Effects of Surface Pressure on the Structure of

4070

Langmuir 2009, 25, 4070-4077

Effects of Surface Pressure on the Structure of Distearoylphosphatidylcholine Monolayers Formed at the Air/Water Interface† Clare M. Hollinshead,‡ Richard D. Harvey,‡ David J. Barlow,*,‡ John R. P. Webster,§ Arwel V. Hughes,§ Anne Weston,| and M. Jayne Lawrence‡ Pharmaceutical Science DiVision, King’s College London, The Franklin Wilkins Building, 150 Stamford Street, London SE1 9NH, U.K. and ISIS Facility, Rutherford Appleton Laboratory, Chilton, Oxon OX11 0QX, U.K. ReceiVed September 1, 2008. ReVised Manuscript ReceiVed NoVember 11, 2008 The structure of the monolayer formed at an air/water interface by the phospholipid distearoylphosphatidylcholine (DSPC) has been determined as a function of the monolayer surface pressure (π) using Brewster angle microscopy and neutron reflectivity. The microscopy studies demonstrate that the DSPC molecules form an extremely homogeneous monolayer on the water surface with no evidence of any domain formation. The neutron reflectivity measurements provide information on the thickness of the DSPC alkyl chains, head groups, and associated solvent distributions, along with the separations between these distributions and the interfacial area per molecule. Partial structure factor analyses of the reflectivity data show that the area occupied by each DSPC molecule decreases from 49 Å2 at π ) 20 mN/m to 44 Å2 at π ) 50 mN/m. There are concomitant increases in the widths of the lipids’ alkyl chains and headgroup distributions (modeled as Gaussians), with the former rising from 18 Å (at π ) 20 mN/m) to 20 Å (at π ) 50 mN/m) and the latter rising from 14 Å (at π ) 20 mN/m) to 18 Å (at π ) 50 mN/m). The compression of the monolayer is also shown to give rise to an increased surface roughness, the principal component of which is found to be the thermal roughness caused by capillary waves. At all surface pressures studied (covering the range from 20 to 50 mN/m), the alkyl chains and head groups of the DSPC are found to have roughly the same orientations, with the alkyl chains tilted with respect to the surface normal by about 34° and the head groups lying parallel to the interface normal, projecting vertically down into the aqueous subphase. Given the various trends noted on how the structure of the DSPC monolayer changes as a function of π, we extrapolate to consider the structure of the monolayer immediately before its collapse.

Introduction Neutron reflection has been widely used in determining the detailed structure of the monolayers formed at the air/water interface by single-chained surfactants,1-6 but there have been comparatively few studies of this type dealing with the insoluble monolayers formed by double-chained surfactants.7-9 As a consequence, our understanding of the interfacial behavior of double-chained surfactants is rather limited. In an attempt to rectify this, neutron reflection experiments have been performed to establish the detailed 3D structure of the monolayers formed at the air/water interface by the double-chained phospholipid † Part of the Neutron Reflectivity special issue. * Author to whom correspondence should be addressed. E-mail: [email protected]. ‡ King’s College London. § Rutherford Appleton Laboratory. | Present address: Cancer Research U.K., 44 Lincoln’s Inn Fields, London WC2 3PX, U.K.

(1) Lu, J. R.; Li, Z. X.; Thomas, R. K.; Staples, E. J.; Tucker, I.; Penfold, J. J. Phys. Chem. 1993, 97, 8012–8020. (2) Lu, J. R.; Simister, E. A.; Thomas, R. K.; Penfold, J. J. Phys. Chem. 1993, 97, 6024–6033. (3) Lu, J. R.; Li, Z. X.; Smallwood, J.; Thomas, R. K.; Penfold, J. J. Phys. Chem. 1995, 99, 8233–8243. (4) Penfold, J.; Staples, E.; Cummins, P.; Tucker, I.; Thompson, L.; Thomas, R. K.; Simister, E. A.; Lu, J. R. J. Chem. Soc., Faraday Trans. 1996, 92, 1773– 1779. (5) Penfold, J.; Staples, E.; Cummins, P.; Tucker, I.; Thomas, R. K.; Simister, E. A.; Lu, J. R. J. Chem. Soc., Faraday Trans. 1996, 92, 1549–1554. (6) Staples, E.; Thompson, L.; Tucker, I.; Penfold, J.; Thomas, R. K.; Lu, J. R. Langmuir 1993, 9, 1651–1656. (7) Barlow, D. J.; Ma, G.; Lawrence, M. J.; Webster, J. R. P.; Penfold, J. Langmuir 1995, 11, 3737–3741. (8) Barlow, D.; Ma, G.; Webster, J. R. P.; Penfold, J.; Lawrence, M. J. Langmuir 1997, 13, 3800–3806. (9) Brumm, T.; Naumann, C.; Sackmann, E.; Rennie, A. R.; Thomas, R. K.; Kanellas, D.; Penfold, J.; Bayerl, T. M. Eur. Biophys. J. 1994, 23, 289–296.

distearoylphosphatidylcholine (DSPC) at a number of surface pressures in the range of 20 to 50 mN/m. Brewster angle microscopy has also been performed to allow visualization of these DSPC monolayers, giving information on the lateral organization of the phospholipid molecules at the air/water interface and complementing the neutron reflection experiments that give information on the organization of the phospholipid molecules along the direction of the interface normal. There have been a number of previous studies conducted to determine the structure of the monolayer formed by DSPC at the air/water interface using the technique of X-ray reflectivity.10-15 These studies have generally been carried out with the DSPC monolayer maintained at a relatively low surface pressure (typically 1, which would seem to indicate that the roughness of the DSPC monolayer has both thermal and structural components. Statistical analyses, however, show that neither of the two slopes is significantly different from 1, so it is more appropriate perhaps to conclude that the intermixing of the alkyl chains and headgroups is primarily due to a capillary wave roughening of the lipid layer. If we accept that this is a reasonable approximation, then the intercepts in eqs 16 and 17 can be used to provide estimates for lzc and lzh, and from these (via eq 13) we have a self-consistent check on δch. Using eq 16, we in fact obtain lzc ) 15.5 ( 0.4 Å, and using

4076 Langmuir, Vol. 25, No. 7, 2009

Hollinshead et al.

Figure 5. Number density profiles for the DSPC monolayer maintained at a surface pressure of 30 mN/m. The profile for the lipid alkyl chains is shown as the solid line, and those for the solvent and head groups are shown as dotted and dashed lines, respectively.

eq 17, we get lzh ) 11.1 ( 0.5 Å. From these results, therefore, (after the necessary scaling of the Gaussian distribution widths by (12/8)) we find that δch should be ∼16 Å, which is the same as calculated on the basis of the structure factor analyses, so we conclude that our estimates for lzc and lzh are reasonable.

From our eq 16 estimate of lzc ) 15.5 ( 0.4 Å, it is deduced that the DSPC alkyl chains are tilted with respect to the interface. Converting this Gaussian distribution width to a dimension that may be more readily related to the C18 chain length (scaling by (12/8)) and taking the fully extended length29 of a C18 chain as ∼23 Å, the angle of tilt is obtained as cos-1(19/23), which is ∼34°. Now because lzc is here assumed to be the same at all of the measured surface pressures, the calculated angle of tilt will remain constant, and this may perhaps seem physically unreasonable. We note, however, that the elastic modulus of the area compressibility for DSPC monolayers is high (∼328 mN/ m)27 so on this basis it is to be expected that there would not be any marked changes in molecular conformation or orientation during compression (in the surface pressure range covered here). Moreover, our neutron reflectivity estimate of the DSPC chain tilt is wholly consistent with the values reported on the basis of X-ray scattering measurements for DSPC monolayers13-15,35 which, over the surface pressure range of 15 to 45 mN/m range from 30 to 36°. Our tilt estimate for DSPC is also consistent with those calculated on the basis of neutron and X-ray reflection

measurements for DPPC34,36 and also with that deduced from spectroscopic measurements for DSPC.37 In considering the interactions between the solvent and the DSPC monolayer, there are informative calculations that can be made using the derived values for the number densities of the monolayer components (nc(z), nh(z), and ns(z); eqs 4 and 5; cf. Figure 5). Specifically, we can determine the levels of headgroup hydration and the degree of wetting of the C18 alkyl chains. For the calculation of headgroup hydration, we integrate over the number density distributions for the headgroup and solvent (Figure 5) and calculate the ratio of these integrals to obtain the number of water molecules associated with each glycerophosphocholine group (within the defined integration limits). At the lowest surface pressure of 20 mN/m, such calculations give the headgroup hydration as 12.3 H2O/headgroup, calculated over a slab of 10.5 Å thickness centered at the midpoint of the headgroup distribution (at z ) 0). However, if these calculations are repeated using a slab of the same thickness but centered at the point of overlap of the C18 chain and headgroup distributions (at z ) -7.5 Å), then we find that there are 7.7 H2O/headgroup, and with the slab centered further toward the bulk, at z ) + 5 Å, there are 28.5 H2O/headgroup (at 20 mN/m). The corresponding sets of values for pressures of 30, 40, and 50 mN/m are (12.5, 7.9, 25.5), (12.1, 8,6, 23.9), and (13.0, 8.7, 24.1), respectively. It is apparent, therefore, that the lipid headgroups have differing levels of hydration depending upon the depth of their insertion into the monolayer (Figure 6), with the value computed here, for the levels of hydration at the center of the headgroup distribution (12-13 H2O/lipid), comparable to that derived on the basis of gravitometric X-ray studies on phospholipid bilayers.38 In previous reflectivity studies of phospholipid monolayers where headgroup hydration has been reported, values of around 5 waters/ headgroup are quoted.34,36 In these studies, however, the reflectivity experiments were performed for just two or three different contrasts, and only average numbers of water molecules per headgroup were reported. The analyses presented here (on

(34) Vaknin, D.; Kjaer, K.; Als-Nielsen, J.; Losche, M. Biophys. J. 1991, 59, 1325–1332. (35) Tardieu, A.; Luzzati, V.; Reman, F. C. J. Mol. Biol.1973, 75, 711733.

(36) Miano, F.; Zhao, X.; Lu, J. R.; Penfold, J. Biophys. J. 2007, 92, 1254– 1262. (37) Brauner, J. W.; Flach, C. R.; Xu, Z.; Bi, X.; Lewis, R. N. A. H.; McElhaney, R. N.; Gericke, A.; Mendelsohn, R. J. Phys. Chem. B 2003, 107, 7202–7211. (38) Jendrasiak, G. L.; Smith, R. L. Cell. Mol. Biol. Lett. 2000, 5, 35–49.

Given that the fully extended length of a glycerophosphatidylcholine group is cited30 as 10.5 Å, our estimate for lzh of 11.1 ( 0.5 Å indicates that the DSPC headgroups must lie parallel to the interface normal, each projecting vertically down into the subphase. The same observation was also noted by Brumm et al.9 and Vaknin et al.34 for DPPC monolayers at π > 30 mN/m. The repulsive electrostatic forces between the charges on neighboring lipid head groups would thus seem to be minimized, not by any in-plane alignment of the headgroup dipoles but rather by out-of-plane disorder arising chiefly from capillary waves.

Structure of Distearoylphosphatidylcholine Monolayers

Figure 6. Schematic illustration of the structure of the DSPC monolayer formed at the air-water interface at a surface pressure of 30 mN/m. Dashed lines mark distances measured along the interface normal, at z ) -8, 0, and -5 Å, measured with respect to the midpoint of the headgroup distribution. The numbers of H2O molecules per lipid are computed from the ratio of the integrated number densities for the headgroups and solvent, with an integrated slab width of 10 Å.

the basis of a simultaneous fit to reflectivity data obtained under seven different contrasts) provide a much more complete picture of phospholipid headgroup hydration. In like fashion, we are able to determine the degree of wetting of the DSPC alkyl chains, and for this we integrate over the number density distributions for the alkyl chains and solvent and calculate the ratio of these integrals to obtain the number of water molecules associated with each pair of C18 chains (within the defined integration limits). By this means, we find that, integrating over a 19-Å-thick slab centered at the midpoint of the C18 chains distribution, there are 1.2, 1.3, 1.7, and 1.9 H2O molecules per DSPC molecule at each of the surface pressures, 20, 30, 40 and 50 mN/m, respectively. It is clear, therefore, that there is always a pronounced overlap between the alkyl chains’ and solvent’s distributions (Figure 5) and that, even at π ) 50 mN/m, there is solvent found at the center of the alkyl chains’ distribution. The number densities of the alkyl chains’ and head groups’ distributions can also be used to provide some further detail on the intermixing of these two components and on the way in which this changes as a function of monolayer compression. At the point of intersection of the alkyl chains’ and head groups’ distributions, zi, the quotient (nh(zi)Vc)/nc(zi)Vh) gives the volume fraction of chains in this part of the monolayer (Vi), and because nh(zi)Vh ) nc(zi)Vc, this is also the volume fraction of the headgroups in this region. At 20 mN/m, zi is -5.2 Å, with the intersection of the alkyl chains’ and head groups’ distributions thus lying 5.2 Å from the midpoint of the solvent distribution, measured in the direction of the air side of the interface. The corresponding values for the monolayer maintained at 30, 40, and 50 mN/m are calculated to be -5.1, -4.8, and -4.5 Å, respectively. We see, therefore, that increasing compression of the monolayer has the

Langmuir, Vol. 25, No. 7, 2009 4077

effect of forcing solvent further into the alkyl chain region, and this is consistent with the earlier calculations relating to C18 chain wetting. The volume fractions of the alkyl chains and head groups, Vi, at each surface pressure are 0.26, 0.27, 0.28, and 0.31, respectively. Despite the increased general wetting of the lipids’ alkyl chains caused by increasing compression, there is also (as would be expected) an increase in the density of the layer, with around half of the layer occupied by lipid at 20 mN/m, rising to roughly three-fifths at 50 mN/m. As the culmination of these various analyses and given the fairly clear trends noted on how the structure of the DSPC monolayer varies as a function of its compression state, we pose the following question: what will be the structure of the monolayer immediately before its collapse? By linearly extrapolating from the data obtained on the monolayer at pressures in the range of 20 to 50 mN/m, we derive estimates for the different structural parameters at the collapse point27,28 of ∼65 mN/m. Taking first the interfacial area occupied by each DSPC molecule, we note that extrapolation gives an area of 42 Å2 at the collapse point, which is precisely what one would expect given that the limiting cross-sectional area for each of the molecules’ two octadecyl chains29 will be 21 Å2. For the headgroup hydration (measured at the center of the headgroup distribution), we extrapolate a value of 14-15 H2O/lipid at the collapse point. The proportion of solvent in the monolayer alkyl chains’ region is predicted to remain around 1-2 H2O/lipid at 65 mN/m, indicating that although its C18 chains are effectively close-packed at this pressure (as judged from the extrapolated a0 of 42 Å2) they are still exposed to solvent. There is no conflict between these two observations, however, because the wetting of the C18 chains is simply a function of the monolayer roughness and rises as a consequence of the increased intermixing of the alkyl chains and head groups brought about by capillary waves. At the collapse, the number density distributions for the alkyl chains and headgroups are extrapolated to intersect at zi ) -3.8 Å, a point, therefore, that is within two methylene groups’ distance from the midpoint of the solvent distribution. The proportion of the monolayer volume occupied by lipid at this point is around two-thirds, with the total volume divided as one-third alkyl chains, one-third head groups, and one-third solvent. Supporting Information Available: Guinier transformations derived from reflectivity profiles for d83-DSPC, d70-DSPC, and d13DSPC monolayers at the air/water interface at various surface pressures. Self-term partial structure factors for hcc, hhh, and hss versus κ at various surface pressures. Cross-term partial structure factors for hch and hcs versus κ at various surface pressures. This material is available free of charge via the Internet at http://pubs.acs.org. LA8028319