11616
Langmuir 2008, 24, 11616-11624
Adsorption Behaviors of DPPC/MO Aggregates on SiO2 Surfaces Zhining Wang and Shihe Yang* Department of Chemistry, The Hong Kong UniVersity of Science and Technology, Clear Water Bay, Kowloon, Hong Kong, China ReceiVed June 4, 2008. ReVised Manuscript ReceiVed July 23, 2008 The adsorption kinetics of extruded 1,2-dipalmitoyl-sn-glycero-3-phosphatidylcholine (DPPC)/1-(cis-9-octadecenoyl)rac-glycerol (monoolein, MO) aggregates on SiO2 surface at 25 °C is investigated in real time, using the dissipative quartz crystal microbalance (QCM) technique. Four adsorption pathways have been identified depending on the molar fraction of MO in the DPPC/MO system: (I) intact vesicle adsorption, (II) vesicle reorganization on a SiO2 surface, (III) supported lipid bilayer (SLB) formation, and (IV) cubosome adsorption. The results can be understood by the fact that DPPC is a lamellar phase-forming lipid, whereas MO prefers the cubic phase. Therefore, the incorporation of MO in DPPC increases the packing parameter. Equally important, MO also increases the mobility of lipid molecules and lateral pressure in the bilayers as a result of the presence of a unique cis- double bond. Before extrusion, the vesicles size increases with the MO content when XMO e 0.7 and cubosomes are formed for XMO g 0.8. The extruded DPPC/MO suspensions consist of reformed vesicles for XMO e 0.7 and filtered cubosomes for XMO g 0.8, all with a uniform diameter of ∼100 nm. Differential scanning calorimetry (DSC) further indicates that the addition of MO lowers the main phase transition temperature of DPPC and thus makes the hydrophobic interior more fluid.
Introduction In general, phospholipids display high rigidity in gel phase and low rigidity in liquid crystal phase, and these phases play a key role in the adsorption of phospholipids vesicles on solid surfaces as well as in many cellular events and deliveries with liposome-based vehicles. Introduction of a single cis- double bond into the chain of a fully saturated phosphatidylcholine (PC) derivative can lower the gel-liquid crystal phase transition temperature and thus facilitate transformations of phospholipid assemblage structures.1-3 This has motivated many studies on supported lipid bilayers (SLBs) using phospholipids with oleoyl chains.4-9 However, phospholipids with saturated acyl chains are also biologically important components and essential to membrane stability at normal and elevated temperatures.10 A balance between saturated and unsaturated phospholipids in membranes is important for many functions such as chilling sensitivity.11 Moreover, saturated phospholipids (mostly 1,2dipalmitoyl-sn-glycero-3-phosphatidylcholine (DPPC), as shown in Figure 1) are important lung surfactants, which play a major * Corresponding author. E-mail address:
[email protected]. (1) Coolbear, K. P.; Berde, C. B.; Keough, K. M. W. Biochemistry 1983, 22, 1466–1473. (2) Niebylski, C. D.; Salem, N. Biophys. J. 1994, 67, 2387–2393. (3) Williams, W. P.; Brain, A. P.; Cunningham, B. A.; Wolfe, D. H. Biochim. Biophys. Acta 1997, 1326, 103–114. (4) Faiss, S.; Lu¨thgens, E.; Janshoff, A. Eur. Biophys. J. 2004, 33, 555–561. (5) Kamo, T.; Nakano, M.; Leesajakul, W.; Sugita, A.; Matsuoka, H.; Handa, T. Langmuir 2003, 19, 9191–9195. (6) Reimhult, E.; Ho¨o¨k, F.; Kasemo, B. Langmuir 2003, 19, 1681–1691. (7) Viitala, T.; Hautala, J. T.; Vuorinen, J.; Wiedmer, S. K. Langmuir 2007, 23, 609–618. (8) Reimhult, E.; Ho¨o¨k, F.; Kasemo, B. J. Chem. Phys. 2002, 117, 7401–7404. (9) Richter, R.; Mukhopadhyay, A.; Brisson, A. Biophys. J. 2003, 85, 3035– 3047. (10) Ishizaki-Nishizawa, O.; Fujii, T.; Azuma, M.; Sekiguchi, K.; Murata, N.; Ohtani, T.; Toguri, T. Nat. Biotechnol. 1996, 14, 1003–1006. (11) Gzyl, B.; Filek, M.; Dudek, A. J. Colloid Interface Sci. 2004, 269, 153– 157. (12) Israelachvili, J. Intermolecular and Surface Forces; Wiley: New York, 1992. (13) Tieleman, D. P.; Marrink, S. J. J. Am. Chem. Soc. 2006, 128, 12462– 12467.
role in the lowering of surface tension of the air/alveolus interface.14,15 Monoglycerides are amphiphilic neutral lipid molecules and can form a reverse micellar phase and three liquid crystalline phases (lamellar, reversed hexagonal, and cubic) depending on temperature and water content.16-18 They are widely used in pharmaceuticals, food processing, paper and pulp factories, and fine chemical industries. Cubic phases of monoglycerides have been shown to deliver small molecule drugs and large proteins by oral and parenteral routes.19 Many proteins entrapped in monoglyceride cubic phases appear to retain their native conformation and bioactivity, because of the protection of biomembrane-like structure of the cubic phase.19-21 Monoolein (MO, Figure 1), a typical monoglyceride, possesses a negative spontaneous curvature and can thus assemble into nonlamellar phases, such as bicontinuous cubic phase and inverted hexagonal phase. The ability of MO to form nonlamellar phases offers many exciting opportunities for studies of membrane lipid organizations and membrane dynamics, for example, membrane fusion.22-26 The MO cubic phase is also very useful for the crystallization of membrane proteins, which occurs at locally formed patches of the highly curved bicontinuous bilayers; it was suggested that this type of bilayer structure could facilitate the arrangement of protein molecules into crystals.27-29 (14) Notter, R. H.; Tabak, S. A.; Mavis, R. D. J. Lipid Res. 1980, 21, 10–22. (15) Amrein, M.; von Nahmen, A.; Sieber, M. Eur. Biophys. J. 1997, 26, 349–357. (16) Larsson, K. J. Phys. Chem. 1989, 93, 7304–7314. (17) Lindblom, G.; Rilfors, L. Biochim. Biophys. Acta 1989, 988, 221–256. (18) Qiu, H.; Caffrey, M. Biomaterials 2000, 21, 223–234. (19) Shah, J. C.; Sadhale, Y.; Chilukuri, D. M. AdV. Drug DeliVery ReV. 2001, 47, 229–250. (20) Sadhale, Y.; Shah, J. C. Int. J. Pharm. 1999, 191, 65–74. (21) Sadhale, Y.; Shah, J. C. Int. J. Pharm. 1999, 191, 51–64. (22) Seddon, J. M. Biochim. Biophys. Acta 1990, 1031, 1–69. (23) Luzzati, V. Curr. Opin. Struct. Biol. 1997, 7, 661–668. (24) Yang, L.; Ding, L.; Huang, H. W. Biochemistry 2003, 42, 6631–6635. (25) Kamo, T.; Nakano, M.; Kuroda, Y.; Handa, T. J. Phys. Chem. B 2006, 110, 24987–24992. (26) Nakano, M.; Kamo, T.; Sugita, A.; Handa, T. J. Phys. Chem. B 2005, 109, 4754–4760. (27) Landau, E. M.; Rosenbusch, J. P. Proc. Natl. Acad. Sci. U.S.A. 1996, 93, 14532–14535.
10.1021/la801723j CCC: $40.75 2008 American Chemical Society Published on Web 09/03/2008
Adsorption of DPPC/MO Aggregates on SiO2
Langmuir, Vol. 24, No. 20, 2008 11617
Figure 1. Chemical structures of DPPC and MO.
Our main interest is to suggest a simple model to quantitatively study the interaction between unsaturated lipids and saturated phospholipids in their composite aggregates by a close examination of the DPPC/MO system. Previous studies revealed that monoglycerides can physically modulate the characteristics of cell membranes, typically modeled with vesicles and monolayers. As reported by Boyle et al.,30 the incorporation of MO in phospholipids increases both the diameter of phospholipid vesicles and the molecular mobility of PC bilayers because of the additional free volume induced by the cis- double bond of MO. In another study, Nakano et al.25,26 showed that the lateral pressure in the acyl chain region increases with the molar fraction of nonlamellar forming lipids (e.g., MO) in the bilayers of 1-palmitoyl-2-oleoylphosphatidylcholine (POPC) and egg yolk PC until the bicontinuous cubic phase takes hold, because the introduction of MO to the bilayers not only gives rise to free volumes but also breaks the size balance between the headgroup and hydrocarbon chain. Up till now, however, no studies have been reported to address the effect of monoglycerides on the (28) Pebay-Peyroula, E.; Rummel, G.; Rosenbusch, J. P.; Landau, E. M. Science 1997, 277, 1676–1681. (29) Chupin, V.; Killian, J. A.; de Kruijff, B. Biophys. J. 2003, 84, 2373–2381. (30) Boyle, E.; Small, D. M.; Gantz, D.; Hamilton, J. A.; Germant, J. B. J. Lipid Res. 1996, 37, 764–772.
adsorption behaviors of phospholipid vesicles at liquid-solid interfaces and on the formation of SLBs. A wide array of surface analytical techniques have been used to study the adsorption kinetics of phospholipids vesicles, including scanning probe microscopy (in situ atomic force microscopy (AFM)),31,32 fluorescence recovery after photobleaching (FRAP),33 attenuated total reflection Fourier transform infrared spectroscopy (ATR-FTIR),34 ellipsometry,35,36 optical waveguide lightmode spectroscopy (OWL),35 surface plasmon resonance (SPR),36 and dissipative quartz crystal microbalance (QCM).31,34,37-40 QCM with dissipation is a powerful technique (31) Seantier, B.; Breffa, C.; Felix, O.; Decher, G. Nano Lett. 2004, 4, 5–10. (32) Richter, R. P.; Berat, R.; Brisson, A. R. Langmuir 2006, 22, 3497–3505. (33) Cremer, P. S.; Boxer, S. G. J. Phys. Chem. B 1999, 103, 2554–2559. (34) Cheng, Y.; Boden, N.; Bushby, R. J.; Clarkson, S.; Evans, S. D.; Knowles, P. F.; Marsh, A.; Miles, R. E. Langmuir 1998, 14, 839–844. (35) Hook, F.; Kasemo, B.; Nylander, T.; Fant, C.; Sott, K.; Elwing, H. Anal. Chem. 2001, 73, 5796–5804. (36) Hook, F.; Voros, J.; Rodahl, M.; Kurrat, R.; Boni, P.; Ramsden, J. J.; Textor, M.; Spencer, N. D.; Tengvall, P.; Gold, J.; Kasemo, B. A. Colloids Surf. B 2002, 24, 155–170. (37) Mueller, P.; Rudin, D. O.; Tien, H. T.; Wescott, W. C. Nature 1962, 194, 979–980. (38) Sackmann, E. Science 1996, 271, 43–48. (39) Sackmann, E.; Tanaka, M. Trends Biotechnol. 2000, 18, 58–64. (40) Seantier, B.; Breffa, C.; Felix, O.; Decher, G. J. Phys. Chem. B 2005, 109, 21755–21765.
11618 Langmuir, Vol. 24, No. 20, 2008
for the studies of molecular and supramolecular deposition at liquid-solid interfaces. It can provide valuable information about both mass and conformational changes of the adsorbed molecules in real time. Furthermore, dissipative QCM is complementary to the real time optical techniques (e.g., ellipsometry, OWL and SPR) because it can provide additional insight into the mechanical and structural properties of the deposition layers. In fact, with the modern instrumental advances, the dissipative QCM technique has been increasingly used to investigate the kinetics of the SLBs formation by simultaneously recording the different changes in both the resonant frequency (f) and energy dissipation (D).6,7,31,34,37-40 It has been established with the dissipative QCM technique that phospholipid vesicles in liquid crystal phase can form SLBs on silica, glass, and mica.7,37-39 In gel phase, however, phospholipid vesicles can survive rupture at water-solid interfaces and form stable supported vesicular layers (SVLs) at experiment temperatures below the phase transition point.31,40 Similar SVLs were also observed when the adsorption occurred on surfaces of a titanium oxide surface,6 because of the weak interaction between the vesicles and the TiO2 surface. Interestingly, the SVLs were found to reorganize, which was generated at the temperature slightly lower than the phase transition temperature (Tm)40 or with the addition of cholesterol derivatives.34 In the present study, we investigate the adsorption of DPPC/ MO aggregates on a SiO2 surface using the dissipative QCM technique. The SiO2 surface is a prototype for studying transformations of phospholipid vesicles to SLBs owing to the lower stability of the latter relative to that of the former on this surface. As can be appreciated from Figure 1, the DPPC/MO aggregates are a kind of ideal saturated/unsaturated binary system, which provides opportunity to study how the tuning of molecular mobility, packing parameter, and lateral pressure shall impinge on vesicle adsorption, fusion and rupture, formation of SLB, and cubosome adsorption with the dissipative QCM technique. Combined with dynamic light scattering (DLS) and differential scanning calorimetry (DSC) measurements, the aggregate adsorption behaviors can be rationalized by some key variables of the binary system. While the molecular mobility facilitates the aggregate transformation kinetics, the packing parameter and lateral pressure affect the energetics of the assemblage structures. All of these variables appear to increase with the ratio of MO to DPPC, and thus profoundly influence the interactions among the aggregates and between the aggregates and solid surfaces, and thus the adsorption and transformation processes.
Materials and Methods Materials. DPPC (purity >99%) was obtained from SigmaAldrich (St. Louis, MO) and 1-monooleoyl glycerol (MO, purity >95%) was purchased from Danisco Cultor (Denmark). These materials were used as received without further purification. Other materials such as chloroform (99.8%, Laboratory-Scan Ltd.), tris(hydroxymethyl)-aminomethane (Tris, 99.9+%, Aldrich), and sodium chloride (NaCl, 99.8%, Riedel-deHae¨n) were of the highest purity available. For the dissipative QCM measurements, we used Tris buffer, which is commonly used for SLB studies.6,25,40 Milli-Q water (Barnstead, compact ultrapure water system) with a resistivity of 18.3 MΩ · cm was used for preparing Tris buffer containing 100 mM NaCl and 10 mM Tris, and the pH was adjusted to 8.0 with 2.0 M HCl solution. Preparation of Vesicle and Cubosomes. Large unilamellar vesicles (LUVs) and cubosomes were prepared for the QCM, transmission electron microscopy (TEM), and DLS measurements as follows. A lipid film was prepared by evaporation of chloroform
Wang and Yang under a N2 stream from a chloroform solution of DPPC/MO with various molar fractions of MO (XMO ) MMO/(MDPPC + MMO), where Mi is the total number of moles of the component i in the mixture. The lipid film was left standing overnight in a vacuum to remove the residual organic solvent, and then resuspended in Tris buffer through vortexing, followed by five freeze-thaw cycles. Uniform LUVs and cubosomes were obtained by extrusion (11 times through 100 nm polycarbonate membranes) through an extruder system (Avanti Polar Lipids, Alabaster, AL). The samples were prepared with a concentration of 0.1 mg/mL. This concentration was chosen because, in the concentration range of 0.01 to 1.0 mg/mL, the pathway for the formation of SLB from the LUVs is concentration independent.40 All of the operations above were performed at 50 °C, which is above the gel-to-liquid crystal phase transition temperature of DPPC (Tm, 41 °C).41 For DSC measurements, the vacuum treated lipid films were resuspended in Milli-Q water with a concentration of 20 mg/mL. Dissipative Quartz Crystal Microbalance. Adsorption of the DPPC/MO aggregates on SiO2-coated AT-cut quartz sensors was monitored by dissipative QCM. The QCM technique is based on the oscillation of a piezoelectric quartz crystal disk with a frequency (f) and an energy dissipation (D), which characterize the mass and the viscoelastic properties of the molecules adsorbed on the crystal surface, respectively. Under ideal conditions (when the adsorbed layer on the crystal surface is rigid), there is a linear relationship between the change in frequency and the adsorbed mass according to the Sauerbrey equation,42 ∆f ) -N · ∆m/Cf, where Cf (17.7 ng/ cm2 · Hz at f ) 5 MHz) is a mass-sensitivity constant, and N (1,3,5,...) is the overtone number of the oscillator. The dissipation factor is defined by ∆D ) Edissipated/2πEstored, where Edissipated is the energy dissipated during one oscillation, and Estored is the energy stored in the oscillating system. If not stated otherwise, dissipations and changes in normalized frequency of the third overtone (n ) 3, i.e., 15 MHz) will be presented. Dissipative QCM measurements were performed with a QCMZ500 (KSV Instruments, Finland) system equipped with a temperature control unit QCM-501.7 QCM sensor crystals (5 MHz) with a coating of silicon dioxide were purchased from Q-SENSE (Gothenburg, Sweden). All of the dissipative QCM measurements were carried out at 25 ( 0.1 °C maintained by the temperature control unit. The solutions were left to stand still in the thermostatted loop before dosing into the actual measurement chamber. Preparation of Substrates. SiO2-coated QCM sensors were washed with copious amounts of Milli-Q water. The quartz crystal sensors were then placed in a freshly prepared “piranha” solution, concentrated sulfuric acid/30% hydrogen peroxide (Riedel-deHae¨n and Sigma, respectively) ) 3:1 v/v, for 10 min. The “piranha” solution was then decanted, and the substrates were collected, rinsed with large amounts of distilled water, and dried under a nitrogen stream. The cleaned substrates were used immediately for dissipative QCM measurements. [Caution: piranha solution must be used with care!] Dynamic Light Scattering. The DLS experiments were performed using a Brookhaven 90plus/BI-MAS DLS instrument (Brookhaven Instrument Corporation, New York). A 15 mW solid-state laser source operated at 659 nm with a fixed detector angle of 90° was used to measure the particle size and size distribution of the samples in the dynamic mode. All measurements were carried out at 25 °C and were performed in triplicate. Negative Stained Electron Microscopy Negative stained samples were prepared as follows. A drop of the LUV or cubosome sample was spread on a 400-mesh copper grid coated with a carbon film, and this was followed by the addition of another drop of the staining solution (2.0 wt.% of uranyl acetate in ethanol solvent). The excess solution droplet was removed immediately using filter paper. After drying in air, the sample was characterized by JEOL 2010 microscopes with an accelerating voltage of 200 kV. Differential Scanning Calorimetry. The DSC measurements were performed by using DSC Q1000 (TA Instruments, U.S.A.) (41) Lasic, D. D. Liposomes: From Physics to Applications; Elsevier: Amsterdam. 1993; pp 63-107. (42) Sauerbrey, G. Z. Phys. 1959, 155, 206–222.
Adsorption of DPPC/MO Aggregates on SiO2
Figure 2. Dissipative QCM (∆f and ∆D) responses for the adsorption/ transformation of the DPPC/MO aggregates with different MO molar fractions on a SiO2 surface in Tris buffer.
with the reference being an empty sealed aluminum pan. In each run, 15 mg of the sample suspension was weighed in a standard aluminum pan and immediately sealed by a gentle press. The suspension sample was rapidly cooled to 10 °C and remained at this temperature for 30 min. Then the sample was heated to 85 °C at a scan rate of 1 °C/min.
Results and Discussion Dissipative QCM Responses to the Adsorption of DPPC/ MO on a SiO2 Surface. The adsorption course of the DPPC/MO aggregates of different proportions on silica surface was tracked by dissipative QCM. Figure 2 shows time-dependent changes in the sensor oscillation frequency and dissipation at the third overtone as the DPPC/MO mixture was adsorbed on the SiO2 surface. Clearly, the adsorption profiles on the SiO2 surface can be simply grouped into four categories with different molar fractions of MO, XMO, in the binary system. The first category of adsorption covers the range of XMO ) 0-0.1, in which ∆f decreases monotonically and settles down at about -900 Hz. On the other hand, ∆D in this XMO range initially increases, then slightly decreases, and finally settles down at 9.9 × 10-6 and 12.2 × 10-6, respectively, for XMO ) 0 and XMO ) 0.1. The trends of the ∆f and ∆D curves in this category are similar to those reported previously for the adsorption of mixed phospholipids7,9 on surfaces of oxidized gold41 and TiO2.6,8,43 This behavior normally corresponds to the adsorption (43) Reimhult, E.; Ho¨o¨k, F.; Kasemo, B. Phys. ReV. E 2002, 66, 051905.
Langmuir, Vol. 24, No. 20, 2008 11619
of intact vesicles on the sensor surface, by which the mass builds up monotonically with adsorption time. In other words, SVLs have been formed here. The slight decrease of ∆D in our study may be caused by the rearrangement of the adsorbed intact vesicles on the SiO2 surface and the interaction between the adsorbed intact vesicles as the adsorption density increases, which tend to make the SVLs more rigid. Alternatively, a small fraction of the adsorbed vesicles may have been ruptured, followed by the formation of some SLB domains. In the second category of adsorption (XMO ) 0.2), ∆f first decreases sharply and finally stabilizes at -690 Hz after a small increase at the time of about 2000 s. Correspondingly, ∆D shows a two step process: it sharply increases and abruptly flattens at about 13 × 10-6, and then increases again at 2000 s, and finally reaches a steady-state value of about 20 × 10-6. The QCM data shows that the mass-uptake (∆f) stays almost constant when ∆D jumps up. This means that the deposited layer becomes even softer in the second step (after 2000 s) but with little mass change. One possibility is a reorganization and fusion of the adsorbed vesicles. This behavior was observed previously in dimyristoyl(DMPC) and DPPC mixtures at the temperatures slightly lower than Tm,40 in which the vesicle adsorption and reorganization were ascribed to the changed rigidity of the vesicles as the lipid composition and temperature are varied before the adsorbed vesicles can form SLBs spontaneously. The adsorption behavior of the third category (XMO ) 0.3-0.7) is epitomized by a sharp downward peak in the ∆f-t profile and a well-defined upward peak in the ∆D-t profile, suggesting an initial adsorption of vesicles and a subsequent change to a supported bilayer. In this case, both ∆f and ∆D quickly settle down to the stabilized values, indicating rupture of the adsorbed vesicles and the attendant formation of SLBs. The frequency changes (∆fstab) for XMO ) 0.3, 0.4, and 0.5 after stabilization are -201, -126, and -103 Hz, respectively; the corresponding dissipation changes (∆Dstab) are 5.2, 2.4, and 1.9, respectively. This suggests an increasing but still incomplete vesicle-to-SLB transformation. In other words, both the SLBs and intact vesicles coexist with different proportions on the SiO2 surface that increase with XMO. For XMO ) 0.6 and 0.7, ∆Dstab is about zero and ∆fstab is about -80 Hz, indicating an almost complete transformation of the adsorbed vesicles to the SLBs. It is useful to compare the stabilized values with those reported previously. For the formation of SLBs, a final stabilized frequency change around -25 Hz was reported by other researchers.6,9,44 Our smaller ∆fstab (-80 Hz) than those mentioned above can be ascribed to the fact that these authors used unsaturated phospholipids, whereas we employed saturated phospholipids in gel phase. On the other hand, the ∆fstab value of -80 Hz is consistent with that reported by Seantier using the same saturated phospholipid (DPPC) in gel phase.31,40 Next, we turn our attention to the minimum in frequency change and the corresponding maximum in dissipation change in the dissipative QCM data for XMO ) 0.3-0.7. As noted by Richter and Brisson, two possible mechanisms could be responsible for such frequency minimum and dissipation maximum.45 The first is a much faster vesicle adsorption than the rupture of isolated vesicles so that the surface density of intact vesicles may rise significantly. The other is the rapid formation of SLBs as soon as a sufficiently high surface vesicle density is reached, which is called the critical vesicle coverage. The latter mechanism was also described previously by Kasemo et al.8,44,46 and Brisson et al.45 In addition, the resulting SLBs were also evidenced by (44) Keller, C. A.; Kasemo, B. Biophys. J. 1998, 75, 1397–1402. (45) Richter, R. P.; Brisson, A. R. Biophys. J. 2005, 88, 3422–3433. (46) Keller, C. A.; Glasmastar, K.; Kasemo, B. Phys. ReV. Lett. 2000, 84, 5443–5446.
11620 Langmuir, Vol. 24, No. 20, 2008
Figure 3. XMO-dependent diameters of the DPPC/MO aggregates (vesicles and cubosomes) before extrusion (squares, left vertical axis) and after extrusion (circles, right vertical axis) measured by DLS. Each point represents the mean ( standard deviation of three samples.
direct AFM imaging.1-3,5,32 Further studies are needed to resolve these mechanisms. In any event, the valley of ∆f is caused by the release of the trapped water in the adsorbed vesicles due to their rupture, while the peak of ∆D is a result of the transition from the adsorbed soft vesicles to a rigid and flat bilayer supported on the substrate. Finally, in the fourth category (XMO g 0.8) of adsorption, ∆f decreases while ∆D increases monotonously. However, the ranges of the variations are considerably smaller than those in the first two categories. Here all of the samples show gradual equilibration of the ∆f and ∆D values, indicating a single phase adsorption on the SiO2 surface. Furthermore, the ranges of both the ∆f drop and the ∆D rise decrease with increasing MO molar fraction in the binary mixture of DPPC/MO. More detailed discussion on this category of adsorption will be given later on after relevant data pertaining to the adsorbed species are presented and their structures and sizes become more apparent. In total, although various vesicle adsorption pathways were observed separately in different systems under different conditions, it is remarkable that the four adsorption pathways are all observed in a single unique binary system by simply tuning the molar ratios. Morphology, Phase, and Size Evolutions of the DPPC/MO Aggregates. In order to appropriately interpret the adsorption behaviors of the DPPC/MO aggregates, DLS and TEM experiments were performed on both the unextruded and the extruded DPPC/MO suspensions. The reason for looking at the unextruded samples (without being subjected to the 100 nm filter membrane extrusion) is to dissect the interaction between DPPC and MO in their aggregates by studying their phase and size. Such interaction is important to understand the adsorption of the DPPC/ MO aggregates. Presented in Figure 3 are results from the DLS measurement. For the unextruded samples, two groups are clearly distinguishable in terms of the particle diameter distributions. Specifically, when XMO e 0.7, the diameters of the aggregates, which are actually vesicles here, increase linearly (from 264 to 1547 nm) with the increase of XMO, whereas a nearly unvarying diameter of 600-700 nm is observed when XMO g 0.8. Considering that MO is a bicontinuous cubic phase forming lipid and it can form small particles in the form of cubic liquid crystals named “cubosomes” in a bulk solution,47-50 it is (47) Barauskas, J.; Johnsson, M.; Joabsson, F.; Tiberg, F. Langmuir 2005, 21, 2569–2577.
Wang and Yang
reasonable to believe that a transition from vesicles to cubosomes shall occur at a certain molar fraction of MO in the DPPC/MO suspensions. This has been verified by TEM images shown in Figure 4. It can be seen from Figure 4A, C that, before extrusion, vesicles are formed at low XMO and cubosomes are formed at high XMO. After the extrusion procedure using a 100 nm filter membrane, however, both the vesicles (Figure 4B) and the cubosomes (Figure 4D) obtained display a nearly constant diameter of about 100 nm. Table 1 summarizes the DLS data on the DPPC/MO system. It should be noticed from Table 1 that, whereas the light scattering intensity of the vesicle suspensions with low molar fractions of MO exhibits little change after extrusion, that of the cubosome suspensions (XMO g 0.8) decreases drastically as a consequence of the extrusion. Furthermore, for extruded cubosome suspensions, the signal intensity decreases with increasing XMO, finally reaching only 4.2 ( 0.6 KCPS (kilocounts per second) for the extruded pure MO dispersion. The results above can be understood as follows. As commonly perceived, vesicles are flexible species, so they can be reformed when forced through the 100 nm filter membrane, and produce uniform 100 nm diameter vesicles regardless of the original vesicle sizes. This is why little light scattering intensity change occurs after extrusion. On the other hand, cubosomes have different feats because they take the bicontinuous cubic structure, which is much stiffer than vesicles.16-18 Instead of reformation in the extruder as for the vesicles, the cubosomes are simply filtered with the 100 nm filter membrane. In other words, only the preformed cubosomes with the diameter less than 100 nm can pass through the filter membrane, whereas those larger than 100 nm will be blocked, which actually account for the majority of the cubosomes in our samples. This explains the drastic decrease in light scattering intensity of the cubosome suspensions after extrusion. It should be cautioned that DLS normally gives a systematically larger mean size of the aggregates at a given composition because larger vesicles scatter light more strongly.51,52 In addition, some multilamellar vesicles in the unextruded samples also contribute to overestimate the mean size of the vesicles. For our purpose in this work, however, the relative mean size is most important, which validates the use of our DLS data. By and large, the identification of vesicles and cubosomes at different molar fractions of MO naturally interpret the distinct difference between the adsorption behaviors of the DPPC/MO suspensions with XMO e 0.7 and with XMO g 0.8. Moreover, the decreasing mass uptake on the SiO2 surface with the increase of XMO for XMO g 0.8 can also be sensibly explicated. Phase Transitions of the DPPC/MO Vesicles. Figure 5 shows typical DSC heating curves of the DPPC/MO vesicles with different values of XMO in DI water. In the absence of MO, DPPC exhibits a sharp endothermic transition at about 41 °C (the main phase transition temperature), which is consistent with the reported value for pure DPPC multilamellar bilayer dispersions.53-55 However, the pretransition, which is reportedly at about 35 °C,53 is not detected in this study. This pretransition is normally (48) Johnsson, M.; Barauskas, J.; Tiberg, F. J. Am. Chem. Soc. 2005, 127, 1076–1077. (49) Wang, Z.; Um, J. Y.; Zheng, L.; Li, G. Chem. Lett. 2004, 33, 416–417. (50) Garg, G.; Saraf, S.; Saraf, S. Biol. Pharm. Bull. 2007, 30, 350–353. (51) Coldren, B.; van Zanten, R.; Mackel, M. J.; Zasadzinski, J. A.; Jung, H.-T. Langmuir 2003, 19, 5632–5639. (52) Egelhaaf, S. U.; Wehrli, E.; Muller, M.; Adrian, M.; Schurtenberger, P. J. Microsc. 1996, 184, 214–228. (53) Ghosh, Y. K.; Indi S., S. S. J. Phys. Chem. B 2001, 105, 10257–10265. (54) Tomoaia-Cotisel, M.; Levin, I. W. J. Phys. Chem. B 1997, 101, 8477– 8485. (55) Ho¨o¨k, F.; Rodahl, M.; Brzezinski, P.; Kasemo, B. Langmuir 1998, 14, 729–734.
Adsorption of DPPC/MO Aggregates on SiO2
Langmuir, Vol. 24, No. 20, 2008 11621
Figure 4. TEM images of the DPPC/MO aggregates: (A) unextruded vesicle; (B) extruded vesicle; (C) unextruded cubosome; (D) extruded cubosome. Table 1. Particle Sizes and Polydispersity Indexes of the DPPC/MO System Determined from DLS Measurementsa before extrusion
after extrusion
XMO
Davb (nm)
P.I.c
S.I. (KCPS)d
Davb (nm)
P.I.c
S.I. (KCPS)d
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
264.3 ( 6.8 443.2 ( 21.4 639.0 ( 57.4 800.6 ( 7.1 966.3 ( 16.6 1212.1 ( 48.8 1314.9 ( 81.1 1547.8 ( 45.1 616.5 ( 7.9 711.7 ( 31.7 713.6 ( 3.6
0.290 ( 0.054 0.397 ( 0.044 0.433 ( 0.052 0.415 ( 0.033 0.352 ( 0.014 0.471 ( 0.025 0.406 ( 0.022 0.429 ( 0.029 0.328 ( 0.014 0.233 ( 0.100 0.354 ( 0.037
287.7 ( 1.3 390.7 ( 3.5 341.5 ( 17.6 513.8 ( 5.9 430.3 ( 7.5 462.8 ( 17.6 404.9 ( 3.4 258.9 ( 1.1 273.8 ( 10.4 215.0 ( 1.4 283.0 ( 2.3
100.8 ( 0.01 90.0 ( 0.3 86.2 ( 0.1 86.3 ( 0.5 91.2 ( 1.6 92.2 ( 0.7 98.3 ( 1.2 98.4 ( 0.8 104.1 ( 1.1 107.2 ( 0.9 102.1 ( 15.9
0.078 ( 0.007 0.075 ( 0.010 0.094 ( 0.015 0.067 ( 0.008 0.066 ( 0.011 0.087 ( 0.014 0.077 ( 0.010 0.081 ( 0.013 0.121 ( 0.001 0.191 ( 0.009 0.602 ( 0.061
411.1 ( 0.5 296.9 ( 6.1 311.4 ( 4.7 355.7 ( 12.1 487.2 ( 3.3 346.5 ( 21.6 233.7 ( 10.1 200.7 ( 10.2 76.0 ( 0.7 10.1 ( 0.8 3.8 ( 0.3
a
Data represent the mean ( standard deviation of three trials.
b
Average diameter from three measurements. c Polydispersity Index.
characterized by a small and broad endothermic peak in the DSC curve, but it escaped our detection perhaps due to the low sensitivity of our DSC instrument or the shorter thermal equilibration time (30 min) in our DSC protocol, which might be insufficient for the ripple formation in the vesicles.53 With increasing amount of MO in the DPPC/MO vesicles, the phase transition point moves monotonically toward lower temperatures. This is a main result of the DSC measurement, which plausibly arises from the insertion of the MO cis-unsaturated chains into the hydrophobic region of the vesicles bilayer. Understandably, the gauche structure of the cis-double bond significantly disrupts efficient packing, makes the hydrophobic interior more fluid, enhances the molecular mobility, and hence results in the observed reduction of the gel-to-liquid crystal phase transition temperature.54 Also noteworthy is the broadening of the phase transition peak as more and more MO is incorporated into the DPPC/MO
d
Signal Intensity.
vesicle bilayers, which could be due to heterogeneous distributions of the constituent molecules and the formation of micro- to nanoscale domains in the DPPC/MO vesicle bilayers.55 A question remains as to how salt affects DSC measurements of the DPPC/MO vesicles. As shown by Sportelli et al.56 and Rudolph et al.57 salt in buffer had very little effect on the main transition temperature of DPPC vesicles (about 1 °C increase of the liquid crystal phase transition even at 3 M NaCl concentration). Therefore, we believe that our DSC results in DI water largely reflect the phase transition temperatures of the DPPC/MO systems in the low salt concentration buffer we used (10 mM Tris, 100 mM NaCl). (56) Sapia, P.; Sportelli, L. Colloids Surf., A 1993, 72, 257–263. (57) Rudolph, A. S.; Goins, B. Biochim. Biophys. Acta 1991, 1066, 90–94.
11622 Langmuir, Vol. 24, No. 20, 2008
Figure 5. DSC thermograms of the DPPC/MO aggregates in aqueous suspensions: (a) pure DPPC, (b) XMO ) 0.2, (c) XMO ) 0.4, (d) XMO ) 0.6.
Figure 6. ∆D vs ∆f plots for the adsorption processes of the MO/DPPC aggregates on a SiO2 surface at different XMO: (A) XMO ) 0-0.2 and XMO ) 0.8-1.0; (B) XMO ) 0.3-0.7.
The Four XMO-Dependent Adsorption Pathways. To facilitate discussion, Figure 6presents the dissipative QCM results in Figure 2 by plotting ∆D versus ∆f (in terms of ∆D-∆f curves). Literally, information about structural changes of vesicles in different immobilization steps can often be obtained by presenting dissipative QCM data in such a time independent manner. This way of analysis essentially concerns the change in damping for
Wang and Yang
Figure 7. Peak values of ∆f and ∆D as a function of XMO obtained from Figure 5B. The peak values are related to the minimum in ∆f (squares, left vertical axis) and the maximum in ∆D (circles, right vertical axis), respectively. Each point represents the mean ( standard deviation of two samples.
every new unit of mass adsorbed, or to put it another way, it estimates how new added mass affects the vesicle structure on the surface. As such, the ∆D-∆f curves allow more direct investigations of the mechanistic details of the vesicle adsorption as a function of XMO. As can be seen from Figure 6, the four adsorption pathways for the DPPC/MO aggregates in aqueous suspensions indeed become more obviously separated into the four ranges of the molar fractions of MO. When XMO ) 0-0.1, the ∆D-∆f curves take an arch shape (Figure 6A). In other words, ∆D slightly decreases after a maximum with the increase of ∆f, indicating rearrangement or rupture of some adsorbed vesicles as discussed above. For XMO ) 0.2, the first half-part of the ∆D-∆f curve (∆D < 13) is similar to that of the first pathways, showing the intact vesicles adsorption. However, as soon as a point is reached, ∆D soars precipitously with ∆f, even though ∆f still shows small fluctuations. As already described above, this seems to indicate the processes of vesicle reorganization and fusion, which are made possible by the incorporation of MO in the DPPC bilayers. Next, the ∆D-∆f curves belonging to the third category of adsorption displayed in Figure 6B all take a “cusp-like” shape.6,7,46 In other words, ∆D initially grows with the decrease of ∆f, and then decreases with the increase of ∆f after an inflection value. The inflection values are denoted by ∆fpeak and ∆Dpeak, respectively, and are plotted in Figure 7 as a function of XMO, the implications of which will be discussed below. It should be pointed out that this type of ∆D-∆f curve usually signals the formation of SLBs: while the SLBs that are formed are mixed with the adsorbed intact vesicles for XMO ) 0.3-0.5, the adsorbed vesicles are completely transformed into SLBs for XMO ) 0.6-0.7. Finally, the ∆D vs ∆f plots in the fourth category of adsorption appear as straight lines (XMO g 0.8, Figure 6A). This is in marked contrast to the three types of ∆D-∆f curves described above, which share a similar rising feature characterizing the initial intact vesicle adsorption. Such a linear ∆D-∆f relationship implies a single-phase adsorption, more specifically, the adsorption of cubosomes.46,58 A further note is that the adsorbed mass gets smaller with the increase of XMO, and, eventually, almost no change of ∆D and ∆f could be detected for the pure MO. As mentioned above, this is simply (58) Israelachvili, J. N.; Mitchell, D. J.; Ninham, B. W. Biochim. Biophys. Acta 1977, 470, 185–201.
Adsorption of DPPC/MO Aggregates on SiO2
Langmuir, Vol. 24, No. 20, 2008 11623
Figure 8. Schematic representation of the adsorption pathways of the DPPC (black)/MO (red) aggregates on a SiO2 surface. (I) XMO ) 0-0.1, intact vesicle adsorption and formation of SVLs; (II) XMO ) 0.2, vesicle reorganization and possible formation of larger vesicles; (III) XMO ) 0.3-0.7, vesicle rupture and formation of SLBs; and (IV) XMO ) 0.8-1.0, cubosome adsorption.
a consequence of the diminished cubosome concentrations after extrusion. The four different adsorption pathways of the DPPC/MO aggregates can now be schematically summarized in Figure 8. Here the key variable is XMO. With increasing XMO, from 0 to 1, (I) intact vesicles adsorption, (II) vesicles reorganization on the SiO2 surface, (III) SLB formation, and (IV) cubosome adsorption take place in that order. Influence of MO on the Aggregate Deposition and Transformation. According to previous studies, the processes of lipid vesicle deposition and SLB formation on solid surfaces can be affected by many factors, including vesicle size,6,8 lipid phase transition temperature,40 charge density,9,32 osmotic pressure,6 lipid concentration40 of the vesicles, and so on. In our case, some factors need be considered as well. The first is of kinetic nature, namely, the increased molecular mobility in the phospholipid bilayer with the addition of MO. This is demonstrated in our DSC result, which shows a significantly decreased melting point when more and more MO is introduced into the DPPC phospholipid bilayer. In fact, Boyle et al.30 also showed previously that the molecular mobility in the phospholipid bilayer increases with the incorporation of MO since in the inner space of the lipid vesicle bilayer, the cis- double bond of MO brings in more free volume for molecule motion. Understandably, the higher molecular mobility makes the reorganization and rupture of the adsorbed vesicles much easier. This argument tallies with the well-documented fact that SLBs are often formed for vesicles in liquid crystal phase, whereas SVLs can be formed for vesicles in gel phase.40 Second, the lateral pressure of the phospholipid hydrocarbon chains increases as a result of the insertion of the cis- double bond of MO.25,26 By forming structures with smaller curvatures, the lateral pressure can be released, and this also explains the MO-induced penchant to form planar lipid bilayers with a zero spontaneous curvature and cubosomes with a negative curvature.25,26 This thermodynamic consideration highlights the effect of the lipid packing parameter P on the adsorption behaviors of the DPPC/MO vesicles. The packing parameter is defined by P ) V/a0lc, where V is the volume of the hydrophobic part of the lipid molecule, a0 is the mean cross-sectional headgroup surface
area, and lc is the length of the extended all-trans alkyl tail. When P ∼ 1, lipid aggregates tend to possess zero spontaneous curvature, exemplified by lamellar phases or vesicles. Double-tailed phospholipid, DPPC, has a packing parameter of around 1 and thus a cylinder-like shape, which is commonly known as a lamellar-forming lipid. On the other hand, the packing parameter of MO is above 1, indicating a larger hydrophobic volume ratio than in DPPC caused by the cis- double bond in MO. As a result, pure MO in water is apt to form cubosomes. One can expect that adding the nonlamellar-forming lipid MO to the DPPC bilayers leads to imbalance between the volume sizes of the headgroup and the acyl chain, and thus an increase in the packing parameter. When XMO e 0.7, the diameters of the vesicles increase with increasing XMO, as verified by DLS for the unextruded samples. Also increased is the lateral pressure of the lipid vesicle bilayer.25,26 As soon as XMO is above 0.7, the spontaneous curvature decreases to such extent that it becomes negative, leading to the formation of the cubic phase in the form of cubosomes (see Figures 3 and 4 above). With the reciprocal relationship of the packing parameter and spontaneous curvature, the adsorption behaviors of the DPPC/ MO vesicles can be qualitatively explained. When XMO ) 0-0.1, MO appears to have little effect on the viscoelastic behavior of the DPPC vesicles due to the insufficient increase of the packing parameter, so the resulting SVLs are thus stable. As XMO is increased to 0.2, the packing parameter has increased to such extent that it could enhance the possibility for the reorganization and fusion of the adsorbed extruded DPPC/MO vesicles around this composition to lower the lateral pressure. The increase of the packing parameter is corroborated by the DLS data: the mean diameter of the unextruded vesicles is found to increase to >600 nm for XMO ) 0.2 from 200 nm for XMO ) 0 and 400 nm for XMO ) 0.1. Still more incorporation of the MO (0.3 e XMO e 0.7) continues to increase the packing parameter such that more and more planar lipid bilayers with a zero spontaneous curvature are eventually formed with the increase of XMO. In the MOdominant regime when XMO g 0.8, the mass uptake (∆f) and viscoelastic property change (∆D) can be both accounted for by
11624 Langmuir, Vol. 24, No. 20, 2008
the adsorption of cubosomes characterized by negative spontaneous curvatures due to the sway of the large packing parameter of the MO. Finally, it is notable in the SLBs formation (pathway III) that ∆fpeak increases and ∆Dpeak decreases linearly with increasing XMO (Figure 7), suggesting a more and more facile transformation from SVLs to SLBs on the SiO2 surface with increasing XMO. This may be interpreted by the decreasing bending rigidity and the increasing molecular mobility of the vesicles as XMO increases. It is known that the DPPC vesicles in gel phase possess a high bending rigidity, but, with the increase of XMO, the binding rigidity is expected to decrease, which is consistent with the decreasing trend of the phase transition temperature we observed. Another possibility is associated with the likelihood for MO with OH end groups to form hydrogen bonds with the SiO2 surface, which could also facilitate the transformation from SVLs to SLBs on the SiO2 surface with increasing XMO. Clearly, this raises more questions for further studies.
Conclusions In summary, the adsorption and subsequent structural changes of the DPPC/MO vesicles on SiO2 surface have been investigated.
Wang and Yang
For the first time, the QCM study has allowed us to identify four distinctive vesicle adsorption pathways in this binary system with broadly varying molar ratios. By simply adjusting the MO molar fraction, we are able to control the pathway of adsorption on the SiO2 surface. To account for the observed four kinetic regimes in the adsorption of the DPPC/MO vesicles, we have considered possible factors of both kinetic and thermodynamic nature, namely, the molecular mobility, lateral pressure, and, to a lesser extent, hydrogen bonding, which all increase with the molar fraction of MO. In this way, we are able to rationalize the structural transformations for the DPPC/MO system on the SiO2 surface, starting from SVLs to reorganized vesicles and SLBs, and to cubosomes as XMO increases from 0 to 1. Work on a more detailed molecular understanding of the vesicle adsorption and transformation pathways is underway in our laboratory and collaborating laboratories, including monolayer studies, in situ imaging, and atomistic/coarse-graining modeling. Acknowledgment. This work is supported by grants of RGC (604206 and 604107) administrated by the UGC of Hong Kong and HKUST (RPC06/07.SC03). LA801723J
