Article pubs.acs.org/Langmuir

Lipid Exchange and Transfer on Nanoparticle Supported Lipid Bilayers: Effect of Defects, Ionic Strength, and Size Jelena Drazenovic, Selver Ahmed, Nicole-Marie Tuzinkiewicz, and Stephanie L. Wunder* Department of Chemistry, Temple University, Philadephia, Pennsylvania 19122, United States S Supporting Information *

ABSTRACT: Lipid exchange/transfer has been compared for zwitterionic 1,2-dimyristoyl-sn-glycero-3-phosphocholine (DMPC) and 1,2-dimyristoyl-d54-sn-glycero-3-phosphocholine (DMPC) small unilamellar vesicles (SUVs) and for the same lipids on silica (SiO2) nanoparticle supported lipid bilayers (NP-SLBs) as a function of ionic strength, temperature, temperature cycling, and NP size, above the main gel-to-liquid crystal phase transition temperature, Tm, using d- and hDMPC and DPPC. Increasing ionic strength decreases the exchange kinetics for the SUVs, but more so for the NP-SLBs, due to better packing of the lipids and increased attraction between the lipid and support. When the NP-SLBs (or SUVs) are cycled above and below Tm, the exchange rate increases compared with exchange at the same temperature without cycling, for similar total times, suggesting that defects provide sites for more facile removal and thus exchange of lipids. Defects can occur: (i) at the phase boundaries between coexisting gel and fluid phases at Tm; (ii) in bare regions of exposed SiO2 that form during NPSLB formation due to mismatched surface areas of lipid and NPs; and (iii) during cycling as the result of changes in area of the lipids at Tm and mismatched thermal expansion coefficient between the lipids and SiO2 support. Exchange rates are faster for NPSLBs prepared with the nominal amount of lipid required to form a NP-SLB compared with NP-SLBs that have been prepared with excess lipids to minimize SiO2 patches. Nanosystems prepared with equimolar mixtures of NP-SLBs composed of d-DMPC (dDMPC-NP-SLB) and SUVs composed of h-DMPC (hDMPC-SUV) show that the calorimetric transition of the “donor” hDMPCSUV decreases in intensity without an initial shift in Tm, indicating that the “acceptor” dDMPC-NP-SLB can accommodate more lipids, through either further fusion or insertion of lipids into the distal monolayer. Exchange for d/hDMPC-NP-SLB is in the order 100 nm SiO2 > 45 nm SiO2 > 5 nm SiO2.

■

lipid bilayers (SLBs).5 While SLBs can be used as surrogates for vesicles and cells6,7 and to study membrane−membrane interactions,8 there are important differences between them, in particular, the presence of an underlying solid support for the SLBs that can affect biological processes. SLBs and nanoparticle (NP) SLBs are under investigation for drug delivery, sensor,6,7 and other biotechnological applications and to make biocompatible interfaces. Lipids in the form of vesicles or NP-SLBs can adsorb hydrophobic organic pollutants9,10 and therefore may be useful for environmental remediation of these toxins. Adsorption of lipids onto inorganic oxides such as pyrite and ferrihydrite protects them from bacterial oxidation11 and dissolution.12 For many of these applications, lipid transfer/ exchange needs to be utilized or understood. Lipid transfer between vesicles and SLBs has been used to change the lipid composition of the bilayer,13−16 often by exchange of oppositely charged lipids.8 These modified surfaces enable an understanding of how lipids on the surface of the SLBs or NP-

INTRODUCTION Lipid transfer or exchange between cell membranes, lipid vesicles, and supported lipid bilayers (SLBs) and between supported bilayers and membranes is important in a variety of areas, including those that have biologically relevance or technological applications. Exchange is an equilibrium process between donor and acceptor structures of equivalent chemical potential, while transfer involves redistribution of lipid between donor and acceptor structures of nonequivalent chemical potentials.1 Exchange/transfer interbilayer kinetics are distinct from intrabilayer (also called flip-flop) kinetics that equilibrate lipids in the two leaflets of the bilayer,2 and simple models have been developed that describe the processes using two independent kinetic parameters, i.e., the rates of transbilayer and interbilayer exchange.3 Flip-flop time scales of between 1 and 90 h depending on the vesicle type (size and composition) have been reported.4 Although these processes in biological systems are often mediated by membrane proteins, spontaneous exchange and flip-flop, which occur without the aid of membrane proteins and without complete fusion of the two bilayers, have been extensively investigated in vesicles1 but less so for supported © 2014 American Chemical Society

Received: October 18, 2014 Revised: November 24, 2014 Published: November 25, 2014 721

DOI: 10.1021/la503967m Langmuir 2015, 31, 721−731

Article

Langmuir

(DPPC, 16:0 PC), 1,2-dipalmitoyl-d62-sn-glycero-3-phosphocholine (DPPC-d62, 16:0 PC), which will be referred to as hDPPC or dDPPC (Avanti Polar Lipids, Alabaster, AL), were used without further purification. All solutions/suspensions were prepared with chloroform and HPLC grade water, with or without 0.1 M PBS buffer (100 mM NaCl solution and 2 mM KCl) or 10 mM NaCl (Fisher Chemicals, Fairlawn, NJ). An Avanti Mini-Extruder was employed for extrusion of the lipids, using 50 or 100 nm pore size polycarbonate filters. SiO2NPs (lot #200109) were a gift from Nissan Chemicals; this SiO2 contains no surfactants. Preparation of Multilamellar (MLV), Small Unilamellar (SUV) Vesicles and Supported Lipid Bilayers (SLBs). The lipids (hDMPC, dDMPC, hDPPC, or dDPPC) were dissolved in chloroform and dried under a nitrogen stream, followed by drying in a vacuum oven at room temperature. The films were hydrated above the lipid Tms (40−50 °C) using HPLC grade water, 10 mM NaCl, or 0.1 M PBS buffer for 1−2 h with constant shaking to produce multilamellar vesicles (MLVs). The MLVs underwent five freeze−thaw cycles followed by extrusion (40×) through a 50 or 100 nm polycarbonate membrane to produce small unilamellar vesicles (SUVs). To indicate the composition of the SUVs, they will be referred to as hDMPC-SUV, dDMPC-SUV, hDPPC-SUV, and dDPPC-SUV. Supported lipid bilayers (SLBs) were prepared using 5, 45, or 100 nm SiO2 beads. To indicate the size of the SiO2-NPs, they will referred to as (5 nm) NP-SLB, (45 nm) NP-SLB, or (100 nm) NPSLB. Thus, hDMPC-(100 nm) SLB will refer to a SLB made from 100 nm SiO2 and h-DMPC lipids. The SUVs and beads were brought to the same temperature before mixing, after which the SUV/bead mixtures were incubated for 1 h at 40 °C for DMPC and 55 °C for DPPC, in water, 10 mM NaCl, or 0.1 M PBS. The amount of SiO2 added is indicated by that required to achieve single bilayer coverage, i.e., where the surface area of the lipids in the SUVs (SASUV) is equal to the nominal surface area of the SiO2 beads (SASiO2), SASUV/SASiO2 = 1/1, or twice that amount, SASUV/ SASiO2 = 2/1. Analysis. Nanodifferential Scanning Calorimetry (Nano-DSC). Nanodifferential scanning calorimetry (nano-DSC) measurements were obtained on a TA Instruments (New Castle, DE) Nano DSC6300. Samples were scanned at cooling/heating rates of 1 °C/min, using 0.5−1 mg of lipid. Dynamic light scattering (DLS) data were obtained for the vesicles and NP-SLBs using a Malvern (Malvern Instrument Ltd. Malvern, U.K.) Zetasizer Nano-ZS. Diameters are reported either as z averages (Dz) or volume/number distributions obtained by cumulant analysis. DLS data for DMPC-SUVs, (Supporting Information Figure 1), indicate Dz = 77.8 nm, with a polydispersity index PDI = 0.081 for hDMPC, and Dz = 79.3 nm, PDI = 0.063 for dDMPC. Exchange Experiments. Two types of exchange experiments were performed. In both, equimolar amounts of h- and d- SUVs or NP-SLBs were mixed. In one experiment, newly mixed samples were placed in the nano-DSC at 4 or 25 °C, ramped (or heated at 2 °C/min) to the prechosen incubation temperature, and held at that temperature for a preset time and the cooling cycle, at 1 °C/min monitored; a new sample was used for each incubation time. For long incubation times (>64 h), the mixtures were incubated in an oven. In the other (cycling) experiment, the separate samples were mixed at 4 °C and transferred to the nano-DSC at 4 °C. The mixed sample was then cycled successively between 4 °C and a temperature higher than that of the highest Tm of the mixture and held at that temperature for a preset time (10−15 min). In the case of DMPC, a Tmax of 35 °C was chosen; higher temperatures produced transfer rates that were too fast. In the case of DPPC, Tmax = 55 °C was chosen for the same reason. In both experiments, fast scan rates (1 °C/min) were used in order to obtain as “static” a picture of the lipid exchange as possible. In separate experiments, 50/50 dDMPC/hDMPC or 50/50 dDPPC/hDPPC MLVs were mixed together initially. SUVs were prepared and adsorbed to the SiO2 as above in order to measure their calorimetric spectra for reference. Similar exchange experiments were also performed using d-NP-SLBs and h-SUVs, added in equimolar lipid ratios. This pairing maximized the initial peak separations.

SLBs interact with lipids on other surfaces or in vesicles and cell membranes. In environmental contexts, it is important to understand the mechanism of exchange/transfer between lipids originally on NPs used for remediation and those subsequently encountered in the environment, where the effects of diurnal temperature cycling may become relevant. The mechanism by which spontaneous exchange/lipid transfer occurs in bilayers has been the subject of much theoretical and experimental research.1 In the case of vesicles in the bulk solvent, it has generally been believed that transfer occurs by diffusion of the monomers through the aqueous media, where the rate-limiting step is desorption of lipid molecules from the donor bilayer into the aqueous phase and is therefore concentration independent.17 The kinetic energy of collisions between the vesicles is transferred between them and thus does not typically give rise to fusion.13 Single molecule diffusion has also been proposed as the mechanism for lipid exchange for SLBs on large SiO2 beads.5 At higher temperatures and concentrations, lipid transfer in vesicles can also occur as the result of collisions.1,18−20 However, two other models have been proposed, both of which require that the two lipid surfaces are near each other, typically as the result of a transient collision or as the result of more prolonged attachment (attachment−transfer−detachment13,15,21−23) of the two bilayers. In the transient collision or attachment−transfer−detachment13,15,21−23 models, monomer diffusion occurs across the ∼1.5 nm hydration barrier between the donor and acceptor vesicles. The mechanism of lipid transfer was suggested to be direct insertion of lipids into the distal monolayer without flip-flop.24 In the hemifusion model, direct contact between the donor and acceptor surfaces allows mixing of the outer (distal) monolayer of lipids.1,21,25 Despite the progress that has been achieved in understanding spontaneous lipid exchange/transfer, there are still interesting questions that have not been answered, particularly for NPSLBs. The effects of interest include ionic strength, defects, temperature cycling, and NP size: (i) Ionic strength might be expected to affect exchange/transfer kinetics since the presence of salt affects the strength of the interactions between the lipid head groups and between the lipids and underlying support; bilayer adsorption/fusion on SLBs is known to depend on the ionic strength of the aqueous media.26,27 (ii) The presence of defects might affect both the rate and mechanism of exchange/ transfer, since defects can be sites for more facile removal, insertion, and flip-flop of lipids28 and/or for further vesicle adsorption and fusion. (iii) Temperature cycling above and below Tm may induce defects and has relevance for NP-SLBs that exist in the environment and experience diurnal/seasonal temperature changes. (iv) Curvature of lipids on NPs can affect their packing parameters, defect density, and thus the ability to desorb from the bilayers. Defect densities in curved membranes, although hypothetical, have been proposed to affect protein binding affinities,29 2H NMR evidence strongly suggests that order parameters in lipid bilayers on silica supports in water have a dependence on curvature,30 and the permeation of cells by vesicles has been found to be dependent on their size (i.e., curvature).31

■

EXPERIMENTAL SECTION

Materials and Methods. Materials. 1,2-Dimyristoyl-sn-glycero-3phosphocholine (DMPC, 14:0 PC), 1,2-dimyristoyl-d54-sn-glycero-3phosphocholine (DMPC-d54, 14:0 PC), which will be referred to as hDMPC or dDMPC, and 1,2-dipalmitoyl-sn-glycero-3-phosphocholine 722

DOI: 10.1021/la503967m Langmuir 2015, 31, 721−731

Article

Langmuir

Figure 1. Time-dependent nano-DSC traces of (top left) h/d DMPC-SUV extruded through 50 nm pore-size filters (40×, after five freeze/thaw cycles) and (bottom left) h/dDMPC-(100 nm)NP-SLB (SASUV/SASiO2 = 1/1) prepared using these h/dDMPC-SUV; both in 10 mM NaCl at 40 °C. Cooling cycles are shown. (right). Time-dependent changes in the mole fractions of the two populations, xI and xII, for h/dDMPC-SUV and h/dDMPC(100 nm)NP-SLB (SASUV/SASiO2 = 1/1) in 10 mM NaCl at 40 °C. Schematic of exchange process is also shown. Analysis of Mixing from Nano-DSC Data. The exchange rate was determined by the method proposed by Bayerl and Sackmann,20 where the mole fractions were calculated from the shift in Tm values at successive incubations times after mixing. These had the transition temperatures of the pure materials (in all cases Tm(d) < Tm(h)), and we will refer to them only as d and h. After an incubation time t, there were also two vesicle populations, I and II, with transition temperatures at time t, TIt and TIIt, which contain combinations of d and h, and where Td < Th, so that TIt increases and TIIt decreases (TIt < TIIt) until they merge, and where xI and xII are the mole fractions of h in the mixtures. The off rate constants, koffd (for dDMPC dissociated from vesicle population I) and koffh (for hDMPC dissociated from vesicle population II), were obtained according to the method of Bayerl and Sackmann,20 which was based on the kinetic model of Thilo.32 The values of koffd and koffh for times when the amount of exchanged lipid was less than 20 mol % were averaged. Values of t1/2 indicate the time when 50% of the lipids were transferred from the vesicles and were obtained from the xI and xII data, and averaged, by fitting the curves to a single exponential, when possible.

■

Figures 3−8), show that the convergence of the separate peaks is more rapid for the SUVs compared with the NP-SLBs. As previously observed, both the d-SUV and d-NP-SLB transitions are shifted ∼4 °C below those of the h-SUV and hNP-SLB transitions,20 and the NP-SLBs peaks are broadened and shifted ∼2−3 °C lower than the corresponding SUVs.33 The corresponding mole fractions (xI and xII) for the populations of mixed d-/hDMPC-SUV and d-/hDMPC-(100 nm)NP-SLB at time t (TI and TII) are shown for 10 mM NaCl at 40 °C in Figure 1 and in water, 10 mM NaCl and PBS at 30, 40, and 50 °C in Supporting Information Figure 11. The values of koff, t1/2, and estimates of the activation energies, ΔEact (obtained by Arrhenius plots of linear fits of ln k vs 1/T (K)), for the complete set of the data are summarized in Table 1. The results indicate that the exchange rates increase with increasing temperature (as expected), decrease with increasing ionic strength, and were greater for the SUVs compared with the NPSLBs at comparable temperatures and ionic strengths. Effect of Temperature Cycling above and below Tm. In order to investigate the effects of temperature cycling above and below the main gel-to liquid crystal phase transition, Tm, the successive scanning method was used. Little exchange is expected below Tm in the short times (5 min) the samples were at those temperatures (the rate of exchange is 1 order of magnitude slower in the gel than in the liquid crystalline phase13,34), so that the only other difference is the effect of cycling. Nano-DSC exchange data (Table 2) for d/hDMPC-(100 nm)NP-SLB in 0.1 M PBS buffer (100 mM NaCl) that have been cycled to 35 °C, held for a specified time, cooled to 10 °C, ramped to 35 °C, and incubated again are presented in Figure 2, left. The times shown in Figure 2 are cumulative. Compared with the data obtained for d/hDMPC-(100 nm)NP-SLB in 0.1 M PBS buffer, where samples held at either 30 or 40 °C (Table 1

RESULTS d -SUV/hDMPC-SUV Compared with dDMPC-(100 nm)NP-SLB/h DMPC -(100 nm)NP-SLB Exchange. For SASUV/SASiO2 = 1/1, the samples precipitated in 0.1 M PBS (100 mM NaCl), while those in water or 10 mM NaCl remained in suspension. At SASUV/SASiO2 = 2/1, all of the samples (in any medium) remained suspended. Mixtures of dDMPC and hDMPC vesicles exhibit complete miscibility with a linear dependence on composition (Supporting Information Figure 2). Time-dependent nano-DSC traces, where a separate sample was mixed for each run, for dDMPC -SUV/hDMPC-SUV and dDMPC-(100 nm)NP-SLB/hDMPC(100 nm)NP-SLB in 10 mM NaCl at 40 °C (Figure 1) and in water, 10 mM NaCl, and 0.1 M PBS buffer at 30, 40, and 50 °C (Supporting Information DMPC

723

DOI: 10.1021/la503967m Langmuir 2015, 31, 721−731

Article

Langmuir Table 1. Values of koff, t1/2, and ΔEact for h/dDMPC-SUV and h/dDMPC-(100 nm)NP-SLB as a Function of Ionic Strength and Temperature Obtained Isothermally koff (s−1)

media

h/dDMPC(100 nm) NP-SLB

4.1 × 10−5 1.2 × 10−4 6.0 × 10−4 108 kJ/mol

3.9 1.1 0.4

6.5 2.8 0.6

3.5 × 10−5 1.4 × 10−4 4.8 × 10−4 130 kJ/mol

no changea no changeb 2.1 × 10−4

5.5 1.7 0.5

a b 0.6

2.7 × 10−5

no changec

5.5

c

2.49 × 10−4 f no changed no changee

2.0 0.5

d e

h/dDMPCSUV

30 40 50

1.6 × 10−4 3.3 × 10−4 6.6 × 10−4 40 kJ/mol

30 40 50

30

water

ΔEact 10 mM NaCl

ΔEact 0.1 M PBS buffer (100 mM NaCl)

35f 40 50 ΔEact

t1/2 (h) h/ dDMPCSUV

T (°C)

−5

3.2 × 10 6 × 10−4 57 kJ/mol

The effects of cycling in the presence of excess lipid can be observed by nano-DSC data for hDMPC-(100 nm)NP-SLB with SASUV/SASiO2 = 2/1 for 10 mM NaCl and 0.1 M PBS (100 mM NaCl) buffer (Table 2). In these experiments, the samples were held at 50 °C for 1 h (the initial incubation). They were then cooled to 10 °C (held for 10 min) and reheated to 50 °C (held for 10 min), and the last two cycles were repeated. The nanoDSC traces (Figure 3) show several interesting features. First, the ratio of the intensities of the SUV peak (at 24 °C) to the intensity of the SLB peak (at ∼22 °C) is always greater for the heating cycles than the cooling cycles. This is partly the result of thermal lag in the calorimeter. However, the differences persist, particularly for the 10 mM NaCl suspensions, but to a lesser extent at slower (0.12°/min) scan speeds (Supporting Information Figure 13). The differences in relative peak intensities for the SUVs and NP-SLBs on the heating and cooling cycles, particularly for the 10 mM NaCl, suggest that when held at 10 °C some lipid is removed, forming SUVs, which are observed in the nano-DSC heating profile. When held at 50 °C for 10 min, these SUVs reattach to the SiO2, so that during cooling, there are less free SUVs. In addition, the SUV/NP-SLB intensity ratios are larger for the 10 mM NaCl than for the 0.1 M PBS on both the heating and cooling cycles. This indicates that for the same amount of lipid (here, SASUV/SASiO2) = 2/1) there is more fusion at higher ionic strengths and, as expected, that the lipids are more tightly bound to the SiO2 for the 0.1 M PBS (100 mM NaCl) buffer compared with the 10 mM NaCl. DLS data for the same system (hDMPC-(100 nm)NP-SLB with SASUV/SASiO2 = 2/1 in 10 mM NaCl and 0.1 M PBS) as well as for the SUVs and SiO2 are given in Figure 3, with all the data summarized in Supporting Information Table 1. For the SUVs (buffer and salt), there is always a slight decrease in size when the samples are cooled from 50 °C to 10 °C (Supporting Information Figure 14), which reflects the shrinkage of the lipids below Tm.35 In the distributions by number (Supporting Information Figure 15), the peaks for the nanosystems with SASUV/SASiO2 = 2/1 (both PBS and 10 mM NaCl) contained tails at larger sizes, but the peak maxima in the distribution decreased going from 50 to 10 °C, also indicating shrinkage of the lipids at Tm.35 In the volume distributions (Figure 3) for SASUV/SASiO2 = 2/ 1 in 10 mM NaCl and 0.1 M PBS buffer, the first incubation cycle (50 °C) showed structures that were larger and/or with a broader size range than subsequent cycles. Control experiments (not shown) in 10 mM NaCl at SASUV/SASiO2 = 1/1 and 1.3/1 showed that the first two cycles were reversible, with increases in size at low temperature. The sizes observed suggest that “twins”, i.e., structures containing on average a NP-SLB and a SUV protrusion or a bilayer that on average entraps two SiO2 NPs, can form during the initial incubation cycle; some of the “twins” separate as the NPs/SUVs are cooled below Tm (as the lipid areas shrink and the SUVs can be pinched off). On subsequent cycles, larger, reversible aggregates form at low temperatures that decrease in size above Tm. In general, sizes were larger at comparable temperatures/cycles for nanosystems in 0.1 M PBS than in 10 mM NaCl. As previously reported for SUVs,35 the NP-SLBs and SUVs adhere to each other at low temperatures and reversibly come apart as the temperature is raised above Tm,

h/dDMPC-(100 nm)NP-SLB

a

Do not exchange for at least 60 h. bDo not exchange for at least 7 h. Do not exchange for at least 5 h. dDo not exchange for at least 6 h. e Do not exchange for at least 6 h. fCycled above and below Tm. c

Table 2. koff Rates for h/dDMPC-(5, 100 nm)SLB, h/dDPPC-(5, 100 nm)NP-SLB, as a Function of SASUV/SASiO2 in PBS (100 mM NaCl) Buffer, Obtained by Temperature Cycling lipid

lipid

SASUV/SASiO2

incubation temp (°C)

SiO2 diam (nm)

koffd (s−1)

prepared with nominal amount of lipid for SASUV/SASiO2 = 1/1 coverage dDMPC dDMPC

hDMPC hDMPC

1/1 1/1

35 35

5 100

1.09 × 10−4 2.49 × 10−4

dDPPC dDPPC

hDPPC hDPPC

1/1 1/1

55 55

5 100

0.90 × 10−4 3.47 × 10−4

prepared with excess lipid (SASUV/SASiO2 = 2/1) and rinsed dDPPC dDPPC dDPPC dDPPC dDPPC

hDPPC hDPPC hDPPC hDPPC hDPPC

2/1 2/1 2/1 2/1 2/1

65 70 50 60 70

100 100 45 45 45

5.35 1.25 1.15 1.18 5.56

× × × × ×

10−5 10−4 10−6 10−5 10−5

and Supporting Information Figures 7 and 8) exhibited no exchange for the times shown in Figure 2, left, the cycled samples show relatively rapid lipid exchange, with koff times of 2.49 × 10−4 s−1. These koff rates are even faster than those reported for DMPC-SUVs (koff = 4.5 × 10−5 s−1 at 35 °C).20 Data (Table 2) were also obtained by temperature cycling for the h/dDMPC-(5 nm)NP-SLB on the 5 nm SiO2 (Figure 2, right) and for the h/dDPPC-(5 nm)NP-SLB and h/dDPPC-(100 nm)NP-SLB on both 5 and 100 nm SiO2 NPs using DPPC (Supporting Information Figure 12). The off rates for all of the these temperature-cycled samples are faster than those previously reported for isothermal exchange with DPPCSUVs (koff = 0.23 × 10−5 s−1 at 55 °C).20 724

DOI: 10.1021/la503967m Langmuir 2015, 31, 721−731

Article

Langmuir

Figure 2. Nano-DSC traces of h/dDMPC-(5, 100 nm)-SLB (SASUV/SASiO2 = 1/1) on (left) 100 nm and (right) 5 nm NPs in 0.1 M PBS buffer (100 mM NaCl) cycled between 10 and 35 °C. Times noted are cumulative. Controls are NP-SLBs prepared with SUVs composed of 1/1 hDMPC/dDMPC.

Figure 3. Successive cooling and heating cycles of for hDMPC-(100 nm)NP-SLB with SASUV/SASiO2 = 2/1 in 10 mM NaCl and 0.1 M PBS (100 mM NaCl); samples initially held at 50 °C/1 h, then repetitively at 10 °C/10 min and 50 °C/10 min: (right) nano-DSC data; (left) DLS data by volume. For diameters see Table 3.

the hDMPC-SUV were suspended, but upon mixing the two (so that SASUV/SASiO2 = 2/1), the precipitated NP-SLBs immedi-

and this effect can be more pronounced the greater the ionic strength, which screens repulsive interactions. Presence of Defects. Nano-DSC data for dDMPC-(5 nm)NP-SLB or dDMPC-(100 nm)NP-SLB (i.e., prepared with 5 and 100 nm SiO2) and incubated with hDMPC-SUV are shown in Figure 4. The dDMPC-(5 nm)NP-SLB or dDMPC-(100 nm)NPSLB (SASUV/SASiO2 = 1/1) themselves were precipitated and

ately resuspended. At the initial stages of transfer for both the 5 and 100 nm dDMPC-(5, 100 nm)NP-SLB, Tm (TII) for the hDMPC-SUV decreased in intensity but did not shift in temperature. This strongly suggests that the hDMPC-SUV 725

DOI: 10.1021/la503967m Langmuir 2015, 31, 721−731

Article

Langmuir

Effect of Defects on Exchange Rate. Exchange rates for h/dDPPC-(45 nm)NP-SLB or h/dDPPC-(100 nm)NP-SLB prepared with (2×) excess (DPPC) lipid, with the residual h/ dDMPC-SUV removed by rinsing (Supporting Information Figure 16), are summarized in Table 2. We previously have shown by thermogravimetric analysis that the use of excess lipids increases lipid coverage of the SiO2-NPs,33 thus reducing defects. The exchange data indicate that removal/reduction in the defects decreases the exchange rates by about an order of magnitude compared with NP-SLBs prepared without excess lipid. For the h/dDPPC-(100 nm) NP-SLB, the exchange rate at 55 °C for samples with defects (prepared with SASUV/SASiO2 = 1/1) is faster than the exchange rates at the higher temperatures of 65 and 70 °C for samples with no/fewer defects (prepared with SASUV/SASiO2 = 2/1) (Table 2). SiO2 Size Effects. In the cycling experiments, values of koff (Table 2) for hDMPC-(5 100 nm)NP-SLB/dDMPC-(5 100 nm)NP-SLB indicate that the exchange rate is slower for lipids on the 5 nm SiO2-NPs than on the 100 nm SiO2-NPs. Unlike the 45 and 100 nm SiO2, where the size of the NP-SLBs are increased by ∼8 nm, approximately the size expected for the addition of a lipid bilayer, the 5 nm SiO2 form large aggregates (∼150 nm, 4000 nm) during SLB formation.36 The exchange rate (Table 2) for h DPPC-(100 nm)NP-SLB/d DPPC-(100

Table 3. Diameters and Polydispersity Indices (PDI) for hDMPC-(100 nm)NP-SLBs with SASUV/SASiO2 = 2/1 that have been temperature cycled, as shown in Figure 3 1st cycle, 50 °C

1st cycle, 10 °C

≥2nd cycle, 50 °C

≥2nd cycle, 10 °C

10 mM NaCl Diameter (nm) PDI

180 0.07

200, 135 0.3

135 0.02

100, 200sh 0.13

0.1 M PBS Diameter (nm) PDI

165 0.05

524 0.22

155 0.03

125, 200sh 0.12

“feed” the dDMPC-NP-SLB, i.e., that there is initially simply lipid transfer from the hDMPC-SUV to the dDMPC-NP-SLB and thus that defects originally existed on the NP-SLBs. In both cases (5 and 100 nm SiO2), the dDMPC-(5, 100 nm)NP-SLB transitions (TI) split (e.g., see t = 0.33 h for the 5 and 100 nm SiO2), indicating differences in the proximal and distal leaflets of the NP-SLBs. The hDMPC-SUV transition (TII) then shifts to lower temperature for the 100 nm SiO2 (at time, t = 0.83 h) signaling that lipid exchange has begun. However, no such shift at this or any later time (for times measured ∼7−8 h) occurs for the 5 nm SiO2.

Figure 4. Lipid transfer between dDMPC-(5 nm)NP-SLB or dDMPC-(100 nm)NP-SLB and hDMPC-SUV in 0.1 M PBS buffer (100 mM NaCl), SASUV/ SASiO2 = 2/1 (i.e., hDMCP-SUV + dDMPC-(5, 100 nm)NP-SLB, SASUV/SASiO2 = 1/1) on (top) 100 nm SiO2 and (bottom) 5 nm SiO2. Dotted line is guide to the eye for the position of Tm for the hDMPC-SUV and dDMPC-(5, 100 nm)NP-SLB. Schematic for the processes of vesicle fusion, lipid insertion, and exchange, at different times, as described in the text. Red indicates lipids on NP-SLBs and blue indicates lipids in SUVs. 726

DOI: 10.1021/la503967m Langmuir 2015, 31, 721−731

Article

Langmuir nm)NP-SLB on 100 nm SiO2 (koff = 1.25 × 10−4 s−1) is faster by a factor of 2 compared with the exchange rate on hDPPC-(45 nm)NP-SLB/dDPPC-(45 nm)NP-SLB for the 45 nm SiO2 (koff = 5.56 × 10−5 s−1). Further, the approximate activation energies obtained from Arrhenius plot of the rate constants, ln(KDPPC) as a function of inverse temperature (1/T) for h/dDPPC-(45 nm)NP-SLB are EA = 179 kJ/mol for the 45 nm SiO2 and for h/dDPPC(100 nm)-NP-SLB are EA = 164 kJ/mol (based on two points) for the 100 nm SiO2 (Supporting Information Figure 17).

remove lipids from the NP-SLB surface, as previously observed.5 This can be attributed to the increased interactions between the lipids and SiO2 that increase with added salt, since it is well established that salt, which screens repulsive electrostatic interactions between them, increases fusion of vesicles onto SiO2. At pH > 3.5 in water, both SiO2 (ζ ∼ −45 mV)40 and DMPC (ζ = −12 ± 1.6 mV) have negative surface charges, which are reduced (become more positive) at 100 mM NaCl (ζ = 0 mV for DMPC).26 Presence of Defects. Dynamic or static structural defects in membrane or vesicle organization, such as can arise at boundaries between phases, by geometric packing constraints in small vesicles, and mismatches in alkyl chain length for twocomponent gel phase lipids, creating pore defects,19 have been suggested to be sites for removal or insertion of lipids.28 Transfer experiments of hDMPC-SUV to planar d-DMPC substrates show that the amount of lipid transferred is greatest near Tm.13 In the case of supported lipid bilayers, there can also be defects in regions where there are exposed surfaces, i.e., where the lipid has not formed a continuous lipid bilayer, as has been observed by AFM for planar SLBs.41 Bare patches of exposed SiO2 can result from the initial bilayer formation since the surface areas of the lipids from the SUVs and the surface area of the SiO2 cannot be matched precisely, particularly if the surface of the SiO2 is rough and since there is always a distribution of vesicles and NP-SiO2 sizes. Therefore, whenever RSUV < RSiO2, there will be a bare SiO2 patch (Figure 5).

■

DISCUSSION When NP-SLBs are prepared for use in biotechnical applications or for remediation, the lipids on them can spontaneously exchange/transfer with other lipids, for example, those in cell membranes or in the environment. However, these NP-SLBs may not have perfect, continuous bilayers, which can affect their interactions with other lipid/vesicles/NPs. The current experiments highlight several variables that can modulate the process of lipid exchange/transfer between NPSLBs or between NP-SLBs and SUVs, with SUV/SUV exchange also investigated for comparison. In the current work, (i) ionic strength decreased the symmetric exchange rates for SUV/SUVs, but more so for NP-SLB/NP-SLBs, where addition of salt (NaCl) increased the attraction of the zwitterionic lipids for the negatively charged SiO2 support; (ii) temperature cycling above and below Tm increased the exchange/transfer process, possibly by the creation of defects on the NP surface; (iii) defects in the bilayers formed on the NP-SiO2 increased the rate of transfer/exchange by approximately an order of magnitude; and (iv) exchange was faster on 100 nm NP-SLBs than on the 45 or 5 nm NP-SLBs, in the latter case possibly due to the aggregate morphology of the NPSLBs. These results can be used to develop models for exchange between NP-SLBs with each other and between NPSLBs with SUVs. SUV/SUV Exchange. The decrease in exchange kinetics for SUVs with increased ionic strength has been previously observed for both zwitterionic and charged lipids,37,38,18 and was originally attributed to the “salting out” of lipid from the aqueous phase.18,38 More recent molecular dynamics simulations/fluorescence correlation spectroscopy39 and nanoindentation26 experiments indicate that there is tight binding of Na+ ions to the carbonyl oxygens of PC lipids, leading to larger complexes with reduced mobility, an increase in bilayer thickness that increases the order parameter of the alkyl chains and makes the membrane more compact. The activation energies (ΔEa) measured here for h/dDMPC-SUV exchange in water are lower than previously reported values (70 kJ/mol between 27 and 45 °C),34 but the t1/2 are in good agreement (2.0 h at 37 °C).34 This may be due to differences in vesicle size. Values ΔE a ,13,28 k off, and t 1/2 for h/d DMPC -SUV exchange18,20 in buffer are of same order of magnitude as those reported here. In general, however, at the same temperature, the time resolution of exchange measured using FTIR13 (10−30 s) will give much better initial rate data (and thus activation energies) than that obtained from nano-DSC experiments, where each scan takes ∼15 min. NP-SLB/SUV and NP-SLB/NP-SLB Exchange/Transfer. The koff values are larger, and the t1/2 values and activation energy (ΔEact) are smaller, for h/dDMPC-SUV exchange compared with h,d DMPC (100nm)NP-SLB/h,d DMPC (100 nm)NP-SLB exchange, indicating that it is more difficult to

Figure 5. Schematic of defect formation and vesicle adsorption. If RSUV > RSiO2, then the excess lipid will form small SUVs. If RSUV ≈ RSiO2, then small defect sites can form where there can be only SUV adsorption, not fusion. If RSUV < RSiO2, then further fusion can occur. SUVs can also adhere to NP-SLBs. Amount of lipid pictured is not quantitative.

Here we have shown that these types of defects can arise and/or be repaired in the presence of excess lipid when NPSLBs are first formed or as the result of temperature cycling. The presence of defects was most clearly observed in temperature cycling experiments where: (i) dDMPC-NP-SLB prepared with “nominal” SASUV/SASiO2 = 1/1 were incubated with hDMPC-SUV (so that the nanosystem contained SASUV/ SASiO2 = 2/1) and where the hDMPC-SUV intensity decreased during cycling as the result of transfer without exchange of lipid from hDMPC-SUV to dDMPC-NP-SLB (Figure 4). (ii) Combined nano-DSC and DLS experiments (Figures 3) for hDMPC-NPSLB (SASUV/SASiO2 = 2/1) indicate that the lipid bilayer can desorb (in the gel phase) and resorb (in the liquid-crystalline 727

DOI: 10.1021/la503967m Langmuir 2015, 31, 721−731

Article

Langmuir

previously been observed on planar (zwitterionic) SLBs by quartz crystal microbalance with dissipation (QCM-D).43 In experiments with dDMPC-(100 nm)NP-SLB/hDMPC-SUV (Figure 4), at the initial stages of exchange/transfer (1/2 h, the TII transition shifts to lower temperatures for the 100 nm SiO2 and a high temperature shoulder appears on the TI transition, although the shifted TI peak persists for ∼7 h. An explanation for the double transition at TI for dDMPC-NP-SLB is based on previous reports of exchange, where it was suggested that the outer monolayer of bilayer vesicles could accommodate additional lipid (∼30%),20 and simulations showing that the proximal leaflet (next to SiO2) is more densely packed than the distal leaflet of SLBs,55 so that the distal leaflet can accommodate additional lipids. Thus, the lower temperature peak (in the TI doublet) would be attributed to the inner leaflet with mostly d-DMPC on the dDMPC-(100 nm)NP-SLB, mixed with some h-DMPC from the SUVs during the initial fusion, while the higher temperature peak (in the TI doublet) would be attributed to the outer leaflet that has incorporated additional h-DMPC. After 1/2 h, the TII transition shifts to lower temperature and slightly decreases in intensity, suggesting that there is both insertion (transfer with no exchange) and exchange (between the hDMPC- SUV and the outer leaflet of the dDMPC-NP-SLB). Since there is no further change in intensity for TII after ∼1 h, the outer leaflet of the dDMPC- (100nm) NP-SLBs may be “full”, i.e., as densely packed as possible, after this time. Exchange can then proceed by single molecule diffusion or mutual insertion/hemifusion of lipids from the contacting dDMPC-(100nm)NP-SLB and hDMPC-SUV, i.e., from the SUVs to the outer leaflet of the NP-SLBs and from outer leaflet of the NP-SLBs to the SUVs. This is evidenced by the change of temperature for both TI (the higher temperature transition in the doublet) and TII. There also appears to be little change in the nano-DSC traces (Figure 4) between 4 and 7 h for the dDMPC-(100 nm)NP-SLB. In particular, after 7 h, the lower temperature transition (TI) of dDMPC-(100 nm)NP-SLB is still observed, indicating that flipflop has not occurred. This is consistent with the view that the initial fusion and then insertion processes have added as many lipids as possible to the NP-SLB, so that it is difficult for “flipflop” as well as removal of lipid (necessary for exchange) to occur in the well packed bilayer. Flip-flop rates have previously been found to be accelerated in incomplete planar SLBs formed by fusion43 and in less perfectly formed bilayers prepared by

phase) to the NP-SiO2, creating and healing defects (bare SiO2 patches) as the temperature is cycled above and below Tm. The results strongly suggest that there are defects in the NP-SLBs, either exposed SiO2 that act as sites for SUV adsorption (Figure 5) or defects that occur at gel/liquid crystal phase boundaries, which can accommodate additional lipids. Previous NMR data for dDMPC-NP-SLB in water on 0.5 and 1.5 μm glass beads showed that the d-DMPC desorbed completely from the beads during passage from the liquid crystalline to the gel state.30 This was attributed to the 5% change of molecular volume of d-DMPC42 that caused a mismatch in area between bilayer and support. In the current case, where the NP-SLBs are in buffer, not water, and the interaction with the substrate is greater, not all of the lipid desorbs in the short time the NP-SLBs are in the gel state. Adsorption of intact vesicles on top of a planar SLB (composed of zwitterionic lipids) in 100 mM NaCl (but not pure water) has previously been reported.43 The desorption of lipids from the NP-SLB could also result in bilayer protrusions (rather than complete lipid removal), as has also been observed.30,44 AFM studies on planar SLB substrates (typically mica or SiO2) have shown the development of domain structure at Tm as well as cracks in the bilayer surface,26 particularly at lipid phase transitions on mica surfaces where the two leaflets are decoupled.45−48 Mechanisms of Exchange/Transfer. In all of the experiments, those where the nanosystems were suspended (SUV/SUV, NP-SLB/SUV, NP-SLB/NP-SLB, SASUV/SASiO2 = 2/1 (10 mM NaCl)), but also precipitated (NP-SLB/NP-SLB, SASUV/SASiO2 = 1/1 (0.1 M PBS 100 mM NaCl)), single molecule diffusion can always be, and probably is, a mechanism for exchange/transfer of lipids. For the precipitated nanosystem, transient fusion during a collision or collisional transfer is not possible. Since concentration-dependent studies were not obtained for the suspended nanosystems, it cannot be ruled out, as it was for SUVs on planar SLBs.13 However, the results both for the NP-SLB/SUV and NPSLB/NP-SLB nanosystems suggest that significant exchange/ transfer occurs when the SUVs are adsorbed to the NP-SLBs, or when NP-SLBs are adjacent to each other (when precipitated), as proposed in the sequential attachment− transfer−detachment (ATD) model for charged (but not low charge density/uncharged) lipids,13,15,22,23 where the mechanism of lipid transfer was suggested to be monomer insertion or hemifusion, with a characteristic time of 3−40 min.21 Evidence that SUVs are adsorbed to the NP-SLBs is indicated by (i) DLS data for nanosystems prepared using SASUV/SASiO2 = 2/1 (Figure 3), where sizes larger than those of SUVs or SiO2 are observed, and where NP-SLBs and SUVs reversibly adsorb to each other as the temperature is lowered, and (ii) resuspension of precipitated NP-SLBs (SASUV/SASiO2 = 1/1) with added SUVs (Figure 4) to form SASUV/SASiO2 = 2/1, where the entropic repulsion lost on the rigid supports36,44 is restored.49 Exposed SiO2 sites have a higher (negative) zeta potential than in regions where the SiO2 is shielded by the lipid. These sites therefore attract other SUVs (in the SUV/NP-SLB experiments) or other NP-SLBs (in the NP-SLB/NP-SLB experiments). The NP-SLBs (SASUV/SASiO2 = 1/1) still have a slightly negative ζ26 and so can also adsorb the neutral SUVs (Figure 5). Adsorption of zwitterionic phosphatidylcholine vesicles has 728

DOI: 10.1021/la503967m Langmuir 2015, 31, 721−731

Article

Langmuir Langmuir−Blodgett methods56 and to occur most frequently at domain interfaces in planar SLBs57 and at phase transitions.58 Here, the time scale for flip-flop is >7 h for lipids on the 100 nm SiO2. For free phospholipid bilayers, “flip-flop” times of τ1/2 ∼ hours to days are typical,59,60 although τ1/2 values are 45 nm > 100 nm for NP-SLBs. A higher Tm suggests better packing of the alkyl chains. Therefore, it might be expected that the off rate for the lipids would be inversely related to SiO2 size, in this range of sizes. These results were obtained for NP-SLBs prepared with SASUV/SASiO2 = 1/1 (i.e., both contain defects) and is thus not in conflict with the results discussed in the previous section, obtained using SASUV/SASiO2 = 2/1.

■

line (PC) lipids between NP-SLBs and SUVs, between NPSLBs and NP-SLBs, and for comparison between SUVs and SUVs. Exchange/transfer was measured using nanodifferential scanning calorimetry, taking advantage of the difference in transition temperatures for d-/h-lipids. The aggregation state of the nanosystems was determined by dynamic light scattering. The presence of defects on NP-SLBs was found to increase both exchange/transfer and flip-flop kinetics. Defects in the NP-SLB were the result of mismatches in gel/liquid crystalline domains (as also occur in vesicles) and exposed patches of SiO2 due to differences in surface areas of the SiO2 and lipids, as expressed by SASUV/SASiO2. These could occur during the original preparation procedure and from temperature cycling of the nanosystems above and below Tm of the lipids, which could result in cracks in the bilayer on the NP surface due to mismatch in the thermal expansion coefficients of the support and lipid bilayer. DLS and nano-DSC data for nanosystems with SASUV/SASiO2 = 2/1 in either 10 mM NaCl or PBS buffer (100 mM NaCl) consisting of nominally equal populations of SUVs and NP-SLBs were consistent with a mechanism where some lipid desorbed from the SiO2-NPs in the form of vesicles, which may have remained attached/adsorbed to the SiO2 during passage from the gel to fluid phase but re-formed the NP-SLB bilayer in the fluid phase. If there were defects in the original NP-SLB, formed at a nominal SASUV/SASiO2 = 1/1 ratio, lipid from the adsorbed vesicles fused to these sites to form a more complete NP-SLB. For NP-SLBs prepared with excess lipid to heal these defects, the exchange rate was considerably slower in NP-SLB/NP-SLB exchange experiments (by approximately an order of magnitude) than for NP-SLBs with defect sites. Increased ionic strength, which formed more stable NP-SLBs and increased shielding between the SUVs and NP-SLBs or between NP-SLBs and NP-SLBs, resulted in slower exchange/transfer rates. The data on size effects suggested that exchange/transfer rates were faster on larger (100 nm SiO2 > 45 nm SiO2 > 5 nm SiO2) NP-SLBs.

■

ASSOCIATED CONTENT

S Supporting Information *

Table 1 and Figures 1−17. This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail [email protected] (S.L.W.). Notes

The authors declare no competing financial interest.

■

ABBREVIATIONS NP, nanoparticle; SLB, supported lipid bilayer; SiO2, silica; MLV, multilamellar vesicle; SUV, small unilamellar vesicle; DMPC, 1,2-dimyristoyl-sn-glycero-3-phosphocholine; DPPC, 1,2-dipalmitoyl-sn-glycero-3-phosphocholine; d, deuterated; h, hydrogenated; dDMPC, deuterated DMPC; hDMPC, hydrogenated DMPC; d-SUV, SUV composed of deuterated lipids; d/h-SUV, SUV composed of deuterated or hydrogenated lipids; d/h-SLB, SLB composed of deuterated or hydrogenated lipids; (100 nm) NP, NP composed of 100 nm SiO2; PBS, phosphate buffer saline; nano-DSC, nanodifferential scanning calorimetry; DLS, dynamic light scattering.

CONCLUSIONS

Exchange or transfer of lipids between bilayer surfaces, where at least one surface is that of a nanoparticle, is important in applications where NPs enter cells for drug delivery applications, since the NPs can be preformed as NP-SLBs or become enveloped by lipids from the cell membrane upon endocytosis. Lipids can coat man-made or naturally occurring NPs, where they will affect their environmental migration and distribution. Here we have investigated factors that affect exchange/transfer of zwitterionic, saturated phosphatidylcho729

DOI: 10.1021/la503967m Langmuir 2015, 31, 721−731

Article

Langmuir

■

two lipid bilayer membranes. J. Phys. Chem. B 2008, 112 (44), 14069− 14074. (22) Sapuri, A. R.; Baksh, M. M.; Groves, J. T. Electrostatically targeted intermembrane lipid exchange with micropatterned supported membranes. Langmuir 2003, 19 (5), 1606−1610. (23) Solon, J.; Streicher, P.; Richter, R.; Brochard-Wyart, F.; Bassereau, P. Vesicles surfing on a lipid bilayer: Self-induced haptotactic motion. Proc. Natl. Acad. Sci. U. S. A. 2006, 103 (33), 12382−12387. (24) Zhdanov, V. P.; Kasemo, B. Lipid exchange during contact between oppositely charged lipid bilayers. J. Phys. Chem. B 2007, 111 (32), 9428−9430. (25) Jahn, R.; Grubmuller, H. Membrane fusion. Curr. Opin. Cell Biol. 2002, 14 (4), 488−495. (26) Garcia-Manyes, S.; Oncins, G.; Sanz, F. Effect of pH and ionic strength on phospholipid nanomechanics and on deposition process onto hydrophilic surfaces measured by AFM. Electrochim. Acta 2006, 51 (24), 5029−5036. (27) Dimitrievski, K.; Kasemo, B. Influence of lipid vesicle composition and surface charge density on vesicle adsorption events: A kinetic phase diagram. Langmuir 2009, 25 (16), 8865−8869. (28) Wimley, W. C.; Thompson, T. E. Exchange and flip-flop of dimyristoylphosphatidylcholine in liquid-crystalline, gel, and twocomponent, two-phase large unilamellar vesicles. Biochemistry 1990, 29 (5), 1296−303. (29) Bhatia, V. K.; Hatzakis, N. S.; Stamou, D. A unifying mechanism accounts for sensing of membrane curvature by BAR domains, amphipathic helices and membrane-anchored proteins. Semin. Cell Dev. Biol. 2010, 21 (4), 381−390. (30) Bayerl, T. M.; Bloom, M. Physical properties of single phospholipid bilayers adsorbed to micro glass beads. A new vesicular model system studied by deuterium nuclear magnetic resonance. Biophys. J. 1990, 58 (2), 357−62. (31) Reinl, H. M.; Hartinger, A.; Dettmar, P.; Bayerl, T. M. Timeresolved infrared ATR measurements of liposome transport kinetics in human keratinocyte cultures and skin reveals a dependence on liposome size and phase state. J. Invest. Dermatol. 1995, 105 (2), 291− 295. (32) Thilo, L. Kinetics of phospholipid exchange between bilayer membranes. Biochim. Biophys. Acta 1977, 469 (3), 326−334. (33) Ahmed, S.; Wunder, S. L. Effect of high surface curvature on the main phase transition of supported phospholipid bilayers on SiO2 nanoparticles. Langmuir 2009, 25 (6), 3682−3691. (34) McLean, L. R.; Phillips, M. C. Kinetics of phosphatidylcholine and lysophosphatidylcholine exchange between unilamellar vesicles. Biochemistry 1984, 23 (20), 4624−4630. (35) Wang, H. R.; Drazenovic, J.; Luo, Z. Y.; Zhang, J. Y.; Zhou, H. W.; Wunder, S. L. Mechanism of supported bilayer formation of zwitterionic lipids on SiO2 nanoparticles and structure of the stable colloids. RSC Adv. 2012, 2 (30), 11336−11348. (36) Savarala, S.; Ahmed, S.; Ilies, M. A.; Wunder, S. L. Formation and colloidal stability of DMPC supported lipid bilayers on SiO2 nanobeads. Langmuir 2010, 26 (14), 12081−12088. (37) Papahadjopoulos, D.; Vail, W. J.; Pangborn, W. A.; Poste, G. Studies on membrane fusion 2. Induction of fusion in pure phospholipid membranes by calcium ions and other divalent metals. Biochim. Biophys. Acta 1976, 448 (2), 265−283. (38) Decuyper, M.; Joniau, M.; Dangreau, H. Intervesicular phospholipid transfer- a free-flow electrophoresis study. Biochemistry 1983, 22 (2), 415−420. (39) Bockmann, R. A.; Hac, A.; Heimburg, T.; Grubmuller, H. Effect of sodium chloride on a lipid bilayer. Biophys. J. 2003, 85 (3), 1647− 1655. (40) Ong, S. W.; Zhao, X. L.; Eisenthal, K. B. Polarization of water molecules at a charged interface - 2nd harmonic studies of the silica water interface. Chem. Phys. Lett. 1992, 191 (3-4), 327−335. (41) Seantier, B.; Breffa, C.; Felix, O.; Decher, G. In situ investigations of the formation of mixed supported lipid bilayers

REFERENCES

(1) Brown, R. E. Spontaneous lipid transfer between organized lipid assemblies. Biochim. Biophys. Acta 1992, 1113 (3−4), 375−389. (2) Kornberg, R. D.; McConnell, H. M. Inside-outside transitions of phospholipids in vesicle membranes. Biochemistry 1971, 10 (7), 1111− 1120. (3) Nakano, M.; Fukuda, M.; Kudo, T.; Endo, H.; Handa, T. Determination of interbilayer and transbilayer lipid transfers by timeresolved small-angle neutron scattering. Phys. Rev. Lett. 2007, 98 (23). (4) Tieleman, D. P.; Marrink, S. J. Lipids out of equilibrium: Energetics of desorption and pore mediated flip-flop. J. Am. Chem. Soc. 2006, 128 (38), 12462−12467. (5) Naumann, C.; Brumm, T.; Bayerl, T. M. Phase transition behavior of single phosphatidylcholine bilayers on a solid spherical support studied by DSC, NMR and FT-IR. Biophys. J. 1992, 63 (5), 1314−19. (6) Sackmann, E. Supported membranes: Scientific and practical applications. Science 1996, 271 (5245), 43−48. (7) Castellana, E. T.; Cremer, P. S. Solid supported lipid bilayers: From biophysical studies to sensor design. Surf. Sci. Rep. 2006, 61 (10), 429−444. (8) Sapuri, A. R.; Baksh, M. M.; Groves, J. T. Membrane interaction studies using microprinted supported lipid membranes and membrane-coated microbeads. Biophys. J. 2002, 82 (1), 814. (9) Wang, H.; Kim, B.; Wunder, S. L. In Supported Lipid Bilayer Nanoparticles for Remediation of Organic Toxins, ACS 2014 Colloid & Surface Science Symposium, Philadelphia, PA, 2014. (10) Wang, H.; Kim, B.; Wunder, S. L. Nanoparticle supported lipid bilayers as an in situ remediation strategy for hydrophobic organic contaminants in soils. Environ. Sci. Technol. 2014, es-2014-04832n. (11) Hao, J.; Murphy, R.; Lim, E.; Schoonen, M. A. A.; Strongin, D. R. Effects of phospholipid on pyrite oxidation in the presence of autotrophic and heterotrophic bacteria. Geochim. Cosmochim. Acta 2009, 73 (14), 4111−4123. (12) Debnath, S.; Hausner, D. B.; Strongin, D. R.; Kubicki, J. Reductive dissolution of ferrihydrite by ascorbic acid and the inhibiting effect of phospholipid. J. Colloid Interface Sci. 2010, 341 (2), 215−223. (13) Reinl, H. M.; Bayerl, T. M. Lipid transfer between small unilamellar vesicles and single bilayers on a solid support: Selfassembly of supported bilayers with asymmetric lipid distribution. Biochemistry 1994, 33 (47), 14091−9. (14) Wacklin, H. P.; Thomas, R. K. Spontaneous formation of asymmetric lipid bilayers by adsorption of vesicles. Langmuir 2007, 23 (14), 7644−7651. (15) Kunze, A.; Svedhem, S.; Kasemo, B. Lipid transfer between charged supported lipid bilayers and oppositely charged vesicles. Langmuir 2009, 25 (9), 5146−5158. (16) Kunze, A.; Sjovall, P.; Kasemo, B.; Svedhem, S. In situ preparation and modification of supported lipid layers by lipid transfer from vesicles studied by QCM-D and TOF-SIMS. J. Am. Chem. Soc. 2009, 131 (7), 2450−2451. (17) McLean, L. R.; Phillips, M. C. Mechanism of cholesterol and phosphatidylcholine exchange or transfer between unilamellar vesicles. Biochemistry 1981, 20 (10), 2893−2900. (18) Jones, J. D.; Thompson, T. E. Mechanism of spontaneous concentration-dependent phospholipid transfer between bilayers. Biochemistry 1990, 29 (6), 1593−1600. (19) Wimley, W. C.; Thompson, T. E. Transbilayer and interbilayer phospholipid exchange in dimyristolyphosphatidylcholine dimyristolyphosphatidylethanolamine large unilamellar vesicles. Biochemistry 1991, 30 (6), 1702−1709. (20) Bayerl, T. M.; Schmidt, C. F.; Sackmann, E. Kinetics of symmetric and asymmetric phospholipid transfer between small sonicated vesicles studied by high-sensitivity differential scanning calorimetry, NMR, electron microscopy, and dynamic light scattering. Biochemistry 1988, 27 (16), 6078−85. (21) Wikstrom, A.; Svedhem, S.; Sivignon, M.; Kasemo, B. Real-time QCM-D monitoring of electrostatically driven lipid transfer between 730

DOI: 10.1021/la503967m Langmuir 2015, 31, 721−731

Article

Langmuir close to the phase transition temperature. Nano Lett. 2004, 4 (1), 5− 10. (42) Bayerl, T. M.; Kochy, T.; Bruckner, S. On the modulation of a high enthalpy pretransition in binary mixtures of DMPC and DMPC by polar headgroup interaction. Biophys. J. 1990, 57 (3), 675−680. (43) Wacklin, H. P. Composition and asymmetry in supported membranes formed by vesicle fusion. Langmuir 2011, 27 (12), 7698− 7707. (44) Dolainsky, C.; Moeps, A.; Bayerl, T. M. Transverse relaxation in supported and unsupported phospholipid model membranes and the influence of ultraslow motions: a phosphorus-31 NMR study. J. Chem. Phys. 1993, 98 (2), 1712−20. (45) Leonenko, Z. V.; Finot, E.; Ma, H.; Dahms, T. E. S.; Cramb, D. T. Investigation of temperature-induced phase transitions in DOPC and DPPC phospholipid bilayers using temperature-controlled scanning force microscopy. Biophys. J. 2004, 86 (6), 3783−3793. (46) Keller, D.; Larsen, N. B.; Muller, I. M.; Mouritsen, O. G. Decoupled phase transitions and grain-boundary melting in supported phospholipid bilayers. Phys. Rev. Lett. 2005, 94 (2). (47) Feng, Z. V.; Spurlin, T. A.; Gewirth, A. A. Direct visualization of asymmetric behavior in supported lipid bilayers at the gel-fluid phase transition. Biophys. J. 2005, 88 (3), 2154−2164. (48) Charrier, A.; Thibaudau, F. Main phase transitions in supported lipid single-bilayer. Biophys. J. 2005, 89 (2), 1094−1101. (49) Helfrich, W. Steric iteraction of fuid membranes in multbilayer systems. Z. Naturforsch., A: Phys. Sci. 1978, 33 (3), 305−315. (50) Bucak, S.; Wang, C.; Laibinis, P. E.; Hatton, T. A. Dynamics of supported lipid bilayer deposition from vesicle suspensions. J. Colloid Interface Sci. 2010, 348 (2), 608−614. (51) Johnson, J. M.; Ha, T.; Chu, S.; Boxer, S. G. Early steps of supported bilayer formation probed by single vesicle fluorescence assays. Biophys. J. 2002, 83 (6), 3371−3379. (52) Weirich, K. L.; Israelachvili, J. N.; Fygenson, D. K. Bilayer edges catalyze supported lipid bilayer formation. Biophys. J. 2010, 98 (1), 85−92. (53) Kasson, P. M.; Pande, V. S. Molecular dynamics simulation of lipid reorientation at bilayer edges. Biophys. J. 2004, 86 (6), 3744− 3749. (54) Anderson, T. H.; Min, Y. J.; Weirich, K. L.; Zeng, H. B.; Fygenson, D.; Israelachvili, J. N. Formation of supported bilayers on silica substrates. Langmuir 2009, 25 (12), 6997−7005. (55) Xing, C. Y.; Faller, R. Density imbalances and free energy of lipid transfer in supported lipid bilayers. J. Chem. Phys. 2009, 131 (17). (56) Crane, J. M.; Kiessling, V.; Tamm, L. K. Measuring lipid asymmetry in planar supported bilayers by fluorescence interference contrast microscopy. Langmuir 2005, 21 (4), 1377−1388. (57) Lin, W. C.; Blanchette, C. D.; Ratto, T. V.; Longo, M. L. Lipid asymmetry in DLPC/DSPC-supported lipid bilayers: A combined AFM and fluorescence microscopy study. Biophys. J. 2006, 90 (1), 228−237. (58) John, K.; Schreiber, S.; Kubelt, J.; Herrmann, A.; Muller, P. Transbilayer movement of phospholipids at the main phase transition of lipid membranes: Implications for rapid flip-flop in biological membranes. Biophys. J. 2002, 83 (6), 3315−3323. (59) Dekruijff, B.; Wirtz, K. W. A. Induction of a relatively fast transbilayer movement of phosphatidylcholine in vesicles-C13 NMR study. Biochim. Biophys. Acta 1977, 468 (2), 318−326. (60) Homan, R.; Pownall, H. J. Transbilayer diffusion of phospholipids. Dependence on headgroup structure and acyl chain length. Biochim. Biophys. Acta 1988, 938 (2), 155−166. (61) Liu, J.; Conboy, J. C. Phase transition of a single lipid bilayer measured by sum-frequency vibrational spectroscopy. J. Am. Chem. Soc. 2004, 126 (29), 8894−8895.

731

DOI: 10.1021/la503967m Langmuir 2015, 31, 721−731