CENTRIFUGATION-BASED ASSAY FOR EXAMINING NANOPARTICLE-LIPID MEMBRANE BINDING AND DISRUPTION

Physical disruption of cellular membranes arising from interactions with engineered nanoparticles is an important, but poorly understood aspect of nanotoxicology and nanomedicine. Model cellular membranes (i.e. lipid bilayers) can be used to identify interaction mechanisms, and most studies have largely focused on lipid bilayers supported on solid planar or spherical substrates. While useful and informative, these systems do not accurately represent an intact cell membrane because they restrict the elastic motion of the bilayer and the capacity for mechanical changes. Free standing bilayers are preferred, but add complexity. Given the importance of nanoparticle–membrane interactions in nanotoxicology and nanomedicine, and the vast range in nanoparticle composition, size, shape, and surface functionalization, there is a need to develop techniques that can rapidly and inexpensively analyze the membranenanoparticle activity by using free standing or unsupported membranes. This work develops a centrifugation-based assay that can analyze the membranenanoparticle activity as a function of nanoparticle surface functionalization, membrane lipid composition, and monovalent salt concentration (NaCl). Free standing, unsupported vesicles were used to gain relevant information on elastic membrane deformation and vesicle destabilization due to nanoparticle binding. Silver nanoparticles were chosen due to their widespread biological applications and surface plasmon resonance (SPR) properties. UV-vis based centrifugation assay, coupled with cryo-TEM and DLS analysis, was proposed to screen nanoparticle-membrane interactions; silver nanoparticles binding ratio RSPR was calculated as a function of Ag nanoparticle coating and vesicle composition. Study showed that strong electrostatic attraction led to significant sedimentation, vesicle / membrane disruption and higher RSPR value; in contrast, systems that exhibited weak or no electrostatic attraction did not show significant sedimentation, membrane disruption or high RSPR value. The centrifugation assay provides a rapid and straightforward way to screen nanoparticle– membrane interactions.


Nanoparticle -membrane interaction
Over the past two decades, nanoparticles have been increasingly used for biological applications such as antimicrobial agents, therapeutics imaging, diagnosis and targeted drug / gene delivery.  For example, silver nanoparticles have been used for disinfection and creating antifouling surfaces. 22 Superparamagnetic iron oxide (SPIO) and gold (Au) nanoparticles have been reported in the field of tumor disgnostics and cancer treatment. 9-10, 16-21 Semiconducting nanocrystals, e.g. quantum dots, were used to improve biological imaging for medical diagnostics, 14 and these crystals were able to offer resolutions up to 1,000 times better than conventional dyes used in many biological tests. Furthermore, multifunctional nanoparticles, which have both diagnostic and therapeutic functions, are able to stimulate gene or drug release at targeted location when triggered by external stimuli, and minimize the risk to normal tissues. [26][27][28][29][30] The introduction of nanoparticles into biological processes leads to new challenges: (1) the characterization of the interaction between nanoparticles and cell membranes; (2) the evaluation of biocompatibility between nanoparticles and cell membranes; (3) the measurement of the cytotoxicity induced by nanoparticles and (4) the prediction of the impact of nanoparticles to biological systems. It has been observed that nanoparticles were able to bind to membrane, causing local changes in membrane curvature. [34][35][36][37] The extent of nanoparticle-induced biophysical and/or biochemical changes on cell membranes would be dependent on the size, charge, surface reactivity, surface chemistry and compositions of nanoparticles. [38][39][40][41][42] It has been studied that nanoparticles may introduce carcinogenic risks, which may be triggered by the production of reactive oxygen species (ROS) by macrophages attempting to destroy foreign materials on the inflammation sites. The ROS produced in this process, may cause DNA damage as well as inflammatory lesions associated with carcinogenesis. [43][44] The broad applications of nanoparticles and their toxicity prompt investigations not only on their functional mechanisms, but also on their cytotoxicity. The size, charge, surface chemistry, and compositions of nanoparticles are important parameters for their physicochemical properties and biological applications. Therefore, there is the urgency to determine how the size, charge, and surface chemistry of nanoparticles influence their functional mechanism and their cytotoxicity. [45][46][47] In this study, nanoparticle -membrane interaction was characterized in order to provide fundamental understanding of the interaction between nanoparticles and cellular systems, and to provide guidance in the design and development of safe nanoparticles for biological applications.

Silver nanoparticle -membrane interaction
In this work, silver nanoparticles (AgNPs) were chosen to study the nanoparticle membrane interaction due to their widespread biological applications and surface plasmon resonance (SPR) property. Firstly, silver nanoparticles are important antimicrobial agents. [48][49][50][51][52] AgNPs are able to destroy bacterial cell walls, to trigger conformational changes of the ion channel, to cause changes of channel opening and dysfunctions. Therefore understanding silver nanoparticle -membrane interactions is essential to understand their toxic effects on both human health and the environment.
Secondly, when silver nanoparticles interact with light, the conduction electrons on the silver surface oscillates at specific wavelength, giving AgNPs the surface plasmon resonance (SPR) property. 53-54 SPR can be assessed by ultraviolet-visible (UV-vis) spectroscopy. Its absorbance and wavelength are functions of AgNP concentrations and aggregation states. Therefore, SPR allows the determination of both AgNP concentrations in supernatant and sediment phases, and AgNP aggregation states in solution and after membrane binding.
Experiments were conducted using anionic, cationic and neutral silver nanoparticles and lipid bilayer vesicles. Unsupported vesicles were used to allow elastic membrane deformation and vesicle destabilization due to nanoparticle binding.
Supernatant and sediment phases were characterized by cryogenic transmission electron microscopy (cryo-TEM) to directly image nanoparticle membrane binding and to connect vesicle stability and structure with the observed centrifugation behavior.

Specific Research Aim and Hypothesis
Aim: Determine nanoparticle -membrane interactions; quantify electrostatic interactions as a function of nanoparticle size, surface chemistry and membrane composition; examine the degree of nanoparticle aggregation at membrane / water interfaces; and the effects of aggregation on membrane disruption.
Hypothesis: Nanoparticle -membrane interactions lead to nanoparticle aggregation at membrane / water interfaces, and cause membrane disruption and pore formation. These phenomena can be examined by employing a centrifugation-based assay.         Table 2-1.

Experimental techniques
For the biomedical applications of nanoparticles, it is necessary to understand cell membrane-nanoparticle interactions and to assess the safety of nanomaterials.
membrane-nanoparticle interaction studies are complicated 4-13 by (1)  nanoparticles at 30 nm show more significant Tm shifts compared to SPIO at 16 nm. 14 A decrease of Tm and a broadening of the transition were observed on supported bilayers which were formed on the 100 nm silica beads. 15 Atomic force microscopy (AFM) was developed to quantitatively measure the binding force of nanoparticles with cell membranes and to study the morphological changes of the membranes due to their interaction with nanoparticles. For example, through AFM studies, it was found that electrostatic interaction drives the binding of nanoparticles to membranes which causes membrane disruption. [16][17] AFM study by Roiter et al. 18 indicated that nanoscale pores were formed on the lipid bilayer when the diameter of nanoparticle was smaller than 22 nm. Furthermore, nanoparticles would be enveloped by the lipid bilayers when the diameter was larger than 22 nm.
Quartz crystal microbalance with dissipation (QCM-D) has been a popular method recently because it is sensitive to frequency changes (Δf) and energy dissipation (ΔD) when nanoparticles bind to membrane. [19][20][21][22] Inductively coupled plasma optical emission spectrometry (ICP-OES) or mass spectrometry (MS) has also been used, for example, to probe the interaction between functionalized Au nanoparticles and silica sphere-supported lipid membranes (SSLMs) by measuring the concentrations of Au nanoparticles both in the aqueous electrolytes (supernatant) and in/on the lipid bilayers. 23 The electrophysiological approach 24 coupled with the droplet-in-oil methodology has been employed to study the interaction between nanoparticles and cell membranes. In the report by De Palnque et al., the droplet-in-oil methodology was first used to create lipid bilayers through the self-assembly of two water droplets coated with a lipid monolayer at water-oil interface. Subsequently, it was found that when silica nanospheres covered as low as 0.02% of the surface of the bilayers, the electrophysiological approach was able to detect bilayer current change caused by nanoparticle adsorption to lipid bilayers. Another electrical approach 25 quantified nanoparticle adsorption to membrane by detecting capacitive increase of suspended planar lipid bilayers.
In addition, in recent years, computer simulation is gaining increasing attention for the study of nanoparticle-membrane interactions. [26][27][28][29][30] These studies have also provided critical information on the relationships between the interactions and the composition, geometry, and physicochemical properties of the nanoparticles.

Centrifugation-based assays
For decades researchers have utilized centrifugation-based assays to determine protein membrane affinity or binding, where the amount of bound protein can be determined by a mass balance taking into account the supernatant (free protein) and sediment (membrane-bound protein) phases. 31 Centrifugation methods to assay proteinmembrane binding affinity have proven to be simple and inexpensive techniques.

Abstract
Centrifugation-based assays are commonly employed to study protein-membrane affinity or binding using lipid bilayer vesicles. An analogous assay has been developed to study nanoparticle-membrane interactions as a function of nanoparticle surface functionalization, membrane lipid composition, and monovalent salt concentration

Introduction
Nanoparticles interact with cell membranes by first binding at the membrane-water interface. Interfacial interactions and the adhesive binding strength are based on nanoparticle surface functionalization and membrane lipid composition, and control the extent to which a nanoparticle will penetrate into the membrane and disrupt lipid organization and membrane structure. 1,2 There is evidence that these nanoparticlemembrane interactions inhibit cellular function and contribute to nanoparticle toxicity. 3-6 A number of experimental techniques have been used to study nanoparticle interactions with model cell membranes, which are commonly employed to investigate binding mechanisms and biophysical changes in membrane structure, including atomic force microscopy, 7-9 fluorescence microscopy, 10 quartz crystal microbalance, [11][12][13][14][15] differential or isothermal scanning calorimetry, [16][17][18][19] electrical capacitance, 20 and cryogenic transmission electron microscopy. 15,21,22 These studies have provided critical information that will be needed to develop approaches that can predict nanoparticlemembrane interactions based on nanoparticle composition, geometry, and physicochemical properties.  Ag-NH and Ag-PEG nanoparticles contain an additional PEI or PEG-grafted PEI coating, respectively. (D) TEM micrograph of Ag-PEG nanoparticles.

Membrane (vesicle) preparation
Vesicles were prepared at 10 mM total lipid concentration in DI water or NaCl solutions (10 mM or 100 mM). Lipids, dissolved in chloroform, and water were added to a round-bottom flask, vortexed for 1 min, and then subjected to rotary evaporation at 50 0 C to remove chloroform. After the chloroform was removed, the flask containing vesicles was transferred to a bath sonicator at 50 0 C and sonicated for 30 min. The vesicles were sized by extrusion through double-stacked polycarbonate membranes with 100 nm pore diameters. Neutral membranes were prepared using DPPC and anionic or cationic membranes were prepared using mixtures of DPPC with DPPG or DPTAP at 3 : 1 or 1 : 1 molar ratios, respectively. A representative cryo TEM images of DPPC/DPPG vesicles is shown in Fig. 3-1B.

Cryogenic transmission electron microscopy
Vitrification of sample specimens for cryo-TEM was performed using a Vitrobot (FEI Company), which is a robotic preparation system with controlled temperature and humidity. Specimens were prepared on Quantifoil grids with 2 mm holey-carbon on 200 square mesh copper (Electron Microscopy Sciences, Hat-field, PA). After the sample was equilibrated within the Vitrobot at 25 o C and 100% humidity for 30 min, the grid was plunged into the sample, withdrawn, and blotted to yield a thin specimen film. The specimen was then vitrified by plunging the grid into liquid ethane, and transferred to liquid nitrogen. Imaging was performed in a cooled stage (model 915, Gatan Inc., Pleasanton, CA) using a JEOL JEM-2100F TEM (Peabody, MA). Image analysis was performed using Image J software.

Dynamic light scattering and zeta potential
Dynamic light scattering (DLS) and zeta potential (z) measurements were

UV-vis spectroscopy
UV-vis spectroscopy was conducted using an Agilent Cary 50 (Santa Clara, CA) spectrophotometer with a Peltier cuvette holder for temperature control. Samples were equilibrated at 25 o C for 3 min in quartz cuvettes (10 mm path length) capped with PTFE lids. Absorbance spectra were conducted in triplicate and the SPR peak height, peak area, and peak position (wavelength) of each spectrum was analyzed by OriginPro software (version 9.0).

Centrifugation assay
A schematic of the UV-vis centrifugation assay is given in Fig. 3 (Table S1 †).
AgNP binding was inferred based on mass balance obtained by UV-vis analysis of the SPR where SPR peak area was a linear function of AgNP concentration. Apparent AgNP binding was determined as the ratio RSPR Where ∆ASPR, NPs+vesicles was the change in the SPR peak area for AgNPs + vesicles before and after centrifugation, and ∆ASPR, NPs was the change for AgNPs. This approach takes into account the inherent sedimentation behavior of the AgNPs. RSPR = 1 indicated that there was no difference in sedimentation relative to the AgNPs alone.

Characterization of vesicles and AgNPs
The hydrodynamic diameters and zeta potentials of the AgNPs and vesicles employed are summarized in Tables 3-1 (Table 3-2). 32

Centrifugation assay
RSPR results for Ag-PEG, Ag-COOH, and Ag-NH vesicle binding are shown in show that the assay is suitable for fluid phase membranes as well as gel phase (DPPC/DPPG) membranes (Fig. 3-4).

AgNP binding and aggregation
Cryo-TEM results are consistent with those in Fig. 3-3. Oppositely charged nanoparticles strongly interact with and bind to vesicles, leading to vesicle aggregation and disruption. There was also evidence that AgNP binding led to nanoparticle aggregation at the membrane-water interface. Shifts in the SPR, ΔλSPR, which are sensitive to AgNP size, aggregation state, and surface functionalization, and the presence of adsorbed molecules, [36][37][38][39] were examined to investigate this further. Fig. 3-7 demonstrates this analysis for DPPC/DPTAP (3 : 1) vesicles where the SPR for AgNPs are compared to AgNP + vesicle mixtures before centrifugation (ΔλSPR = λSPR, NP+V -λSPR, NP). All shifts in SPR were 'red-shifts' (Table 3-3) and correlated with electrostatic AgNP-vesicle attraction (

Effect of salt concentration
Electrostatic interactions were further probed by varying salt concentration. The studies focused on anionic DPPC/DPPG (1 : 1), which represents a model bacterial membrane, Ag-PEG nanoparticles, and monovalent NaCl. Ag-PEG was selected because PEG coatings are commonly employed in nanomedicine and provide a protective coating that resists protein adsorption. [43][44][45] In conjunction with centrifugation results (RSPR, Fig. 3-8A), cryo-TEM micrographs were analyzed (Fig. 3-8B and C) to determine apparent AgNP membrane-water partition coefficients, K, and to compare the aggregate number for bound and unbound AgNPs (Fig. 3-8D). and K was ~1, denoting an even distribution of AgNPs between the aqueous phase and the membrane. Aggregates were also observed bound to vesicles in the sediment (C2) and significantly distorting the membranes (C3 where A132 is the effective Hamaker constant based, R are radii, k is the Boltzmann constant, T is temperature, and h is the surface separation distance based on dh.
Subscripts 1, 2, and 3 denote the AgNP, vesicle, and water, respectively. A132 was Given that the AgNPs contained a thick polymer coating, A22 for polyethylene glycol was used. 46 Velec was calculated as = 4 0 1 2 ( 1 2 where εr is the dielectric constant of water, ε0 is the permittivity of free space, Φ are the surface potentials (taken as ζ in DI water), and k is the inverse Debye length. In DI water, k was based on the Na + counterion concentration of DPPG (4 mM), and this concentration was added to the 10 and 100 mM NaCl solutions (Fig. 3-9).
For DPPC/DPPG + Ag-PEG, an energy barrier exists near h = 0.5 nm due to electrostatic repulsion. This barrier decreases with increasing NaCl concentration due to charge screening, consistent with the increasing RSPR observed in Fig. 3-8. While this barrier did hinder Ag-PEG binding, it did not prevent it based on the cryo-TEM results.
For DPPC/DPPG + Ag-NH, strong electrostatic attraction was observed at all NaCl concentrations despite charge screening. This analysis explains why there was little change in RSPR with Ag-NH in salt solution. For 0 mM NaCl, the Na+ counterion concentration (4 mM) associated with DPPG was used to determine 1/k.

Conclusion
A UV-vis based centrifugation assay, coupled with cryo-TEM and DLS analysis, was introduced as a method for examining nanoparticle-membrane interactions. In analogous protein-membrane centrifugation assays, one can directly measure bound and unbound protein concentrations. This is not as straightforward for the nanoparticle-membrane assay. As opposed to a direct measurement, the nanoparticle-membrane assay reflects changes in the colloidal stability of a sample due to heteroaggregation that is dependent upon the degree of nanoparticle-membrane binding. AgNP + vesicles systems that exhibited strong electrostatic attraction led to significant sedimentation and vesicle/membrane disruption. In contrast, systems that exhibited minimal or no electrostatic attraction did not show significant changes in sedimentation behavior or membrane disruption. This suggests that additional analysis (e.g. imaging) may be needed in conjunction with this assay when examining weakly interacting vesicle-nanoparticle systems. Further optimization of the assay, including centrifugation conditions, vesicle size, nanoparticle concentration, may also improve the ability to examine such systems.
Collectively, the trends observed for nanoparticle binding and membrane disruption as a function of nanoparticle surface chemistry and lipid composition are consistent with previous studies that have demonstrated nanoparticle binding and deformation in small vesicles, 13 giant unilamellar vesicles, 6, 40 planar bilayers, 4 and lipid monolayers 41 ; nanoparticle partitioning to supported lipid bilayers; 20 and nanoparticle binding and leakage from vesicles. 42 By analyzing shifts in SPR wavelength and comparing to cryo-TEM micrographs, it was possible to discern different modes of modes of AgNP binding; individual AgNP binding followed by aggregation at membrane/water interfaces due to charge neutralization, or aggregate AgNP binding due to aggregation in solution caused by charge screening.

CHAPTER 4 CONCLUSION
This study proposes an easy assay analysis to measure surface plasmon resonance RSPR binding and predict interactions of silver nanoparticles with vesicles.
RSPR value categorizes three types of binding: oppositely charged particles and vesicles demonstrate destructive interaction, causing vesicles disrupted or totally destroyed.
Oppositely charged nanoparticles have strong interactions with vesicles, the binding between nanoparticless and vesicles cause nanoparticle aggregation, and nanoparticle aggregations with vesicles lead to significant sedimentation; similarly charged particles and vesicles show very weak or zero binding, in the between that is moderate binding.
Salt concentration does not influence interactions between oppositely charged particles and vesicles because stronger charge maintain nanoparticle stability, however, salt plays significant role for moderate binding of Ag-PEG particles with DPPC/DPPG, higher salt concentration makes nanoparticles aggregate , particles aggregates interact with the vesicles rather than individual particles; nanoparticle aggregates are capable of penetrating into the vesicles and inducing local changes in membrane curvature.
Compared to protein-membrane centrifugation assays, nanoparticle-membrane assays do not allow one to directly quantify the degree of nanoparticle binding or the membrane/water partition coefficient of the nanoparticle. Rather, the nanoparticlemembrane assay reflects the change in the colloidal stability of the vesicle with nanoparticle sample due to nanoparticle-membrane binding. Two aspects are clear,