694

Cell Membranes and Multilamellar Vesicles: Influence of pH on Solvent Induced Damage Myra K. Jacobsohna, b,*, Melanie M. Lehman b and Gert M. Jacobsohn a aDepartment of Biological Chemistry, Hahnemann University, Philadelphia, Pennsylvania 19102 and bDepartment of Biology, Beaver College, Glenside, Pennsylvania 19038

Pigment leakage from sheep and horse erythrocytes and from red beet tissue induced by non-polar solvents was determined as a function of pH. The results were compared to disruption of multilamellar vesicles (MLV) composed of phospholipids with equimolar cholesterol under identical conditions of solvent exposure and pH. Solvent access to cholesterol was used to measure vesicle disruption. MLV were made from 1,2-dioleoyl phosphatidylethanolamine, sphingomyelin (SP) and various phosphatidylcholines to simulate the major Hpid components of membranes. Pigment leakage from erythrocytes caused by petroleum hydrocarbon (b.p. 60-80~ was maximal at pH 2-4 and at pH 10, but minimal at pH 6.8; alcohols caused less pigment leakage than petroleum hydrocarbon. Betacyanin leakage from beet tissue induced by petroleum hydrocarbon was maximal at pH 2, with very little leakage at pH 4, 6.6 and pH 10. Alcohols caused minimal damage to beet tissue above pH 2. Cholesterol removal by petroleum hydrocarbon from MLV of miYed lipid composition was maximal at pH 2-4, reduced at pH 6.8 and minimal at pH 10. Lipid mixtures in which fatty acyl side chains of one phospholipid were of a different length than the other lost more sterol than miYtures in which the acyl side chains were of identical chain length. MLV with more than 25% SP lost more sterol than those with less or no SP. Results show that in mixtures of phospholipids, SP exposes the hydrocarbon phase of a bilayer to solvent extraction, a property that was also observed in native membranes. Erythrocyte membranes, which contain SP, were more severely damaged by petroleum hydrocarbon than beet cells, which have none. Membranes from erythrocytes were more prone to solvent disruption at pH 10 than MLV, but they were more resistant at physiological pH. It is suggested that conformational changes in membrane proteins due to shifts in pH cause exposure of hydrophobic portions of surrounding lipids to the environment. At neutral pH, the native conformation of proteins is expected to stabilize the bilayer of membranes. Lipids 27, 694-700 (1992).

phase, cholesterol would become accessible to the nonpolar solvent and the sterol would be extracted. This method was used previously to study the influence of side chain mobility, head group type and pH on the stability of bilayers (1,2). Results showed that at moderately acid to neutral pH, phospholipids with mobile side chains and]or those with ethanolamine-containing head groups form unstable bilayers from which cholesterol and phospholipids are readily removable, which may be a manifestation of reverse hexagonal phase formation (3). Conditions which would cause the head groups to occupy more space, Le., like charges and increased hydration, would increase the stability of bilayers. In the present study, we compared the effect of organic solvents upon pigment leakage from whole cells to disruption of the lipid structure of multilamellar vesicles (MLV) caused by the same solvents. Leakage of betacyanin from beet root cells and of hemoglobin from erythrocytes were used as indicators of disruption of cell membranes. Pigment leakage was compared to cholesterol extraction from mixed lipid MLV under similar conditions of pH and solvent exposure. Cell membranes were expected to differ from model systems because native membranes include substantial amounts of protein which may be subject to conformational changes by pH and because the lipid compositions of model vesicles were not exact duplicates of those in membranes. MATERIALS AND METHODS

Fresh beets were purchased from a local market. Sheep and horse erythrocytes were obtained from Environmental Diagnostics {Burlington, NC). Blood cells were washed at least three times with isotonic phosphate buffer, pH 6.8, prior to use and were made up to a uniform 50% suspension in the same buffer. Cells were held at 4~ until us~ Dioleoyl-sn-glycero-3-phosphocholine (DOPC) and dioleoyl-sn-glycero-3-phosphoethanolamine (DOPE) were purchased from Avanti Polar Lipids (Alabaster, AL); bovine brain sphingomyelin (SP) was obtained from P-L We have applied a simple physico-chemical procedure Biochemicals {Milwaukee, WI); dipalmitoyl-sn-glyceroto study the stability of bilayers formed from single or 3-phosphocholine (DPPC) was from Sigma Chemical Co. mixed phospholipids. The Criterion for stability was the {St. Louis, MO). Cholesterol was recrystallized twice from capacity of bilayers to retain cholesterol upon exposure methanol. [4-14C]Cholesterol was obtained from Amerto a non-polar solvent. The assumption was that if phos- sham {Arlington Heights, IL). The purity of phospholipids pholipids form stable bilayers, the polar head groups was checked by thin-layer chromatography on silica would repel the solvent and allow for only limited extract- gel G plates developed with chloroform]methanol/water ability of cholesterol. If the packing of the bilayer was (65:25:4, vol/vol/vol) and visualized with iodine vapors. All unstable, or if the lipids formed a reverse hexagonal other substances and solvents were reagent grade and were used without further purification. *To whom correspondence should be addressed. Stability of cell membranes. The stability of beet root Abbreviations: CAPS, cyclohexylaminopropanesulfonicacid; C/M, membranes was determined by a modification of the chloroform/methanol 2:1 (vol]vol);DOPC, 1,2-dioleoyl-sn-glycero- method described by Grunwald (4). Cores were cut from 3-phosphocholine;DOPE, 1,2-dioleoyl-sn-glycero-3-phosphoethanol- beets with a cork borer 1#4) and sectioned freehand with amine; DPPC, 1,2-dipslmitoyl-sn-glycerc~3-phosphocholine;Hip reversehexagonalphase; MES, 2-(N-morpholino)ethanesulfonicacid; a razor blade into 3-mm slices. The end slices from each MLV, multilamellar vesicles; PC, phosphatidylcholine; PE, core were discarded and the remaining slices were washed phosphatidylethanolamine; PS, phosphatidylserine; SP, under cold running tap water for several hours or oversphingomyelin; TLC, thin-layer chromatography. night. To initiate an experiment, 10 root slices were placed LIPIDS, Vol. 27, no. 9 (1992)

695

SOLVENT STRESS AND PH EFFECTS ON MEMBRANES in each of fifteen 100-mL beakers, with 9.0 or 10.0 mL of a designated sucrose-buffer solution. Buffers were 0.4 molal with respect to sucrose. The buffers used were 0.1 M KH2PO4 and phosphoric acid at pH 2.0; 0.1 M 2-(N-morpholino)ethanesulfonic acid (MES)/NaOH at pH 4.0; KH2PO4/NaOH at pH 6.6 or 6.8; 0.1 M cyclohexylaminopropanesulfonic acid (CAPS)/NaOH at pH 10.0. For controls, beet root slices were incubated in 10.0 mL of a sucrose-buffer; for solvent treated samples, they were incubated in 9.0 mL of a sucrose~buffer with 1.0 mL of either methanol, ethanol, a 2:1 (vol/vol) mixture of chloroform/ methanol, or petroleum hydrocarbon (b.p. 60-80~ The organic solvents were added at the start of an experiment. At time zero and periodically thereafter, the absorbance of betacyanin at 535 nm was determined. All incubates were kept at 22-23~ in a constant temperature bath with continuous shaking. Incubations continued for 2 h, and all experiments were run in duplicate~ For experiments on erythrocytes, incubations were carried out at room temperatur~ The same series of buffers described for the beet cell experiments were used except that pH 6.8 was substituted for pH 6.6 and the buffers were made isotonic with NaC1. Experiments were initiated by addition of 200 ~L of a 50% red cell suspension. The contents of each tube were mixed immediately after addition of red cells and were agitated by vortexing briefly e v e r y 20 s. After 10 min, the tubes were centrifuged for 10 or 20 min in a benchtop centrifuge to sediment unlysed cells, membranes and denatured hemoglobin. Absorbance of supernatants was read at 590 nn~ All experiments were run in triplicate Stability of multilarnellar vesicles. The effect of pH on pigment leakage from cells or vesicles was compared to the loss of cholesterol from bilayers prepared from pure lipids. MLV were prepared from phospholipids in approximately similar proportions to those reported in the literature for that type of cell. Phosphatidylcholine {PC), phosphatidylethanolamine {PE) and SP lipids with equimolar amounts of sterol, to which [4-14C]cholesterol (22,000 cpm) had been added, were incorporated into MLV. The procedure for making and incubating MLV and extracting them with petroleum hydrocarbon has been described previously (2). A slight change in the procedure was made in that lipid films were suspended in 0.4 mL of appropriate buffer solutions for 10 rain prior to vortexing with glass beads, and the samples were left undisturbed for 45 to 60 min prior to extraction with solvent. In some of the experiments, the three sequential petroleum hydrocarbon extracts of each sample were assayed individually. Extracted cholesterol was determined by scintillation counting of t h e residue left, after evaporation of the solvent, in Ecoscint fluid {National Diagnostics, Manville, N J). The pattern of cholesterol removal from the vesicles revealed information about the environment of the hydrocarbon components of the bilayers. RESULTS

Betacyanin leakage from beet root cells. Incubation of b e e t root slices in sucrose-buffer solutions with and without organic solvents for a period of 2 h showed that near maximum amounts of pigment leakage due to chloroform/ methanol (C/M) occurred within 30 min after the start of the experiment. Data for pH 6.6 are shown in Figure 1.

Pigment leakage with other solvents increased linearly over the 2 h time span with leakage caused by solvents only slightly higher than that observed in samples with no solvent (Fig. 1, insert}. In the beet cell and later in the erythrocyte experiments, we observed that pigment leaking from cells was affected by pH. Hemoglobin turned brown and precipitated at low pH, while betacyanin slowly turned yellow at high pH. Denatured hemoglobin exhibited maximum absorbance at 590 nm. Betacyanin was measured immediately at the end of incubation. Alterations in the pigments due to pH were reflected in changes in absorbance. In order to compare samples at one pH with the others, we assigned pigment leakage due to C/M as the maximum {100%} that could be obtained from cells at any given pH. Pigment leakage from all other samples, including controls with no organic solvent, were compared to C/M samples at the same pH and expressed as percent of maximum leakage. We calculated the percentage of maximum damage based upon pigment leakage after 30 min of incubation. The experiment was repeated twice and yielded the same pattern of results each time but absolute values for pigment leakage differed with different batches of beets. Results for one experiment are depicted in Figure 2. At pH 2.0, the samples without organic solvent showed considerable pigment leakage. Addition of any organic solvent increased the amount of betacyanin leakage" but ethanol and petroleum hydrocarbon caused the most damage. At pH 4.0 and 6.6, none of the samples with or without organic solvent showed any significant amount

3

.f.~"

E

I/"

c-

cO 03 tO

~ 2

I / /

0

I

/

i,_ 09

o 1o

.':~'J ~=;

,

/

c-

o

0.15

i

0.05

/ / /


r

E

40

.•

t

0

I t

300

p.

=...I

E

400

\

0

20 0 O~ r

% ";,, \

\

2

2oo

o 0

100

o)

\r

0 ,,_1

I

2

4

6

8

10

,

2

pH FIG. 2. Effect of solvents on h e t a c ~ - i - leRkage from beet root tissue as function of pH at 30 min. Absorbance at 535 nm of samples exposed to an organic solvent at a given pH was compared to absurbance at the same pH and wavelength of samples exposed to chloroform/methanol (2:1, vol/vol). Organic solvents were present to the extent of 10% the volume of the suspending medium. Conditions are described under Materials and Methods. All points are averages for three independent experiments. Standard deviations are included within symbols or are shown by vertical bars. &, Buffer alone; O, methanol; V, ethanol; e, petroleum hydrocarbon.

I

4

,

!

I

6

|

8

,

\ F"I

10

j

I

12

pH FIG. 3. Cholesterol extraction from MLV of DOPC/cholesterol (I:I, molar ratio), DOPE/cholesterol (1:I, molar ratio), DOPC/DOPE/cholesterol (1:1:2, molar ratio) or SP/DOPC/DOPE/cholesterol (0.5:0.5:1:2, molar ratio) related to the pH of the suspension media. MLV were prepared from pure phospholipids with [4-14C]cholesterol equimolar to total phospholipids. The MLV were extracted three times with petroleum hydrocarbon. Details are given in the Methods section. Plots show the sum of cholesterol removed in all extracts. Experiments were done in triplicate and averages are indicated. Standard deviations are included in the symbols or are shown by vertical bars. [2, DOPC; O, DOPE; A, DOPE/DOPC; e , SP/DOPC/DOPE.

of pigment leakage compared with leakage caused by C]M. At pH 10.0, pigment leakage from samples without organic solvent increased slightly, but organic solvents 500 caused little, if any, additional leakag~ Except at pH 2.0, organic solvents did not sigm'ficantly aggravate damage due to changes in pH. In order to compare the behavior of beet cells with that , ~ 400 of pure lipid bflayers, MLV of lipids similar to those found in beet root tissue were prepared. Phospholipid analysis of beet root tissue reported by Beiss (5) showed that PC 300 constituted 35% of the lipid and PE 17%, with phosphatidylserine (PS) and phosphatidic acid mnklng up the balanc~ A combination of DOPCJDOPE/cholesterol (1:1:2, 200 molar ratio) was used as an analog to beet cell membranes. We have previously found the extractability characteristics of MLV made from PC with two 18:1 side chains 0 100 (DOPC) to be very similar to those of PC with one 16:0 0j= and one 18:1 side chain (2}. These MLV lost decreasing amounts of cholesterol upon extraction as the pH increase~ At pH 2.2, about 150 pg or 30% of the sterol were 2 4 6 8 10 12 lost; at pH 4.6 and 6.8, 100 ~g or 20% was lost and at pH 10.4, less than 50 ~g or 10% was lost (Fig. 3). While this pH pattern of data was similar to that observed with whole cells, a minimum of cholesterol extraction did not occur in the physiological range, nor were the properties of the FIG. 4. Cholesterol extraction from MLV of DPPC/cholesterol (I:I, ratio), DOPE/cholesterol (1:I, molar ratio) and DPPC/DOPE! mixed lipid vesicles an intermediate between the charac- molar cholesterol (1:1.2 molar ratio) as function of pH of suspending medial teristics of the single lipid components. To determine Legend as in Figure 3. [::], DPPC; O, DOPE; &, DPPC/DOPE. whether similarity of head groups or of acyl side chains was responsible for the characteristics of the lipid vesicles, MLV of DPPC and DOPE (1:1) withequimolar cholesterol to cholesterol extraction as those prepared from DOPE were prepared and studied for cholesterol loss to nonpolar alone. More than 450 ~g or 90% of the cholesterol was exsolvent. Results are shown in Figure 4. These mixed lipid tracted at pH values below 6.8. At pH 10.4, 150 ~g or 30% vesicles had much the same characteristics with respect of the cholesterol was extracted, but this was more than

f

LIPIDS, Vol. 27, no. 9 (1992)

,

697

SOLVENT STRESS AND PH EFFECTS ON MEMBRANES

the amount extracted from single species vesicles. Data from these experiments and those reported earlier (2) indicate that acyl side chain length and unsaturation are important parameters in determining the stability of lipid bilayers. Hemoglobin leakage from erythrocytes. Hemoglobin leakage from sheep and horse erythrocytes was determined after 10 rain of incubation in isotonic buffer solutions of pH comparable to those used in the beet root experiments. Results for sheep erythrocytes are shown in Figure 5. At pH 2.0, buffered samples showed about 40% of maximum pigment leakage. Organic solvents did not greatly increase the leakage of pigment, except for petr~ leum hydrocarbon, which caused a maximum of 55% leakage. At pH 4.0 the differences were much more prc~ nounced. At this pH the buffered red cells exhibited less pigment leakage than at pH 2.0, but about 20% more than was observed with betacyanin leakage from beet root cells. Methanol and ethanol did not increase pigment leakage from erythrocytes, but petroleum hydrocarbon showed nearly twice as much leakage as buffer incubated samples (40% vs. 20%}. At pH 6.8, untreated and ethanol breated samples showed no pigment leakage, while methanol and petroleum hydrocarbon caused 8 to 10% leakage. At pH 10.0, there was virtually no leakage from samples with alcohols, but the ones incubated with petroleum hydrocarbon showed almost as much pigment loss as the C/M exposed samples {>80%}. Results with horse erythrocytes {Fig. 6) were similar to those with sheep erythrocytes with some notable exceptions. At pH 4.0, no hemoglobin leaked from buffered samples and very little leakage was observed with meth-

80

I

anol or ethanol. Petroleum hydrocarbon, however, caused even more leakage than at pH 2.0. At pH 6.8, neither the buffered nor methanol or ethanol samples showed any pigment leakage, while petroleum hydrocarbon caused about 10% of maximum pigment leakage At pH 10.0 there was no pigment leakage from buffered samples and very little from methanol or ethanol, but the leakage caused by petroleum hydrocarbon rose to 35%. This was less than half the damage observed in sheep erythrocytes with the same solvent. The lipid compositions of membranes from sheep and horse erythrocytes differ considerably. The predominant phospholipids in sheep erythrocyte membranes have been reported to be SP {51.0%} and PE {26.2%} (6). We confirmed the absence of PC by isolating sheep erythrocyte ghosts, extracting the lipids with C/M, and analyzing them by thin-layer chromatography (TLC) (Jacobsohn, M.K., Lehman, M.M., and Jacobsohn, G.M., unpublished observations}. Horse erythrocyte membranes contain 42.4% PC, 13.5% SP and 22.4% PE (6). The remainder of the phospholipid was mainly phosphatidylserine (PS). We measured cholesterol extractability from MLV of SP and mixtures of SP with PE or PC plus equimolar amounts of sterol and compared the data to hemoglobin leakage from erythrocytes. Data on petroleum hydrocarbon extraction of cholesterol from vesicles of SP are shown in Figure 7. A mixture of SP/DOPE/cholesterol (1:1:2, molar ratio} allowed petroleum hydrocarbon to extract more than 480 ~g or 95% of the sterol at pH