Phospholipid Bilayer Softening Due to Hydrophobic Gold Nanoparticle

Subscriber access provided by University of Sunderland

Biological and Environmental Phenomena at the Interface

Phospholipid Bilayer Softening due to Hydrophobic Gold Nanoparticle Inclusions Saptarshi Chakraborty, Akram Abbasi, Geoffrey D. Bothun, Michihiro Nagao, and Christopher L. Kitchens Langmuir, Just Accepted Manuscript • DOI: 10.1021/acs.langmuir.8b02553 • Publication Date (Web): 16 Oct 2018 Downloaded from http://pubs.acs.org on October 20, 2018

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 39 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

1

Phospholipid

2

Nanoparticle Inclusions

3

Saptarshi Chakraborty 1, Akram Abbasi2, Geoffrey D. Bothun2, Michihiro Nagao 3, 4, Christopher

4

L. Kitchens 1

5

1

Department of Chemical and Biomolecular Engineering, Clemson University, SC 29634, USA

6

2

Department of Chemical Engineering, University of Rhode Island, Kingston, RI 02881, USA

7

3

NIST Center for Neutron Research, National Institute of Standards and Technology,

8

Gaithersburg, MD 20899, USA

9

4

Bilayer

Softening

due

to

Hydrophobic

Center for Exploration of Energy and Matter, Department of Physics, Indiana University,

10

Bloomington, IN 47408, USA

11

1

Corresponding author Email: [email protected]; Phone: 864-656-2131

12 13 14 15 16 17 18 19

ACS Paragon Plus Environment

Gold

Langmuir 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

20

ABSTRACT

21

Liposome-Nanoparticle Assemblies (LNAs) are vital in the context of novel targeted drug delivery

22

systems, in addition to investigating nanoparticle-lipid bilayer interactions. Quantifying membrane

23

structural properties and dynamics in presence of nanoparticle inclusions provides a simple model

24

to elucidate nanoparticle effects on membrane biophysical properties. We present experimental

25

evidences of bilayer softening due to small hydrophobic gold nanoparticle inclusions. LNA

26

structure has been investigated by a combination of cryo-TEM, DLS and SANS. Neutron spin

27

echo (NSE) spectroscopy demonstrated a remarkable ~15% bending modulus decrease for LNAs

28

relative to pure liposomes. Clear dependence of bending modulus on AuNP diameter and

29

concentration was observed from our observations. Our findings point towards local bilayer

30

fluidization by nanoparticle inclusions leading to an overall bilayer softening. These findings add

31

valuable information to liposomal drug-delivery vehicle design and membrane biophysics

32

research.

33

INTRODUCTION

34

Nanoparticle (NPs) applications in conjugation with biological entities has increased

35

manifold with the advent of nanoparticle drug delivery systems and liposome-nanoparticle

36

assemblies (LNAs) as stimuli responsive drug delivery vehicles.1-4 From an environmental

37

impact standpoint, increased exposure to NPs from consumer products has also led to increased

38

nanoparticle-cell membrane interactions5 resulting in NP uptake by endocytosis,6 attachment,7

39

cell membrane penetration and changes in cell membrane permeability.8-9 Specifically, gold

40

nanoparticles (AuNPs) have been extensively applied for gene therapy,10 targeted drug

41

delivery,11-13 contrast agents in bioimaging,14-15 biosensors,16 hyperthermia agent17-19 and other

42

applications.


Page 2 of 39

Page 3 of 39 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

43

Cell membranes are in most cases, the first contact point between any engineered NP and

44

a biological target. Cell membranes are complex systems containing a variety of lipids and

45

embedded proteins which present difficulty in examining underlying biophysical properties.

46

Single or multi-component lipid bilayers in the unilamellar liposomal form, provides a simple cell

47

membrane model that can be used to examine NP-lipid bilayer interactions. Generally, NPs interact

48

with liposomes in two ways: a) hydrophilic NPs adsorbed internally or externally to the bilayer

49

inner or outer leaflet respectively,3, 9, 20-24 or b) hydrophobic NP inclusions that reside inside the

50

bilayer acyl core.2, 25-28 In the current work, we focus on the latter, where AuNP inclusions reside

51

in the bilayer core. Prior studies involving LNAs have focused on formation, stability, structural

52

changes, and phase behavior associated with AuNP incorporation.25, 29, 30 Previously, we have

53

demonstrated effects on structure and phase behavior of lipid bilayers with embedded stearylamine

54

(SA) coated AuNPs.31 Lipid bilayer thickening was observed in addition to increase in transition

55

temperature on embedding SA-AuNPs in bilayer as shown by small angle neutron scattering

56

(SANS) and differential scanning calorimetry (DSC) respectively.

57

From the perspective of designing efficient stimuli-responsive LNAs, it is essential to

58

quantify the bio-mechanical effects brought about in the lipid bilayer due to AuNP inclusions.

59

Similarly, in the context of AuNP toxicity, it is important to focus on lipid bilayer biomechanical

60

parameters which dictate many biologically relevant processes. Membrane bending modulus is

61

one of the most vital membrane property related to liposomal drug release characteristics and

62

directly relates to permeability and bilayer phase behavior.32 Bending fluctuations in lipid bilayers

63

control biological phenomena such as budding33, stalk formation34 and gene delivery35. Bending

64

modulus maintenance is thus directly linked to a cell’s stability, proper functioning and its survival.

65

Prior studies to quantify lipid bilayer bending modulus were limited to giant unilamellar vesicles


Langmuir 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 39

66

(GUVs), when micropipette aspiration36 or flickering spectroscopy37 was utilized. Aforementioned

67

methods are not feasible for measurements in the length scale of small unilamellar vesicles (SUVs)

68

(< 100 nm diameter). Dynamic light scattering (DLS) can be used to study collective bilayer

69

undulations and translational dynamics, however bending fluctuations of SUV bilayers are much

70

smaller than length scales probed by light scattering and cannot be isolated solely through DLS.38-

71

40

72

length scale and is suited for vibrational, rotational and diffusion dynamics in lipid bilayers.41-42

73

Solid-state NMR has been used to access the bending fluctuation domain, but suffers from sample

74

geometry restrictions, sensitivity and lack of spatial correlations.39, 43 Invasive techniques like

75

atomic force microscopy (AFM) requires supported lipid bilayers which greatly differ in

76

mechanical properties from liposomal/cellular bilayer structure in solution.44

Nuclear magnetic resonance (NMR) probes much smaller length scales than bending dynamics

77

Bending fluctuations in SUV bilayers have pico to nano second time scale and 0.1 nm to

78

10 nm length scale which overlaps the time and length regimes probed by neutron spin echo (NSE)

79

spectroscopy.20, 45-49 NSE is a non-invasive technique that can measure an ensemble average of

80

bending dynamics providing greater statistical advantage over other methods. Neutrons as probes,

81

provide greatly enhanced contrast between protonated lipids and deuterated solvent (D2O) to

82

directly measure bending fluctuations. NSE spectroscopy has been used previously to measure

83

bending modulus for bilayers including multicomponent bilayers,49-51 multi-bilayers,47 stacked

84

bilayers in complex fluids,52 and mechanical properties of nanoscopic lipid domains.53 NSE has

85

been successfully utilized in investigating the effects of foreign molecules and inclusions like

86

cholesterol,54 local anesthetic,55 non-steroidal anti-inflammatory drugs (NSAIDs),56-57 and

87

inorganic NPs attached to the outer leaflet of bilayers.20 In this work, bending modulus of LNA

88

bilayers

composed

of

zwitterionic

dipalmitoylphosphatidylcholine


(DPPC)

and

Page 5 of 39 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

89

dipalmitoylphosphatidylglycerol (DPPG) with dodecanethiol-functionalized AuNP (AuNP-DDT)

90

inclusions, has been measured through a combination of DLS and NSE. Effect of AuNP size and

91

concentration has been examined at two different lipid to nanoparticle (L/N) ratios in the fluid

92

phase. LNA structures were investigated extensively with cryo-transmission electron microscopy

93

(cryo-TEM) and small angle neutron scattering (SANS).

94

RESULTS AND DISCUSSION

95

AuNP Synthesis and LNA formation. Hydrophobic dodecanethiol-functionalized

96

AuNPs (DDT-AuNPs) were synthesized with average core diameters of 3.0 ± 0.6 nm and 5.5 ±

97

1.5 nm (henceforth referred to as 3 nm AuNP and 5 nm AuNP), as determined by TEM (Figure

98

1). The AuNP diameters correspond to a size smaller and larger than the average bilayer thickness

99

for fluid phase DPPC/DPPG bilayers (≈ 3.8 nm).


Langmuir 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

100

Figure 1. TEM images of (A) 3 nm AuNP and (B) 5.5 nm AuNP. Distribution of particle size for (C) 3 nm AuNP and (D) 5 nm AuNP. Errors represent one standard deviation from mean diameter.

101

LNAs were synthesized for both particle sizes at a theoretical lipid to NP (L/N) ratio of

102

7500:1 (low AuNP conc.) and 5000:1 (high AuNP conc.). LNAs obtained after final extrusion

103

through 100 nm diameter pore size membrane were stable and remained dispersed for weeks

104

without any visible sedimentation or aggregation. Pure liposomes (no AuNPs) exhibited a

105

characteristic bluish hue after extrusion and LNA dispersions were dark brown (Supplementary

106

Information, Figure S1).

107

A key result often excluded in prior LNA research is the AuNP inclusion efficiency based

108

on AuNP size. This was accomplished by drying stable LNA aliquots and extracting the AuNP


Page 6 of 39

Page 7 of 39 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

109

inclusions in toluene. A two tailed t-test (95% confidence level) of particle size distributions from

110

TEM images (Supplementary Information, Figure S2) demonstrates no significant difference

111

between 3 nm diameter AuNP inclusions (3.0 ± 0.8 nm for low concentration and 3.0 ± 0.9 nm for

112

high concentration) and actual 3 nm diameter AuNP (3.0 ± 0.6 nm). Hence, this depicts successful

113

inclusion of the entire AuNP size range in LNA bilayers. The 5 nm AuNP (5.5 nm ± 1.6 nm)

114

however, showed successful incorporation of smaller AuNPs only (3.8 nm ± 1.2 nm for low

115

concentration and 3.7 nm ± 1.1 nm for high concentration) in the bilayer, which are statistically

116

different. LNAs formed with 5 nm AuNPs are henceforth referred to as 3.8 nm AuNP LNAs and

117

results have been summarized in Table 1. Theoretical studies previously conducted with CdSe

118

quantum dots in DOPC bilayers showed a threshold diameter of 6.5 nm below which CdSe

119

quantum dot bilayer inclusion is successful, and above which lipid stabilized micelle structures

120

were formed.58 Similarly, a threshold (AuNP + DDT) diameter was observed, above which AuNPs

121

are preferentially stabilized by lipid monolayers or form aggregates. The post-extrusion LNA L/N

122

ratio was determined from elemental gold and phosphorus concentrations measured by inductively

123

coupled plasma-optical emission spectroscopy (ICP-OES). TEM and ICP-OES analysis of AuNPs

124

inclusions and LNAs are summarized in Table 1. The AuNP concentration is lower than theoretical

125

for all LNAs, due to AuNP aggregation and loss during hydration and extrusion (Supplementary

126

Information, Figure S3). The 3 nm AuNP has a loading efficiency of 77% and 68% for the low

127

and high concentrations respectively. The 5 nm AuNP loading efficiency was much lower (43%

128

and 42% for high and low 5 nm AuNP concentrations respectively) due to additional large AuNP

129

exclusion from lipid bilayers. Assuming a homogeneous AuNP distribution in all LNAs, AuNP

130

inclusion concentration can also be estimated as AuNP per LNA as shown in Table 1 (see

131

Supplementary Information, Section S4 for details). However, discussions in the following section


Langmuir 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

132

will show that depicting AuNP inclusion concentrations in AuNP per LNA terms is a crude

133

average.

134

Table 1. Summary of AuNP inclusions in LNAs and actual AuNP loading in lipid a bilayers. measured by TEM analysis of AuNPs extracted from stable LNAs. b measured by ICP-OES analysis of stable LNAs. Errors in diameter represent one standard deviation from the mean.

135

Cryo-TEM Characterization of LNAs. Formation of stable SUVs with AuNP inclusions

136

in the bilayer core was visualized by cryo-TEM imaging of LNAs (Figure 2 and Supplementary

137

Information S5). Analysis of pure liposomes (no AuNP), showed presence of unilamellar

138

liposomes only (average diameter 85 ± 24 nm) with no multilamellar liposomes (> 1000 pure

139

liposomes analyzed).

140

observed previously30. The LNAs were either partially loaded with isolated AuNPs (Figure 2C and

141

2F), partially loaded with small AuNP clusters (Figure 2D and E), completely loaded with AuNPs

142

(Figure 2C and 2E) or apparently bare (henceforth mentioned as bare liposomes, observed in the

143

background, out of focus, in Figure 2B, C and E). Clusters containing a few AuNPs were observed

144

in bilayers for all LNAs (Figure 2E). On statistical analysis of all LNA cryo-TEM images (> 500

145

liposomes counted), LNAs with AuNP clusters and apparent bare liposomes were the most

The AuNP distribution in LNAs was not homogeneous, as similarly


Page 8 of 39

Page 9 of 39 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

146

prevalent liposomes at > 80%. Figure 2E shows an LNA completely packed with AuNPs but

147

demonstrates a homogenous NP distribution in the bilayer. Interestingly, cryo-TEM shows that

148

LNAs with completely packed bilayers had greater sphericity and smaller average diameter (for

149

example, 52 ± 31 nm for 3.8 nm LNA (high) compared to the bare LNAs (87 ± 14 nm) or LNAs

150

with small AuNP clusters (89 ± 21 nm) in the same sample. Similar observations have been made

151

previously with iron oxide NPs in DSPC liposomes27 and CdSe NPs in polymersomes.59 Figure

152

2D shows the presence of larger AuNPs and possible AuNP aggregates in the lipid bilayer. Position

153

of these aggregates/large AuNPs could not be judged with certainty through cryo-TEM. Some disc

154

shaped/rod/worm like AuNP clusters were also observed at high AuNPs loadings (Figure 2B).

155

Figure 2. Cryo-TEM images showing (A) pure liposomes; 3 nm AuNP LNAs: (B) high (C) low; 5 nm AuNP LNAs: (D) high and (E) low; (F) shows isolated AuNP in LNAs. Arrows in (C) denote partially packed LNAs; Arrow in (B) shows existence of possible disc/rod/worm like AuNP aggregates. Arrow in (D) shows presence of large AuNP aggregates and large AuNPs. Arrow in (E) depicts small AuNP clusters in lipid bilayer.


Langmuir 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

156

AuNP distribution in LNAs is attributed to non-homogeneous mixing in bilayers due to

157

cluster generation during AuNP-lipid bilayer thin film formation. Stable incorporation of

158

hydrophobic AuNPs in lipid bilayers while maintaining structural integrity occurs by unzipping of

159

lipid bilayers and AuNP insertion. Unzipping of bilayer by an isolated AuNP is an energy intensive

160

process, thus AuNPs form clusters to reduce generated void spaces in the bilayers during LNA

161

sample preparation.30, 58, 60 We envision three distinct regions in the dry AuNP-lipid film during

162

LNA preparation: 1) bilayer regions with large AuNP clusters, 2) bilayer regions with small AuNP

163

clusters and 3) apparently bare lipid bilayer regions/isolated AuNPs in bilayer. LNAs with fully

164

packed AuNPs would be generated from region 1, where a single cluster can form a low curvature

165

LNA itself. The detachment of bilayers containing AuNP clusters occurs at the interface with

166

region 3 (devoid of AuNP), during hydration or the successive extrusion cycles. LNAs with small

167

AuNP clusters are formed from region 2 (AuNP clusters are not large enough to form fully packed

168

LNAs on their own due to high curvature limitations). Additionally, due to clustering, isolated

169

AuNP inclusions in LNAs (Figure 2F) were rarely observed.

170

Bilayer Thickness from SANS. The average bilayer thickness was determined by small

171

angle neutron scattering (SANS) spanning temperature ranges from 40 oC to 60 oC for the pure

172

liposomes and all LNAs in D2O (solvent). SANS spectra were collected over a q range of 0.003

173

Å-1 ≤ q ≤ 0.523 Å-1 and bilayer thickness was determined by fitting the scattering data to a spherical

174

vesicle model61 over a q range of 0.02 Å-1 ≤ q ≤ 0.523 Å-1 as shown in Figure 3 (Data fitting and

175

analysis details in Supplementary Information S6). SANS provides an ensemble averaged bilayer

176

thickness for all liposomes and LNAs. The measured bilayer thickness should be interpreted

177

qualitatively only due to the presence of various AuNP distributions in the bilayer. An increase in

178

bilayer thickness was observed for all liposomes on transitioning from fluid phase to gel phase as


Page 10 of 39

Page 11 of 39 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

179

observed previously31,

180

temperature. Pure liposome bilayer thickness increased from 3.8 nm (60oC) to 4.3 nm (40oC)

181

across the transition temperature of 41oC (Figure 3B). Changes in bilayer thickness for LNAs with

182

low and high concentrations of 3 nm AuNP was not significantly different from pure liposomes.

183

However, 3.8 nm AuNP LNAs showed an increase in bilayer thickness compared to pure

184

liposomes at both concentrations as evidenced by the SANS spectra shift towards low-q for 3.8

185

nm AuNP LNAs (Supplementary Information Section S6). Cryo-TEM image analysis

186

demonstrated that completely packed LNAs were smaller in size than the LNAs with AuNP

187

clusters and bare liposomes in the same sample. In addition, statistical analysis of cryo-TEM

188

images yielded a > 80% population of LNAs with small AuNP clusters and bare liposomes. To

189

obtain a true statistical analysis of the sample population, SANS spectra were analyzed in the

190

intermediate to low q regime. No distinguishing features were observed between spectra obtained

191

from pure liposomes and LNAs in the low and intermediate q regime. The spherical vesicle model

192

fit was extended to intermediate and low q regime and yielded similar overall diameters for pure

193

liposomes and LNAs (≈ 90 nm mean liposome diameter). Unlike statistical analysis of cryo-TEM

194

images which is subjected to imaging biases, SANS demonstrates the prevalence of LNAs with

195

small AuNP clusters and bare liposomes with a negligible population of completely packed smaller

196

liposomes. However, analyzing LNA size distribution through SANS should be approached with

197

caution due to investigation of a limited low-q region.

62-64

, which is attributed to greater lipid ordering below transition


Langmuir 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 39

o

198

Figure 3. (A) SANS curves for pure liposomes and LNAs in fluid phase (60 C); curves have been offset by amounts mentioned in the plot for visual clarity. (B) Bilayer thickness calculated at different temperatures from spherical vesicle model fit of scattering data. Error bars represent one standard deviation from the mean. Dashed lines are for visual aid only.

199

Dynamic Light Scattering. Overall liposome motion can be classified into liposome

200

translational motion due to diffusion and shape fluctuation motions (bending, etc.). Translational

201

motion can be effectively measured at the appropriate q range using DLS which yields the

202

exponentially decaying autocorrelation function g2 (t) 65:

203

𝑔𝑔 2 (𝑡𝑡) = 𝐵𝐵 + 𝛽𝛽 𝑒𝑒𝑒𝑒𝑒𝑒 (−2𝛤𝛤𝑇𝑇 𝑡𝑡)

(1)

204

where B is the autocorrelation baseline at infinite delay, β is the correlation function amplitude

205

and decay rate ΓT is related to diffusion constant (DT) as:

206

𝛤𝛤𝑇𝑇 = 𝐷𝐷𝑇𝑇 𝑞𝑞 2

207 208 209

(2)

Pure liposome and LNA hydrodynamic radius (Rh) can be directly calculated from the StokesEinstein equation: 𝑘𝑘 𝑇𝑇

𝐵𝐵 𝑅𝑅ℎ = 6𝜋𝜋𝜋𝜋𝐷𝐷 ,

(3)

𝑇𝑇


Page 13 of 39 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

210

where T is the temperature, kB is the Boltzmann constant and η is the viscosity of D2O (function of

211

temperature). Autocorrelation curves demonstrates nearly identical decay pattern at 25 oC for pure

212

liposomes and LNAs (Supplementary Information, Section S7). Decay curve is described by a

213

single exponential decay which confirms a narrow LNA size distribution. This further corroborates

214

our assumption that smaller, completely-packed LNAs are a negligible population. LNAs are

215

primarily similar in size as pure liposomes. Diffusion coefficients obtained from fitting the

216

autocorrelation curve to Equation 1 and 2, depicts similar DT at 25 oC (0.47 Å2 ns-1 for pure

217

liposomes vs ~0.44 Å2 ns-1 and 0.43 Å2 ns-1 for 3 nm AuNP LNAs and 3.8 nm AuNP LNAs

218

respectively). DT corresponds to a hydrodynamic radii of ~ (40, 44 and 45) nm for pure liposomes,

219

3 nm AuNP LNAs and 3.8 nm AuNP LNAs respectively based on Equation 3 (Supplementary

220

Information, Figure S7). DT has been measured at very dilute lipid concentrations through DLS at

221

25 oC. Accurate determination of DT at NSE concentrations cannot be investigated through DLS

222

and actual DT lies between 0 < DT (Actual) < DT (DLS). Similarly, DT at various temperatures have been

223

estimated based on Equation 3 and solvent viscosity changes and does not represent diffusion

224

coefficients accurately, existing at temperature and concentration conditions during NSE

225

experiments.

226

NSE Measurements of Bilayer Bending Modulus. NSE was used to measure normalized

227

intermediate scattering functions (ISF) I(q, t)/I(q, 0) (q and t represent wave vector and Fourier

228

time, respectively) over a q range of 0.039 Å-1 to 0.1 Å-1 and Fourier time up to 100 ns with neutron

229

wavelengths of 8 Å and 11 Å. For pure membrane undulations (assuming no translational

230

diffusion), Zilman-Granek (ZG) model66 for single membrane fluctuations describes the bending

231

fluctuations in a membrane as:

232

𝐼𝐼(𝑞𝑞,𝑡𝑡)

𝐼𝐼(𝑞𝑞,0)

2

= exp[−(𝛤𝛤𝑏𝑏 𝑡𝑡)3 ]

(4)


Langmuir 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

233

234 235 236 237 238 239

Page 14 of 39

where Γb is the bending relaxation rate given by:

𝛤𝛤𝑏𝑏 =

1

𝑘𝑘 𝑇𝑇 2 𝑘𝑘 𝑇𝑇 0.025𝛾𝛾 � 𝐵𝐵𝜅𝜅� � � 𝐵𝐵𝜂𝜂 � 𝑞𝑞 3

= 𝛤𝛤𝑍𝑍𝑍𝑍 𝑞𝑞 3

(5)

where 𝜅𝜅̃ is the effective bending modulus, η is the solvent viscosity (D2O viscosity as a function of temperature in this work), kB is the Boltzmann constant, T is the temperature and 𝛾𝛾 approaches

unity when 𝜅𝜅̃ ≫ 𝑘𝑘𝐵𝐵 𝑇𝑇 . On incorporating intermonolayer frictions67-68 into the ZG model as

proposed by Watson and Brown,69 the effective bending modulus 𝜅𝜅̃ , can be substituted by intrinsic bending modulus κ, given by 𝜅𝜅̃ = 𝜅𝜅 + 2ℎ2 𝑘𝑘𝑚𝑚 where h is the height of neutral surface from bilayer

241

mid-plane and km is the monolayer area compressibility modulus. km is further depicted as 𝑘𝑘𝑚𝑚 =

242

denoted as hc.70 As discussed in an earlier paper71, monolayer bending modulus can be expressed

240

243 244 245 246

12𝜅𝜅𝑚𝑚 /ℎ𝑐𝑐2 , where 𝜅𝜅𝑚𝑚 is the monolayer bending modulus and monolayer hydrocarbon thickness is ℎ

2

as bilayer bending modulus as 𝜅𝜅𝑚𝑚 = 𝜅𝜅/2, which yields 𝜅𝜅̃ = 𝜅𝜅(1 + 48 �2ℎ � ), and an assumption 𝑐𝑐

of h/2hc = 0.5 to consider the neutral surface is at the hydrophobic-hydrophilic interface leads to 𝜅𝜅̃ = 13𝜅𝜅. Equation 5 has thus been used in the following form to calculate the intrinsic bending modulus:

1

𝑘𝑘 𝑇𝑇 2 𝑘𝑘 𝑇𝑇 0.0069 � 𝐵𝐵𝜅𝜅 � ( 𝐵𝐵𝜂𝜂 )𝑞𝑞 3

= 𝛤𝛤𝑍𝑍𝑍𝑍 𝑞𝑞 3

247

𝛤𝛤𝑏𝑏 =

248

Figure 4A shows a typical NSE decay curve for 3 nm AuNP LNAs (high) when no translation

249

diffusion is taken into account. NSE decay curves are fit to Equation 4 to yield a bending decay

250

rate which demonstrates a q3 dependence in the whole q range.


(6)

Page 15 of 39 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

251

Langmuir

Figure 4. Typical normalized intermediate scattering function (ISF) at different q for 3 nm o AuNP LNAs (high) at 60 C for (A) No translational diffusion assumed; solid lines are fit to Equation 4. (B) Effect on ISF due to 3 nm AuNP inclusions (high) relative to pure liposomes o at 60 C (No diffusion, solid lines fit to Equation 4). Errors in all figures represent one standard deviation from the mean.

252

Figure 4B shows a direct comparison of AuNP inclusion effect (3 nm AuNP high conc.)

253

on the DPPC/DPPG lipid bilayer I(q, t)/I(q, 0) at two q values at 60 oC. AuNP inclusions lead to

254

slightly larger relaxation rates (slightly faster relaxation of bilayer bending) compared to pure

255

liposomes, which is a clear indication of membrane softening according to ZG theory. NSE yields

256

an ensemble average bending moduli for all LNAs (fully packed LNAs, partially packed LNAs,

257

and vacant liposomes) and should be qualitatively interpreted with respect to pure liposome

258

bending modulus. Figure 5 shows the bending modulus calculated for all LNAs above the bilayer

259

transition temperature (summarized in Table 2). The bending modulus measured for the pure lipid

260

bilayer in the current work at 60 oC (≈ 14.7 kBT) is comparable to the bending modulus obtained

261

for pure DPPC bilayer (≈ 9.5 kBT) at 60 oC previously by NSE51 and low angle X-ray spectroscopy

262

(≈ 16.3 kBT).72 The bending modulus is an order of magnitude higher at temperatures near and

263

below the transition temperature (Figure 5 inset). This is expected as bilayer stiffening occurs due


Langmuir 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

264

to greater lipid ordering in the gel phase than in fluid phase.51, 71, 73 3 nm AuNP LNAs (high) show

265

an average 15% reduction in bending modulus relative to pure liposomes in the fluid phase. Change

266

in bending modulus was insignificant for 3 nm AuNP LNAs (low) in the fluid phase. 3.8 nm AuNP

267

LNAs however, shows ca. 15% bending modulus reduction at both low and high AuNP loadings

268

in the fluid phase. No bending modulus hysteresis was observed on reheating the 3 nm AuNP

269

LNAs (high) from gel phase to fluid phase, thus ruling out any hysteresis on membrane elastic

270

properties upon phase transition. NSE decay data must also be analyzed by taking translational

271

diffusion into account. However, incorporating diffusion contributions as measured through DLS

272

(at extremely dilute lipid concentrations) does not present an accurate analysis of the NSE data.

273

This account has been presented in detail in Supplementary Information, Section S8. Figure S8

274

depicts similar membrane softening as observed in Figure 5. On accounting for DT, membrane

275

softening in 3 nm AuNP LNAs (low) is also resolved. Bending moduli depicted in Table 2 (in

276

parenthesis) demonstrates an extreme value, absolute value lies between the two extreme cases as

277

summarized in Table 2.


Page 16 of 39

Page 17 of 39

22 20

Pure Liposomes 3 nm AuNP LNAs (low) 3 nm AuNP LNAs (high) 3.8 nm AuNP LNAs (low) 3.8 nm AuNP LNAs (high)

18

κ (kBT)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

16 14 12 50

52

54

56

58

60

T (oC)

278

Figure 5. Intrinsic bending moduli κ of pure DPPC/DPPG liposomes and LNAs assuming no diffusion contribution to NSE decay. Inset shows bending moduli over fluid and gel phase. Dashed lines are for visual aid only. Error bars represent one standard deviation from the mean.


Langmuir 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 18 of 39

Table 2. Bilayer thickness (d) measured from SANS data and bending o modulus (κ) at 60 C obtained from NSE measurements (values in parenthesis depict bending modulus when translational decay contributions are applied). Errors represent one standard deviation from mean.

279 280

NSE findings demonstrate both AuNP concentration and AuNP size have a significant

281

impact on lipid bilayer bending modulus. However, the impact of AuNP diameter in bending

282

modulus reduction appears more prominent as seen in case of 3.8 nm AuNP LNAs. A qualitative

283

assessment of the bilayer area compressibility modulus (KA) with AuNP inclusions was conducted,

284

based on SANS thickness measurements and bending modulus obtained from NSE. Bilayer

285

bending results in compression and stretching of the lipid leaflets and can be described as a thin

286

elastic sheet model relating κ and KA.74 Bending modulus κ is directly proportional to area

287

compressibility modulus KA as:

288

𝐾𝐾𝐴𝐴 = 𝛽𝛽𝛽𝛽/(2ℎ𝑐𝑐 )2

289

(8)

where hc is the monolayer thickness and β is a coupling constant that depends upon the degree of

290

coupling between the two monolayers. β ranges from 12 (strong coupling between monolayer

291

leaflets) to 48 (no coupling and leaflets can slide freely).75-76 Increase in hydrophobic thickness


Page 19 of 39 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

292

and reduction in bending modulus implies a drastic reduction in bilayer compressibility modulus

293

KA for LNAs from Eq. 8. Definitive quantification of KA was not undertaken due to insufficient

294

knowledge about the change in β in the presence of inclusions. Decrease in KA is a clear indicator

295

of an increase in thickness fluctuation dynamics49 in lipid bilayer which plays a pivotal role in

296

membrane protein functioning77 and enzyme catalysis78. Thickness fluctuation dynamics

297

measurement with AuNP inclusions in tail deuterated lipid bilayers is a potentially interesting

298

study to quantify thickness fluctuation dynamics, amplitude and membrane visco-elastic

299

properties.49, 71, 79

300

Decrease in bilayer bending modulus (decrease in area compressibility modulus) with

301

AuNP inclusions can be explained based on contribution from multiple possible factors. It has

302

been previously shown that bending modulus is directly related to bilayer thickness.80-81 However,

303

in the present case, the average bilayer thickening is a result of membrane unzipping to situate the

304

AuNP/ AuNP cluster inclusion in the bilayer core. However, coupling between monolayers should

305

be severely compromised on introducing isolated AuNP inclusions or cluster inclusions.

306

Additionally, theoretical80 and X-ray diffraction82 studies on surfactant lamellas have

307

demonstrated dramatic decreases in bending modulus with increasing area per molecule (bulkier

308

head groups for surfactants). Increased area per molecule directly implies a less confined chain

309

and smaller loss of conformational entropy during lamellar bending. Similarly, total bilayer

310

volume in the vicinity of AuNP and AuNP cluster inclusion is greatly increased, but number of

311

lipids in the specific volume remains constant. Thus, increase in bilayer volume can be

312

compensated only through stretching and conformation change of lipid tails, in other words

313

increasing the area per lipid. Previous researchers64, 83 have well documented the phenomena of

314

increasing area per lipid with increasing temperature (transition from gel to fluid phase). The


Langmuir 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

315

AuNP inclusions can thus be envisioned to generate higher local fluidity compared to the bulk

316

bilayer leading to an average bilayer softening. It has been shown previously, that small alkyl chain

317

additives at relatively low mole fractions can cause substantial reduction in bending modulus.80-81

318

AuNP inclusions in the present research is functionalized with small alkyl chain dodecanethiol,

319

which might substantially contribute towards disrupting conformational packing, leading to

320

membrane softening.

321

CONCLUSION

322

An investigation into effects of hydrophobic AuNP inclusions on lipid bilayer bending

323

modulus was conducted through NSE spectroscopy. AuNPs with average diameters of ~ 3 nm and

324

~ 5.5 nm were embedded into DPPC/DPPG lipid bilayers. It was observed that AuNPs above a

325

particular threshold diameter were not included in the DPPC/DPPG lipid bilayer, resulting in

326

decreased loading efficiency for larger AuNPs. Non-homogeneous lipid and AuNP mixing during

327

dry film formation leads to apparently bare, partially packed and completely packed LNAs.

328

Average LNA bilayer thickening was observed depending on size and concentration of AuNP

329

inclusions. On investigating LNA membrane dynamics with NSE spectroscopy, a general

330

membrane softening was observed relative to pure liposomes which manifested in the form of

331

reduced bending modulus. AuNP concentration and AuNP size were both factors in determining

332

magnitude of membrane softening. On considering translational contribution to NSE decay,

333

similar bilayer softening trends are retained. Our experimental results qualitatively point towards

334

a decrease in area compressibility modulus and subsequently an increased thickness fluctuation

335

dynamics. We demonstrate NSE to be a crucial tool in determining inclusion effects on membrane

336

dynamics and mechanical properties. These studies have potential implications for membrane

337

biophysical characterization and designing innovative drug delivery vehicles.


Page 20 of 39

Page 21 of 39 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

338

EXPERIMENTAL METHODS

339

Materials.

340

Hydrogen tetrachloroaurate (HAuCl4.3H2O, 99.9% purity), dodecanethiol (

Phospholipid Bilayer Softening Due to Hydrophobic Gold Nanoparticle

Recommend Documents