Measurements of Ion Binding to Lipid-Hosted Ionophores by Affinity

Jun 24, 2019 - Stark, G.; Ketterer, B.; Benz, R.; Läuger, P. The Rate Constants of .... Wahab, M. F.; Anderson, J. K.; Abdelrady, M.; Lucy, C. A. Pea...
1 downloads 0 Views 2MB Size
Download PDF
Article pubs.acs.org/Langmuir

Cite This: Langmuir 2019, 35, 9410−9421

Measurements of Ion Binding to Lipid-Hosted Ionophores by Affinity Chromatography Eric E. Ross,* Bridget Hoag, Ian Joslin,† and Taylor Johnston† Department of Chemistry & Biochemistry, Gonzaga University, Spokane, Washington 99258, United States

Downloaded via BUFFALO STATE on July 24, 2019 at 05:59:21 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

S Supporting Information *

ABSTRACT: The binding affinity between antibiotic ionophores and alkali ions within supported lipid bilayers was evaluated using affinity chromatography. We used zonal elution and frontal analysis methods in nanovolume liquid chromatography to characterize the binding selectivity of the carrier and channel ionophores valinomycin and gramicidin A within different phosphatidylcholine bilayers. Distinct binding sensitivity to the lipid phase, both in affinity and selectivity, is observed for valinomycin, whereas gramicidin is less sensitive to changes in a membrane environment, behavior that is consistent with ion binding occurring within the interior of an established channel. There is good agreement between the chromatographic retention and the reported binding selectivity measured by other techniques. Surface potential near the binding site affects ion retention and the apparent association binding constants, but not the binding selectivity or enthalpy measurements. A model accounting for the surface potential contributions of retained ions during frontal analyses yields values close to intrinsic binding constants for gramicidin A (KA for K+ between 70 and 120 M−1) using reasonable estimates of the initial potential that is postulated to arise from the underlying silica.

■

INTRODUCTION Lipophilic ionophores that transport specific ions across membranes have widespread applications as antibiotics,1−3 chemical sensing elements,4−6 models in studies of ion transport,7,8 and for potential therapeutic purposes.9−12 Ion complexation is a key determinant of selectivity for carrier-type ionophores such as valinomycin (Val)13,14 and is a factor in that of channel-formers such as gramicidin as well.15−17 Affinity measurements can provide insight into the mechanisms of transport activity and the effect of synthetic modifications on function.3,18−20 An evaluation of the effects on binding caused by diverse lipid environments is experimentally challenging because of various aspects of the techniques capable of probing the low affinity interaction occurring at the bilayer interface or within ion channels. Here, we describe a unique approach for the direct detection of ion/ ionophore complexation through ion retention on affinity chromatography columns containing solid-supported lipid bilayers. Solvents are often substituted for lipid environments in affinity characterization by spectrometric titrations, where detection of the complex is facilitated by increased stability and the absence of lipids. Binding is highly sensitive to solvent polarity, and different solvents can mimic various regions of the lipid bilayer. For example, alcohols and dimethylsulfoxide are often used to represent the lipid interface where carrier-type ionophores bind ions before diffusing across the hydrophobic bilayer core.21,22 Trends from these data generally correlate with binding and transport selectivity measured in lipid © 2019 American Chemical Society

membranes. However, differences in quantitative selectivity and transport mechanism are observed between measurements in bulk solvents and lipid environments,23−26 reflecting the former’s limited replication of the lipid bilayer chemical environment. Notably, the ion-conductive channel conformation of gramicidin does not form in solvents.27,28 Because lipid bilayer properties are sensitive to membrane composition, phase, and solution factors, these yield potential effects on the binding environment. It has been suggested that altered binding stability from interfacial membrane properties could explain differences in mediated transport rates observed under different solution conditions.20,29 Consequently, determining the effects that membrane and solution variables have on binding strength and selectivity could complement information about transport rate and complex structure gained by other techniques. NMR can reveal detailed mechanisms of binding within lipid environments but determination of the affinity may involve deuterated30,31 or isotopically enriched ionophores,32,33 micellar rather than lamellar lipid structures for resolution enhancement,34 or restrict methodology to ions with NMRactive isotopes or competitive assays with those that are.19 Improving upon the sensitivity of NMR techniques has been cited as one reason for the development of new techniques for the measurement of ionophore/ion selectivity.35 Practical Received: May 2, 2019 Revised: June 20, 2019 Published: June 24, 2019 9410

DOI: 10.1021/acs.langmuir.9b01301 Langmuir 2019, 35, 9410−9421

Langmuir

■

factors can limit a researcher’s use of a particular technique for these measurements including access to instrumentation, low experimental throughput, or high technical expertise requirements. Conductance and water permeability studies36,37 have returned binding constants within lipid systems but the difficulty of working with fragile free-standing lipid films has motivated the development of electrochemical methods based on more robust electrode-supported lipids and tethered membranes38,39 which have not, to our knowledge, been used to investigate alkali ion binding to ionophores. This work examines the hypothesis that the retention of alkali ions on columns containing supported lipid bilayers will accurately reflect the ion-binding properties and selectivity of ionophores within lipid environments measured by other techniques. Minor differences in the weak affinity of solutes for a stationary phase can produce distinctive retention times, hence the potential for evaluating low affinity ionophore/ion interactions.40,41 Solute retention is directly related to a partition or binding constant with stationary phases.42,43 Column and elution conditions are identical for different solutes which are often injected simultaneously, producing a robust measure of the relative affinity for the stationary phase. Given the variety of detection options in chromatography, chemical labels are seldom required and the experimental throughput is generally high compared to spectroscopic techniques.44 In this journal, chromatography has been used to characterize interactions between a number of solutes and interfaces including: amino acids and carbon nanotubes;45 peptides and titanium dioxide;46 surfactants and gold;47 and peptides and conjugated polymer nanoparticles with commercial hydrophobic stationary phases.48,49 Affinity chromatography with biological membranes and membrane mimics has a long history, but the technique has not yet been applied to the study of membrane ionophores.50−52 Lipid bilayers can be immobilized as liposomes53 or adsorbed directly on silica column materials.54,55 The attachment of liposomes requires chemical modification of the surface and reactive bilayer components that might contribute to nonspecific ion retention or eliminate suitable void markers for this work. Here, 250 nm Stöber silica spheres are aggregated into micron-scale particles and used to support lipid bilayers formed by the rehydration of solvent-cast lipids.55,56 This procedure is part of a process that results in a microcapillary chromatography column with a particular lipid and ionophore formulation within 2 h. The present work investigates ion binding to the ionophores Val and gramicidin A within liquid- and gel-phase lipid bilayer environments. Both ionophores selectively transport potassium over sodium ions through lipid bilayers, albeit by different mechanisms. Val is a cyclic peptide that chelates select metal ions at the lipid interface and diffuses across the hydrophobic portion of the bilayer.57,58 Gramicidin A is a linear peptide that can form a β6.3-helical dimerized channel which interacts with alkali ions near the channel entrance points.59 As two of the most extensively studied natural ionophores, ion affinity and selectivity have been measured by NMR for both peptides within lipid environments, providing comparative reference points for this work.31,34 A brief review of the relationship between chromatographic retention and affinity binding constants is given, and we discuss the effect that surface potential has on different parameters and values returned from the data.

Article

EXPERIMENTAL SECTION

This study focuses on two ionophores, Val and gramicidin (gA) (Figure 1), hosted within lipid bilayers supported on silica colloids. A

Figure 1. Structure of Val (left), a naturally occurring dodecadepsipeptide carrier ionophore. Gramicidin A (right) is shown in the dimerized channel structure determined within hydrated DMPC bilayers; it is a linear peptide with a structure formyl-Val-Gly-Ala-DLeu-Ala-D-Val-Val-D-Val-Trp-D-Leu-Trp-D-Leu-Trp-D-Leu-Trp-ethanolamine. Image from the RCSB PDB (rcsb.org) of PDB ID: 1MAG (Ketchem, R. R.; Lee, K. C.; Huo, S.; Cross, T. A. Macromolecular Structural Elucidation with Solid-State NMR-Derived Orientational Constraints. J. Biomol. NMR, 1996, 8, 1−14).

pH 7.0, 50 mM Tris buffer system devoid of sodium was the aqueous media. Figure 2 contains the structure of the primary lipids used, POPC (1-palmitoyl-2-oleoyl-glycero-3-phosphocholine, Tm = −2 °C) and DMPC (1,2-dimyristoyl-sn-glycero-3-phosphocholine Tm = 24 °C), and a schematic representation of the experimental configuration and stationary phases. 1,2-Dilauroyl-sn-glycero-3-phosphocholin (DLPC) (Tm = −2 °C) and 1,2-dipalmitoyl-sn-glycero-3-phosphocholine (DPPC) (Tm = 41 °C) which have two fewer and two more carbons in each saturated tail, respectively, then DMPC were also used. Also shown is an electron microscopy image of the silica stationary phases, termed Stöber-agglomerated particles (SAPs), composed of 250 nm Stöber silica spheres. The apparent size of the SAPs, as evaluated by scanning electron microscopy, is 11.4 ± 3.6 μm. Bilayers were formed by rehydrating dried, solvent-cast films on SAPs which were then slurry packed into 10−12 cm fused silica columns. Thermogravimetric analyses (TGA) and microbalance measurements on SAPs and columns show them to contain an average 110 ± 4 nL of mobile phase, 136 ± 8 μg of silica, 7.3 ± 0.6 nmol lipid per centimeter of column. The ionophore/lipid ratio was 9:1 unless noted otherwise. The eluent, typically Rb+ > Li+ > Na+, largely matching the trend observed by NMR measurements albeit with the order of Li+ and Na+ reversed.34 Within different lipids and solvents, gA is known to adopt a number of different conformational dimer states.28,65,66 A nonchannel, double-helical form binds to Rb+ strongly, but not to K+ or Li+.28 Cross and co-workers did not observe Na+ binding to a nonchannel form within DMPC bilayers.65 Here, gA binding to K+ is stronger than Rb+, and there is clear binding to Na+ and Li+, observations that support the presence of the channel conformation within SAPsupported bilayers. These data indicate that the chromatographic retention order is well correlated with the relative binding affinity of membrane ionophores within lipid environments determined by other techniques. The bar graphs in Figure 3 contain the retention factors measured with the ionophores in different phosphatidylcholine lipid environments. Binding to Val was nearly undetectable in the shorter-chain DLPC bilayers at 20 °C, and much weaker

just above the phase transition temperature (25 °C) of DMPC than below it (20 °C). Interestingly, binding to Val in DPPC at these temperatures only occurred under specific circumstances which are described later, and binding was too weak to observe above this lipid’s characteristic melting point. Within fluidphase bilayers, the apparent binding decreases on the order DMPC > POPC > DLPC. In solvents, Val complexation with alkali ions is highly sensitive to solvent polarity.67 In lipid bilayers, it complexes alkali ions near the membrane surface in a process that requires conformational changes.58 It follows that a change in apparent ion binding to Val within different lipids may provide a measure of feedback on the local polarity and motional restriction of the binding environment. The frontal analysis studies discussed later indicate that changes to the apparent binding constant (KA) rather than to the number of binding sites is the principle factor affecting retention with Val. Conversely, frontal analyses indicate that the retention differences for gA-columns are primarily due to the number of binding sites formed in the films. The two ionophores equilibrated at different rates within the lipid films. Ion retention to Val-lipid columns did not change appreciably after hydration but retention on gA-lipid columns increased over several days and stabilized near day 5 within POPC (Figure S4). Again, frontal analysis data (discussed later) indicate that the retention increase arises from an increasing number of binding sites rather than a change in binding affinity. Further increase in retention after five days was rare in any of the lipids, so it was used as the storage period (at temperatures above the lipid melting point) for gA data in Figure 3. Within the saturated lipids, gA equilibration may be extremely slow on SAPs, or it may not favor the same distribution of gA conformations. The conformation of gA within lipid phases is affected by solvent history and lipid composition. Cross et al. found that rehydration of dried DMPC and gA at an 8:1 M ratio after cosolubilization in chloroform resulted in stable, nonchannel conformations that did not interact with Na+ despite sonication and heat treatments, but that resolubilization from trifluoroethanol (TFE) quickly resulted in the channel form of gA.65 Here, gA solvent-cast from CHCl3 routinely showed little initial ion retention, and inconsistent retention increases after week-long incubation periods. When TFE is used to cast gA, 9413

DOI: 10.1021/acs.langmuir.9b01301 Langmuir 2019, 35, 9410−9421

Article

Langmuir

Figure 4. (a) Temperature dependence of K+ (circle) and Rb+ (square) retention for Val (top) and gA (bottom) within DMPC (solid markers) and POPC (open markers); (b) chromatograms for Rb+ on Val-DMPC (3 mol %) columns; and (c) sample van’t Hoff plot for gA-DMPC.

Effect of Temperature on Ion Retention. Plots of kvalues versus temperature for the ionophores in POPC and DMPC bilayers are shown in Figure 4a. Most notably, ion retention to Val-DMPC declines sharply between 18 and 25 °C. This range spans the enthalpic transition observed by differential scanning calorimetry (DSC) with DMPC on SAPs, which is not detected with 10 mol % Val, but is with a reduced ionophore fraction (3 mol %, Figure S3). Figure 4b contains a set of chromatograms observed for such a column between 10 and 28 °C, and the decline in retention begins about one degree higher and is somewhat sharper than for columns with more Val (Figure S8). The ion peaks on Val-DMPC columns broaden asymmetrically below 20 °C, which is indicative of slowing mass transfer kinetics.70 This likely reflects a decrease in dissociation rate (kd) for the conformationally dynamic interaction resulting from an increasingly viscous lipid environment. A treatment of the peak widths to estimate kd values (described in the Supporting Information with Figure S9) suggests that the rates increase by a factor of 200 over the temperature range, but are several orders of magnitude smaller than rates measured in free-standing fluid bilayers. For Val-POPC and gA within both lipids, the changes with temperature appear fairly linear (Figure 4a). This prompted a van’t Hoff evaluation to estimate the binding enthalpy in these lipid systems (Figures 4c and S10−S12). Enhanced ion concentration due to surface potential should not influence the enthalpy determination because it causes a common offset of the van’t Hoff plot without affecting the slope. The plots for gA appear to have slight curvature, suggesting that the enthalpy is not constant over the temperature range, which is not surprisingly given the changes to bilayer viscosity that might affect the ionophore. Taken over the entire range, the enthalpies for gA-POPC are −24.2 ± 1.9, −21.5 ± 2.9, and −18.5 ± 4.1 kJ/mol for K+, Rb+, and Na+, respectively. For gADMPC, they are −31.0 ± 2.6, −29.4. ± 0.6, and −20.4 ± 2.5 kJ/mol for K+, Rb+, and Na+, respectively. These are somewhat larger than the values determined by Hinton and co-workers for gA in lysoPC by NMR (of −20.4, −16.7, −14.8 kJ/mol for K+, Rb+, and Na+, respectively).34 The plot curvature is evident when the low- and hightemperature regions are evaluated separately (see Figures S10 and S11); the data between 10 and 20 °C return lower enthalpies and a value of −19.6 ± 1.5 kJ/mol for K+ with gAPOPC which is very close to the NMR-obtained value. The higher temperature range returns values for K+ and Rb+ that are 40−60% larger with both DMPC and POPC. The better agreement with the literature values at lower temperatures

initial retention is higher and consistent in the retention increase over a five-day period. Altogether, the observations support the presence of the channel conformation in lipid SAPs but indicate that it forms much more slowly than in unsupported lipid systems, at least at the high gA mole fraction investigated. Differences between the observed and predicted retention (from intrinsic KA and eq 1) could be caused by lipid environment, temperature, or negative surface potentials that increase the interfacial concentration of injected cations. The values referenced for Val were measured at 50 °C in DMPC bilayers, or at lower temperatures in monoglyceride bilayers. The intrinsic binding constants for gA reported by Hinton et al. were measured in lysoPC from egg yolk above 34 °C. For the exothermic interactions, affinity will increase at lower temperatures. The solid support may affect properties of the surrounding lipids to which the binding is sensitive. To our knowledge, there are no precedent studies involving ion binding to these ionophores within gel-phase lipids for comparison. A negative surface potential may arise from anion adsorption or ionized silanol groups on the silica surface near the bilayers. Sintering the SAPs at 1100 °C reduces but does not completely eliminate the surface coverage of silanols, and partial rehydroxylation can occur in water.68 In the NMR studies on alkali ion−Val binding by Meers and Feigenson, thiocyanate adsorption to the lipids substantially increased apparent KA values regardless of ionic strength.31 Here, chloride is present at 45 mM in the mobile phase, but it binds to phospholipids with much lower affinity than the thiocynate ion.69 Increasing the mobile phase buffer, and hence chloride concentration, decreased ion retention with both ionophores (Figure S5). The chloride concentration cannot be increased otherwise without introducing a different cation that could potentially compete with the alkali ions. The decrease in retention suggests that chloride adsorption contributes minimally to a surface potential which likely originates from the presence of residual acidic silanol groups that are increasingly screened at higher ionic strengths. Factors affecting uncertainty in k-values and their relationship to KA are further discussed in the Supporting Information (with Figures S6 and S7). Last, we note that neither ionophore was included in the mobile phase, but no decline in ion retention was observed over the period in which the columns were used that would suggest that ionophore loss was a factor in these analyses. 9414

DOI: 10.1021/acs.langmuir.9b01301 Langmuir 2019, 35, 9410−9421

Article

Langmuir

Meers and Feigenson reported that K+ complexation caused Val to move toward the lipid interface in DMPC.31 Freezing the cationic complex near the DPPC headgroups may increase its accessibility during subsequent ion injections and contribute to the increased retention, but it is doubtful that Val would be completely embedded within the tail region of the bilayer in the absence of alkali ions in light of its aforementioned behavior in saturated lipids. More likely, either gelation of DPPC pushes uncomplexed Val out of the bilayer so far as to reduce its binding constant to an unmeasurably low value, or it restricts the conformational changes it needs to bind ions. Freezing around the more lipophilic Val−ion complex72 may position it within an environment where it has higher binding affinity. The incubation condition had no effect on the binding behavior of Val-DMPC, revealing a distinction in how Val is affected by the two gel-phase lipid environments. Chromatographic Determination of Ionophore Binding Selectivity. A negative surface potential should increase the interfacial concentration of like-charged alkali ions equally; thus, it would not bias a measure of the chromatographic selectivity (α) ratio for two ions, α = k2/k1, which is equivalent to the binding selectivity, that is, the KA ratio for two ions when nI/VMP is common for both. Table 1 contains the average α values determined from the background-corrected retention factors within individual chromatograms and the reported data from NMR measurements. Rather than calculating α from the mean k-values of different columns, it was determined from those within individual chromatograms, which normalizes some sources of error because the relative standard deviations are substantially smaller than those for the mean k values from separate column preparations. There is general agreement in the Rb+ and K+ selectivity values for both ionophores measured by the two techniques in spite of differences in lipid environment and measurement temperatures. For brevity, this discussion focuses on Val and gA, but the retention measured with a third ionophore, nigericin, is also consistent with the established binding selectivity (Figure S13). The thermodynamic origin of the group I binding selectivity for gA and Val involves both ionophore and cation properties and has been previously discussed.13,17,34,72 The smaller ions have larger hydration energies that oppose ligand binding but they offer larger potential ion−ligand dipole interactions. Binding is favored for the larger ions, K+ and Rb+, over smaller ones for both Val and gA. The binding behavior for the gA channel was shown to closely follow the pattern of ion transfer enthalpies between a water and amide environment.34 With Val, cavity size constraints of the ionophore play a role in binding, producing a much greater difference in its selectivity for K+ and Rb+ over both smaller (Na+) and larger (Cs+) ions.72 The αK/Rb measured for gA in saturated-tail lipids at 20 °C is very close to that observed by Hinton et al. in egg lysoPC at 34 °C,34 whereas the αK/Rb within POPC is slightly higher. Although a phase transition is not observed by DSC with gADMPC, the steady-state fluorescence emission of laurdan undergoes a red shift above 20 °C (Figure S14) which is characteristic of its behavior upon lipid melting.73 There was no significant change in αK/Rb gA-DMPC with temperature (Figure S15), indicating that the environmental changes affecting laurdan emission do not influence gA selectivity in DMPC. This is consistent with an observation of channel conductivity being largely unaffected by a lipid phase transition74 and that ion binding to the channel induces, and

could plausibly be associated with the relative changes to temperature-dependent bilayer properties better matching those occurring in the referenced lysoPC phases in this range. The impact of the large gA mole fraction on these properties is likely substantial. Finally, the intercepts in the van’t Hoff plots are affected by surface potential and the substitution of k for KA, so only relative comparisons of entropy can be made. The trend in the entropic opposition to complexation is the same as in lysoPC, binding to K+ being more negative than to Rb+. For Val-POPC, the enthalpies for Rb+ and K+ were 38.4 ± 2.9 and 31.3 ± 3.3 kJ/mol, respectively, with K+ more opposed entropically; however, these values are not particularly robust because of their sensitivity to the background correction and the reduced reliability of the smaller k values. Plot curvature was less apparent. Correction with lipid-only background retention rather than bare-SAP retention reduced the calculated ΔH by nearly 23% for Rb+ but only 7% for K+, whereas both values change by less than 5% with gA (Figure S10). The methodology shows potential to probe binding energetics, but uncertainty in the effect of the solid support and high ionophore mole fraction on bilayer properties are factors when comparing results with those from unsupported systems. Further, background correction uncertainty limits the accuracy obtainable with weaker binding ionophores. Ion Binding to Val-DPPC Columns. DPPC has two more carbon atoms in each tail than DMPC and exhibits a phase transition near 38 °C on SAPs.55 In liposomes, Val shows lesser effect on the phase transition for DPPC than DMPC which suggests that it remains closer to the bilayer interface in both liquid and gel phases in DPPC.71 Despite the expected similarity in the interfacial chemical environments of these lipids below their phase transition temperatures, the binding behavior of Val was substantially different within DPPC. Unexpectedly, no binding occurs with Val-DPPC unless Rb+ or K+ ions (10−50 mM) are present in the mobile phase during the cooling cycle through the transition temperature, at which point retention with extensive peak tailing is observed (Figure 5). Retention declines in subsequent injections before

Figure 5. Chromatograms for Val-DPPC at 20 °C (top) with no Rb+ present during the cooling cycle from 45 °C, and (bottom) with 50 mM Rb+ present. The inset shows the initial retention decline during repeated injections, followed by a decrease from the temperature change.

stabilizing and then decreases significantly when heated to the apparent phase transition temperature as it does with DMPC (inset, Figure 5b). Incubation of Val-DPPC with Rb+ and K+ at 20 °C has no effect, but binding returns repeatedly if ions are present through the phase transition. 9415

DOI: 10.1021/acs.langmuir.9b01301 Langmuir 2019, 35, 9410−9421

Article

Langmuir Table 1. Alkali Ion Selectivity of Gramicidin A and Vala gramicidin A αK/Rb (gA); αRb/K (Val) αK/Na (gA)

Val

DMPC

DPPC

POPC

lysoPCb (NMR)

DMPC

DMPC 25 °C

POPC

DMPCc (NMR)

1.16 ± 0.04 3.4 ± 0.1

1.18 ± 0.03 4.2 ± 0.1

1.26 ± 0.03 3.8 ± 0.1

1.19 1.84

1.3 ± 0.1

1.75 ± 0.06

2.6 ± 0.2

1.69

a

Ratio of background-corrected retention factors compared within individual chromatograms, each from independent lipid-SAP preparations (n = 3), at 20 °C unless noted. bReferences 34, 34 °C. cReferences 31, 50 °C. Reference 58 provides a value of 2.63 in monooleate bilayers at 10 °C.

consequently requires, minimal conformation changes.75 In POPC, αK/Rb declines from 1.27 to 1.20 in the same temperature range. Minor effects on the gA channel structure from lipid environment changes have been observed,33 and the higher αK/Rb values in POPC relative to saturated lipids may reflect such an effect. The αK/Na values are significantly larger for both lipid systems than observed in lysoPC phases. This could originate from the solid support influence on bilayer properties that affects conformational shifts needed to bind to the smaller Na+ ion. The selectivity values measured for Val by other techniques are 1.69 within DMPC at 50 °C by NMR and 2.63 at 10 °C within glycerol monooleate bilayers by electrochemical measurements (Table 1). The former value is very close to the chromatographic αRb/K observed for Val in DMPC above 25 °C. We observe lower αRb/K values within the gel phase of DMPC. An upward trend in αRb/K begins at 22 °C (Figure S15) that coincides with the phase transition for DMPC on SAPs observed by DSC, suggesting that membrane fluidity correlates with factors affecting Val binding selectivity. Val ion transport ceases in the gel phase, and it does not appear that the impact of bilayer gelation on ion binding has been reported previously. The αRb/K value within POPC bilayers is much higher than that within DMPC and is the same as that observed with monooleate bilayers (Table 1). When k-values of ionophore-free lipid-SAPs are used in place of those for bare SAPs for background correction, a minor reduction in the calculated α values involving Rb+ and K+ occurs (2−5%) and somewhat larger reduction occurs for αK/Na (up to 13%, Table S1), but the changes are not significant enough to alter the implications of the data. The similarities between the chromatographic α values and those from other techniques provide support for the eq 2 model between retention factors and binding affinity. Increasing Apparent KA with Anionic Lipids. Column retention is increased when a small fraction of negatively charged lipid (POPS, 1-palmitoyl-2-oleoyl-sn-glycero-3-phospho-L-serine) is added to the bilayer formulation which increases the concentration of ions near the ionophore binding sites (eq 2). Intrinsic binding constants for alkali ions to phosphatidylserine are very small,76 and the alkali retention observed on lipid-only SAPs containing 9:1 POPC/POPS is only slightly greater than that for POPC alone. Figure 6 contains chromatograms for Val-POPC, Val-POPC + POPS (1:8.1:0.9 mole ratio), and the SAP-columns containing only the lipid. When PS is included at 10 mol % in a POPC-only column, background ion retention increases by a factor of 2. Specific binding to Val increases more substantially however, and the background-corrected k-values increase from 0.07 to 0.37 for K+ and from 0.18 to 0.83 for Rb+, producing a value of 2.24 for αRb/K in this lipid system. It is uncertain whether the inclusion of POPS modifies bilayer properties enough to increase Val−ion binding through an effect other than that on the interfacial concentration of alkali ions. By a somewhat

Figure 6. (Top) Effect of the anionic lipid POPS on Val retention in POPC. Chromatograms for Val within POPC bilayers (top) and within POPC-POPS (8.1:0.9) bilayers (bottom).

smaller factor (3.1 vs 4.6), the specific retention to gA-POPC columns containing only 3 mol % gA was increased through the addition of PS (Figure S16). In view of the relative insensitivity of gA binding to lipid environment, this is expected to arise from the enhanced ion concentration factor and illustrates the potential to increase method sensitivity to very weak binding or columns at lower ionophore. Frontal Analysis of Ion Binding. The masses of lipid and ionophore on a column are known with reasonable accuracy through mass and TGA measurements, but the fraction of ionophore that is accessible to ions is not. Frontal analysis generates a binding isotherm where both the apparent KA and number of binding sites (nI, eq 3) are obtained from double reciprocal plots of bound ion versus ion concentration. Figure 7 shows data sets for K+ binding to gA-POPC measured over a five day period at 20 °C and with Val-DMPC at 15 and 22.5 °C. The changes in ion retention to Val arise from a change in apparent KA, which reduces from nearly 2500 M−1 with 1 nmol of binding sites to 226 M−1 with 1.5 nmol of binding sites. The changes with gA-columns arise from increases in the number of binding sites, and go from only 1.7 to nearly 4 nmol, whereas the KA values of about 300 M−1 are all within 15%. In all cases, the number of binding sites is a fraction of the ionophore on the column (8 nmol) and the apparent KA values are substantially higher than reported intrinsic ones. Substantial changes to the calculated values occur when the infused ion concentration is corrected using surface potential (ψo) contributions from the retained ions. The increasing surface charge from bound cations will cause the interface-tobulk ion ratio to change across the binding isotherm. Assuming that the cationic complex is distributed evenly on the surface area of the SAPs within the column, the surface charge per unit area (C/m2) that arises from binding can be calculated and used to model the effect of increasing ψo on the interfacial concentration of cations. The Grahame equation relates surface charge to ψo in an electrolyte solution, and a 9416

DOI: 10.1021/acs.langmuir.9b01301 Langmuir 2019, 35, 9410−9421

Article

Langmuir

As discussed previously, a negative potential near the binding sites likely exists from ionized silanol group on the SAPs. A surface density of 0.2 ionized silanol groups per nm2, based on data from thermally annealed silica,68 would generate a surface potential of approximately −55 mV in the 50 mM buffer system utilized, but a majority of the binding sites should be distanced several nanometers from the surface charges and nearer to the membrane interface where the potential would be diminished. Figure 8 contains plots of the apparent KA and nI predicted using different initial ψo values (no bound cation) to which the charge density from bound cations is added. In all cases, the reciprocal plots remain linear. The surface concentration of K+ at higher injected concentrations is affected in greater proportion than that at lower concentrations which results in smaller intercepts (larger nI) and steeper slopes of the reciprocal plots. Although the actual ψo is unknown, similar values of nI result from modeled values between zero and −20 mV, with higher values causing sharper divergence to lower nI values, or in the case of the day-five gAPOPC data, to unrealistically high estimates of nI. The predicted values for KA decrease with increasingly negative surface potential starting points. Given a Debye length of about 1.4 nm in the buffer, an interfacial potential of less than 25 mV at the membrane seems reasonable. The consequence for the evaluation of gA-POPC binding is that the predicted fraction of accessible ionophore increases to nearly 80% of the 8 nmol of ionophore on the column (after five days of hydration) from less than 50% when surface potential effects are neglected. The apparent KA values, greater than 300 M−1 without correcting for ψo, immediately decrease to within a factor of 2 of the intrinsic KA determined by Hinton et al. (extrapolated to 20 °C)34 and match it when the initial surface potential is modeled to be about −20 mV. The KA values for gA evaluated on day 1 and day 5 agree within 10% throughout modeled potentials to −30 mV, whereas the number of binding sites consistently differs by >100%. An interpretation that is consistent with this data is that a slow conversion of the peptide to the channel conformation occurs within the supported bilayers. This is further supported by the blue shift in laurdan emission from gA-POPC SAPs over the incubation period (Figure S17),

Figure 7. (Top) Set of frontal plateaus from continuous injection of 0.125−4 mM K+ ion on gA-POPC. (Middle) Double reciprocal plots of K+ frontal data for a gA-POPC at different hydration periods, and (bottom) data sets for K+ on a Val-DMPC column at two temperatures where retention is substantially different.

Boltzmann relationship determines the interfacial to bulk concentration ratio of monovalent ions in relation to it (see Supporting Information Section S1.8). This approach has been used previously in the study of peptide binding to charged vesicles.77

Figure 8. Corrected values for KA and nI from the frontal analysis data sets in Figure 7 arising from different starting point surface potentials modified by ion binding. (a,b) Data for gA-POPC at three time periods post-hydration, and (c,d) Val-DMPC at two temperatures. Values in black fill are the apparent values determined without consideration of surface potential. Each column nominally contains 8 nmol of ionophore. 9417

DOI: 10.1021/acs.langmuir.9b01301 Langmuir 2019, 35, 9410−9421

Article

Langmuir

interactions or binding to ionophores at lower loading densities. Particle-supported lipid bilayers provide a convenient platform to evaluate the effect of different lipid compositions and conditions on binding interactions; however, they may induce binding-behavior artifacts for interactions that are lipidphase sensitive through various physical influences on the bilayer.79−82 The materials used here may be viewed as a starting point for evaluating the efficacy of other lipid systems with this methodology. Liposomes are more amenable to surface potential characterization than the lipid-SAP particles, and their properties would be less influenced by the solid support which could increase the relatability of the data to that from other techniques, albeit at the cost of greater preparation time and perhaps, additional lipid formulation constraints. Alternatively, modified supports in the form of cushioned-lipid bilayers may reduce the substrate influence on the lipid phase.83 Nevertheless, the lipid-SAP system as demonstrated is capable of distinguishing the sensitivity of ion-binding interactions with Val and gA to temperature-induced changes in the lipid environment and revealing characteristic differences in ionophore accessibility in the bilayers through evaluation of binding site availability.

indicating that the peptide is increasingly impacting the lipid in a manner that is consistent with the known effect that gA channels have on diacyl-lipid bilayer properties.78 Zonal elution retention to gA in DMPC and DPPC was reduced relative to POPC (Figure 3). For both lipids, frontal data return apparent KA values similar to that with POPC, but with a smaller number of binding sites (Figure S18). The few columns evaluated beyond the five day incubation period did not show significant increases in binding, but it is not clear whether the time or storage conditions are sufficient to equilibrate these supported gA-lipid systems. For Val within DMPC at 15 °C, the calculated value of about 1.1 nmol for nI changes little, whereas the derived KA values decrease steadily with more negative initial ψo (Figure 8). For binding at 22.5 °C, where Val-DMPC bilayers are transitioning between the gel and fluid phase on SAPs, the apparent nI increases to about 3 nmol and there is a steep decrease for KA. These data indicate that there are roughly three times fewer Val binding sites in the gel-phase DMPC than in the more fluid phase, but binding occurs with substantially larger affinity. That only about 40% (approximately 3 out of 8 nmol) of the Val is binding-accessible at the higher temperature that may indicate that it is not readily diffusing across the supported lipid bilayers. At higher temperatures, binding is too weak to generate meaningful frontal analysis data sets. Within the gel phase, populations of the ionophore may be inaccessible within the hydrophobic core of the bilayer or at the membrane interface proximal to silica, or due to conformational restrictions in the viscous bilayer.

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.langmuir.9b01301.

■

Detailed descriptions of methods, materials, procedures, and calculations; representative TGA of lipid-SAPs, scanning electron microscopy images of SAPs, chromatogram showing void maker elution with mobile phase additive dip, chromatograms for bare silica SAPs, chromatograms showing effect of Cs+ in the mobile phase, DSC thermogram for DMPC-SAP, chromatograms for three gA-POPC over 5 day period, graph of ionic strength dependence of retention, discussion on sources of retention factor uncertainty, comparison of zonal and frontal methods for linear condition determination with gA and Val, plot of temperature versus k for 10 and 3 mol % Val-DMPC, DMPC-Val chromatograms with NLC and EMG fits and plot of kd versus temperature with discussion, van’t Hoff plots for three data sets each for gA and Val with tabulated data, chromatogram for nigericin-POPC column, laurdan GP plots versus temperature, plots of selectivity versus temperature, table of selectivity with values calculated alternative background correction, chromatogram for gA-POPC at 3 mol % with and without POPS, fluorescence spectra of laurdan in gA-POPC SAPs, and frontal dataset for gA-DMPC and DPPC (PDF)

SUMMARY AND CONCLUSIONS These studies have demonstrated the feasibility of using affinity chromatography to probe ion/ionophore binding within lipid bilayers. Given the general correlation, both positive and negative, of binding affinity with transport activity for many ionophore classes and the dearth of techniques suitable for measuring the interaction within lipid bilayers, this approach might serve to compliment other techniques used in the development of membrane ionophores by providing a means to quickly examine the effect of different lipid environments, or membrane and solution components on binding selectivity and relative affinity. The lipid columns can be prepared in about the same time as liposomes and can directly observe binding selectivity within the period of a single chromatogram. In comparison to solution NMR, the methodology requires far less costly equipment, less lipid and ionophore, and no isotopically labeled components. Frontal analysis isotherms allow discrimination of changes in binding affinity from changes in ionophore accessibility which could provide structural insight into ionophore-bilayer interactions. Although uncertainty in the surface potential confounds the direct evaluation of intrinsic binding constants by zonal elution, the selectivity of ionophores within different lipid environments and measurements of binding enthalpy are possible without its consideration. The KA values determined by frontal analysis for gA approach intrinsic values determined by NMR when surface potential effects are modeled, though the unknown starting potential near the binding sites precludes a definitive determination of the value. Surface potential can be exploited to increase the apparent affinity of binding events at the interface, enabling the detection of weaker binding

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. ORCID

Eric E. Ross: 0000-0003-1820-5170 Author Contributions †

I.J. and T.J. equivalent contributions.

9418

DOI: 10.1021/acs.langmuir.9b01301 Langmuir 2019, 35, 9410−9421

Article

Langmuir Notes

(17) Tian, F.; Cross, T. A. Cation transport: an example of structural based selectivity 1 1Edited by I. B. Holland. J. Mol. Biol. 1999, 285, 1993−2003. (18) Clarke, H. J.; Howe, E. N. W.; Wu, X.; Sommer, F.; Yano, M.; Light, M. E.; Kubik, S.; Gale, P. A. Transmembrane Fluoride Transport: Direct Measurement and Selectivity Studies. J. Am. Chem. Soc. 2016, 138, 16515−16522. (19) Sham, S. S.; Shobana, S.; Townsley, L. E.; Jordan, J. B.; Fernandez, J. Q.; Andersen, O. S.; Greathouse, D. V.; Hinton, J. F. The Structure, Cation Binding, Transport, and Conductance of Gly15-Gramicidin A Incorporated into SDS Micelles and PC/PG Vesicles. Biochemistry 2003, 42, 1401−1409. (20) Hamidinia, S. A.; Tan, B.; Erdahl, W. L.; Chapman, C. J.; Taylor, R. W.; Pfeiffer, D. R. The Ionophore Nigericin Transports Pb2+ with High Activity and Selectivity: A Comparison to Monensin and Ionomycin. Biochemistry 2004, 43, 15956−15965. (21) Paulus, E. F.; Kurz, M.; Matter, H.; Vértesy, L. Solid-State and Solution Structure of the Salinomycin−Sodium Complex: Stabilization of Different Conformers for an Ionophore in Different Environments. J. Am. Chem. Soc. 1998, 120, 8209−8221. (22) Busschaert, N.; et al. Towards Predictable Transmembrane Transport: QSAR Analysis of Anion Binding and Transport. Chem. Sci. 2013, 4, 3036−3045. (23) Xie, Q.; Gokel, G.; Hernandez, J.; Echegoyen, L. Efficient Sodium Cation Transport Across Liposome Membranes using Synthetic Carriers. J. Am. Chem. Soc. 1994, 116, 690−696. (24) Ovchinnikov, Y. A. Physico-Chemical Basis of Ion Transport through Biological Membranes: Ionophores and Ion Channels. Eur. J. Biochem. 1979, 94, 321−336. (25) Taylor, R. W.; Chapman, C. J.; Pfeiffer, D. R. Effect of Membrane Association on the Stability of Complexes between Ionophore A23187 and Monovalent Cations. Biochemistry 1985, 24, 4852−4859. (26) Riddell, F. G.; Arumugam, S.; Brophy, P. J.; Cox, B. G.; Payne, M. C. H.; Southon, T. E. The Nigericin-Mediated Transport of Sodium and Potassium Ions through Phospholipid Bilayers Studied by Sodium-23 and Potassium-39 NMR Spectroscopy. J. Am. Chem. Soc. 1988, 110, 734−738. (27) Pascal, S. M.; Cross, T. A. High-Resolution Structure and Dynamic Implications for a Double-Helical Gramicidin A Conformer. J. Biomol. NMR 1993, 3, 495−513. (28) Chen, Y.; Wallace, B. A. Binding of Alkaline Cations to the Double-Helical Form of Gramicidin. Biophys. J. 1996, 71, 163−170. (29) Painter, G. R.; Pollack, R.; Pressman, B. C. Conformational dynamics of the carboxylic ionophore lasalocid A underlying cation complexation-decomplexation and membrane transport. Biochemistry 1982, 21, 5613−5620. (30) Feigenson, G. W.; Meers, P. R. 1H NMR study of valinomycin conformation in a phospholipid bilayer. Nature 1980, 283, 313−314. (31) Meers, P.; Feigenson, G. W. Location and Ion-Binding of Membrane-Associated Valinomycin, a Proton Nuclear Magnetic Resonance Study. Biochim. Biophys. Acta 1988, 938, 469−482. (32) Jing, N.; Urry, D. W. Ion Pair Binding of Ca2+ and Cl− Ions in Micellar-Packaged Gramicidin A. Biochim. Biophys. Acta, Biomembr. 1995, 1238, 12−21. (33) Separovic, F.; Gehrmann, J.; Milne, T.; Cornell, B. A.; Lin, S. Y.; Smith, R. Sodium Ion Binding in the Gramicidin A Channel. SolidState NMR Studies of the Tryptophan Residues. Biophys. J. 1994, 67, 1495−1500. (34) Hinton, J. F.; Whaley, W. L.; Shungu, D.; Koeppe, R. E., II; Millett, F. S. Equilibrium Binding Constants for the Group I Metal Cations with Gramicidin A Determined by Competition Studies and 205 Tl Nuclear Magnetic Resonance Spectroscopy. Biophys. J. 1986, 50, 539−544. (35) Blair, S. M.; Kempen, E. C.; Brodbelt, J. S. Determination of Binding Selectivities in Host-Guest Complexation by Electrospray/ Quadrupole Ion Trap Mass Spectrometry. J. Am. Soc. Mass Spectrom. 1998, 9, 1049−1059.

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This research was partially supported by the National Science Foundation under grant nos. 1214145 and 0959377. The research was supported in part by a grant to Gonzaga University from the Howard Hughes Medical Institute through the Undergraduate Science Education Program and by the O’Leary Scholars Program at Gonzaga. We thank Dr. James Brozik for helpful discussion pertaining to the manuscript.

■

REFERENCES

(1) Pressman, B. C. Biological Applications of Ionophores. Annu. Rev. Biochem. 1976, 45, 501−530. (2) Gokel, G. W.; Carasel, I. A. Biologically Active, Synthetic Ion Transporters. Chem. Soc. Rev. 2007, 36, 378−389. (3) De Riccardis, F.; Izzo, I.; Montesarchio, D.; Tecilla, P. Ion Transport through Lipid Bilayers by Synthetic Ionophores: Modulation of Activity and Selectivity. Acc. Chem. Res. 2013, 46, 2781−2790. (4) Bühlmann, P.; Pretsch, E.; Bakker, E. Carrier-Based Ion-Selective Electrodes and Bulk Optodes. 2. Ionophores for Potentiometric and Optical Sensors. Chem. Rev. 1998, 98, 1593−1688. (5) Xie, X.; Bakker, E. Determination of Effective Stability Constants of Ion-Carrier Complexes in Ion Selective Nanospheres with Charged Solvatochromic Dyes. Anal. Chem. 2015, 87, 11553. (6) Kolusheva, S.; Shahal, T.; Jelinek, R. Cation-Selective Color Sensors Composed of Ionophore−Phospholipid−Polydiacetylene Mixed Vesicles. J. Am. Chem. Soc. 2000, 122, 776−780. (7) Kelkar, D. A.; Chattopadhyay, A. The Gramicidin Ion Channel: A Model Membrane Protein. Biochim. Biophys. Acta, Biomembr. 2007, 1768, 2011−2025. (8) Andersen, O. S.; Koeppe, R. E.; Roux, B. Gramicidin Channels. IEEE Trans. NanoBioscience 2005, 4, 10−20. (9) Huang, X.; Borgström, B.; Stegmayr, J.; Abassi, Y.; Kruszyk, M.; Leffler, H.; Persson, L.; Albinsson, S.; Massoumi, R.; Scheblykin, I. G.; Hegardt, C.; Oredsson, S.; Strand, D. The Molecular Basis for Inhibition of Stemlike Cancer Cells by Salinomycin. ACS Cent. Sci. 2018, 4, 760−767. (10) Busschaert, N.; Park, S.-H.; Baek, K.-H.; Choi, Y. P.; Park, J.; Howe, E. N. W.; Hiscock, J. R.; Karagiannidis, L. E.; Marques, I.; Félix, V.; Namkung, W.; Sessler, J. L.; Gale, P. A.; Shin, I. A Synthetic Ion Transporter that Disrupts Autophagy and Induces Apoptosis by Perturbing Cellular Chloride Concentrations. Nat. Chem. 2017, 9, 667−675. (11) Kevin, D. A., II; Meujo, D. A.; Hamann, M. T. Polyether Ionophores: Broad-Spectrum and Promising Biologically Active Molecules for the Control of Drug-Resistant Bacteria and Parasites. Expert Opin. Drug Discovery 2009, 4, 109−146. (12) Valkenier, H.; Davis, A. P. Making a Match for Valinomycin: Steroidal Scaffolds in the Design of Electroneutral, Electrogenic Anion Carriers. Acc. Chem. Res. 2013, 46, 2898−2909. (13) Varma, S.; Rogers, D. M.; Pratt, L. R.; Rempe, S. B. Design principles for K+ selectivity in membrane transport. J. Gen. Physiol. 2011, 137, 479−488. (14) Antonenko, Y. N.; Yaguzhinsky, L. S. The ion selectivity of nonelectrogenic ionophores measured on a bilayer lipid membrane: nigericin, monensin, A23187 and lasalocid A. Biochim. Biophys. Acta, Biomembr. 1988, 938, 125−130. (15) Lockless, S. W. Determinants of Cation Transport Selectivity: Equilibrium Binding and Transport Kinetics. J. Gen. Physiol. 2015, 146, 3−13. (16) Gokel, G. W.; Daschbach, M. M. Coordination and Transport of Alkali Metal Cations through Phospholipid Bilayer Membranes by Hydraphile Channels. Coord. Chem. Rev. 2008, 252, 886−902. 9419

DOI: 10.1021/acs.langmuir.9b01301 Langmuir 2019, 35, 9410−9421

Article

Langmuir (36) Wang, K. W.; Tripathi, S.; Hladky, S. B. Ion-Binding Constants for Gramicidin a obtained from Water Permeability Measurements. J. Membr. Biol. 1995, 143, 247−257. (37) Eisenman, G.; Sandblom, J.; Neher, E. Interactions in cation permeation through the gramicidin channel. Cs, Rb, K, Na, Li, Tl, H, and effects of anion binding. Biophys. J. 1978, 22, 307−340. (38) Cranfield, C. G.; Bettler, T.; Cornell, B. Nanoscale Ion Sequestration to Determine the Polarity Selectivity of Ion Conductance in Carriers and Channels. Langmuir 2015, 31, 292−298. (39) Atanasov, V.; Knorr, N.; Duran, R. S.; Ingebrandt, S.; Offenhäusser, A.; Knoll, W.; Köper, I. Membrane on a Chip: A Functional Tethered Lipid Bilayer Membrane on Silicon Oxide Surfaces. Biophys. J. 2005, 89, 1780−1788. (40) Zopf, D.; Ohlson, S. Weak-Affinity Chromatography. Nature 1990, 346, 87−88. (41) Tateno, H.; Nakamura-Tsuruta, S.; Hirabayashi, J. Frontal Affinity Chromatography: Sugar-Protein Interactions. Nat. Protoc. 2007, 2, 2529−2537. (42) Hage, D. S. High-Performance Affinity Chromatography: A Powerful Tool for Studying Serum Protein Binding. J. Chromatogr. B: Anal. Technol. Biomed. Life Sci. 2002, 768, 3−30. (43) Schriemer, D. C. Biosensor Alternative: Frontal Affinity Chromatography. Anal. Chem. 2004, 76, 448A. (44) Kim, H. S.; Wainer, I. W. Rapid Analysis of the Interactions between Drugs and Human Serum Albumin (HSA) using HighPerformance Affinity Chromatography (HPAC). J. Chromatogr. B: Anal. Technol. Biomed. Life Sci. 2008, 870, 22−26. (45) Iwashita, K.; Shiraki, K.; Ishii, R.; Tanaka, T.; Hirano, A. Liquid Chromatographic Analysis of the Interaction between Amino Acids and Aromatic Surfaces using Single-Wall Carbon Nanotubes. Langmuir 2015, 31, 8923−8929. (46) Gertler, G.; Fleminger, G.; Rapaport, H. Characterizing the Adsorption of Peptides to TiO2 in Aqueous Solutions by Liquid Chromatography. Langmuir 2010, 26, 6457−6463. (47) Király, Z.; Findenegg, G. H.; Mastalir, Á . Adsorption of Dodecyltrimethylammonium Bromide and Sodium Bromide on Gold Studied by Liquid Chromatography and Flow Adsorption Microcalorimetry. Langmuir 2006, 22, 3207−3213. (48) Tsai, C.-W.; Ruaan, R.-C.; Liu, C.-I. Adsorption of Antimicrobial Indolicidin-Derived Peptides on Hydrophobic Surfaces. Langmuir 2012, 28, 10446−10452. (49) Urbano, L.; Clifton, L.; Ku, H. K.; Kendall-Troughton, H.; Vandera, K.-K. A.; Matarese, B. F. E.; Abelha, T.; Li, P.; Desai, T.; Dreiss, C. A.; Barker, R. D.; Green, M. A.; Dailey, L. A.; Harvey, R. D. Influence of the Surfactant Structure on Photoluminescent πConjugated Polymer Nanoparticles: Interfacial Properties and Protein Binding. Langmuir 2018, 34, 6125−6137. (50) Lundahl, P.; Yang, Q. Liposome chromatography: liposomes immobilized in gel beads as a stationary phase for aqueous column chromatography. J. Chromatogr. 1991, 544, 283−304. (51) Pidgeon, C.; Venkataram, U. V. Immobilized artificial membrane chromatography: Supports composed of membrane lipids. Anal. Biochem. 1989, 176, 36−47. (52) Zhang, Y.; Xiao, Y.; Kellar, K. J.; Wainer, I. W. Immobilized Nicotinic Receptor Stationary Phase for on-Line Liquid Chromatographic Determination of Drug-Receptor Affinities. Anal. Biochem. 1998, 264, 22−25. (53) Moravcová, D.; Planeta, J.; Wiedmer, S. K. Silica-Based Monolithic Capillary Columns Modified by Liposomes for Characterization of Analyte-Liposome Interactions by Capillary Liquid Chromatography. J. Chromatogr. A 2013, 1317, 159−166. (54) Hu, W.; Haddad, P. R.; Hasebe, K.; Mori, M.; Tanaka, K.; Ohno, M.; Kamo, N. Use of a Biomimetic Chromatographic Stationary Phase for Study of the Interactions Occurring between Inorganic Anions and Phosphatidylcholine Membranes. Biophys. J. 2002, 83, 3351−3356. (55) Ross, E. E.; Hoag, C.; Pfeifer, Z.; Lundeen, C.; Owens, S. Metal Ion Binding to Phospholipid Bilayers Evaluated by Microaffinity Chromatography. J. Chromatogr. A 2016, 1451, 75−82.

(56) Ross, E. E.; Mok, S.-W.; Bugni, S. R. Assembly of Lipid Bilayers on Silica and Modified Silica Colloids by Reconstitution of Dried Lipid Films. Langmuir 2011, 27, 8634−8644. (57) Stark, G.; Ketterer, B.; Benz, R.; Läuger, P. The Rate Constants of Valinomycin-Mediated Ion Transport through Thin Lipid Membranes. Biophys. J. 1971, 11, 981−994. (58) Hladky, S. B.; Leung, J. C.; Fitzgerald, W. J. The Mechanism of Ion Conduction by Valinomycin: Analysis of Charge Pulse Responses. Biophys. J. 1995, 69, 1758−1772. (59) Woolf, T. B.; Roux, B. The Binding Site of Sodium in the Gramicidin A Channel: Comparison of Molecular Dynamics with Solid-State NMR Data. Biophys. J. 1997, 72, 1930−1945. (60) Mclaughlin, S. The Electrostatic Properties of Membranes. Annu. Rev. Biophys. Biophys. Chem. 1989, 18, 113−136. (61) Fornstedt, T.; Guiochon, G. Theoretical Study of HighConcentration Elution Profiles and Large System Peaks in Nonlinear Chromatography. Anal. Chem. 1994, 66, 2116−2128. (62) Wahab, M. F.; Anderson, J. K.; Abdelrady, M.; Lucy, C. A. Peak Distortion Effects in Analytical Ion Chromatography. Anal. Chem. 2014, 86, 559−566. (63) Ng, E.; Schriemer, D. C. Emerging Challenges in Ligand Discovery: New Opportunities for Chromatographic Assay. Expert Rev. Proteomics 2005, 2, 891−900. (64) Droge, S. T. J. Dealing with Confounding pH-Dependent Surface Charges in Immobilized Artificial Membrane HPLC Columns. Anal. Chem. 2016, 88, 960−967. (65) LoGrasso, P. V.; Moll, F., III; Cross, T. A. Solvent History Dependence of Gramicidin A Conformations in Hydrated Lipid Bilayers. Biophys. J. 1988, 54, 259−267. (66) Patrick, J. W.; Gamez, R. C.; Russell, D. H. The Influence of Lipid Bilayer Physicochemical Properties on Gramicidin A Conformer Preferences. Biophys. J. 2016, 110, 1826−1835. (67) Izatt, R. M.; Pawlak, K.; Bradshaw, J. S.; Bruening, R. L. Thermodynamic and kinetic data for macrocycle interactions with cations and anions. Chem. Rev. 1991, 91, 1721−2085. (68) Zhuravlev, L. T. The surface chemistry of amorphous silica. Zhuravlev model. Colloids Surf., A 2000, 173, 1−38. (69) Macdonald, P. M.; Seelig, J. Anion Binding to Neutral and Positively-Charged Lipid Membranes. Biochemistry 1988, 27, 6769− 6775. (70) Talbert, A. M.; Tranter, G. E.; Holmes, E.; Francis, P. L. Determination of Drug−Plasma Protein Binding Kinetics and Equilibria by Chromatographic Profiling: Exemplification of the Method using L-Tryptophan and Albumin. Anal. Chem. 2002, 74, 446−452. (71) Sankaram, M. B.; Easwaran, K. R. K. Location of Valinomycin in Lipid Vesicles. J. Biosci. 1984, 6, 635−642. (72) Varma, S.; Sabo, D.; Rempe, S. B. K+/Na+ Selectivity in K Channels and Valinomycin: Over-coordination Versus Cavity-size constraints. J. Mol. Biol. 2008, 376, 13−22. (73) Parasassi, T.; De Stasio, G.; d’Ubaldo, A.; Gratton, E. Phase Fluctuation in Phospholipid Membranes Revealed by Laurdan Fluorescence. Biophys. J. 1990, 57, 1179−1186. (74) Boheim, G.; Hanke, W.; Eibl, H. Lipid phase transition in planar bilayer membrane and its effect on carrier- and pore-mediated ion transport. Proc. Natl. Acad. Sci. U.S.A. 1980, 77, 3403−3407. (75) Tian, F.; Lee, K.-C.; Hu, W.; Cross, T. A. Monovalent Cation Transport: Lack of Structural Deformation upon Cation Binding. Biochemistry 1996, 35, 11959−11966. (76) Rostovtseva, T. K.; Aguilella, V. M.; Vodyanoy, I.; Bezrukov, S. M.; Parsegian, V. A. Membrane Surface-Charge Titration Probed by Gramicidin A Channel Conductance. Biophys. J. 1998, 75, 1783− 1792. (77) Wenk, M. R.; Seelig, J. Magainin 2 Amide Interaction with Lipid Membranes: Calorimetric Detection of Peptide Binding and Pore Formation. Biochemistry 1998, 37, 3909−3916. (78) de Planque, M. R. R.; Greathouse, D. V.; Koeppe, R. E.; Schäfer, H.; Marsh, D.; Killian, J. A. Influence of Lipid/Peptide Hydrophobic Mismatch on the Thickness of Diacylphosphatidylcho9420

DOI: 10.1021/acs.langmuir.9b01301 Langmuir 2019, 35, 9410−9421

Article

Langmuir line Bilayers. A2H NMR and ESR Study Using Designed Transmembrane α-Helical Peptides and Gramicidin A. Biochemistry 1998, 37, 9333−9345. (79) Ahmed, S.; Madathingal, R. R.; Wunder, S. L.; Chen, Y.; Bothun, G. Hydration Repulsion Effects on the Formation of Supported Lipid Bilayers. Soft Matter 2011, 7, 1936−1947. (80) Ahmed, S.; Nikolov, Z.; Wunder, S. L. Effect of Curvature on Nanoparticle Supported Lipid Bilayers Investigated by Raman Spectroscopy. J. Phys. Chem. B 2011, 115, 13181−13190. (81) Roiter, Y.; Ornatska, M.; Rammohan, A. R.; Balakrishnan, J.; Heine, D. R.; Minko, S. Interaction of Nanoparticles with Lipid Membrane. Nano Lett. 2008, 8, 941−944. (82) Charrier, A.; Thibaudau, F. Main Phase Transitions in Supported Lipid Single-Bilayer. Biophys. J. 2005, 89, 1094−1101. (83) Wong, J. Y.; Majewski, J.; Seitz, M.; Park, C. K.; Israelachvili, J. N.; Smith, G. S. Polymer-Cushioned Bilayers. I. A Structural Study of various Preparation Methods using Neutron Reflectometry. Biophys. J. 1999, 77, 1445−1457.

9421

DOI: 10.1021/acs.langmuir.9b01301 Langmuir 2019, 35, 9410−9421