5878
Langmuir 2007, 23, 5878-5881
DHA-Induced Changes of Supported Lipid Membrane Morphology Dorota Thid,* Jason J. Benkoski, Sofia Svedhem, Bengt Kasemo, and Julie Gold Department of Applied Physics, Chalmers UniVersity of Technology, 412 96 Go¨teborg, Sweden ReceiVed February 22, 2007. In Final Form: April 3, 2007 Docosahexaenoic acid (DHA) is a polyunsaturated long fatty acid known to have fundamental effects on cell membrane function. Here, the effect of DHA on phosphocholine-supported lipid bilayers was measured using the quartz crystal microbalance with dissipation monitoring (QCM-D) technique. Above a concentration of 60 µM (i.e., near the critical micelle concentration), DHA had drastic effects on the viscoelastic properties of the supported membranes, suggesting a more complex process and structure than simple insertion of molecules in the bilayer. Fluorescence microscopy revealed the spontaneous formation of elongated out-growths from the bilayers, which were remarkable for their length (∼100 µm) and extensive coverage of the surface. These results demonstrate the applicability of QCM-D as a method to screen for conditions where membrane remodeling occurs but also that complementary techniques are required to describe in more detail the changes in viscoelastic properties of the membrane. These results are highly relevant for the present rapid development in the field of model lipid membranes aiming toward increased knowledge about processes occurring at biological surfaces.
Introduction The considerable effort devoted to the study of cell membraneassociated functions underlines the extent to which membranes actively participate in cellular processes. One group of molecules that has been reported to have a fundamental influence on membrane function is the polyunsaturated long fatty acids. Docosahexaenoic acid (DHA) is the longest, most unsaturated of the fatty acids found in biological systems, and it was chosen as a model molecule in the present study because its functions are currently being debated.1 More specifically, DHA has been reported to alter lipid packing, changing, for example, membrane transition temperature, curvature, inter-leaflet lipid flip-flop rate, and permeability.1,2 Also reported was the phase segregation into DHA-rich rafts, which are notable for their exclusion of cholesterol.1,2 Although DHA is normally stored in membranes as the fatty acid portion of certain phospholipids, free DHA has been suggested to possess transcellular activity.3 The few reports on the effects of free DHA using model membranes were performed on unilamellar vesicles and showed increased membrane fluidity and permeability.3,4 In the present study, 1-palmitoyl-2-oleyl-sn-glycero-3-phosphocholine (POPC) supported bilayers were used in combination with surface sensitive techniques to study the effects of DHA on lipid membranes. The quartz crystal microbalance with dissipation monitoring (QCM-D) technique is powerful in this context due to its sensitivity to small changes in both mass (via changes in resonant frequency (∆f) of the sensor) and viscoelastic properties of an adsorbed film (via the damping of the sensor oscillation due to dissipative losses (∆D)).5 For low ∆D/∆f ratios (rigid films), the change in mass (∆mSauerbrey) scales linearly with ∆f (Hz) by -17.7/n ng/(cm‚Hz), where n specifies the harmonic. In the following, all ∆f values have been normalized to the fundamental frequency. When ∆D/∆f is large (non-rigid * Corresponding author. E-mail:
[email protected]. Tel: +46 31 772 6117. Fax: +46 31 772 3134. (1) Stillwell, W.; Shaikh, S. R.; Zerouga, M.; Siddiqui, R.; Wassall, S. R. Reprod., Nutr., DeV. 2005, 45, 559. (2) Stillwell, W.; Wassall, S. R. Chem. Phys. Lipids 2003, 126, 1. (3) Onuki, Y.; Morishita, M.; Chiba, Y.; Tokiwa, S.; Takayama, K. Chem. Pharm. Bull. 2006, 54, 68. (4) Ehringer, W.; Belcher, D.; Wassall, S. R.; Stillwell, W. Chem. Phys. Lipids 1990, 54, 79. (5) Rodahl, M.; Ho¨o¨k, F.; Krozer, A.; Brzezinski, P.; Kasemo, B. ReV. Sci. Instrum. 1995, 66, 3924.
films), the linear relation (∆mSauerbrey) underestimates the true mass and viscoelastic models have to be used.6,7 QCM-D senses immobilized macromolecules and water associated with them within submicrometer distances from the sensor surface. Fluorescence microscopy complements QCM-D by offering the possibility to visualize fluorescently labeled supported membranes on the micrometer scale. Supported bilayers were formed by spontaneous fusion of small, unilamellar POPC vesicles on SiO2coated QCM-D sensors by the typical two-stage process,8 resulting in ∆f ∼ -25 Hz and ∆D ∼0.11 × 10-6 (Figure 1A,B). The effect of DHA on supported bilayers varied with fatty acid concentration. From 20 µM to 60 µM, only small increases in frequency (loss of mass; ∆f e 2.0 Hz) and dissipation (∆D e 0.3 × 10-6) were observed (Table 1, Figure 1A). In contrast, DHA concentrations of 100 µM and 200 µM caused a substantial decrease in frequency (mass uptake) of ∆f ) -7.0 and -22 Hz, respectively. The increase in dissipation was similarly dramatic, with values of ∆D ) 1.3 × 10-6 and ∆D ) 5.7 × 10-6, respectively (Table 1, Figure 1B). The resulting ∆D/∆f ratios were considerably higher (10× and 30×, respectively) as compared to the nontreated bilayer (Table 1), and the normalized frequency shifts varied with the harmonic (presented as Sauerbrey mass differences in Figure 1C) indicating a conversion of the rigid bilayer into a more complex and more dissipative film. In addition, the modeled QCM-D mass was larger than the Sauerbrey mass (Figure 1C), which is a typical consequence of viscoelastic losses.6,7 The QCM-D responses were largely reversed upon rinsing with pure buffer. For DHA concentrations of e100 µM, only a small net increase in frequency of ∼5 Hz and a small loss of dissipation of ∼ -0.25 × 10-6 remained relative to the values for the bilayer before treatment (Table 1). The DHA treatment could be repeated on the same membrane at least 3 times (data not shown), with a reversible effect on ∆D, indicating a planar membrane structure was existing between the DHA injections. However, for each DHA addition, a small net loss of mass was observed upon subsequent rinsing. Although net loss of lipid material and defect formation cannot be excluded, an escalating effect on increases in mass and dissipation shifts was observed (6) Larsson, C.; Rodahl, M.; Ho¨o¨k, F. Anal. Chem. 2003, 75, 5080. (7) Voinova, M. V.; Rodahl, M.; Jonson, M.; Kasemo, B. Phys. Scr. 1999, 59, 391.
10.1021/la700523x CCC: $37.00 © 2007 American Chemical Society Published on Web 04/25/2007
Letters
Figure 1. QCM-D data of supported POPC bilayers treated with (A) 20 µM or (B) 200 µM of DHA (O ∆f3/3, 0 ∆D3): 1, buffer; 2, POPC vesicle solution; 3, 20 µM DHA; and 4, 200 µM DHA. (C) QCM-D data in panel B expressed as mass, where solid lines are linear transformations of (O) ∆f3, (0) ∆f5, (]) ∆f7, and the dashed line is mass from modeled data.
with each subsequent DHA addition, pointing toward a systematically altered membrane composition (i.e., some POPC molecules were likely to have been replaced by DHA molecules). Only the 200 µM concentration maintained a residual increase in dissipation after rinsing, ∆D ∼0.4 × 10-6. Consequently, the ∆D/∆f ratios remained somewhat elevated after the 200 µM DHA exposure (Table 1), demonstrating at this concentration, as opposed to the lower concentrations, a non-reversible rearrangement of the bilayer structure. Fluorescence microscopy facilitated interpretation of the QCM-D response that accompanied the exposure of supported bilayers to DHA. Upon addition of 200 µM DHA, the stable, planar, rhodamine-doped bilayers instantaneously formed elongated lipid structures (Figure 2A), the number and length of which increased with time (Figure 2B). Observe how individual outgrowths in Figure 2B go in and out of focus as they extend out of the focal plane. Although the height of the protrusions could not be accurately measured by alternately focusing on their tips and bases, the difference between the two positions clearly indicated that the lipid outgrowths were three-dimensional. The lipid protrusions contained fluorescently labeled lipids, indicating mixing of the three lipid types present in the system, and stayed attached to the underlying bilayer. The worm-like structures are consistent with the considerable ∆f and ∆D shifts observed at this concentration (Table 1). Fluorescence microscopy indicated a persisting bilayer structure on the surface, as fluorescence recovered after photobleaching, with no formation of larger defects in the membrane. The possibility of POPC being replaced by DHA has to be considered, especially at 200 µM, where the effect of DHA is not fully reversible (Figure 1,
Langmuir, Vol. 23, No. 11, 2007 5879
Table 1). In line with the QCM-D data, fluorescence microscopy showed that rinsing with buffer led to partial reversal of the structural change, leaving areas enriched in the fluorescent lipid variant after rinsing (Figure 2C). The lipid extensions were most likely removed by shearing during the rinsing step and remaining material collapsed into compact assemblies. The latter can account for the slightly elevated, non-reversible ∆D response (Table 1), which is normally observed in QCM-D for intact vesicles.8 Dynamic light scattering gave further insight into the restructuring effect of DHA on the lipid membranes. It revealed the presence of ∼250 nm aggregates (presumably vesicles) in bulk solution at a DHA concentration of 200 µM. As expected, the 200 µM DHA solution interacted only weakly with bare, SiO2-coated QCM-D sensors (∆f ) 1 Hz, ∆D ) 0.16 × 10-6), as both DHA and SiO2 are negatively charged under present experimental conditions. Our results show that DHA induces dramatic structural changes on supported bilayers. Although this is the first observation, to our knowledge, of extended lipid structures growing from planar bilayers in response to fatty acids, similar behavior has been reported previously for other stimuli. For example, lipid tubes have been formed by applying mechanical forces to giant vesicles.9 Also, chemical stimuli have been shown to induce such transformations. There are at least two reports where anionic supported bilayers remodeled in a similar fashion to the present study when exposed to a cationic amphiphilic peptide10 or a cationic polymer.11 We also note that threadlike lipid/detergent structures are commonly observed when reconstituting membrane proteins.12 A suggested morphology of lipid extensions presented in Figure 2D is based on our fluorescent images (Figure 2B) and their similarity to pictures of recently reported tubular outgrowths from supported membranes.10 However, the possibility of elongated micellar structures, sometimes observed in bulk, cannot be excluded due to the limitation in resolution of the microscope. Formation of worm-like outgrowths from the supported bilayer membrane is accompanied by a several-fold increase in lipid membrane area. Since fluorescence microscopy images, similar to other reports,10 do not point toward defect formation, a substantial incorporation of DHA into the POPC membranes is necessary. In the current experimental setups, sufficient amounts of DHA are present to fulfill the area expansion. The possible mechanisms of DHA incorporation into the POPC bilayer include (i) partitioning of solubilized DHA monomers into the membrane, (ii) transfer of monomers from aggregates adsorbed onto the target membrane, or (iii) fusion of aggregates with the target membrane.13 Despite the fact that the reservoir of DHA in bulk solution exceeded the mass of the POPC bilayer for all experiments, DHA only had a strong effect on the bilayer above its critical micelle concentration (CMC ∼60 µM14). However, based on the escalating effect of DHA at repetitive additions, most likely monomers of DHA were incorporated into the membrane at all concentrations, whereas, above the CMC, this process was largely facilitated by the presence of DHA aggregates. The QCM-D responses cannot easily distinguish between outgrowth of lipid structures, and the adsorption of (8) Keller, C. A.; Kasemo, B. Biophys. J. 1998, 75, 1397. (9) Karlsson, A.; Karlsson, R.; Karlsson, M.; Cans, A. S.; Stromberg, A.; Ryttsen, F.; Orwar, O. Nature 2001, 409, 150. (10) Domanov, Y. A.; Kinnunen, P. K. J. Biophys. J. 2006, 91, 4427. (11) Rossetti, F. F.; Reviakine, I.; Csucs, G.; Assi, F.; Voros, J.; Textor, M. Biophys. J. 2004, 87, 1711. (12) Knol, J.; Sjollema, K.; Poolman, B. Biochemistry 1998, 37, 16410. (13) Chen, I. A.; Szostak, J. W. Biophys. J. 2004, 87, 988. (14) Serth, J.; Lautwein, A.; Frech, M.; Wittinghofer, A.; Pingoud, A. EMBO J. 1991, 10, 1325.
5880 Langmuir, Vol. 23, No. 11, 2007
Letters
Table 1. ∆f3/3 and ∆D Changes Relative the Underlying Bilayer upon Exposure to DHA and Re-exposure to Buffera in equilibrium with DHA DHA (µM) 20 40 50 60 100 200 a
∆f (Hz)
∆D
(10-6
)
0.8 0.23 0.6 0.19 2.0 0.22 2.0 0.33 -7.0 1.34 -22 5.70 nontreated underlying bilayer
upon re-exposure to buffer
∆Dtot × 10 /-∆ftot
∆f (Hz)
∆D (10-6)
∆Dtot × 106/-∆ftot
0.016 0.014 0.014 0.014 0.044 0.125
1.6 2.5 5.1 4.3 3.9 4.1
-0.05 -0.13 -0.08 -0.24 -0.25 0.41
0.005 0.001 0.000 0.011 0.007 0.025 0.004
6
∆Dtot × 106/-∆ftot is based on total responses due to the underlying bilayer (∆f ∼ -25 Hz, ∆D ∼0.11 × 10-6) and treatment.
Figure 2. Fluorescently labeled supported POPC membranes (A) directly after addition of 200 µM DHA, (B) 20 min later, and (C) after re-exposure to buffer. Note how the outgrowths in panel B extend out of focus, providing direct evidence that the outgrowths are threedimensional. Scale bar is 20 µm and valid for images A-C. (D) A suggestion for bilayer morphological changes in equilibrium with 200 µM DHA (the fluorescent lipids were omitted as they constituted
