Water Interface and Its

Nov 1, 1996 - a Millipore Q purification system (Millipore Corporation, France). The purified ... plate method,20 using a filter paper (20 × 15 mm2) ...
0 downloads 0 Views 445KB Size
5856

Langmuir 1996, 12, 5856-5862

Bovine Milk Sphingomyelin at the Air/Water Interface and Its Interaction with Xanthine Oxidase Dorthe Kristensen,† Tommy Nylander,*,‡,§ Jan Trige Rasmussen,| Marie Paulsson,§ and Anders Carlsson⊥ Department of Analytical and Pharmaceutical Chemistry, Royal Danish School of Pharmacy, Universitetsparken 2, DK-2100 Copenhagen Ø, Denmark, Physical Chemistry 1, and Food Technology, University of Lund, P.O. Box 124, S-22100 Lund, Sweden, Protein Chemistry Laboratory, Department of Molecular Biology, University of Aarhus, The Science Park, Gustavs Wieds Vej 10, DK-8000 Aarhus C, Denmark, and Scotia LipidTeknik AB, P.O. Box 6686, S-113 84 Stockholm, Sweden Received March 18, 1996. In Final Form: September 3, 1996X The monolayer properties of two different sphingomyelin samples from bovine milk (one pure with respect to sphingomyelin (>99%) and one sphingolipid fraction containing 68% sphingomyelin) at the air/water interface and the interaction between the lipid monolayer and xanthine oxidase (XO) from the milk fat globule membrane were investigated by the film balance technique. For comparison, similar measurements were performed on distearoylphosphatidylcholine (DSPC) monolayers. Bovine milk sphingomyelin formed monolayers comparable with those of other natural sphingomyelins. A liquidcondensed phase transition at 20-22 mN/m reflected the high amount of long saturated fatty acids, as in the case for monolayers of egg sphingomyelin. Monolayers of bovine milk sphingomyelin were metastable, as opposed to those of DSPC. The observed decrease in surface area with time at constant pressure indicated that dissolution/expulsion into the subphase took place. A pure sphingomyelin monolayer was significantly more stable at a surface pressure of 10 mN/m than at 20 mN/m. The discontinuity in the surface area change versus time and the low solubility of the lipid show that the instability cannot be explained by simple desorption of lipid monomers. The presence of XO in the subphase increased the maximal surface pressure at an area per molecule of sphingomyelin of 30 Å2 (maximal compression) by 15 mN/m for the sphingolipid sample and 20 mN/m for pure sphingomyelin, indicating a stabilization of sphingomyelin monolayers in the presence of XO at high surface pressure. This effect was not observed for DSPC monolayers, which suggests a specific interaction between sphingomyelin and XO.

Introduction There is an increasing interest in the use of natural sphingomyelins (N-acylsphingosine-1-phosphocholines) in pharmaceutical and cosmetic applications. The low level of unsaturation has been suggested to make sphingomyelins less susceptible for electrophilic attack compared to phosphoroglycerides. Sphingomyelin is a precursor to the ceramides, which are found in large amounts in the human skin surface epidermal cells. Hence, one of the major commercial potentials of sphingomyelins is believed to be within cosmetics.1 Milk is a convenient source of sphingomyelins, where they comprise 24% of the phospholipids in the milk fat globule membrane.2 Sphingomyelins are one of the major groups of polar lipids in the plasma membranes of mammalian cells. Sphingomyelins are composed of a phosphorylcholine headgroup, a sphingoid long-chain base, and a fatty acid linked to an amide nitrogen of the base. The most common base is sphingosine, a C18 amine diol with a double bond between carbon number 4 and 5. The fatty acid composition depends on the origin, but the majority of the naturally occurring sphingomyelins contain * To whom correspondence should be addressed at Physical Chemistry 1, Lund University. † Royal Danish School of Pharmacy. ‡ Physical Chemistry 1, University of Lund. § Food Technology, University of Lund. | University of Aarhus. ⊥ Scotia Lipidteknik AB. X Abstract published in Advance ACS Abstracts, November 1, 1996. (1) Holleran, W. M. in Advances in Lipid Research; Elias, P. M., Ed.; Academic Press: San Diego, CA, 1991; Vol. 24, pp 119-139. (2) Christie, W. W.; Noble, R. C.; Davies, G. J. Soc. Dairy Technol. 1987, 40, 10-12.

S0743-7463(96)00259-4 CCC: $12.00

saturated chains in the range C14 to C24.3 Sphingomyelins resemble the common membrane lipid phosphatidylcholine (PC), having the same phosphorylcholine polar headgroup. Both form lamellar phases spontaneously on addition of water. The phase behavior and the properties of liposomes and emulsions of bovine milk sphingomyelin have been investigated and compared with those of the corresponding phosphatidylcholine systems.4 In a recent study, the incorporation of lysozyme, β-lactoglobulin, and R-lactalbumin in the gel phase composed of sphingomyelin from bovine milk and solubilized sodium palmitate was investigated.5 Xanthine oxidoreductase occurs in mammal tissue and fluids in two interconvertible forms, as a dehydrogenase (EC 1.1.1.204) (XD) or as an oxidase (EC 1.1.3.22) (XO) and is one of the enzymes involved in purine degradation.6 The enzyme in its active form is a dimer of two identical subunits, each with a molecular weight of approximately 148 000 g/mol. XO as a monomer has four redox centers in the active site: molybdenum, two 2Fe/2S sites, and a FAD center.7 Milk is one of the richest sources of XO mainly associated to the milk fat globule membrane, representing approximately 10% of the total protein components in the membrane.8 The function of XO in milk, especially the occurrence in such a large amount associated with the milk fat globule (3) Barenholz, Y.; Thompson, T. E. Biochem. Biophys. Acta 1980, 604, 129-158. (4) Malmsten, M.; Bergenståhl, B.; Nyberg, L.; Odham, G. J. Am. Oil Chem. Soc. 1994, 71, 1021-1026. (5) Minami, H.; Nylander, T.; Carlsson, A.; Larsson, K. Chem. Phys. Lipids 1996. 79, 65-70. (6) Parks, D. A.; Granger, D. N. Acta Physiol. Scand. 1986, Suppl. 548, 87-99. (7) Hille, R.; Massey, V. In Molybdenum Enzymes; Spiro, T. G., Ed.; Wiley & Sons: New York, 1985; pp 443-518. (8) Briley, M. S.; Eisenthal, R. Biochem. J. 1975, 147, 417-423.

© 1996 American Chemical Society

Bovine Milk Sphingomyelin and Xanthine Oxidase

membrane, is still unclear.9,10 XO is most likely associated with butyrophilin, a major polypeptide in the membrane. They can only be separated from each other under reducing conditions.11 Butyrophilin is an acidic glycoprotein specific to mammary glands. It is an intrinsic membrane protein with a single membrane-spanning domain.12 A putative complex between butyrophilin and XO constitutes a major portion of the dense 10-50 nm thick coat structure sandwiched between the milk fat globule membrane and the outer shell of the fat droplet.13 Recently it has been shown that these two proteins occur in constant molar proportions in the membrane, independent of the fat content of the milk and of the diameter of the fat globules, but vary in amount with breed and stage of lactation.14 These results suggest that these proteins alone may not solely be responsible for anchoring the membrane to the fat globule surface. The present study was undertaken to investigate the interaction between XO and sphingomyelin preparations from milk at the air/water interface. Numerous studies on the interaction of proteins with lipids have been published.15,16 In order to study lipid-protein interactions, spread monolayers of the lipids have often been used.17-19 For this purpose the monolayer properties of sphingomyelin preparations as well as the interfacial behavior of XO were investigated. The interaction was studied by the film balance technique, as was the effect of the presence of XO in the subphase on the Π-A isotherms of sphingomyelin. For comparison, measurements were also performed on monolayers of phosphatidylcholine, which is another constituent in the milk fat globule membrane. Experimental Section Materials. A sphingolipid fraction prepared from bovine milk fat (batch T40706) was obtained from Scotia LipidTeknik AB, Stockholm, Sweden. This sample was composed of 94% polar lipids: sphingomyelin (68%), monohexosylceramide (7%), dihexosylceramide (8%), and phosphatidylcholine (11%). Neutral lipids and lactose constituted the rest (6%). The fatty acid composition was dominated by C16:0 (29%), C22:0 (11%), C23:0 (18%), and C24:0 (10%). Sphingomyelin from bovine milk (batch T50619) with a purity of >99% was obtained from Scotia LipidTeknik AB, Stockholm, Sweden. The major fatty acids were C16:0 (15.7%), C22:0 (19.4%), C23:0 (31.2%), and C24:0 (18.8%). The fatty acid composition was analyzed by Scotia LipidTeknik AB. The molecular weights of the sphingolipid fraction and bovine sphingomyelin calculated from the fatty acid composition are 8435 and 782 g/mol, respectively. Synthetically prepared distearoylphosphatidylcholine (DSPC) (Lot B9537) was donated by Genzyme Pharmaceuticals and Fine Chemicals, Haverhill, U.K. Chromatographic column materials were purchased from Pharmacia, Uppsala, Sweden. A 40 mM phosphate buffer (pH 7.4) containing 0.1 M sodium chloride (PBS-buffer, ionic strength ) 0.2) was used unless stated otherwise. (9) McPherson, A. V.; Kitchen, B. J. J. Dairy Res. 1983, 50, 107-133. (10) Keenan, T. W.; Dylewski, D. P. In Advanced Dairy Chemistry, Vol. 2: Lipids; Fox, P. F., Ed.; Chapman & Hall: London, 1995; pp 89-130. (11) Valivullah, H. M.; Keenan, T. W. Int. J. Biochem. 1989, 21, 103107. (12) Jack, L. J. W.; Mather, I. H. J. Biol. Chem. 1990, 265, 1448114486. (13) Freudenstein, C.; Keenan, T. W.; Eigel, W. N.; Sasaki, M.; Stadler, J.; Franke, W. W. Exp. Cell Res. 1979, 118, 277-294. (14) Mondy, B. L.; Keenan, T. W. Protoplasma 1993, 177, 32-36. (15) Reynolds, J. A. In Lipid-Protein Interactions; Jost, P. C., Griffith, O. H., Eds.; John Wiley & Sons: New York, 1982; Vol. 2, pp 193-224. (16) Nylander, T.; Ericsson, B. In Food Emulsions; Friberg, S., Larsson, K., Eds.; Marcel Dekker: New York, in press. (17) Colacicco, G. J. Colloid Interface. Sci. 1969, 29, 345-364. (18) Verger, R.; Pattus, F. Chem. Phys. Lipids 1982, 30, 189-227. (19) Ahlers, M.; Mu¨ller, W.; Reichert, A.; Ringsdorf, H.; Venzmer, J. Angew. Chem., Int. Ed. Engl. 1990, 29, 1269-1285.

Langmuir, Vol. 12, No. 24, 1996 5857 The water used was deionized, distilled, and passed through a Millipore Q purification system (Millipore Corporation, France). The purified water had a resistance > 18 MΩ. All other chemicals were Merck or Sigma products of analytical grade and used without further purification. Methods. A KSV 5000 LB film balance system (KSV Chemicals, Helsinki, Finland) was used to measure the surface pressure-area (Π-A) isotherms and surface pressure versus time (Π-t). The surface tension was measured by the Wilhelmy plate method,20 using a filter paper (20 × 15 mm2) to ensure complete wetting even at the high surface pressure of the lipid monolayer. Prior to the measurements the paper was prewetted with the subphase buffer for at least 30 min. Dynamic surface pressure isotherms (Π-A) were recorded as follows: Distilled water or PBS-buffer was poured into a rectangular Teflon trough. The maximum area was 22500 mm2 (50 × 450 mm2). The surface was cleaned by sweeping it with the Teflon barrier, and surface active contaminants were removed by suction (aspiration) of the interface. The sphingolipid fraction and sphingomyelin (25 µg dissolved in 50 µL of spreading solvent) were spread on the surface using a Hamilton microsyringe, from a chloroform/methanol (2:1, v/v) or a hexane/ethanol (9:1, v/v) solution. DSPC was spread from a chloroform/methanol (2:1, v/v) solution. The solvent was allowed to evaporate for 10 min before compression/expansion of the film was initiated. Three different barrier speeds were used: 12.5, 25, and 50 mm/min. The stabilities of the sphingolipid fraction and the pure sphingomyelin monolayers were investigated by measuring the change in surface area at constant surface pressures of 10 and 20 mN/m. The effects of compression speed and spreading solvent were investigated at both surface pressures. XO was purified essentially as described by Rajagopalan,21 but with the following modifications. Bovine milk cream was washed twice in 0.2 M potassium phosphate buffer, pH 7.8, 1 mM benzamidine (500 mL/L of milk) prior to churning in the same buffer (1:1, v/v) containing 2 mM dithioerythriol. After treating the buttermilk at 37 °C for 60 min in the presence of 50 mM EDTA, 5 mM sodium salicylate, and 0.2 M NaHCO3, ammonium sulfate precipitation was performed at 0 °C in the presence of butanol (15%, v/v) and the precipitate obtained at 26-52% saturation was collected. The sample was resuspended in 5 mM NH4HCO3 buffer, pH 7.8, 0.1 mM EDTA, and 1 mM sodium salicylate, and the remaining butanol was removed by either dialysis or chromatography on a Sephadex G25 column using the same buffer. The material was then applied onto a Q-Sepharose high performance column and eluted with a linear gradient of 0-1 M NaCl in 10 mM NH4HCO3 buffer, pH 7.8, 0.1 mM EDTA, and 1 mM sodium salicylate. XO-containing fractions were identified by monitoring the absorption, color and enzyme activity and then further purified on a Sephacryl S300 gelfiltration column in 50 mM NH4HCO3 buffer, pH 7.8. The enzyme-containing fractions were concentrated by ultrafiltration using an Amicon cell with a PM10 filter and stored in aliquots at -80 °C. The typical yield obtained by this method ranged from 130 to 150 mg of XO from 24 L of milk. The molecular weight of XO was determined to be 148 000 g/mol by SDS-PAGE electrophoresis, using Coomasie Brilliant Blue R250 staining, and the purity of the sample was estimated to be at least 95%, as judged from the intensity of the electrophoretic bands. The adsorption experimentsssurface pressure (Π) versus timesat different XO concentrations were performed according to the procedure employed by Bos and Nylander.22 After the protein solution was poured in the trough, the surface was cleaned by sweeping the interface with the Teflon barrier and any surface active materials were removed by suction of the compressed film. The Wilhelmy plate was immediately lowered into the solution, and the adsorption of XO versus time was followed by recording the increase in surface pressure (Π). New protein solutions were prepared immediately before each experiment, by diluting an aliquot of the stock solution. The adsorption of XO into a phospholipid monolayer was performed as follows. The lipid was spread directly at a cleaned (20) MacRitchie, F. Chemistry at Interfaces; Academic Press: San Diego, CA, 1990. (21) Rajagopalan, K. V. In CRC Handbook of Methods for Oxygen Radical Research, 2nd ed.; Greenwald, R. A., Ed.; CRC Press: Boca Raton, FL, 1986; pp 21-23. (22) Bos, M. A.; Nylander, T. Langmuir 1996, 12, 2791-2797.

5858 Langmuir, Vol. 12, No. 24, 1996

Kristensen et al.

Figure 1. Π-A isotherms of (1) the sphingolipid fraction, (2) sphingomyelin, and (3) DSPC spread on PBS-buffer. A compression speed of 12.5 mm/min was used. interface of the protein solution. The influence of the protein incorporation into the phospholipid monolayer on the compression/expansion Π-A isotherms was investigated by varying the time span between spreading and the beginning of the compression and expansion cycle. All experiments were performed at 25 ((0.4) °C, and the deviation between duplicate experiments was e1 mN/m for Π versus time measurements and e0.5 mN/m for Π versus area isotherms.

Results and Discussion Π-A Isotherms of Phospholipids. The Π-A isotherms for the sphingolipid fraction, purified sphingomyelin, and DSPC are shown in Figure 1. The isotherm for DSPC is typical for a fully condensed monolayer, whereas the isotherms for the pure sphingomyelin preparation show a marked plateau at a surface pressure of about 21 mN/ m, indicating a phase transition from expanded to condensed state. This has also been observed for other sphingomyelins, egg yolk sphingomyelin and N-palmitoyl sphingomyelin.23,24 The transition is more pronounced for the enriched sphingomyelin preparation compared to the sphingolipid fraction, which is expected due to the heterogeneous composition of the sphingolipid fraction. The influence of the subphase composition on the sphingolipid fraction and the sphingomyelin and DSPC Π-A isotherms is shown in Figure 2, where the isotherms recorded on PBS-buffer and water are compared. The isotherms for sphingolipid and DSPC monolayers are shifted to a lower area per molecule when the lipid is spread on the subphase with higher ionic strength compared to water. This can be attributed to a decrease in repulsion between the charged headgroups. Since the monolayers are metastable (see below), it cannot be ruled out that the ionic strength affects the stability of the monolayers. It is interesting to note that little difference between the isotherms recorded on water and on buffer is observed for the pure sphingomyelin preparation. The reason for this is unknown. However, the phase transition for both sphingomyelin-containing preparations is more evident on the buffer subphase compared to the water subphase. The collapse pressures of DSPC monolayers on water and buffer are 64 and 62 mN/m, respectively. The collapse pressures of the sphingomyelin-containing monolayers are not evident in these experiments. In Table 1 the phase transition pressure as well as the mean molecular areas at 5, 15, 20, and 30 mN/m for the investigated phospholipids at the used subphase compositions are listed together with earlier published data for corresponding lipids. The observed variations in the reported mean molecular areas for the sphingomyelin and DSPC monolayers can be explained by the different (23) Ibdah, J. A.; Phillips, M. C. Biochemistry 1988, 27, 7155-7162. (24) Lund-Katz, S.; Laboda, H. M.; McLean, L. R.; Phillips, M. C. Biochemistry 1988, 27, 3416-3423.

Figure 2. Π-A isotherms of (A) the sphingolipid fraction, (B) the sphingomyelin fraction, and (C) DSPC spread on PBS-buffer and water. A compression speed of 12.5 mm/min was used.

temperatures and subphases employed, as well as the inherent biological variance in the samples. It should be noted, as discussed further below, that the isotherms are dependent on the compression speed at surface pressures above 22 mN/m. Thus the given values from the present study, recorded at a speed of 50 mm/min, will be lower at a lower compression speed. Good agreement is observed between the present results from the DSPC study and earlier work, when comparing the mean molecular areas at the different surface pressures. Stability of Monolayers. The influence of compression speed on the Π-A isotherms for the sphingolipid fraction and sphingomyelin is shown in Figure 3. Collapse of the monolayers, at 65 mN/m for both preparations, is only observed with a barrier speed of 50 mm/min, e.g. the highest speed employed in these experiments. The recorded compression-expansions isotherms (see Figures 6 and 7, where examples of both compression and expansion isotherms are given) showed significant hysteresis. It was also found that the hysteresis was significantly reduced when the compression was stopped at lower pressures. This suggests loss of lipid molecules from the monolayer at the air/water interface at high surface pressure. However, it cannot be ruled out that rearrangements within the monolayer may also have an effect. Sphingomyelins have previously been reported to form metastable monolayers.25,33 In agreement with the (25) Yedgar, S.; Cohen, R.; Gatt, S.; Barenholz, Y. Biochem. J. 1982, 201, 597-603.

Bovine Milk Sphingomyelin and Xanthine Oxidase

Langmuir, Vol. 12, No. 24, 1996 5859

Table 1. Mean Molecular Areas at Various Π and Phase Transitions of the Sphingolipid Fraction, Shingomyelin, and DSPC in Pure Monolayers on PBS-Buffer and Water at 25 °C and Comparison with Previous Reports for Monolayers of Various Natural, Semisynthetic, and Synthetic Sphingomyelins and DSPC at the Air/Water Interface

type of sphingolipida

buffer

bovine brain bovine brain bovine brain bovine brain bovine brain bovine brain bovine milk sphingolipid bovine milk sphingolipid bovine milk bovine milk beef heart hydrogenated beef heart egg egg egg sigma sigma N-palmitoyl-DN-palmitoyl DL-erythroN-oleyl-DN-oleylN-stearoyl-DL-erythroN-stearoylN-lignoceryl-DL-erythro3-OH-stearoyl-DL-erythro3-0-Me-N-stearoyl-DL-erythro3-O-Et-N-stearoyl-DL-erythro3-deoxy-N-stearoyl-DL-erythro 3-OH-egg3-OTHP-egg-DL-d 3-OH-N-(R-OH-pamitoyl)-DLerythro3-deoxy-2-O-stearoyl 3-deoxy-N-stearoylDSPCe DSPCe DSPCe DSPCe DSPCe

50 mM Tris-HCl, 5 mM MgSO4 10 mM phosphate, 150 mM NaCl sodium citrate sodium phosphate sodium borate 50 mM Tris-HCl, 140 mM NaCl 40 mM phosphate, 100 mM NaCl H20 40 mM phosphate, 100 mM NaCl H20 20 mM NaCl/10 mM CaCl2 20 mM NaCl/10 mM CaCl2 10 mM phosphate, 150 mM NaCl 8.70 mM phosphate, 80 mM NaCl 50 mM Tris, 140 mM NaCl sodium citrate sodium phosphate 10 mM phosphate, 150 mM NaCl 50 mM Tris-HCl, 5 mM MgSO4 10 mM phosphate, 150 mM NaCl H2O 50 mM Tris-HCl, 5 mM MgSO4 H2O 50 mM Tris-HCl, 5 mM MgSO4 H2O H2O H2O H2O H2O H2O H2O

mean molecular area (Å2/molecule) phase at various Π (mN/m) transition pH T (°C) 5 15 20 30 (mN/m) 7.4 7.6 3 6 8 7.5 7.4 7.4 7.4 7.4 5.6 5.6 7.6 7 7.5 3 6 7.6 7.4 7.6 7.4 7.4

H2O H2O 8.70 mM phosphate, 80 mM NaCl 7 10 mM phosphate 7 H2O 40 mM phosphate, 100 mM NaCl 7.4 H2O

23 25 20 20 20 22 25 25 25 25 25 25 25 25 22 20 20 25 23 25 22 23 22 23 25 25 25 25 25 25 25

210 80 69 69 72

185

78 86 79 76 78 64 83 74

58 66 64 61 61 49

70 70 81 70 87

58 58

53

48

63

45 53 51 57 41 58 90 47

22 22 25 25 25 25 25

80 80 56 55 56 52 55

63 58 54 52 51 50 51

57 57 59

64

52

150 58 52 52 54 54 50 58 60 52 55 47 62 61 59 53 51 61 49 68 74 45 52 45

65 51 46 46 46

≈20 14 ndf nd none

40 48 42 40 48 42 54 49

22 none 22 15-20 40c 40c 19 ( 1 22

47 47 52 45 61

43

nd nd 19 ( 1 12 ( 1 none nd 4(1 nd 11 ( 1

58 52 52 51 50 49 50

50 48 51 49 48 48 48

nd nd none none none none none

43

ref 25b 24 26b 26b 26b 27b present present present present 28b 28b 24 23b 27b 29b 29b 24 25b 24 30 25b 30 25b 31 31 31 31 31 31 31 32b 32b 23b 22b 22b present present

a Refers to sphingomyelin unless stated otherwise. b Values read from graphs. c Phase transition determined by sprinkling of talc particles on the monolayer. d Egg 3-O-tetrahydropyranoyl. e Distearoylphosphatidylcholine. f nd, not detectable.

present study, it was shown, from constant area-Π measurements, that monolayers of sphingomyelins were stable up to 20 mN/m.25 To give further information concerning the instability of the sphingomyelin monolayer, additional experiments were performed. Before discussing these results, it is important to note that the spreading in itself was quantitative, since spreading of different amounts resulted in the same Π-A isotherm. Equilibrium spreading pressures (ESPs) of the two sphingomyelin preparations were determined on water and buffer at 25 °C by placing an excess of lipid crystals at the surface. The lipid will spread until the surface pressure reaches a constant pressure, the ESP. For both preparations the values of the ESP were 0 mN/m, thereby resembling phosphatidylcholines, which only spread from the liquid crystal at temperatures above the gel-liquid crystal transition temperatures.20 In contrast to the case for sphingomyelin, the Π-A isotherm of DSPC was unaffected by the compression speed (data not shown). The DSPC monolayer was very stable and could be compressed to and held at high surface pressure (40 mN/m) without any change in area. Furthermore almost no hysteresis upon expansion was observed (Figure 8). In relation to these results, it is noteworthy that the isotherms for the pure sphingomyelin (26) Casas, M.; Min˜ones, J.; Iribarnegaray, E. Colloids Surf. 1991, 59, 345-359.

preparation are much more dependent on the compression speed above the transition point (plateau) at 22 mN/m than those of the sphingolipid fraction that contains 11% phosphatidylcholine and 15% ceramides (Figure 3). This suggests that the pure monolayers are less stable than the ones formed from the heterogeneous preparation. The compressibility is also much higher and the areas at high surface pressure are smaller for the sphingomyelin preparation. In other words, one or several of the other components in the sphingolipid fraction stabilize the sphingomyelin portion in the monolayer. Other lipids have been shown to interact with sphingomyelin; for instance, cholesterol reduced the mean molecular areas of mixed monolayers, which was interpreted as condensation of sphingomyelin.27,31 The lower stability of sphingomyelin monolayers compared to that of the DSPC ones can be attributed to differences in the hydrogen-bonding capacity of the belt region of the molecules, which connects the apolar region with the phosphorylcholine group.3 It is (27) Slotte, J. P. Biochemistry 1992, 31, 5472-5477. (28) Shah, D. O.; Schulman, J. H. Lipids 1966, 2, 21-27. (29) Casas, M.; Min˜ones, J. Colloid Polym. Sci. 1992, 270, 485-491. (30) Mattjus, P.; Slotte, J. P. Chem. Phys. Lipids 1994, 71, 73-81. (31) Gro¨nberg, L.; Ruan, Z.; Bittman, R.; Slotte, J. P. Biochemistry 1991, 30, 10746-10754. (32) Bittman, R.; Kasireddy, C. R.; Mattjus, P.; Slotte, J. P. Biochemistry 1994, 33, 11776-11781. (33) Lusted, D. Biochim. Biophys. Acta 1973, 307, 270-278.

5860 Langmuir, Vol. 12, No. 24, 1996

Kristensen et al.

Figure 4. Rate of desorption as log area (log(A)) versus square root of time (t1/2) of sphingomyelin from monolayers at contant surface pressures of 10 and 20 mN/m. Data for the monolayer area at the surface pressure 40 mN/m are indicated. The lipid is spread from a hexane/ethanol (9:1, v/v) solution on PBSbuffer. A compression speed of 12.5 mm/min was used.

Figure 3. Π-A isotherms (A) of the sphingolipid fraction and (B) sphingomyelin spread on PBS-buffer. Results from three different compression speeds, 12.5, 25, and 50 mm/min, are shown.

generally accepted that sphingomyelins are more polar than phosphatidylcholines with the same chain length. Several studies concerning the stability of insoluble monolayers of ionic substances, e.g. carboxylic acids and amines, have been reported.34 Information on monolayer stability has generally been derived from Π-A isotherms as the effect on the collapse pressure. Studies on films held at constant area or constant surface pressure reveal instability at Π values considerable below the collapse pressure.34 In constant area experiments it is difficult to separate the monolayer instability from relaxation processes and the film is also subject to various phase transitions during the compression of the film. This is avoidable when the pressure is kept constant. Furthermore, registration of the area change versus time permits kinetic analyses. The change in relative monolayer area of the sphingolipid fraction and sphingomyelin, respectively, at constant surface pressure was recorded versus time. The monolayer stability of the pure sphingomyelin is more dependent on the surface pressure compared to that of monolayers of the sphingolipid fraction (data not shown). At 10 mN/m, well below the phase transition area, the monolayer of pure sphingomyelin is reasonable stable, but at 20 mN/m the area decreases significantly with time. This trend is more pronounced at even higher surface pressures (40 mN/m). The monolayer area of the sphingolipid fraction at surface pressures of 10 and 20 mN/m versus time is identical during the first 30 min. However, at higher surface pressure at larger decrease in monolayer area was observed. A linear decrease of the logarithm of the monolayer area with the square root of time at constant surface pressures is indicative of desorption from the monolayer.20 Examples of such plots are given in Figure 4. No linear relation is observed for monolayers of pure sphingomyelin at 10 mN/m. Monolayers held at 20 mN/m show marked (34) Binks, B. P. Adv. Colloid Interface Sci. 1991, 34, 343-432.

discontinuities during the first 10-20 min of the experiment. This indicates that the monolayer instability cannot be explained by simple desorption of lipid monomers to the subphase. It should also be borne in mind that the monomer aqueous solubility of this type of molecules is extremely low (≈10-10 M35); thus, dissolution of lipid monomers into the subphase is unlikely to have a significant effect. Hexane/ethanol (9:1, v/v) has frequently been used as spreading solvent for sphingolipids.23,24,31 It has been demonstrated that the composition of the spreading solvent can influence the loss from monolayers of phospholipids, long-chain alcohols, and fatty acids into the subphase.36 However, we did not observe any significant difference in the results whether the lipids were spread from hexane/ethanol (9:1, v/v) or chloroform/methanol (2:1, v/v). The compression speed has been reported to determine the initial decrease in octadecylamine monolayer area in constant surface pressure experiments, due to structural rearrangements occurring within the monolayer.37 In the present study no effect of the compression speed on the relative area decrease with time at constant surface pressure could be observed. This suggests that the rearrangement processes are not the main reason for the monolayer instability. Furthermore, the observed areas per molecule from the Π-A isotherm are less than expected for a condensed phospholipid monolayer although the spreading was quantitative. However, our results on the stability of monolayers held at different surface pressures demonstrate that the kinetics of loss of material from the monolayer are not simple desorption. These results clearly demonstrate that information regarding monolayer stability, complementary to what is obtained from Π-A isotherms, can be drawn from constant surface pressure data. To our knowledge no work on the stability of monolayers of natural sphingomyelins using this approach has previously been published. The present study also demonstrates that care has to be taken when comparing areas per molecule obtained from Π-A isotherms with those obtained with other methods, like X-ray diffraction experiments,4,5 as the isotherms can be strongly dependent on the compression speed. Adsorption of XO. Figure 5 shows Π as a function of time for XO in PBS-buffer at different concentrations. A protein concentration of 5 mg/L was sufficiently low to (35) Small, D. M. Handbook of Lipid Research, Vol. 4. The Physical Chemistry of Lipids: From Alkanes to Phospholipids; Plenum Press: New York, 1986. (36) Gericke, A.; Simon-Kutscher, J.; Hu¨hnerfuss, H. Langmuir 1993, 9, 2119-2127. (37) Vikholm, I.; Teleman, O. J. Colloid. Interface. Sci. 1994, 168, 125-129.

Bovine Milk Sphingomyelin and Xanthine Oxidase

Langmuir, Vol. 12, No. 24, 1996 5861

Figure 5. Surface pressure (Π) versus time for XO dissolved in PBS-buffer at various concentrations.

ensure proper spreading of the phospholipids without any surface pressure barrier from the adsorbed protein molecules present. However it must be borne in mind that even though no surface pressure is observed a substantial amount of protein can be present at the interface and an increase in surface pressure occurs when the surface concentration approaches that of a close packed monolayer.38 The adsorption kinetics of XO, typical for macromolecules, demonstrate that a long time was needed before the surface pressure reached a constant (equilibrium) value. The plateau level of Π was approximately 21 mN/m after 12 h for 5 mg/L. Similar values has been reported for other proteins; for instance, the surface pressure of the whey protein β-lactoglobulin reaches a plateau level of about 20 mN/m.22 Π-A Isotherms of Mixed Lipid-Protein Monolayers. As expected the Π-A isotherms of the phospholipid monolayers on the protein (e.g. XO) solution will largely depend on the time elapsed between spreading and compression. Already after a few minutes the isotherms were shifted toward larger area. As the films were compressed, the isotherms asymptotically approached that of the pure phospholipid monolayer. Figures 6A, 7A, and 8A depict the time dependencies for compression curves for the sphingolipid fraction, sphingomyelin, and DSPC, respectively, while Figures 6B, 7B, and 8B show the corresponding expansion curves. For the sphingolipid monolayer the compression curves with incorporation of XO for 5, 10, and 20 min coincide at 45 mN/m, signifying the same extent of expulsion of protein from the mixed protein-lipid monolayer. The corresponding values for sphingomyelin and DSPC monolayer experiments performed under similar conditions are 53 and 52 mN/m. For all three lipids the results for the dynamic Π-A isotherms after 3 h of XO incorporation deviate. Not all of the protein seems to be expelled, as the compression curves did not coincide with the other Π-A compression curves. Furthermore, Π did not reach zero after maximum expansion of the film area. Our results for the increase in surface pressure of the XO solution in the presence of the phospholipid monolayer as compared to a clean interface show the same trend (data not shown). The surface pressures are more or less identical after 2 h. After 3 h the Π of the sphingomyelin monolayers with adsorbed XO is actually about 2 mN/m lower than that when XO is adsorbed into its own monolayer. One possible explanation for these observations of mixed protein-lipid monolayers at long equilibrium times might be an increase of the monolayer viscosity with time. This can be caused by the formation of intra- and intermolecular disulfide bridges. XO has been reported to have four free sulfhydryl groups per XO dimer.39 For, β-lactoglobulin, which is (38) De Feijter, J. A.; Benjamins, J. In Food Emulsions and Foams; Dickinson, E., Ed.; Royal Society of Chemistry: London, 1987; pp 7285.

Figure 6. Dynamic surface pressure (Π) as a function of the molecular area per lipid molecule for (A) compression and (B) expansion isotherms of pure sphingolipid fraction monolayers and mixed XO/lipid monolayers recorded at different times elapsed between spreading and compression. The protein concentration in the subphase, PBS-buffer, was 5 mg/L. The compression and expansion speed was 12.5 mm/min.

Figure 7. Dynamic surface pressure (Π) as a function of the molecular area per lipid molecule for (A) compression and (B) expansion isotherms of pure sphingomyelin monolayer and mixed XO/lipid monolayers recorded at different times elapsed between spreading and compression. The protein concentration in the subphase, PBS-buffer, was 5 mg/L. The compression and expansion speed was 12.5 mm/min.

another milk protein which contains sulfhydryl groups, the presence of free SH groups gives rise to slow polymerization of the protein in the adsorbed layer via sulf-

5862 Langmuir, Vol. 12, No. 24, 1996

Figure 8. Dynamic surface pressure (Π) as a function of the molecular area per lipid molecule for (A) compression and (B) expansion isotherms for pure DSPC monolayer and mixed XO/ lipid monolayers recorded at different times elapsed between spreading and compression. The protein concentration in the subphase, PBS-buffer, was 5 mg/L. The compression and expansion speed was 12.5 mm/min.

hydryl-disulfide interchange.40 Another possibility is that a lipid-protein complex might eventually be formed, which facilitates the solubilization of the lipid in the subphase. Hence, lipid molecules might be lost from the monolayer before it is possible to expel the protein. The Π-A isotherms of the mixed XO-sphingolipid or XO-sphingomyelin monolayers did not, even at high surface pressures, coincide with the ones recorded for the pure lipid monolayers; see Figures 6-8. Furthermore a significant increase in Π at the area per molecule 30 Å2 (maximum compression) was observed in the presence of XO. For the sphingolipid fraction ∆Π is 15 mN/m, and it is approximately 20 mN/m for the sphingomyelin. The collapse pressure for DSPC was not affected by XO in the subphase. It should be noted that sphingomyelin and phosphatidylcholine are zwitterionic and that XO carries a slightly positive net charge at pH 7.4, as the isoelectric point of the protein is 7.7 on the basis of on the amino acid composition.41 The stabilizing effect of XO for sphingomyelin-containing monolayers might be due to different packing and orientation of individual molecules in a monolayer as a consequence of the interaction with XO. The exact mechanism of this interaction and the location of XO require further investigation. In a monolayer study of protein-lipid interactions, Vikholm and Teleman37 showed that antibodies in the subphase counteracted the dissolution of octadecylamine from a monolayer, even when the surface pressures were held above surface pressures where the protein could be incorporated. From their data they concluded that expelled protein molecules remained in an adjacent layer and collapse pressure (39) Cheng, S. G.; Koch, U.; Brunner, J. R. J. Dairy Sci. 1988, 71, 901-916. (40) Dickinson, E.; Matsumara, Y. Int. J. Biol. Macromol. 1991, 13, 26-30. (41) Berglund, L.; Rasmussen, J. T.; Andersen, M. D.; Rasmussen, M. S.; Petersen, T. E. J. Dairy Sci. 1996, 79, 198-204.

Kristensen et al.

increased due to a specific interaction between the protein and the lipid monolayer. Lipid-protein Interactions at Liquid Interfaces. Most studies on lipid-protein interactions are carried out at constant area. The majority of these studies employ only phosphoroglycerides, but some studies also include sphingomyelins.23,42 Another method involves measurement of protein adsorption into a lipid monolayer at constant surface pressure by recording the increase in area, ∆A versus time. This method is believed to be more appropriate for mimicking the conditions in biological membranes. This approach has been used to study the interaction between β-lactoglobulin and distearoylphosphatidic acid (DSPA), distearoylphosphatidylcholine (DSPC), and dipalmitoylphosphatidic acid (DPPA) at the air/water interface.22 For studies of lipid-protein interactions at constant area or at constant Π, an interaction is recognized to take place if the resulting surface pressure is above the adsorption plateau level of the protein itself, e.g. for XO above Π ) 21 mN/m. However, in this surface pressure region bovine milk sphingomyelin monolayers are metastable, and therefore results obtained by this approach should be evaluated with certain precautions. We chose to perform dynamic Π-A experiments to get qualitative information about the sphingomyelin-XO interaction. A study with mixed films of sphingomyelin and pepsin on a subphase of polymerized silicic acid, which increased the stability of the system, also involved dynamic Π-A measurements.29 As opposed to the case in our work, pepsin was spread at the interface prior to spreading of the lipid. The results suggested that pepsin and sphingomyelin are miscible at the interface. Conclusion This study shows that bovine milk sphingomyelin forms metastable monolayers and that the high proportion of long saturated fatty acids is reflected in a liquid-condensed phase transition at 22 mN/m, features also reported for other natural sphingomyelins. In this study we found indications of an interaction between sphingomyelin and XO. It is interesting to note this result in an increase in Π at maximum compression, which reflects an increase in the stability of the sphingolipid/sphingomyelin monolayers. The fact that this effect was not observed for DSPC monolayers, as well as the substantially larger effect observed for pure sphingomyelin monolayers compared to the ones formed by the sphingolipid sample, indicates that the interaction is dependent on the presence of sphingomyelin. Although it is difficult to make a direct comparison between the simplified “model membrane” used in this study and such a complex system as the milk fat globule membrane it, cannot be ruled out that this type of interaction can take place and even be of significance for the stabilization of the membrane. Acknowledgment. Dr. Martin Bos is thanked for his valuable help in the experimental work and discussions during this study. Prof. Kåre Larsson and Prof. Petr Dejmek are thanked for comments. D.K. acknowledges the Ph.D. stipend from the Danish Agricultural and Veterinary Research Council and the financial support from the Danish Dairy Board and the Danish Research Academy. T.N. acknowledges financial support from the Swedish Council for Forestry and Agricultural Research. LA960259O (42) Schubert, D.; Herbst, F.; Marie, H.; Rudloff, V. In Protides of biological fluids. Proc. 29th Coll. 1981; Peeters, H., Ed.; Pergamon Press: Oxford, 1982; pp 121-124.