Poly(Ethylene Glycol)-Cholesterol Inhibits L-Type Ca2+ Channel Currents and Augments Voltage-Dependent Inactivation in A7r5 Cells

Cholesterol distributes at a high density in the membrane lipid raft and modulates ion channel currents. Poly(ethylene glycol) cholesteryl ether (PEG-cholesterol) is a nonionic amphipathic lipid consisting of lipophilic cholesterol and covalently bound hydrophilic PEG. PEG-cholesterol is used to formulate lipoplexes to transfect cultured cells, and liposomes for encapsulated drug delivery. PEG-cholesterol is dissolved in the external leaflet of the lipid bilayer, and expands it to flatten the caveolae and widen the gap between the two leaflets. We studied the effect of PEG-cholesterol on whole cell L-type Ca2+ channel currents (I Ca,L) recorded from cultured A7r5 arterial smooth muscle cells. The pretreatment of cells with PEG-cholesterol decreased the density of I Ca,L and augmented the voltage-dependent inactivation with acceleration of time course of inactivation and negative shift of steady-state inactivation curve. Methyl-β-cyclodextrin (MβCD) is a cholesterol-binding oligosaccharide. The enrichment of cholesterol by the MβCD:cholesterol complex (cholesterol (MβCD)) caused inhibition of I Ca,L but did not augment voltage-dependent inactivation. Incubation with MβCD increased I Ca,L, slowed the time course of inactivation and shifted the inactivation curve to a positive direction. Additional pretreatment by a high concentration of MβCD of the cells initially pretreated with PEG-cholesterol, increased I Ca,L to a greater level than the control, and removed the augmented voltage-dependent inactivation. Due to the enhancement of the voltage-dependent inactivation, PEG-cholesterol inhibited window I Ca,L more strongly as compared with cholesterol (MβCD). Poly(ethylene glycol) conferred to cholesterol the efficacy to induce sustained augmentation of voltage-dependent inactivation of I Ca,L.


Introduction
Cholesterol, a rigid lipid, embedded in the hydrophobic core of the lipid bilayer with its polar single hydroxyl group on the surface of the membrane, stabilizes the structure of the bilayer [1,2]. It flip-flops between the external and internal leaflets of the bilayer to establish equilibrated distribution [3,4]. Cholesterol and sphingomyelin accumulate on lipid rafts with channels and signaling proteins to construct platforms of cellular signaling [5]. Enrichment and depletion of cholesterol utilizing methyl-b-cyclodextrin (MbCD), a cholesterol-binding oligosaccharide [6] have provided evidence to establish that cholesterol is indispensable in the regulation of ion channel function [7]. It regulates the channel activity in lipid media by controlling the physical properties of the bilayer and at the lipid-protein interface by direct interaction with the channel protein [8].
Poly(ethylene glycol) cholesteryl ether (PEG-cholesterol), is a nonionic amphiphile consisting of hydrophobic cholesterol and covalently bound hydrophilic PEG [9]. PEG-cholesterol is watersoluble and is used to formulate lipoplexes to transfect cultured cells [10], and liposomes for encapsulated drug delivery [11]. Since PEG moiety decelerates flip-flop, PEG-cholesterol is accumulated in the outer leaflet of the bilayer in the human skin fibroblast [12], and flattens the caveolae in the K562 human leukemic cell line [13]. The accumulation of PEG-cholesterol produces bumpy protrusions of the external leaflet of the bilayer in human erythrocytes [13,14]. PEG-cholesterol inhibits raft-dependent endocytosis in HT-1080 human fibrosarcoma cells [9], fibroblasts [12] and leucocytes [13]. The effect of PEG-cholesterol on the function of ion channels has not yet been reported to the best of our knowledge.
L-type Ca 2+ channel currents (I Ca,L ) through opened Ca V 1.2 channels supply Ca 2+ into the muscle to initiate contraction of arterial smooth muscles (ASMCs). In the ASMCs, I Ca,L is not generated by action potentials but by moderate sustained depolarization induced by neurotransmitters, hormones, autacoids and mechanical stress [15]. The I Ca,L during the sustained depolarization is called window current (I WD ), which is determined by the number of Ca V 1.2 channels and their voltage-dependent activation and inactivation (VDI) [16,17]. The enrichment of cholesterol of swine coronary ASMCs by in vitro manipulation with cholesterol-saturated MbCD (cholesterol (MbCD)) induces inhibition of I Ca,L [18]. Voltage-gated Ca 2+ channel currents of murine pancreatic b-cells are also inhibited by the enrichment with cholesterol (MbCD) [19]. However, how VDI of Ca V 1.2 channels are modulated by the cholesterol enrichment has been little clarified.
We showed the effects of PEG-cholesterol on I Ca,L of cultured ASMCs, using A7r5 cell-lines derived from embryonic rat thoracic aorta [20]. Loading of PEG-cholesterol induced a long-lived inhibition of I Ca,L, similar to the enrichment of cholesterol by cholesterol (MbCD). However, although the cholesterol enrichment did not augment VDI, PEG-cholesterol did augment VDI, and induced an extensive inhibition of I WD . PEG-moiety replacing hydroxyl group of cholesterol is responsible for the augmentation of VDI. The characteristic effect of PEG-cholesterol on VDI was explained by hypothesizing its localization in the external leaflet of the bilayer.

Cultured A7r5 cells and pretreatment by the lipids
A7r5 smooth muscle cells were purchased from the American Type Culture Collection (Manassas, VA, USA) and maintained under 5% CO 2 at 37uC in Dulbecco's modified Eagle's medium with L-glutamine and 4,500 mg/L glucose containing 10% fetal bovine serum (Invitrogen, Carlsbad, CA, USA). They were subcultured every 3 or 4 days with trypsin (0.25%) and EDTA (0.02%). A portion of the detached cells were transferred to normal Tyrode's (NT) solution and kept at 4uC in micro-tubes. For patch clamp, the cells in the micro-tube were pretreated at room temperature (21uC) with NT solution as a control, or various concentrations of PEG-cholesterol, cholesterol (MbCD), MbCD or 10 mM PEG dissolved in NT solution for 90-300 min. In the experiment to examine the reversibility of the effects of PEGcholesterol, cells pretreated with 1 mM PEG-cholesterol for 90 min were further treated with 10 or 30 mM MbCD in the presence of 0.9 or 0.7 mM PEG-cholesterol for longer than 90 min.

Patch-clamp
The A7r5 cells in the micro-tube were dispersed on cover-glass in a chamber mounted on an inverted microscope (IX72, Olympus, Tokyo, Japan). After the attachment of cells to the cover-glass, the chamber was super-fused, first with NT, then with 10 mM Ba 2+ solution for total ,40 min (range 12-180 min) before recording the first I Ca,L . Patch pipettes were pulled from hard glass capillary tubing containing a glass filament using a micropipette puller (P-97 Sutter Instrument Co., Novato, CA, USA) coated with silicone elastomer (Sylgard 184, Dow Corning Co., Midland, MI, USA), and fire-polished using a microforge (MF-830, Narishige, Tokyo, Japan). The pipette resistance was ,10 MV when filled with pipette solution. The voltage-clamp amplifier (Axopatch 200B, Molecular Devices, Sunnyvale, CA, USA) was driven by Clampex 10 software via a digital interface (Digidata 1400, Molecular Devices). Currents were filtered at 2 kHz using the amplifier's low-pass 8-pole Bessel filter, and digitized at 10 kHz before being stored on the computer hard drive for later analysis. Membrane capacitance was obtained by applying a negative-going ramp step, and also by using the built-in  Table 1. (G), the r 500 , the ratio of the amplitude of terminal I Ca,L to that of peak I Ca,L . Statistical comparison was performed using 2way ANOVA followed by Dunnett's test; ##, p, 0.01, ###, p,0.001, ####, ****, p,0.0001. Inset, simplified chemical structure of PEG-cholesterol. n, an average number of repeat, was 13.6. doi:10.1371/journal.pone.0107049.g001 PEG-Cholesterol and Inhibition of I Ca,L PLOS ONE | www.plosone.org program in Clampex. The averaged membrane capacitance was 66 pF. Holding potential (HP) was 280 mV. The I Ca,L for the I/ V relationship was obtained by applying 500 ms depolarization steps in 10 mV increments at 0.2 Hz from 240 to 50 mV preceded by a 50 or 70 ms pre-pulse to 240 mV to inactivate Ttype Ca 2+ channel currents that exhibited small and variable amplitude. Quasi steady-state inactivation curves (f ' /V) were obtained using a gapped double-pulse protocol at 0.1 Hz. A 2 s conditioning pulse to potentials between 2100 and 30 mV from a HP of 280 mV was followed by a 50 or 70 ms step to 240 mV, and then a 500 ms test pulse to 0 mV. The I Ca,L was quantified after subtracting the background current.

Data analysis
The I Ca,L was analyzed after low-pass filtering with a cut-off frequency of 1 kHz by Clampfit 10 software (Molecular device). Statistical analysis was conducted by GraphPad Prism (V.6, GraphPad Software, San Diego, CA, USA). Igor Pro (V.6, Wavemetrics, Portland, OR, USA) was used for curve fitting and illustration. The I/V relationship of I Ca,L was fitted with the following equation adapted from the Boltzmann equation: where V is the membrane potential, G max is the maximal conductance, E rev is the reversal potential, V 0.5 is the half activation potential, and k is the slope factor. The f ' (availability)/V relationship was fitted with the following Boltzmann equation: The amplitude of I Ca,L was estimated after subtracting the background current and plotted against voltage (V). Figure 1F shows the I/V relationship of the peak I Ca,L density. The maximum density was obtained at 0 mV. It was 6.360.6 pA/pF (mean 6 S.E.M., n = 37) in the control, decreased with the pretreatment with PEG-cholesterol and cholesterol (MbCD), and increased with 10 mM PEG. The I/V relationship was fitted by the Boltzmann equation. The reversal potential of I Ca,L (E rev ) was 48.462.2 mV in the control, and was little affected by the pretreatments ( Table 1). The amplitude of I Ca,L in the curve fitting was maximal between 0 and 210 mV, except for that from 4 mM cholesterol (MbCD)-pretreated cells ( Figure 1F). G max , the maximal conductance of I Ca,L , was 128.9611.3 pS/pF in the control, decreased by 14% with 0.1 mM PEG-cholesterol, 49% with 10 mM PEG-cholesterol, 43% with 4 mM cholesterol (MbCD), and increased by 15% with 10 mM PEG ( Table 1). The V 0.5 for activation was 215.361.3 mV in the control, and shifted to a depolarizing direction by 2.2 mV with 4 mM cholesterol (MbCD) ( Table 1). The rate of the current decay during the pulse was quantified by r 500 , the ratio of I Ca,L at 500 ms of the pulse to the peak amplitude. The r 500 was little affected with 4 mM cholesterol (MbCD), but decreased with PEG-cholesterol in a concentrationdependent manner to ,50% with 10 mM PEG-cholesterol, and also with 10 mM PEG600 with a statistical significance when compared with the control ( Figure 1G).
Effect of PEG-cholesterol, cholesterol (MbCD) and PEG on steady-state inactivation of I Ca,L Figure 2 illustrates the effects of the pretreatments with PEGcholesterol and others on quasi-steady-state voltage-dependent inactivation. The I Ca,L was gradually inactivated by the increase in Table 1. The effect of PEG-cholesterol, PEG, cholesterol (MbCD) and MbCD on parameters of the I/V relationship. depolarization of the pre-pulse. The pre-pulse-dependent inactivation increased steeply between 240 mV and 220 mV (Figures 2 A-E). The peak current amplitude of the I Ca,L normalized by the maximal amplitude was plotted against the pre-potential to obtain the quasi-steady-state inactivation relationship (f ' /V) (Figure 2F). The relationships were sigmoidal, with a small voltageindependent component that was not inactivated by large depolarization. The f ' /V relationships were well-fitted by the Boltzmann equation with the parameters given in Table 2. In the control, the V 0.5 was 229.160.4 mV and the voltage-independent availability (c 0 ) was 0.1160.01. PEG-cholesterol shifted the curve to the left, making the V 0.5 more negative at 234.2 mV at 0.1 mM, and 239.3 mV at 10 mM, and decreased the c 0 in a concentration-dependent manner ( Table 2). In contrast, 4 mM cholesterol (MbCD) increased the c 0 and shifted the curve in a depolarizing direction. 10 mM PEG slightly shifted the curve in a hyperpolarizing direction (V 0.5 , 230.3 mV), steepened the slope of the curve, and decreased the c 0 . From the I/V (Figure 1) and the f ' /V relationships, the I WD was calculated as the product of the simulated curves (I?f ' /V) at a voltage range between 240 and 20 mV ( Figure 2G). In the control, the I WD density was maximal at a slightly negative voltage of 210 mV, as large as ,80% of the maximal density at 220 mV, and ,30% at 230 mV. PEG-cholesterol strongly inhibited I WD in a concentration-dependent manner. It inhibited the I WD more extensively compared with its inhibition of I Ca,L , i.e. the ratio of maximal I Ca,L to maximal I WD was 0.12 in the control and 0.09 and 0.06 in 0.1 and 10 mM PEG-cholesterol pretreated cells respectively. 4 mM cholesterol (MßCD) induced relatively small inhibition of the I WD with the maximal I WD to maximal I Ca,L ratio of 0.23, as it shifted the f ' /V relationship to the right and increased the c 0 . Figure 3 summarizes the changes in the values of I Ca,L , r 500 , V 0.5 and I WD obtained with several concentrations of PEG-cholesterol, PEG and cholesterol (MbCD). I Ca,L given by the maximal current density was inhibited by PEG-cholesterol in a concentration-dependent manner ( Figure 3A). It was slightly increased by 10 mM PEG, and was inhibited by 1.3 and 4 mM cholesterol (MßCD) almost to the same extent. Correspondingly, the G max that was almost proportional to the maximal current density of I Ca,L was slightly increased by 10 mM PEG, decreased by PEG-cholesterol in a concentration-dependent manner, and was decreased by cholesterol (MbCD (partly shown in Table 1). The r 500 estimated at 0 mV decreased with PEG and 1, 3 and 10 mM PEG-cholesterol, with statistical significance, but was not significantly affected by 1.3 and 4 mM cholesterol (MbCD) ( Figure 3B). The V 0.5 of f ' /V relationship was shifted to more negative potentials in a concentration-dependent manner by PEGcholesterol but was slightly shifted in a depolarizing direction by 1.3 and 4 mM cholesterol (MbCD), and was slightly shifted in a negative direction by 10 mM PEG ( Figure 3C, Table 2). The I WD was concentration-dependently inhibited by PEG-cholesterol more strongly when compared with the inhibition of I Ca,L , while it was not affected with statistical significance by 1.3 and 4 mM cholesterol (MbCD) and 10 mM PEG pretreatment ( Figure 3D).

Effects of MbCD on I Ca,L modulated by PEG-cholesterol
Pretreatment with MßCD, a scavenger of cholesterol, increased I Ca,L (Figures 4B-D, G, Table 1) and slowed the time of inactivation ( Figure 4H) in a concentration-dependent manner. The pretreatment with MbCD also shifted the f ' /V relationship in the depolarizing direction and increased the c 0 in a concentration-dependent manner ( Figure 4I, Table 2), i.e., in the 10 and 30 mM MbCD pretreated cells, I Ca,L was only slightly affected by the conditioning pulse at 240 mV, and large currents appeared even after the conditioning by a 220 mV pre-pulse (Figure 4b). PEG-cholesterol (1 mM) induced ,50% inhibition of I Ca,L amplitude ( Figure 4G), accelerated the current decay ( Figure 4A, H), shifted the f ' /V relationship to the left by 10 mV, and decreased the c 0 ( Figure 4I, Table 2). We examined the effects of 10 and 30 mM MbCD on PEG-cholesterol-induced modulation of I Ca,L . Additional pretreatment with 10 mM MbCD of the 1 mM PEG-cholesterol pretreated cells did not reverse the inhibition of the current density (Figure 4Ea, G), the acceleration of the time course of decay ( Figure 4H), or the negative shift of the V 0.5 ( Figure 4I). However, the pretreatment with 30 mM MbCD  Table 2. (G), the I WD /V relationship. The I WD was obtained by multiplying values of I Ca,L from simulated I/V relationships ( Figure 1F) and f ' from simulated f ' /V relationships ( Figure 2F). Data were obtained from the same cells in Fig. 1 (Table 1). Furthermore, 30 mM MbCD reversed PEG-cholesterol-induced acceleration of the time course of inactivation ( Figure 4F, H) and partially reversed the PEGcholesterol-induced leftward shift of the V 0.5 ( Figure 4I) associated with a recovery of the c 0 to the control value (Table 2). Reflecting the changes of the I/V and f ' /V relationships, the I WD was largely inhibited by 1 mM PEG-cholesterol, markedly increased by MbCD in a concentration-dependent manner, and the PEGcholesterol-induced inhibition was not recovered by the additional pretreatment by 10 mM MbCD, but was increased to more than the control by that of 30 mM MbCD ( Figure 4J).

Discussion
Amphipathic PEG-cholesterol inhibited I Ca,L , augmented VDI and strongly inhibited the I WD . PEG-moiety covalently bound to cholesterol was responsible for the augmentation of VDI, since cholesterol enrichment by cholesterol (MbCD) did not augment VDI. The effects of PEG-cholesterol on I Ca,L were long-lived, and were apparently reversed by MbCD.

Inhibition of I Ca,L by cholesterol and PEG-cholesterol
Enrichment of cholesterol using cholesterol (MbCD) induces strong inhibition of I Ca,L in guinea pig gallbladder smooth muscle cells [21]. I Ca,L of cholesterol-enriched coronary ASMCs from swine with hypercholesterolemia, and those of cholesterol (MbCD)-treated ASMCs from normal swine, are ,50% of the control in the current density without change in the voltagedependence of activation, and the inhibition is reversible by MbCD [18]. In the present study, the enrichment of cholesterol by cholesterol (MbCD) and the loading of PEG-cholesterol decreased the maximal amplitude of I Ca,L and G max by ,50%. Contrastingly, the removal of cholesterol by MbCD from the membrane clearly increased I Ca,L . Voltage-gated Ca 2+ channel curren is augmented by MbCD also in chick cochlear hair cells [22]. The cholesterol (MbCD)-induced inhibition and MbCDinduced augmentation suggest that endogenous cholesterol constitutively inhibits I Ca,L as proposed by the MbCD-induced augmentation of Ca 2+ -permeable transient receptor potential melastatin (TRPM)3 activity [23]. The mechanisms of the cholesterol enrichment-and PEG-cholesterol-induced inhibition of I Ca,L are elusive. The thickness of the lipid bilayer of rabbit ASMC increases with dietary cholesterol enrichment [24]. Single channel conductance of BK Ca (hSlo a-subunit) channels in planar lipid bilayers decreases with the increase of bilayer thickness [25]. PEG-cholesterol loaded on the external leaflet increases the bilayer thickness, as the PEG-tail extends over the surface of the external leaflet and PEG-cholesterol insertion expands the external leaflet to induce a separation of the bilayer [13]. The thickening of the bilayer produced by cholesterol (MbCD) and PEG-cholesterol may decrease the unitary conductance of Ca V 1.2 channels. The conductance and the open probability of the single Ca V 1.2 channel currents in the cholesterol (MbCD) and PEG-cholesterol treated cells should be clarified.
The Ca V 1.2 channel complex that generates I Ca,L is consisted of a 1 , b and a 2 d subunits in ASMCs, cardiac myocytes, neurons and endocrine cells [26]. The a 1 subunit is the primary subunit with a voltage-dependent gate and ion-selective pore, and the b and the a 2 d are accessory subunits to modulate the gating of the a 1 subunit and assist its trafficking to the plasma membrane. The a 2 subunit is exposed over the external surface and the d subunit is bound to the a 2 subunit by a disulfide bond at one end, and is fixed to the outer leaflet at the other end [27,28]. Pregabalin, an a 2 d ligand, induces the inhibition of I Ca,L in rat cerebral ASMCs [29] and the a 2 d -1 subunit is essential for the plasma membrane expression of the a1 subunit [28,29]. The PEG-cholesterol solubilized in the outside of the membrane as well as in the external leaflet may directly interact with the a 2 d subunit to down-regulate its regulatory roles to maintain I Ca,L .
PEG-cholesterol-induced inhibition was apparently reversible by 30 mM (but not by 10 mM) MbCD. The necessity of the high concentration may be explained by the difficulty to remove membrane-embedded PEG-cholesterol by MbCD, since the size of PEG-cholesterol is larger than that of cholesterol and PEG is hydrophilic. However, the MbCD -induced reversal could be produced by the removal of cholesterol and its non-specific effects to remove the membrane lipids other than cholesterol [6]. PEG600 applied at a high concentration of 10 mM as negative control of PEG-cholesterol did not induce inhibition of I Ca,L .

Modulation of VDI by cholesterol (MbCD) and PEGcholesterol
The VDI of N-type Ca 2+ channel currents (I Ca,N ) is not affected by the enrichment of cholesterol by cholesterol (MbCD) in the neuroblastoma-glioma hybrid cells [30]. The time course of inactivation of voltage-gated Ca 2+ channel currents of murine pancreatic b-cells is not affected by the cholesterol enrichment by cholesterol (MbCD) [19]. However, exposure of IMR32 neuroblastoma cells to a cholesterol-enriched medium with tetrahydrofuran for 20-24 hours shifts the V 0.5 of the steady-state inactivation-voltage relationship of I Ca,N ,20 mV in a depolarizing direction [2]. In the present study, both the enrichment and the depletion of cholesterol performed using MbCD shifted the V 0.5 in a depolarizing direction associated with an increase of the c 0 ( Table 2). The increase of the c 0 was reflected in the slow current decay in the MbCD-pretreated cells (Figure 4). MbCD could scavenge membrane lipids other than cholesterol by nonspecific binding to counteract the VDI [6]. Nevertheless, since the enrichment and the depletion of cholesterol using MbCD modulated the density of I Ca,L differently, we consider that the cholesterol enrichment does not augment the VDI. The pretreatment by a high concentration of PEG600 unexpectedly accelerated the time of inactivation (Figure 1 and Figure 3), although it only slightly affected the I/V and f ' /V relationships. The detailed effects of PEG600 and the underlying mechanism should be studied more in the future.
PEG-cholesterol augmented the VDI of I Ca,L, manifested with the decrease of the r 500 , the decrease of the c 0 and the negative shift of the f ' /V relationship ( Figure 2, Table 2). The augmentation of the VDI could arise both from conformational changes of the Ca V 1.2 channel complex and the changes of the expression of subunits of the complex. The hypothetical PEG-cholesterolinduced thickening of the hydrophobic core of the bilayer in the presence of constant hydrophobic length of the a 1 subunit induces a lipid-protein hydrophobic mismatch [31]. The mismatch forces the bending of the bilayer adjacent to the a 1 subunit to re-align the lipid bilayer hydrophobic core to the subunit's hydrophobic exterior. It induces a change of configuration of the a 1 subunit, which could result in the augmentation of the VDI. Co-expression of the a 2 d subunit with the a 1 and b subunits augments the VDI in Ca V 1.2 and Ca V 2.1 channels extrinsically expressed in HEK293 cells [32]. PEG-cholesterol pretreatment could compromise the facilitatory action of the a 2 d subunit in the a 1 subunit trafficking to the membrane [28,29] to result in a relative abundance of the a 2 d subunit, and thus the augmentation of the VDI. As another mechanism, PEG-cholesterol or PEG-cholesterol-induced expansion of the external leaflet [12] could physically stimulate the a 2 d subunit to augment the VDI.
Triton X-100 (TX-100) is an amphiphile possessing poly [oxyethylene glycol] chain, analogous to PEG. TX-100 reversibly inhibits I Ca,L in non-neuronal cells, including rat mesenteric artery ASMCs, with IC 50 of low mmoles [33]. Moreover, it rapidly and reversibly augments the VDI of I Ca,N with ,20 mV negative shift of steady-state inactivation curve in IMR32 neuroblastoma cells [2]. TX-100 and other amphiphiles, such as capsaicin, shift the V 0.5 of the Na channel current in a hyperpolarizing direction in  (H), the r 500 /V relationship. Statistical comparison was performed using 2way ANOVA followed by Dunnett's test; **, **, p,0.01, ***, p,0.001, ****, ****, ####, p,0.0001. (I), the f ' /V relationship. The I/V and f ' /V relationships were fitted to the Boltzmann equations and the constants obtained from the fitting are given in Table 1 and Table 2. (J), the I WD /V relationship. The I WD was obtained by multiplying simulated I Ca,L /V and f ' /V relationships. doi:10.1371/journal.pone.0107049.g004 HEK293 cells with an extrinsic expression of Na + channels [34]. However, the observation that water-soluble PEG-cholesterol can be encapsulated in liposomes constructed by phosphatidylcholine indicates that PEG-cholesterol does not have detergent-like activity [13]. The pretreatment with 10 mM PEG shifted the f ' /V relationship in a hyperpolarizing direction only slightly ( Figure 2, Table 2), and its acute application did not affect the f ' / V relationship in a preliminary experiment. Therefore, the covalent coupling with cholesterol was necessary for PEG to induce the inhibition of I Ca,L and the augmentation of the VDI.
To summarize the mechanism, solvation of PEG-cholesterol into the outer leaflet of the lipid bilayer expands the leaflet and increases thickness of the bilayer. It can induce a hydrophobic mismatch between the Ca V 1.2a 1 subunit and the bilayer in A7r5 cells. The mismatch could induce configurational change of the a 1 subunit. Also, the configuration of the a 2 d subunits may be changed by direct interaction with PEG-cholesterol or the expansion of the external leaflet. We hypothesize that these changes of Ca V 1.2 configurations and those of the subunit expression induce the PEG-cholesterol-induced augmentation of the VDI.
In conclusion, PEG-cholesterol inhibited I Ca,L and augmented its voltage-dependent inactivation (VDI). The augmentation of the VDI contributed to the PEG-cholesterol-induced strong inhibition of I WD . Poly(ethylene glycol) conferred to cholesterol the efficacy to induce sustained augmentation of VDI of I Ca,L .