Langmuir 1996,11, 3948-3952
3948
Hydration of DPPC Monolayers at the Air/Water Interface and Its Modulation by the Nonionic Surfactant C 1 2 E 4 : A Neutron Reflection Study C. Naumann,? T. Brumm,? A. R. Rennie,t J. Penfold,$ and T. M. Bayerl*9+ Physik Department E22, Technische Universitat Miinchen, 0-85748 Garching bei Miinchen, Germany, Cavendish Laboratory, University of Cambridge, Cambridge CB3 OHE, U.K., and Rutherford Appleton Laboratory, Chilton Didcot Oxon OX11 OQX, U.K. Received March 6, 1995. I n Final Form: May 23, 1995@ The hydration of a phospholipidmonolayer of dipalmitoylphosphocholine(DPPC)at the airlwater interface is studied for different lateral pressures which correspond the the fluidlike I and LE phases and to the solidlike S phase using the neutron reflection technique. This method enables for the first time a precise determination of the number (n,,) of water molecules per DPPC headgroup as a function of monolayer phase state. Furthermore, the effect of the presence of the nonionic surfactant poly(oxyethyleneglyco1)n-dodecyl ether (C12E4)on the monolayer hydration was determined for the I and LE phases. For the pure DPPC monolayer, we observe a drastic dependence of n,’ on the phase state, with the highest hydration of n,’ = 22 f 2 measured for the I phase and the lowest hydration of n,‘ = 4 i 1 for the solidlike S phase. In between these extremes,we obtain nw’= 10 f 1 for the fluidlike LE phase. These values correlate with those determined for the headgroup volume of DPPC at the corresponding phase states. The presence of C&4 in the monolayer at a molar lipidsurfactant ratio of 2 causes a significant increase of the monolayer hydration by 35-70% for the fluidlike I and S phases. However, the surfactant exhibits a significantly less pronounced dependence of its hydration on the lateral pressure as compared t o DPPC. A comparison of our pure DPPC monolayer hydration data with published values for DPPC bilayers reveals striking differences with respect to nw’values for apparently similar phase states in monolayers and bilayers. We conclude that an identification of the corresponding monolayer and bilayer phases solely on the basis of headgroup hydration is not applicable.
Introduction Two major model systems are used in the field of membrane biophysics for studying the various aspects of structure, dynamics, and interaction of complex biological membranes. These are the lipid bilayer systems (e.g., vesicles, multilayers, black lipid membranes) and lipid monolayers at the aidwater interface. The hydration state of phospholipids is generally acknowledged to be of major importance for the structure and dynamics of phospholipid bilayers. 1-3 This is because lipid hydration directly modulates a t least one of the basic molecular forces, which ultimately determines the structure of all lipid assemblies. This socalled hydration force dominates the interbilayer interaction over all the other forces resulting from van der Waals and electrostatic interactions a t veryoclose separation distances between the lipid layers ( < 10A). Its modulation is a prerequisite for the fusion of membranes as well as for the coupling of proteins to the bilayer surface. Its origin is still a matter of c o n t r ~ v e r s y . ~ , ~ A wealth of information is available about the hydration of lipid bilayer systems, based on methods including NMR, X-ray diffraction, neutron scattering, infrared spectroscopy, and dielectric relaxation technques5-12 In Technische Universitat Miinchen.
* University of Cambridge. +
5 Rutherford Appleton Laboratory.
Abstract published in Advance A C S Abstracts, September 1, 1995. (1)Jiirgens, E.;Hohne, G.;Sackmann,E. Ber. Bunsenges. Phys. Chem. 1983,87,95. (2) Rand, R. P.; Parsegian, V. A. Biochem. Biophys. Acta 1980,988, @
---. 251
(3) Israelachvili, J. N. Intermolecular and Surface Forces; Academic: London, 1992. (4) Israelachvili, J . N.; Wennerstrom, H. Langmuir 1990,6,873. (5) Finer, E.G.; Darke, A. Chem. Phys. Lipids 1974,12, 1. (6)Klose, G.; Gawrisch, K. Stud. Biophys. 1981,84,21.
0743-746319512411-3948$09.00/0
contrast, there is virtually no work available dealing with the lipid hydration in monolayers in their different phase states. The reason for this lack of information is mostly technical in nature. The methods employed for studying lipid hydration in bilayers are not applicable to monolayers. The neutron reflection technique has the potential to provide information about monolayer hydration by exploiting the drastic scattering contrast which can be obtained by substituting H2O for D2O as the monolayer subphase.13-16 Beside that, only the synchrotron X-ray scattering enables the measurement of monolayer hydration as demonstrated for the phosphatidylethanolamine mon01ayer.l~However, the latter method cannot take advantage of HzO/D20 substitution and selective deuteration of the monolayer. There are two aims of this work: First, we wish to demonstrate the applicability of the neutron reflection method as a sensitive tool for measuring the hydration state of a phospholipid monolayer of DPPC and to provide for the first time detailed monolayer hydration data as a (7) Hauser, H.; Pascher, I.; Pearson, R. H.; Sundell, S. Biochim. Biophys. Acta 1981,650,21. (8)Biildt, G.; Gally, H. U.; Seelig, J.; Zaccai, G. J . Mol. Biol. 1979, 134,673. (9)Konig, S.;Pfeiffer, W.; Bayerl, T. M.; Richter, D.; Sackmann, E. J . Phys. (Paris) 1992,2,1589. (10) Konig, S.; Sackmann, E.; Richter, D.; Zorn, R.; Carlile, C.; Bayerl, T. M.J . Chem. Phys. 1994,100 (4), 3307. (11)Fringeli, U. P.; Guenthard, H. H. Biochim. Biophys.Acta 1976, 450,101. (12) Enders, A,; Nimtz, G. Ber. Bunsenges. Phys. Chem. 1984,88, 512. (13) Penfold, J.; Thomas, R. K. J . Phys. Condens. Matter 1990,2, 1369. (14) Russell, T.P. Mat. Sei. Rep. 1990,5,171. (15) Bayerl, T. M.;Thomas, R. K.; Penfold, J.; Rennie, A. R.; Sackmann, E. Biophys. J . 1990,57,1095. (16) Vaknin, D.; Kjaer, K.; Als-Nielsen, J.; Losche, M. Biophys. J . 1991,59,1325. (17) Helm, C. A,; Tippmann-Krayer, P.; Mohwald, H.; Als-Nielsen, J.; Kjaer, K. Biophys. J . 1991,60,1457.
0 1995 American Chemical Society
Hydration of DPPC Monolayers function of lateral pressure. Second, we have studied the effect of the incorporation of a nonionic surfactant of the poly(oxyethy1ene)variety into the DPPC monolayer on its hydration state. From a previous paper devoted to the structural effects of this surfactant on the DPPC monolayer, a significant effect of the surfactant incorporation on the lipid headgroup hydration was predicted on the basis of the surfactant headgroup structure in the mixed monolayer.18
Materials and Methods The selectivity chain-deuterated lipid DPPC-d62was obtained from Avanti Polar Lipids (Alabaster, AL). The protonated nonionicsurfactant C12E4-hwas purchased from Nikko Chemicals Inc. (Tokyo, Japan). The DzO was provided by Isotec Inc. (Miamisburg, OH). The neutron reflection measurements were performed on the CRISP spectrometer of the Rutherford Appleton Laboratories (Didcot,U.K.).19 We used a home-built,thermostated film balance with a trough size of 15 x 26 cm (volume400 mL). The barrier position was controlled by a computer, and the film pressure was measured by a Wilhelmy plate pressure detector (accuracy 0.1 mN/m). The temperature was kept tixed at 20 "C. The monolayer preparation and instrumental setup have been described in detail previously.Z0 All measurements were done on a subphase of DzO, and two different monolayers were studied: (1)DPPC-d62 on DzO; (2) DPPC-d&lzE& on DzO (molar ratio: rws = 2). DPPC-ds2 consists of two fully deuterated palmitoylchains and a protonated lecithin (cholin)headgroup. The protonated surfactant (C12E4h ) was used for all measurements. Data Interpretation Procedure. For analyzing the reflectivity data, we used the well-established opticalmatrix method.z1 First, we attempted to fit the reflectivity data by a one-layer model. Satisfactory fits to the data were obtained for the measurements at the lowest lateral pressure (1.4mN/m)in terms of this simple model. At this pressure, the scattering length density is similar in both the hydrophilic and the hydrophobic parts of the monolayer. This can be seen as a first indication of the monolayer hydration. To obtain satisfactory fits at all pressures studied, we had to apply a two-layer model. A severe limitation ofthis approachis the large number offree parameters, because some of them may change simultaneously. Therefore, we imposed two constraints to the two-layerfits (1)the molecular area (AM)must agree with the value measured independently using the film balance (withAM=ALfA&ws;A~andAs are the lipid and surfactant molecular area, and T-IB is the lipidsurfactant ) agree molar ratio), and (2) the total layer thickness ( d ~must with that obtained previously for fully deuterated DPPC-& on a subphase of null-reflecting water. The data analysis was initiated by changing the parameter of the hydrophobiclayer thickness step by step and by varying the thickness of the headgroup region and scattering length density of the hydrophilic and hydropobic regions. At this stage, interfacial roughnesses were set to zero. As a result, we obtained a number of satisfactory fits. Assuming that no water can penetrate into the hydrophobicregion, the corresponding areas per DPPC molecule were determined. Then that result was selected from the set of possible solutions for which the area per molecule (AM)agreed well with that determined by the film balance technique.1s,z2In each case, only one physically meaningful result remained. It should be noted that the error of the film balance measurement is less than 5%. At this stage of the analysis procedure, we had to introduce the interfacial roughness as an additional parameter with the (18)Naumann, C.; Dietrich, C.; Lu, J. R.; Thomas, R. K.; Rennie, A. R.; Penfold, J.; Bayerl, T. M. Lungmuir 1994, 10, 1919. ( 19)Penfold, J.;Ward, R. C.; Williams, W. G. J . Phys. E.: Sci. Znstrum. 1987,20, 1411. (20) Johnson, S.J.;Bayerl, T. M.; Weihan, W.; Noack, H.; Penfold, J.; Thomas, R. K.; Kanellas, D.; Rennie, A. R.; Sackmann, E . Biophys. . . 45. 1991, 60,1017. (21) Born, M.; Wolf, E.PrinciplesofOptics; Pergamon: London, 1959. (22) Brumm, T.;Naumann, C.; Sackmann, E.; Rennie, A. R.;Thomas, R. K.; Kanellas, D.; Penfold, J.;Bayerl, T. M. Eur. Biophys. J . 1994,23, 289.
Langmuir, Vol. 11, No. 10, 1995 3949
I
30
I\
5 "/m
4
(S phase)
E
L
20 :
=E c
10
0
I
50
60
I
I
1
I
70
80
90
100
area per molecule [A2] Figure 1. Pressure-area diagram of DPPC-d,jz on a DzO
subphase. The lateral pressures which were used for the neutron reflectivity experiments are indicated. condition that each interface has the same roughness value. We varied it in such a manner that the total monolayer thickness ( d ~agreed ) with the results we obtained recently with totally deuterated DPPC-d75 on null-reflecting water (cmw).18~22 The latter measurements can be considered as the most accurate determination of dT owing to the contrasts and the applicability of a one-layer mode. With the included constraints described above,the two-layer data analysis procedure provides a unique solution for each reflectivity curve.
Results (1) Hydration in Pure DPPC Monolayers. The pressure-area diagram of DPPC-d6z on DzO at 25 "C is shown in Figure 1, and the three positions where neutron reflection measurements were performed are indicated. With the exception of the tilted liquid condensed phase (LC phase), we performed measurements at each phase state observed on the DPPC monolayer according to the nomenclature given by Albrecht et 1.4 mN/m, isotropic fluid phase (I phase); 9.0mN/m, anisotropic fluid phase (LE phase); 35 mN/m, solid phase (S phase). At this point, it should be noted that the existence of two different liquid-phase states of DPPC is still in discussion. Therefore, it may well be possible that the I and LE phases are the same. The phase diagram in Figure 1 is qualitatively similar to that obtained for DPPC-dsz on a HzO subphase.16 However, the isotope substitution of the subphase has a distinct effect on the fluid phases as well as on the coexistence region between the LE phase and the LC phase (the plateau in Figure 11, with the corresponding pressure-area values being 15-20% lower for the HzO subphase.16 In general, these effects are minor compared to those observed due to the isotope substitution of the parts of lipid molecule.22 Since the present work concentrates solely on DPPC-d62 monolayers on a DzO subphase, the latter changes are of no relevance here. I Phase. The reflectivity curve of D P P C - ~on ~Z D20 a t JC = 1.4mN/m and T = 25 "C is shown in Figure 2, together with two-layer model simulations obtained for the parameters given in Table l (cf. Materials and Methods for details). The free parameter in the simulations is the number of water molecules per lipid located in the hydrophilic headgroup layer of the DPPC , The significant variations show the remarkable effect headgroup hydration exhibits on the reflectivity in the Q range accessible to the experiment. The best fit of the data according to (23) Albrecht, 0.; Gruler, H.; Sackmann, E . J.Phys. fPurisll978,39, 301.
Naumann et al.
3950 Langmuir, Vol. 11, No. 10, 1995
- . n wr
---
...
0 nw=5 n w =10
nwd5 nw=20
-.
...... n w . = 3 6 ... n, ~ 4
- best flt: n
0 I
22
I
indicates a drastic increase of hydration to 30 f 2 D20 molecules per lipid, compared to 22 k 2 obtained for the pure DPPC monolayer. From this, we can estimate that each surfactant molecule binds to 16 DzO molecules at this lateral pressure, Le., 4 per ethoxy group. Figure 6 shows data and simulations for x = 9.0 mN/m, and the best fit gives 17 f 2 wateD molecules per lipid. Considering that each DPPC binds 10 D2O molecules for the case of the pure DPPC monolayer, a hydration of 14 D20 molecules per C12E4 molecule can be estimated.
Discussion n,=O I
I
I
I
0.1
0.2
0.3
0.4
momentum transfer Q [ A " ] Figure 2. Experimental and theoretical neutron reflectivity profiles of a DPPC-dsz monolayer (subphase DzO) at a lateral pressure of 1.4mN/m (I phase). Data points (asterisk) and simulations (lines)are presented. The simulations were done assuming different numbers (n,,) of DzO molecules in the hydrophilic headgroup of the lipid as indicated.
the two-layer model-fittingprocedure (fitting parameters given in Table 1)yields 22 i 2 D20 molecules per DPPC molecule. For the sake of clarity, the fitted curve is not shown in Figure 1. LE Phase. A similar representation of experimental results and simulations is given in Figure 3 for the LEphase data a t 9 mN/m. Again, the high sensitivity of the simulated curve with respect to the lipid headgroup hydration is obvious. The best fit results (fitting parameters shownin Table 1)give 10 i 1D2O (9 mN/m) per lipid headgroup. A slight discrepancy between data and simulation can be observed for Q > 0.2 which was not observed a t lower pressures (I phase). We have no conclusive explanation for this effect, though we can rule from the way the experiment was performed (stepwise increase of x from low to high values) that it is not due to the formation of a n additional interface. One possibility is that a t this pressure, close to the main phase transition, the monolayer might exhibit some inhomogeneitiesin the lateral direction that can obscure the reflectivity curve. S Phase. The transition from the LE to the S phase via the LC phase can be expected to have drastic consequences for the lipid headgroup hydration.24 Figure 4 shows the representation of experimental data and simulations for the S phase of the monolayer a t 35 mN/m. Here the best fit gives 4 i 1 D2O molecules per lipid. (2) Effect of C12E4on the Hydration in DPPC-dsz Monolayers. The effect of the nonionic surfactant C12E4 was studied a t two pressures in the I and LE phases (1.4 and 9 mN/m), but no measurements were done for the S phase. This is because the surfactant is completely squeezed out from the monolayer into the subphase a t the transition between the LE phase and the S phase.ls Thus, it is the I-phase state where the most pronounced effects of the surfactant on the monolayer hydration can be expected. Since we used the protonated surfactant in our experiments, the reflectivity changes due to its incorporation into the DPPC monolayer are negligible compared to the contribution which arises from the additional D2O uptake of the headgroup layer by the presence of poly(oxyethylene) groups. The reflectivity of the mixed monolayer a t 1.4 mN/m and a lipidsurfactant mixing ratio ruS = 2 is shown along with the simulations in Figure 5. The best fit result (24) Nagle, J. F.;Wiener,M.Biophys. J.Biochim.Biophys.Acta1988, 942, 1.
Our results provide evidence for a drastic reduction of the DPPC monolayer hydration a t the transition from the I phase (1.4 mN/m) to the S phase (35 mN/m) via the LE and LC phases by a factor of 5-6. Furthermore, it is clearly demonstrated that the presence of C12E4 (rus = 2) in the monolayer (I and LE phases) causes a significant increase of the monolayer hydration by 35-70% as compared to the pure DPPC monolayer a t the same pressure. So far, lipid hydration has been studied mainly for bilayer systems by a variety of methods, as summarized in the Introduction. Since all these methods detect water in the vicinity of the lipid headgroup in different ways, a comparison with our results requires clarification ofwhat hydration measured by the neutron reflection method really means. All information about the hydration in our experiments is extracted from two-layermodel simulations which consider different water contents of the headgroup layer while assuming that no water can enter the hydrophobic chain region a t any phase state. Hence, this approach considers only those water molecules (n,,)located in between adjacent DPPC headgroups, while those layered on top of the headgroup region (n,,,) do not significantly contribute. As a consequence, this method can be expected to be very sensitive to the changes of water contents (n,O between the headgroups owing to changes of the headgroup spacing with the lateral pressure (x). Most of the other methods do not distinguish water in between the headgroups from that on top of the headgroup layer which may be still associated with the membrane but rather give a total water content nw = n,, n,,,. Because of different lateral pressures, the comparison of hydration data for monolayers and bilayers is generally complicated. A severe limitation is the fact that a fundamental aspect of hydration in multibilayers is lacking in monolayers: the interbilayer interaction comprising a balance between attractive van der Waals and repulsive hydration and steric interactions. On the other hand, this gives us the rare opportunity to consider the hydration of a single lipid layer as a function of its phase state in the absence of interbilayer contributions. Unfortunately, neutron reflectivity does not provide information about the binding strength of the water molecules located in the headgroup region. Pure Monolayer Hydration. DPPC monolayers exhibit a remarkable phase polymorphism. Four different phase states (I, LE, LC, and S ) have been observed.23 The hydration value obtained for the I phase agrees well with that obtained from a one-layer fit, which was used to check the data. The application of the simplest model was possible because of comparable magnitudes of the scattering length density values of the hydrophilic and hydrophobic parts of the monolayer at this pressure (Table 2). Additionally, we note that the hydration value obtained for the S phase a t 35 mN/m agrees well with that reported by other authors.16
+
Hydration of DPPC Monolayers
Langmuir, Vol. 11, No. 10, 1995 3951
Table 1. Optical Matrix Fitting Results (Two-LayerModel Fit) for DPPC-de2 and Mixed DPPC-dedClzEr-h (m= 2) Monolayere dcH, A DPPC-dsz on D2O (1.4mN/m) DPPC-ds2 on D2O (9.0mN/m) DPPC-dsz on DzO (35 mN/m) DPPC-d62 C12E4-h on D2O (1.4mN/m) DPPC-ds2 C12E4-h on D2O (9.0mN/m)
+
+
@CH, A-2
6.5 x
9 12 17 9 12
AM, A 2
101 77
6.3 x 7.3 x 4.9 x 4.9 x
46 127 95
dH, A
@H,A-2 4.6 x 4.0 x
11.4 9.5 10.9
roughness,A
n w D P P C molecule
3 3 4 3 3
22 10 4 30 17
3.4 x
11.0
4.9 x 4.5 x 10-6
9.9
a ~ C H thickness , of the hydrophobic monolayer region; @CH, scattering length density of the hydrophobic monolayer region;AM,area per lipid molecule;dH, thicknessofthe hydrophilic monolayer region; @H, scatteringlength density ofthe hydrophilic monolayer region; roughness, interfacial roughness used for all interfaces. 6. 4.
3. 2.
... n w = 1 5 --nw=O nw=4 nw= 8 .._ nw=12 . nw=16 - -nw=20
loJ: 6. 5. 4.
-
.
...
-
3.
beet ilk n N data
2-
nw=20 n w 125 - n w =35 -nw=40
E
= 30
best fn: n N data
10
....... I 71-7 1
0.05
0.10
0.15
0.20
I
I
I
0.25
0.30
0.35
0.05
momentum transfer o [A'']
Figure 3. Experimental and theoretical neutron reflectivity profiles of the system from Figure 2 but at a lateral pressure of 9.0 mN/m (LE phase).
I
I
I
I
I
1
1
0.10
0.15
0.20
0.25
0.30
0.35
0.40
momentum transfer Q / A ' Figure 5. Experimental and theoretical neutron reflectivity (rus = 2) monolayer on profiles of a mixed DPPC-d&12E& a subphase of DzO at a lateral pressure of 1.2mN/m (I phase).
Data (asterisk) and simulations (lines) are presented. The number (n,,) corresponds to the total number of D2O molecules associated with the hydrophilic region of the monolayer and is calculated per lipid molecule.
...... n w = 2 nw,=6
_.. nW=B ... n w = 1 0
-
best fik n wI I dah
0.05
E
f.
4
I
I
I
I
I
0.10
0.15
0.20
0.25
0.30
momentum transfer Q [ A ' ] Figure 4. Experimental and simulated neutron reflectivity profiles of the system from Figure 2 but at a lateral pressure of 35.0 mWm (S phase).
Further insight into the hydration behavior is provided by consideringthe headgroup volumes for the three phase states, as indicated by our fit data: 1150 A3 (I phase a t 1.4 mN/m), 730 A3(LE phase a t 9 mN/m), and 500 A3(S phase a t 35 mN/m). Thus, we have 420 A3for the I-to-LE transition and 230 A3for the LE-to-Stransition. Assuming a volume of 30 A3 DzO molecule, we are able to explain the drastic changes of the hydration by the concomitant changes of the headgroup volumes. The hydration chan e by 14 water molecules as calculated from AV = 420 from the I to the LE phase agrees well with the observed difference of 12water molecules. The 7-8 water molecules calculated from AV = 230 A3 (LE-to-S phase) are comparable to the 6 DzO molecules, which we obtained from the fits. Therefore, we conclude that the hydration state is determined by the headgroup volume accessible to the water. Our data indicate that the change ofthe monolayer hydration with the lateral pressure is fairly continuous,
i3
1 0.05
0.10
0.15
r 0.20
1
I
I
0.25
0.30
0.35
momentum transfer Q I A ' Figure 6. Experimental and theoretical neutron reflectivity profiles of the system from Figure 5 but at 9.0 mN/m.
but measurements at more pressures would be required to prove this. Acomparison of our hydration data with those obtained for the DPPC-bilayer by volumetric and scattering methods shows striking differences (Table 3).24 These discrepancies illustrate the complications which arise by comparing monolayer and bilayer data for apparently similar phase states (Le., LE and La/Sand Lc). By simply identifying the n,, of the monolayer with that of the bilayer in the Laphase, we would obtain a lateral pressure per monolayer leaflet of the bilayer of 35 mN/m. Such a value was proposed by B l ~ m e .On ~ ~the other hand, the S and La phases are structurally very different. Therefore, we conclude that the identification of corresponding phases by the hydration state is not applicable ~~
(25) Blume, A. Biochim. Biophys. Acta 1979, 557, 32.
3952 Langmuir, Vol. 11, No. 10, 1995
Naumann et al.
Table 2. Hydration Values of Pure DPPC-dez and Mixed DPPC-de&&-h (m= 2) Monolayers (Bulk Solution: D20) as Obtained from the Best Fitting Resultsa nwiDPPC-dsz + nwIEO nw/DPPC-dsz C12E4-h nwlClzE4-h group 1.4mN/m 22f2 30 f 2 16f4 4 f l 9 mN/m 10 f 2 17 f 2 14f4 3.5i 1 35 mN/m 4fl a The number of DzO molecules per surfactant molecule and per ethylene oxide (EO) group were calculated using a lipid surfactant molar ratio (rus)of 2.
Table 3. Comparison of the Hydration ( n ~for ) Monolayers and Bilayers of DPPC" phase state (monolayer)
I LE LC S
nw monolayer
22 10 4
bilayer
phase state (bilayer)
4-7 1.3 0.6
La phase Lp phase L, phase
The bilayer d a t a were taken from Nagle and Wiener (1988).24
for the bilayer-monolayer comparison, mainly because headgroup volume, molecular area, and phase states differ. Despite the above discussed restrictions for a comparison ofmonolayer and bilayer hydration, some similarities exist. The result for the LE phase a t n = 9 mN/m of n,, = 10 f2 water molecules per lipid agrees well with the results obtained by NMR methods for the bound water fraction in fully hydrated fluid lecithin bilayer^.^^^ Since the DPPC molecular area a t n = 9 mN/m is A = 77 A2, we estimate, using a n average cross-sectional area per water molecule of 8 A2, that about one layer of water is associated with the DPPC headgroup. Time-resolvingspectroscopic methods such as NMR relaxation and dynamic neutron scattering distinguish dynamically two different classes of bound water in DPPC bilayers a t a hydration of n, = 12.9J0 However, since the neutron reflection method averages over long times compared to the correlation times of the water motion, we obtain an average over all bound water molecules in the headgroup layer. In contrast, our value of n, = 22 & 2 molecules ofwater obtained a t n = 1.4 mN/m clearly exceeds the maximum amount of bound water that was observed for lecithin bilayers by NMR. For the latter, 12 molecules of bound water and another 11 molecules of quasi-free water trapped between the bilayers were observed. The sum of these two gives a value n, = 23, which compares well with our result a t n = 1.4 mN/m. Moreover, a value of n, = 23 was also obtained by X-ray measurements for fluid DPPC multilayers, though other authors report a value of n, = 37.26927We can conclude that the amount ofwater in the headgroup layer at very low lateral pressure (I phase) is similar to what is found for fully hydrated bilayers when the quasi-free trapped water is included. Taking into account the above discussion, we can assume%thatno totally free water is incorporated in the hydrophilic region of the monolayer (independent on the lateral pressure). Mixed Monolayers of DPPC and C12E4. In a previous work on this mixed system, we have shown that C12E4 readily associates with the fluid (LE) domains of a DPPC monolayer but gets expelled into the subphase a t high lateral pressure (S phase).18 It was furthermore shown in this work that the oxyethylene group of C12E4 adopts (26) Ruocco, M. J.; Shipley, G. G. Biochim. Biophys. Acta 1982,691, 309. (27) Lis, L. J.; McAlister, M.; Fuller, N.; Rand, R. P.; Parsegian, V. A.Biophys. J . 1982, 37,657.
a random coiled conformationin the monolayer headgroup region in the LE phase, thereby causing a significant increase in the lateral pressure (n)a t a fxed value of A. Since the poly(oxyethy1ene)groups is highly hydrophilic, a significant increase of hydration can be expected for the mixed monolayer, which is supported experimentally by our results (Table 2). The hydration per C12E4 molecule, estimated on the assumption that the DPPC headgroup hydration is not significantly changed as compared to the pure DPPC monolayer, is 16 k 4 (1.4 mNlm) and 14 f 4 (9 mN/m), corresponding to an average hydration per EO group of 4 & 1(1.4 mN/m) and 3.5 k 1(9 mN/m). Taking into account the used concentrations of the mixtures (far below the cmc of the surfactant), good agreement exists for the hydration values obtained for pure surfactant monolayer^.^^^^^ From this, we see that the surfactant shows a significantly less pronounced n dependence of hydration than observed for DPPC. For the latter, hydration is reduced by 55%for a change of n from 1.4 to 9 mN/m, while for the surfactant this reduction is only 13%. This difference may indicate a major difference between the headgroups of the lipid and surfactant regarding their structural changes with n. While the choline headgroup may accommodate a reduced area a t high n by changing its headgroup orientation connected with a drop of its hydration, the more randomly coiled E4 group seems not to be straightened by a pressure increase but rather gets squeezed out of the monolayer a t the transition to the crystalline phase. The observed effects of C12E4 on monolayer hydration differ markedly from what is known for bilayers. It is well-established that the presence of poly(oxyethy1ene) groups in multilayers causes a reduction of the quasi-free water as well as of the lipid-bound water and, as a consequence, a closer apposition of adjacent bilayers.30 This is thought to be the basic mechanism for the socalled chemically induced fusion of lipid bilayers.31
Conclusion We have shown that neutron reflection provides a very sensitive tool for obtaining unique information about the hydration of lipid monolayers a t the airlwater interface. The crucial point is that the measurements are performed on a subphase of D2O and that the molecular area of the lipids is known from the independently obtained film balance data. The employment of the contrast variation technique by H2O/D20 substitution is what makes the neutron reflection method particularly suited for monolayer hydration studies. The data obtained clearly support the idea of striking structural differences of phospholipid monolayers and bilayers for apparently comparable phase states. The experiments presented in this work provide a basis for studying questions of high biological relevance such as the effect of the coupling ofwater-soluble proteins with the monolayer on the hydrate structure of the latter. Such measurements are currently being performed in our laboratory.
Acknowledgment. This work was supported by grants from the Deutsche Forschungsgemeinschaft and the BMFT. LA950176S (28) Lu, J. R.; Lee, E. M.; Thomas, R. K.; Penfold, J.;Flitsch, S. L. Langmuir 1993, 9, 1352. (29) Lu, J. R.; Li, Z. X.; Su,T. J.;Thomas, R. K.; Penfold, J. Langmuir 1993, 9,2408. (30)Arnold, K.; Gawrisch, K. Methods Enzymol. 1993,220, 143. (31) Arnold, K.; Hermann, A,;Gawrisch, K.; Pratsch, L. InMoZecuZar Mechanisms ofMembrane Fusion; Ohki, S., Doyle, D., Flanagan, T. D., Hui, S. W., Mayhew, E., Eds.; Plenum: New York, 1988; p 255.