Volatile anaesthetic effects on calcium conductance of planar lipid bilayers formed with synthetic lipids or extracted lipids from sarcoplasmic reticulum

British Journal of Anaesthesia 1997; 78: 66–74

Volatile anaesthetic effects on calcium conductance of planar lipid bilayers formed with synthetic lipids or extracted lipids from sarcoplasmic reticulum T. ANDOH, T. J. J. BLANCK, I. NIKONOROV AND E. RECIO-PINTO

Summary Volatile anaesthetics are known to increase leakage of calcium from the light fraction of skeletal sarcoplasmic reticulum (L-SR) which has no calcium release channels. To explore the role of the lipid environment, we have examined the effect of volatile anaesthetics on calcium conductance (gCa) of lipid membranes. Planar lipid bilayers were formed with a mixture of synthetic phospholipids and cholesterol, resembling the composition of SR membranes, or with lipids extracted from skeletal L-SR. gCa was estimated by calculating the calcium transference number (tCa) using diffusion potential measurements. Membranes formed with L-SRextracted lipids had a higher gCa than membranes formed with synthetic lipids. Volatile anaesthetics increased total conductance and gCa in a dosedependent manner, but did not affect tCa or membrane specific capacitance. In membranes formed with L-SR-extracted lipids, isoflurane induced the largest increase in gCa (1260 (SEM 304) % increase, n
4, 0.94 mmol litre1), followed by enflurane (264 (75) %, n
5, 1.88 mmol litre1) and halothane (53 (33) %, n
5; 1.54 mmol litre1). In membranes formed with synthetic lipids, volatile anaesthetic-induced increases in gCa followed the same trend but were larger. Volatile anaesthetics increased gCa without changing the ionic selectivity of membranes. However, the magnitude of the increase in gCa in the presence of volatile anaesthetics cannot account for the previously observed calcium leakage from L-SR vesicles. Therefore, the volatile anaesthetic-induced increase in calcium leakage in L-SR vesicles must be mediated via other pathways involving membrane proteins. (Br. J. Anaesth. 1997; 78: 66–74) Key words Ions, calcium. Ions, ion channels. Anaesthetics volatile. Membrane, lipid bilayer. Theories of anaesthetic action, lipid bilayer.

The volatile anaesthetics halothane, enflurane and isoflurane alter calcium homeostasis in many cells. Volatile anaesthetics alter cellular Ca2 in part by acting on the sarcoplasmic reticulum (SR). They have been shown to increase the rates of Ca2

release1–5 and of Ca2 uptake5 6 from the SR. The former effect results from a volatile anaestheticinduced increase in the open probability of calcium release channels,7 increase in the rate of ATP breakdown1 5 and decrease in SR calcium pump activity.8–10 The increase in the rate of Ca2 uptake results from an increase in SR calcium pump activity when volatile anaesthetics are present at low concentrations.1 5 9 Volatile anaesthetics also increase Ca2 leakage in vesicles isolated from the light SR (L-SR), which lacks calcium release channels.1 The latter effect might be caused by a volatile anaesthetic action on the SR lipid membrane, enhancement of leakage pathways through membrane proteins or action on the interface between membrane lipids and proteins. We have investigated one of these possibilities, the effect of volatile anaesthetics on Ca2 permeability of SR lipid membranes. It has been shown previously in liposomes that halothane increases membrane permeability to monovalent cations (K, Rb) and protons.11 12 Therefore, it is expected that volatile anaesthetics should also increase Ca2 permeability of lipid membranes. Using the planar lipid bilayer system, we have investigated to what extent volatile anaesthetics affected Ca2 permeability of lipid membranes formed with either synthetic or L-SR-extracted lipids. We found that at the concentrations studied, volatile anaesthetics did not affect membrane-specific capacitance or ionic selectivity but significantly increased overall membrane ionic permeability (up to 10-fold). However, the magnitude of this increase was too small to account for the previously observed volatile anaesthetic-induced increase in Ca2 leakage in isolated L-SR-vesicles.

TOMIO ANDOH, MD*, THOMAS J. J. BLANCK, MD, PHD†, IGOR NIKONOROV, MS, ESPERANZA RECIO-PINTO, PHD, Cornell University Medical College, Anaesthesiology Department, LC2, BOX50, 1300 York Avenue, New York, NY 10021, USA. Accepted for publication: August 29, 1996. *Present address: Yokohama City University, School of Medicine, Department of Anaesthesiology, 3–9 Fukuura, Kanazawa-Ku, Yokohama 236, Japan. †Address for correspondence: Department of Anaesthesiology, The Hospital for Special Surgery, 535E 70th St, New York, NY 10021, USA.

Ca2+ conductance of planar lipid bilayers

Materials and methods LIPID COMPOSITION OF PLANAR LIPID BILAYERS

Two mixtures of synthetic lipids were used: one consisted of a mixture of the neutral lipids palmitoyllinoleoyl phosphatidylcholine, palmitoyl-arachidonyl phosphatidylethanolamine and cholesterol at a molar fraction of 7.5:2.5:1. This ratio was chosen to mimic the reported lipid composition of muscle SR.13 14 The second lipid mixture contained the same components as the first mixture and 10% phosphatidylinositol. The latter was selected because SR lipids contain negatively charged phospholipids (approximately 10%), the main one being phosphatidylinositol.13 15 Lipids were purchased from Avanti Polar Lipids, Inc. The lipid mixture was dissolved in decane (5% w/v) to form planar–lipid bilayers, as described previously.16 L-SR vesicles were isolated from rabbit skeletal muscle,1 resuspended with distilled water (approximately 8.3 mg protein/ml) and lyophilized. Lipids were extracted17 18 and stored in chloroform/ methanol (2/1) at 20⬚C. The phospholipid content was measured by a phosphorus assay.19 The phosphorus content of lipid extract was 0.72 mm/mg SR protein. To form planar–lipid bilayers, the chloroform–methanol was evaporated and the lipids dissolved in decane (5% w/v). Decane was used as it has been shown that it has little effect on blocking nerve impulses and on membrane thickness.20 The data shown were obtained using a single preparation of L-SR-extracted lipids. Other L-SR-extracted lipid preparations were used in pilot experiments. The bilayer chamber contained two compartments (1.24 ml capacity) separated by a partition with a hole (0.2–1.0 mm diameter) where lipid bilayers were formed. The front and back compartments are referred to as cis and trans, respectively. Silver–silver chloride electrodes were placed in each compartment; the reference electrode was placed in the trans compartment. The compartments contained 1 ml of a given solution. Salt bridges were not used during the experiments with CaCl2 solutions. When asymmetrical CaCl2 solutions were used, the applied potential was corrected for the estimated electrode–liquid potential differences.21 For the experiments using sucrose solution, KCl 3 mol litre1 salt bridges were used to stabilize the electrode offset. MEMBRANE CONDUCTANCE AND CAPACITANCE MEASUREMENTS

Total membrane conductance (gm), that is conductance caused by all ion species present in the solution, was measured using symmetrical solutions of CaCl2 200 mmol litre1. Slope conductances were obtained from linear fits of the current–voltage relationships constructed by applying 25-mV voltage steps between 50 and 50 mV. Holding potentials were maintained for 1–2 min to eliminate the contribution of slow capacitive currents. Membrane capacitances were calculated by measuring capacitive currents after applying 1-mV triangular voltage steps at a

67 frequency of 100 Hz. Membranes were seen with a 120-times magnification and their diameter, limited by a visible torus, was measured with a graduated grid (2.5 ␮m/division). Membrane surface area was calculated assuming that the shape of the membrane was an ellipse. All solutions contained HEPES 2 mmol litre1 at pH 6.0. Free acid HEPES of the best grade was used (Sigma Chemical Co., St Louis, MO, USA). We used a pH value of 6.0 to reduce proton-hydroxide conductance (gHOH).22 Total membrane conductance (gm) was measured using symmetrical solutions of CaCl2 0.2 mol litre1. In all cases it was assumed that the conductance caused by the anion form of HEPES was negligible. DIFFUSION POTENTIAL MEASUREMENTS AND ESTIMATIONS OF TRANSFERENCE NUMBER

To obtain the relative membrane ionic selectivity for calcium over chloride, we estimated the transference number of Ca2 (tCa) by measuring diffusion potentials as described previously.22–24 The zero current level was first measured in symmetrical CaCl2 20 mmol litre1 by applying 0 mV. The concentration of CaCl2 was then increased in the cis compartment in a stepwise manner to 60, 120 and 200 mmol litre1 by addition of CaCl2 1 mol litre1, and the potential required to re-establish the zero current level (diffusion potential) was determined by interpolation of the current measurements obtained at 2–3 different membrane potentials. Changes in diffusion potential were obtained at every dose of volatile anaesthetic for membranes formed with synthetic lipids, and at the median and high doses of volatile anaesthetics for membranes formed with L-SR-extracted lipids. The tCa was calculated by equation (1), assuming that only Ca2 and Cl ions carried the ionic current (see Appendix for derivation). Vd = (1 − 1.5t Ca )× ( RT /F )× l n (a cis / a trans ) (1) where Vd
measured diffusion potential, R
gas constant, T
temperature in Kelvin, F
Faraday constant, acis
mean salt activity of cis chamber and atrans
mean salt activity in trans chamber. Values for mean salt activity were interpolated from previously reported values.25 In most cases, tCa values were estimated using a 10-fold concentration gradient. In some cases, tCa values were obtained using a six-fold concentration gradient. ADDITION OF VOLATILE ANAESTHETICS AND MEASUREMENT OF VOLATILE ANAESTHETIC CONCENTRATIONS

Stock solutions of saturated volatile anaesthetics were prepared1 in either sucrose 0.6 mol litre1, CaCl2 20 mmol litre1 or CaCl2 200 mmol litre1. A small volume (20–100 ␮l) of the bath solution was first removed and then replaced by an equal volume of the bath solution saturated with a volatile anaesthetic. After each addition of volatile anaesthetic (to both compartments), the chamber compartments were covered with glass coverslips and solutions stirred continuously for 1 min. The volatile anaesthetic concentration in the stock and chamber solutions was determined by gas chromatography.26

68

British Journal of Anaesthesia

Table 1 Estimated values for the hypothetical MAC at 25⬚C (H-MAC) and the corresponding hypothetical aqueous concentration at 25 ⬚C (H-EC50). H-MAC and H-EC50 were obtained using equations (2) and (3) as described under materials and methods

Isoflurane Halothane Enflurane

MAC(30) (37 ⬚C)

H-MAC (25 ⬚C)

H-EC50 (25 ⬚C)

2.1% 1.4% 2.9%

0.93% 0.62% 1.29%

0.47 0.32 0.66

As experiments were performed at room temperature (approximately 25⬚C), the magnitude of volatile anaesthetic effects at 37⬚C are uncertain. To attempt a comparison of the relative potency of the volatile anaesthetics, we also present the volatile Table 2 Membrane areas and specific capacitance values for membranes formed with synthetic neutral lipids (Synthetic) and with L-SR-extracted lipids (L-SR) in the absence and presence of volatile anaesthetic (mean (SEM)). (In all cases the partition hole was 0.2 mm in diameter)

Anaesthetic (mmol litre1) Synthetic Control Isoflurane 0.40 Isoflurane 0.72 Isoflurane 0.94 Control Halothane 0.32 Halothane 0.78 Halothane 1.54 Control Enflurane 0.80 Enflurane 1.32 Enflurane 1.88 L-SR Control Isoflurane 0.40 Isoflurane 0.72 Isoflurane 0.94 Control Halothane 0.32 Halothane 0.78 Halothane 1.54 Control Enflurane 0.80 Enflurane 1.32 Enflurane 1.88

anaesthetic concentration as the ratio of the hypothetical aqueous concentration (H-EC50) corresponding to the hypothetical MAC at 25⬚C. This was done in the following way. First, we estimated the hypothetical MAC value at 25⬚C (H-MAC (25⬚C)) using equation (2) derived from Franks and Lieb.27 H-MAC (Tc)/MAC (37⬚C)
(2) exp (− 20.2 (37 − Tc) / 273.15 + Tc)) Then, using H-MAC (25⬚C) we estimated the corresponding hypothetical aqueous concentrations (mmol litre1) (H-EC50) using equation (3).27 (3) H-EC50 (mmol litre1 ) = 0.44614␣MAC (%) where ␣
Bunsen water/gas partition coefficient. At 25⬚C, ␣ values are similar for halothane (1.20) and isoflurane (1.08),28 but there is no reported value for enflurane. As the three volatile anaesthetics have similar ␣ values at 37⬚C,28 29 we assumed that this also held true at 25⬚C, and used a mean ␣ value of 1.14 for all three volatile anaesthetics. We used MAC values from rabbits30 as the extracted SR lipids were from rabbit skeletal muscle (table 1). STATISTICS

Membrane area ((cm2)104)

Membrane capacitance (␮F cm2)

n

1.76 (0.12) 1.91 (0.18) 1.96 (0.22) 2.02 (0.20) 1.79 (0.17) 1.74 (0.17) 1.81 (0.18) 1.76 (0.16) 1.64 (0.20) 1.52 (0.23) 1.56 (0.24) 1.52 (0.22)

0.61 (0.11) 0.66 (0.09) 0.69 (0.51) 0.70 (0.39) 0.61 (0.15) 0.62 (0.16) 0.67 (0.15) 0.72 (0.14) 0.73 (0.07) 0.72 (0.78) 0.69 (0.79) 0.68 (0.80)

6 6 6 6 4 4 4 4 5 5 5 5

0.93 (0.35) 0.88 (0.32) 0.84 (0.34) 0.90 (0.36) 1.06 (0.23) 0.87 (0.18) 0.99 (0.26) 0.97 (0.20) 0.86 (0.17) 0.94 (0.18) 0.66 (0.22) 0.90 (0.23)

1.11 (0.18) 1.19 (0.27) 1.23 (0.20) 1.30 (0.20) 0.75 (0.08) 0.77 (0.07) 0.79 (0.07) 0.77 (0.13) 0.99 (0.24) 0.70 (0.06) 0.62 (0.07) 0.89 (0.25)

6 6 6 6 6 6 6 6 4 4 4 4

Results are presented as mean (SEM). Differences between values obtained from synthetic and L-SRextracted lipids in the absence of anaesthetic were analysed using the Mann–Whitney rank sum test. Differences between control and anaesthetics values were analysed by Friedman’s test followed by the non-parametric Dunnett’s test based on rank sum as post hoc analysis. In all cases, P
0.05 was considered significant.

Results SPECIFIC CAPACITANCE

Under control conditions, membranes formed with L-SR-extracted lipids were smaller and had higher specific capacitance values than membranes formed with synthetic lipids (tables 2, 3). Such differences could reflect the presence of negatively charged phospholipids, as membranes formed with synthetic lipids containing negatively charged phospholipids had membrane areas and specific capacitance values resembling those of membranes formed with L-SR-extracted lipids (table 3).

Table 3 Electrical properties and calculated calcium permeability in membranes formed with synthetic and L-SRextracted lipids (mean (SEM)). n
number of membranes. For synthetic neutral lipids n
15 except for tCa and gmsucr, n
4. For L-SR-extracted lipids, n
16 except for tCa, n
3. For synthetic neutralcharged lipids, n
6. Am
membrane area, Cm
specific membrane capacitance; gm
total specific membrane conductance in CaCl2; tCa
transference number for calcium; gCa
specific calcium conductance; PCa
calcium permeability; gmsucr
total specific membrane conductance in sucrose. *Significantly different from synthetic neutral lipid, P
0.05. **Significantly different from synthetic neutral lipid, P
0.01 Synthetic neutral lipids Am (cm2) Cm (␮F cm2) gm (S cm2) tca gCa (S cm2) PCa (cm s1) gmuscr (S cm2)

1.73 (0.01)104 0.65 (0.06) 1.31 (0.17)108 0.40 (0.02) 0.52 (0.07)108 1.61 (0.22)1012 1.34 (0.01)108

L-SR-extracted lipids 0.96 (0.16)104** 0.88 (0.07)** 3.96 (0.58)108** 0.56 (0.01)* 2.22 (0.32)108** 6.87 (0.99)1012**

Synthetic neutralcharged lipids 0.71 (0.13)104* 1.05 (0.04)* 2.43 (0.32)108*

Ca2+ conductance of planar lipid bilayers

69

Figure 1 Current–voltage relationships of membranes formed with synthetic neutral lipids measured in the presence of CaCl2 200 mmol litre1 before and after addition of isoflurane. Each point represents the mean value from 3–5 membranes, in the absence (■) and presence of isoflurane 0.72 mmol litre1 ( ) and 0.92 mmol litre1 (●). Inset shows the current–voltage relationship for control at an expanded scale. The lines represent linear regression fits for each condition and have the following values: y
0.005x0.001, r2
0.998 for control; y
0. 187x0.113, r2
1.000 for isoflurane 0.72 mmol litre1 and y
0.901x0.333, r2
1.000 for isoflurane 0.92 mmol litre1.

The area and specific capacitance of membranes formed with synthetic or L-SR-extracted lipids were not changed significantly by isoflurane, halothane or enflurane (table 2). Therefore, within the concentration range studied, these volatile anaesthetics did not affect membrane dimensions (area and specific capacitance). TOTAL MEMBRANE CONDUCTANCE IN CACL2

In the absence and presence of anaesthetic, current– voltage relationships were linear from 50 to 50 mV (fig. 1). gm was significantly higher in membranes formed with L-SR-extracted lipids than in membranes formed with synthetic neutral lipids (table 3). When phosphatidylinositol was added to the mixture of lipids, gm values were between those for membranes formed with synthetic neutral lipids and L-SR-extracted lipids (table 3). All three volatile anaesthetics increased gm in a dose-dependent manner; in all cases, the dose–response curves were linear with the log of gm (fig. 2). The strongest effect was shown by isoflurane and the weakest effect by halothane; the effect of enflurane was intermediate. The magnitude of the volatile anaesthetic-induced increase in gm was lower in membranes formed with L-SR-extracted lipids than in membranes formed with synthetic lipids. TRANSFERENCE NUMBER FOR CALCIUM

In order to determine the ionic selectivity of membrane

Figure 2 Effects of volatile anesthetics on total specific membrane conductance. Total specific membrane conductance (gm) measurements were performed in symmetrical solutions of CaCl2 200 mmol litre1 in membranes formed with synthetic neutral lipids (A) and with L-SR-extracted lipids (B). The anaesthetic concentration in the aqueous solution is given in mmol litre1 and as the ratio of the hypothetical-aqueous concentration (H-EC50) corresponding to the hypothetical MAC at 25⬚C (insets). The data were fitted to a single exponential function. The fits of the data collected in membranes formed with synthetic lipids gave the following values: y
1.31108102.53X (r2
0.99), y
1.60108100.31X (r2
0.97), and y
0.98108100.50X (r2
0.99) for isoflurane (●) (six membranes), halothane (■) (four membranes) and enflurane (▲) (five membranes), respectively for log gm vs mmol litre1; and y
1.31108101.19X (r2
0.99), y
1.60108100.10X (r2
0.97) and y
0.98108100.33X (r2
0.99) for isoflurane, halothane and enflurane, respectively for log gm vs xH-EC50. The fits of the data collected in membranes formed with L-SR-extracted lipids gave the following values: y
4.27108101.20X (r2
0.99), y
3.56108100.17X (r2
0.92) and y
3.17108100.28X (r2
0.96) for isoflurane (!) (six membranes), halothane (
) (six membranes) and enflurane (䉭) (four membranes), respectively, for log gm vs mmol litre1; and y
4.27108100.56X (r2
0.98), y
3.56108100.06X (r2
0.92) and y
3.17108100.19X (r2
0.96) for isoflurane, halothane and enflurane, respectively for log gm vs xH-EC50. The data points represent mean (SEM). The observed increases in gm were significant at all doses of volatile anaesthetic except that observed at the lowest dose of halothane in membranes formed with L-SR-extracted lipids.

70

British Journal of Anaesthesia

Figure 3 Diffusion potentials in various CaCl2 concentration gradients of membranes formed with synthetic neutral and L-SR-extracted lipids. Diffusion potentials were measured as described under materials and methods and plotted against the logarithmic ratio of salt activity of cis to trans solution. ●
Mean of four membranes formed with synthetic lipids and !
mean of three membranes formed with L-SR-extracted lipids. The regression line gave the following values: y
23.28x0.008 (r2
0.998) and y
9.72x0.04 (r2
0.999) for synthetic and L-SR- extracted lipids, respectively. The Nerst potential values for Cl (dotted line) and Ca2 (broken line) are indicated. Data points are mean (SEM). Lipid membranes were initially formed in symmetrical CaCl2 20 mmol litre1, then the concentration of CaCl2 in the cis compartment was increased. Membrane potential (Vm) was obtained by adding to the command potential (Vcmd) estimated electrode–liquid potential differences (ER) (20.1, 32.2 and 40.1 mV for 60/20, 120/20 and 200/20 mmol litre1 of CaCl2, cis/trans) (e.g. in CaCl2 60/20, gradient for Vcmd
50 mV, Vm
50(20)
30 mV). ER results from the electrodes being exposed to different concentrations of Cl and was calculated using the following expression: ER
2.303(RT/F ) log (aClcis/aCltrans) where aCl
activity of Cl in either the cis or trans compartments.21 25

Table 4 Calcium transference number (tCa) in the absence and presence of volatile anaesthetic (mean (SEM)). n
3 membranes in all cases except for control with synthetic lipids where n
4 membranes Anaesthetic (mmol litre1)

Synthetic lipids tCa

L-SR-extracted lipids tCa

Control Isoflurane 0.40 Isoflurane 0.72 Isoflurane 0.94 Halothane 0.32 Halothane 0.78 Halothane 1.54 Enflurane 0.80 Enflurane 1.32 Enflurane 1.88

0.40 (0.02) 0.45 (0.03) 0.37 (0.02) 0.40 (0.03) 0.47 (0.08) 0.45 (0.02) 0.52 (0.05) 0.51 (0.05) 0.49 (0.04) 0.40 (0.01)

0.56 (0.01) 0.56 (0.01) 0.48 (0.08) 0.60 (0.01)

conductance, we estimated the transference number for calcium (tCa
gCa/gm) by assuming that under the conditions measured, the only two permeant ions were Cl and Ca2. This was done by measuring the diffusion potential at different CaCl2 concentrations (see materials and methods). The diffusion potential was a linear function of the log of the activity ratio (log acis /atrans) across the membrane in synthetic and

Figure 4 Effect of volatile anaesthetics on calcium permeability in membranes formed with synthetic (A) and L-SR-extracted lipids (B). The log of membrane calcium permeability (PCa) was plotted against the anaesthetic concentration in the aqueous solution in mmol litre1 and against the ratio of the hypotheticalaqueous concentration (xH-EC50) corresponding to the hypothetical MAC at 25⬚C (insets). The number of membranes used are the same as in figure 2. Data were fitted to a single exponential function. The fits of the data collected in membranes formed with synthetic lipids gave the following values: y
1.661012102.49X (r2
0.99), y
1.651012100.37X (r2
0.97), and y
1.341012100.50X (r2
0.99) for isoflurane (●), halothane (■) and enflurane (▲), respectively, for log PCa vs mmol litre1; and y
1.661012101.17X (r2
0.99), y
1.651012100.12X (r2
0.97) and y
1.3410120.33X (r2
0.99) for isoflurane, halothane and enflurane, respectively, for log PCa, vs xH-EC50. The fits of the data collected in membranes formed with L-SR-extracted lipids gave the following values: y
7.391012101.20X (r2
0.98), (y
5.801012100.14X (r2
0.87) and y
5.601012100.30X (r2
0.97) for isoflurane (!), halothane (") and enflurane (䉭), respectively, for log PCa vs mmol litre1; and y
7.391012100.56X (r2
0.98), y
5.801012100.04X (r2
0.87) and y
5.601012100.20X (r2
0.97) for isoflurane, halothane and enflurane, respectively, for log PCa vs xH-EC50. Data points represent mean (SEM). The observed increases in PCa were significant at all doses of volatile anaesthetic except that observed at the lowest dose of halothane in membranes formed with L-SR-extracted lipids.

Ca2+ conductance of planar lipid bilayers L-SR-extracted lipids and intercepted 0 at both axes, as is expected for diffusion potentials (fig. 3). Figure 3 also indicates the diffusion potential for a purely chloride-selective membrane and a purely calciumselective membrane. The diffusion potentials of membranes formed with synthetic lipids were closer to the values expected for a purely chloride-selective membrane than the diffusion potentials of membranes formed with L-SR-extracted lipids. In fact, the calculated tCa value was larger in membranes formed with L-SR-extracted lipids than in membranes formed with synthetic lipids (table 3). The volatile anaesthetics did not produce significant changes in tCa (table 4). Therefore, volatile anaesthetics induced an increase in gm without changing the ionic selectivity of the membrane pathway. Figure 3 shows that the diffusion potential changed linearly with log (acis /atrans), indicating that the tCa values are similar at various concentration gradients. CALCIUM CONDUCTANCE AND PERMEABILITY OF LIPID MEMBRANES

Specific calcium conductance (gCa) and permeability (PCa) were estimated in CaCl2 200 mmol litre1 by assuming that the proton and hydroxide conductances were negligible, and by using equations (4) and (5): (4) g Ca = g m t Ca PCa =RTgCa /(z2F 2[Ca2+ ]) (5) where z
valence of calcium, [Ca2]
concentration of calcium (mol cm3). gCa and PCa values were approximately 4.5 times larger in membranes formed with L-SR-extracted lipids than in membranes formed with synthetic lipids (table 4). Volatile anaesthetics did not affect tCa (table 4) but increased PCa (fig. 4). The observed increase in PCa was therefore caused by an increase in gm (fig. 2) rather than a change in ionic selectivity (table 4). Isoflurane produced the largest increase in PCa, the effects of enflurane and halothane were smaller.

71

Figure 5 Effect of isoflurane on total specific membrane conductance in sucrose. The total specific membrane conductance in sucrose (gmsucr) was measured in symmetrical solutions of sucrose 0.6 mol litre1 in membranes formed with synthetic lipids (electrodes with salt bridges). gmsucr was plotted against the anaesthetic concentration in the bath solution (mmol litre1). Data were fitted to a single exponential function. The fits of the data collected in sucrose 0.6 mol litre1 (!) gave the following values: y
1.43108101.28X (r2
0.99) (four membranes). ●
Data collected in CaCl2 solutions shown in figure 2A: y
1.31108102.53X (r2
0.99).

absence of a gradient for H and OH. There are two possible explanations why in the absence of volatile anaesthetic, gm was equal to gmsuc. First, sucrose could increase PH+OH − and second, CaCl2 could decrease. PH+OH − .

Discussion At the concentrations studied, we found that volatile anaesthetics increased gCa, and gCl without affecting ionic selectivity or specific capacitance of planar lipid bilayers. We found also that volatile anaesthetics induced a significant increase in PCa of lipid membranes, such an increase depended on the lipid composition, but its magnitude was too small to account for the previously reported volatile anaestheticinduced increase in Ca2 leakage in isolated L-SR vesicles.

MEMBRANE CONDUCTANCE IN THE PRESENCE OF SUCROSE 0.6 MOL LITRE1

MEMBRANE SPECIFIC CAPACITANCE

Total membrane conductance measured in CaCl2 (gm) is the sum of the conductance for each ion present in the bath solution. We attempted to estimate the contribution of proton-hydroxide conductance to total membrane conductance by measuring membrane conductance in the presence of symmetrical solutions of sucrose 0.6 mol litre1. In the absence of volatile anaesthetic, gmsucr and gm were the same (fig. 5). Isoflurane increased gmsucr of membranes in a dose-dependent manner (fig. 5). However, the isoflurane-induced increase in gmsucr was approximately 25% of the isoflurane-induced increase in gm. This indicates that volatile anaesthetics induce an increase not only in PH + OH − but also in PCa or Pcl, or both. If one assumes that gmsucr is only a result of PH + OH − then gm of these membranes in the absence of volatile anaesthetic should be entirely a result of PH + OH − . This is unlikely as we were able to measure diffusion potentials in gradients of CaCl2 and in the

Membranes formed with L-SR-extracted lipids and synthetic lipids containing negatively charged phospholipids were smaller and had higher specific capacitance values than membranes formed with synthetic neutral lipids. Therefore, at least in the presence of high Ca2 concentrations, negatively charged phospholipids appear to be responsible, in part, for the formation of a larger membrane torus (smaller membrane area) and providing the membrane with a higher specific capacitance value. A potential artefact in our capacitance measurements might result from residual decane in the lipid phase. This was unlikely because the specific capacitance values obtained in this study were similar to those reported previously for biological cell membranes.31 Moreover, it has been reported that when actual membrane areas are measured, decane does not significantly affect membrane thickness and it appears to accumulate mainly at the membrane

72

British Journal of Anaesthesia

Table 5 Calcium currents (ICa) in the absence and presence of volatile anaesthetic. *The concentrations of volatile anaesthetic were: halothane 1.54, enflurane 1.88 and isoflurane 0.94 mmol litre1. **This value was calculated using Ca2+ leakage measurements from SR of skinned cardiac myocytes (1.24 mmol litre1)33 in the absence of ATP and anaesthetic. It was assumed that cell volume was 24 pl, SR volume per myocyte was 8.4 pl and SR surface area per myocyte was 29.28×103 ␮m2 .34 ***This value was calculated from Ca2+ leakage measured in isolated L-SR-vesicles from skeletal muscle after ATP depletion.1 It was assumed that 1 mg protein from L-SR membranes corresponded to 2.8103 cm2 SR membrane area35 ICa (mol cm2 s1) Synthetic lipid L-SR-extracted lipid L-SR-extracted lipidhalothane* L-SR-extracted lipidenflurane L-SR-extracted lipidisoflurane SR in skinned muscle cells** Isolated L-SR-vesicles***

2.01017 8.31017 11.21017 26.51017 137.01017 5.91013 3–111014

torus.20 The lack of a volatile anaesthetic effect on either membrane area or specific capacitance indicates that, at the concentrations studied, volatile anaesthetics did not affect membrane geometry. MEMBRANE CONDUCTANCE AND TCa VALUES

Total membrane conductance (gm) and Ca2 selectivity (tCa) was higher in membranes formed with L-SR-extracted lipids than in membranes formed with synthetic neutral lipids. This difference reflects, in part, the presence of negatively charged lipids in the L-SR-extracted lipid preparation, as gm and tCa values were higher in membranes formed with synthetic lipids containing negatively charged lipids than in those containing only neutral lipids (table 3). The higher tCa value is consistent with previous observations that lipid membranes become more cationic-selective when they have negatively charged lipids.24 At the concentrations studied, volatile anaesthetics increased gm without changing the tCa value of the membrane, that is volatile anaesthetics increased the rate of ion transport across the lipid membrane pathway without affecting the ionic selectivity of such a pathway. As there were no membrane proteins present, ion transport across these membranes must be determined by lipid interactions. Our study, however, did not allow the molecular mechanism of these interactions in the presence of volatile anaesthetics to be defined. ESTIMATION OF Ca LEAKAGE CURRENTS

The magnitude of the Ca2 leakage currents were estimated using the Goldman flux equation,32 and by assuming that the transmembrane potential across the SR membrane was 5 mV, intraluminal and cytoplasmic Ca2 concentrations were 10 mmol litre1 and 0.1 ␮mol litre1, respectively, and that PCa was not affected by changes in Ca2 ion concentration. ICa =(−PCa 2FEm / RT )×([Ca2+ ]in −[Ca2+ ]out )× exp (−2FEm/RT )/(1− exp(−2FEm/RT ) where ICa
Ca2 flux (mol cm2 s1), Em


transmembrane potential (V) and [Ca2]in and [Ca2]out
intraluminal and myoplasmic Ca2 concentrations (mol cm3), respectively. The estimated ICa for lipid membranes are three or four orders of magnitude smaller than those estimated for SR in skinned cardiac myocytes33 (table 5). The high Ca2 leakage value of SR in skinned cardiac myocytes could reflect the presence of non-selective ion channel proteins. Non-selective chloride channels that allow passage of Ca2 ions and contribute significantly to Ca2 leakage have been demonstrated in isolated SR vesicles.36 If Ca2 leakage through the SR in intact cells is as high as that reported for SR in skinned cells, then the observed volatile anaestheticinduced increase in ionic permeability of lipid membranes will not contribute significantly to the overall SR leakage currents. Estimated Ca2 leakage from isolated L-SRvesicles1 was two or three orders of magnitude higher than that estimated for lipid membranes in the absence and presence of volatile anaesthetics (table 5). Measurements were performed at room temperature in lipid membranes (this study) and at 37 ⬚C in isolated L-SR vesicles.1 However, such differences in temperature are unlikely to account for the different Ca2 leakage values. In fact, it has been reported that in liposomes formed with SR-extracted lipids, increasing the temperature from 25 to 40 ⬚C increased the rate constant of Ca2 efflux by less than one order of magnitude.37 As discussed in the next section, our estimated PCa values for lipid membranes, if anything, represent overestimations because of the differences in other experimental conditions (e.g. calcium concentration, pH, etc.). Therefore Ca2 leakage from isolated L-SR vesicles, both in the absence and presence of volatile anaesthetic, involves paths other than SR membrane lipids. These paths may involve inhibition of Ca2 pump activity, Ca2 efflux through membrane proteins such as the Ca2 pump (slippage through ATPase channels)1 9 or through ion channels other than Ca2 release channels (e.g. non-selective ion channels),36 or Ca2 leakage through the proteinlipid interface (boundary lipids). The volatile anaesthetic-induced increase in PCa of lipid bilayers and in Ca2 leakage from L-SR vesicles showed different properties. While volatile anaesthetics increased the PCa of lipid bilayers in a dosedependent manner, such dose dependency was not observed for the volatile anaesthetic-induced increase in Ca2 leakage from L-SR vesicles. The volatile anaesthetic-induced increase in PCa of lipid bilayers was greater for isoflurane than for halothane or enflurane. This potency difference was not observed for the volatile anaesthetic-induced increase in Ca2 leakage from L-SR vesicles. Therefore, the volatile anaesthetic-induced increase in PCa of membranes formed with L-SR-extracted lipids did not appear to be directly responsible for the volatile anaesthetic-induced increase in nonspecific Ca2 leakage from isolated L-SR vesicles. Indirectly, however, the observed volatile anaesthetic-induced increase in membrane PCa may serve as a trigger to activate calcium release channels; the volatile anaesthetic-induced increase in SR

Ca2+ conductance of planar lipid bilayers membrane chloride permeability may increase the activation of non-selective ion channels,36 and the Ca2 currents through these non-selective channels may also serve as a trigger for calcium release channels. EFFECTS OF THE EXPERIMENTAL CONDITIONS AND ASSUMPTIONS ON THE ESTIMATED PCa

In this study several assumptions were made when estimating PCa. First, that increases in Ca2 did not affect the PCa of lipid membranes. At high concentrations Ca2 is known to bind to phospholipids and produce phase separation of lipid mixtures.38–40 To our knowledge, however, there are no reports indicating that increases in Ca2 affect the PCa of lipid membranes. Such increases in Ca2 would not be expected to decrease PCa, if anything they may increase PCa. Second, we used pH 6.0 to reduce proton conductance through lipid membranes. Changes in pH are known to strongly affect the function of membrane proteins, however, our experiments were performed in the absence of membrane proteins. In the absence of membrane proteins, ions permeate only through the lipid membrane. If H and Ca2 are using the same lipid pathway, reduction of protons should lead to an increase in the conduction of calcium. In this case, PCa is overestimated when using a low pH. Third, the assumption that gHEPES was negligible compared with gCa and gCl and the lack of correction for the contribution by gH/OH tends to overestimate the calculated PCa. The presence of decane in the lipid membranes should increase PCa. All of the above assumptions tend to overestimate PCa. In spite of this, our estimated control PCa values agreed closely with those obtained with Ca2 flux measurements in liposomes formed with various phospholipids or SR-lipid extracts, in which low concentrations of Ca2 were used (1 mmol litre1), no decane was present, pH was 8.0 and flux measurements were not contaminated by gH/OH.37 These observations suggest that in the absence of proteins our estimated PCa values were close to those obtained using lipid vesicles at low calcium concentration and pH 8.0,37 and that under our conditions the contribution of gH/OH was not significant. Even if there is overestimation of PCa, our main conclusion remains valid, that is that the volatile anaesthetic-induced increases in PCa, although significant, were too small to account for the previously observed volatile anaesthetic-induced increases in Ca2 leakage in skinned SR and isolated L-SR vesicles (table 5). Therefore, membrane paths other than SR membrane lipids must be involved. In SR membranes, plasmalogens (alkyl ester phospholipids) account for 72% of phosphatidylethanolamine (PE) and 12% of phosphatidylcholine (PC) (20% of total phopholipids).14 In this study we used diacyl PE and PC because plasmalogens were not available commercially. To our knowledge there are no reports on how plasmalogens affect membrane ionic permeability. It is possible, therefore, that the presence of plasmalogens in membranes formed with L-SR-extracted lipids contributes in part to the differences observed

73 between membranes formed with synthetic lipids and L-SR-extracted lipids. Halothane which, of the volatile anaesthetics studied, is the most potent trigger of malignant hyperthermia, produced the smallest increase in gCa of lipid membranes and hence the smallest predicted increase in cytoplasmic Ca2. The latter suggests that the halothane-induced increase in PCa of SR lipid membranes by itself does not appear to account for halothane-induced malignant hyperthermia. In summary, at the concentrations studied, volatile anaesthetics increased the rate of ion transport across lipid membranes without affecting the ionic selectivity of the membrane or its dimensions. The strongest effect was exhibited by isoflurane and the weakest effect by halothane; the effect of enflurane was intermediate. This volatile anaesthetic effect cannot account for the volatile anaestheticinduced increase in Ca2 leakage observed in isolated L-SR vesicles.

Acknowledgements We are grateful to Dr Bernd W. Urban for critically reading the manuscript, Dr Anthony Scotto for providing the method for measurement of phosphorus, Dr David D. Thomas for providing the method for lipid extraction, Jin Zhang for assistance in the preparation of the manuscript and Irusia Kocka for editorial assistance. Supported in part by National Institutes of Health grant GM30799 (T. J. J. B.).

Appendix In our CaCl2 solutions, the diffusion potential (Vd) or potential at which the current equals zero, was determined by the following equation: Vd = gCaECa / gm+gClECl / gm+gH/ OH E H/ OH / gm+gHEPES EHEPES / gm (1) where gCa
calcium conductance; gCl
chloride conductance; gH/OH
proton–hydroxide conductance; gHEPES
HEPES conductance; ECa, ECl, EH/OH and EHEPES
equilibrium potentials (Nernst potential) for each of the ions. As there was no pH or HEPES gradient, EH/OH and EHepes are zero, and equation (1) becomes: Vd = t Ca ECa + t Cl E cl (2) where tCa
transference number for Ca2 and tCl
transference number for Cl. This equation can be rewritten as follows: Vd= − t Ca × ( RT / 2F )× ln([Ca 2+ ]cis /[Ca 2+ ]trans ) + (3) t Cl ( RT /F )× ln ([Cl]cis /[Cl]trans ) Vd=(t Cl − t Ca / 2)( RT /F )× ln([CaCl 2 ]cis /[CaCl 2 ]trans ) (4) Vd=(t Cl − t Ca / 2)( RT /F )× ln(a cis /a trans ) (5) where R
gas constant; T
temperature in Kelvin scale; F
Faraday constant; [ ]cis and [ ]trans
ion or salt concentration in the cis (front) and trans (back) compartments; acis and atrans
mean salt activity in the cis and trans compartments (interpolated from previously published values).25 When assuming that only Ca2 and Cl ions carry the current, then: t Ca+ t Cl = 1 (6) and equation (5) becomes: (7) Vd=(1 − 1.5t Ca) × ( RT /F )× ln(a cis / atrans )

References 1. Blanck TJJ, Peterson CV, Baroody B, Tegazzin V, Lou J. Halothane, enflurane, and isoflurane stimulate calcium leakage from rabbit sarcoplasmic reticulum. Anesthesiology 1992; 76: 813–821. 2. Su JY. Effects of halothane on functionally skinned rabbit soleus muscle fibres: A correlation between tension transient and 45Ca release. Pflügers Archiv 1980; 388: 63–67. 3. Su JY, Bell JG. Intracellular mechanism of action of isoflurane and halothane on striated muscle of the rabbit.

74 Anesthesia and Analgesia 1986; 65: 457–462. 4. Nelson TE, Sweo T. Ca2 uptake and Ca2 release by skeletal muscle sarcoplasmic reticulum: Differing sensitivity to inhalational anesthetics. Anesthesiology 1988; 69: 571–577. 5. Frazer MJ, Lynch C. Halothane and isoflurane effects on Ca2 fluxes of isolated myocardial sarcoplasmic reticulum. Anesthesiology 1992; 77: 316–323. 6. Karon BS, Thomas DD. molecular mechanism of CaATPase activation by halothane in sarcoplasmic reticulum. Biochemistry 1993; 32: 7503–7511. 7. Connelly TJ, Coronado R. Activation of the Ca2 release channel of cardiac sarcoplasmic reticulum by volatile anesthetics. Anesthesiology 1994; 81: 459–469. 8. Karon BS, Mahaney JE, Thomas DD. Halothane and cyclopiazonic acid modulate Ca-ATPase oligomeric state and function in sarcoplasmic reticulum. Biochemistry 1994; 33: 13928–13937. 9. Inesi G, de Meis L. Regulation of steady state filling in sarcoplasmic reticulum. Journal of Biological Chemistry 1989; 264: 5929–5936. 10. Louis CF, Zaulkernan K, Roghair T, Mickelson JR. The effects of volatile anesthetics on calcium regulation by malignant hyperthermia-susceptible sarcoplasmic reticulum. Anesthesiology 1992; 77: 114–125. 11. Pang KY, Chang TL, Miller KW. On the coupling between anesthetic induced membrane fluidization and cation permeability in lipid vesicles. Molecular Pharmacology 1979; 15: 729–738. 12. Barchfeld GL, Deamer DW. The effect of general anaesthetics on the proton and potassium permeabilities of liposomes. Biochimica et Biophysica Acta (Netherlands) 1985; 819: 161–169. 13. Lau YH, Caswell AH, Brunschwig JP, Baerwald RJ, Garcia M. Lipid analysis and freeze-fracture studies on isolated transverse tubules and sarcoplasmic reticulum subfractions of skeletal muscle. Journal of Biological Chemistry 1979; 254: 540–546. 14. Gross RW. Identification of plasmalogen as the major phospholipid constituent of cardiac sarcoplasmic reticulum. Biochemistry 1985; 24: 1662–1668. 15. Van Winkle WB, Bick RJ, Tucker DE, Tate CA, Entman ML. Evidence for membrane microheterogeneity in sarcoplasmic reticulum of fast twitch skeletal muscle. Journal of Biological Chemistry 1982; 257: 11689–11695. 16. Recio-Pinto E, Duch DS, Levinson SR, Urban BW. Purified and unpurified sodium channels from eel electroplax in planar lipid bilayers. Journal of General Physiology 1987; 90: 375–395. 17. Hidalgo C, Ikemoto N, Gergely J. Role of phospholipids in calcium-dependent ATPase of the sarcoplasmic reticulum. Enzymatic and ESR studies with phospholipid-replaced membranes. Journal of Biological Chemistry 1976; 251: 4224–4232. 18. Bigelow DJ, Thomas DD. Rotational dynamics of lipid and Ca-ATPase in sarcoplasmic reticulum. Journal of Biological Chemistry 1987; 262: 13449–13456. 19. Chen PS, Toribara TY, Warner H. Microdetermination of phosphorus. Analytical Chemistry 1956; 28: 1756–1758. 20. Haydon DA, Hendry BM, Levinson SR, Requena J. The molecular mechanisms of anaesthesia. Nature (London) 1977; 268: 356–358. 21. Finkelstein A, Mauro A. Physical principles and formalisms of electrical excitability. In: Kandel ER, ed. Handbook of Physiology. The Nervous System, vol. 1. Bethesda: American Physiological Society, 1977; 168–169. 22. Gutknecht J. Proton/hydroxide conductance and perme-

British Journal of Anaesthesia

23. 24. 25. 26.

27. 28. 29.

30.

31. 32.

33. 34. 35. 36.

37.

38.

39.

40.

ability through phospholipid bilayer membranes. Proceedings of the National Academy of Sciences USA 1987; 84: 6443–6446. Hopfer U, Lehninger AL, Lennarz WJ. The effect of the polar moiety of lipids on the ion permeability of bilayer membrane. Journal of Membrane Biology 1970; 2: 41–58. Fuks B, Homble F. Permeability and electrical properties of planar lipid membranes from thylakoid lipids. Biophysical Journal 1994; 66:1404–1414. Lide DR, Frederikse HPR. CRC Handbook of Chemistry and Physics, 74th Edn. Boca Raton: CRC press, 1993; 5–94. Schlame M, Hemmings HC. Inhibition by volatile anesthetics of endogenous glutamate release from synaptosomes by a presynaptic mechanism. Anesthesiology 1995; 82: 1406–1416. Franks NP, Lieb WR. Selective actions of volatile general anaesthetics at molecular and cellular levels. British Journal of Anaesthesia 1993; 71: 65–76. Smith RA, Porter EG, Miller KW. The solubility of anaesthetic gases in lipid bilayers. Biochimica Biophysica Acta 1981; 645: 327–338. Allott PR, Steward A, Flook V, Mapleson WW. Variation with temperature of the solubilities of inhaled anaesthetics in water, oil and biological media. British Journal of Anaesthesia 1973; 45: 294–300. Firestone LL, Miller JC, Miller KW. Tables of physics and pharmacological properties of anesthetics. In: Roth SH, Miller KW, eds. Molecular and Cellular Mechanisms of Anesthetics. New York: Plenum Medical Book Co. 1986; 455–470. Fettiplace R, Andrews DM, Haydon DA. The thickness, composition and structure of some lipid bilayers and natural membranes. Journal of Membrane Biology 1971; 5: 277–296. Inesi G. Transport across sarcoplasmic reticulum in skeletal and cardiac muscle. In: Giebisch G, Tosteson DC, Ussing HH, eds. Transport Across Single Biological Membranes. New York: Springer-Verlag, 1978; 357–393. Kawai M, Konishi M. Measurement of sarcoplasmic reticulum calcium content in skinned mammalian cardiac muscle. Cell Calcium 1994; 16: 123–136. Page E. Quantitative ultrastructural analysis in cardiac membrane physiology. American Journal of Physiology 1978; 235: C147–C158. Malan NT, Sabbadini R, Scales D, Inesi G. Functional and structural roles of sarcoplasmic reticulum protein components. FEBS Letters 1975; 60: 1220–1225. Sukhareva M, Morrissette J, Coronado R. Mechanism of chloride-dependent release of Ca2 in the sarcoplasmic reticulum of rabbit skeletal muscle. Biophysical Journal 1994; 67: 751–765. de Boland AR, Jilka RL, Martonosi AN. Passive Ca2 permeability of phospholipid vesicles and sarcoplasmic reticulum membranes. Journal of Biological Chemistry 1975; 250: 7501–7510. Altenbach C, Seelig J. Ca2 binding to phosphatidylcholine bilayers as studied by deuterium magnetic resonance. Evidence for the formation of a Ca2 complex with two phospholipid molecules. Biochemistry 1984; 23: 3913–3920. Roux M, Bloom M. Ca2, Mg2, Li, Na and K distributions in the headgroup region of binary membranes of phosphatidylcholine and phosphatidylserine as seen by deuterium NMR. Biochemistry 1990; 29: 7078–7090. Ohki K, Sekiya T, Yamauchi T, Nozawa Y. Physical properties of phosphatidylcholine–phosphatidylinositol liposomes in relation to a calcium effect. Biochimica Biophysica Acta 1981; 644: 165–174.