- Email: [email protected]
Does cholesterol discriminate between sphingomyelin and phosphatidylcholine in mixed monolayers containing both phospholipids?
CPL ELSEVIER
Chemistry and Physics of kipids 81 (1996) 69 80
CHEMISTRY AND PHYSICS OF LIPIDS ,
Does cholesterol discriminate between sphingomyelin and phosphatidylcholine in mixed monolayers containing both phospholipids? Peter Mattjus*, J. Peter Slotte Department qf Biochemistry and Pharmacy Abo Akademi University, FIN-20520 Turku, Finhmd
Received 24 November 1995: re~.ised 5 February 1996: accepted 7 February 1996
Abstract
The objective of this work was to examine the interaction of cholesterol with both phosphatidylcholines, l-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC) or 1,2-dipalmitoyl-sn-glycero-3-phosphocholine IDPPC), and sphingomyelins, N-oleoyl-D-sphingomyelin(O-SPM) or N-palmitoyl-o-sphingomyelin(P-SPM), in monolayers at an air/water interlace. We used cholesterol oxidase to probe for the relative strength of sterol-phospholipid interaction, and fluorescence microscopy to visualize lateral domain formation in the mixed monolayers. The ternary mixed monolayers, which contained cholesterol, POPC, and O-SPM had a twofold higher average oxidation rate than the corresponding system containing DPPC and P-SPM. This difference in oxidation rate between saturated and unsaturated systems was observed irrespective of the ratio between phosphatidylcholine and sphingomyelin in the monolayer. With either the saturated or the unsaturated systems, however, the rate of oxidation was influenced by the ratio of phosphatidylcholine to sphingomyelin. As the monolayer content of phosphatidylcholine increased and the sphingomyelin content decreased correspondingly (to maintain a constant cholesterol-to-phospholipid molar ratio), an increase in the average oxidation rate was seen in both saturated and mono-unsaturated monolayer systems. The relationship between the rate of cholesterol oxidation and the phosphatidylcholine/sphingomyelinratio was not linear, suggesting a preferential interaction of cholesterol with sphingomyelin even when phosphatidylcholine was present in the monolayer. The formation and stability of cholesterol-rich lateral (liquid-condensed) domains in the monolayers, as determined by monolayer fluorescence microscopy, was found to be highly influenced by the phospholipid class, the degree of acyl chain saturation, and by the ratio of phosphatidylcholine to sphingomyelin in the monolayer. The differences in cholesterol oxidation rates and lateral domain formation, as a function of the ratio of two phospholipids in the monolayers, apparently derived from differences in the hydrophobic interactions between the lipids. Keywords: Monolayers; Lipid domains; Cholesterol; Phosphatidylcholine; Sphingomyelin; Lipid interactions: Cholesterol oxidase; Fluorescence microscopy Abbreviations: TRITC-PE, tetramethylrhodamine isothiocyanate-phosphatidylethanolamine; POPC, l-palmitoyl-2-oleoyl-sn-glyccro-3-phosphocholine: DPPC, 1,2-dipalmitoyl-sn-glycero-3-phosphocholine:O-SPM, N-oleoyl-D-sphingomyelin:P-SPM, N-palmitoyl-D-sphingomyelin. * Corresponding author, Department of Biochemistry and Pharmacy, Abo Akademi University, P.O. Box 66, FIN-20521 Turku, Finland. Fax: + 358 21 2654 745.
11009-3084/96/ $15.00 ~ 1996 Elsevier Science lreland Ltd. All rights reserved PII S0009-3084(96)02535-2
70
P. Mattjus, J.P. Slotte / Chemistry and Physics ~/' LipMs 81 (1996) 69 80
I. Introduction
In biological as well as in model membranes, cholesterol interacts with phospholipids, thereby affecting their physicochemical behavior (for a review of the literature, see [1]). The interaction of cholesterol with phospholipids is affected by both the acyl chain composition of the phospholipids, and by the nature of their head-group [2]. The specific interaction of cholesterol with sphingomyelins and phosphatidylcholines is of biological importance, due to the high concentration of these lipids in the plasma membrane compartment of most cells [3,4]. The condensing effect of cholesterol on the lateral packing of phosphatidylcholine and sphingomyelin in monolayers have been shown to be similar [5], although the strength of interaction between cholesterol and phosphatidylcholine or sphingomyelin is known to differ [6]. This difference in interaction is most clearly seen when cholesterol desorption rates from phosphatidylcholine- or sphingomyelin-containing membranes are measured [7,8]. Sphingomyelin monolayers are reported to have a smaller surface potential compared to phosphatidylcholine monolayers, probably due to the lack of the carbonyl ester dipole in sphingomyelin [9,10]. The compressibility modulus has also been shown to be much higher in cholesterol-sphingomyelin monolayers compared to cholesterolphosphatidylcholine monolayers [11]. When cholesterol oxidase was used to probe for the strength of cholesterol interaction with phosphatidylcholines and sphingomyelins, is was shown that the oxidation susceptibility of cholesterol in sphingomyelin monolayers was reduced compared to the rate determined in phosphatidylcholine monolayers [12,13]. It was also shown that the presence of low or moderate amounts sphingomyelin in a ternary monolayer with cholesterol and phosphatidylcholine could influence the stoichiometry at which cholesterol became more resistant to oxidation by cholesterol oxidase [13]. Finally, a recent study of binary mixed monolayers containing cholesterol and either phosphatidylcholine or sphingomyelin, using monolayer fluorescence microscopy, showed a marked difference in the properties of the lateral
sterol-rich domains formed in phosphatidylcholine- and sphingomyelin-containing monolayers [14]. These and other results [15,16] suggest that cholesterol is more strongly associated with sphingomyelin than with phosphatidylcholine in monolayers and bilayers. However, an earlier study using the X-ray diffraction technique and differential scanning calorimetry failed to show a preferential interaction between cholesterol and either N-palmitoyl sphingomyelin or dipalmitoyl phosphatidylcholine, when all three lipid classes were present in model bilayer membranes [17]. A preferential interaction between cholesterol and either sphingomyelin or phosphatidylcholine was also not seen in equilibrium exchange experiments using erythrocyte ghost membranes and sonicated vesicles [181. In order to address the question of selective interaction between cholesterol and sphingomyelin or phosphatidylcholine, we have examined the interaction of cholesterol with these phospholipids in binary (cholesterol and one phospholipid) and ternary (cholesterol together with both phospholipids) mixed monolayers at the air/water interfaces. The association of cholesterol with the phospholipids was determined using cholesterol oxidase as a probe for the strength of interaction. We also observed the monolayers visually using monolayer fluorescence microscopy to gain possible information about the formation of lateral cholesterol-rich domains and the effects of phospholipid type and properties on domain formation.
2. Material and methods
2.1. Materials
Cholesterol, DPPC, POPC, O-SPM and PSPM were obtained from Sigma Chemicals (St. Louis, MO). The cholesterol was at least 99% pure. The phospholipids gave all a single spot when analyzed by thin layer chromatography. Tetra-methylrhodamine isothiocyanate phosphatidylethanol amine (TRITC-PE) was purchased from Molecular Probes Inc. (Eugene, OR) and was used as delivered. It was stored in
P, Mat(]us, J.P. Slotte / Chemistry and Ph)sics 0[ LipMs 81 (1996) 69 80
aliquots in the dark, at - 2 5 ° C , under an argon atmosphere. Stock solutions of the phospholipids were prepared in hexane/2-propanol (3:2 v/v), and stored in the dark at - 2 5 ° C . The solutions were warmed to room temperature before use. The concentration of the various phospholipid stock solutions was determined by the method described by Bartlett [19]. Cholesterol oxidase (3fl-hydroxysteroid oxidase, EC 1.1.3.6, Streptomyces cinnamoneus) was purchased from Calbiochem (San Diego, CA) and was dissolved in a Tris-buffer containing 50 mM Tris-HC1 (pH 7.5) and 140 mM NaCI. The enzyme was dispensed into 160/~1 aliquots, stored at - 2 5 ° C , and used within 2 h after thawing to 0°C. Buffer salts were of pro analysis grade and the water was purified with reverse osmosis, and further passed through a Milli-Q U F + filtering system to a resistivity of 18.2 M~2/cm.
2.2. Oxidation of cholesterol in ternary mixed monolayers The oxidation of monolayer cholesterol was performed in a zero-order-type teflon trough using a KSV 3000 surface barostat (KSV Instruments Ltd., Helsinki, Finland) as described previously [12]. The monolayers were prepared to a molar ratio of 1.5:1 of cholesterol to phospholipid, and the phospholipids used were either DPPC and P-SPM or POPC and O-SPM. The ratio of the two phospholipids was varied in increments of 10%, while the cholesterol-to-phospholipid ratio was constant. The subphase (Trisbuffer) in the reaction compartment (28.3 cm 2, 30 ml volume) was magnetically stirred (100 revs./ min), and thermostated to 30°C. The lipid solution (containing cholesterol and the phospholipids) was prepared to include 300 nmol lipids in a 1.8 ml glass vial, and was mixed immediately prior to spreading on the air/water interface using a Hamilton syringe. After the monolayer had been compressed to a lateral surface pressure of 20 raN/m, and allowed to stabilize (about 5 rain from the spreading), cholesterol oxidase was added to the reaction compartment (in 40 /tl, to a final concentration of 23 mU/ml). Constant surface pressure was maintained by
71
computer-controlled compensatory barrier movement throughout the experiment. Three different kinetic measurements were made with each film composition. Calculation of the average oxidation rate was done as described previously 1112].
2,3. Monolayer preparation and Hsualization with epifluorescence microscope The lipid solution (300 nmol total lipid, with 0,5 mol% TRITC-PE) was prepared in a glass vial immediately prior to spreading. The monolayer was allowed to stabilized for about 3 min, at an available mean molecular area of 133.5/k 2, before compression was started. The monolayers were subjected to symmetric compression at a speed not exceeding 4 A2/molecule/min, at ambient temperature (about 20°C). The surface texture of the monolayers were documented at a surface pressure of 0.2 mN/m both before and after a compression/expansion cycle using a KSV Mini systems apparatus (KSV Instruments Ltd., Helsinki) mounted on an Olympus IMT-2 fluorescence microscope, as described previously [14]. During the compression/expansion cycle, the monolayers were compressed to 35 raN/m, at which pressure the monolayers were allowed to stabilize for 1 min, before relaxing back to a low surface pressure (about 0.2 mN,m). At least two runs were performed with each lipid composition, and one representative micrograph was chosen from a number of similar images.
3. Results
3,1. Oxidation of cholesterol in mixed monolayers The relative strength of the interaction between cholesterol and different phospholipids in mixed monolayers at the air/water interface was assayed using cholesterol oxidase as a probe. The susceptibility to oxidation of cholesterol has been shown to depend on the ability of neighboring phospholipid to interact with and 'shield' the cholesterol molecule from being exposed to the enzyme at the water/lipid interface [12,13]. The various monolayer systems which were examined were prepared
P. MatOus, J.P. Slotte / Chemistry and Ph.vsics of Lipids 81 (1996) 69-80
72
to have a cholesterol to phospholipid mol ratio of 1.5:1. Although this ratio is unphysiologically high, it was chosen because otherwise a comparison between phosphatidylcholines and sphingomyelins would not have been possible. This is because cholesterol oxidation in sphingomyelinmembranes below 50 mol% cholesterol is too slow to be measurable. The average oxidation rates of cholesterol present in membranes comprising saturated or mono-unsaturated phospholipids are shown in Fig. 1. The monolayers containing mono-unsaturated phospholipids (varying proportions of POPC and O-SPM) had on average a twofold higher oxidation rate as compared with monolayers containing saturated phospholipids (DPPC and P-SPM; Fig. 1). As the proportion of phosphatidylcholines increased, and the concentration of sphingomyelins decreased correspondingly, an increase in the average oxidation rate
%
O0 90
gO
70
U'hbJqorfs} e l i n
60
50
zc
5C
20
1©
0
30 40 .50 60 70 % PhosphcJtidyl¢ holinc
80
90
I CO
13 ~4 ""
i.2
-
1 i ©
b8 +S
O7 L~ b 05
3 o4 ,', 5 ~:,' <:
02 0 i
;
q
10
/C'
Fig. 1. Average oxidation rate versus the variation in the phospholipid content. Mixed monolayers were prepared to have a cholesterol-to-phospholipid tool ratio of 1.5, and the ratio of phosphatidylcholine to sphingomyelin was varied in increments of 10%. The monolayers where exposed to 23 m U / m l cholesterol oxidase at 30°C at a surface pressure of 20 m N / m . The filled circles represent cholesterol with POPC and O-SPM, and the open circles cholesterol with DPPC and P-SPM. The dotted line represents the oxidation rate for ideal mixtures of cholesterol and phospholipids. Values are averages _+ S.D. from three experiments at each phospholipid increment.
was seen for both systems. The observable oxidation rate was not, however, linearly proportional to the fractional phospholipid composition, since a negative deviation from an ideal situation was observed (Fig. 1). It appeared that in both phospholipid systems (i.e., saturated and mono-unsaturated monolayers), the retarding effect of sphingomyelin on the oxidation rate was greater than would be expected based on ideal mixing of cholesterol with either phospholipid type. Therefore, these novel results suggest that cholesterol in these monolayer membranes interacted preferentially with sphingomyelin even when both phospholipids were in the same membrane. 3.2. SurJace texture of mixed monolayers To study the ability of cholesterol to form lateral domains with phosphatidylcholines and sphingomyelins, four different ternary mixed monolayer systems were examined visually with monolayer fluorescence microscopy. The monolayer concentration of cholesterol was either 20 or 33 mol%, with the phospholipids being either saturated (DPPC and P-SPM) or mono-unsaturated (POPC and O-SPM). All micrographs were taken at a surface pressure of about 0.2 mN/m, both during initial compression and after a compression/expansion cycle. The fluorescent probe, TRITC-PE, was included in each monolayer at 0.5 mol%. This probe, like other fluorescent lipid probes, partition preferentially into expanded domains, giving these a bright fluorescence, while being effectively excluded from condensed domains. 3.2.1. 20 mol% cholesterol During initial compression of a monolayer with 20 tool% cholesterol in DPPC only, cholesterolrich laterally condensed domains were formed. These were circular in shape (i.e. liquid) but unequal in their size distribution (Fig. 2A). The addition of P-SPM to the monolayer (still keeping the cholesterol mol% at 20) led to the formation of much larger cholesterol-rich domains during initial compression (Fig. 2B depicts the monolayer with 70% DPPC and 30% P-SPM). Only when the monolayer phospholipid composition approached
P. Mattjus, J.P. SIotte / Chemistry and Physics (~f Lipids 81 (1996) 69 80
73
A
Fig. 2. Domain formation in mixed saturated monolayers. The mixed monolayers contained 20 tool% cholesterol and 80 tool%, phospholipids (DPPC and P-SPM), together with 0.5 mol% TRITC-PE. The micrographs were documented at a surface pressure of 0.2 mN/m. The top row shows monolayers during initial compression, whereas the bottom row depicts mc~nolayers after a compression/expansion cycle. Panels A and D are with 100% DPPC and no sphingomyelin, panels B and E are with 70% DPPC and 30% P-SPM, while panels C and F are 100% P-SPM and no phosphatidyleholine. The scale bar represents 100 pm.
pure P-SPM did the cholesterol-rich domains again decrease in diameter (Fig. 2C). However, if the monolayers were compressed and expanded, the domain properties were remarkably similar, irrespective of the phospholipid composition (Fig. 2D-E). For the rnono-unsaturated phospholipid monolayers at 20 mol% cholesterol (during initial compression), the cholesterol-rich domains were smallest for O-SPM (Fig. 3D), whereas they were
larger with POPC present (Fig. 3A-C). The domains became very small in all mixed monolayers after a compression-expansion cycle(Fig. 3E H). 3.2.2. 33 mol % cholesterol
Increasing the cholesterol concentration to 33 mol% resulted in the formation of partly coalesced, cholesterol-rich domains during initial compression. With DPPC and cholesterol, larger liquid-condensed domains against a liquid-ex-
B
Fig. 3. Domain formation in mixed unsaturated monolayers. The mixed monolayers contained 20 mol% cholesterol and 80 mol% phospholipids (POPC and O-SPM), together with 0.5 mol% TRITC-PE. The micrographs were documented at a surface pressure of 0.2 mN/m. The top row shows monolayers during initial compression, whereas the bottom row depicts monolayers after a compression/expansion cycle. Panels A and E are with 100% POPC and no sphingomyelin, panels B and F are with 90% POPC and 10% O-SPM, panels C and G are with 10% POPC and 90% O-SPM, while panels D and H are 100% O-SPM and no phosphatidylcholine. The scale bar represents 100 /~m.
A
t~
.5
e~
5~
P. Mat(/us, J.P. Slotte / Chemistry and Physics ()f Lipids 81 (1996) 69 8 0
A
B
75
C
J Fig. 4. Domain formation in mixed saturated monolayers. The mixed monolayers contained 33 mol% cholesterol and 67 tool% phospholipids (DPPC and P-SPM), together with 0.5 mol% TRITC-PE. The micrographs were documented at a surface pressure of 0.2 mN/m. The top row shows monolayers during initial compression, whereas the bottom row depicts mon01ayers after a compression/expansion cycle. Panels A and D are with 100% DPPC and no sphingomyelin, panels B and E are with 90% DPPC and 10% P-SPM, while panels C and F are 100% P-SPM and no phosphatidylcholine. The scale bar represents 100/~m.
panded phase were observed (Fig. 4A). Adding P-SPM to the monolayer (with a corresponding decrease in DPPC concentration) led to smaller extent of domain fusion (Fig. 4B depicts a monolayer with 90% DPPC and 10% P-SPM). Cholesterol in a P-SPM monolayer (at 33 mol%) also had only a limited extent of fusion of cholesterol-rich domains (Fig. 4C). After a compression/expansion cycle, all monolayers examined displayed completely coalesced condensed domains (Fig. 4D E).
In monolayers with 33 tool% cholesterol and mono-unsaturated phospholipids, the lateral domains formed during initial compression were markedly different, going from pure POPC to pure O-SPM as the phospholipid component. Whereas small circular domains were seen in cholesterol/ POPC monolayers, no clear domains could be resolved in cholesterol/O-SPM monolayers (Fig. 5A and E). When POPC was added to the cholesterol/O-SPM mixed monolayer, condensed do-
R
#-,
iJ
E
Fig. 5. Domain formation in mixed unsaturated monolayers. The mixed monolayers contained 33 mol% cholesterol and 67 mol% phospholipids (POPC and O-SPM), together with 0.5 mol% TRITC-PE. The micrographs were documented at a surface pressure of 0.2 mN/m. The top row shows monolayers during initial compression, whereas the bottom row depicts monolayers after a compression/expansion cycle. Panels A and F are with 100% POPC and no sphingomyelin, panels B and G are with 90% POPC and 10% O-SPM, panels C and H are with 50% POPC and 50% O-SPM, panels D and I are 10% POPC and 90% O-SPM, while panels E and J are with 100% O-SPM and no phosphatidylcholine. The scale bar represents 100 #m.
A A~
I
Y.,"
C~
P. Matt~us J.P. Slotte / Chemistry and Physics ol l.grids 81 (1996) 69 80
% P E
I (3C1 '~} 3
30
"-.
i
77] ~ -
~ [) v-
~:o,
r
:;
l:? I
121
2C
r
I
-
.....
1C
.i
•
-
.'f ]
E
c
'
)
t:"
:,Jrl (-:
--
:
::11, '
~
)',
!1)
T
-
[
.
,
Jl
,!~
i
!
,].~
,1
"
,:4
]
s ~:,
~'
"
)[:
"-4
:, i'~
[,PV I~ 121:'
77
2 ,::
.
I
-
!:
]
; i
IC' t
0
!O
i__
0
5~
i 40
~,()
~.
&
~
ii/3
7('
~(]'
L i{)
. "
:::
!.
.
i/'. r:
. I "
.
. 2
t :3
:0
1C
,°4 DF'FC
;'
!
_ ?
_ )
L
qr
:C)l
Fig. 6. Stability of cholesterol-rich domains. The phase transformation pressure was determined visually during compression of the monolayers, and panel A shows transformation pressures for saturated monolayers whereas panel B gives results with mono-unsaturated monolayers. The dotted lines represent monolayers with 20 tool% cholesterol, the solid lines monolayers with 33 mol% cholesterol. Values are averages + S.D. from two monolayers.
mains started to form, and these had also partially coalesced (Fig. 5B-D). The elongated form of the domains seen in (Fig. 5D; 10% POPC and 90% O-SPM; 33 mol% cholesterol) was at least in part due to excitation light, since photochemical effects are known to disrupt domain boundary stability close to the phase transformation pressure [20,21]. After the monolayers were subjected to a compression/expansion cycle, the condensed domains in the cholesterol/POPC monolayer became slightly smaller than they were during initial compression (Fig. 5F). All monolayers that contained O-SPM in various amounts displayed fused liquid-condensed domains, with inclusions of liquid-expanded domains with variable size (Fig. 5G-J).
and when the cholesterol concentration was higher (Fig. 6A). The contribution of P-SPM to domain stability was disproportionate relative to its concentration. Therefore, P-SPM appeared to have a large stabilizing effect on the cholesterolrich domains even at relatively low concentrations (Fig. 6A). With mixed monolayers containing mono-unsaturated phospholipids, the highest domain stability was seen in pure POPC/cholesterol monolayers, whereas the stability was considerably lower in cholesterol/O-SPM monolayers (Fig. 6B). The contribution of POPC to domain stability was disproportionate relative to its concentration in the ternary monolayer.
4. Discussion
3.3. Domain boundary stability The stability of the macroscopic cholesterolrich domains was observed to depend on the lateral surface pressure of the monolayers, on the cholesterol concentration, and on the phosphatidylcholine/sphingomyelin ratio and degree of saturation (Fig. 6). With saturated phospholipids, the stability of the domain boundary line was higher when P-SPM dominated in the monolayer,
Many lines of evidence suggest that the molecular interaction between cholesterol and phosphatidylcholine on one hand, and cholesterol and sphingomyelin on the other, is different [6,7,1116], although some contradictory reports also exists [17,18]. In this study we have attempted to determine whether a selective interaction between cholesterol and sphingomyelin exist in mixed monolayers, in which both sphingomyelin and
78
P. Mattjus, J.P. Slotte / Chemistry and Physics oj Lipids 81 (1996) 69 80
phosphatidylcholine are simultaneously present together with cholesterol. We have used cholesterol oxidase to determine the oxidation susceptibility of monolayer cholesterol, since it is known that sphingomyelin is much more effective in preventing cholesterol oxidation compared to phosphatidylcholine [12]. The cholesterol oxidation experiments were performed at a physiologically relevant lateral surface pressure (i.e. 20 mN/m). However, the cholesterol/phospholipid tool ratio in the monolayer membranes had to be increased to an unphysiologically high value of 1.5, because the cholesterol oxidation rate in a sphingomyelin monolayer below 50 mol% sterol is undetectable within reasonable experimental time. Our finding that cholesterol was oxidized faster in mono-unsaturated phospholipid monolayers as compared to saturated phospholipid monolayers, is consistent with our previous reports on binary monolayers with cholesterol and one phospholipid class [12,13]. This finding suggests that cholesterol interacts more favorably with saturated phospholipids, since mono-unsaturated phospholipids are not as effective as saturated ones in preventing or retarding cholesterol oxidation. A similar effect of phospholipid unsaturation on cholesterol exchange kinetics has also been reported [7,8]. The novel observation at this point was the finding that in ternary mixed monolayers, with both phosphatidylcholine and sphingomyelin present, the effect of sphingomyelin on retarding cholesterol oxidation was more pronounced than would be expected based on its fractional presence in the monolayer. This important finding shows either that cholesterol interacted preferentially with sphingomyelin, or that sphingomyelin cooperatively influenced the interaction of cholesterol with phosphatidylcholine molecules, thus effecting the slower oxidation of cholesterol observed in the ternary monolayers. The interaction of cholesterol with phospholipids in monotayers is known to lead to the formation of cholesterol-rich domains, which may coexist with cholesterol-poor domains provided that the membrane cholesterol concentration is fairly low (5-35 mol%, this range is also influenced by the type of the phospholipid; [14,20]). These cholesterol-rich domains can be visualized
by monolayer fluorescence microscopy, because of the selective partitioning of a fluorescent probe into liquid-expanded phases [21-23]. The results of this study show that the properties of the cholesterol-rich domains where influenced not only by the phospholipid type and the degree of unsaturation, but also by the proportions of phosphatidylcholine and sphingomyelin present in the monolayers. The formation of cholesterol-rich liquid-condensed domains in binary monolayers containing cholesterol and either DPPC or PSPM [14], as well as in ternary monolayers with cholesterol and both PC and SPM (this study), was shown to increase with increasing concentrations of cholesterol (going from 20 to 33 mol%). Domain properties in the ternary mixed monolayers were also sensitive to the history of the monolayer. With no history of compression (i.e. during initial compression), major differences in liquid-condensed domain sizes were seen as a function of the phosphatidylcholine/sphingomyelin ratio. However, after a compression/expansion cycle, the cholesterol-rich domains in the 20 mol% monolayers were similar with a given phospholipid pair. It is believed that the domains formed after a compression/expansion cycle are closer to equilibrium shapes compared to the domains observed during initial compression [20,24,25]. Therefore, on the macroscopic level, equilibrium domains formed in monolayers with 20 mol% cholesterol were similar in any mixture of phosphatidylcholine and sphingomyelin. However, as the oxidation data indicated, it is likely that differences still existed on a microscopic level [26 28]. At higher cholesterol saturation (i.e. 33 mol%), especially with mono-unsaturated phospholipids, differences in the properties of cholesterol-rich domains were seen at different phosphatidylcholine/sphingomyelin ratios, even after a compression/expansion cycle. It appeared that the susceptibility of the cholesterol-rich domains to coalesce was increased with increasing sphingomyelin content in the ternary monolayers. This finding in ternary monolayers containing monounsaturated phospholipids is contrary to what was observed in binary monolayers containing cholesterol and saturated phospholipids (eithero-
P. Mattjus, J.P. Slotte / Chemistry and Physics of Lipids 81 (1996) 69 80
DPPC or P-SPM; [14]). It therefore appears that the susceptibility of the cholesterol-rich domains to coalesce to a great extent depends on the degree of saturation of the monolayer phospholipid. It has been reported that fusion of condensed domains is prevented or retarded by long-range dipole-induced repulsions [22,29]. Consequently, it has to be elucidated in further studies how these dipole-induced repulsions are affected by the molecular properties of the phospholipids (i.e. degree of unsaturation), and how cholesterol by associating with the phospholipid may change the overall dipole properties of the phospholipids. The stability of the domain boundary line of cholesterol-rich domains is known to be a function of the molecular properties of the phospholipids [20,30,31]. There is a critical point (i.e. a specific lateral surface pressure for a given monolayer composition) at which the circular shape of the liquid-condensed domains is destabilized, and the line boundary begin to fluctuate, finally leading to a complete dissipation of the line boundary as the lateral surface pressure increases (phase transformation pressure). The stability of the line boundary has been reported to be influenced by changes in the dipole density difference (between condensed and expanded domains) and the interfacial line tension [20,30]. Both line tension and dipole repulsion result from the molecular properties of the lipids involved in domain formation in the monolayers. The dipole density differences may arise from the net dipole of the ternary mixtures, due to the fact that sphingomyelin and phosphatidylcholine do not have the same dipole characteristic [9,10]. Thence, differences in domain stability should reflect differences in molecular interactions or differences in molecular properties resulting from intermolecular associations. The phase transformation pressures recorded in this study for the different cholesterol phospholipid ternary mixed monolayers were" visually observed, and represent the onset of the domain shape changes. The pressures reported for monolayers containing either DPPC or P-SPM with 20 or 33 tool% cholesterol are slightly lower than the corresponding values reported previously
79
(which were pressures giving complete dissipation of domain boundaries [14]). The stability of the cholesterol-rich domains observed in this study was clearly affected by the properties of the phospholipids. In saturated phospholipid monolayers, the cholesterol-rich domains were more stable when the ratio of P-SPM to DPPC was high, whereas the opposite was true for monounsaturated phospholipid monolayers. In addition, the domain stability was not a linear function of the ratio of sphingomyelin-to-phosphatidylcholine, which again may suggest that there existed selective interactions between any two of the three monolayer lipid components. In conclusion, our results obtained from experiments in which we used cholesterol oxidase as a probe, suggest that cholesterol did discriminate between sphingomyelin and phosphatidylcholine, even when both were present simultaneously. The general observation that sphingomyelin retards cholesterol desorption from membranes to a greater extent than phosphatidylcholine [7,8], and our findings that sphingomyelin prevents of retards enzyme-catalyzed oxidation of cholesterol better than phosphatidylcholine (this study, [13]), together imply that cholesterol interacts more favorably with sphingomyelin than with phosphatidylcholine. Our monolayer fluorescence microscopy results, although not directly comparable to our oxidation data, also suggest that the formation of cholesterol-rich domains in the mixed monolayers was influenced by the type of phospholipid present in the monolayer. It has been suggested that the more favorably interaction between cholesterol and sphingomyelin would arise from more favorable hydrophobic interactions between these two lipids [6,32 34]. Even though sphingomyelins have more functional groups which can participate in the formation of hydrogen bonds, it is unclear what addition hydrogen bonding in the polar region make to the molecular interactions between cholesterol and sphingomyelin [35-38].
Acknowledgements This study was supported by generous grants from the Academy of Finland, the Sigrid Juselius
80
P. Mattjus, J.P. Slotte / Chemistry and Physics of Lipids 81 (1996) 69 80
Foundation, the Magnus Ehrnrooth a n d the O s k a r O f l u n d F o u n d a t i o n .
Foundation,
References Ill P. Yeagle (1993) in: L.X. Finegold (Ed.), Cholesterol in Model Membranes, CRC Press, Boca Raton, FL, pp. 2-12. [2] P.L. Yeagle (1985) Biochim. Biophys. Acta 822, 267 287. [3] S. Patton (1970) J. Theor. Biol. 29, 489-491. [4] Y. Lange, M.H. Swaisgood and T.L. Steck. (1989) J. Biol. Chem. 264, 3786-3793. [5] J.M. Smaby, H.L. Brockman and R.E. Brown (1994) Biochemistry 33, 9135 9142. [6] S. Lund-Katz, H.M. Laboda, L,R. McLean and M.C. Phillips (1988) Biochemistry 27, 3416-3423. [7] R. Bittman (1993) in: L.X. Finegold (Ed.), Cholesterol in Model Membranes, CRC Press, Boca Raton, FL, pp. 45-65. [8] M.C. Phillips, W.H. Johnsson and G.H. Rothblat (1987) Biochim. Biophys. Acta 906, 223-276. [9] D.O. Shah and J.H. Schulman (1967) Lipids 2, 21 27. [10] J.M. Smaby and H.L. Brockman (1990) Biophys. J. 58, 195 204. [11] D. Needham and D.S. Nunn (1990) Biophys. J. 58, 997 1009. [12] J.P. Slotte (1992) Biochemistry 31, 5472-5477. [13] P. Mattjus and J.P. Slotte (1994) Chem. Phys. Lipids 71, 73 81. [14] J.P. Slotte (1995) Biochim. Biophys. Acta 1235, 419 427. [15] R.A. Demel, J.W.C.M. Jansen, P.W.M. van Dijck and LL.M. van Deenen (1977) Biochim. Biophys. Acta 465, 1-10. [16] P.W.M. van Dijck (1979) Biochim. Biophys. Acta 555, 89-101. [17] W.1. Calhoun and G.G. Shipley (1979) Biochemistry 18, 1717-1722.
[18] Y. Lange, J.S. D'Alessandro and D.M. Small (1979) Biochim. Biophys. Acta 556, 388-398. [19] G.R. Bartlett (1959) J. Biol. Chem. 234, 466-468. [20] S. Subramaniam and H.M. McConnell (1987) J. Phys. Chem. 91, 1715-1718. [21] R. Peters and K. Beck (1983) Proc. Natl. Acad. Sci. USA 80, 7183-7187. [22] H.M. McConnell, T.K. Tamm and R.M. Weis (1984) Proc. Natl. Acad. Sci. USA 81, 3249-3253. [23] K. Nag and K.M.W. Keough (1993) Biophys. J. 65, 1919 1926. [24] H. Yu and S.W. Hui (1992) Chem. Phys. Lipids 62, 69 78. [25] D.J. Keller, J.P. Korb and H.M. McConnell (1987) J. Phys. Chem. 91, 6417-6422. [26] J.L. Rubinstein, J.C. Owicki and H.M. McConnell, H.M. (1980) Biochemistry 19, 569 573. [27] F. Schroeder, J.R. Jefferson, A.B. Kier, J. Knittel, T.J. Scallen, W.G. Wood and I. Hapala (1991) Proc. Soc. Exp. Biol. Med. 196, 235 252. [28] M.B. Sankaram and T.E. Thompson (1991) Proc. Natl. Acad. S¢i. USA 88, 8686-8690. [29] M. L6sche, E. Sackmann and H. M6hwald (1983) Ber. Bunsenges. Chem. 87, 848-852. [30] P.A. Rice and H.M. McConnell (1989) Proc. Natl. Acad. Sci. USA 86, 6445 6448. [31] J.P. Slotte (1995) Biochim. Biophys. Acta 1238, 118-126. [32] L. Salem (1962) J. Phys. Chem. 37, 2100 2113. [33] R.A. Demel, K.R. Bruckdorfer and L.L.M. van Deenen (1972) Biochim. Biophys. Acta 255, 311 320. [34] M. Shinitzky and Y. Barenholz (1974) J. Biol. Chem. 249, 2652-2657. [35] R. Bittman, C.R. Kasireddy, P. Mattjus and J.P. Slotte (1994) Biochemistry 33, 11776-11781. [36] J.M. Boggs (1987) Biochim. Biophys. Acta 906, 353 404. [37] C.-C. Kan, Z.-s. Ruan and R. Bittman (1991) Biochemistry 30, 7759-7766. [38] L. Gr6nberg, Z.-s. Ruan, R. Bittman and J.P. Slotte (1991) Biochemistry 30, 10746 10754.
Chemistry and Physics of kipids 81 (1996) 69 80
CHEMISTRY AND PHYSICS OF LIPIDS ,
Does cholesterol discriminate between sphingomyelin and phosphatidylcholine in mixed monolayers containing both phospholipids? Peter Mattjus*, J. Peter Slotte Department qf Biochemistry and Pharmacy Abo Akademi University, FIN-20520 Turku, Finhmd
Received 24 November 1995: re~.ised 5 February 1996: accepted 7 February 1996
Abstract
The objective of this work was to examine the interaction of cholesterol with both phosphatidylcholines, l-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC) or 1,2-dipalmitoyl-sn-glycero-3-phosphocholine IDPPC), and sphingomyelins, N-oleoyl-D-sphingomyelin(O-SPM) or N-palmitoyl-o-sphingomyelin(P-SPM), in monolayers at an air/water interlace. We used cholesterol oxidase to probe for the relative strength of sterol-phospholipid interaction, and fluorescence microscopy to visualize lateral domain formation in the mixed monolayers. The ternary mixed monolayers, which contained cholesterol, POPC, and O-SPM had a twofold higher average oxidation rate than the corresponding system containing DPPC and P-SPM. This difference in oxidation rate between saturated and unsaturated systems was observed irrespective of the ratio between phosphatidylcholine and sphingomyelin in the monolayer. With either the saturated or the unsaturated systems, however, the rate of oxidation was influenced by the ratio of phosphatidylcholine to sphingomyelin. As the monolayer content of phosphatidylcholine increased and the sphingomyelin content decreased correspondingly (to maintain a constant cholesterol-to-phospholipid molar ratio), an increase in the average oxidation rate was seen in both saturated and mono-unsaturated monolayer systems. The relationship between the rate of cholesterol oxidation and the phosphatidylcholine/sphingomyelinratio was not linear, suggesting a preferential interaction of cholesterol with sphingomyelin even when phosphatidylcholine was present in the monolayer. The formation and stability of cholesterol-rich lateral (liquid-condensed) domains in the monolayers, as determined by monolayer fluorescence microscopy, was found to be highly influenced by the phospholipid class, the degree of acyl chain saturation, and by the ratio of phosphatidylcholine to sphingomyelin in the monolayer. The differences in cholesterol oxidation rates and lateral domain formation, as a function of the ratio of two phospholipids in the monolayers, apparently derived from differences in the hydrophobic interactions between the lipids. Keywords: Monolayers; Lipid domains; Cholesterol; Phosphatidylcholine; Sphingomyelin; Lipid interactions: Cholesterol oxidase; Fluorescence microscopy Abbreviations: TRITC-PE, tetramethylrhodamine isothiocyanate-phosphatidylethanolamine; POPC, l-palmitoyl-2-oleoyl-sn-glyccro-3-phosphocholine: DPPC, 1,2-dipalmitoyl-sn-glycero-3-phosphocholine:O-SPM, N-oleoyl-D-sphingomyelin:P-SPM, N-palmitoyl-D-sphingomyelin. * Corresponding author, Department of Biochemistry and Pharmacy, Abo Akademi University, P.O. Box 66, FIN-20521 Turku, Finland. Fax: + 358 21 2654 745.
11009-3084/96/ $15.00 ~ 1996 Elsevier Science lreland Ltd. All rights reserved PII S0009-3084(96)02535-2
70
P. Mattjus, J.P. Slotte / Chemistry and Physics ~/' LipMs 81 (1996) 69 80
I. Introduction
In biological as well as in model membranes, cholesterol interacts with phospholipids, thereby affecting their physicochemical behavior (for a review of the literature, see [1]). The interaction of cholesterol with phospholipids is affected by both the acyl chain composition of the phospholipids, and by the nature of their head-group [2]. The specific interaction of cholesterol with sphingomyelins and phosphatidylcholines is of biological importance, due to the high concentration of these lipids in the plasma membrane compartment of most cells [3,4]. The condensing effect of cholesterol on the lateral packing of phosphatidylcholine and sphingomyelin in monolayers have been shown to be similar [5], although the strength of interaction between cholesterol and phosphatidylcholine or sphingomyelin is known to differ [6]. This difference in interaction is most clearly seen when cholesterol desorption rates from phosphatidylcholine- or sphingomyelin-containing membranes are measured [7,8]. Sphingomyelin monolayers are reported to have a smaller surface potential compared to phosphatidylcholine monolayers, probably due to the lack of the carbonyl ester dipole in sphingomyelin [9,10]. The compressibility modulus has also been shown to be much higher in cholesterol-sphingomyelin monolayers compared to cholesterolphosphatidylcholine monolayers [11]. When cholesterol oxidase was used to probe for the strength of cholesterol interaction with phosphatidylcholines and sphingomyelins, is was shown that the oxidation susceptibility of cholesterol in sphingomyelin monolayers was reduced compared to the rate determined in phosphatidylcholine monolayers [12,13]. It was also shown that the presence of low or moderate amounts sphingomyelin in a ternary monolayer with cholesterol and phosphatidylcholine could influence the stoichiometry at which cholesterol became more resistant to oxidation by cholesterol oxidase [13]. Finally, a recent study of binary mixed monolayers containing cholesterol and either phosphatidylcholine or sphingomyelin, using monolayer fluorescence microscopy, showed a marked difference in the properties of the lateral
sterol-rich domains formed in phosphatidylcholine- and sphingomyelin-containing monolayers [14]. These and other results [15,16] suggest that cholesterol is more strongly associated with sphingomyelin than with phosphatidylcholine in monolayers and bilayers. However, an earlier study using the X-ray diffraction technique and differential scanning calorimetry failed to show a preferential interaction between cholesterol and either N-palmitoyl sphingomyelin or dipalmitoyl phosphatidylcholine, when all three lipid classes were present in model bilayer membranes [17]. A preferential interaction between cholesterol and either sphingomyelin or phosphatidylcholine was also not seen in equilibrium exchange experiments using erythrocyte ghost membranes and sonicated vesicles [181. In order to address the question of selective interaction between cholesterol and sphingomyelin or phosphatidylcholine, we have examined the interaction of cholesterol with these phospholipids in binary (cholesterol and one phospholipid) and ternary (cholesterol together with both phospholipids) mixed monolayers at the air/water interfaces. The association of cholesterol with the phospholipids was determined using cholesterol oxidase as a probe for the strength of interaction. We also observed the monolayers visually using monolayer fluorescence microscopy to gain possible information about the formation of lateral cholesterol-rich domains and the effects of phospholipid type and properties on domain formation.
2. Material and methods
2.1. Materials
Cholesterol, DPPC, POPC, O-SPM and PSPM were obtained from Sigma Chemicals (St. Louis, MO). The cholesterol was at least 99% pure. The phospholipids gave all a single spot when analyzed by thin layer chromatography. Tetra-methylrhodamine isothiocyanate phosphatidylethanol amine (TRITC-PE) was purchased from Molecular Probes Inc. (Eugene, OR) and was used as delivered. It was stored in
P, Mat(]us, J.P. Slotte / Chemistry and Ph)sics 0[ LipMs 81 (1996) 69 80
aliquots in the dark, at - 2 5 ° C , under an argon atmosphere. Stock solutions of the phospholipids were prepared in hexane/2-propanol (3:2 v/v), and stored in the dark at - 2 5 ° C . The solutions were warmed to room temperature before use. The concentration of the various phospholipid stock solutions was determined by the method described by Bartlett [19]. Cholesterol oxidase (3fl-hydroxysteroid oxidase, EC 1.1.3.6, Streptomyces cinnamoneus) was purchased from Calbiochem (San Diego, CA) and was dissolved in a Tris-buffer containing 50 mM Tris-HC1 (pH 7.5) and 140 mM NaCI. The enzyme was dispensed into 160/~1 aliquots, stored at - 2 5 ° C , and used within 2 h after thawing to 0°C. Buffer salts were of pro analysis grade and the water was purified with reverse osmosis, and further passed through a Milli-Q U F + filtering system to a resistivity of 18.2 M~2/cm.
2.2. Oxidation of cholesterol in ternary mixed monolayers The oxidation of monolayer cholesterol was performed in a zero-order-type teflon trough using a KSV 3000 surface barostat (KSV Instruments Ltd., Helsinki, Finland) as described previously [12]. The monolayers were prepared to a molar ratio of 1.5:1 of cholesterol to phospholipid, and the phospholipids used were either DPPC and P-SPM or POPC and O-SPM. The ratio of the two phospholipids was varied in increments of 10%, while the cholesterol-to-phospholipid ratio was constant. The subphase (Trisbuffer) in the reaction compartment (28.3 cm 2, 30 ml volume) was magnetically stirred (100 revs./ min), and thermostated to 30°C. The lipid solution (containing cholesterol and the phospholipids) was prepared to include 300 nmol lipids in a 1.8 ml glass vial, and was mixed immediately prior to spreading on the air/water interface using a Hamilton syringe. After the monolayer had been compressed to a lateral surface pressure of 20 raN/m, and allowed to stabilize (about 5 rain from the spreading), cholesterol oxidase was added to the reaction compartment (in 40 /tl, to a final concentration of 23 mU/ml). Constant surface pressure was maintained by
71
computer-controlled compensatory barrier movement throughout the experiment. Three different kinetic measurements were made with each film composition. Calculation of the average oxidation rate was done as described previously 1112].
2,3. Monolayer preparation and Hsualization with epifluorescence microscope The lipid solution (300 nmol total lipid, with 0,5 mol% TRITC-PE) was prepared in a glass vial immediately prior to spreading. The monolayer was allowed to stabilized for about 3 min, at an available mean molecular area of 133.5/k 2, before compression was started. The monolayers were subjected to symmetric compression at a speed not exceeding 4 A2/molecule/min, at ambient temperature (about 20°C). The surface texture of the monolayers were documented at a surface pressure of 0.2 mN/m both before and after a compression/expansion cycle using a KSV Mini systems apparatus (KSV Instruments Ltd., Helsinki) mounted on an Olympus IMT-2 fluorescence microscope, as described previously [14]. During the compression/expansion cycle, the monolayers were compressed to 35 raN/m, at which pressure the monolayers were allowed to stabilize for 1 min, before relaxing back to a low surface pressure (about 0.2 mN,m). At least two runs were performed with each lipid composition, and one representative micrograph was chosen from a number of similar images.
3. Results
3,1. Oxidation of cholesterol in mixed monolayers The relative strength of the interaction between cholesterol and different phospholipids in mixed monolayers at the air/water interface was assayed using cholesterol oxidase as a probe. The susceptibility to oxidation of cholesterol has been shown to depend on the ability of neighboring phospholipid to interact with and 'shield' the cholesterol molecule from being exposed to the enzyme at the water/lipid interface [12,13]. The various monolayer systems which were examined were prepared
P. MatOus, J.P. Slotte / Chemistry and Ph.vsics of Lipids 81 (1996) 69-80
72
to have a cholesterol to phospholipid mol ratio of 1.5:1. Although this ratio is unphysiologically high, it was chosen because otherwise a comparison between phosphatidylcholines and sphingomyelins would not have been possible. This is because cholesterol oxidation in sphingomyelinmembranes below 50 mol% cholesterol is too slow to be measurable. The average oxidation rates of cholesterol present in membranes comprising saturated or mono-unsaturated phospholipids are shown in Fig. 1. The monolayers containing mono-unsaturated phospholipids (varying proportions of POPC and O-SPM) had on average a twofold higher oxidation rate as compared with monolayers containing saturated phospholipids (DPPC and P-SPM; Fig. 1). As the proportion of phosphatidylcholines increased, and the concentration of sphingomyelins decreased correspondingly, an increase in the average oxidation rate
%
O0 90
gO
70
U'hbJqorfs} e l i n
60
50
zc
5C
20
1©
0
30 40 .50 60 70 % PhosphcJtidyl¢ holinc
80
90
I CO
13 ~4 ""
i.2
-
1 i ©
b8 +S
O7 L~ b 05
3 o4 ,', 5 ~:,' <:
02 0 i
;
q
10
/C'
Fig. 1. Average oxidation rate versus the variation in the phospholipid content. Mixed monolayers were prepared to have a cholesterol-to-phospholipid tool ratio of 1.5, and the ratio of phosphatidylcholine to sphingomyelin was varied in increments of 10%. The monolayers where exposed to 23 m U / m l cholesterol oxidase at 30°C at a surface pressure of 20 m N / m . The filled circles represent cholesterol with POPC and O-SPM, and the open circles cholesterol with DPPC and P-SPM. The dotted line represents the oxidation rate for ideal mixtures of cholesterol and phospholipids. Values are averages _+ S.D. from three experiments at each phospholipid increment.
was seen for both systems. The observable oxidation rate was not, however, linearly proportional to the fractional phospholipid composition, since a negative deviation from an ideal situation was observed (Fig. 1). It appeared that in both phospholipid systems (i.e., saturated and mono-unsaturated monolayers), the retarding effect of sphingomyelin on the oxidation rate was greater than would be expected based on ideal mixing of cholesterol with either phospholipid type. Therefore, these novel results suggest that cholesterol in these monolayer membranes interacted preferentially with sphingomyelin even when both phospholipids were in the same membrane. 3.2. SurJace texture of mixed monolayers To study the ability of cholesterol to form lateral domains with phosphatidylcholines and sphingomyelins, four different ternary mixed monolayer systems were examined visually with monolayer fluorescence microscopy. The monolayer concentration of cholesterol was either 20 or 33 mol%, with the phospholipids being either saturated (DPPC and P-SPM) or mono-unsaturated (POPC and O-SPM). All micrographs were taken at a surface pressure of about 0.2 mN/m, both during initial compression and after a compression/expansion cycle. The fluorescent probe, TRITC-PE, was included in each monolayer at 0.5 mol%. This probe, like other fluorescent lipid probes, partition preferentially into expanded domains, giving these a bright fluorescence, while being effectively excluded from condensed domains. 3.2.1. 20 mol% cholesterol During initial compression of a monolayer with 20 tool% cholesterol in DPPC only, cholesterolrich laterally condensed domains were formed. These were circular in shape (i.e. liquid) but unequal in their size distribution (Fig. 2A). The addition of P-SPM to the monolayer (still keeping the cholesterol mol% at 20) led to the formation of much larger cholesterol-rich domains during initial compression (Fig. 2B depicts the monolayer with 70% DPPC and 30% P-SPM). Only when the monolayer phospholipid composition approached
P. Mattjus, J.P. SIotte / Chemistry and Physics (~f Lipids 81 (1996) 69 80
73
A
Fig. 2. Domain formation in mixed saturated monolayers. The mixed monolayers contained 20 tool% cholesterol and 80 tool%, phospholipids (DPPC and P-SPM), together with 0.5 mol% TRITC-PE. The micrographs were documented at a surface pressure of 0.2 mN/m. The top row shows monolayers during initial compression, whereas the bottom row depicts mc~nolayers after a compression/expansion cycle. Panels A and D are with 100% DPPC and no sphingomyelin, panels B and E are with 70% DPPC and 30% P-SPM, while panels C and F are 100% P-SPM and no phosphatidyleholine. The scale bar represents 100 pm.
pure P-SPM did the cholesterol-rich domains again decrease in diameter (Fig. 2C). However, if the monolayers were compressed and expanded, the domain properties were remarkably similar, irrespective of the phospholipid composition (Fig. 2D-E). For the rnono-unsaturated phospholipid monolayers at 20 mol% cholesterol (during initial compression), the cholesterol-rich domains were smallest for O-SPM (Fig. 3D), whereas they were
larger with POPC present (Fig. 3A-C). The domains became very small in all mixed monolayers after a compression-expansion cycle(Fig. 3E H). 3.2.2. 33 mol % cholesterol
Increasing the cholesterol concentration to 33 mol% resulted in the formation of partly coalesced, cholesterol-rich domains during initial compression. With DPPC and cholesterol, larger liquid-condensed domains against a liquid-ex-
B
Fig. 3. Domain formation in mixed unsaturated monolayers. The mixed monolayers contained 20 mol% cholesterol and 80 mol% phospholipids (POPC and O-SPM), together with 0.5 mol% TRITC-PE. The micrographs were documented at a surface pressure of 0.2 mN/m. The top row shows monolayers during initial compression, whereas the bottom row depicts monolayers after a compression/expansion cycle. Panels A and E are with 100% POPC and no sphingomyelin, panels B and F are with 90% POPC and 10% O-SPM, panels C and G are with 10% POPC and 90% O-SPM, while panels D and H are 100% O-SPM and no phosphatidylcholine. The scale bar represents 100 /~m.
A
t~
.5
e~
5~
P. Mat(/us, J.P. Slotte / Chemistry and Physics ()f Lipids 81 (1996) 69 8 0
A
B
75
C
J Fig. 4. Domain formation in mixed saturated monolayers. The mixed monolayers contained 33 mol% cholesterol and 67 tool% phospholipids (DPPC and P-SPM), together with 0.5 mol% TRITC-PE. The micrographs were documented at a surface pressure of 0.2 mN/m. The top row shows monolayers during initial compression, whereas the bottom row depicts mon01ayers after a compression/expansion cycle. Panels A and D are with 100% DPPC and no sphingomyelin, panels B and E are with 90% DPPC and 10% P-SPM, while panels C and F are 100% P-SPM and no phosphatidylcholine. The scale bar represents 100/~m.
panded phase were observed (Fig. 4A). Adding P-SPM to the monolayer (with a corresponding decrease in DPPC concentration) led to smaller extent of domain fusion (Fig. 4B depicts a monolayer with 90% DPPC and 10% P-SPM). Cholesterol in a P-SPM monolayer (at 33 mol%) also had only a limited extent of fusion of cholesterol-rich domains (Fig. 4C). After a compression/expansion cycle, all monolayers examined displayed completely coalesced condensed domains (Fig. 4D E).
In monolayers with 33 tool% cholesterol and mono-unsaturated phospholipids, the lateral domains formed during initial compression were markedly different, going from pure POPC to pure O-SPM as the phospholipid component. Whereas small circular domains were seen in cholesterol/ POPC monolayers, no clear domains could be resolved in cholesterol/O-SPM monolayers (Fig. 5A and E). When POPC was added to the cholesterol/O-SPM mixed monolayer, condensed do-
R
#-,
iJ
E
Fig. 5. Domain formation in mixed unsaturated monolayers. The mixed monolayers contained 33 mol% cholesterol and 67 mol% phospholipids (POPC and O-SPM), together with 0.5 mol% TRITC-PE. The micrographs were documented at a surface pressure of 0.2 mN/m. The top row shows monolayers during initial compression, whereas the bottom row depicts monolayers after a compression/expansion cycle. Panels A and F are with 100% POPC and no sphingomyelin, panels B and G are with 90% POPC and 10% O-SPM, panels C and H are with 50% POPC and 50% O-SPM, panels D and I are 10% POPC and 90% O-SPM, while panels E and J are with 100% O-SPM and no phosphatidylcholine. The scale bar represents 100 #m.
A A~
I
Y.,"
C~
P. Matt~us J.P. Slotte / Chemistry and Physics ol l.grids 81 (1996) 69 80
% P E
I (3C1 '~} 3
30
"-.
i
77] ~ -
~ [) v-
~:o,
r
:;
l:? I
121
2C
r
I
-
.....
1C
.i
•
-
.'f ]
E
c
'
)
t:"
:,Jrl (-:
--
:
::11, '
~
)',
!1)
T
-
[
.
,
Jl
,!~
i
!
,].~
,1
"
,:4
]
s ~:,
~'
"
)[:
"-4
:, i'~
[,PV I~ 121:'
77
2 ,::
.
I
-
!:
]
; i
IC' t
0
!O
i__
0
5~
i 40
~,()
~.
&
~
ii/3
7('
~(]'
L i{)
. "
:::
!.
.
i/'. r:
. I "
.
. 2
t :3
:0
1C
,°4 DF'FC
;'
!
_ ?
_ )
L
qr
:C)l
Fig. 6. Stability of cholesterol-rich domains. The phase transformation pressure was determined visually during compression of the monolayers, and panel A shows transformation pressures for saturated monolayers whereas panel B gives results with mono-unsaturated monolayers. The dotted lines represent monolayers with 20 tool% cholesterol, the solid lines monolayers with 33 mol% cholesterol. Values are averages + S.D. from two monolayers.
mains started to form, and these had also partially coalesced (Fig. 5B-D). The elongated form of the domains seen in (Fig. 5D; 10% POPC and 90% O-SPM; 33 mol% cholesterol) was at least in part due to excitation light, since photochemical effects are known to disrupt domain boundary stability close to the phase transformation pressure [20,21]. After the monolayers were subjected to a compression/expansion cycle, the condensed domains in the cholesterol/POPC monolayer became slightly smaller than they were during initial compression (Fig. 5F). All monolayers that contained O-SPM in various amounts displayed fused liquid-condensed domains, with inclusions of liquid-expanded domains with variable size (Fig. 5G-J).
and when the cholesterol concentration was higher (Fig. 6A). The contribution of P-SPM to domain stability was disproportionate relative to its concentration. Therefore, P-SPM appeared to have a large stabilizing effect on the cholesterolrich domains even at relatively low concentrations (Fig. 6A). With mixed monolayers containing mono-unsaturated phospholipids, the highest domain stability was seen in pure POPC/cholesterol monolayers, whereas the stability was considerably lower in cholesterol/O-SPM monolayers (Fig. 6B). The contribution of POPC to domain stability was disproportionate relative to its concentration in the ternary monolayer.
4. Discussion
3.3. Domain boundary stability The stability of the macroscopic cholesterolrich domains was observed to depend on the lateral surface pressure of the monolayers, on the cholesterol concentration, and on the phosphatidylcholine/sphingomyelin ratio and degree of saturation (Fig. 6). With saturated phospholipids, the stability of the domain boundary line was higher when P-SPM dominated in the monolayer,
Many lines of evidence suggest that the molecular interaction between cholesterol and phosphatidylcholine on one hand, and cholesterol and sphingomyelin on the other, is different [6,7,1116], although some contradictory reports also exists [17,18]. In this study we have attempted to determine whether a selective interaction between cholesterol and sphingomyelin exist in mixed monolayers, in which both sphingomyelin and
78
P. Mattjus, J.P. Slotte / Chemistry and Physics oj Lipids 81 (1996) 69 80
phosphatidylcholine are simultaneously present together with cholesterol. We have used cholesterol oxidase to determine the oxidation susceptibility of monolayer cholesterol, since it is known that sphingomyelin is much more effective in preventing cholesterol oxidation compared to phosphatidylcholine [12]. The cholesterol oxidation experiments were performed at a physiologically relevant lateral surface pressure (i.e. 20 mN/m). However, the cholesterol/phospholipid tool ratio in the monolayer membranes had to be increased to an unphysiologically high value of 1.5, because the cholesterol oxidation rate in a sphingomyelin monolayer below 50 mol% sterol is undetectable within reasonable experimental time. Our finding that cholesterol was oxidized faster in mono-unsaturated phospholipid monolayers as compared to saturated phospholipid monolayers, is consistent with our previous reports on binary monolayers with cholesterol and one phospholipid class [12,13]. This finding suggests that cholesterol interacts more favorably with saturated phospholipids, since mono-unsaturated phospholipids are not as effective as saturated ones in preventing or retarding cholesterol oxidation. A similar effect of phospholipid unsaturation on cholesterol exchange kinetics has also been reported [7,8]. The novel observation at this point was the finding that in ternary mixed monolayers, with both phosphatidylcholine and sphingomyelin present, the effect of sphingomyelin on retarding cholesterol oxidation was more pronounced than would be expected based on its fractional presence in the monolayer. This important finding shows either that cholesterol interacted preferentially with sphingomyelin, or that sphingomyelin cooperatively influenced the interaction of cholesterol with phosphatidylcholine molecules, thus effecting the slower oxidation of cholesterol observed in the ternary monolayers. The interaction of cholesterol with phospholipids in monotayers is known to lead to the formation of cholesterol-rich domains, which may coexist with cholesterol-poor domains provided that the membrane cholesterol concentration is fairly low (5-35 mol%, this range is also influenced by the type of the phospholipid; [14,20]). These cholesterol-rich domains can be visualized
by monolayer fluorescence microscopy, because of the selective partitioning of a fluorescent probe into liquid-expanded phases [21-23]. The results of this study show that the properties of the cholesterol-rich domains where influenced not only by the phospholipid type and the degree of unsaturation, but also by the proportions of phosphatidylcholine and sphingomyelin present in the monolayers. The formation of cholesterol-rich liquid-condensed domains in binary monolayers containing cholesterol and either DPPC or PSPM [14], as well as in ternary monolayers with cholesterol and both PC and SPM (this study), was shown to increase with increasing concentrations of cholesterol (going from 20 to 33 mol%). Domain properties in the ternary mixed monolayers were also sensitive to the history of the monolayer. With no history of compression (i.e. during initial compression), major differences in liquid-condensed domain sizes were seen as a function of the phosphatidylcholine/sphingomyelin ratio. However, after a compression/expansion cycle, the cholesterol-rich domains in the 20 mol% monolayers were similar with a given phospholipid pair. It is believed that the domains formed after a compression/expansion cycle are closer to equilibrium shapes compared to the domains observed during initial compression [20,24,25]. Therefore, on the macroscopic level, equilibrium domains formed in monolayers with 20 mol% cholesterol were similar in any mixture of phosphatidylcholine and sphingomyelin. However, as the oxidation data indicated, it is likely that differences still existed on a microscopic level [26 28]. At higher cholesterol saturation (i.e. 33 mol%), especially with mono-unsaturated phospholipids, differences in the properties of cholesterol-rich domains were seen at different phosphatidylcholine/sphingomyelin ratios, even after a compression/expansion cycle. It appeared that the susceptibility of the cholesterol-rich domains to coalesce was increased with increasing sphingomyelin content in the ternary monolayers. This finding in ternary monolayers containing monounsaturated phospholipids is contrary to what was observed in binary monolayers containing cholesterol and saturated phospholipids (eithero-
P. Mattjus, J.P. Slotte / Chemistry and Physics of Lipids 81 (1996) 69 80
DPPC or P-SPM; [14]). It therefore appears that the susceptibility of the cholesterol-rich domains to coalesce to a great extent depends on the degree of saturation of the monolayer phospholipid. It has been reported that fusion of condensed domains is prevented or retarded by long-range dipole-induced repulsions [22,29]. Consequently, it has to be elucidated in further studies how these dipole-induced repulsions are affected by the molecular properties of the phospholipids (i.e. degree of unsaturation), and how cholesterol by associating with the phospholipid may change the overall dipole properties of the phospholipids. The stability of the domain boundary line of cholesterol-rich domains is known to be a function of the molecular properties of the phospholipids [20,30,31]. There is a critical point (i.e. a specific lateral surface pressure for a given monolayer composition) at which the circular shape of the liquid-condensed domains is destabilized, and the line boundary begin to fluctuate, finally leading to a complete dissipation of the line boundary as the lateral surface pressure increases (phase transformation pressure). The stability of the line boundary has been reported to be influenced by changes in the dipole density difference (between condensed and expanded domains) and the interfacial line tension [20,30]. Both line tension and dipole repulsion result from the molecular properties of the lipids involved in domain formation in the monolayers. The dipole density differences may arise from the net dipole of the ternary mixtures, due to the fact that sphingomyelin and phosphatidylcholine do not have the same dipole characteristic [9,10]. Thence, differences in domain stability should reflect differences in molecular interactions or differences in molecular properties resulting from intermolecular associations. The phase transformation pressures recorded in this study for the different cholesterol phospholipid ternary mixed monolayers were" visually observed, and represent the onset of the domain shape changes. The pressures reported for monolayers containing either DPPC or P-SPM with 20 or 33 tool% cholesterol are slightly lower than the corresponding values reported previously
79
(which were pressures giving complete dissipation of domain boundaries [14]). The stability of the cholesterol-rich domains observed in this study was clearly affected by the properties of the phospholipids. In saturated phospholipid monolayers, the cholesterol-rich domains were more stable when the ratio of P-SPM to DPPC was high, whereas the opposite was true for monounsaturated phospholipid monolayers. In addition, the domain stability was not a linear function of the ratio of sphingomyelin-to-phosphatidylcholine, which again may suggest that there existed selective interactions between any two of the three monolayer lipid components. In conclusion, our results obtained from experiments in which we used cholesterol oxidase as a probe, suggest that cholesterol did discriminate between sphingomyelin and phosphatidylcholine, even when both were present simultaneously. The general observation that sphingomyelin retards cholesterol desorption from membranes to a greater extent than phosphatidylcholine [7,8], and our findings that sphingomyelin prevents of retards enzyme-catalyzed oxidation of cholesterol better than phosphatidylcholine (this study, [13]), together imply that cholesterol interacts more favorably with sphingomyelin than with phosphatidylcholine. Our monolayer fluorescence microscopy results, although not directly comparable to our oxidation data, also suggest that the formation of cholesterol-rich domains in the mixed monolayers was influenced by the type of phospholipid present in the monolayer. It has been suggested that the more favorably interaction between cholesterol and sphingomyelin would arise from more favorable hydrophobic interactions between these two lipids [6,32 34]. Even though sphingomyelins have more functional groups which can participate in the formation of hydrogen bonds, it is unclear what addition hydrogen bonding in the polar region make to the molecular interactions between cholesterol and sphingomyelin [35-38].
Acknowledgements This study was supported by generous grants from the Academy of Finland, the Sigrid Juselius
80
P. Mattjus, J.P. Slotte / Chemistry and Physics of Lipids 81 (1996) 69 80
Foundation, the Magnus Ehrnrooth a n d the O s k a r O f l u n d F o u n d a t i o n .
Foundation,
References Ill P. Yeagle (1993) in: L.X. Finegold (Ed.), Cholesterol in Model Membranes, CRC Press, Boca Raton, FL, pp. 2-12. [2] P.L. Yeagle (1985) Biochim. Biophys. Acta 822, 267 287. [3] S. Patton (1970) J. Theor. Biol. 29, 489-491. [4] Y. Lange, M.H. Swaisgood and T.L. Steck. (1989) J. Biol. Chem. 264, 3786-3793. [5] J.M. Smaby, H.L. Brockman and R.E. Brown (1994) Biochemistry 33, 9135 9142. [6] S. Lund-Katz, H.M. Laboda, L,R. McLean and M.C. Phillips (1988) Biochemistry 27, 3416-3423. [7] R. Bittman (1993) in: L.X. Finegold (Ed.), Cholesterol in Model Membranes, CRC Press, Boca Raton, FL, pp. 45-65. [8] M.C. Phillips, W.H. Johnsson and G.H. Rothblat (1987) Biochim. Biophys. Acta 906, 223-276. [9] D.O. Shah and J.H. Schulman (1967) Lipids 2, 21 27. [10] J.M. Smaby and H.L. Brockman (1990) Biophys. J. 58, 195 204. [11] D. Needham and D.S. Nunn (1990) Biophys. J. 58, 997 1009. [12] J.P. Slotte (1992) Biochemistry 31, 5472-5477. [13] P. Mattjus and J.P. Slotte (1994) Chem. Phys. Lipids 71, 73 81. [14] J.P. Slotte (1995) Biochim. Biophys. Acta 1235, 419 427. [15] R.A. Demel, J.W.C.M. Jansen, P.W.M. van Dijck and LL.M. van Deenen (1977) Biochim. Biophys. Acta 465, 1-10. [16] P.W.M. van Dijck (1979) Biochim. Biophys. Acta 555, 89-101. [17] W.1. Calhoun and G.G. Shipley (1979) Biochemistry 18, 1717-1722.
[18] Y. Lange, J.S. D'Alessandro and D.M. Small (1979) Biochim. Biophys. Acta 556, 388-398. [19] G.R. Bartlett (1959) J. Biol. Chem. 234, 466-468. [20] S. Subramaniam and H.M. McConnell (1987) J. Phys. Chem. 91, 1715-1718. [21] R. Peters and K. Beck (1983) Proc. Natl. Acad. Sci. USA 80, 7183-7187. [22] H.M. McConnell, T.K. Tamm and R.M. Weis (1984) Proc. Natl. Acad. Sci. USA 81, 3249-3253. [23] K. Nag and K.M.W. Keough (1993) Biophys. J. 65, 1919 1926. [24] H. Yu and S.W. Hui (1992) Chem. Phys. Lipids 62, 69 78. [25] D.J. Keller, J.P. Korb and H.M. McConnell (1987) J. Phys. Chem. 91, 6417-6422. [26] J.L. Rubinstein, J.C. Owicki and H.M. McConnell, H.M. (1980) Biochemistry 19, 569 573. [27] F. Schroeder, J.R. Jefferson, A.B. Kier, J. Knittel, T.J. Scallen, W.G. Wood and I. Hapala (1991) Proc. Soc. Exp. Biol. Med. 196, 235 252. [28] M.B. Sankaram and T.E. Thompson (1991) Proc. Natl. Acad. S¢i. USA 88, 8686-8690. [29] M. L6sche, E. Sackmann and H. M6hwald (1983) Ber. Bunsenges. Chem. 87, 848-852. [30] P.A. Rice and H.M. McConnell (1989) Proc. Natl. Acad. Sci. USA 86, 6445 6448. [31] J.P. Slotte (1995) Biochim. Biophys. Acta 1238, 118-126. [32] L. Salem (1962) J. Phys. Chem. 37, 2100 2113. [33] R.A. Demel, K.R. Bruckdorfer and L.L.M. van Deenen (1972) Biochim. Biophys. Acta 255, 311 320. [34] M. Shinitzky and Y. Barenholz (1974) J. Biol. Chem. 249, 2652-2657. [35] R. Bittman, C.R. Kasireddy, P. Mattjus and J.P. Slotte (1994) Biochemistry 33, 11776-11781. [36] J.M. Boggs (1987) Biochim. Biophys. Acta 906, 353 404. [37] C.-C. Kan, Z.-s. Ruan and R. Bittman (1991) Biochemistry 30, 7759-7766. [38] L. Gr6nberg, Z.-s. Ruan, R. Bittman and J.P. Slotte (1991) Biochemistry 30, 10746 10754.