pubs.acs.org/Langmuir © 2009 American Chemical Society
The Location of the Biosurfactant Surfactin in Phospholipid Bilayers Supported on Silica Using Neutron Reflectometry Hsin-Hui Shen,*,† Robert K. Thomas,† and Phil Taylor‡ †
Physical and Theoretical Chemistry Laboratory, South Parks Road, Oxford, OX1 3QZ, United Kingdom and ‡ ISIS Pulsed Neutron and Muon Source, STFC, Rutherford Appleton Laboratory, Didcot, OX11 0QX, United Kingdom Received May 6, 2009. Revised Manuscript Received October 18, 2009
We have investigated the formation of supported surfactin-phospholipid mixed bilayers using neutron reflectometry. Micellar mixtures of phospholipid (diphosphatidyl choline, DPPC), surfactin, and β-D-dodecyl maltoside were used to make the deposition. When the surfactin concentration is at its critical micelle concentration (CMC = 6 10-6 M) in the bulk solution, there is no adsorption at all on the silica. When the surfactin concentration is lowered below the CMC, a mixed bilayer of surfactin and DPPC is formed. Since surfactin does not adsorb on silica from solutions of surfactin alone, this shows that there is a strong attraction between surfactin and DPPC. The variation of adsorbed amount, composition, and structure of the adsorbed layer are consistent with the attractive interaction between surfactin and DPPC and with their respective negative and positive affinities for the silica surface. Three phospholipid isotopic contrasts were measured and used to define the composition and structure of the surfactin-phospholipid bilayer. The maximum amount of surfactin in the bilayer reaches a mole fraction of about 0.2 and this is located in the outer leaflet of the bilayer within the headgroup and part of the adjacent chain region.
Introduction Surfactin is a water-soluble, highly surface-active lipopetide produced by certain strains of Bacillus subtilis. It consists of a cyclic heptapetide (four leucines, valine, and one each of glutamic and aspartic acids) closed by a lactone group of a β-fatty acid of chain length about 14 carbons. A measure of its high surface activity is its ability to reduce the surface tension at the air/water interface to about 27 mN/m. The high amphiphilicity of the molecule enables it to interact strongly with membranes, and this brings with it a number of biological activities, for example, antiviral,1 antimycoplasmic,2,3 antibacterial,4 and hemolytic.5 However, the detailed molecular mechanism of this action is not understood, although several suggestions have been made.6-9 The interaction of surfactin with biological membranes must involve some kind of insertion into the lipid bilayers, causing permeability changes and/or membrane disruption.10 A range of surfactin-lipid systems at surfaces have been studied by Fourier Transform Infrared Spectroscopy (FT-IR),11 atomic force microscopy (AFM),9 NMR,10 and computer simulation.7 These studies *To whom correspondence should be addressed. E-mail: hsin-hui.shen@ csiro.au. (1) Vollenbroich, D.; Ozel, M.; Vater, J.; Kamp, R. M.; Pauli, G. Biologicals 1997, 25, 289. (2) Nissen, E.; Vater, J.; Pauli, G.; Vollenbroich, D. In Vitro Cell. Dev. Biol.: Anim. 1997, 33, 414. (3) Vollenbroich, D.; Pauli, G.; Ozel, M.; Vater, J. Appl. Environ. Microbiol. 1997, 63, 44. (4) Makovitzki, A.; Avrahami, D.; Shai, Y. Proc. Natl. Acad. Sci. U.S.A. 2006, 103, 15997. (5) Kracht, M.; Rokos, H.; Ozel, M.; Kowall, M.; Pauli, G.; Vater, J. J. Antibiot. 1999, 52, 613. (6) Sheppard, J. D.; Jumarie, C.; Cooper, D. G.; Laprade, R. Biochim. Biophys. Acta 1991, 1064, 13. (7) Nicolas, J. P. Biophys. J. 2003, 85, 1377. (8) Deleu, M.; Pauot, M. C. R. Chim. 2004, 7, 641. (9) Deleu, M.; Nott, K.; Brasseur, R.; Jacques, P.; Thonart, P.; Dufrene, Y. F. Biochim. Biophys. Acta Biomembr. 2001, 1513, 55. (10) Heerklotz, H.; Wieprecht, T.; Seelig, J. J. Phys. Chem. B 2004, 108, 4909. (11) Ferre, G.; Besson, F.; Buchet, R. Spectrochim. Acta, Part A 1997, 53, 623.
320 DOI: 10.1021/la9034936
have concluded that surfactin can penetrate into a phospholipid monolayer at the air-water interface. However, a monolayer is not at all representative of the bilayer structure of a membrane, and in the present paper we determine the location and concentration of surfactin incorporated into phospholipid bilayers supported on silica using neutron reflectometry. In a subsequent paper we will examine the mechanism by which a phospholipid bilayer can be destroyed by solubilization by surfactin in the external solution. The reason for choosing a supported bilayer as the model membrane is that it is more accessible to advanced techniques for probing interfaces than the corresponding bilayer in a vesicle. We have previously used neutron reflectometry to study the structure and packing of surfactin at the air/water and hydrophobic solid/aqueous interfaces. At both air/water and hydrophobic/aqueous interfaces surfactin was shown to assemble in a compact globular structure forming layers of thickness 14-15 A˚ thick with limiting areas per molecule in the region of 145 A˚2.12 For such a compact globular structure to occur the side chain must fold back toward the heptapeptide ring. We also showed that the aggregation number of a surfactin micelle is in the region of 20, that is, it forms very small micelles. Finally, we showed that surfactin does not adsorb at all at the silica/water interface. In separate experiments we have established the effectiveness of neutron reflectometry for the study of supported bilayers and especially for determining how small proteins bind and penetrate into the layer.13 This sensitivity relies on the judicious use of isotopic substitution (contrast variation) to highlight the small protein, and we use the same method to study the binding and incorporation of surfactin into a supported bilayer of diphosphatidyl choline (DPPC). (12) Shen, H.-H.; Thomas, R. K.; Chen, C.-Y.; Darton, R. C.; Baker, S. C.; Penfold, J. Langmuir 2009, 25, 4211–4218. (13) Vacklin, H. P.; Tiberg, F.; Fragneto, G.; Thomas, R. K. Langmuir 2005, 21, 2827.
Published on Web 11/11/2009
Langmuir 2010, 26(1), 320–327
Shen et al.
Article
In a neutron reflectivity measurement the specular reflectivity R is measured as a function of Q (where Q is the wave-vector transfer normal to the surface or interface and is defined as Q = 4π sin θ/λ, where θ is the grazing angle of incidence and λ is the neutron wavelength). R(Q) is approximately given by RðQÞ ¼
16π2 jFðQÞj2 Q2
where F(Q) is the one-dimensional Fourier transform of F(z), the average neutron scattering length density distribution in the direction normal to the surface. An adsorbed film is characterized in terms of its neutron scattering length density profile F(z), defined as FðzÞ ¼
X
nj ðzÞbj
j
where bj is the neutron scattering length of nucleus j and nj(z) is the number density of nuclei in the direction normal to the interface. Through these two relationships there is a direct relation between the reflectivity and the structure of the interface along the normal direction. The method is equally applicable to mixed layers because, for example, for a mixture of two components, A and B, and water (w), the scattering length density will be Flayer ¼ φA FA þ φB FB þ φW FW and isotopic substitution can be used to discriminate between the different F. The conventional method for calculating reflectivities uses the optical matrix method, which was originally developed for the reflection of light from interfaces.14 In this method the layer is divided into as many sublayers as is convenient or appropriate and a characteristic matrix is evaluated for each sublayer, from which the whole reflectivity is calculated exactly (in contrast to the approximate, but physically more clear, derivation given above). It is then compared with the observed reflectivity, and the process is repeated until a satisfactory fit is obtained. The values of the scattering lengths of the sublayers so obtained are then used to reconstruct the distribution of the constituents of the interfacial layer. Any ambiguities can usually be resolved by isotopic substitution of either water or of the constituents of the interface (see, e.g., ref 15).
Experimental Details Neutron Reflectometry. The neutron reflection measurements were carried out on the SURF reflectometer at the ISIS pulsed neutron source at the Rutherford Appleton Laboratory, UK. All measurements were done in the time-of-flight mode, using wavelengths, λ, in the range 0.5-6.6 A˚. The instruments and the procedure for making measurements have been fully described elsewhere.16 The solution was held in a Teflon container of volume approximately 13 mL, which was clamped against a silicon block between temperature-controlled aluminum and magnetic stirrer plates. Sample changes were made by injection by syringe via plastic tubes through inlet and outlet ports located on opposite sides of the container. The solution was stirred by a magnetic flea placed in a compartment in the wall of the trough (14) Zhou, X. L.; Chen, S. H.; Felcher, G. P. J. Phys. Chem. 1991, 95, 9025. (15) Lee, E. M.; Thomas, R. K. Phys. B 1989, 156, 525. (16) Holzwarth, J. F.; Kell, H.; Couderc-Azouani, S.; Vater, J.; Dietrich, U.; Heenan, R.; Gutberlet, T.; Penfold, J.; Boettcher, C. Abstr. Pap. Am. Chem. Soc. 2004, 227, U853.
Langmuir 2010, 26(1), 320–327
Table 1. Scattering Length Densities of Different Isotopic Compositions (Contrasts) of Dipalmitoylphosphatidylcholine (DPPC) properties -5
-1
btotal (10 A˚ ) Ftotal (10-6 A˚-2) bheadgroup (10-5 A˚-1) Vheadgroup (A˚3) Fheadgroup (10-6 A˚-2) bchains (10-5 A˚-1) Vchains (A˚3) Fchains (10-6 A˚-2)
h-DPPC
d31-DPPC
d75-DPPC
27.5 0.23 60.0 326 1.84 -32.5 889 -0.37
415.4 3.42 60.0 326 1.84 355.3 889 4.00
808.4 6.65 195.4 326 5.99 613.0 889 6.89
Table 2. Scattering Length Densities of Surfactin (in CmSi and D2O) properties btotal (10-5 A˚-1) M (g mol-1) Ftotal (10-6 A˚-2) bheadgroup (10-5 A˚-1) Vheadgroup (A˚3) Fheadgroup (10-6 A˚-2) bchains (10-5 A˚-1) Vchains (A˚3) Fchains (10-6 A˚-2)
surfactin (CmSi)
surfactin (D2O)
17.75 1036 1.17 19.1 1143 1.67 -1.38 373 -0.37
24.69 1036 1.63 24.0 1143 2.10 -1.38 373 -0.37
and a small magnetic stirrer plate clamped behind the sample housing The single crystals of silicon were all 125 50 25 mm3 with the long faces parallel to the (111) planes and were polished on one of the long faces. In the absence of any special precautions the surface of a (111) face always has a coating of oxide between 10 and 25 A˚ thick. This surface was cleaned using “piranha” acid, which is a 5:4:1 mixture of H2O:H2SO4:H2O2. This treatment was done for 15 min at 85 °C and was followed by copious rinsing with water (UHQ, resistivity >18 MΩ cm-1). The surface was finally treated by UV/ozone following a modification of the procedure first used by Brzoska.17 Short wavelength UV is produced by a quartz mercury vapor lamp inside a glass tube and the silicon crystal placed inside the tube, through which flows a slow current of oxygen. The silicon was left in the UV/ozone cell for 30 min. This technique removes final traces of organic impurity and also improves the hydrophilicity of the surface. Two isotopic compositions of water (two “contrasts”) were used in the neutron reflectometry experiments, D2O (98% isotopic purity) and CmSi (contrast matched to silicon). CmSi is made by mixing H2O (scattering length density (SLD) = -0.57 10-6 A˚-2) and D2O (SLD = 6.35 10-6 A˚-2) in a ratio to give a scattering length density of 2.07 10-6 A˚-2, which is the same as the SLD of silicon. The SLDs of the three isotopic DPPCs were calculated from volumes of the fragments estimated from molecular dynamics simulations.18,19 These have been shown to be in good agreement with experimental data from density and X-ray/ neutron diffraction measurements. Table 1 lists the SLDs of each isotopic DPPC contrast, which were used as the basis for the fitting of the neutron reflectometry data. The SLDs calculated for surfactin in CmSi and D2O in Table 2 are based on the estimated volume of surfactin, its composition, and the assumption that all labile hydrogens of the amino acid residues exchange with the solvent.
Phospholipid Bilayers and Their Codeposition with Surfactin. DPPC was chosen as the phospholipid for the supported bilayer partly because it is known to form high quality supported bilayers and partly because it is commercially available in four different isotopic forms. Three of these isotopes were used in the (17) Brzoska, J. B.; Shahidzadeh, N.; Rondelez, F. Nature 1992, 360, 719. (18) Armen, R. S.; Uitto, O. D.; Feller, S. E. Biophys. J. 1998, 75, 734. (19) Petrache, H. I.; Feller, S. E.; Nagle, J. F. Biophys. J. 1997, 72, 2237.
DOI: 10.1021/la9034936
321
Article
Shen et al.
present work, h-DPPC, d31-DPPC (one chain deuterated), and d75-DPPC (fully deuterated), all of which were purchased from Avanti Polar Lipids and stored at -20 °C before use. Surfactin is a naturally occurring cyclic lipopeptide produced from Bacillus subtilis. In our previous paper we used protonated and deuterated samples prepared and analyzed by ourselves. Here, because the phospholipids are readily available in various deuterated forms, sufficient contrast variation can be obtained by using just the protonated form. This was purchased from Sigma-Aldrich (purity >98%) and was used without further purification. The sample was characterized by surface tension and found to be very similar in CMC and limiting surface tension to our own preparations, which had the advantage of having been very fully characterized by mass spectrometry.12 The formation of phospholipid bilayers on solid supports has been investigated by neutron and X-ray reflection.13,20 Here we used a recently developed method in which the bilayer is deposited from micellar solutions based on the nonionic surfactant, β-Ddodecylmaltoside (DDM) as a solubilizing agent. Vacklin et al.13 have shown that DDM has no affinity for the SiO2 surface on its own and that DDM is also an efficient solubilizing agent for the deposition of phospholipid.13 Here, the supported bilayer was formed by coadsorption of DPPC with DDM and surfactin from a given molar ratio at an overall concentration of approximately 0.141 g/dm3 in D2O/CmSi, followed by rinsing with water, and then adsorption from 10 and then 100 times diluted solutions. Each concentration was allowed to equilibrate for 1.5 h with the silicon surface and then rinsed with 30 mL of water to help remove DDM before injecting the next diluted concentration. Dilution of the mixed micellar bulk solution leads to an exchange between surface and bulk aggregation which enriches the bilayer with insoluble lipid. For DPPC on its own a high density bilayer can generally be obtained after three steps of dilution, during which the DDM is eliminated.13 The reflectivity of the bilayer was measured at each stage. Phosphate buffered saline (PBS) buffer at a concentration at pH 7.4 was used as the supporting solution.
Results and Discussion DPPC-DDM-Surfactin Mixtures in the Ratio of 1:6:0.2. The incorporation of surfactin into the supported bilayer was studied using three contrasts, h-DPPC/D2O, d31-DPPC/D2O and d75-DPPC/CmSi and following the progressive dilution procedure for deposition described above with the DPPC/DDM/ surfactin in the molar ratio 1:6:0.2. At the highest total concentration, and for all three contrasts, nothing adsorbed on the silicon surface at all, even after 1.5 h, and the neutron reflectivity profile remained exactly the same as the original D2O/clean surface profile (see Figure 1a). The surfactin concentration at this point (6.0 10-6 M) is slightly above the CMC of surfactin on its own.12 Since it has already been established that a largely DPPC bilayer would be deposited in the absence of surfactin13,20 we can conclude that the mixed micellar solution contains sufficient surfactin to prevent any adsorption on the SiO2 surface. Dilution of the starting solution to 1/10th of its original concentration gives a total concentration of 2.4 10-5 M with a surfactin concentration of 6.0 10-7 M, now well below the surfactin CMC. At this concentration the neutron reflectivity profiles showed that there was rapid adsorption, faster than the time scale of the experiment (about 1 h). Following this step the 100 times diluted concentration was injected and the reflectivity from the final bilayer was recorded after rinsing with PBS buffer. The reflectivity profiles of the final bilayers (see below in Figure 3) were generally similar to those from the 10 times diluted concentration and we therefore use the results from the final bilayers, shown in Figure 1a, for a more detailed analysis. (20) Vacklin, H. P.; Tiberg, F.; Thomas, R. K. Biophys. J. 2003, 84, 177a.
322 DOI: 10.1021/la9034936
Figure 1. (a) Neutron reflectivity profiles of the final mixed DPPC-surfactin bilayers for the three DPPC contrasts after rinsing with PBS buffer at 25 °C. The concentrations of DPPC/ DDM/surfactin in the deposition process were in the ratio of 1:6:0.2. (b) The effect of further surfactin incorporation at a bulk concentration of 10-5 M on the final h-DPPC/surfactin from panel a.
The thickness of the adsorbed layer is characteristic of a phospholipid bilayer, and the established model of a bilayer was therefore used as a starting point for the analysis. Thus, the structure was separated into three sublayers, the two headgroup regions and the central chain region. The nominal phospholipid coverage of the silica is characterized by the volume fraction of DPPC chains at the center of the layer. Since the coverage of the silica surface by bilayer is found to be much less than one, the layer must be in patches. The edges of the patches will introduce an average disorder into the system, which is incorporated as two adjustable parameters, (i) the fraction of heads in the chain region, to allow for mixing disorder, and (ii) the interfacial roughness. In addition, any asymmetry in surfactin distribution will induce an asymmetry in the phospholipid headgroup distributions, and this was allowed for in part by an additional parameter of the ratio of volume fraction of head 1 (next to the silica) to that of head 2. The surfactin was characterized by the width of its distribution and the position of the center of this distribution. We have previously reported the dimensions of surfactin adsorbed at the air/water and hydrophobic solid/water interfaces and found it to be a compact ball-like structure with a diameter in the region of 1315 A˚.12 Its robustness to the hydrophobic environment suggests that its dimensions will not change in the bilayer, but here we are concerned with whether it is distributed throughout the layer and Langmuir 2010, 26(1), 320–327
Shen et al.
Article
Table 3. Fitted Parameters and Calculated Bilayer Properties for DPPC-Surfactin Supported Bilayers into Which Surfactin Has Been Incorporated properties head 1 thickness (A˚) chain thickness (A˚) head 2 thickness (A˚) phospholipid coverage ratio heads 1 to 2 fraction heads in chain surfactin thickness (A˚) surfactin penetration (A˚) surfactin coverage interfacial roughness (A˚)
h-DPPC surfactin
d31-DPPC surfactin
d75-DPPC surfactin
added SF above CMC
8.25 ( 1 38.5 ( 1 8.5 ( 1 0.48 ( 0.05 1.9 ( 0.5 0.1 ( 0.05 16 ( 0.5 13 ( 1 0.30 ( 0.05 3.2 ( 0.5
8.0 ( 1 40 ( 1 8.0 ( 1 0.53 ( 0.05 1.8 ( 0.5 0.10 ( 0.05 16 ( 0.5 14 ( 1 0.33 ( 0.05 5.1 ( 0.5
8.5 ( 1 42.5 ( 1 8.5 ( 1 0.51 ( 0.05 1.9 ( 0.5 0.08 ( 0.1 15 ( 1 14 ( 1 0.34 ( 0.05 4.1 ( 0.5
7.5 ( 1 33 ( 1 7.5 ( 1 0.44 ( 0.05 5.6 ( 3 0.11 ( 0.05 26 ( 1 15 ( 1 0.23 ( 0.05 4.7 ( 0.5
we therefore retained an adjustable width. It was also assumed that the DDM is completely rinsed away with water, as has been found in the absence of surfactin.13,20 The various parameters were adjusted until a satisfactory fit was obtained using the component volumes and scattering length densities given in Tables 1 and 2. The fitted results for the three DPPC contrasts, presented in Table 3, are consistent with each other, suggesting that the model we have used is qualitatively correct. The phospholipid chain volume fraction is about 0.45 ( 0.05, which is much smaller than would be obtained for DPPC on its own, and the surfactin volume fraction in its part of the layer is 0.30 ( 0.03 with dimensions similar to surfactin at other interfaces and a penetration of about 15 A˚ into the bilayer. The ratio of head groups between inner and outer layers is in the range of 2:1 and shows that surfactin displaces a significant fraction of phospholipids molecules from the outer layer. Since the surfactin lies partly at the level of the DPPC chains, we have assumed that surfactin also displaces chains from this region in such a way that the total volume fraction of DPPC and surfactin in the chain layer is constant across the layer (we discuss this assumption further below). When surfactin and the effects of head mixing are included in the calculation of the overall coverage of the surface, the values in Table 3 give an approximate overall coverage about 30% higher than that indicated by the DPPC volume fraction at the center of the layer, that is, about 0.6. The overall molar fraction of surfactin in the bilayer is approximately 0.2, the same as molar ratio in the bulk solution. The errors quoted in Table 3 are based on least-squares fits of the individual profiles in the region of the local minimum. The question of uniqueness of the fit is always important for such complex structures and is never easy to establish. The main questions here are whether or not the surfactin is all in the outer part of the layer and whether or not its composition of 0.25 mol fraction is well determined. The sensitivity of the experiment to the concentration and position of the surfactin is most easily demonstrated by simulations of the reflectivity profiles for different configurations of the layer. Figure 2 shows the sensitivity of simulations starting with the parameters of Table 3 to the major parameters of interest, phospholipid coverage (Figure 2a), surfactin coverage (Figure 2b) and surfactin penetration (Figure 2c). In Figure 2a the main effect is in the sharp changes in the fringe at Q values in the range from 0.03 to 0.06 A˚-1. The peak is in exactly the same position and its height increases with coverage. The reflectivity from the d31-DPPC-surfactin supported bilayer retains the high sensitivity to decreasing phospholipid coverage. The effects of surfactin coverage and surfactin penetration are shown in Figure 2b,c. Two main fringes, at 0.04-0.1 A˚-1 and 0.13-0.2 A˚-1, are observed. The original surfactin coverage is 0.34 and the simulated data shows that increasing the surfactin Langmuir 2010, 26(1), 320–327
Figure 2. Simulated reflectivity profiles of the d31-DPPC-surfactin bilayer. All other parameters are as given in Table 3 for the d31DPPC-surfactin layer.
coverage to 0.5 shifts the stronger fringe to lower Q with a higher reflectivity. Figure 2c shows that surfactin penetration in the range 0-30 A˚ makes major contributions to both fringes, and all DOI: 10.1021/la9034936
323
Article
the peaks are shifted significantly for each 5 A˚ of change. The sensitivity of these profiles to surfactin penetration is consistent with our conclusions that surfactin is located in the outer part of in the membrane. The most reliable experimental value of the surfactin penetration is 15 A˚ from the d31-DPPC data, which locates the surfactin in the outer part of the outer leaflet of the DPPC bilayer. It should be noted that one of the defects of neutron reflectometry is that it is not generally sensitive to lateral inhomogeneity. The mixed bilayer must be patchy because of the coverage and the distribution of surfactin may also be patchy. Thus, the experiment would not be able to distinguish patches of bilayer decorated around the edge of their outer leaflets by surfactin from uniform incorporation of surfactin into the outer part of the bilayer. However, the high value of the mole fraction of surfactin at 0.2 would seem to be much too high to be accommodated only as a decoration around the DPPC patches. For ideal mixing in the adsorbed layer the proportions adsorbed depend on the strength of the adsorption of the individual components. For example, Vacklin et al. have shown that in mixtures of phospholipids there is strong fractionation between bulk solution and surface which is controlled by the interaction of the individual phospholipids in the mixture with the silica surface. 13,20 Surfactin does not adsorb on the silica surface on its own. Thus, the presence of a fraction of surfactin comparable with the bulk concentration shows that there is an attractive interaction between surfactin and DPPC. This is consistent with the isothermal calorimetry (ITC) experiments done by Razafindralambo et al.21 who found a strong interaction between surfactin and POPC vesicles. Table 3 shows the results from a further experiment on the deposited bilayer. Starting with the final mixed bilayer obtained in the deposition process (h-DPPC bilayer), a surfactin concentration of 10-5 M was introduced into the bulk solution and the neutron reflectivity recorded. The profile presented in Figure 1b was recorded continuously for 3 h after addition of the higher surfactin concentration. The reflectivity after an hour shows a decrease in the phospholipid coverage to 0.37, and extra surfactin in a 14 A˚ surfactin layer on top of the bilayer (Table 3). The supported bilayer already has surfactin in the bilayer, and the additional surfactin forms a diffuse layer on the top of this mixed DPPC-surfactin bilayer. After a further 2 h the whole bilayer had disappeared, showing that surfactin had solubilized the premixed bilayer. The bilayer apparently becomes unstable because it cannot accommodate more surfactin in the bilayer without desorption. The presence of surfactin on top of the DPPC layer suggests that this is the first adsorption step. Premixed DPPC-DDM-Surfactin Layers at Other Compositions. The premixed h-DPPC-DDM-surfactin bilayers were also investigated using three surfactin compositions in molar ratios of 1:6:0.2, 1:6:0.1 and 1:6:0.05. The initial hDPPC-DDM mixture had an overall concentration of 2.1 10-4 M and the surfactin concentrations were 6.0 10-6, 3.0 10-6, and 1.5 10-6 M, respectively. The sequence of dilutions was the same as described above. The reflectivity profiles were recorded after rinsing with PBS buffer at pH 7.5 and are shown in Figure 3a-c. The discussion in the previous section has shown that there is no adsorption at the highest surfactin concentration at the ratio of 1:6:0.2 when the surfactin concentration is just above its CMC. However, for the other two compositions, this is not the case and supported bilayers form at the highest concentration and increase in coverage with dilution. The neutron (21) Razafindralambo, H.; Dufour, S; Paquot, M.; Deleu, M. J. Therm. Anal. Calorim. 2009, 95 817.
324 DOI: 10.1021/la9034936
Shen et al.
Figure 3. Reflectivity profiles of premixed DPPC-DDMsurfactin layers: (a) concentration ratio of 6:1:0.2, (b) concentration ratio of 6:1:0.1, and (c) concentration ratio at 6:1:0.05.
reflectivity profiles from the first adsorption step show this result clearly. At the 1:6:0.1 ratio there is a low coverage of bilayer and this increases for the 1:6:0.05 solution. There is no adsorption at all for the 1:6:0.2 solution. The 10 and 100 times diluted concentrations have similar reflectivity profiles for all three ratios, and the final bilayer profiles have higher coverages than in the initial step. The results from the three different ratios of DPPC-DDMsurfactin supported bilayers after the first rinse are compared directly in Figure 4. The reflectivity profile at 1:6:0.2 is the same as pure D2O because the mixed solution does not adsorb on the silica at all, but adsorption occurs at the more dilute surfactin concentrations. The reflectivity profiles show that a higher bilayer coverage results from a lower surfactin concentration. The parameters given in Langmuir 2010, 26(1), 320–327
Shen et al.
Article Table 4. Fitted Parameters and Calculated Properties of Surfactin-75% DPPC-25% DDM bilayers. There Is No Adsorption at 6:1:0.2 Ratios
Figure 4. Formation of mixed DPPC/surfactin bilayers via coadsorption with DDM at concentrations of 6.0 10-6, 3.0 10-6, and 1.5 10-6 M.
Figure 5. Reflectivity profiles of the final h-DPPC-surfactin bilayers in PBS after all dilution and rinsing steps.
Table 4 have been obtained using the constraint that there is 25% of DDM mixed with the h-DPPC (this is accounted for by adjusting the scattering length density of the lipid). Vacklin et al. studied supported DPPC-DDM bilayers on silica at various contrasts and found that 25-28% DDM remained inside the DPPC bilayer after the first injection. Because the scattering length density of DDM is quite close to h-DPPC, the effect of 25% DDM is small for this contrast. Thus, the calculated scattering length densities with 75% h-DPPC and 25% DDM in the bilayer change from 1.84 10-6 and -0.37 10-6 to 1.66 10-6 A˚-2 in the head and -0.37 10-6 A˚-2 in the chain region, respectively. For the other lipid contrasts, the scattering length densities of d75DPPC and d31-DPPC would be more strongly affected. h-DPPC is the optimum contrast for studying the effect of composition of solution on composition of the layer, and the following discussion therefore focuses only on this contrast. The fitted results for h-DPPC show that the phospholipid coverage drops sharply from 0.83 through 0.60 and 0.38 to 0 with 0, 1.5, 3, and 6.0 10-6 M surfactin, respectively, in the bulk solution. Surfactin therefore has a very powerful solubilization effect on the DPPC bilayer. Figure 5 shows the final bilayer for the three surfactin compositions after rinsing with D2O-PBS buffer. The pure h-DPPC bilayer has a coverage of 0.9 but it decreases strongly when more surfactin is present, that is, a small amount of surfactin in the Langmuir 2010, 26(1), 320–327
properties
6:1:0
6:1:0.05
6:1:0.1
head 1 thickness (A˚) chain thickness (A˚) head 2 thickness (A˚) phospholipids coverage ratio heads 1 to 2 fraction heads in chain surfactin thickness (A˚) surfactin penetration (A˚) surfactin coverage interfacial roughness (A˚)
6.5 ( 1 32 ( 1 7.5 ( 1 0.83 ( 0.05 1.2 ( 0.2 0.07 ( 0.05 0 0 0 6 ( 0.5
7.25 ( 1 35 ( 1 7.0 ( 1 0.54 ( 0.05 1.8 ( 0.5 0.1 ( 0.05 25 ( 0.5 15 ( 1 0.21 ( 0.05 3.6 ( 0.5
7.0 ( 1 34 ( 1 7.0 ( 1 0.37 ( 0.05 2.1 ( 0.5 0.1 ( 0.05 27 ( 0.5 13 ( 1 0.25 ( 0.05 5.9 ( 0.5
mixture has a large effect on adsorption at the solid-liquid interface (see Table 5). The bilayer also accommodates a significant amount of surfactin, which shows that the phospholipid bilayer interacts attractively with surfactin, enabling adsorption of surfactin on the silica surface. Discussion. Because surfactin does not adsorb on silica on its own, its adsorption must be coupled to an attractive interaction with the zwitterionic DPPC. As already noted, this is consistent with the ITC experiments done by Razafindralambo et al.21 These authors further concluded that the interaction between surfactin and POPC vesicles was driven in part by a charge (surfactin)-dipole (PC zwitterionic group) interaction. This again is consistent with the position of surfactin in the bilayer. Its position allows it to interact with both headgroup and hydrophobic chains. A number of studies have reported adsorption of mixed anionic (sodium dodecyl sulfate, SDS) and nonionic (CnEm, e.g., C12E612) surfactants at the hydrophilic silica surface. Like surfactin, the anionic SDS does not adsorb at all at the silicawater interface, but in the presence of the nonionic surfactant CnEm it is coadsorbed from solutions rich in CnEm because of the known strong interaction between the head groups of the two surfactants. The repulsive interaction of the SDS with the surface has two consequences. First, the total adsorption drops as the ratio of SDS in the external solution increases. For example, a solution composition of SDS/C12E6/0.1 M NaCl/ pH 2.4 at 10/90 and 20/80 shows strong adsorption, whereas the data for 40/60 and 50/50 show very little adsorption.12 Second, the SDS is predominantly, though not exclusively, located in the outer part of the surface bilayer. The repulsive interaction is minimized if the SDS is not immediately adjacent to the solid surface. These effects are qualitatively the same for the surfactin/DPPC mixtures and almost certainly have the same physical origin. There is one subtle difference in the behavior for surfactin/DPPC, which probably reflects a weaker interaction between the two components than in the SDS/nonionic case. In the experiments of ref 12, the combined effects of total adsorption increasing as the bulk SDS concentration decreases led to the situation that the amount of adsorbed SDS increased steadily as its bulk concentration decreased. Clearly, at a low enough concentration this trend must be reversed, that is, there must actually be a maximum at some low SDS concentration. Examination of the composition of surfactant mixtures at the air/water interface shows that very strongly interacting systems maintain a good level of each component in the adsorbed layer over a greater range of concentration than weakly interacting systems.22 This will have the effect of shifting the maximum to lower concentration of SDS. Tables 4 and 5 show a clear maximum for the surfactin/DPPC at (22) Hines, J. D.; Thomas, R. K.; Garrett, P. R.; Rennie, G. K. J. Phys. Chem. B 1998, 102, 8834.
DOI: 10.1021/la9034936
325
Article
Shen et al. Table 5. Fitted Parameters and Calculated Properties of Surfactin-h-DPPC Bilayers properties
6:1:0.0
6:1:0.05
6:1:0.1
6:1:0.2
head 1 thickness (A˚) chain thickness (A˚) head 2 thickness (A˚) phospholipids coverage ratio heads 1 to 2 fraction heads in chain surfactin thickness (A˚) surfactin penetration (A˚) surfactin coverage interfacial roughness (A˚)
8(1 39.5 ( 1 8.25 ( 1 0.90 ( 0.05 1.7 ( 0.5 0.06 ( 0.05 0 0 0 5.9 ( 0.5
7.5 ( 1 37.5 ( 1 7.5 ( 1 0.61 ( 0.05 1.9 ( 0.5 0.14 ( 0.05 22.5 ( 0.5 14.5 ( 1 0.34 ( 0.05 5.0 ( 0.5
8.0 ( 1 41 ( 1 8.0 ( 1 0.49 ( 0.05 2.1 ( 0.5 0.14 ( 0.05 19 ( 0.5 14.5 ( 1 0.31 ( 0.05 4.9 ( 0.5
8.25 ( 1 38.5 ( 1 8.5 ( 1 0.48 ( 0.05 1.9 ( 0.5 0.1 ( 0.05 16 ( 0.5 13 ( 1 0.30 ( 0.05 3.2 ( 0.5
the 1:6:0.1 ratio, and this suggests that the interaction between these two amphiphiles is less than that between SDS and C12En. While there are some similarities between the surfactin/DPPC bilayer formation and those of more conventional surfactant systems, there is also a striking difference. Adsorption increases as the chemical potential (which varies as ln (bulk monomer concentration)) of the adsorbate increases in the bulk solution. Adsorption of surfactant therefore increases with concentration up to the CMC before it reaches a plateau where the chemical potential is approximately constant because the monomer concentration is constant. In surfactant mixtures there may be subtle changes of the chemical potential of the two components at the mixed CMC, which can bring about changes in the composition of the adsorbed layer, but the main effect is that adsorption is approximately constant above the CMC. In that adsorption decreases to zero above the CMC of surfactin, the behavior of the surfactin/DPPC mixtures is not only opposite to that of conventional surfactant mixtures but also contrary to expectation. Since chemical potential generally increases with concentration and since an adsorbed mixed DPPC/surfactin layer is stable at lower concentration, it should be relatively more stable at a higher bulk concentration. A possible explanation is as follows. Above the CMC of surfactin the nature of the combined bulk solution aggregates changes to give aggregates that interact unfavorably with the surface. For example, a structure of aggregates of surfactin embedded in DPPC bilayer fragments such that the surfactin spans the width of the bilayer would interact repulsively with the silica surface because of charge repulsion. However, this would not be enough on its own to bring about the observed lack of adsorption above the CMC. The rearrangement of the aggregate to a stable adsorbed layer would also have to be too slow to occur on the time scale of the contact of the aggregate with the surface. In addition adsorption would have to be dominated by direct interaction of aggregates with the surface rather than through interaction with monomer units, which is not generally the case. The situation is further complicated by the presence of the DDM. We will return to this unusual problem in a subsequent paper. Figure 6a shows the distributions of the various fragments along the surface normal for the final supported DPPC-surfactin bilayer on the silica surface deposited from the ratio of 1:6:0.2. This shows clearly the more or less complete segregation of the surfactin into the outer layer. A possible intermediate stage of surfactin solubilization is seen in the distribution profile of Figure 6b which was obtained when the already formed mixed bilayer was exposed to 6 10-6 M surfactin in the bulk solution. A diffuse surfactin monolayer formed on top of the bilayer in the first hour, and this was followed by strong disruption and complete removal of the bilayer. It has been suggested that surfactin forms pores in membranes.6 However, our results do not show surfactin forming 326 DOI: 10.1021/la9034936
Figure 6. Distribution profiles of (a) supported surfactin-DPPC bilayers formed after deposition on a silica surface, (b) a diffuse surfactin monolayer on top of the supported layer when surfactin is added to the mixed supported bilayer at a concentration of 10-5 M.
pores that extend through the supported bilayer. The reflectivity profiles are totally inconsistent with surfactin being distributed in a more symmetrical way. Other small proteins or peptides show both symmetrical adsorption in the bilayer and attachment only to the outer leaflet. Thus, neutron reflectometry showed that melittin inserts vertically into the supported bilayers and spans the entire thickness of the membrane,23 whereas snake venom phospholipase A2 (PLA2) attaches to and penetrates only the outer leaflet.13 The most common penetrating mechanism by membrane proteins is the barrel-stave mechanism, in which peptide molecules initially adsorb horizontally along the bilayer surface. This perturbs the membrane packing and allows peptide (23) Vacklin, H. P. Ph.D. Thesis. Univeristy of Oxford, 2004.
Langmuir 2010, 26(1), 320–327
Shen et al.
Article
molecules to insert vertically into the bilayers, which leads to pore formation by bundles of helices. Surfactin must do something of this sort but the reflectometry experiment cannot distinguish any in-plane inhomogeneity of the surfactin. Segregation of the surfactin within the outer plane is in principle possible using off-specular scattering but this is not an experiment that is yet possible with the required sensitivity.
Conclusions The deposition of mixtures of surfactin and the phospholipids, DPPC, on silica using the surfactant β-D-dodecyl maltoside as a carrier leads to supported mixed bilayers provided that the concentration of the surfactin is below its CMC. If the surfactin concentration is above its CMC either during the deposition or after deposition, the bilayer either does not deposit or is removed into the solution, respectively. Since surfactin does not adsorb on silica on its own, these results show that there is a strong attractive interaction between surfactin and DPPC in the deposited bilayer.
Langmuir 2010, 26(1), 320–327
The strong solubilization effect also shows that there is an attractive interaction in the mixed solution. From neutron reflectometry, the maximum amount of surfactin in the deposited bilayer was found to be about 0.2 (mole fraction). The mixed bilayer has a structure consistent with a DPPC bilayer but with surfactin located in the outer leaflet of the bilayer. The major part of the surfactin is located in the headgroup region but some penetrates into the hydrophobic chain region, and the surfactin retains the quasispherical structure previously observed at the air/water interface and at the hydrophobic/water interface. The location of surfactin in the outer part of the layer resembles the distribution in mixtures of anionic and nonionic surfactants and suggests that this is the effect of the negatively charged surfactin interacting unfavorably with the silica. Acknowledgment. H.H.S. thanks the Swire foundation for a scholarship.
DOI: 10.1021/la9034936
327
