Chemistry and Physics of Lipids, 30 ( t 982) 189-227 189 Elsevier/North-Holland Scientific Publishers Ltd. L I P I D - P R O T E I N I N T E R A C T I O N S IN M O N O L A Y E R S R. VERGER and F. PATTUS Centre de Biochimie et de Biologie Moldculaire du CNRS, 31, Chemin Joseph-Aiguier, 13277 Marseille Cedex 09 (France) We restricted ourselves to a few examples of the different methodological aspects of the investigation of lipid-protein interactions in monotayer assemblies. Experiments with monolayers have the unique advantage that the arrangement and packing of the molecules can be easily measured and controlled. The first section is devoted to proteins which do not degrade lipids. Soluble proteins are generally injected in the subphase below a preformed lipid monolayer and measurements are performed either at constant surface area or at constant surface pressure. These experiments can give information on the penetration capacity into the interface, its lipid specificity with a direct access to the area of the protein segment interacting with lipids. Reconstitution of functional enzymatic complexes can be achieved as well as the determination of the orientation of the protein at the interface. Most intrinsic membrane proteins are insoluble in water. In the absence of detergent they aggregate and display no affinity for lipid interfaces. These proteins can be spread from an organic solvent solution but with the risk of being denatured. In order to circumvent this difficulty a method for spreading an aqueous suspension of lipoproteins or natural membrane vesicles was developed. This spreading method allows the formation of lipoprotein films retaining biological activities and native membrane constituents. Formation of functional transmembrane complexes in planar bilayers from such lipoprotein films is the most fascinating application of this spreading technique. In the second section, we reviewed the use of pure or mixed lipid monolayers as substrates for lipolytic enzymes. Either long-chain lipids were used and their surface density was not controlled; or short-chain lipids were applied, again without control of surface density; or short-chain lipids were used at constant surface density. Recently a method was developed to study the hydrolysis of long chain lipids with control of surface density. The monolayer technique allows us to study accurately the influence of surface pressure and protein cofactors on the hydrolysis velocity and lag time in lipolysis. Two types of mixed lipid monolayers can be formed at the air/water interface: either by spreading, from a volatile organic solvent, a mixture of water insoluble lipids or by injecting a detergent solution into the water subphase covered with a preformed pure lipid monolayer. These techniques are ideally suited for the study of the mode of action of lipolytic enzymes on controlled mixed interfaces. Keywords.. lipid monolayers; protein interaction; enzyme kinetics; constant area; constant pressure; methodology. Introduction A large n u m b e r o f biological reactions take place at interfaces. These interfaces 0009-3084/82/0000-0000/$02.75 © 1982 Elsevier/North-Holland Scientific Publishers Ltd. 190 R. Verger and F. Pattus, Lipid-protein interactions in monolayers made of lipid and proteins are not only of great importance in the functioning of natural membranes but are also involved in a wide variety of biological processes such as intestinal fat digestion, blood clotting or lipoprotein metabolism. In order to understand the function of such complicated processes model systems have been developed. There are three main types of model system: liposomes or lipid vesicles [1]; planar bilayer membranes [2] ; monomolecular films at gas-liquid or liquid-liquid interfaces. The monolayer technique has been extensively used to investigate the interfacial behavior of phospholipids and to investigate the lipid-lipid interactions at the interface [3,4]. Experiments with monolayers have the advantage over the other two model systems that the arrangement of the molecules can be controlled easily by changing the molecular area and the surface pressure of the monolayer. The data obtained by variation of these parameters can give unique information about lipid-lipid and lipid-subphase interactions. Proteins can also be spread at interfaces. Unfortunately, most experiments were carried in absence of lipids [5,6] or at low surface pressure [7,8] or with soluble proteins which do not interact with lipid in vivo. Although experiments carried out at the air-water interface give information about flexibility and tensio-activity of proteins, unfolding and denaturation of proteins occurs at the air-water interface and at low surface pressure [14,17]. In order to extrapolate from model systems to biological structures one must take care that proteins at the interface maintain their 'native' structure. In nature, membrane proteins are embedded in a lipid matrix. Membrane proteins are first inserted in preformed membranes during biosynthesis [9] or interact after complete biosynthesis with lipoprotein water interfaces [ 10]. In monomolecular films, the interracial free energy and lipid packing of molecules can be varied, Extrapolation of monolayer data to biological structures needs an evaluation or a correspondence between the interracial free energy of natural structures (membranes, lipoproteins, emulsions) and the interracial free energy in monolayers. This problem has always been a subject of debate. In a recent meeting [11] Phillips answered this difficult question in a way very close to our own opinion, which is now supported by increasing experimental evidence: 'How do we make a comparison between the model and the real system? There is no clear answer to this question. The best way, I think, to make the comparisons is to consider molecular areas and I like to think in terms of the area for phospholipid molecules.' Although interracial free energy of natural structures cannot be determined, an 'equivalent surface pressure' can be defined. Recently, Blume [12] re-investigated this problem with monolayers and bilayers made of saturated phospholipids. This author concluded that 'the behavior of the bilayer system is very similar to that of the respective monolayer system at a lateral pressure of 30 dyn/cm because at this pressure the absolute area and the area change in both systems are the same'. Demel et al. [ 13] comparing the action of phospholipases on erythrocyte R. Verger and F. Pattus, Lipid-protein interactions in monolayers 191 membranes and on lipid monolayers found a similar equivalent surface pressure. This 'equivalent surface pressure' must be defined for every biological interface. Any fluctuation in the structure and composition of the lipid-water interface may change the lateral forces. The purpose of this review is not to cover the whole literature on lipid-protein interactions in monolayers. The reader is referred to different reviews [8,14-19] for earlier work and detailed information. We restrict ourselves to the different methodological aspects of the investigation of lipid-protein interactions in monolayer assemblies. The interaction of lipolytic enzymes with lipid monolayers is considered in a separate section since lipids are the substrates for these enzymes. Experimental For detailed information the reader is referred to the excellent book by Gaines [19] on experimental methods and properties of monolayers. The most commonly and easily measured property of a monolayer is its surface pressure Or). Surface pressure is defined as the decrease in interfacial tension (or surface free energy) produced by the monolayer, 7r = 3 ' 0 - 3'm; where 3'o and 3'm are the interfacial tension in the absence and presence of the monolayer, respectively. Quantitative measurements of film pressure are made by two types of device: the Wilhelmy and the Langmuir 'film balance', respectively [19]. The Wilhelmy-type balance is easier to handle. There are, however, certain disadvantages (see Ref. 18). Different types of trough and automatic apparatus have been designed [ 19-22]. They are composed of a trough, linear or circular [23,24] with one or several compartments. A mobile barrier can compress or expand the film. A regulatory unit drives the barrier in order to maintain surface pressure constant [21,22]. Surface potential (AV) of monomolecular films is also commonly measured. AV is the change of Volta potential which occurs when a monolayer is spread at the air-water interface. However, interpretation of the data are somewhat difficult (see Ref. 18). In order to measure surface concentration of proteins in surface films, radioactivity is quite useful. For certain isotopes (131I, 14C, 3Ss, 32p, 3H) surface radioactivity can be directly measured with a thin end-window Geiger-MOller counter placed just above the film [25,26]. An alternative is to measure radioactivity by conventional techniques after quantitative recovery of the film [23]. Spectroscopic techniques have been used to study monomolecular assemblies such as fluorescence [27,28], UV [29] and CD spectroscopy [30] (for a review see Ref. 31). However, these methods often involve transfer of the floating monolayer to a solid surface. Deposition of single monolayers or multilayers (as many as 100) first developed by Blodgett and Langmuir [32,33] is rather easy with lipids such as barium stearate. The deposition 192 R. Verger and F. Pattus, Lipid-protein interactions in monolayers of lipoprotein monolayers on solids is more difficult. One can be satisfied with 5 successive depositions of lipoprotein monolayers. This technique has also been applied to electron microscopic studies [34,35]. PART I: PROTEINS OTHER THAN LIPOLYTIC ENZYMES Proteins which can be isolated in a soluble form Belonging to this class of proteins, some membrane proteins [ 3 6 - 3 8 ] , serum apolipoproteins [39,40], phospholipid exchange proteins [41], proteins involved in blood clotting [42] and toxins [43] have been studied by the monolayer technique. Experiments at constant surface, area Penetration capacity, lipid specificity Most studies were carried out at constant surface area. The protein is injected below a phospholipid fdm at an initial surface pressure (Tri). An increase of surface pressure is observed due to the penetration of protein segments into the interface. The final increase of surface pressure An = nfina 1 - ~ri and the initial velocity of surface pressure increase (dzr/dt)t= 0 give information on the affinity of the protein for the lipid-water interface. In Fig. 1, An is plotted against 7ri for the penetration o f apolipoproteins apo A1, apo E and bovine serum albumin injected beneath an egg phosphatidylcholine (PC) 25 z 15 • o APO E • APO A-| ~. BSA o N 5 5 15 LIPID 25 35 INITIAL ~ / r n N m -~ Fig. 1. Change of surface pressure (Art) when protein is injected beneath a monolayer of egg yolk PC spread at different initial pressures. (Reprinted with permission of the New York Academy of Sciences from Phillips and Sparks, Ann. N.Y. Aead. Sci., 348 (1960) 122.) R. Verger and F Pattus, Lipid-protein interactions in monolayers 193 monolayer at different initial surface pressures (from Phillips and Sparks [39]). The linearity between An and rti is a general property of protein penetration. Extrapolation of the linear plot to An = 0 gives a measure of the relative 'penetration capacity' of proteins into egg PC monolayers. These data indicate that bovine serum albumin (BSA) is excluded from egg PC monolayer when the lipid molecular area is less than 85 ~2 while the equivalent figures for apo A1 and apo E are 70/~2 and 65 ~ 2 respectively. The average molecular area for phospholipids in the surface of human lipoproteins has been estimated as about 65 ~2. This explains why BSA is not a component of serum lipoproteins and suggests that apo A1 and apo E would only just be able to remain at the lipoprotein surface. When lipoprotein lipase acts on triglyceride-rich lipoproteins, accumulation of surface active products of lipolysis would lead to a compression of the surface molecules. From the monolayer data one should expect sequential ejection of apo A~ before apo E. Such a mechanism could control the apoprotein composition of the apolipoproteins and could be important in the interconversion oflipoproteins. By changing the nature of the lipid, one can explore the lipid specificity of the protein. For example, we found a rather sharp lipid specificity with colicin A injected beneath different lipid monolayers. Colicin A is a basic protein (tool. wt. 60 000) produced by different strains of bacteria, which kills E. coli cells. Colicin A binds to a receptor on the outer 17. coli membrane and then by an unknown mechanism depolarizes the inner membrane. Colicin A penetrates phosphatidyl glycerol monolayers until the collapse pressure of the film, while the interaction stops at 20 and 25 dyn/cm for PC and phosphatidylethanolamine (PE), respectively. This capacity of proteins to penetrate lipid layers is not directly correlated with the overall charge or hydrophobicity of the protein. We found that the basic colicin A (pl = 9.5) phospholipid interaction is highly dependent on a protein residue(s) with a pK of 5.8. The cardiotoxin Ill from Na]a mossambica mossambica venom, a cytolytic toxin, is highly surface-active at the air-water interface when reduced and carboxymethylated [44]. However, in contrast with the native toxins, the penetration capacity of the denatured protein in phospholipid monolayers is negligible. This indicates that the three-dimensional 'native' structure of the protein is necessary for efficient penetration. The increase of surface pressure after protein injection is not a quantitative measurement of protein penetration. For example, with negatively charged monolayers and in absence of Ca 2~ ions, two superimposed inverse effects can be observed [44]: protein penetration which tends to increase surface pressure, and negative charge neutralization which tends to condense the phospholipid monolayer and to decrease surface pressure. This latter effect is abolished when the phospholipid monolayer is precondensed with Ca z÷ ions before protein injection. One should remember that in monolayer systems the surface/volume ratio is small (-~1 cm2/cm 3) as compared to liposomes or emulsions (5 × 103-104 cruZ/ cm3). Even for a protein with high affinity for lipid-water interfaces, only a small fraction of the protein injected binds to the monolayer. 194 R. Verger and F. Pattus, Lipid-protein interactions in monolayers When A3, is proportional to penetration, and when the binding is reversible and the adsorbant behaves ideally, the Gibbs equation (I' = -- ~-f 1 . d~//d log C) can be used to estimate the surface concentration and apparent molecular area of the adsorbant at the interface. In other cases, direct measurements with radiolabeled proteins are necessary. The influence of ionic strength, divalent and chaotropic ions and temperature on the increase of surface pressure, and the amount of protein bound to the film can give information on the nature of the protein-lipid interaction. Fluorescence spectroscopy, when it can be applied, can give information on the orientation of the protein at the interface. Recently Teissi~ [44b] used energy-transfer fluorescence quenching to observe the binding of cytochrome c to lipid monolayers. The probe (donor), dansyl PE, was dispersed either in dipalmitoyl phosphatidylcholine, in phosphatidic acid, or in the mixture of the two lipids. The heme of the protein was the acceptor. This author confirms the previous studies of Quinn and Dawson [7] which showed that the attraction between cytochrome c and phospholipid monolayers was a diffusion-limited process and linked mainly to electrostatic interactions. Furthermore, the orientation of the heine group of the protein with respect to the plane of the lipid interface could be measured. Reconstitution of functional enzymatic complexes Romeo et al. [45] studied the re-assembly of the galactosyl transferase system of Salmonella typhimurium in monolayers at the air-water interface. These authors showed that formation of a functional enzymatic complex needs the following sequential scheme (LPS, lipopolysaccharides): LPS + PE -* LPS-PE Enzyme + LPS + PE ~ E n z y m e - L P S - P E This sequence cannot be reversed without altering the transferase activity. It was shown that the lipid specificity of the functional complex originates from the LPS-phospholipid interaction, while the cofactor activity of Mg2÷ ions is at the level of the protein LPS-PE interaction. Mg2÷ ions mediate the penetration of the transferase into the LPS-PE interface. Determination of the protein segments interacting with lipids London et al. [46] studied the specific interaction of the bovine myelin A1 basic protein with lipid monolayers. The interaction was measured by recording simultaneously the change in surface pressure and surface radioactivity using basic protein lalI-labeled on two tyrosine residues [68,134]. This protein showed a specific interaction with cerebroside sulphate, an acidic lipid extracted from myelin. After surface pressure equilibration the films were transferred on to a clean R. Vergerand F. Pattus, Lipid-protein interactions in monolayers 195 subphase, and proteolytic enzymes such as trypsin, chymotrypsin or pronase were injected into the subphase. A decrease in surface pressure and a concomitant decrease in surface radioactivity were observed. After tryptic hydrolysis the A1 basic protein-cerebroside sulphate complex was collected from the interface. Peptide maps showed that specific regions of the protein molecule were protected from proteolysis after interaction with lipids. London et al. showed that the N-terminal part of the protein (position 20-113) containing one of the ~alI-labeled tyrosines, interacted with the lipid phase while the C-terminal region was released into the subphase by trypsin digestion. On the basis of these experiments and the known amino acid sequence of the protein, London et al. could draw a structural model of the Al-basic protein bound to the lipid surface suggesting specific lipid binding regions. Experiments at constant surface pressure Experiments at constant surface pressure have been extensively used for studying lipolytic enzymes (see Part II). With other soluble proteins this type of experiment has seldom been used. Bougis et al. [44] combined experiments at constant surface area and constant surface pressure to study the mode of interaction of snake venom cardiotoxins with lipid interfaces. These highly basic proteins interact preferentially with negatively charged phospholipids [43,47]. Although electrostatic interactions predominate, hydrophobic interactions seem to occur as well. The classification of cardiotoxins with respect to their penetrative capacity into phosphatidylserine (PS) films agrees with that established by spectrofluorometry in a liposome system, or by cell hemolysis. Experiments at constant surface pressure with 12SI-labeled cardiotoxin III allowed these authors to determine directly the apparent molecular area of the protein in the phospholipid film. This apparent molecular area was calculated from the increase in the surface area of the film, and the amount of radiolabeled protein in the film. In order to reduce experimental errors, one has to measure AS/S values lower than 15%. Extrapolation to zero surface increase ( A S / S = O ) gives the apparent molecular area of the first protein penetrating the pure phospholipid film. Bougis et al. found a similar molecular area at high surface pressure by this direct determination method or using the Gibbs equation (see above). These authors found, surprisingly, that the molecular area of cardiotoxin III gave two characteristic values of 1400 /~2 and 550 /~2 depending on the surface pressure (Fig. 2). From these results, together with the known sequence of this toxin and the three-dimensional structure of a related toxin, a mechanism of the lytic activity of the toxin was proposed. Protein insoluble in water: the vesicle spreading technique Most intrinsic membrane proteins are insoluble in water. In the absence of 196 R. Verger and F. Pattus, Lipid-protein interactions in monolayers a 0 ~J "IT = 32.2 1500 500 0 .... Dynes/cm o~ I ~ E ~<~ v 1000 0 ® <C k. $ 10 15 ~sls(%) k.. 0 soo :3 U O~ 0 0 b 10 30 50 Surface pressure.(dyneslcm) Fig. 2. (a) Variation with surface pressure of the apparent molecular area of diiodo-CTX IIl (0.9 X 10 -7 M) in dialuroylphosphatidylglycerol (DLPG) (o), dilauroylphosphatidylserine (DLPS) (o) and PS (~) films. Insert: variation of the apparent molecular area for different values of the relative surface increase (AS/S) obtained at 0.7, 0.9 and 1.8 X 10 -7 M of diiodo-CTX III (DLPG film at 32.2 dyn/cm). (b) Drawing of three-dimensional structure of snake short neurotoxin, 'flat' (left) and 'edgewise' (right) according to Feldmann. The hatched area represents the transition region (20-30 dyn/cm) between the 'flat' and 'edgewise' orientations [44]. (Reprinted with permission from Biochemistry, 20 (1981) 4915. Copyright (1981) American Chemical Society.) R. Vergerand F. Pattus, Lipid-protein interactions in monolayer8 197 detergents they aggregate and display no affinity for lipid interfaces. The methods described in the first section cannot be applied to the study of such proteins. A method to study lipid-protein interactions in monolayers was used recently by R.A. Klein et al. (unpublished). These authors spread the protein from a cholate solution directly on a preformed phospholipid monolayer at zero or low surface pressure. The detergent was then removed by changing the subphase. In a few cases membrane proteins are soluble in an organic solvent such as hexane and can be spread at the air-water interface together with lipids [48-50]. However, this is not a general situation: organic solvents may also irreversibly denature proteins. In order to tackle this problem we developed a method for spreading an aqueous suspension of lipoproteins or natural membrane vesicles at the air-water interface. The advantage of this technique is that it can be applied to any kind of lipoprotein complex using mild conditions. Principle o f the technique, composition o f the films obtained The principle of the technique is described in Fig. 3 [51 ]. This is an adaptation of the Trurnit method [52] used to spread proteins alone. The spreading at zero surface pressure is schematically represented on the left-hand side of Fig. 3. The aqueous suspension of vesicles or liposomes is added dropwise at a rate of 30 pl/ rain with an Agla Microliter Syringe Unit along a clean, wet glass rod (5 mm in diameter) positioned in compartment I. The membranous suspension drips from the rod and spreads over the water surface. When the water surface is large enough, no rise in surface pressure is observed. The glass rod is then removed and the mobile barrier moved to compress the surface film. The film then passes through compartment II where it is rinsed, and when further compressed passes completely to the surface of the third compartment. The spreading at a constant surface pressure is schematically represented on the right-hand side of Fig. 3. The glass rod is here located in compartment III and a mobile barrier (A) placed close to the glass rod. As the membranous material spreads over the surface, there is an immediate rise in surface pressure which is automatically compensated by the movement of the barrier. A barostat [22] drives the mobile barrier in order to expand the surface film and hence maintains the surface pressure around the end point value (-+1 dyn/cm). When all the membranous material has spread, a second barrier (B) is added; the two barriers are then linked and pulled back over compartment I. One must keep in mind that less than 100% of the molecules spread at the interface. Depending on the nature and concentration of the membrane material used, spreading yields fluctuate between 30 90% of the material added to the glass rod [51,54-58]. The spreading yield of the membrane components is independent of the size, or the orientation of the glass rod. Better spreading yields are obtained with concentrated suspensions. However, the suspension must not be too viscous. 198 R. Verger and F. Pattus, Lipid-protein interactions in monolayers SPREADING OF PIG INTESTINAL BRUSH BORDERVESICLES AT THE AIR/WATER INTERFACE AT ~ ~ ZERO SURFACE PRESSURE ~ ~ ill I ~ i[' ill li [ LI ~Liiil AT CONSTANTSURFACE PRESSURE lii I I III Lll I Fig. 3. Principle of the two different spreading techniques used. The thermostated (25°C) Teflon trough is composed of four compartments connected by narrow surface canals. The subphases of compartments I, II and III are mixed by magnetic stirrers at 250 rev./min. Compartment II is used as a rinsing compartment. The surface pressure recording device is based on the principle of the Langmuir balance and is located in compartment IV [51]. In contrast with claims by Korenbrot and Pramik [53], an osmotic shock is not required for the vesicles to spread. Moreover, a low ionic strength subphase is unfavorable for spreading. In our hands 0.1 M or higher salt concentrations were necessary to obtain good spreading yields. Using phospholipid vesicles [54], we showed that spreading at low surface pressure (Tr < 5 dyn/cm) was required to obtain true monomolecular films at the a i r - w a t e r interface. When spreading was performed at a higher surface pressure, the vesicles retained part o f their internal contents, and the compression isotherms deviated from a monomolecular film. Another spreading technique was used by Schindler [57]. Films were formed by vesicle adsorption and spontaneous spreading at the interface from a lipoprotein suspension in the subphase. A detailed structural analysis revealed that below 5 dyn/cm any vesicle reaching the interface disintegrated and formed a monolayer. Above 5 d y n / c m the sublayer o f the monolayer becomes highly enriched in vesicles whose lipids exchange with those of the monolayer. Biochemical R. Vergerand F. Pattus, Lipid-protein interactions in monolayers 199 analysis of the films obtained after spreading natural membrane vesicles of different origin, showed that their lipid composition is identical to the composition of the starting membrane material. The protein content is not always identical. Some extrinsic proteins are released during spreading [55]. However, intrinsic membrane proteins spread as well as lipids. Biological activities and structure o f membrane films One advantage of the sp,eading of vesicle suspensions is that many membranebound enzymatic activities are preserved. The intestinal brush-border hydrolases, E. coli NADH oxidase [55], and erythrocyte acetylcholine esterase activities [56] could be measured either after film recovery, or by injecting the substrate into the subphase below the enzyme film. Care must be taken, however, that the film is maintained at low surface pressure for short periods of time only. We found experimentally that below 15 dyn/cm slow expansion of the film occurred due to slow unfolding and protein denaturation at the interface. Above 15 dyn/cm surface pressure, enzymatic activities are stable over hours [56]. The main advantage of the monolayer technique is that the packing of the molecules at the interface can be controlled. Figure 4 shows the influence of surface pressure on two different membranebound enzymes. The upper curve shows that the intestinal aminopeptidase activity is independent of surface pressure. This result is in agreement with the mode of integration of the protein into the membrane. The aminopeptidase is immobilized by a small hydrophobic peptide with the active site far from the lipid interface [58,59]. By contrast, erythrocyte acetylcholine esterase activity is highly influenced by surface pressure. This behavior indicates that the enzyme is buried in the lipid matrix. One may note that the specific activity of acetylcholine esterase at 30 dyn/cm is equivalent to its specific activity in erythrocytes. This surface pressure was given as the 'equivalent surface pressure' of the erythrocyte membrane [13]. All the membranous enzymatic activities mentioned so far belong to proteins distributed on one side of the membrane. It is to be expected that these enzymes would remain active in lipid monolayers. One can ask about the orientation and activity of 'spanning' proteins with functional sites on both sides of the bilayer. Surprisingly we found that Ca2÷-ATPase from sarcoplasmic reticulum was activatable by Ca 2÷ ions at the air-water interface [56]. Moreover, Trissl and Graber [60] recently studied thylakoid membrane layers obtained by spreading chloroplast material at a heptane-water interface. They measured the flash-induced potential change at the interface with a capacitive electrode. These authors showed that chloroplast films are asymmetrically orientated. The photovoltage from the interfacial layer showed similar characteristics to the photosynthetic primary charge separation. Kinetics of the charge separation were investigated. Although the structure of such films was not extensively studied, Trissl and Graber gave R. Verger and F. Pattus, Lipid-protein interactions in monolayers 200 • • • •w • • • • • -i~ J o~ 1.0 / 2.5 C.O -13 n 2,0 n ~, -- 0 W 0 ~ o=' Z "Zm 5" o - , \ ,, 0.s o o" & 6,,. t " £ i )0 I 15 I 20 L 25 I 30 35 (dynes/cm) Fig. 4, Specific activity of aminopeptidase and acetyl choline esterase as a function of surface pressure. Pig intestinal brush-border m e m b r a n e s (o e) or h u m a n e r y t h r o c y t e m e m b r a n e s (o---o) were spread at 'zero surface pressure'. Enzymatic activities were measured by injection of a chromogenic substrate below the film. Specific activities were calculated by dividing these activities by the a m o u n t of protein in the film [56]. arguments that this thylakoid membrane layer is presumably a monolayer of oriented membrane constituents. The structure of such membrane films is still, however, a subject of debate and needs further study. According to Hwang et al. [48] who spread bacteriorhodopsin fragments from hexane, and Korenbrot and Pramik [53] who spread reconstituted cattle rhodopsin by 'osmotic lysis', such films consist of bilayer patches surrounded by a sea of phospholipids in a monomolecular array. Under these conditions, however, the spreading pressure was not controlled and the spreading yield was very low. With lipid vesicles there is good evidence that films are either pure monolayers (spreading at zero surface pressure) [54] or a monolayer with a sublayer enriched in lipid vesicles [54,57]. With erythrocyte membrane films spread under controlled conditions and observed by freeze etching after recovery on a mica sheet, we found ~o evidence for patch formation [56]. Discrepancies between our results and those of Hwang et al. may also be due to the peculiar structure of purple membrane fragments. Such fragments are, for example, not dissociated in hexane. R. Vergerand F. Pattus, Lipid-protein interactions in monolayers s,s%s,° I so,,a . . . . t 201 ,o,mat,oof Fig. 5. Principle of bilayer formation from two monolayers [66]. Formation o f planar bilayers from two monolayers This is the most fascinating perspective offered by the monolayer technique. Protein as well as lipids are asymmetrically distributed within the two halves of the membrane. With membrane models such as liposomes, or the planar bilayers first foreseen by Langmuir and Waugh [61] and developed by Mueller and Rudin [2], only few proteins could be re-incorporated asymmetrically. In 1967 Takagi et al. [62] paved the way for the successful formation of asymmetric model membranes by publishing a short note on a method for the formation of planar lipid bilayers from two monolayers containing rhodopsin. Five years later this technique emerged in the work of Montal and Mueller [63]. In 1978 a further breakthrough in the development of a general method for reconstitution of biological membrane function in an asymmetrical environment was achieved by Schindler and Rosenbush [64]. These authors applied the vesicle spreading technique to form planar bilayers from two monolayers. The advantage of forming the monolayers by vesicle adsorption from the subphase (Fig. 5) is that the bilayer is at equilibrium surface pressure with the vesicles, and that this surface pressure is maintained constant during membrane formation. The other advantage is that the number of proteins per bilayer can be adjusted at will by dilution with lipid vesicles [65]. One drawback, however, is the need for more biological material compared to the Pattus' method. Schindler and Quast [66] achieved successful reconstitution of the acetylcholine receptor function by this technique. Recent studies indicate that functional asymmetric planar bilayers can be formed with a wide variety of membrane complexes. Recently, Korenbrot and Hwang [67] succeeded in incorporating bacteriorhodopsin in bilayers as large as 0.7 cm z by deposition of two monolayers on electrically conductive nitrocellulose supports. PART I1: LIPID MONOLAYERS AS SUBSTRATES FOR LIPOLYTIC ENZYMES Introduction Naturally occurring lipids are important building blocks for biological membranes. They are water-insoluble and spontaneously form molecular aggregates, 202 R. Vergerand F. Pattus, Lipid-protein interactions in monolayers O Fig. 6. Schematic picture of lipolysis. such as monomolecular films, bilayers, emulsions, liposomes or micelles. In addition, a number of soluble enzymes that play an important role in such biological events as digestion, lipid transport and lipid metabolism are known and have been isolated. The fact that it is the substrate itself that forms molecular aggregates makes lipolysis a very attractive system for studies of interfacial enzyme kinetics. The highly schematic picture of lipolysis illustrated by Fig. 6, in which a soluble enzyme, E, transforms an insoluble substrate, S, into partly soluble products, P, raises several questions. For general aspects of enzymatic lipolysis the reader is referred to previous reviews [68-74]. Activation o f lipolytic enzymes by interfaces One of the most characteristic and intriguing features of lipolytic enzymes is their activation by interfaces. This phenomenon was recognized very early by Holwerda et al. [75] and Sch~bnheyder and Volqvarts [76] who, using tricaproin as a substrate for pancreatic lipase, showed clearly that the rate of breakdown of a molecular solution of this glyceride is very slow but that once the substrate solubility is exceeded, the enzymic activity increases dramatically. In 1958, when highly purified preparations of the enzyme became available, the problem was re-investigated by Sarda and Desnuelle [77] in a more quantitative way. They clearly demonstrated a fundamental difference between an ordinary weU-known esterase and pancreatic lipase based upon an ability or inability to be activated by interfaces. The lipases appear to constitute a special class of esterases capable of hydrolyzing plurimolecular aggregates at a high rate. This peculiar behavior of pancreatic lipase turned out not to be limited to emulsions. Brockman et al. reported a 1000-fold increase in the rate of hydrolysis of tripropionin by pancreatic lipase in the presence of siliconized glass beads [78]. The interpretation given by these investigators was that a spherical glyceride monolayer formed around the beads to give rise to an interface similar to that in micelles or emulsions. R. Verger and F. Pattus, Lipid-protein interactions in rnonolayers 203 Pieterson et al. [79] and Misiorowski and Wells [80] have shown that other lipolytic enzymes are also strongly activated by substrate aggregation. The latter group found an increase of about 104-fold in enzyme activity in a diethyl ether medium in which small water droplets containing phospholipase A2 and Ca2+ were coated by a monolayer of dioctanoyllecithin. In pure aqueous systems, below the critical micelle concentration (CMC) where only monomeric lecithin molecules are present, phospholipiase A activity remains very low and seems to follow normal Michaelis-Menten kinetics. When the monomers aggregate above the CMC, a strong increase in lipolytic activity occurs indicating that the miceilar aggregates are a much better substrate for the enzyme than the molecules dispersed in water. The naturally occurring zymogen of pancreatic phospholipase A2 has about the same activity as the enzyme on monomeric short-chain lecithins; however, above the CMC the formation of substrate aggregates does not accelerate hydrolysis. As one might expect, not all interfaces are equivalent in this effect. From bulk studies with pancreatic phospholipase A2, it has been shown that homologous lecithins with acyl chain lengths varying from 6 to 10 carbon atoms are hydrolyzed at very different rates [81]. For example, under similar conditions of ionic strength the enzyme hydrolyzes the dioctanoyl derivative with a specific activity of 6000 /amol min -1 mg -1, whereas the didecanoyUecithin is not hydrolyzed at all. In contrast, it was found by the monolayer technique that lecithins with acyl chain lengths varying from 8 to 12 carbon atoms were hydrolyzed at about the same rate [82]. This difference in response to chain length can be assumed to be related not to the different techniques used (bulk and monolayer), but to a strong dependence of enzyme activity on the 'quality' of the interfacial structure of the substrate. Once the 'quality' of the interface is equalized, the enzyme acts with the same velocity. Interfaces occur in nature and can be prepared experimentally as emulsions, micelles, liposomes and monolayers. In this review, we limit ourselves to the use of monomolecular films as substrates for lipolytic enzymes. Why use lipid monolayers as substrates ? Spreading of a lipid at the air-water interface as a monomolecular film has been intensively studied over 60 years. For classical references, see the books of Adam [83], Davis and Rideal [84], and Gaines [18]. There are at least four major reasons for using lipid monolayers as substrates for lipolytic enzymes: 1. The technique is highly sensitive and very little lipid is needed to obtain kinetic measurements. This argument can often be decisive when one uses synthetic or rare lipids, although the advantage of high sensitivity may be partly offset by the requirement for pure substances and for extreme cleanliness. 2. During the course of the reaction, one can monitor one of several physicochemical parameters characteristic of the monolayer film: surface pressure, 204 R. Verger and F. Pattus, Lipid-protein interactions in rnonolayers potential, radioactivity, etc. These variables often give unique information. 3. A rather common observation reported by many authors working on the kinetics of lipolytic enzymes is the presence of a lag period in the hydrolysis of both emulsions, liposomes, micelles and monolayers [78,85-96[. Because such lag periods in general prevent measurements of initial velocity, one attempts to suppress them, e.g. by the addition of surface-active agents such as bile salts or Triton. It seems evident, however, that a study of such pre-steady states might yield valuable information on the initial stages of the reaction. Such studies should preferably be done on monolayers of short chain lipids where the perturbing influence of increasing amounts of reaction products can be minimized. 4. Probably the most important reason fundamentally is the possibility of varying the 'quality of interface' determined by the nature of the lipids forming the monolayer, the orientation of the molecules, molecular and charge density, water structure, fluidity, etc. One further advantage of the monolayer technique as compared to bulk methods is the possibility of transferring the film from one aqueous subphase to another. By contrast, a shortcoming of the monolayer technique is the adsorption or denaturation of many enzymes at lipid-water interfaces [1437,97-104]. This phenomenon is related to the interfacial free energy. Surface films at high interfacial energy (low surface pressure, ~r) are highly denaturing, whereas films at low interfacial energy (high ~r) are much less denaturing. Different solid-water interfaces used as containers for dilute enzyme solutions can cause adsorption and denaturation of protein. Hydrophobic materials having high interfacial free energies, e.g. Teflon, can adsorb large quantities of proteins [105,106]. Fortunately, this phenomenon occurs slowly [23]. As already discussed in part I, a major difference between the monolayer and the bulk system lies in the ratio of interfacial area to volume (I/V), which are different by orders of magnitude. In the monolayer system, this ratio is usually about 1 cm -~ depending upon the depth of the trough, whereas in the bulk system it can be as high as l0 s cm -1, depending upon the amount of lipid used. As a consequence, bulk conditions allow the adsorption of nearly all the enzyme at the interface, whereas with a monolayer only one enzyme molecule out of 100 may be at the interface [69,82]. As a consequence of this situation, a small but unknown amount of enzyme, responsible for the observed hydrolysis rate, is adsorbed to the monolayer. In order to circumvent this limitation two different methods were proposed to recover and measure the quantity of enzymes adsorbed at the interface [23,107a,b]. Pure lipid monolayers as substrates A new field of investigation was opened in 1935 when Hughes [108] used the monolayer technique for the first time to study enzymic reactions. He observed R. Vergerand F. Pattus, Lipid-protein interactions in monolayers 205 that the rate of the phospholipase A-catalyzed hydrolysis of a lecithin film measured by the decrease of surface potential was considerably reduced when the number of lecithin molecules per square centimeter was increased. Since this early work, several laboratories have used the monolayer technique to follow lipolytic activities, mainly with glycerides and phosphofipids as substrates. These studies can be tentatively divided into four groups. Either long-chain lipids were used and their surface density (number of molecules per square centimeter)was not controlled; or short-chain lipids were applied, again without control of surface density; or short-chain lipids were used at constant surface density. Recently, a method was developed in our laboratory to study the hydrolysis of long-chain lipids with control of surface density. As we shall see, it is not surprising that this classification is roughly chronological. Group 1. Long-chain lipid monolayers without control o f surface density Bangham and Dawson [109] and Dawson [110] investigated several types of phospholipase that catalyze the hydrolysis of acyl ester bonds in glycerophosphatides. Quinn and Barenholz [111] compared the activity of phosphatidylinositol phosphodiesterase against substrate in dispersions and as monolayers at the air-water interface. All these authors followed the enzymic hydrolysis of monolayers of 32P-labeled phospholipids by measuring the loss of surface radioactivity that occurs as the water-soluble 32p-labeled reaction products leave the surface and are no longer detected by the surface counter. Dawson et al. noticed that at surface pressures above 30 dyn/cm enzymic hydrolysis of the lecithin film did not occur unless a small amount of an anionic amphipathic substance was introduced into the film. This result led the authors [112] to the conclusion that the sign of the zeta potential on the substrate surface was critical for the initiation of enzyme activity.* Following the original technique described by Hughes in 1935, Colacicco and Rapport [113] and Shah and Schulman [114] measured the action of snake venom phospholipase A on lecithin monolayers. Both groups spread the lipid film on an enzyme solution and followed the fall in surface potential. Assuming that the reaction products remain in the film, the fall in surface potential will be proportional to the number of lecithin molecules transformed to lysolecithin and fatty acid, and it can be used to express the activity of the venom. Depending upon the degree of unsaturation of the lecithin used, optimal activity of the venom phospholipase was encountered at surface pressures between 12 and 25 dyn/cm, values that are different from the 2 9 - 3 3 dyn/cm found by Dawson [110] using 32P-labeled lecithin. *It has been shown that the hydrolysis of PC suspensions by C1. welchii phospholipase C has an absolute requirement for Ca2÷ as well as being affected by the surface charge on the substrate (Chem. Phys. Lipids, 15 (1975) 15). 206 IL Vergerand F. Pattus, Lipid-protein interactions in monolayers For the surface-potential technique to be valid, however, it is essential that the products of the reaction remain in the film, whereas for correct surface radioactivity measurements using 32p-labeled lecithin it is imperative that the lysolecithin molecules rapidly leave the film. With natural long-chain phospholipids, it is clear that these two extreme situations are never encountered. In our opinion, the reliability of both techniques is doubtful. If at low rr, the products of the reaction remain in the film, what is their influence on the rate of enzymic reaction? If at high rr the products leave the surface, is this desorption process quantitative and not rate limiting? The validity of both techniques in following enzymic hydrolysis of phospholipid monolayers has been discussed by Colacicco [115]. Another example showing the limitations of monolayer kinetics obtained with long-chain phospholipids are studies of the action of phospholipase D on surface films of [14C]choline-labeled lecithin [87]. The hydrolytic reaction was followed by the surface radioactivity technique. Upon cleavage of the polar headgroup, phosphatidic acid was formed and remained in the film. The presence of negatively charged phospholipid appeared to have a profound influence on the enzymic reaction. Especially at high surface pressure, the authors observed an increase of the hydrolysis rate with time. An interesting attempt to follow more quantitatively the phospholipase C hydrolysis of [3H]choline-labeled lecithin by direct surface counting, has been reported by Miller and Ruysschaert [98]. They showed that the diffusioncontrolled escape rate of radioactive phosphorylcholine from the film was much higher than the rate of enzymic cleavage. Therefore, the latter step is rate limiting and controls the decrease in surface radioactivity. The authors showed experimentally that, in a limited range of substrate concentrations, V is linearly dependent on the surface concentration. Deviations from linearity were found at lower surface concentration, and the velocity dropped with time: this behavior was explained by irreversible adsorption leading to inactivation of the enzyme at the interface. Also at higher surface concentrations much lower enzyme activities were found, and this was interpreted as resulting from steric hindrance in the substrate-enzyme interaction. Unfortunately, the use of an impure enzyme preparation and a racemic substrate did not allow more quantitative conclusions to be drawn. Group 2. Short-chain lipid monolayers without control o f surface density In order to overcome the technical difficulties inherent in the use of long-chain lipids, Olive [116], Lagocki et al. [117], Garner and Smith [118], and Dervichian [119] used synthetic short-chain lipids which upon enzymic hydrolysis yield readily soluble products. Since then, several groups have used short-chain lipids as substrates in monolayer studies of lipolytic enzymes. Garner and Smith [118] studied the action of pancreatic lipase on a series of octanoyl glycerides and octanoate esters spread as monolayers at the air-water R. Verger and F. Pattus, Lipid-protein interactions in monolayers 207 interface, With the exception of trioctanoin all substrates yielded soluble reaction products, and the authors followed the course of the reaction at constant area by the fall in surface pressure with time. The rate of hydrolysis was determined near the collapse pressure, where the decrease in film pressure was found to be linear with respect to time. Another technique follows the hydrolysis of monomolecular films of glyceryl tri[1-14C]octanoate by measurement of the decrease in surface radioactivity that occurs in the presence of lipoprotein tipase [120-123]. Interestingly, the hydrolysis rate of a variety of substrates measured by the monolayer technique [118] correlated with the corresponding rates measured by Derbesy and Naudet [124] using the classical bulk titrimetric procedure. A similar monolayer study has been reported by Lagocki et al. [125] who investigated the kinetics of lipase hydrolysis of monolayers of trioctanoin and dioctanoin. They reported that the rate of the reaction was independent of the surface pressure of the substrate. Although these authors showed that during the hydrolysis of trioctanoin the formation of insoluble 1,2-diglyceride did not greatly affect the hydrolysis of the triglyceride in the mixed film, we want to emphasize that in general the production of insoluble reaction products should be avoided. In our discussion on long-chain monolayers, we have seen how the presence of insoluble reaction products can influence the rate of lipolysis; if such products are at the same time substrates for the enzyme in a consecutive, albeit slower reaction, interpretation of the kinetic results becomes more complicated [125]. A second point of discussion concerns the technique used by the above groups, i.e. studying the rate of hydrolysis at constant surface area. This method is simpler than working at constant surface pressure. However, the former technique can be used only in those few cases where the rate constant of hydrolysis is independent of the surface pressure. Brockman et al. [78] studied the enhancement of hydrolysis of tripropionin by pancreatic lipase in the presence of siliconized glass beads. This is a very interesting interfacial system that can be considered as a spherical monolayer of tripropionin at a liquid-solid interface. The binding of the enzyme to the surface was shown to be reversible and diffusion controlled. In addition, the hydrolytic reaction on the surface appeared to be first order with respect to the amount of adsorbed enzyme, and first order with respect to the concentration of tripropionin at the solid-liquid interface. This latter result is in agreement with the findings of Miller and Ruysschaert [98] (see below). Brockman et al. reported an enhancement of the velocity on the surface of the siliconized glass beads of three orders of magnitude compared to the homogeneous reaction. They ascribed this activating effect to the increased local concentration of the substrate at the interface. It should be noted, however, that although these authors measured directly the fraction of enzyme adsorbed at the interface this determination was done in the absence of tripropionin. From their results, it can be calculated that the ratio of molecules present as monomers in solution to tripropionin molecules adsorbed to the interface is about 300. Taking into account that the K m of lipase for tripropionin monomers 208 R. Vergerand F. Pattus, Lipid-protein interactions in monolayers is of the order of 10 mM [126], we see that the large amount of substrate in the bulk phase could act as a 'competitive inhibitor' [127]. Group 3. Short-chain lipid monolayers with controlled surface density Dervichian [119], using short-chain lipids spread at the air-water interface, described for the first time a technique for following the rate of enzyme reactions at constant surface pressure as shown schematically in Fig. 7. The kinetics of the surface reaction were followed by continuously recording the decrease in area of the film with time. Because most lipolytic reactions taking place at the air-water interface have been shown to be surface pressure dependent, this technique seems to be optimally suited to study interfacial kinetics. Moreover, use of a fully automated barostat as proposed by Verger and de Haas [22] makes the technique rather simple and accurate. The technique is limited, however, to those cases where the substrate forms a stable surface film at the air-water interface, and where all enzymatic reaction products are freely water soluble and diffuse immediately away into the bulk phase. Therefore, until now only acyl esterases have been studied in detail with this method. Because of the requirement for highly purified lipolytic enzymes, it is not surprising that most reports deal with the lipolytic enzymes, lipase and phospholipase A2. Conditions of solubility explain why surface studies on (phospho)lipase action have been limited generally to short-chain glycerides or phospholipids containing 8 - 1 2 carbon atoms in the acyl chain. To bring enzyme and substrate together, two techniques are extensively used, namely, spreading of the lipid film on a homogeneous enzyme solution or injection .)J. ~1 . , Fig. 7. Principle of the technique for measuring enzyme activity on short-chain hpid monolayers with controlled surface density. R. Verger and F.. Pattus, Lil- M-protein interactions in monolayers 209 of the enzyme under a preformed, stabilized surface monolayer. The injection technique requires efficient stirring of the subphase in order to obtain a homogeneous bulk solution rapidly. In the spreading technique, stirring is not needed. This might be considered as advantageous for the stability of the film; however, in both cases the possibly rate-limiting step of product desorption from the interface can be overcome by efficient stirring [82]. A possible drawback of the spreading technique is the adsorption and/or denaturation that can occur with some enzymes at interfaces during the time required to clean the surface. Consequently, there is a possibility of trapping the enzyme at the interface during spreading that would lead in general to higher reaction rates than those observed with the injection techrdque [82]. Several types of trough have been used to study enzyme kinetics. The most generally used is a simple rectangular one giving non-linear kinetics. To obtain rate constants, a semilogarithmic transformation of the data is required. This drawback was overcome by a new trough design consisting of two compartments connected by a narrow surface canal. As shown on the right side of Fig. 8, the recorded kinetics obtained with this trough are linear in contrast to the non-linear plots obtained with the usual one-compartment trough. Brockman et al. [128] investigated the rate of lipase hydrolysis of trioctanoin and 1,2-dioctanoin films at constant surface pressure. Their method is based upon measurement of the amount of substrate that must be added to a monolayer to maintain constant surface pressure during the course of the enzymic reaction. Although the idea of an 'open' system in which substrate molecules disappearing from the monolayer are continuously replaced is attractive, the presence of organic solvents at the surface might influence the surface parameters of the film. This is especially the case for phospholipids, where less volatile solvents must be used to spread the film and such effects cannot be excluded. Moreover, surface heterogeneity may influence the kinetics at higher surface pressures where the films are no longer in an ideal liquid-expanded state. On the other hand, the usefulness of this technique was demonstrated clearly in the lipase-catalyzed hydrolysis of a surface film of trioctanoin. The experiments not only confirmed the surface pressure dependence of the enzymic reaction, but also showed the complexity of the kinetics in lipolytic reactions where the first hydrolysis product is a substrate in a consecutive lipolytic reaction. Kinetic equations were derived for a two-step reaction yielding an insoluble intermediate and soluble final products. In a series of papers, Dervichian and Barque [ 129-131 ] investigated the kinetics of pancreatic lipase acting on surface films of 1,3-didecanoylglycerol. In agreement with previous studies [82,88], it was found that a rapid establishment of the adsorption-desorption equilibrium required efficient stirring of the subphase. However, in contrast to most other monolayer studies using lipolytic enzymes Dervichian and Barque stopped stirring after mixing the enzyme sample with the aqueGus bulk subphase. The authors considered that the extreme slowness of diffusion, far from being a drawback, was an advantage because once the agitation was stopped, practically no new molecules of enzyme came to the surface and R. Verger and F. Pattus, Lipid-protein interactions in monolayers 210 o o i00 ~ o ~: 0----- - 0 I-- o o o. 0 0 0 ;E W w I-- o . I: ::::::::::::::::::::::::::::::::::::::::::::: 0 o . . . . . . . . '~ 0 . [] of e~ .n .~ [] 0--o o o 0 °i ~ ~ . . . . CO ',D ~" ---71 7 o / ~o_ o ° ~: 0,1 / o o I ~ :i cO ~ 1o17 / ..... _ tOllO o o 8 o o o o ~= ~ ~ Q .~ R. Vergerand F. Pattus, Lipid-protein interactions in monolayers 211 only those enzyme molecules already adsorbed take part in the lipolysis reaction. Thus, with the limited stirring procedure used, substrate and enzyme together form a segregated well-defined system on the surface. A limitation of this method, however, could be the rate limiting step of product desorption from the monolayer when no stirring is applied. Barque and Dervichian [131] further studied lipase action on monolayers at constant surface area, following the decrease in surface pressure with time. Because of the limited stirring procedure described above, the authors assumed that the total amount of adsorbed enzyme (E* + E*S in Ref. 88) remained constant. Upon decreasing the surface pressure, i.e. substrate density, the equilibrium E* + S -~ E*S is supposed to shift to the left and when [E*]gets higher than the corresponding value in equilibrium with the bulk enzyme concentration, the enzyme would desorb from the monolayer. Unfortunately no direct experimental data support this interesting interpretation. Nevertheless the main conclusion of Dervichian and Barque, based solely on kinetic evidence, states that both the amount of adsorbed enzyme and its specific activity increase with increasing packing density of the substrate molecules. Influence of surface pressure on enzyme velocity In agreement with the results of studies on long-chain phospholipids [87,108, l 13,114] most investigators using soluble short-chain lipids also report an optimum in the velocity-Tr profile [13,82,132-134]. The exact value of this optimum varies considerably with the particular enzyme-substrate combination used, and differences in the experimental technique such as spreading or injection may also cause variations in the optimal pressure. Several qualitative interpretations have been given to explain this phenomenon. Agreeing with the original interpretation given by Hughes [ 108], later workers tried to explain the existence of an optimum in the velocity-Tr profile by emphasizing the importance of molecular orientation in the monolayer for an optimal fit between the active site of the enzyme and the substrate molecule. From both physical and kinetic data, it was proposed that the glycerol moiety of the substrate may exist in two discrete conformations, only one of which is reactive [107b,165]. The distribution of substrate between the conformers could be a function of lipid packing density. The authors support the hypothesis that a packing-dependent conformational distribution of substrate head groups is a factor in the regulation of lipolysis. Another interpretation of the n-optimum was given by Esposito et al. [133]. These authors supposed that fipase upon adsorption at the interface acquires a functional conformation for intermediary values of the interfacial free energy and consequently of the film pressure n. Lower and higher values would lead to inactive forms due to denaturation or insufficient conformational change in the enzyme molecule. From the early work of Shah and Schulman [114] as well as from results of Verger et al. [134] and a study by Pattus et al. [91], it is clear that this hypothesis cannot be correct. The observed maxima in velocity-It profiles disappear when they are related to the interracial excess of enzyme. 212 R. Vergerand F. Pattus', LipM-protein interactions in monolayers Shen et al. [135] reported that the Crotalus adamanteus phospholipase A2 catalyzed the hydrolysis of dioctanoyl lecithin monolayers independently of the surface pressure regardless of the method of measurement, whether under constant surface area or pressure. This contrasts sharply with the results reported for pancreatic phospholipase A2 [82,136]. Although such an independence of the reaction rate may be found over a limited range of surface pressure, it is definitely not a general phenomenon; this has been demonstrated by several laboratories. Interestingly, Shen et al. [135] concluded that the active dimeric form of the Crotalus aclamanteus phospholipase A2 is in a true equilibrium with enzymatically inactive subunits. However, one can argue that particularly at low enzyme concentrations (10 -9 M) and low surface pressure (8 dyn/cm), irreversible denaturation and/ or adsorption can occur giving underestimated experimental velocities. Cohen et al. [ 101 ] working at variable surface pressure, reported the existence of two enzyme classes with respect to the effect of surface pressure on the rate of hydrolysis of substrate monolayers. They found two enzymes that showed continuously increasing activity with increasing surface pressure, and two that had extensive regions of pressure independence. They further showed that the pressureindependent enzymes were quite resistant to surface denaturation, while the pressure-dependent enzymes underwent rapid surface denaturation at air-water interfaces. On the basis of this correlation, they attributed the pressure dependency to the denaturation of the enzyme in surface domains that were not covered by substrate molecules. Finally, Cohen et al. [101] stated that 'surface denaturation of enzymes is sufficient to explain the phenomenon of surface pressure dependency'. Such a conclusion is weakened, however, by a limitation inherent to the monolayer technique. The amount of enzyme really involved in catalysis is unknown. If, in a velocitym profile, the amount of enzyme adsorbed to the monolayer was a function of 7r then the plotted velocities would be apparent values and the ~ optimum would have no obvious meaning. Verger et al. [103] and Pattus et al. [91] have re-investigated this problem using radio-labeled pancreatic phospholipase A2. They measured simultaneously the enzyme activity and the amount of radioactive enzyme in excess at the interface in the presence of a lecithin monolayer. This enzyme was found to be resistant to interfacial inactivation. Figure 9 summarizes the main results obtained at different surface pressures for the film. The amount of radioactive enzyme present in, or close to the interface decreases linearly with increasing surface pressure. The ratio of observed enzyme activity to the amount of protein as determined by surface radioactivity gives the specific activity of the enzyme. The specific activity increases continuously with 7r, reaching an optimum value in the range of surface pressures investigated. This view was recently challenged by Momsen and Brockman [107b]. These authors showed that both the extent of enzyme adsorption and the catalytic rate constant of the adsorbed pancreatic lipase molecules increased linearly with 1,3-didecanoyl glycerol monolayer packing [107b]. However, these authors R. Vergerand F. Pattus, Lipid-protein interactions in monolayers 05_ 10 213 25 -t ~9 x 04_ 8 6 ///'~"'"~l 20 J \\/ 15 _ ~o3_ c E 4 _ --__0.2 _ 10 ~3_g 5 I ~ 1_~ I I I 6 8 10 j 12 14 16 SURFACE PRESSURE Idynes/cmJ I I I 12 13 14 PACKING DENSITY x l 0 II{moiecule$/cm2) Fig. 9. Influence of surface pressure and packing density on the activity (o- - -o), lag time (e --) and surface density (• o) of 1~5I amidinated phospholipase A2 acting on didecanoyllecithin monolayers. (Reprinted with permission from Pattus et al., Biochemistry, 13 (1979) 2691. Copyright (1979) American Chemical Society.) determined the amount of adsorbed enzyme by measuring the lipase activity remaining after film recovery on an hydrophobic paper sheet, while Rietsch et al. [23] measured the amount of radiolabeled enzyme after aspiration of the surface film. In the former technique, the yield of active enzyme recovered was assumed to be identical at all surface pressures used and taken as 80%, value found for the tri[1J4C]oleoylglycerol. Another limitation of this technique is that hydrophobic surfaces are known to be highly denaturing for the pancreatic lipase [23,99-104]; as a consequence only residual active enzyme molecules are detected by this assay procedure. Furthermore it is well established that the denaturation rate constant of pancreatic fipase is surface pressure dependent [23]. All these experimental limitations, due to the hydrophobic paper recovery technique, weaken the conclusions of Momsen and Brockman [107b] concerning the increase with surface pressure of the amount of adsorbed pancreatic lipase to monomolecular lipid films. Nevertheless at high surface pressures (Tr > 15 dyn/cm), where surface denaturation is minimal for pancreatic lipase, the authors confirmed that as little 214 R. Vergerand F. Pattus, Lipid-protein interactions in monolayers as few percent of active enzyme molecules are adsorbed to the glyceride monolayer with turnover numbers close to those observed previously [103,154] .This interesting approach of Momsen and Brockrnan reinforces the idea that lipid packing density and conformation regulate both the interfacial enzyme adsorption and the reactivity of substrate molecules. Influence o f protein cofactors on hydrolysis o f lipid monolayers Using lipid monolayers at low surface pressure, it was shown that lipases irreversibly denature at interfaces. Thus, one must be careful in interpreting data obtained at low surface pressure with lipolytic enzymes. Depending upon the experimental conditions used, very high apparent activations can be observed. Often, this activation is only apparent and reflects the protective effects of the factors against interfacial enzyme denaturation. The rate of hydrolysis, measured by the decrease in surface radioactivity of a monomolecular film of glyceryl tri[1-14C]octanoate, was enhanced five-fold by 10 mM apo CI! from human very low density lipoprotein [121]. The rate of hydrolysis increased as a function of apo CII concentration and reached a maximum at a concentration of apo Cll corresponding to a molar ratio of enzyme to apo Cll of about 1 : 1 [122]. However, these conclusions were based on the assumption that all the enzyme and apo Cii are adsorbed to the monolayer. Rietsch et al. [23] using 1,2-didecanoylglycerol monolayers, showed that the decrease in activation with surface pressure for colipase follows the decrease in the inactivation rate constant of the pancreatic lipase alone. The strong but not absolutely specific protective effect of colipase, most visible at low surface pressure, can account for the higher enzyme activity in the presence of colipase. Verger et al. [103] showed that colipase anchors lipase under circumstances where the lipase alone is unable to bind to the lipid interface (highly packed, negatively charged phospholipid monolayers). Jackson et al. [137] observed comparable behavior of milk lipoprotein lipase acting on 1,2-didecanoylglycerol monolayers. Below 15 dyn/cm the apolipoproteins apo CII , apo A I and apo CII 1 protected lipoprotein lipase non-specifically against surface denaturation. More specific effects of apo Eli were found only above 15 dyn/cm. Lag periods in lipolysis A rather common observation reported by many authors working on the kinetics of lipolytic enzymes is the presence of lag periods in the hydrolysis of both emulsions, liposomes, micelles and monolayers [78,85-96]. Brockman et al. [78] studying lipase hydrolysis of tripropionin adsorbed to siliconized glass beads, described a short lag period in the range of 15 s, and they found alinear relationship between the lag time and enzyme concentration. They suggested that this short pre-steady state is due to diffusion-controlled enzyme adsorption. Usually, however, such lag phases are much longer. The origin of the slow hydrolysis phase has R. Vergerand F. Pattus, Lipid-protein interactions in monolayers 215 not been extensively studied. Very often the interpretation is given that the products of the enzyme reaction such as lysolecithin or fatty acid in the case of phospholipase A, or phosphatidic acid in the case of phospholipase D, 'activate' the enzyme by improving the substrate dispersion or promote the enzyme substrate interaction through charge effects. The zero-order trough, consisting of two compartments connected by a narrow surface canal (see Fig. 8) gives, as expected, linear kinetics after injection of the venom enzyme under the monolayer [88]. However, as shown in Fig. 10 (left panel), this is not the case when pancreatic phospholipase A is injected under a film of dinonanoyl lecithin. In fact, one observes that the velocity given by the slope of the recorded curve increases with time and seems to approach an asymptotic limit indicated by the dashed line: the intercept between the asymptote and the time axis is the induction time, 7-. This behavior is in sharp contrast with the kinetics shown in Fig. 10. right panel, which is obtained by injection of pure phospholipase A2 from snake or bee venom under the same lecithin film. What is the reason for the lag period observed for the pancreatic enzyme? One can immediately rule out the possibility that it is due to slow mixing of the injected enzyme below the surface fdm. It has been shown that complete mixing is attained in less than 30 s. A slow diffusion of enzyme from the bulk to the surface through an unstirred layer can be excluded as well. Langmuir and Schaeffer [138] have shown that with stirring of the subphase the thickness of the unstirred layer is of the order of 0.01 ram. Using the diffusion constant [ 139] of pancreatic phospho*/. SUBSTRATEREMAINING O O O O O 0 0 D D O 0 0 O O 0 0 0 0 0 ii,';Li 80] i'I! IPANCREATIC 60i PHQSPH~IPASE J I ! i '1 I i / L 401 20] t ,, 1 i i I ,i ~000o0ooooooooooooooo 0 5 10 15 0 5 10 TIME (MIN) Fig. 10. Kinetics of the hydrolysis of a dinonanoyllecithin film upon injection of phospholipase A2 from different sources. The continuous curves are the tracings from the barostat recorder. The points and the broken lines are computed values. (From Verger et al. [88].) 216 R. Verger and F. Pattus, Lipid-protein interactions in monolayers hpase A2, 1.35 X 10 -6 cm2/s, one can estimate that under these conditions the lag period would be of the order of seconds. Secondly, the venom phospholipases which have been reported [140,141] to possess diffusion constants similar to that of the pancreatic enzyme, do not show a lag period. Finally, the hydrolysis by pancreatic phospholipase A2 of other substrates such as phosphatidylglycerol and its lysyl derivative, is not characterized by a lag period [22,88]. To explain the unusually long induction time observed during the hydrolysis of lecithin films by pancreatic enzyme, one may assume that the enzyme penetrates slowly into the monolayer. In other words, taking into account the initial model, the establishment of the equilibrium E ~ E* could be the rate-limiting step in the overall enzymic process. Once the enzyme has penetrated, the formation of the interfacial Michaelis complex and the liberation of products starts immediately. As an alternative explanation one can also imagine that the formation of this complex ( E * + S ~ E'S), or an enzyme conformational change could be ratelimiting. In order to discriminate between these two possibilities, Pattus et al. [91 ] measured simultaneously as a function of time the enzyme activity and the amount of radioactive enzyme in excess at the interface. Under conditions of short (1.4 min) or long (11 min) induction times, the amount of radioactive enzyme present in, or close to the monolayer increased continuously up to saturation value. This increase was found to be closely related to the increase in enzyme activity during the same period of time. As a consequence, the specific activity of the enzyme remained constant throughout the kinetic experiment. These data demonstrate clearly that the induction times observed during the hydrolysis of lipid monolayers by lipolytic enzymes are related to a slow enzyme penetration into the interface.* Wieloch et at. [142] studied the action of pancreatic lipase and colipase on rac-l,2-dilaurin monolayers. Biphasic kinetics were observed under conditions of high lipid packing. Similar kinetics had earlier been reported using phospholipid emulsified triolein droplets [94]. These kinetics are characterized by a lag time r d, which is dependent on the products accumulating at the substrate-water interface. This lag time is differentiated from the previously described enzyme concentration independent lag time ri, in systems of low lipid packing (Verger et at. [88]). Both r i and r d reflect a rate limiting step due to the slow enzyme penetration into the substrate interface. The variation of r d under different conditions (change in pH and concentration of Ca2+, enzyme, BSA and lipolytic products) led the authors to propose a model for product activation during lipolysis. The pancreatic lipase-colipase system can be used to probe the lipid *Phospholipase C hydrolysis of PC liposomes has an induction period which is dependent on substrate concentration i.e. surface area, as well as Ca2+ levels and pre-incubation of the enzyme (Chem. Phys. Lipids, 15 (1975) 15). Slow penetration or adsorption of the enzyme at the interface prior to hydrolysis taking place, would also be a satisfactory interpretation of the lag phase phenomenon in this system. R. Vergerand F. Pattus, Lipid-protein interactions in monolayers 217 packing of emulsified triglyceride particles and lipoproteins using rd as a reference value. Recently, Verheij et al. [ 143,144] measured the influence of surface pressure on the induction time of several phospholipases A2. These authors could classify these enzymes unambiguously according to their penetration power. For each enzyme, there is a characteristic critical packing density of the substrate molecules above which the enzyme cannot penetrate the film. This type of information can be of practical importance in choosing a lipolytic enzyme to degrade the lipid moiety of biological membranes [13], or to predict the anticoagulant properties of a given phospholipase AE [144]. Group 4. Long-chain lipids with control of surface density Rothen et al. (pers. comm.) studied the hydrolysis of long-chain phospholipid monolayers by different phospholipases A2. In both cases a large excess of serum albumin was present in the aqueous subphase in order to solubilize the products. This step was found not to be rate-limiting. The authors could obtain with natural long-chain lipids linear kinetics very similar to those described for short-chain lipids using a zero-order trough and the barostat technique [22]. As shown in Fig. 11 Rothen et al. could measure the velocity-Tr profiles for egg lecithin, PE and PG derived from egg lecithin, as well as for brain PS. However, albumin is tensio-active and prevents measurements below 22-23 dyn/cm g~ u-; >, ~ 0.5 0.4 0.3 0.2 Z~ z~ 0.1 r 10 2JO 30 40Surface Pressure (dyneslcm) Fig. 11. Influence of surface pressure on the hydrolysis of long chain phospholipids by the v); phospholipase A 2 from Vipera berus (3 pg/l): 37°C; BSA, 400 rag/l; egg yolk PG (v egg yolk PE (e el;egg yolk PC (o o);brain PS (~ ~). 218 R. Verger and F. Pattus, Lipid-protein interactions in monolayers when the barostat technique is used. Nevertheless, the upper surface pressure range is considered to be closer to the lipid packing existing in natural membranes, e.g. chylomicrons, making this method valuable for the study of quantitative lipolytic enzyme kinetics with natural lipids. Scow et al. [145] measured the action of purified lipoprotein lipase on tri[3H] oleoylglycerol and the effect of albumin on the movement of lipolytic products. The amount of trioleoylglycerol applied was 14 times that needed to cover the surface of the aqueous subphase with a monolayer. It was concluded that lipolytic products immediately located and spread throughout the interface, displacing substances with lower spreading pressures from the interface. Addition of albumin to the aqueous subphase accelerated markedly the desorption of oleic acid and mono-oleoylglycerol from the interface and thereby enhanced lipolysis. When albumin was not in contact with the site of hydrolysis, oleic acid and monooleoylglycerol readily moved along the interface to the area in contact with albumin where they were desorbed from the interface. These findings support the hypothesis that transport of lipolytic products occurs by lateral movement in cell membranes [146]. Mixed lipid monolayers as substrates Most studies of tipolytic enzyme kinetics have been done primarily in vitro with pure lipids as substrates. In reality, however, virtually all biological interfaces are composed of complex mixtures of lipids and proteins. Thus, intestinal lipolysis, blood chylomicra and the hydrolysis of lipoproteins, as well as intracellular lipolytic activities, all involve the simultaneous participation of several classes of lipids, for example glycerides, phospholipids, cholesterol derivatives, and bile salts. Studies of membrane phospholipid hydrolysis by phospholipases [147] also show this characteristic of lipid complexity. Demel et al. [13] compared the action of purified phospholipases on monomolecular films at various interfacial pressures with the action on erythrocyte membranes. The monolayer technique is ideally suited for the study of the mode of action of lipolytic enzymes at interfaces using controlled mixtures of lipids. Two types of mixed lipid monolayers can be formed at the air-water interface: either by spreading a mixture of water-insoluble lipids from a volatile organic solvent, or by injecting a detergent solution into the aqueous subphase covered with preformed lipid monolayer. Water-insoluble lipids forming a mixed monolayer In the chapter devoted to long chain lipid monolayers, we mentioned the early work of Bangham and Dawson [109], Shah and SchuUman [114], and Quarles and Dawson [87] on the influence of neutral or charged lipid additives on the hydroly- R. Vergerand F. Pattus, Lipid-protein interactions in monolayers 219 sis of phospholipid monolayers by phospholipases. One attempt to study the hydrolysis of mixed films at constant surface pressure was made by Zografi et al. [82] using a first-order trough with L- and D-dioctanoyllecithin mixtures. The authors concluded that inhibition of the phospholipase by inhibitors present in mixed films cannot be studied by the monolayer technique. Chatelain et al. [148,149] have studied the hydrolysis by phospholipases C and A2 of mixed monolayers composed of a spin probe (tempamine derivative of palmitic acid) and dipalmitoyl-DL-phosphatidylcholine spread at the air-water interface. The changes in the kinetics of enzymatic hydrolysis of the mixed monolayers confirmed the observation that phospholipase activity is strongly dependent on lipid packing. The authors concluded that the mean minimal distance separating two cluster edges seems to be a critical parameter which modulates enzymatic activity. Hirasawa et al. [95] studied the hydrolysis of phosphatidylinositol monolayers by the phosphatidylinositol phosphodiesterase of pig brain. As the monolayer pressure was increased a sharp cut off in the rate of enzymic hydrolysis occurred at 33 dyn/cm. The addition of either phosphatidic acid, phosphatidylglycerol or oleyl alcohol increased the film pressure at which cut-off occurred, as well as increasing the rate of hydrolysis at lower pressures. The rate of hydrolysis but not the cut-off pressure, was markedly increased by oleic acid and slightly increased by PE. PC, palmitoylcholine and octadecylamine decreased the cut-off pressure as well as the enzymic activity below this pressure. Stearic acid and stearyl alcohol had no effect on either the cut-off pressure or the activity. All activators decreased the length of the lag phase before enzyme activity began, and PC increased it. Clearly, all the activating substances when mixed with phosphatidyl inositol were probably introducing discontinuities into the film packing thus allowing penetration of the active center of the enzyme, even at high film pressures. These effects could be of great importance in vivo, as in addition to illustrating obvious possible regulatory mechanisms, they suggest a potential self-amplifying effect for phosphatidylinositol hydrolysis that may have physiological implications in the enhanced breakdown of phosphatidylinositol after agonist-receptor interaction [150,151]. Hendrickson et al., [151b] showed that the water insoluble amphiphile, dicetyl phosphate, inhibited the hydrolysis of didecanoylphosphatidylcholine monolayers by both pancreatic and Crotalus adamanteus phospholipases A2 below their normal cut-off pressures. Dicetyl phosphate, however, enhanced enzyme penetration and thus activated the pancreatic enzyme above its normal cut-off pressure. Because of their ready penetration into lipid interfaces, cationic surface-active local anesthetics produce a positively charged interface. Hendrickson [ 152] showed that the degree of penetration of different anesthetics into lecithin monolayers correlated with the degree of anesthetic potency. In order to completely eliminate subphase effects, mixed lecithin films containing a long-chain tetracaine analog 220 R. Verger and F. Pattus, Lipid-protein interactions in monolayers were spread over the reaction and reservoir compartments of a zero order trough [153]. The authors concluded that the inhibition of phospholipase As by the anesthetic analog and octadecylamine was due to surface charge effects alone, and not to any effects related to the structure or molecular spacing of the monolayer. Unfortunately, phospholipase action on mixed lecithin-amine monolayers necessarily resulted in an increase in the mole fraction of amine in the monolayer as lecithin was hydrolysed. Thus, the actual mole fraction of amine giving a specified percentag e inhibition might be as much as 50% greater than stated, and the induction times might be underestimated due to the increasing inhibition. Bhat and Brockman [107a] studied the enzymatic synthesis/hydrolysis of cholesteryl o r a t e in mixed surface films. Lecithin, if present in molar excess relative to the sum of free and esterified cholesterol, is inhibitory. Inhibition is associated with division of the substrate into reactive and unreactive pools which are not exchangeable. Bile salts and other surfactants reverse the inhibition presumably by disrupting the unreactive lecithin-substrate complex. Experiments were also performed using film aspiration instead of hydrophobic paper to collect the monolayer. This was done to determine if the method of recovery affected the results obtained. Within error, the authors showed that with oleic acid films the measurement of adsorbed enzyme and its specific activity are independent from the method of collection. Bhat and Brockman (Biochemistry, in press) further investigated, using initial enzyme adsorption flux, the role of oleic acid in the regulation of the hydrolysis of cholesteryl oleate in lipid films. The rate constant for the cholesterol esterase adsorption is markedly dependent upon both the concentration of oleic acid headgroups and the acyl chain packing density in the film. In contrast to adsorption, catalysis by the adsorbed enzyme is pH independent between 5.5 and 7.5. These results indicate, as in the case of pancreatic phospholipase [69], that adsorption and catalysis occur at functionally, if not physically, distant sites on the protein. The strong adsorption of the enzyme to a hydrolysis product, oleic acid constitutes a kind of product activation which operates also in the case of pancreatic lipase [94,1421. A new application of the zero-order trough was proposed by Pi~roni and Verger [154] for studying the hydrolysis of mixed-monomolecular films at constant surface density and constant lipid composition as schematically shown in Fig. 12. It was found that the amount of bound pancreatic lipase decreases linearly with increasing proportions of lecithin in mixed triglyceride-lecithin f'dms kept at constant pressure. This suggested that the penetration of mixed films by pancre~ttic lipase shows a specificity for the triglyceride; i.e. enzyme penetration into an interface is affected by 'interfacial quality'. When a pure triglyceride film was progressively diluted with lecithin, the specific activity of pancreatic lipase exhibited a bell shaped curve ; a mixed film containing only 20% trioctanoylglycerol was hydrolyzed at the same rate as a monolayer of pure triglyceride. Using the same technique as described in Fig. 12, Barenholz and Verger (unpublished experiments) have studied the hydrolysis of lecithin in lecithin- R. Vergerand F. Pattus, Lipid-protein interactions in monolayers o.. !-.!.i-!'?? ,°'oil-'', ° "°'1/" "o "°1/ © i © mU n 221 "'" "''" i m mm • atom i . • o m • m u m • "° "21""l/"""" I" • on muo m mum n n u m m nun m m u un m m m m m m 0 mmmmm mo Substrate • • m mmmmmmmm ~ non Substrate .L Soluble products Fig. 12. Principle of the method for the study of enzymic lipolysis of mixed monomolecular films. (From G. Pi6roni and R. Verger [154].) sphingomyelin mixed monomolecular films. The authors observed that the activity of the phospholipase A2 from Vipera berus venom was a function of the lecithin concentration. Sphingomyelin acted as a substrate-diluent, but did not change the activity-surface pressure profile. Similar results were obtained with pancreatic phospholipase A2 at low surface pressure (10 dyn/cm). In contrast the mixed monolayers show phase separation at higher pressures, and pancreatic phospholipase activity was 10- to 16-fold increased over the expected value predicted by substrate dilution alone. This increase was explained mainly by a larger amount of enzyme that penetrated the interface. It seems that the packing defects which occur in membranes during phase transitions and phase separations give rise to changes in interactions with proteins including lipolytic enzymes [93,142,155-161]. Partly water-soluble amphiphiles forming a mixed monolayer In this paragraph we will .only consider the situation in which a a partly water-soluble lipid is present in the aqueous subphase over water-insoluble lipid monolayer has been spread. This situation is related to certain conditions prevailing in vivo (e.g. gastrointestinal drug administration). detergent, or which a pure more closely tract or after 222 R. Verger and F. Pattus, Lipid-protein interactions in monolayers Hendrickson [152] first reported the penetration of local anesthetics into lecithin monolayers, and further studied the effects of surface and subphase concentration of the anesthetic on the inhibition of pancreatic phospholipase A2 action on didecanoylphosphatidylcholine monolayers [162]. It was proposed that the most surface-active anesthetics inhibit by surface effects, while the less surfaceactive anesthetics (e.g. lidocaine and procaine) inhibit by interaction with the enzyme in the subphase, thus preventing enzyme penetration of the monolayer interface. Some studies have shown that various drugs may produce an abnormal accumulation of lipid material in cells of many tissues. These lipidoses are due either to an inhibition of phospholipid degradation, or to a stimulation of biosynthesis. Chatelain et al. [163] and Defrise Quertain et al. [164] studied by surface pressure and surface radioactivity measurements two possible modes of action of diazafluoranthen and some amino-piperazine derivatives, involving either the direct inactivation of the phospholipases or the formation of a drug-phospholipid complex which inhibited the action of the enzymes. The capacity to inhibit enzymatic activity depends on the amphiphilic balance of the drug. The diazafluoranthen derivative forms a very stable drug-lipid complex and inhibits completely the hydrolysis of dipalmitoyl-DL-phosphatidylcholine monolayers by phospholipase A2. Because the aminopiperazine derivatives are more water-soluble, the complex with the phospholipid monolayer is less stable. In this case only a delay in the lipid hydrolysis was observed. Momsen et al. [ 165] determined surface pressure-area isotherms for 1,3-dodecanoylglycerol as a function of the concentration of taurodeoxycholate in the subphase. Subjecting the monolayers to hydrolysis by pancreatic lipase yielded kinetic data which, together with physical studies, supported a model based on the glycerol moieties of the diglyceride molecules undergoing discrete changes of state. The more condensed state exhibits an area of 38 A:/diglyceride molecule and is hydrolyzed at a rate proportional to its concentration in the monolayer. Taurodeoxycholate at 0.05-0.6 mM shifts the apparent area of the expanded state to 360 A2/diglyceride molecule. Lairon et al. [166] reported a new technique for following the lipolysis of monomolecular films in the presence of bile salts by using a 'zero order' trough. Under these conditions pancreatic lipase hydrolysis of rac-l,2-didodecanoylglycerol monolayers, maintained at constant surface pressure in the presence of sodium taurodeoxycholate in the subphase, can be followed without any modification of the film composition with time. This technique was used to measure simultaneously both the enzyme activity and adsorption onto a lipid substrate under conditions of variable, but controlled 'interfacial quality'. The authors utilized this method to obtain precise lipolysis kinetics under conditions closer to those prevailing in the intestinal lumen during fat digestion. The results obtained are compatible with the model previously presented [167] in which R. Verger and F. Pattus, Lipid-protein interactions in monolayers 223 pancreatic colipase would be required for the formation of the lipase-bile lipoprotein complex, and bile lipids would be required to direct the adsorption of this lipolytic entity towards the emulsified substrate. Using the same technique, Wieloch et al. [168] studied the interactions o f colipase and its pro-form, and their lipase-activating properties, with a monomolecular rac-l,2-didodecanoylglycerol film in the presence of taurodeoxycholate. In this system at high surface pressure the authors observed an all-or-none effect on lipase activation by colipase and procolipase, respectively. The results support the idea that trypsin activation o f procolipase yields a colipase molecule with better lipid-binding properties [ 169]. Perspectives It is probable that future work using monomolecular films as substrates for lipolytic enzymes will develop in the following directions: (i) Spreading of natural membranes at the air-water interface as the substrate for soluble lipolytic enzymes. (ii) Studies of the synergism between lipolytic enzymes using mixing lipid films and the 'zero order' trough. (iii) Studies on membrane-bound lipolytic enzymes and lipid in a membranous continuum. (iv) Possible implication of lipolytic enzymes in membranous fusion processes. Acknowledgements Thanks are due to Dr. C. Rothen (Bern, Switzerland) and Dr. Y. Barenholz (Jerusalem, Israel) for personal communications prior to publication. Professor G.H. de Haas (Utrecht University, The Netherlands) kindly gave permission to reproduce original material from the work cited in reference [69]. The authors are grateful to Mme M. Th. Nicolas for typing the manuscript. References 1 A.D. Bangham, M.M. Standish and J.C. Watkins, J. Mol. Biol., 13 (1965) 238. 2 P. MueUer, D.O. Rudin, H.T. Tien and W.C. Wescott, Nature, 194 (1962) 979.. 3 D.A. Cadenhead, in: S. Abrahamson and I. Paschle (Eds.), Structure of Biological Membranes, Plenum Press, New York, 1977, p. 63. 4 F. Milller-Landau and D.A. Cadenhead, Chem. Phys. Lipids, 25 (1979) 299, 315, 329. 5 D.E. Graham and M.C. Phillips, J. Colloid Interface Sci., 70 (1979) 403; 76 (1980) 227. 6 B.W. Shen and A.M, Scanu, Biochemistry, 19 (1980) 3643. 7 P.J. Quinn and R.M.C. Dawson, Biochem. J., 116 (1970) 671. 224 8 9 10 11 12 13 R. Verger and F. Pattus, Lipid-protein interactions in monolayers G. Colacicco, Lipids, 5 (1970) 636. G. Blobel, B. Dobberstein, J. Cell. Biol., 67 (1975) 835. W. Wickner, Annu. Rev. Biochem., 48 (1979) 23. M.C. Phillips, Ann. N.Y. Acad. Sci., 348 (1980) 138. A. Blume, Biochim. Biophys. Acta, 557 (1979) 32. R.A. Demel, W.S.M. Geurts van Kessel , R.F.A. Zwaal, B. Roelofsen and L.L.M. van Deenen, Biochim. Biophys. Acta, 406 (1975) 97. 14 L.K. James and L.G. Augenstein, Adv. Enzymol., 28 (1966) 1. 15 I.R. Miller, in: J.F. DanieUi, M.D. Rosenberg and D.A. Cadenhead (Eds.), Progress in Surface and Membrane Science, Vol. 4, Academic Press, New York, 1971, p. 299 16 M.C. Phillips, M.T.A. Evans and H. Hauser, Adv. Chem. SeE, 144 (1975) 217. 17 F. Macrichtie, Adv. Protein Chem., 32 (1978) 283. 18 G.L. Gaines, Jr., in: I. Prigogine (Ed.), Insoluble Monolayers at Liqid-Gas Interfaces, Intersciences Publishers, New-York, 1966. 19 N. Gershfeld, in: E.D. Korn (Ed.), Methods in Membrane Biology, Vol. 1, Plenum Press, New York, 1974, p. 69. 20 L.I. Rothfield and V.A. Fried, in: E.D. Korn (Ed.), Methods in Membrane Biology, Vol. 4, Plenum Press, New York, 1975, p. 277. 21 P. Frornherz, Rev. Sci. Instrum., 46 (1975) 1380. 22 R. Verger and G.H. de Haas, Chem. Phys. Lipids, 10 (1973) 127. 23 J. Rietsch, F. Pattus, P. Desnuelle and R. Verger, J. Biol. Chem., 252 (1977) 4313. 24 P. Fromherz, Biochim. Biophys. Acta, 25 (1971) 382. 25 M. Maramatsu, in: E. Matigevic (Ed.), Surface and Colloid Science, Vol. 6, Wiley New York, 1973,p. 101. 26 M.A. Frommer and I.R. Miller, J. Colloid Interface Sci., 21 (1966) 245. 27 J. Tessi6, J.F. Tocanne and A. Baudras, FEBS Lett., 70 (1976) 123. 28 F. Kopp, P.A. Cuendet, K. Mtihlethaler and H. Zuber, Biochim. Biophys. Acta, 553 (1979) 438. 29 R.E. Hirsch and S.S. Brody, Eur. J. Biochem., 89 (1978) 281. 30 R.G. Cornell, J. Colloid Interface Sck, 70 (1979) 167. 31 H. Kuhn, D. Moebius and H. Biicher, in: A. Weissberger and B. Rassiter (Eds.), Physical Methods of Chemistry, Vol. 1, Pt. 313, J. Wiley, New York, 1972, p. 577. 32 K.B. Blodgett, J. Am. Chem. Soc., 57 (1935) 1007. 33 K.B. Blodgett and I. Langmuir, Physiol. Rev., 51 (1937) 964. 34 H.E. Ries and W.A. Kimball, J. Phys. Chem., 59 (1955) 94. 35 J.A. Spink, J. Colloid Interface Sci., 23 (1967) 9. 36 E.J.F. van Zoelen, R.F.A. Zwaal, F.A.M. Reuvers, R.A. Demel and L.L.M. van Deenen, Biochim. Biophys. Acta, 464 (1977) 482. 37 L. Ter-Minassian-Saraga, J. Colloid Interface Sci., 70 (1979) 245. 38 P.J. Quinn and R.M.C. Dawson, Biochem. J., 113 (1969) 791. 39 M.C. Phillips and C.E. Sparks, Ann. N.Y. Acad. Sci., 348 (1980) 122. 40 R.L. Jackson, F. Pattus and R.A. Demel, Biochim. Biophys. Acta, 556 (1979) 369. 41 R.A. Demel, K.W.A. Wirtz and L.L.M. van Deenen, Nature, 246 (1973) 102. 42 M.F. Lecompte, I.R. Miller, J. Elion and R. Benarous, Biochemistry, 19 (1980) 3434. 43 O.S. Ksenzhek, V.S. Gevod, A.M. Omel'Chenko, S.N. Semenov, A.I. Sotnichenko and A.I. Miroshnikov, Mol. Biol. (Moscow), 12 (1978) 1057. 44 P. Bougis, H. Rochat, G. Pi6roni and R. Verger, Biochemistry, 20 (1981) 4915. 44bJ. Tessi6, Biochemistry, 20 (1981) 1554. 45 D. Romeo, A. Girard and T. Rothfield, J. Mol. Biol., 53 (1970) 475. 46 Y. London, R.A. Demel, W.S.M. Geurts van Kessel, F.G.A. Vassenberg and L.L.M. van Deenen, Biochim. Biophys. Acta, 311 (1973) 507. 47 J. Dufourcq and J.F. Faucon, Biochemistry, 17 (1978) 1170. tL Vekgeiahd F_-Pattus, Lipid-protein interactions in monolayers 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 225 S.B. Hwang, J.l. Korenbrot and W. Stockenius, J. Membr. Biol., 36 (1977) 36. A. Darszon, M. Philipp, J. Zarco and M. Montal, J. Membr. Biol., 43 (1978) 71. M. Montal, A. Darszon and H. Schindler, Q. Rev. Biophys., 14 (1981) 1. R. Verger and F. Pattus, Chem. Phys. Lipids, 16 (1976) 285. H.J. Trurnit, J. Colloid Interface Sci., 15 (1960) 1. J.l. Korenbrot and M.J. Pramik, J. Membr. Biol., 37 (1977) 235. F. Pattus, P. Desnuelle and R. Verger, Biochim. Biophys. Acta, 507 (1978) 62. F. Pattus, M.C.L. Piovant, C.J. Lazdumki, P. DesnueUe and R. Verger, Biochim. Biophys. Acta, 507 (1978) 71. F. Pattus, C. Rothen, M. Streit and P. Zahler, Biochim. Biophys. Acta, 647 (1981) 29. H. Schindler, Biochim. Biophys. Acta, 555 (1979) 316. D. Louvard, S. Maroux and P. Desnuelle, Biochim. Biophys. Acta, 389 (1975) 389. F. Pattus, R. Verger and P. Desnuelle, Biochem. Biophys. Res. Commun., 69 (1976) 718. H.W. Trissl and P. Gr~iber, Bioelectrochem. Bioenerg., 7 (1980) 167. I. Langmuir and D.F. Waugh, J. Gen. Physiol., 21 (1938) 745, M. Takagi, K. Azuma and V. Kishimoto, Annu. Rep. Biol. Works, Fac. Sci., Osaka Univ., 13 (1967) 107. M. Montal and P. Mueller, Proc. Natl. Acad. Sci. U.S.A., 69 (1972) 3561. H. Schindler and J. Rosenbush, Proc. Natl. Acad. Sci. U.S.A., 75 (1978) 3751. H. Schindler, FEBS Lett., 122 (1980) 77. H. Schindler and U. Quast, Proc. Natl. Acad. Sci. U.S.A., 77 (1980) 3052. J.J. Korenbrot and S.B. Hwang, J. Gen. Physiol., 76 (1980) 649. H. Brockerhoff and R.G. Jensen, Lipolytic Enzymes, Academic Press, New York, 1974. R. Verger and G.H. de Haas, Annu. Rev. Biophys. Bioeng., 5 (1976) 77. B. Borgstr6m, C. Erlanson-Albertsson and T. Wieloch, J. Lipid Res., 20 (1979) 805. M. S6m~riva and P. Desnuelle, Adv. Enzymol, 48 (1979) 319. R. Verger, Methods Enzymol., 64B (1980) 340. E.A. Dennis, P.L. Darke, R.A. Deems, C.R. Kensil and A. PlOckthun, Mol. Cell. Biochem., 36 (1981) 37. A.J. Slotboom, H.M. Verheij and G.H. de Haas, in: M. Florkin and E.H. Stotz (Eds.), Comprehensive Biochemistry, Elsevier, Amsterdam, 1982. K. Holwerda, P.E. Verkade and A.H.A. de Wilhingen, Red. Trav. Chim. Pays-Bas, 55 (1936) 43. F. SchCnheyder and K. Volqvartz, Acta Physiol. Scand., 9 (1945) 57. L. Sarda and P. Desnuelle, Biochim. Biophys. Acta, 30 (1958) 513. H.L. Brockman, J.H. Law and F.J. K~zdy, J. Biol. Chem., 248 (1973) 4965. W.A. Pieterson, J.C. Vidal, J.J. Volwerk and G.H. de Haas, Biochemistry, 13 (1974) 1455. R.L. Misiorowski and M.A. Wells, Biochemistry, 13 (1974) 4921. G.H. de Haas, P.P.M. Bonsen, W.A. Pieterson and L.L.M. Van Deenen, Biochim. Biophys. Acta, 239 (1971) 252. G. Zografi, R. Verger and G.H. de Haas, Chem. Phys. Lipids, 7 (1971) 185. N.K. Adam, The Physics and Chemistry of Surfaces. Dover, New York, 1968. J.T. Davis and E. Rideal, Interfacial Phenomena, Academic Press, New York, 1963. O.A. Roholt and M. Schlamowitz, Arch. Biochem. Biophys., 94 (1961) 364. A.F. Rosenthal and M. Pousada, Biochim. Biophys. Acta, 164 (1968) 226. R.H. Quarles and R.M.C. Dawson, Biochem. J., 113 (1969) 697. R. Verger, M.C.E. Mieras and G.H. de Haas, J. Biol. Chem., 248 (1973) 4023. M.K. Jain and R.C. Apitz-Castro, J. Biol. Chem., 253 (1978) 7005. D.O. Tinker and J. Wet, Can. J. Biochem., 57 (1979) 97. F. Pattus, A.J. Slotboom and G.H. de Haas, Biochemistry, 13 (1979) 2691. G.C. Upreti, S. Rainier and M.K. Jain, J. Membr. Biol., 55 (1980) 97. 226 93 94 95 96 R. Vergerand F. Pattus, Lipid-protein interactions in monolayers G.C. Upreti and M.K. Jain, J. Membr. Biol., 55 (1980) 113. B. Borgstr6m, Gastroenterology, 78 (1980) 954. K. Hirasawa, R.F. Irvine and R.M.C. Dawson, Biochem. J., 193 (1981) 607, M. Menashe, D. Lichtenberg, C. Gutierrez-Merino and R.L. Biltonen, J. Biol. Chem., 256 (1981) 4541. 97 A.D. McLaren and L. Packer, Adv. Enzymol., 33 (1970) 245. 98 I.R. Miller and J.M. Ruysschaert, J. Colloid Interface Sci., 35 (1971) 340. 99 H. Brockerhoff, J. Biol. Chem., 246 (1971) 5828. 100 A. Vandermeers, M.C. Vandermeers-Piret, J. Rath~ and J. Christophe, Biochim. Biophys. Acta, 370 (1974) 257. 101 H. Cohen, B.W. Shen, W.R. Snyder, J,H. Law and F.J. K~zdy, J. Colloid Interface Sci., 56 (1976) 240. 102 W.E. Momsen and H.L. Brockman, J. Biol. Chem., 251 (1976) 378. 103 R. Verger, J. Rietsch and P. Desnuelle, J. Biol. Chem,, 252 (1977) 4319. 104 P. Canioni, R. Julien, J. Rathelot and L. Sarda, Lipids, 12 (1977) 393. 105 S.R. Turner, M. Litt and W.S. Lynn, J. Colloid Interface Sci., 48 (1974) 100. 106 J.L. Brash and D.J. Lyman, The Chemistry of Biosurfaces, Dekker, New York, 1971. 107a S.G. Bhat and H,L. Brockman, J. Biol. Chem., 256 (1981) 3017. 107bW.E. Momsen and H.L. Brockman, J. Biol. Chem., 256 (1981) 6913. 108 A. Hughes, Biochem. J., 29 (1935) 437. 109 A.D. Bangham and R.M.C. Dawson, Biochem, J., 75 (1960) 133. 110 R.M.C. Dawson, Biochem. J., 98 (1966) 53. 111 P.J. Quinn and Y. Barenholz, Biochem. J., 149 (1975) 199. 112 R.M.C. Dawson, Methods Enzymol., 14 (1969) 633. 113 G. Colacicco and M,M. Rapport, J, Lipid Res., 7 (1966) 258. 114 D.O. Shah and J.H. Schulman, J. Colloid Interface Sci., 25 (1967) 107. 115 G. Colacicco, Nature (London), 233 (1971) 202. 116 J. Olive, Contribution ~ l'Enzymologie Superficielle des Lipides, Ph.D. Thesis, University of Paris, 1969. 117 J.W. Lagocki, N.D. Boyd, J.H. Law and F.J. K~zdy, J. Am. Chem. Soc., 92 (1970) 2923. 118 C.W. Garner and L.C. Smith, Biochem. Biophys. Res. Commun., 39 (1970) 672. 119 D.G. Dervichian, Biochimie, 53 (1971) 25. 120 L.C. Smith, J. Lipid Res., 13 (1972) 769. 121 A.L. Miller and L.C. Smith, J. Biol. Chem., 248 (1973) 3359. 122 J. Chung and A.M. Scanu, J. Biol. Chem., 252 (1977) 4202. 123 L.C. Smith, J.C. Voyta, A.L. Catapano, P.K.J. Kinnunen, A.M. Gotto, Jr. and J,T. Sparrow, Ann. N.Y. Acad. Sci., 348 (1980) 213. 124 M. Derbesy and M. Naudet, Rev. Fr. Corps Gras, 4 (1972) 225. 125 J.W. Lagocki, J.H. Law and F.J. K~zdy, J. Biol. Chem., 248 (1973) 580. 126 B. Entressangles and P. DesnueUe, Biochim. Biophys. Acta, 341 (1974) 437. 127 M.A. Wells, Biochemistry, 13 (1974) 2248. 128 H.L. Brockman, F.J. K~zdy and J.H. Law, J. Lipid Res., 16 (1975) 67. 129 D.G. Dervichian and J.P. Barque, J. Lipid Res., 20 (1979) 437. 130 J.P. Barque and D.G. Dervichian, J. Lipid Res., 20 (1979) 447. 131 J.P. Barque and D.G. Dervichian, J. Lipid Res., 20 (1979) 599. 132 J. Olive and D.G. Dervichian, Biochimie, 53 (1971) 207. 133 S. Esposito, M. S6m6riva and P. Desnuelle, Biochim. Biophys. Acta, 302 (1973) 293. 134 R. Verger, J. Rietsch, M.C.E. Van Dam-Mieras and G.H. de Haas, J. Biol. Chem., 251 (1976) 3128. 135 B.W. Shen, F.H.C. Tsao, J.H. Law and F.J. K6zdy, J. Am. Chem. Soc., 97 (1975) 1205. 136 R. Verger, J. Rietsch, F. Pattus, F. Ferrato, G. Pi6roni, G.H. de Haas and P. DesnueUe, Enzymes of Lipid Metabolism, Plenum Pubhshing Corporation, New York, 1978, p. 79. R. Vergerand F. Pattu's, Lipid-protein interactions in monolayers 227 137 R.L. Jackson, F. Pattus and G.H. de Haas, Biochemistry, 2 (1980) 373. t38 1. Langmuir and V.F. Schaeffer, J. Am. Chem. Soc., 59 (1937) 2400. 139 G.H. De Haas, N.M. Postema, W. Nieuwenhuizen and LL.M. Van Deenen, Biochim. Biophys. Acta, 159 (1968) 103. 140 M.A. Wells and D.J. Hanahan, Biochemistry, 8 (1969) 414. 141 R.A. Shipolini, G.L. Callewaert, R.C. Cottrell, S. Doonan, C.A. Vernon and B.E.C. Banks, Eur. J. Biochem., 20 (1971) 459. 142 T. Wieloch, B. Borgstr6m, G. Pi6roni, F. Pattus and R. Verger, J. Biol. Chem., (1982) in press. 143 H.M. Verheij, M.R. Egmond and G.H. de Haas, Biochemistry, 20 (1981) 94. 144 H.M. Verheij, M.C. Boffa, C. Rothen, I~..C. Brijkaert, R. Verger and G.H. de Haas, Eur. J. Biochem., 112 (1980) 25. 145 R.O. Scow, P. Desnuelle and R. Verger, J. Biol. Chem., 254 (1979) 6456. 146 R.O. Scow, E.J. Blanchette-Mackie and L.C. Smith, Circ. Res., 39 (1976) 149. 147 B. Roelofsen and R.F.A. Zwaal, Meth. Membr. Biol., 7 (1976) 147. 148 P. Chatelain, F. Defrise-Quertain and J.M. Ruysschaert, J. Colloid Interface Sci., 72 (1979) 287. 149 P. Chatelain, R. Brasseur, J.M. Ruysschaert and M. Delmelle, J. Colloid Interface Sci., 72 (1979) 294. 150 R.F. Irvine and R.M.C. Dawson, Biochem. Soc. Trans., 8 (1980) 27. 151 R.F. Irvine and R.M.C. Dawson, Biochem. Soc. Trans., 8 (1980) 376. 15 lb H.S. Hendrickson, W.M. Trygstad, T.L. Loftness and S.L. Sailer, Arch. Biochem. Biophys., 212 (1981) 508. 152 H.S. Hendrickson, J. Lipid Res., 17 (1976) 393. 153 C. Willman and H.S. Hendrickson, Arch. Biochem. Biophys., 191 (1978) 298. 154 G. Pi6roni and R. Verger, J. Biol. Chem., 254 (1979) 10090. 155 J.A.F. Op den Kamp, J. de Gier and L.L.M. Van Deenen, Biochim. Biophys. Acta, 345 (1974) 253. 156 J.A.F. Op den Kamp, M.T. Kauerz and L.L.M. Van Deenen, Biochim. Biophys. Acta, 406 (1975) 169. 157 P.N. Strong and R.B. Kelly, Biochim. Biophys. Acta, 469 (1977) 231. 158 D.O. Tinker, A.D. Purdon, J. Wei and E. Mason,Can J. Biochem., 56 (1978) 552. 159 J.C. Wilschut, J. Regts, H. Westenberg and G. Scherphof, Biochim. Biophys. Acta, 508 (1978) 185. 160 R. Cohen and Y. Barenholz, Biochim. Biophys. Acta, 509 (1978) 181. 161 R. Kensil and E.A. Dennis, J. Biol. Chem., 254 (1979) 5843. 162 H.S. Hendrickson and M.C.E. van Dam-Mieras, J. Lipid Res., 17 (1976) 399. 163 P. Chatelain, C. Berliner, J.M. Ruysschaert and J. Jaffa, Biochim. Biophys. Acta, 419 (1976) 540. 164 F. Defrise-Quertain, P. Chatelain and J.M. Ruysschaert, J. Pharm. Pharmacol., 30 (1978) 608. 165 W.E. Momsen, J.M. Smaby and H.L. Brockman, J. Biol. Chem., 254 (1979) 8855. 166 D. Lairon, M. Charbonnier-Augeire, G. Nalbone, J. L6onardi, J.C. Hauton, G. Pi6roni, F. Ferrato and R. Verger, Biochim. Biophys. Acta, 618 (1980) 106. 167 D. Lairon, G. Nalbone, H. Lafont, J. L6onardi, N. Domingo, J.C. Hauton and R. Verger, Biochemistry, 17 (1978) 5 263. 168 T. Wielcoh, B. Borgstrt~m, G. Pi6roni, F. Pattus and R. Verger, FEBS Lett., 128 (1981) 217. 169 b. Borgstr6m, T. Wieloch and C. Erlanson-Albertsson, FEBS Lett., 108 (1979) 407.