the ADP/ATP carrier of mitochondrial membranes [14]. The reason for its wide use is that it is capable of binding to the nucleotide binding site of these enzymes, although in the case of the PMCA it was found that eosin inhibition was not competitive with ATP, suggesting an alternative binding site in the case of this enzyme [15,16]. In virtually all the ion pump studies [5–7,10,12,13] the change in fluorescence intensity of eosin has been used to monitor shifts between the E1 and E2 state. However, recently a ratiometric method was developed based on the wavelength shift in the eosin excitation spectrum [11]. Although the intensity-based studies have delivered much useful information, the ratiometric approach offers the additional advantage of probing the environment of the binding site and gaining information on its polarity. In this paper we report a calibration of eosin's fluorescence excitation spectrum and provide quantitative information on the polarity of ATP binding sites of the Na+,K+-ATPase, H+,K+-ATPase and the sarcoplasmic reticulum Ca2+-ATPase under different experimental conditions. Because the cytoplasmic extramembrane lysine-rich N-terminal tail of the Na+,K+-ATPase has been implicated in determining the distribution of this enzyme between the E1 and E2 states [11], we have also carried out eosin experiments using N-terminally truncated enzyme. Comparison of these results with those of native Na+,K+-ATPase and H+,K+-ATPase (which also possesses a lysine-rich N-terminal tail) and with those of the Ca2+-ATPase (which lacks a lysine-rich N-terminal tail) supports the conclusion that the N-terminus of the Na+,K+and H+,K+-ATPases does play a key role in determining the conformation of these enzymes.

2.2. Proteolytic cleavage of the Na+,K+-ATPase Specific cleavage of the N-terminus of the Na+,K+-ATPase α-subunit was carried out as described previously [29]. Briefly, membranebound Na+,K+-ATPase was incubated with trypsin, from bovine pancreas, at a trypsin to protein weight ratio of 1:50 for 10 min on ice in the presence of 20 mM histidine, 130 mM NaCl and 1 mM EDTA (pH 7.0). The reaction was then stopped by the addition of soy bean trypsin inhibitor at twice the amount of trypsin and subsequently washed three times via ultracentrifugation (75,000g for 1 h at 24 °C) using an Optima XE-100 ultracentrifuge and SW 32 Ti Swinging-Bucket Rotor (Beckman Coulter) to remove any residual trypsin and trypsin inhibitor. The protein contents before and after N-terminal truncation were determined by the Peterson modification [19] of the Lowry method [20], using bovine serum albumin as a standard. The specific ATPase activities before and after N-terminal truncation were determined by the pyruvate kinase/lactate dehydrogenase coupled assay at 37 °C [27] as described previously [28]. At least 99% of the activity could be blocked by the addition of 50 μM ouabain, a specific inhibitor of the Na+,K+ATPase. The specific ATPase activity of the preparation after N-terminal truncation was 870 μmol ATP hydrolysed h−1 (mg of protein)−1 (in comparison to a value of 1400 before truncation). The observed rate constants, kobs, of the E1Na+ 3 → E2P transition before and after Nterminal truncation of Na+,K+-ATPase were measured in E1 buffer (30 mM imidazole, 130 mM NaCl, 5 mM MgCl2, 1 mM EDTA, pH 7.2) using the fluorescent probe RH421 in combination with the stopped-

2. Materials and methods 2.1. Enzymes Na+,K+-ATPase-containing membrane fragments from pig kidney outer medulla were purified as described by Klodos et al. [17]. The specific ATPase activity at 37 °C and pH 7.0 was measured according to Ottolenghi [18]. The activity of the preparation used was 1400 μmol ATP hydrolysed h−1 (mg of protein)−1 at saturating substrate concentrations in a buffer containing 20 mM histidine, 250 mM sucrose and 2

BBA - Biomembranes 1862 (2020) 183138

K.R. Hossain, et al.

flow method via an SF-61SX2 stopped-flow spectrofluorimeter (TgK Scientific, Bradford on Avon, UK) as described previously [28,30–32]. For the N-truncated enzyme, stabilized in E1, kobs was 48 ± 5 s−1, which was significantly lower than the values for Na+,K+-ATPase before truncation of 119 ± 8 s−1. These values of kobs agree well with those previously reported by Cornelius et al. [29] for the enzyme before and after truncation. The position of trypsin cleavage of the N-terminus is referred to as the T2 site and has been determined to be between lysine-30 (30K) and glutamic acid-31 (31E) of the conserved 28L29K30K31E motif [33]. The sequence numbering is based here on the pig sequence (Sus scrofa).

E1-stabilizing buffer composed of 30 mM imidazole, 130 mM NaCl, 5 mM MgCl2, 1 mM EDTA (pH 7.2) and an E2-stabilizing buffer composed of 30 mM imidazole, 10 mM KCl and 1 mM EDTA (pH 7.2) were used for both the Na+,K+-ATPase and N-truncated Na+,K+-ATPase measurements, with the buffer compositions being chosen based on previously published partial reaction data [28,34]. For the H+,K+ATPase, the E1-stabilizing buffer contained 25 mM imidazole, 5 mM MgCl2, 1 mM EDTA (pH 8.5) and the E2-stabilizing buffer contained 5 mM imidazole, 240 mM KCl and 1 mM EDTA (pH 8.5). These buffer compositions were chosen based on measurements of Diller et al. [35]. For SERCA, the E1-stabilizing buffer was composed of 25 mM imidazole, 50 mM KCl, 2 mM CaCl2, 5 mM MgCl2, 2 mM EDTA (pH 7.0) and the E2-stabilizing buffer contained 25 mM imidazole, 50 mM KCl, 2 mM EDTA (pH 7.0). The SERCA buffer compositions were based on results reported by Inesi et al. [4].

2.3. Materials The origins of the reagents used were: imidazole (≥99%, Sigma, Castle Hill, Australia), tris(hydroxymethyl)aminomethane (99%, Alfa Aesar, Heysham, UK), eosin Y (C.I. 45,380, BDH, Kilsyth, Australia), trypsin from bovine pancreas, Type XI, lyophilized powder, (≥6000 BAEE units/mg protein) (≥99%, Sigma, Castle Hill, Australia), soy bean trypsin inhibitor (Roche Diagnostics Australia, North Ryde, Australia), methyl-β-cyclodextrin (mβCD) (cell-culture-tested, 1.5–2.1 methyl per mol glucose, Sigma), NaCl (suprapure, Merck, Kilsyth, Australia), KCl (analytical grade, Merck), L-histidine (99%, Sigma), MgCl2∙6H2O (analytical grade, Merck), EDTA (99%, Sigma), HCl (0.1 N Titrisol solution, Merck), 4-morpholinepropanesulfonic acid (MOPS) (99%, Sigma), sucrose (99%, Ajax Finechem (Australia), ATP disodium∙3H2O salt (special quality, Roche), NADH disodium salt (approximately 100%, grade 1, Roche), phospho(enol)pyruvic acid cyclohexylammonium (≥97%, Sigma), pyruvate kinase (PK)/lactate dehydrogenase (LDH) from rabbit muscle (900–1400 units/ml of LDH, 600–1000 units/ml of PK, Sigma), ouabain octahydrate (Sigma), NaOH (analytical grade, Merck), HCl (0.1 N Titrisol® solution, Merck), ethanol (99.7%, Merck), methanol (99%, Redox Chemicals, Wetherill Park, Australia), 1-propanol (96%, Ajax Finechem, Scoresby, Australia), 1butanol (99.5%, Ajax Finechem), 1-pentanol (98%, Ajax Finechem) and 1-hexanol (98%, Merck). N-(4-Sulfobutyl)-4-(4-(dipentylamino)phenyl) butadienyl)-pyridinium inner salt (RH421) was obtained from Molecular Probes (Eugene, OR) and was used without further purification. RH421 was added to Na+,K+-ATPase-containing membrane fragments from an ethanolic stock solution. The dye spontaneously partitions into the membrane fragments.

2.5. Computational docking The structure of the eosin dianion (see Fig. 1) was modelled in MOE (Molecular Open Environment, version 2018.01; Chemical Computing Group, Montreal, Canada) and its conformational energy minimized using Molecular Mechanics with the MMFF94s force field. The following protein crystal structures for Na+,K+-ATPase (pig) and SERCA (rabbit) were obtained from the Protein Databank: Na+,K+ATPase/E2, PDB ID: 3B8E [36]; SERCA/E1, PDB ID: 1SU4 [37]; SERCA/E2, PDB ID: 2AGV [38]. All entries other than the protein entity (the α subunit in the case of Na+,K+-ATPase) were deleted and the structures were prepared for docking using AutoDockTools (version 1.5.6, The Scripps Research Institute) [39]. A docking box size of 18 Å3 with a grid point separation of 1 Å proved to be sufficiently large to fully accommodate the eosin molecule when probing the ATP binding sites. The centre of the box was positioned at the oxygen atom of the ribose ring of the respective (deleted) nucleotide derivative present in the crystal structure. To facilitate a convenient comparison of the docking results, the sequences of the N-domains of all protein structures were aligned and their structures superimposed on top of each other. Docking runs were executed at the default settings of AutoDock Vina using the highest level of exhaustiveness [40]. 2.6. Amino acid sequence analysis Sequences of the main catalytic α1 subunit of the Na+,K+-ATPase, the gastric H+,K+-ATPase and the sequence of the sarcoplasmic reticulum Ca2+-ATPase (all from Homo sapiens) were obtained from the protein database of the National Center for Biotechnology Information (https://www.ncbi.nlm.nih.gov/protein/). The sequences were aligned using the MUSCLE program [41] within the MEGA7 suite of evolutionary genetics programs [42].

2.4. Eosin fluorescence measurements All fluorescence spectral measurements were conducted with an RF5301 PC spectrofluorophotometer (Shimadzu, Kyoto, Japan) using 1 cm pathlength quartz semimicrocuvettes (for measurements of eosin in solvents) or microcuvettes (for measurements in the presence of proteins). All fluorescence excitation spectra were measured at an emission wavelength, λem, of 555 nm (bandwidth 5 nm) with an OG530 cutoff filter (Schott, Mainz, Germany) in front of the photomultiplier. At each of the excitation wavelengths, 490 nm and 538 nm, the apparent background fluorescence was subtracted before calculating the fluorescence ratio, R = F490/F538. The background level was determined at an excitation wavelength of 400 nm, at which eosin doesn't undergo excitation. For the eosin experiments in which the probe was titrated with different P-type ATPases, the titrations were carried out as follows. 99 μL of either E1-stabilizing buffer or E2-stabilizing buffer for each of the different P-type ATPases were added to the cuvette followed by 1 μL of eosin (2.9 μM in water) and titrated with protein up to a concentration of 400 μg/ml by consecutive additions of protein to the cuvette. The total final volume of protein suspension added was never ˃10 μL (i.e., 110 μL total volume in the cuvette at the end of the titration) and the measured fluorescence intensities were corrected for the increasing volume in the cuvette over the course of the titration. An

3. Results 3.1. Calibration of eosin excitation spectra in alcohols It has already been reported that the absorbance and fluorescence emission spectra of eosin undergo a red shift, i.e. to longer wavelengths, with decreasing solvent polarity [43]. In order to utilise the fluorescence excitation spectrum of eosin as a quantitative probe of the polarity of the ATP-binding pocket of P-type ATPases, it is first necessary to measure spectra in media of known polarity. For this we have utilized a homologous series of alcohols, i.e. water, methanol, ethanol, propanol, butanol, pentanol and hexanol. In addition, mixtures of water and methanol of compositions 20, 40 and 70% w/w were used to obtain solutions with polarities between that of pure water and pure methanol, and expected dielectric constants for these mixtures were obtained from the literature [44]. To avoid any effects of eosin dimerization on the observed spectra [45] a low eosin concentration of 29 nM was used for 3

BBA - Biomembranes 1862 (2020) 183138

K.R. Hossain, et al.

Fig. 2. Normalized fluorescence excitation spectra of eosin (29 nM) in water (blue), methanol (red), ethanol (magenta), propanol (green), butanol (black), pentanol (violet), and hexanol (orange). The emission wavelength was 555 nm (+OG530 cutoff filter). The bandwidths for both excitation and emission were 5 nm. The vertical dotted lines at 490 nm and 538 nm have been included to highlight the two wavelengths, either side of the fluorescence maxima, which were used to define the fluorescence excitation ratio, R. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

1.2 1.0 0.8 0.6

F/Fmax 0.4 0.2 0.0 480

500

520

540

Excitation wavelength / nm Fig. 3. Effect of solvent polarity, as expressed by the dielectric constant, on the fluorescence excitation ratio, R, of eosin (29 nM). R is defined as the fluorescence intensity ratio using excitation wavelengths of 490 and 538 nm, i.e. R = F490/ F538, at an emission wavelength of 555 nm. The individual points at increasing dielectric constants are joined by straight lines to aid the eye of the reader. Each point is the average of 3 measurements. The errors in R are smaller than the size of the symbols used for the points and hence the error bars are not visible.

1.4 1.2 1.0 0.8

R 0.6 0.4 0.2 0.0 10

20

30

40

50

60

70

80

90

Dielectric constant all measurements. Normalized fluorescence excitation spectra in a range of solvents are shown in Fig. 2. There the red shift of the fluorescence excitation spectrum of eosin with decreasing polarity of the solvent (from water to hexanol) can be clearly seen. Rather than quantify the shift via the change in excitation wavelength maximum, we prefer to use a more sensitive measure of the shift, i.e., the fluorescence excitation ratio R = F490/F538 (as described under Materials and Methods). The change in R with increasing dielectric constant, ε, of the solvent is shown in Fig. 3. It can be seen that the dependence of R on ε is not completely monotonic. In particular, there appears to be a jump in the R value on going from pure ethanol (ε = 24.3) to pure methanol (ε = 32.6). This could possibly be explained by a contribution from more specific hydrogen bonding interaction between the solvents and eosin, rather than a pure polarity or dipole-dipole interaction. For this reason we have not attempted to fit a particular function form to the data. Based on this calibration, measurements of eosin's R value in different local environments should allow the polarity of the environment to be estimated. As a test of the method we have chosen a simple model system, namely methyl-β-cyclodextrin (mβCD), a cyclic oligosaccharide possessing a central hydrophobic cavity to which halogenated xanthene dyes similar to eosin are known to bind and form an inclusion complex [45].

3.2. Polarity of the methyl-β-cyclodextrin cavity Titration of eosin with increasing concentrations of mβCD in aqueous solution causes a gradual red shift of the probe's fluorescence excitation spectrum and a decrease in the R value, as more eosin binds to the mβCD cavity (see Fig. 4). To quantify the polarity of the mβCD cavity we have fitted a hyperbolic binding curve to the measured data, as described by the following equation:

c ⎞ R = R 0 + (R∞ − R 0)⋅⎛ K + c⎠ ⎝

(1)

where R0 represents the R value in the absence of any added mβCD, R∞ represents the value at an infinite mβCD concentration where all eosin molecules would be expected to be bound within the mβCD cavity and K is the half-saturating concentration of mβCD. The polarity of the mβCD cavity can be estimated from the value of R∞ of 0.20 ( ± 0.01). Based on the calibration shown in Fig. 3, this corresponds to a local dielectric constant of approximately 18, i.e., similar to that of butanol. Previous studies using other probes [45,46] have indicated a cyclodextrin cavity microenvironment close to that of ethanol (ε ≈ 24). The slightly lower polarity estimated here could simply be due to our extrapolation to infinite mβCD concentration. The lowest experimentally determined R value at 200 mM mβCD of 0.26, corresponds to an environment close to that of ethanol (ε ≈ 24). Thus, there is reasonably good agreement with previous studies indicating a dielectric constant of 4

BBA - Biomembranes 1862 (2020) 183138

K.R. Hossain, et al.

Fig. 4. Effect methyl-β-cyclodextrin concentration on the fluorescence excitation ratio, R, of eosin (29 nM). The solid line represents a fit to the data of a hyperbolic binding curve (see Eq. 1). The fit parameters determined were: R0 = 1.17 ( ± 0.01), R∞ = 0.20 ( ± 0.01), and K = 16.4 ( ± 0.9) mM. All measurements were performed in MilliQ water. Each point represents an average of four measurements and the errors bars show the standard deviation.

1.4 1.2 1.0 0.8

R 0.6 0.4 0.2 0.0 0

50

100

150

200

[CD] / mM

with a much lower affinity to the E2 state [47,48]. If one makes the assumption that only one eosin molecule binds per protein, the halfsaturating protein concentrations can be converted to dissociation constants by dividing by the protein molecular weight. Utilizing a value of 147,000 g mol−1 for an αβ complex of the Na+,K+-ATPase [49], yields Kd values of 1.1 ( ± 0.2) μM and 0.17 ( ± 0.01) μM for the E2 and E1 states, respectively. The value for the E1 state is very similar to values reported for ATP itself of 0.12–0.63 μM [47,48]. The value for the E2 state suggests that eosin binds more strongly to the E2 state than ATP (for which Kd values in the range 71–450 μM have been reported; [47,48]). The fitted curves indicate that although binding of eosin to the protein is much weaker in E2 than in E1, the polarity of the binding site is very similar in both states. In E1-buffer a value for R∞ of 0.368 ( ± 0.008) and in E2-buffer a value of 0.30 ( ± 0.06) was calculated. Based on the calibration (see Fig. 3) this corresponds to a local dielectric constant of approximately 25, i.e., quite similar to that of ethanol, or slightly more polar than that of the mβCD cavity.

the cyclodextrin cavity in the region of 20. The eosin ratiometric method described here can, thus, be used with a certain degree of confidence in the analysis of the polarity of ATP binding sites of ATPases. 3.3. Eosin binding to the Na+,K+-ATPase Similar to the results obtained with mβCD, titration of eosin with Na+,K+-ATPase-containing membrane fragments causes a significant drop in the R value (see Fig. 5). The drop can be reversed by the addition of ATP (in the absence of Mg2+ in the buffer to avoid phosphorylation), as one would expect for eosin binding to the enzyme's ATP binding site. However, the concentration dependence of the drop depends strongly on the buffer composition, which controls the enzyme's conformational state. As described under Materials and Methods, we have utilized two different buffer compositions which are known to specifically stabilise the E1 (Na+-selective) and E2 (K+-selective) conformational states. The results (see Fig. 5) indicate that a significantly higher protein concentration is required to observe saturation in eosin binding to the protein when it is in the E2 state than in the E1 state. Fitting Eq. (1) to the observed data yielded a half-saturating protein concentration of 160 ( ± 30) μg/ml in an E2-stabilizing buffer, whereas in an E1-stabilizing buffer the value was only 26 ( ± 2) μg/ml. This behaviour parallels that of the enzyme's natural substrate, ATP, which is known to bind with a high affinity to the E1 state, but

3.4. Eosin binding to the gastric H+,K+-ATPase and sarcoplasmic reticulum Ca2+-ATPase For comparisons with the Na+,K+-ATPase, titrations of eosin have also been carried out with two other P-type ATPases, i.e., the gastric H+,K+-ATPase and the sarcoplasmic reticulum Ca2+-ATPase. The Fig. 5. Titration of eosin (29 nM) with Na+,K+-ATPase-containing membrane fragments in either an E1-stabilizing buffer (solid spheres) or an E2-stabilizing buffer (solid squares). The solid lines represents fits to the data of a hyperbolic binding curve (see Eq. (1)). For E1-stabilizing conditions the fit parameters determined were: R0 = 1.04 ( ± 0.01), R∞ = 0.368 ( ± 0.008), and K = 26 ( ± 2) μg/ml. For E2-stabilizing conditions the fit parameters determined were: R0 = 1.02 ( ± 0.02), R∞ = 0.30 ( ± 0.06), and K = 160 ( ± 30) μg/ml. Each point is an average of three separate samples, with error bars indicating the standard deviation.

1.2 1.0 0.8 R

E2

0.6

E1 0.4 0.2 0

100

200 +

300

400

+

[Na ,K -ATPase] / ug/ml

5

BBA - Biomembranes 1862 (2020) 183138

K.R. Hossain, et al.

Fig. 6. Titration of eosin (29 nM) with either the H+,K+-ATPase (above) or the sarcoplasmic reticulum Ca2+-ATPase (below). In both cases solid spheres represent the results obtained in an E1-stabilizing buffer and solid squares represent the results obtained in an E2-stabilizing buffer. The solid lines represents fits to the data of a hyperbolic binding curve (see Eq. (1)). For the H+,K+-ATPase the fit parameters determined were R0 = 1.13 ( ± 0.02), R∞ = 0.38 ( ± 0.01), and K = 11 ( ± 1) μg/ml for the E1 state, and R0 = 1.19 ( ± 0.02), R∞ = 0.25 ( ± 0.02), and K = 47 ( ± 5) μg/ml for the E2 state. For the Ca2+-ATPase the fit parameters determined were R0 = 1.057 ( ± 0.008), R∞ = 0.212 ( ± 0.005), and K = 24 ( ± 1) μg/ml for the E1 state, and R0 = 1.03 ( ± 0.01), R∞ = 0.201 ( ± 0.006), and K = 19 ( ± 1) μg/ml for the E2 state. Each point is an average of three separate samples, with error bars indicating the standard deviation.

1.2 +

+

H ,K -ATPase

1.0 0.8 R

E2 0.6 E1

0.4 0.2 0

100

200 +

300

400

+

[H ,K -ATPase] / ug/ml

1.2 1.0

2+

Ca -ATPase

0.8 R

0.6 E1 0.4 E2 0.2 0

100

200

300

400

2+

[Ca -ATPase] / ug/ml

suggest a slightly lower polarity in this state, with dielectric constant closer to 20. Titrations of eosin with the sarcoplasmic reticulum Ca2+-ATPase are shown in the lower panel of Fig. 6. In contrast to the results obtained with the Na+,K+- and H+,K+-ATPases, although eosin still binds to the Ca2+-ATPase, there appears to be very little difference in the binding affinity between the E1 and E2 states. The values of K determined were 24 ( ± 1) μg/ml in E1 and 19 ( ± 1) μg/ml in E2. Based on the molecular weight of the Ca2+-ATPase of 110,458 g mol−1 [49] and a single eosin molecule binding per protein, the corresponding Kd values are 0.22 ( ± 0.01) μM and 0.17 ( ± 0.01) μM in the E1 and E2 states, respectively. If anything, the affinity of eosin to the protein appears to be slightly higher in E2 than in E1, i.e., the opposite of the situation with the Na+,K+- and H+,K+-ATPases. The polarity of the binding site also seems to be lower in the Ca2+-ATPase than in the Na+,K+- and H+,K+-ATPases. The values of the R∞ determined for the Ca2+-ATPase were 0.212 ( ± 0.005) and 0.201 ( ± 0.006) for the E1 and E2 states respectively. This suggests a local binding site dielectric constant of approximately 18 in both states. A possible explanation for the significant differences in behaviour between the Ca2+-ATPase and both the Na+,K+- and H+,K+-ATPases could be that they are related to the proteins' N-termini. Both the Na+,K+-ATPase and the H+,K+-ATPase have extended extramembrane lysine-rich N-termini on the cytoplasmic side of membrane [51]. From tryptic digestion patterns [33,49,52,53] and experiments in which the N-terminus was removed either by proteolytic cleavage [29,33] or by

results obtained are shown in Fig. 6. For both enzymes a significant drop in the eosin R value was observed with increasing enzyme concentration, and, as in the case of the Na+,K+-ATPase, the addition of ATP to a suspension of eosin saturated with either the H+,K+-ATPase or the sarcoplasmic reticulum Ca2+-ATPase caused an increase in the R value and a shift of the spectrum back towards that of eosin in aqueous solution. This is again consistent with eosin binding to the ATP binding site of both enzymes. However, significant differences between the enzymes are obvious when one compares the observed concentration dependencies of the titrations under E1- and E2-promoting conditions. The results obtained with the H+,K+-ATPase are very similar to those obtained with the Na+,K+-ATPase (see Fig. 5). This is not surprising, because the H+,K+-ATPase is the most closely related ion pump to the Na+,K+-ATPase in the entire P-type ATPase family. In both cases eosin binds significantly more strongly to the E1 state than the E2 state. For the H+,K+-ATPase the half-saturating protein concentrations, K, in the E1 and E2 states were 11 ( ± 1) μg/ml and 47 ( ± 5) μg/ml, respectively. Based on the molecular weight of an αβ complex of the H+,K+-ATPase of 147,346 g mol−1 [50] and again assuming a single eosin molecule binds per protein complex, the halfsaturating protein concentrations correspond to dissociation constants, Kd, of 0.075 ( ± 0.007) μM and 0.32 ( ± 0.03) μM in the E1 and E2 states, respectively. In the E1 state the R∞ of 0.38 ( ± 0.01) is also very similar to that of the Na+,K+-ATPase, indicating a local binding site dielectric constant of approximately 25. In the E2 state, the value of R∞ determined for the H+,K+-ATPase was 0.25 ( ± 0.02), which would

6

BBA - Biomembranes 1862 (2020) 183138

K.R. Hossain, et al.

Fig. 7. Titration of eosin (29 nM) with native and N-terminally truncated Na+,K+-ATPase in either an E1-stabilizing (above) or E2-stabilizing buffer (below). In both cases solid symbols represent the results obtained for the native enzyme and open symbols represent the results obtained for N-terminally truncated enzyme. Solid lines represents fits to the native enzyme data of a hyperbolic binding curve (see Eq. (1)), whereas dashed lines represent corresponding fits for the N-terminally truncated enzyme. For enzyme in E1-stabilizing buffer the fit parameters determined were R0 = 1.04 ( ± 0.01), R∞ = 0.368 ( ± 0.008), and K = 26 ( ± 2) μg/ml for the native enzyme, and R0 = 1.06 ( ± 0.02), R∞ = 0.19 ( ± 0.04), and K = 81 ( ± 11) μg/ml for the N-terminally truncated enzyme. For the enzyme in E2-stabilizing buffer the fit parameters determined were R0 = 1.02 ( ± 0.02), R∞ = 0.30 ( ± 0.06), and K = 160 ( ± 30) μg/ml for the native enzyme, and R0 = 1.078 ( ± 0.009), R∞ = 0.15 ( ± 0.04), and K = 190 ( ± 20) μg/ml for N-terminally truncated enzyme. For the non-truncated enzyme, each point is an average of three separate samples, with error bars indicating the standard deviation. For the truncated enzyme, each point represents an average of three repeat measurements on the same sample. The error bars shown represent relative errors of 10%, which was estimated based on the measurements of non-truncated enzyme.

1.2 1.0 0.8 R

E1 truncated

0.6 E1 0.4 0.2 0

100

200 +

300

400

300

400

+

[Na ,K -ATPase] / ug/ml

1.2 1.0

E2 truncated

0.8 E2 R

0.6 0.4 0.2 0

100

200 +

+

[Na ,K -ATPase] / ug/ml

affinity when the enzyme is in the E2 state. The half-saturating concentration, K, was 160 ( ± 30) μg/ml for the native enzyme and 190 ( ± 20) μg/ml for the N-terminal truncated enzyme, which correspond to Kd values of 1.1 ( ± 0.2) μM and 1.3 ( ± 0.1) μM for the native and the truncated enzymes, respectively. Thus, the results support the conclusion that the major reason for the higher binding affinity of eosin to the enzyme in the E1 state over the E2 state is an N-terminus-associated phenomenon specific to the E1 state.

mutagenesis [54], it is known that the N-terminus undergoes significant movement during the E2-E1 conformational transition. In contrast, the extramembrane N-terminus of the sarcoplasmic reticulum Ca2+-ATPase is much shorter. To test whether or not the extended N-terminus of the Na+,K+-ATPase could account for the different behaviour, we have, therefore, carried out titrations of eosin with Na+,K+-ATPase from which the N-terminus has been cleaved by controlled trypsin treatment. 3.5. Eosin binding to N-terminally truncated Na+,K+-ATPase

3.6. Docking results Comparisons of titrations of eosin with native and N-terminally truncated Na+,K+-ATPase are shown in Fig. 7. There it can be seen that N-terminal truncation significantly reduces the affinity of eosin for the protein when it is in the E1 state. The half-saturating concentration, K, increases from 26 ( ± 2) μg/ml for the native enzyme to 81 ( ± 11) μg/ ml for the N-terminal truncated enzyme. The corresponding Kd values, calculated in the same way as in previous sections, are 0.55 ( ± 0.08) μM and 0.17 ( ± 0.01) μM for the truncated and non-truncated enzymes, respectively. Thus, N-terminal truncation shifts the affinity towards the value obtained in the E2 conformation. This result would seem to support the suggestion made in the previous section, that the lack of any significant difference in eosin binding affinity between the E1 and E2 states of the Ca2+-ATPase is due to this enzyme's much shorter extramembrane N-terminus. In contrast to the E1 results of the Na+,K+-ATPase, N-terminal truncation appears to have no significant effect on the eosin binding

To visualize eosin binding to Na+K+ATPase and SERCA at the molecular level, we performed docking simulations with the enzymes for which high-resolution X-ray crystal structures were available. Only the top-ranked solution from each docking run in AutoDock Vina was considered for further analysis. The docking-predicted binding affinity of eosin for SERCA was slightly higher in the enzyme's E2 conformation (−7.5 kcal mol−1 or − 31 kJ mol−1) than in its E1 conformation (−7.1 kcal mol−1 or − 30 kJ mol−1). An inspection of the docked eosin poses revealed that eosin was buried more deeply in the binding pocket if the enzyme was in the E2 conformation (see Fig. 8), allowing for π/π-stacking interactions with Phe 487 as well as a hydrogen bond with Arg 678, which was absent in the E1 conformation. Another indicator of the extent of interactions taking place was the size of the protein surface area in direct proximity of the ligand, which was larger in the E2 than in the E1 conformation (4242 Å2 versus 2471 Å2). 7

BBA - Biomembranes 1862 (2020) 183138

K.R. Hossain, et al.

Fig. 8. Top row: Eosin docked into the E1 conformation (left) and E2 conformation (right) of SERCA. Non-polar surface areas are coloured in green, polar areas are coloured in purple. Bottom row: Docking-predicted ligand/receptor interactions and solvent exposure (areas shaded in blue). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

to the cytoplasm [48,55,56]. In contrast, the affinities of eosin towards the E1 and E2 states of the sarcoplasmic reticulum Ca2+-ATPase were very similar (see Fig. 6), with, if anything, a slightly higher affinity towards the E2 state than the E1 state. As far as we are aware, this would also seem to be consistent with what is known about the interaction of ATP with the Ca2+-ATPase. In the case of this enzyme, the transition from the E2 to the E1 state has been reported to occur via a spontaneous deprotonation of the E2 state [57,58], with no reports of any involvement of ATP in the transition or of any significant change in ATP affinity. The experiments reported here have focussed solely on unphosphorylated E1 and E2 forms of the ATPases concerned. However, in principle the ratiometric eosin method described here may also be applicable to investigating the polarity of the ATP binding site of phosphorylated forms of these enzymes, i.e. E1P and E2P. Because eosin binds to the ATP binding site of the E1 and E2 states and blocks ATP phosphorylation, any investigation of the E1P and E2P states would require phosphorylation by ATP before the addition of eosin or the use of ATP and phosphate analogues [13]. In an attempt to account for the experimentally determined binding affinities of eosin for the Na+,K+-ATPase and SERCA at the molecular level, we performed computational docking simulations. These were limited to the cases for which high resolution X-ray crystal structures were available, namely for SERCA in the E1 and E2 conformations and for the Na+,K+-ATPase in the E2 conformation. The lower affinity of

Consistent with the character of the eosin molecule, the contact surface areas in both enzyme conformations were amphiphilic, with 57% and 54% of the surface area being non-polar in E1 and E2, respectively (see Fig. 8; top row). A comparison of the eosin/E2 complex of Na+,K+-ATPase with that of SERCA revealed that the dye bound in the same location, but in a somewhat shifted, more solvent-exposed position (see Fig. 9). This resulted in a smaller contact area (2919 Å2) in comparison to SERCA (3965Å2). Consistent with the less favourable ligand binding pose in the E2 form of Na+,K+-ATPase, the docking-predicted binding affinity dropped by about one order of magnitude, with a calculated interaction energy of −6.4 kcal mol−1 or − 27 kJ mol−1. 4. Discussion Eosin has been shown here to be a sensitive probe of the ATP binding site of the Na+,K+-ATPase, the gastric H+,K+-ATPase and the sarcoplasmic reticulum Ca2+-ATPase. The titration data of eosin with the Na+,K+- (see Fig. 5) and H+,K+-ATPases (see Fig. 6) indicate that eosin binds more strongly to the E1 conformational state of these enzymes than the E2 state. This parallels the behaviour in both enzymes of ATP itself, which is known to bind with high affinity to the E1 state, where it is required for phosphorylation, and with low affinity to the E2 state, where it accelerates the E2-E1 transition required for K+ release 8

BBA - Biomembranes 1862 (2020) 183138

K.R. Hossain, et al.

Fig. 9. Left: Docking-predicted eosin binding poses in the E2 conformations of Na+,K+-ATPase (gold) and SERCA (blue). Right: 2D representation of eosin in the binding site of Na+,K+-ATPase in the E2 conformation, indicating a high level of solvent exposure (areas shaded in blue). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

the purpose of the N-terminal lysine residues of the Na+,K+- and H+,K+-ATPases is still not certain and requires further research, several of them are conserved across vertebrate species [51], which is indicative of an important role, either in enzyme function or its regulation. It is well known, particularly from research on antimicrobial peptides, that lysine residues promote membrane binding [60–63] and investigations using peptide fragments of the Na+,K+-ATPase N-terminus have confirmed membrane interaction [64]. Therefore, a possible role of the lysine residues is to promote interaction of the N-terminus with the neighbouring lipid membrane surface. In spite of the differences in eosin affinity to the Na+,K+-ATPase in the E1 and E2 states, extrapolation to infinite protein concentration, where all of the eosin molecules would be expected to be proteinbound, indicates that there is no significant difference in the local polarity of the eosin binding site between the two states. Thus, we have no clear evidence for any significant structural rearrangement within the binding site which might account for the change in binding affinity. In contrast, as just discussed, there is evidence that the N-terminus, located far from the actual ATP binding site, is implicated in the higher affinity of the E1 state. A possible explanation is that the higher affinity of the E1 state is due largely to an entropic effect rather than an enthalpic one. Theoretical predictions suggest that the N-terminus of both the Na+,K+-ATPase and the H+,K+-ATPase are likely to be disordered [51]. Probably for this reason, the N-terminus of the Na+,K+-ATPase could not be resolved in any of the published X-ray crystal structures [36,65–67]. However, just because a protein segment can't be resolved crystallographically doesn't mean that it is unimportant for protein function. It has recently been recognised that intrinsically disordered regions of proteins can play an important part in the thermodynamics of protein conformational transitions [68,69]. Changes in protein backbone and side-chain conformational entropy are thought to make major contributions to protein folding and to the regulation of protein activity [70,71]. In fact Baxa et al. [71] developed a server which allows the

SERCA for the dye when in the E1 as opposed to the E2 conformation (K values of 0.22 μM versus 0.17 μM) was correctly predicted (binding affinities of −30 kJ mol−1 versus −31 kJ mol−1) and could be rationalized by additional binding interactions occurring in the E2 conformation as the result of a deeper burial of the dye in the enzyme's binding pocket (see Fig. 8). Likewise, a comparison of predicted dye/ enzyme interactions for the E2 conformations of SERCA and Na+,K+ATPase accounted for the experimentally observed lower affinity of Na+,K+-ATPase for the dye (K value of 1.1 μM) because of an altered, more solvent-exposed position in the binding site (interaction energy of −27 kJ mol−1). This is also in agreement with the experimental R∞ values obtained using eosin, which indicated a local dielectric constant of 25 in the Na+,K+-ATPase binding site in comparison with a value of 18 for SERCA. Due to the lack of a crystal structure of Na+,K+-ATPase in a “true” E1 conformation, a direct comparison to SERCA was unfortunately not possible. Based on experiments performed here with N-terminally truncated Na+,K+-ATPase, it appears that a likely reason for the differences in behaviour between the Na+,K+- and H+,K+-ATPases, on the one hand, and the sarcoplasmic reticulum Ca2+-ATPase on the other, lie in differences in the structures of their N-termini. The cytoplasmic N-termini of the α1 subunits of Homo sapiens Na+,K+-ATPase (accession no. P05023) and H+,K+-ATPase (accession no. P20648) possess 82 and 97 amino acid residues, respectively, up to the start of the first transmembrane segment (see Fig. 10). Both contain high frequencies of lysine residues, i.e., 13.41% for the Na+,K+-ATPase and 13.40% for the H+,K+-ATPase [51]. In contrast, the N-terminus of Homo sapiens sarcoplasmic reticulum Ca2+-ATPase (SERCA1, fast twitch skeletal muscle isoform, accession number O14983) only possesses 47 residues prior to the start of the first transmembrane sequence. Although it still has a relatively high lysine frequency of 8.51% (calculated using the ProtoParam tool, https://web.expasy.org/protparam/ within the ExPASy server of the Swiss Institute of Bioinformatics [59]), this is significantly lower than that of both the Na+,K+- and the H+,K+-ATPases. Although

Fig. 10. Sequence alignment of the N-termini of the α1 isoform of the catalytic α-subunit of the Na+,K+- and H+,K+-ATPases with that of the N-terminus of the sarcoplasmic reticulum Ca2+-ATPase (SERCA1, fast twitch skeletal muscle form). All sequences are those of Homo sapiens. Conserved residues are highlighted in yellow. Lysine residues (K) are shown in bold red font. The numbering shown is only that of the Na+,K+-ATPase. The alignment is based on the entire polypeptide sequence of each protein, but only the alignments up to the start of the first transmembrane region is shown. 9

BBA - Biomembranes 1862 (2020) 183138

K.R. Hossain, et al.

Council (Discovery Grants DP121003548, DP150101112 and DP170101732). F.C. acknowledges financial support of the Danish Medical Research Council and Novo Nordisk Foundation. S.P. is grateful for a Senior Scholar fellowship from the American-Australian Fulbright Commission. The authors thank Dr. Michael Baxa, University of Chicago, for assistance with entropy calculations and Prof. Helge Rasmussen, Kolling Institute, University of Sydney, for the kind gift of rabbit leg muscle tissue for Ca2+-ATPase preparation. The authors acknowledge the support received from the Bosch Institute's Molecular Biology Facility, and the expert help of Facility staff, especially Dr. Donna Lai and Dr. Sheng Hua.

change in conformational entropy for the transition between an ordered structure and a completely unfolded one to be estimated (https:// godzilla.uchicago.edu/cgi-bin/PLOPS/PLOPS.cgi). Use of the PLOPS server requires the input of a pdb file of the protein or protein segment concerned. Because no crystal structures of the N-terminus of either the Na+,K+-ATPase or the H+,K+-ATPase are available, we have carried out a prediction of their secondary structures using the QUARK server (http://zhanglab.ccmb.med.umich.edu/ QUARK) [72] and inputted the resultant pdb file obtained for each protein into the PLOPS server. For both proteins we used the entire Nterminus up to the start of the first transmembrane helix for the structure prediction, i.e., up to the motif TPPPTTP for Homo sapiens Na+,K+-ATPase and the motif RPPRGTP for Homo sapiens H+,K+-ATPase (see Fig. 10). Sequence alignment of the Na+,K+-ATPase, H+,K+ATPase and the Ca2+-ATPase, shows that the cytoplasmic N-termini of the Na+,K+-ATPase and the H+,K+-ATPase are both significantly longer than that of the Ca2+-ATPase. The N-terminus of the Ca2+-ATPase does not start until two residues after the conserved 30L31K32K33E of the Na+,K+-ATPase or two residues after the 45M46K47K48E motif of the H+,K+-ATPase. Calculating the total entropy contribution to the change in free energy of a complete unfolding of the predicted Na+,K+ATPase and H+,K+-ATPase N-termini structures up to the start of the Ca2+-ATPase sequence (i.e. up to residue 34 for the Na+,K+-ATPase and 49 for the H+,K+-ATPase) yields values of –TΔS of −234 kJ mol−1 for the Na+,K+-ATPase and −322 kJ mol−1 for the H+,K+-ATPase. These values far exceed the expected free energy of hydrolysis of ATP under physiological conditions of −55 kJ mol−1 [73]. Although it is highly unlikely that the E2 to E1 transition of the Na+,K+-ATPase or the H+,K+-ATPase involves a conversion of their N-termini from a fully structured to a fully disordered state and the calculated values of –TΔS could be partially compensated by entropy changes of solvating water molecules which are not included in the model, this calculation nevertheless shows that changes in the degree of structure or the degree of disorder of the N-terminus of both of these proteins could make a considerable contribution to the stability of the E1 state and hence the overall kinetics and thermodynamics of their pumping mechanisms. Based on previous data suggesting a membrane interaction of the Na+,K+-ATPase N-terminus in the E2 state [11,64,74], a possible scenario for a change in order of the N-terminus could be a release of the N-terminus from the membrane accompanying the E2-to-E1 transition. If the direct physical connection of the ATP binding site through the αsubunit polypeptide chain to the N-terminus causes ATP binding to further enhance the release of the N-terminus from the membrane and increase its entropy in the E1 state, this would be expected to lower the free energy of the ATP-bound state and hence increase the E1 state's ATP binding affinity. Thus, it seems feasible that changes in N-terminus order could make a significant contribution to ATP binding affinity, even with only minor structural changes in the binding site itself. This would be similar to the situation reported by Tzeng and Kalodimos [70] for binding of DNA to the catabolite activator protein (CAP). Furthermore, if the N-terminus were in fact converted into a completely disordered state in the E1 conformation, resulting in an entropy change of the order calculated above, the resultant energy stabilisation would be expected to completely block the enzyme in the E1 state. This could, therefore, represent a very effective mechanism of enzyme regulation, switching off ion pumping completely.

References [1] J.H. Kaplan, Biochemistry of Na,K-ATPase, Annu. Rev. Biochem. 71 (2002) 511–535. [2] G. Sachs, J.M. Shin, O. Vagin, N. Lambrecht, I. Yakubov, K. Munson, The gastric H,K ATPase as a drug target: past, present and future, J. Clin. Gastroenterol. 41 (2007) S226–S242. [3] J.V. Møller, C. Olesen, A.-M. Winther, P. Nissen, The sarcoplasmic reticulum Ca2+ATPase: design of a perfect chemi-osmotic pump, Q. Rev. Biophys. 43 (2010) 501–566. [4] G. Inesi, D. Lewis, C. Toyoshima, A. Hirata, L. de Meis, Conformational fluctuations of the Ca2+-ATPase in the native membrane environment. Effects of pH, temperature, catalytic substrates, and thapsigargin, J. Biol. Chem. 283 (2008) 1189–1196. [5] J.C. Skou, M. Esmann, Eosin, a fluorescent probe of ATP binding to the (Na+ + K+)-ATPase, Biochim. Biophys. Acta 647 (1981) 232–240. [6] J.C. Skou, M. Esmann, The effects of Na+ and K+ on the conformational transitions of (Na+ + K+)-ATPase, Biochim. Biophys. Acta 746 (1983) 101–113. [7] M.R. Montes, R.M. Gonzalez-Lebrero, P.J. Garrahan, R.C. Rossi, Eosin fluorescence changes during Rb+ occlusion in the Na+/K+-ATPase, Biochemistry 45 (2006) 13093–13100. [8] J.T. Organ, M.S. Reifenberger, M.A. Milanick, C. Gatto, Kinetic characterization of Na,K-ATPase inhibition by eosin, Blood Cells Mol. Dis. 38 (2007) 229–237. [9] M. Havlíková, M. Huličak, V. Bazgier, K. Berka, M. Kubala, Fluorone dyes have binding sites on both cytoplasmic and extracellular domains of Na,K-ATPase, Biochim. Biophys. Acta 1828 (2013) 568–576. [10] D.A. Middleton, N.U. Fedosova, M. Esmann, Long-range effects of Na+ binding in Na,K-ATPase reported by ATP, Biochemistry 54 (2015) 7041–7047. [11] Q. Jiang, A. Garcia, M. Han, F. Cornelius, H.-J. Apell, H. Khandelia, R.J. Clarke, Electrostatic stabilization plays a central role in autoinhibitory regulation of the Na+,K+-ATPase, Biophys. J. 112 (2017) 288–299. [12] M.L. Helmich-de Jong, J.P.M. van Duynhoven, F.M.A.H. Schuurmans Stekhoven, J.J.H.H.M. De Pont, Eosin, a fluorescent marker for the high-affinity ATP site of (K+ + H+)-ATPase, Biochim. Biophys. Acta 858 (1986) 254–262. [13] N.A. Saffioti, M. de Sautu, M.S. Ferreira-Gomes, R.C. Rossi, J. Berlin, J.P.F.C. Rossi, I.C. Mangialovori, Biochim. Biophys. Acta Biomembr. 1861 (2019) 366–379. [14] E. Majima, N. Yamaguchi, H. Chuman, Y. Shinohara, M. Ishida, S. Goto, H. Terada, Binding of the fluorescein derivative eosin Y to the mitichondrial ADP/ATP carrier: characterization of the adenine nucleotide binding site, Biochemistry 37 (1998) 424–432. [15] C. Gatto, M.A. Milanick, Inhibition of the red cell calcium pump by eosin and other fluorescein analogues, Am. J. Cell Physiol. – Cell Physiol. 264 (1993) C1577–C1586. [16] C. Gatto, C.C. Hale, W. Xu, M.A. Milanick, Eosin, a potent inhibitor of the plasma membrane Ca pump, does not inhibit the cardiac Na-Ca exchanger, Biochemistry 34 (1995) 965–972. [17] I. Klodos, M. Esmann, R.L. Post, Large-scale preparation of sodium-potassium ATPase from kidney outer medulla, Kidney Int. 62 (2002) 2097–2100. [18] P. Ottolenghi, The reversible delipidation of a solubilized sodium-plus-potassium ion-dependent adenosine triphosphatase from the salt gland of the spiny dogfish, Biochem. J. 151 (1975) 61–66. [19] G.L. Peterson, A simplification of the protein assay method of Lowry et al. which is more generally applicable, Anal. Biochem. 83 (1977) 346–356. [20] O.H. Lowry, N.J. Rosebrough, A.L. Farr, R.J. Randall, Protein measurement with the Folin phenol reagent, J. Biol. Chem. 193 (1951) 265–275. [21] G. Sachs, H.H. Chang, E. Rabon, R. Schackman, M. Lewin, G. Saccomani, A nonelectrogenic H+ pump in plasma membranes of hog stomach, J. Biol. Chem. 251 (1976) 7690–7698. [22] M.R. Montes, A.J. Spiaggi, J.L.E. Monti, F. Cornelius, C. Olesen, P.J. Garrahan, R.C. Rossi, Rb+ occlusion stabilized by vanadate in gastric H+/K+-ATPase at 25°C, Biochim. Biophys. Acta Biomembr. 1808 (2011) 316–322. [23] E.S. Baginski, P.P. Foa, B. Zak, Determination of phosphate: study of labile organic phosphate interference, Clin. Chim. Acta 14 (1967) 155–158. [24] F. Cornelius, Y.A. Mahmmoud, Direct activation of gastric H,K-ATPase by N-terminal protein kinase C phosphorylation. Comparison of the acute regulation mechanisms of H,K-ATPase and Na,K-ATPase, Biophys. J. 84 (2003) 1690–1700. [25] C.M. Jakobsen, S.R. Denmeade, J.T. Isaacs, A. Gady, C.E. Olsen, S.B. Christensen, Design, synthesis, and pharmacological evaluation of thapsigargin analogues for targeting apoptosis to prostatic cancer cells, J. Med. Chem. 44 (2001) 4696–4703.

Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Acknowledgement R.J.C. acknowledges financial support from the Australian Research 10

BBA - Biomembranes 1862 (2020) 183138

K.R. Hossain, et al.

[52] P.L. Jørgensen, Purification and characterization of (Na+,K+)-ATPase. V. Conformational changes in the enzyme transitions between the Na-form and the Kform studied with tryptic digestion as a tool, Biochim. Biophys. Acta 401 (1975) 399–415. [53] P.L. Jørgensen, E. Skriver, H. Hebert, A.B. Maunsbach, Structure of the Na,K-pump: crystallization of pure membrane-bound Na,K-ATPase and identification of functional domains of the α-subunit, Ann. N. Y. Acad. Sci. 402 (1982) 207–225. [54] R. Scanzano, L. Segall, R. Blostein, Specific sites in the cytoplasmic N terminus modulate conformational transitions of the Na,K-ATPase, J. Biol. Chem. 282 (2007) 33691–33697. [55] I.M. Glynn, D.E. Richards, Occlusion of rubidium ions by the sodium-potassium pump: its implications for the mechanism of potassium transport, J. Physiol. 330 (1982) 17–43. [56] W.W. Reenstra, J. Crothers Jr., J.G. Forte, The conformation of H,K-ATPase determines the nucleoside triphosphate (TNP) selectivity for active proton transport, Biochemistry 46 (2007) 10145–10152. [57] H. Rui, A. Das, R. Nakamoto, B. Roux, Proton countertransport and coupled gating in the sarcioplasmic reticulum calcium pump, J. Mol. Biol. 430 (2018) 5050–5065. [58] N. Tsunekawa, H. Ogawa, J. Tsueda, T. Akiba, C. Toyoshima, Mechanism of the E2 to E1 transition in Ca2+ pump revealed by crystal structures of gating residue mutants, Proc. Natl. Acad. Sci. U. S. A. 115 (2018) 12722–12727. [59] E. Gasteiger, C. Hoogland, A. Gattiker, S. Duvand, M.R. Wilkins, R.D. Appel, A. Bairoch, Protein identification and analysis tools on the ExPASy server, in: J.M. Walker (Ed.), The Proteomics Protocols Handbook, Humana Press, New York, 2005, pp. 571–607. [60] M. Zasloff, Antimicrobial peptides of multicellular organisms, Nature 415 (2002) 389–395. [61] L. Li, I. Vorobyov, T.W. Allen, The different interactions of lysine and arginine side chains with lipid membranes, J. Phys. Chem. B 117 (2013) 11906–11920. [62] A. Hädicke, A. Blume, Binding of the cationic peptide (KL)4K to lipid monolayers at the air-water interface: effect of lipid headgroup charge, acyl chain length and acyl chain saturation, J. Phys. Chem. B 120 (2016) 3880–3887. [63] A. Gorman, K. R. Hossain, F. Cornelius, R. J. Clarke, Penetration of phospholipid membranes by poly-L-lysine depends on cholesterol and phospholipid composition, Biochim. Biophys. Acta Biomembr., (in press). [64] K. Nguyen, A. Garcia, M.-A. Sani, D. Diaz, V. Dubey, D. Clayton, G. Dal Poggetto, F. Cornelius, R.J. Payne, F. Separovic, H. Khandelia, R.J. Clarke, Interaction of Nterminal peptide fragments of the Na+,K+-ATPase with membranes, Biochim. Biophys. Acta Biomembr. 1860 (2018) 1282–1291. [65] T. Shinoda, H. Ogawa, F. Cornelius, C. Toyoshima, Crystal structure of the sodiumpotassium pump at 2.4 Å resolution, Nature 459 (2009) 446–450. [66] R. Kanai, H. Ogawa, B. Vilsen, F. Cornelius, C. Toyoshima, Crystal structure of a Na+-bound Na+,K+-ATPase preceding the E1P state, Nature 502 (2013) 201–206. [67] M. Nyblom, H. Poulsen, P. Gourdon, L. Reinhard, M. Andersson, E. Lindahl, N. Fedosova, P. Nissen, Crystal structure of Na+,K+-ATPase in the Na+-bound state, Science 342 (2013) 123–127. [68] A.J. Wand, The dark energy of proteins comes to light: conformational entropy and its role in protein function revealed by NMR relaxation, Curr. Opin. Struct. Biol. 23 (2013) 75–81. [69] J.A. Drake, B.M. Pettitt, Thermodynamics of conformational transitions in a disordered protein backbone model, Biophys. J. 114 (2018) 2799–2810. [70] S.-R. Tzeng, C.G. Kalodimos, Protein activity regulation by conformational entropy, Nature 488 (2012) 236–240. [71] M.C. Baxa, E.J. Haddadian, J.M. Jumper, K.F. Freed, T.R. Sosnick, Loss of conformational entropy in protein folding calculated using realistic ensembles and its implications for NMR-based calculations, Proc. Natl. Acad. Sci. U. S. A. 111 (2014) 15396–15401. [72] D. Xu, Y. Zhang, Ab initio protein structure assembly using continuous structure fragments and optimized knowledge-based force field, Proteins 80 (2012) 1715–1735. [73] R.J. Clarke, M. Catauro, H.H. Rasmussen, H.-J. Apell, Quantitative calculation of the role of the Na+,K+-ATPase in thermogenesis, Biochim. Biophys. Acta Bioenerg. 1827 (2013) (2013) 1205–1212. [74] A. Garcia, P.R. Pratap, C. Lüpfert, F. Cornelius, D. Jacquemin, B. Lev, T.W. Allen, R.J. Clarke, The voltage-sensitive dye RH421 detects a Na+,K+-ATPase conformational change at the membrane surface, Biochim. Biophys. Acta Biomembr. 1859 (2017) 813–823.

[26] M. Lape, C.E. Lam, M. Versluis, R. Kempton, S. Paula, Molecular determinants of sarco/endoplasmic reticulum calcium ATPase inhibition by hydroquinone-based compounds, Proteins 70 (2008) 639–649. [27] A. K. Schwartz, M. Nagano, M. Nakao, G. E. Lindenmayer, J. C. Allen, H. Matsui, The sodium- and potassium-activated adenosine triphosphatase system, in: A. Schwartz (Ed.), Methods in Pharmacology, Vol. 1, Chapter 12, Appleton-CenturyCrofts, New York, pp. 361–388. [28] A. Garcia, B. Lev, K.R. Hossain, A. Gorman, D. Diaz, T.H.N. Pham, F. Cornelius, T.W. Allen, R.J. Clarke, Cholesterol depletion inhibits Na+,K+-ATPase activity in a near-native membrane environment, J. Biol. Chem. 294 (2019) 5956–5969. [29] F. Cornelius, Y.A. Mahmmoud, L. Meischke, G. Cramb, Functional significance of the shark Na,K-ATPase N-terminal domain. Is the structurally variable N-terminus involved in tissue-specific regulation by FXYD proteins? Biochemistry 44 (2005) 13051–13062. [30] A. Pilotelle-Bunner, F. Cornelius, P. Sebban, P.W. Kuchel, R.J. Clarke, Mechanism of Mg2+ binding in the Na+,K+-ATPase, Biophys. J. 96 (2009) 3753–3761. [31] M. Khalid, F. Cornelius, R.J. Clarke, Dual mechanisms of allosteric acceleration of the Na+,K+-ATPase by ATP, Biophys. J. 98 (2010) 2290–2298. [32] S.L. Myers, F. Cornelius, H.J. Apell, R.J. Clarke, Kinetics of K+ occlusion by the phosphoenzyme of the Na+,K+-ATPase, Biophys. J. 100 (2011) 70–79. [33] P.L. Jørgensen, J.H. Collins, Tryptic and chymotryptic sites in sequence of α-subunit of (Na+ + K+)-ATPase from outer medulla of mammalian kidney, Biochim. Biophys. Acta 860 (1986) 570–576. [34] C. Lüpfert, E. Grell, V. Pintschovius, H.-J. Apell, F. Cornelius, R.J. Clarke, Rate limitation of the Na+,K+-ATPase pump cycle, Biophys. J. 81 (2001) 2069–2081. [35] A. Diller, O. Vagin, G. Sachs, H.-J. Apell, Electrogenic partial reactions of the gastric H,K-ATPase, Biophys. J. 88 (2005) 3348–3359. [36] J.P. Morth, B.P. Pedersen, M.S. Toustrup-Jensen, T.L.-M. Sørensen, J. Petersen, J.P. Andersen, B. Vilsen, P. Nissen, Crystal structure of the sodium-potassium pump, Nature 450 (2007) 1043–1049. [37] C. Toyoshima, M. Nakasako, H. Nomura, H. Ogawa, Crystal structure of the calcium pump of sarcoplasmic reticulum at 2.6 Å resolution, Nature 405 (2000) 647–655. [38] K. Obara, N. Miyashita, C. Xu, I. Toyoshima, Y. Sugita, G. Inesi, C. Toyoshima, Structural role of countertransport revealed in Ca2+ pump crystal structure in the absence of Ca2+, Proc. Natl. Acad. Sci. U. S. A. 102 (2005) 14489–14496. [39] G.M. Morris, R. Huey, W. Lindstrom, M.F. Sanner, R.K. Belew, D.S. Goodsell, A.J. Olson, AutoDock4 and AutoDockTools4: automated docking with selective receptor flexibility, J. Comput. Chem. 30 (2009) 2785–2791. [40] O. Trott, A.J. Olson, AutoDock Vina: improving the speed and accuracy of docking with a new scoring function, efficient optimization, and multithreading, J. Comput. Chem. 31 (2010) 455–461. [41] R.C. Edgar, MUSCLE: multiple sequence alignment with high accuracy and high throughput, Nucleic Acids Res. 32 (2004) 1792–1797. [42] S. Kumar, G. Stecher, K. Tamura, MEGA7: molecular evolutionary genetics analysis version 7.0 for bigger datasets, Mol. Biol. Evol. 33 (2016) 1870–1874. [43] M. Chakraborty, A.K. Panda, Spectral behaviour of eosin Y in different solvents and aqueous surfactant media, Spectrochim. Acta B 81 (2011) 458–465. [44] P.S. Albright, L.J. Gosting, Dielectric constants of the methanol-water system from 5 to 55°, J. Am. Chem. Soc. 68 (1946) 1061–1096. [45] L. Flamigni, Inclusion of fluorescein and halogenated derivatives in α-, β- and γcyclodextins. A steady-state and picosecond time-resolved study, J. Phys. Chem. 97 (1993) 9566–9572. [46] M. Hansen, E. Pines, G.R. Fleming, Excited-state proton transfer of protonated 1aminopyrene complexed with β-cyclodextrin, J. Phys. Chem. 96 (1992) 6904–6910. [47] A. Pilotelle-Bunner, J.M. Matthews, F. Cornelius, H.-J. Apell, P. Sebban, R.J. Clarke, ATP binding equilibria of the Na+,K+-ATPase, Biochemistry 47 (2008) 13103–13114. [48] R.J. Clarke, Mechanism of allosteric effects of ATP on the kinetics of P-type ATPases, Eur. Biophys. J. Biophys. Lett. 39 (2009) 3–17. [49] P.L. Jørgensen, J.P. Andersen, Structural basis for E1-E2 conformational transitions in Na,K-pump and Ca-pump proteins, J. Membr. Biol. 103 (1988) 95–120. [50] I. Dach, C. Olesen, L. Signor, P. Nissen, M. le Maire, J.V. Møller, C. Ebel, Active detergent-solubilized H+,K+-ATPase is a monomer, J. Biol. Chem. 287 (2012) 41963–41978. [51] D. Diaz, R.J. Clarke, Evolutionary analysis of the lysine-rich N-terminal cytoplasmic domains of the gastric H+,K+-ATPase and the Na+,K+-ATPase, J. Membr. Biol. 251 (2018) 653–666.

11