Competitive cation binding to phosphatidylinositol-4,5-bisphosphate domains revealed by X-ray fluorescence

Abstract

Phosphatidylinositol-4,5-bisphosphate (PIP₂) is an important signaling phospholipid in the inner leaflet of the cell membrane. Due to the high negative charge of its headgroup, PIP₂ strongly interacts with cellular cations. We have used synchrotron diffraction and fluorescence techniques to determine preferential cation binding to two-dimensional PIP₂ templates. The natural, highly unsaturated PIP₂ is manipulated as a Langmuir monolayer on a physiological buffer containing 100 mM KCl and varying amounts of Ca²⁺ and Mg²⁺. X-ray fluorescence shows an 800% surface enhancement in K⁺ concentration (bound to PIP₂) compared to bulk concentration. Adding physiological levels of Ca²⁺( 1–100 μM) results in gradual replacement of K⁺ by Ca²⁺ ions, leading to a significant change in the organization of the PIP₂ model membrane, while higher concentrations (100–1000 μM) lead to three orders of magnitude increase in surface [Ca²⁺]. Similar experiments with Mg²⁺ ions also show strong ion binding to PIP₂ at physiological levels (1 mM) with a lesser structural effect. For mixed solutions of Mg²⁺ and Ca²⁺ we find that Ca²⁺ occupies the majority of binding sites. Remarkably we find that, with both 1 mM Mg²⁺ and 1 mM Ca²⁺ in the subphase there is still a 400% surface enrichment of K⁺ ions at the headgroup region.

---

1 Introduction

Although a minor phospholipid in cell membranes, phosphatidylinositol-4,5-bisphosphate (PIP₂) is implicated in nearly all aspects of cell physiology.¹ Its versatile functions stem from the highly charged and phosphorylated inositol headgroup which is involved in numerous enzyme regulation events. For example, PIP₂ binds the tumor suppressor PTEN and activates it to prevent uncontrolled cell growth.^2–4 Through its highly phosphorylated headgroup, PIP₂ acts as a template for Ca²⁺ signaling and inwardly rectifying potassium channels.^5,6 PIP₂ is highly negatively charged at neutral pH and much of its signaling occurs via electrostatic interactions with various ions and charged moieties of proteins.^7,8 Specific cations interact with PIP₂ to mediate or reduce its interactions with signaling partners. For instance, it has been proposed that the highly cationic protein MARKS regulates PIP₂ by engulfing it with cationic lysine residues, thus neutralizing PIP₂ from binding to other targets until MARKS is removed via a calmodulin mediated mechanism.^9,10 Other polyvalent cations, such as neomycin and spermine, have been found to inhibit PIP₂ mediated activation of the ATP-sensitive potassium channel.¹¹

Ca²⁺ has been found to interact strongly with PIP₂ and cause clustering that potentially enhances its signaling roles.^12,13 Clustering has been observed in AFM imaging of Langmuir monolayers of PIP₂ transferred to solid support and by condensation of PIP₂ monolayers upon addition of Ca²⁺ to the physiologically prepared solution.¹³ In contrast, limited cluster formation is observed for monolayers of PIP₂ that are spread on Mg²⁺ solutions.¹³ On mixed Ca²⁺/Mg²⁺ solutions, a higher intrinsic binding constant is found for Ca²⁺ than Mg²⁺, and Ca²⁺ replaces Mg²⁺.¹³ It is well established that Ca²⁺ flux is commonly exploited by the cell to trigger signaling events, potentially via PIP₂.¹⁴

Here, we report on the competition between K⁺, Mg²⁺, and Ca²⁺ for PIP₂ binding sites in a two-dimensional template formed by a PIP₂ monolayer that is prepared at the vapor/solution interface as a model environment of the cytosol. The use of a lipid monolayer of natural, highly unsaturated, PIP₂ that forms a uniform headgroup template allows direct evaluation of specific cation–PIP₂ interactions without the complications of mixing with other lipids that may introduce non-specific ion binding. The pure PIP₂ monolayers used in this study resemble the domain formation of PIP₂ in model mixed-lipid membranes that may exist in living cells,^{12,13,15–17} especially in regions of active PIP₂ synthesis. Here we employ a combination of X-ray fluorescence and reflectivity to directly characterize cation binding to PIP₂. A previous study utilized X-ray reflectivity and grazing incidence X-ray diffraction to determine the influence of Ca²⁺ on a mixed DPPC/PIP₂ monolayer with emphasis on the mechanism of synaptic vesicle fusion.¹⁸ Due to the mixed nature of the lipid-film no quantitative information on the specific interaction of Ca²⁺ with PIP₂ is possible from those data. Our current study extends other work^13,18–21 in a few ways, in particular, by providing a quantitative measure of the number of bound ions per molecule and by systematically and specifically monitoring competition among cytoplasmic ions.

2 Materials and methods

2.1 Materials

L-α-Phosphatidylinositol-4,5-bisphosphate (brain, porcine–triammonium salt) [brain PI(4,5)P₂] was purchased from Avanti Polar Lipids (Birmingham, AL) in powder form and dissolved in a 20 : 9 : 1 by volume mixture of chloroform, methanol, and water to form a lipid stock. Concentration was determined by carefully weighing the lipid powder on an analytical balance before dissolving. CaCl₂, MgCl₂, KCl, HCl, Tris, and EDTA of purity > 99% were purchased from Sigma Aldrich (St. Louis, MO), and the Ca/Mg/K salts were at least 99.99% pure based on trace metal basis. Buffer subphase contained 10 mM Tris and 100 mM KCl, along with varying amounts of CaCl₂, MgCl₂, and EDTA. Buffer pH was set to pH 7.2 ± 0.1 using high purity HCl. 100 mM KCl was used to replicate physiological salt levels. For buffers with no divalent cations, 0.1 mM EDTA was used to chelate any trace divalent cation impurities. MilliQ water was obtained from a MilliQ filter system and was determined to have 18.2 MΩ resistivity.

2.2 Experimental set-up and data analysis

X-ray reflectivity (XR) and near-total-reflection fluorescence (XNTRF) measurements were conducted on the liquid surface spectrometer at beamline 9ID-C, Advanced Photon Source, Argonne National Laboratory (E = 8.0 keV; wavelength λ = 1.5497 Å). ESI Fig. S1† provides a sketch of the experimental setup. The monolayer is formed by drop-wise deposit (at a molecular area of 85–100 Ų per molecule) of PIP₂ stock solution onto the aqueous surface of the subphase solution that is contained in a Langmuir trough and maintained at constant temperature of 20 °C. The Langmuir trough is encapsulated in an enclosure with thin Kapton windows and purged with water-saturated helium gas to minimize background scattering from air and potential X-ray radiation damage to the samples. An oxygen sensor (S101, Qubit System Inc.) in the trough is used to monitor oxygen levels in the enclosure. The monolayer is compressed by a Teflon barrier to a desired surface pressure of 30 ± 2 mN m⁻¹. Surface pressure is monitored by a microbalance made of a Wilhelmy microbalance and filter paper plate. The X-ray measurements are started after the oxygen–helium exchange reaches equilibrium, where the oxygen content inside the enclosure is reduced by a factor of at least 100 with respect to that of the ambient atmosphere. Radiation damage is monitored by systematically translating the trough laterally to probe fresh surfaces and examine measurement reproducibility.

The specular reflection is collected by a Bicron scintillation detector at an X-ray exit angle α_f = α_i, α_i being the incident angle. The angular-dependence of both XR and XNTRF are expressed as functions of Q_z, the z-component (surface normal) of the scattering (wave vector transfer) vector Q. Q_z is related to α_i via Q_z = (4π/λ)sin α_i, and Q_c represents the Q_z that corresponds to the critical angle for total reflection, α_c (= 0.154° at E = 8 keV).

The fluorescence signals are collected at a series of α_i near the critical angle α_c by a Vortex energy dispersive detector (EDD, silicon-drift, Vortex-90EX) with a collimator in the front end of the EDD probe, which accepts the X-ray fluorescence of emitted photons in the direction along the surface normal at ∼1° angular resolution.

X-ray reflectivity data, R(Q_z), are normalized to the calculated Fresnel reflectivity (R_F), for an ideally smooth, flat air–water interface.^22,23 The R/R_F data for a surface monolayer on an aqueous surface are accounted for in terms of a simplistic, two-box (i.e., two-layer) structural model, where one box contains the hydrophilic head group, and the other contains the hydrocarbon tail.^22,24,25 Each box is characterized by a uniform electron density (ρ) and vertical height (l) relative to the interface. Given (ρ_H, l_H) and (ρ_T, l_T), in which subscripts “H” and “T” represent the headgroup and the hydrocarbon tail respectively, a step-like, discrete ED profile across the interfaces, ρ₀(z), is constructed and its corresponding reflectivity R₀(Q_z) is calculated by the Parratt recursion method and smeared by a Gaussian^22,25 (see ESI†). The structural refinements are carried out through a least squares optimization method.²⁵ Surface fluorescence studies have been performed in the past^26,27 and analytic routines to quantify the surface enrichment of ions have been established^28,29 and are discussed in detail in the ESI.†

3 Results

3.1 Ca²⁺ reorients PIP₂ in model lipid membranes

Fig. 1(a) shows normalized XR, R/R_F, from natural PIP₂ monolayers at physiologically relevant subphase conditions in the absence and presence of Ca²⁺. Addition of Ca²⁺ results in dramatic changes in the structure of PIP₂. There is a systematic shift of the R/R_F minimum to lower momentum transfers, Q_z, and an overall rise in reflectivity as Ca²⁺ concentration increases. Qualitatively, these features indicate an increase in film thickness and a higher electron density in the film in the presence of Ca²⁺. Fig. 1(b) shows the corresponding electron (scattering) densities (ED) across the lipid layer calculated from the best fit parameters (see Table 1) to the reflectivity curves. The increase in ED is mainly in the headgroup region and the increase in thickness occurs in the acyl-chain region (see Table 1).²⁴ The increased density in the headgroup region is due to Ca²⁺ binding and screening of PIP₂ charge leading to a reduction of the molecular area due to reduced repulsion. As a result the hydrocarbon chains are able to pack more tightly (i.e. relative order increases, but they are still disordered as we observe no diffraction from the monolayer (data not shown)), and this results in a subsequent condensation of the PIP₂ monolayer in agreement with the condensing effect of divalent cations on PIP₂.^13,19–21 The XR data allows for an estimation of the number of Ca²⁺ bound to the headgroup of PIP₂, based on a space-filling model,^25,28,29 and the estimated electron density of the Ca²⁺ ion (see Table 1 and ESI†).

Fig. 1 X-ray reflectivity of the PIP₂ monolayer in the presence of varying [Ca²⁺]. (a) X-ray reflectivity measured from a PIP₂ monolayer spread on a pH 7.2 ± 0.1 subphase containing 10 mM Tris, 100 mM KCl, and varying amounts of CaCl₂ (from 0 to 1000 μM). (b) Electron density versus distance from the PIP₂ headgroup/aqueous interface for modeling the corresponding reflectivities shown in (a). Electron density profiles are calculated from a 2-box refinement to the normalized reflectivity with the parameters that are listed in Table 1.

Table 1 PIP₂ monolayer structural parameters derived from a 2 layer model fit to the reflectivity curves in Fig. 1(a)^a

Subphase (CaCl2)	0 mM	1 μM	10 μM	100 μM	1 mM
a Structural parameters derived from the reflectivity curves for a PIP2 monolayer spread on a pH 7.2 ± 0.1 subphase containing 10 mM Tris, 100 mM KCl, as well as 1 mM, 10 mM, 100 mM, or 1 mM CaCl2 respectively. lH (Å) describes the length of the headgroup region, ρH (e Å^−3) is the electron density of the headgroup region, lT (Å) is the length of the acyl chain region, ρT (e Å^−3) is the electron density of the acyl chain region, σ (Å) is the surface roughness, Amol (Å^2) is the area per molecule in the monolayer calculated from the X-ray reflectivity, and Amol2 (Å^2) is the area per molecule in the monolayer determined from the isotherms. Ca^2+/PIP2 is the number of surface-bound calcium ions (Ca^2+) per molecule (PIP2).
lH (Å)	10.3(7)	9.2(9)	7.6(4)	7.2(4)	8.2(6)
ρH (e Å^−3)	0.51(1)	0.56(2)	0.65(2)	0.72(2)	0.70(3)
lT (Å)	13.4(3)	14.7(4)	16.2(2)	18.1(2)	18.4(3)
ρT (e Å^−3)	0.323(3)	0.321(4)	0.322(2)	0.322(2)	0.307(4)
lH + lT (Å)	23.6(4)	24.0(5)	23.8(2)	25.2(2)	26.7(3)
σ (Å)	4.2(1)	4.3(2)	4.6(1)	4.7(1)	4.5(1)
Amol (Å^2)	64–69	58–62	54–55	48–49	47–49
Amol2 (Å^2)	70(5)	57(6)	48(10)	45(10)	42(10)
Ca^2+/PIP2	n/a	0.3 ± 0.5	0.6 ± 0.3	0.9 ± 0.4	1.6 ± 0.9

---

3.2 Mg²⁺ affects the structure of PIP₂ to a lesser degree

Fig. 2(a) shows X-ray reflectivity from PIP₂ monolayers on buffer subphases containing 1 mM Mg²⁺ and various Ca²⁺ concentrations. It should be noted that the XR is not very sensitive to accumulation of the lighter Mg²⁺ ion, as its replacement with water molecules in the headgroup region does not affect the ED appreciably. While the addition of 10 μM Ca²⁺ (to the 1 mM Mg²⁺) is hardly noticeable in the XR data, increasing Ca²⁺ concentration has a more significant effect, indicating the exchange of Mg²⁺ by Ca²⁺ even at inferior Ca²⁺ concentrations, as may occur during Ca²⁺ signaling. More quantitatively, Fig. 2(b) shows the ED profiles that best fit the reflectivity (solid lines Fig. 2(a)) with the structural parameters listed in Table 2. The XR data allows for an estimation of the number of divalent cations (M²⁺) bound to the headgroup of PIP₂, based on a space-filling model^25,28,29 and the estimated electron density of the Mg²⁺ ion (see Table 2 and ESI†). The volume and electron density of the ion is dependent on the hydration state. If we assume a dehydrated Mg²⁺ ion we calculate a minimum of ∼2 Mg²⁺ ions per PIP₂ headgroup. However, Slochower et al. have recently suggested that the Mg²⁺ ion remains hydrated while bound to PIP₂.³⁰ If we assume a hydration shell of six water molecules around each Mg²⁺ ion, we calculate a much lower Mg²⁺ density of ∼0.6 ions/PIP₂.

Fig. 2 X-ray reflectivity of the PIP₂ monolayer in the presence of 1 mM Mg²⁺ and varying [Ca²⁺]. (a) X-ray reflectivity measured from a PIP₂ monolayer spread on a pH 7.2 ± 0.1 subphase containing 10 mM Tris, 100 mM KCl, 1 mM MgCl₂, and varying amounts of CaCl₂ (from 0 to 1000 μM). (b) Electron density versus distance from the PIP₂ headgroup/aqueous interface for modeling the corresponding reflectivities shown in (a). Electron density profiles are calculated from a 2-box refinement to the normalized reflectivity with the parameters that are listed in Table 2.

Table 2 PIP₂ monolayer structural parameters derived from a 2 layer model fit to the reflectivity curves in Fig. 2(a)^a

Mg^2+ 1 mM	0 μM Ca^2+	10 μM Ca^2+	100 μM Ca^2+	1 mM Ca^2+
a Structural parameters derived from the reflectivity curves for a PIP2 monolayer spread on a pH 7.2 ± 0.1 subphase containing 10 mM Tris, 100 mM KCl, 1 mM MgCl2, as well as 0 μM, 10 μM, 100 μM, or 1 mM CaCl2 respectively. Parameters are the same as those for Table 1.
lH (Å)	10(1)	9(1)	9(2)	10.0(8)
ρH (e Å^−3)	0.57(4)	0.60(4)	0.63(5)	0.63(3)
lT (Å)	15.4(6)	16.0(6)	16.9(8)	17.0(4)
ρT (e Å^−3)	0.303(7)	0.308(8)	0.30(1)	0.300(5)
lH + lT (Å)	25.4(6)	25.3(6)	25.8(9)	27.0(4)
σ (Å)	4.2(2)	4.3(2)	4.3(3)	4.3(2)
Amol (Å^2)	55–64	53–57	50–55	51–58
Amol2 (Å^2)	54(6)	55(6)	56(6)	57(6)
Mg^2+/PIP2	2.0 ± 1.8	n/a	n/a	n/a

---

3.3 X-ray fluorescence

A more direct way to determine cation binding to PIP₂ model membranes is X-ray fluorescence. Fig. 3(a) shows X-ray fluorescence spectra (Q_z < 0.0217 Å⁻¹) for a pure 100 mM KCl buffer interface (no PIP₂ present, black squares) compared to the same interface in the presence of PIP₂ (red squares). The highly negative surface charge of PIP₂ attracts K⁺ from the buffer solution as expected from the Poisson–Boltzman theory of the double layer. The K⁺ surface concentration is enriched by a factor of eight to 0.8 M when PIP₂ is present as compared to 0.1 M in the bulk (surface concentration is based on the total volume calculated from the penetration depth of the X-ray (D ∼ 52 Åat Q_z = 0.01 Å⁻¹​) and the molecular area of the PIP₂ molecule (A ∼ 50 Ų, taken from isotherms and consistent with published results (Slochower et al.,²¹ see Fig. 1)). As we increase the [Ca²⁺] in the bulk, the surface [Ca²⁺] increases dramatically (green (1 μM), blue (0.1 mM), and orange (1 mM) data points) with a simultaneous reduction in [K⁺], as Ca²⁺ displaces K⁺ from the headgroup interface due to the higher binding constant for Ca²⁺ compared to that of K⁺.

Fig. 3 X-ray fluorescence reveals specific cation binding to the PIP₂ monolayer. (a) X-ray fluorescence signal from surface K⁺ and Ca²⁺ ions in the absence (black) or with a PIP₂ monolayer (red, labeled “pure subphase”) spread on a pH 7.2 ± 0.1 subphase containing 10 mM Tris, 100 mM KCl, and varying amounts of CaCl₂ as indicated (from 0 to 1000 μM). The monolayer is held at a constant surface pressure of 30 ± 2 mN m⁻¹. The fluorescence signal is integrated over Q_z values from 0.010 to 0.021 Å⁻¹, probing surface ions in the penetration depth of the evanescent beam (approximately 52 Å at Q_z ​= 0.01 Å⁻¹​). (b) Same as (a) but the solution includes additional 1 mM Mg²⁺ (dark green).

Fig. 3(b) compares the surface fluorescence (Q_z < 0.0217 Å⁻¹) for a PIP₂ monolayer on a subphase with 1 mM Mg²⁺, 100 mM K⁺ and varying amounts of Ca²⁺ with that of the pure subphase with no monolayer (black squares). Data from a 100 mM K⁺ (labelled “pure subphase”) subphase covered with a PIP₂ monolayer is shown as well (red squares). This and the bare subphase data are the same in Fig. 3(a and b). We note that the fluorescence signals from Mg²⁺ ions are too low in energy to be detected by the Vortex EDD, so we take advantage of the decrease in K⁺ signals to indirectly monitor Mg²⁺ exchange for K⁺. Indeed for 1 mM Mg²⁺ in the buffer solution we observe a significant decrease in the interfacial [K⁺] (dark green triangles) indicating Mg²⁺ migration to the interface ([K⁺] changes from 0.8 M to 0.5 M at the PIP₂ headgroup). By increasing concentrations of Ca²⁺ to the Mg²⁺ and K⁺ containing solution, prominent Ca²⁺ signals emerge (light green diamonds and light blue triangles) in accordance with our reflectivity results. The Ca²⁺ fluorescence saturates at roughly 1 mM Ca²⁺ (cyan data vs. green data).

We now treat the fluorescence data more quantitatively to determine the ion concentration and the number of bound ions per PIP₂ molecule for the various conditions described above (see ESI† for methods). Fig. 4(a) shows the ion density (K⁺ and Ca²⁺) at the PIP₂ headgroup for the different subphase conditions investigated. Multiplying the ion density by the area per lipid molecule (determined during preparation of the PIP₂ monolayer) we obtain the number of ions per lipid, Fig. 4(b). Fig. 4(a) and (b) thus summarize the observed ion density and ions/lipid with respect to Ca²⁺ concentration, with or without 1 mM Mg²⁺. In the absence of Ca²⁺ (or Mg²⁺) K⁺ density is high, with roughly 1.6 K⁺ per PIP₂ molecule (Fig. 4(b), black data). As the [Ca²⁺] increases, the K⁺ density drops off as it is replaced by Ca²⁺. Even for 1 mM Ca²⁺ there are still ∼0.5 K⁺ per PIP₂ molecule, compared to ∼1.6 Ca²⁺ per PIP₂.

Fig. 4 Calculated ion density for K⁺ and Ca²⁺ at the PIP₂ monolayer. (a) Ion density plotted vs. [Ca²⁺] for K⁺ and Ca²⁺ ions at the surface of the PIP₂ monolayer within a depth of ∼52 Å (at Q_z ​= 0.01 Å⁻¹​) of the headgroup/solution interface. Ion density of K⁺ and Ca²⁺ is compared for samples in the presence (blue and magenta respectively) and absence (black and red) of 1 mM Mg²⁺. (b) Number of ions per lipid plotted vs. [Ca²⁺] for K⁺ and Ca²⁺. Measured points in the absence of Ca²⁺ and Mg²⁺ are included at 10⁻³ M. Lines are guides to the eye.

As noted above, Mg²⁺ ion density could not be directly detected. However, the presence of 1 mM Mg²⁺ leads to a measurable reduction in K⁺ (blue data in Fig. 4(a and b)). In the absence of Ca²⁺, 1 mM Mg²⁺ reduces the K⁺ density to similar levels as observed for 1 mM Ca²⁺ indicating that the binding energies of both ions, although different, are of the same order of magnitude. We estimate ∼1 Mg²⁺ per PIP₂ at this concentration based on the observed decrease in K⁺ signal and the increase in ED, as compared to 1.6 Ca²⁺ per PIP₂.

Interestingly, a persistent K⁺ signal is observed in both cases (i.e. 1 mM Mg²⁺ alone and with increasing [Ca²⁺]), suggesting that some K⁺ ions reside in regions of PIP₂ domains that even large amounts of Ca²⁺ ions cannot replace. The observation of remaining surface signal for K⁺, even with high concentrations of Ca²⁺ or Mg²⁺, suggests that K⁺ plays an important role in PIP₂ signaling.

4 Discussion

The X-ray experiments provide a qualitative and quantitative picture of the interaction of PIP₂ domains with mono- and di-valent cations. They clearly show a hierarchy of associated ions at the interface.

4.1 K⁺ concentration increases 8-fold in the presence of PIP₂

Based on electrostatics, there is a significant surface enhancement of K⁺ in the presence of PIP₂ (approximately 8-fold, assuming a uniform K⁺ distribution within the X-ray penetration depth). At pH 7.2, PIP₂ has a single remaining exchangeable proton and carries an equivalent of four electron charges.¹⁶ Our experiment shows that ∼1.6 K⁺ ions per PIP₂ molecule associate with the headgroup. Based on a PIP₂ charge of −4,^16,31 this results in an effective charge of −2.4 for each PIP₂. By effective charge we refer to the charge experienced by PIP₂ targets within 1 nm from the charged monolayer. These remaining charges are compensated by ions in solution including the protons that are competing for space at the interface. The same is true with Mg²⁺ and Ca²⁺, i.e. the charges on PIP₂ that are not compensated by these ions are compensated by protons, either as bound ions or as a distribution within 1–2 nm from the interface.³² The overall K⁺ ion concentration at the interface increases from 0.10 ± 0.01 to 0.8 ± 0.1 M, consistent with what Toner et al. have suggested based on zeta potential measurements (Toner et al. determined 0.7–1.5 K⁺ ions/PIP₂ or 0.3–0.7 M (assuming a molecular area of PIP₂ of 67 Ų per molecule)), although our number, determined directly, is a bit higher.³³

4.2 Ca²⁺ interacts strongly and partially replaces K⁺

For Ca²⁺ concentrations that are two to five orders of magnitude lower than the K⁺ concentration, we find that a significant number of Ca²⁺ ions bind and replace K⁺ ions at the interface. Based on the K/Ca exchange at the interface, we estimate that the binding constant for Ca²⁺ is at least 100 times greater than that of K⁺. With 1 mM Ca²⁺ and 100 mM KCl, 1.6 Ca²⁺ and 0.5 K⁺ ions bind per PIP₂, reducing the effective charge of the PIP₂ to −0.3. If we account for the increased deprotonation of PIP₂ in the presence of divalent cations³¹ this charge increases to roughly −0.4. While the charge of PIP₂ is always neutralized over a distance of 1 nm or so from the interface, the presence of these tightly bound Ca²⁺ ions may reduce the effect of the PIP₂ charge and electrostatic interactions between PIP₂ and protein targets. This is similar to the effect Ca²⁺ ions have on DMPA as reported recently.²⁹

4.3 Mg²⁺ is replaced by Ca²⁺ but significant [K⁺] remain at the PIP₂ headgroup

Mg²⁺, despite its similarities to Ca²⁺ in terms of charge, plays a distinctly different role in cells. While [Ca²⁺] ranges from nano to micro-molar and varies dramatically, [Mg²⁺] remains relatively constant at ∼1 mM free Mg²⁺.³⁴ Although we could not directly detect Mg²⁺ ions, by monitoring K⁺ fluorescence signals at the interface we estimate the number of Mg²⁺ ions at ∼1 Mg²⁺ ion per PIP₂ at the physiological concentration of 1 mM Mg²⁺. This is consistent with our rough estimate of 0.6 Mg²⁺ per PIP₂ based on the increased electron density in the headgroup region observed in the XR data (see Fig. 1(b) and assuming a hydrated Mg²⁺ ion). At a subphase composition of 1 mM Mg²⁺ and 100 mM K⁺ we find ∼1 Mg²⁺ and 0.7 K⁺ ions bound per PIP₂ resulting in an effective charge of −1.3, a noticeably higher charge as compared to the charge observed in the presence of 1 mM Ca²⁺. As we add Ca²⁺ to the 1 mM Mg²⁺ solution, we observe very little change in the interfacial K⁺ concentration (see Fig. 4(a) and (b)) consistent with the fact that there are K⁺ ions that are tucked in deeply in the PIP₂ headgroup interface and are not replaced by either one or both divalent ions. These K⁺ ions may play a role in protein–PIP₂ interactions, altering the local charge and shape of PIP₂ as perceived by the protein's lipid-binding pocket. For a 1 : 1 Mg²⁺/Ca²⁺ ratio the Ca²⁺ almost completely replaces the Mg²⁺, while ∼0.5 K⁺ ions per PIP₂ remain at the headgroup region.

4.4 Biological implications

The levels of ion binding determined in this study likely match those found in vivo for several reasons. First, our ionic strength matches that of the cytosol of mammalian cells, and while changes in ionic strength and exact ion concentrations in the cytosol will impact the exact number of bound ions, the 100 mM KCl concentration provides a realistic value. Second, our pure PIP₂ monolayers match the in plane density that is found in 2D PIP₂-rich domains in the membrane since the surface pressure of our experiments was chosen to match that determined for lipid bilayers. Additionally, we previously showed that the global concentration of PIP₂ in model membranes has little impact on the charge of PIP₂, likely due to nano-clustering of PIP₂ resulting in PIP₂ residing in regions of high local concentration.¹⁶ Binding of ions to PIP₂ clusters induced by other lipid components (e.g. cholesterol and PI) is similarly not expected to lead to major deviations from the values we report here. Cholesterol and phosphatidylinositol (PI) were both shown to induce formation of macroscopic PIP₂ rich clusters in model lipid membranes.^15,31 However, in both cases a relatively small impact is observed on the charge of PIP₂, suggesting that ion binding may also remain consistent with the values that we report here.^31,35

The varying Ca²⁺ concentration in the cell may have a significant effect on PIP₂ mediated signaling events. Our results suggest that Ca²⁺ may have a dual effect on PIP₂-mediated signalling: (1) Ca²⁺ may trigger clustering of PIP₂ to bring PIP₂ binding partners in close proximity allowing for increased binding (signaling) and (2), Ca²⁺ may neutralize the PIP₂ charge leading to an effective sequestering effect similar to the MARCKS protein thus reducing the available pool of functional PIP₂ in the membrane.³⁶ This dual role of Ca²⁺ may be tightly regulated by the oscillating calcium concentration of the cell.¹⁴ In a resting cell, Ca²⁺ concentrations are quite low and only reach the micromolar range. According to our results, at these concentrations (1–10 μM) some Ca²⁺ binds to PIP₂ at a ratio of roughly 0.1 ions/PIP₂. With these reduced levels of Ca²⁺ we see an effect on the lipid packing in the presence of Ca²⁺ in the form of condensation of the PIP₂ monolayer. This condensation is likely a result of clustering of PIP₂ triggered by a screening effect of the calcium ions reducing the charge–charge repulsion between PIP₂ headgroups and enabling increased hydrogen bond formation between PIP₂ headgroups.¹⁶ Additionally, the calcium ions may play a ‘bridging role’ by linking adjacent headgroups. The increased interaction between PIP₂ headgroups drives the reduced molecular area and results in the increased acyl chain length (see Tables 1 & 2). Ca²⁺-induced clustering of PIP₂ at low [Ca²⁺] may lead to increased binding (signaling) of PIP₂ binding proteins due to the increased local PIP₂ concentration. Such binding can be stabilized by hydrogen bond interactions as summarized in the electrostatic-hydrogen bond switch model previously proposed for phosphatidic acid.³⁷ Previous studies have also suggested that 1 μM Ca²⁺ is enough to cause formation of small PIP₂ clusters.^12,13 However, these studies have not always accounted for the large number of competing cations (such as K⁺ and Mg²⁺, etc.) within the cytosol, which may reduce the impact of 1 μM Ca²⁺ and prevent the formation of Ca²⁺ induced PIP₂ clustering. Indeed, we find a reduced Ca²⁺ induced condensation in the presence of 1 mM Mg²⁺. During Ca²⁺ signaling events, large amounts of Ca²⁺ are released via calcium channels, and Ca²⁺ concentrations reach 10–100 μM.¹⁴ It is possible that near calcium channels local concentrations may even reach millimolar values. This can lead to large amounts of Ca²⁺ binding to PIP₂ to a peak of around 1.6 Ca²⁺ ions per lipid. At these high Ca²⁺ concentrations, Ca²⁺ will start to neutralize the PIP₂ charge and may play more of a sequestering role, while Ca²⁺ induced clustering will also be promoted. Many proteins that bind to PIP₂ rely on electrostatic interactions to dock to PIP₂, and the presence of Ca²⁺ ions may reduce or eliminate binding between a protein and PIP₂. Alternatively, the additional Ca²⁺ ions might be shared between two PIP₂ molecules, acting as a bridge to promote stronger PIP₂ clustering. Thus, we suggest that Ca²⁺ may facilitate PIP₂-mediated signalling events in different ways depending on its local concentration. At moderate [Ca²⁺] PIP₂ signaling might be facilitated due to clustering and enhanced binding, while at higher [Ca²⁺] PIP₂ signaling may be inhibited due to charge neutralization.

K⁺ concentration in the cytosol is between 100 and 150 mM, and that of Na⁺ one or two orders of magnitude lower (the reverse is found on the outside of a typical mammalian cell). Coupled with the intracellular location of PIP₂, the observation that the PIP₂ binding constant for K⁺ is ∼100 fold lower than the Ca²⁺ binding constant, and the fact that the fluorescence signals for Na⁺ are too weak to be detected by the Vortex EDD is the reason we focused on K⁺ in this work. The PIP₂ binding affinity of Na⁺ is expected to be similar to that of K⁺. Our results show that as many as 1.6 K⁺ ions per PIP₂ are present in PIP₂ rich domains. The cytosol also commonly contains roughly 1 mM free Mg²⁺. At this concentration roughly one Mg²⁺ ion binds to each PIP₂, replacing approximately one K⁺ ion, but maintaining a significant amount (∼0.5 ions) of K⁺ bound to the PIP₂ headgroup.

5 Conclusions

Employing surface sensitive (synchrotron) X-ray diffraction and spectroscopic techniques we determine the structure of natural PIP₂ in the form of a monolayer on physiological solutions with various relevant ions to quantify ion adsorption and binding to PIP₂ domains in plasma membranes. We find that in buffer solutions with 100 mM KCl approximately 1.6 K⁺ ions are associated with each headgroup of PIP₂ increasing by eight-fold the concentration at the surface compared to that in the bulk. Upon addition of Ca²⁺ or Mg²⁺ our results show a few effects: first, the divalent ions replace the K⁺, and second, binding leads to increased packing of the PIP₂ monolayer, consistent with recent observations.^13,19,21,31 Remarkably, we find that at least 0.5 K⁺/PIP₂ remain at the PIP₂ interface even at physiological levels of Ca²⁺ or Mg²⁺. K⁺ ions tightly bound to PIP₂ effectively change the binding partner for PIP₂ binding proteins from just PIP₂ to PIP₂–K⁺. This should be taken into account when considering protein–PIP₂ interactions, e.g. in MD simulations.

For mixed Mg²⁺/Ca²⁺ we demonstrate that Ca²⁺ has a stronger affinity to binding than Mg²⁺ as it replaces most of the bound Mg²⁺ ions. Our results also show that the divalent ions have measurable effect on the structure of PIP₂ in domains in particular on the molecular area that can decrease by almost 20% in the presence of 1 mM Ca²⁺ or Mg²⁺ in physiological solutions.²¹ Such an increase in in-plane density may be related to clustering of PIP₂ in cells and PIP₂ clusters may form important platforms for PIP₂ signaling,^38–41 for example by increasing protein binding affinity as was recently demonstrated for PIP₂ clusters induced by cholesterol.¹⁵

Acknowledgements

ZTG and EEK gratefully acknowledge financial support from NSF through grant CHE 1216827. The work at the Ames Laboratory was supported by the Office of Basic Energy Sciences, U.S. Department of Energy under Contract No. DE-AC02-07CH11358. Use of the Advanced Photon Source was supported by the U. S. Department of Energy, Office of Science, Office of Basic Energy Sciences, under Contract No. DE-AC02-06CH11357.

Notes and references

1. G. Di Paolo and P. de Camilli, Nature, 2006, 443, 651–657.
2. N. R. Leslie, I. H. Batty, H. Maccario, L. Davidson and C. P. Downes, Oncogene, 2008, 27, 5464–5476.
3. R. E. Redfern, D. Redfern, M. L. Furgason, M. Munson, A. H. Ross and A. Gericke, Biochemistry, 2008, 47, 2162–2171.
4. A. H. Ross and A. Gericke, Proc. Natl. Acad. Sci. U. S. A., 2009, 106, 1297–1298.
5. B. Hille, E. J. Dickson, M. Kruse, O. Vivas and B. C. Suh, Biochim. Biophys. Acta, 2015, 1851, 844–856.
6. M. Soom, R. Schonherr, Y. Kubo, C. Kirsch, R. Klinger and S. H. Heinemann, FEBS Lett., 2001, 490, 49–53.
7. M. Langner, D. Cafiso, S. Marcelja and S. McLaughlin, Biophys. J., 1990, 57, 335–349.
8. R. V. Stahelin, J. L. Scott and C. T. Frick, Chem. Phys. Lipids, 2014, 182, 3–18.
9. A. Arbuzova, J. Wang, D. Murray, J. Jacob, D. S. Cafiso and S. McLaughlin, J. Biol. Chem., 1997, 272, 27167–27177.
10. S. McLaughlin, G. Hangyas-Mihalyne, I. Zaitseva and U. Golebiewska, Biochem. Soc. Symp., 2005, 189–198.
11. Z. Fan and J. C. Makielski, J. Biol. Chem., 1997, 272, 5388–5395.
12. M. J. Sarmento, A. Coutinho, A. Fedorov, M. Prieto and F. Fernandes, Biochim. Biophys. Acta, 2013, 1838, 822–830.
13. Y. H. Wang, A. Collins, L. Guo, K. B. Smith-Dupont, F. Gai, T. Svitkina and P. A. Janmey, J. Am. Chem. Soc., 2012, 134, 3387–3395.
14. M. D. Bootman, P. Lipp and M. J. Berridge, J. Cell Sci., 2001, 114, 2213–2222.
15. Z. Jiang, R. E. Redfern, Y. Isler, A. H. Ross and A. Gericke, Chem. Phys. Lipids, 2014, 182, 52–61.
16. E. E. Kooijman, K. E. King, M. Gangoda and A. Gericke, Biochemistry, 2009, 48, 9360–9371.
17. I. L. Salvemini, D. M. Gau, J. Reid, L. A. Bagatolli, A. Macmillan and P. D. Moens, Chem. Phys. Lipids, 2014, 177, 51–63.
18. S. K. Ghosh, S. Castorph, O. Konovalov, T. Salditt, R. Jahn and M. Holt, Biophys. J., 2012, 102, 1394–1402.
19. I. Levental, A. Cebers and P. A. Janmey, J. Am. Chem. Soc., 2008, 130, 9025–9030.
20. I. Levental, D. A. Christian, Y. H. Wang, J. J. Madara, D. E. Discher and P. A. Janmey, Biochemistry, 2009, 48, 8241–8248.
21. D. R. Slochower, Y. H. Wang, R. Radhakrishnan and P. A. Janmey, Phys. Chem. Chem. Phys., 2015, 17, 12608–12615.
22. J. Als-Nielsen and D. McMorrow, Elements of modern X-ray physics, Wiley, Hoboken, 2nd edn, 2011.
23. P. S. Pershan and M. L. Schlossman, Liquid Surfaces and Interfaces: Synchrotron X-ray methods, Cambridge University Press, New York, 2012.
24. K. Kjaer, Phys. B, 1994, 198, 100–109.
25. D. Vaknin, in Characterization of Materials, ed. E. N. Kaufmann, John Wiley & Sons, New York, 2012, vol. 2, pp. 1393–1423.
26. J. Daillant, L. Bosio, J. J. Benattar and C. Blot, Langmuir, 1991, 7, 611–614.
27. W. B. Yun and J. M. Bloch, J. Appl. Phys., 1990, 68, 1421–1428.
28. W. Bu, K. Flores, J. Pleasants and D. Vaknin, Langmuir, 2009, 25, 1068–1073.
29. W. Wang, N. A. Anderson, A. Travesset and D. Vaknin, J. Phys. Chem. B, 2012, 116, 7213–7220.
30. D. R. Slochower, P. J. Huwe, R. Radhakrishnan and P. A. Janmey, J. Phys. Chem. B, 2013, 117, 8322–8329.
31. Z. T. Graber, A. Gericke and E. E. Kooijman, Chem. Phys. Lipids, 2014, 182, 62–72.
32. W. Bu, D. Vaknin and A. Travesset, Langmuir, 2006, 22, 5673–5681.
33. M. Toner, G. Vaio, A. McLaughlin and S. McLaughlin, Biochemistry, 1988, 27, 7435–7443.
34. B. Alberts, Molecular Biology of the Cell: Reference edition, Garland Science, 2008.
35. Z. T. Graber, Z. Jiang, A. Gericke and E. E. Kooijman, Chem. Phys. Lipids, 2012, 165, 696–704.
36. J. Y. Wang, S. McLaughlin and D. Murray, Biophys. J., 2003, 84, 461a.
37. E. E. Kooijman, D. P. Tieleman, C. Testerink, T. Munnik, D. T. Rijkers, K. N. Burger and B. de Kruijff, J. Biol. Chem., 2007, 282, 11356–11364.
38. X. Gao, P. R. Lowry, X. Zhou, C. Depry, Z. Wei, G. W. Wong and J. Zhang, Proc. Natl. Acad. Sci. U. S. A., 2011, 108, 14509–14514.
39. A. Honigmann, G. van den Bogaart, E. Iraheta, H. J. Risselada, D. Milovanovic, V. Mueller, S. Mullar, U. Diederichsen, D. Fasshauer, H. Grubmuller, S. W. Hell, C. Eggeling, K. Kuhnel and R. Jahn, Nat. Struct. Mol. Biol., 2013, 20, 679–686.
40. K. Kwiatkowska, Cell. Mol. Life Sci., 2010, 67, 3927–3946.
41. L. Picas, J. Viaud, K. Schauer, S. Vanni, K. Hnia, V. Fraisier, A. Roux, P. Bassereau, F. Gaits-Iacovoni, B. Payrastre, J. Laporte, J. B. Manneville and B. Goud, Nat. Commun., 2014, 5, 5647.

---

Footnotes

† Electronic supplementary information (ESI) available: Calculation of molecular area from reflectivity data, and calculation of # of divalent ions bound to PIP₂. See DOI: 10.1039/c5ra19023a

‡ Currently at: Department of Chemistry, 231 S 34th street, University of Pennsylvania, Philadelphia, PA 19104.

---