Polymer Coating and Lipid Phases Regulate Semi-Conductor

Oct 19, 2018 - Support. Get Help · For Advertisers · Institutional Sales; Live Chat. Partners. Atypon · CHORUS · COPE · COUNTER · CrossRef · CrossChec...
0 downloads 0 Views 2MB Size
Subscriber access provided by The University of Texas at El Paso (UTEP)

Article

Polymer Coating and Lipid Phases Regulate Semi-Conductor Nanorods’ Interaction with Neuronal Membranes: a Modeling Approach Barbara Salis, Giammarino Pugliese, Teresa Pellegrino, Alberto Diaspro, and Silvia Dante ACS Chem. Neurosci., Just Accepted Manuscript • DOI: 10.1021/acschemneuro.8b00466 • Publication Date (Web): 19 Oct 2018 Downloaded from http://pubs.acs.org on October 24, 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 30 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

ACS Chemical Neuroscience

Polymer Coating and Lipid Phases Regulate Semi-Conductor Nanorods’ Interaction with Neuronal Membranes: a Modeling Approach Barbara Salis1,2*, Giammarino Pugliese3, Teresa Pellegrino3, Alberto Diaspro2,4, Silvia Dante2* 1Dipartimento

di Informatica, Bioingegneria, Robotica e Ingegneria dei Sistemi, Università di Genova, Italy Istituto Italiano di Tecnologia, Genova, Italy 3Nanomaterials for Biomedical Applications, Istituto Italiano di Tecnologia, Genova, Italy 4Dipartimento di Fisica, Università di Genova, Italy 2Nanoscopy&NIC@IIT,

*Corresponding author: [email protected]; [email protected]

Abstract The interplay between nanoparticles (NPs) and cell membranes is extremely important with regard to using NPs in biology applications. With the aim of unravelling the dominating factors on the molecular scale, we have studied the interaction between polymer-coated semiconductor nanorods (NRs) made of Cadmium Selenium/Cadmium Sulfur and model lipid membranes. The zeta potential (ζ) of the NRs was tuned from having a negative value (-24 mV) to having a positive one (+11 mV) by changing the amine content in the polymer coating. Supported Lipid Bilayers (SLBs) and Lipid Monolayers (LMs) were used as model membranes. Lipid mixtures containing anionic or cationic lipids were employed in order to change the membrane ζ from -77 mV to + 49 mV; lipids with saturated hydrophobic chains were used to create phase-separated gel domains. The NRs’ adsorption to the SLBs was monitored by quartz crystal microbalance with dissipation monitoring; interactions with LMs with the same lipid composition were measured by surface pressure-area isotherms. The results showed that the NRs only interact with the model membrane if the mutual Δζ is higher than 70 mV; at the air-water interface, positively charged NRs remove lipids from the anionic lipid mixtures and the negative ones penetrate the space between the polar heads in the cationic mixtures. However, the presence of gel domains in the membrane inhibits this interaction. The results of the Derjaguin-Landau-Verwey-Overbeek model frame indicate that the interaction occurs not only due to electrostatic and van der Waals forces, but also to steric and/or hydration forces. Keywords: nanorods, semiconductor nanoparticles, lipid bilayers, QCM-D, Langmuir, DLVO, zeta potential

ACS Paragon Plus Environment

ACS Chemical Neuroscience 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

Introduction Nanostructured materials and, in particular, inorganic nanoparticles (NPs) with nanoscale dimensions (between 1 and 100 nm) are largely employed in biomedical applications today. Engineered surface modified NPs enable us to direct NPs’ effects specifically towards living organisms and cells, rendering them suitable for use in cancer treatments, bio-imaging and drug delivery techniques1-4. Evaluating the NPs’ interactions in a physiological environment, as well as assessing the effects they have on human beings, can be challenging due to the following complexities with regard to the systems that are under study: there are a broad variety of nanomaterial properties that have to be considered (namely the nanoparticles’ size, surface charge density and shape5-9); there are great variations in the biological targets (namely cell lines and biomarkers7); the living organisms are extremely intricate; and the conditions (such as the NP concentration8) under which the interactions are examined. All these variables affect the interactions and, in turn , could cause possible hazardous effects. NPs can also play active roles in mediating biological effects: it has been shown that the cell uptake of fluorescent Cadmium Selenium/Cadmium Sulfur (CdSe/CdS) nanorods (NRs) by Hydra vulgaris, a simple model organism, can be tuned by modifying the NR surface charge.10 This interaction can lead to biological responses in the living animal (a tentacle writhing response)6. Recently, we reported that the NPs’ interaction with the neuron cell membrane (regardless of the NPs composition, size and shape) was driven by the surface charge of the NPs and, in the case of negatively charged particles, it has elicited a neuronal electrical response7, whereas no interaction was observed in cells carrying a static membrane potential (such as glial cells). These events are reasonably due to the difference in superficial potential between the neuronal membrane and the external functionalization of the NRs. In this work, we investigated this phenomenon using bioengineered NRs and model membranes. We investigated the role of membrane complexity and superficial charge using nanoparticles that have been shown to affect the neuronal electrophysiology7 and, therefore, they could be envisaged as action potential modulators. The complexity of the cellular membrane and the variety of NP properties that are available make the prediction of these interactions challenging. Using model membranes that can mimic a basic plasma membrane structure provides a system that is suitable for studying membrane-NP interactions.11, 12 The use of such a system can simplify the study, since experimental studies on the cytotoxicity of engineered nanomaterials have been reported in literature using model membranes13, 14. These studies have helped us to identify a number of mechanisms by which nanomaterials induce toxicity, including membrane damage, such as the formation of thinned regions and holes14, 15.

ACS Paragon Plus Environment

Page 2 of 30

Page 3 of 30 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

ACS Chemical Neuroscience

Many studies have been performed so far in the field of NP-cell and model membrane interaction, especially with regard to gold spherical NPs16-18, but there have not been many on rod-shaped nanostructured materials; the importance of aspect ratio in the NP dimensions that are related to membrane cytotoxicity was investigated by Lins et al. 19, and they demonstrated that the NPs’ size (i.e. rods with higher aspect ratio) influences the elasticity of the model membranes. Furthermore, Stellacci et al. 16 underlined the importance of the polymeric ligands outside the metal core in their study of these interactions. In some other studies also the effect of the so called protein corona has been taken into account20: here we avoided on purpose the presence of proteins into the buffer in order to specifically address the effect of the polymer coating. In this work, we focus on the behavior of polymer coated CdSe/CdS NRs (which were developed by our group and used in refs6, 7) with model lipid membranes composed by different mixtures of 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC), 1-palmitoyl-2-oleoyl-sn-glycero-3-phospho-L-serine (sodium salt) (POPS), 1,2dipalmitoyl-sn-glycero-3-phosphocholine (DPPC) and N-[1-(2,3-Dioleoyloxy)propyl]N,N,N-trimethylammonium chloride (DOTAP) in order to elucidate the interaction mechanism on a molecular scale. We performed a systematic study of the interaction between the NRs using different surface charges and Supported Lipid Bilayers (SLBs). Although it is known that the effect of the solid substrate could non-negligibly influence some properties of the bilayer, such as its bending modulus21, SLBs are reliable model systems for cell membranes thanks to their similarities with biological ones (such as their lateral fluidity, ability to incorporate proteins, and impermeability to ionic species11, 22). The surface potential of the model membranes was modulated by changing the membrane composition, mixing zwitterionic lipids with cationic or anionic ones. The presence of gel domains in fluid membranes was also investigated by introducing lipids with saturated alkyl chains to the mixtures. At room temperature, saturated lipids are in a gel phase, generating gel nanosized domains that are similar to the plasmatic membrane’s lipid rafts23, 24. The presence of the nanosized domains in the SLBs obtained by vesicle fusion has been well documented in the literature by atomic force microscopy both for anionic/zwitterionic mixtures25-28 and for cationic/zwitterionic mixtures29. We used a Quartz Crystal Microbalance with Dissipation monitoring (QCM-D) technique to investigate the SLB, and we used LMs to perform pressure-area isotherms (π-A) in a Langmuir trough30. The results were interpreted using the DLVO theory in which the net interaction energy between two surfaces is calculated as the sum of the electrostatic and van der Waals energies.31 Not every interaction that the system predicted actually occurred, indicating that the interactions are driven by more than just electrostatic forces.

ACS Paragon Plus Environment

ACS Chemical Neuroscience 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 30

Results And Discussion LUV and NR characterization: DLS and ζ. Large unilamellar vesicles (LUVs) of different compositions were fabricated by extrusion with a diameter of 100 nm. The main component of the mixtures was the zwitterionic POPC, which was mixed either with cationic DOTAP or anionic POPS. POPC was chosen as the main component of the mixtures since it is the main component of plasmatic cell membranes32. The mixture containing POPC/POPS 90:10 mol:mol was chosen to mimic the physiological charge of a cell membrane. The mixtures containing DOTAP are not physiological; however, DOTAP is widely used as a cationic lipid in transfection experiments33 and its interaction with NPs can be of interest. POPC, POPS, and DOTAP have unsaturated hydrocarbon chains and are in the fluid phase at room temperature. The molar composition of the mixtures enables us to tune the ζ of the vesicles (measured in PBS 1.4 mM) within the range of 77 mV and +49 mV (Table 1), with the respective standard deviation. It is important to underline that the effective ζ of the employed SLB systems could be different from the ζ calculated by DLS for the corresponding vesicles: this is due to the compositional asymmetry by which the bilayer on solid supports could be characterized34. To investigate the effect of the lipid phases, lipid mixtures containing 5, 10 and 20% molar amounts of saturated DPPC were fabricated, keeping the molar percentage of anionic/cationic lipid constant. The vesicle dispersions containing DPPC were identical in size (with a 100 nm diameter) and in ζ to their corresponding in the fluid phase (Table 1). The NRs’ functionalization resulted in negatively (NR-) and positively (NR+) charged NRs, as is depicted in table 2, along with their relative ζ and amount of EDC, PEG and DMEDA used with respect to the number of particles. Table 1. ζ in mV of lipid mixtures studied Fluid lipid mixtures (mol:mol) POPC/POPS 75:25 POPC/POPS 90:10 POPC POPC/DOTAP 75:25 POPC/DOTAP 50:50 Phase separated lipid mixtures (mol:mol:mol) POPC/POPS/DPPC 85:10:5 POPC/POPS/DPPC 80:10:10 POPC/POPS/DPPC 70:10:20 POPC/DOTAP/DPPC 45:50:5 POPC/DOTAP/DPPC 40:50:10

ACS Paragon Plus Environment

+/+/+ + + +

ζ (mV) -77 ± 3 -54 ± 2 -23 ± 1 13 ± 1 49 ± 2 ζ (mV) -54 ± 2 -54 ± 2 -54 ± 2 49 ± 2 49 ± 2

Page 5 of 30 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

ACS Chemical Neuroscience

POPC/DOTAP/DPPC 30:50:20

+

49 ± 2

Table 2. Ratio of molecules to EDC, PEG or DMEDA per number of NR particles used in tuning the NRs’ superficial charge Sample name NR+ NR-

ζ (mV) +11 -24

EDC/NRs 5e+05 3e+05

PEG/NRs 500 500

DMEDA/NRs 5000 1000

QCM-D measurements: fluid bilayers. The NRs’ adsorption to the SLB was monitored as a function of the lipid charge by employing fluid bilayers. The SLB formation process was monitored by QCM-D measurements which showed differences according to the composition of the lipid vesicles. POPC, POPC/POPS 90:10 and POPC/POPS 75:25 (mol:mol ratios) LUVs showed the usual SLB formation kinetics after all of the vesicles had been adsorbed and subsequently fused on the SiO2 substrate35. The adsorption of entire vesicles on the surface was indicated by a large negative shift in the frequency, i.e., Δf7/7 ≈ -72 Hz (where Δf7/7 is the measured shift in frequency relative to the 7th overtone, normalized by the overtone number), and a dissipation value ΔD ≈ 12e+06. When the vesicles started to fuse, the frequency decreased until a stable value was reached (Δf7/7 ≈ -26 Hz) and the dissipation value was close to 0. The fusion process lasted 40 min. For positively charged LUVs, the formation of SLB was faster and there was no adsorption of entire vesicles. However, an SLB was directly formed in 20 min, as was indicated by a Δf7/7 ≈ -16 Hz and a constant value of ΔD = 0. In the case of SLBs containing DOTAP, there was no full coverage of the sensor’s surface, as can be understood by the differences in the Δf7/7 at the end of the SLBs’ formation process between POPC/POPS 90:10 and POPC/DOTAP 50:50 (mol:mol ratios) (Figure S1). The NRs’ behavior was also tested on bare SiO2 substrates, and we discovered that the NR+ were adsorbing on the solid substrate, whereas NR- did not interact with the sensor’s surface (Figure S2). After the formation of SLBs, the interaction with both NRs was systematically tested. After the interaction of NRs with SLB (ΔDNR- ≈ 0.5e+06, ΔDNR+ ≈ 2.5e+06), ΔD values were low enough to consider the system as a rigid layer, allowing the Sauerbray equation36 to be used to calculate the adsorbed mass. Figure 1 shows the mass of the NRs that was adsorbed on the SLBs as a function of the amount of charged lipids that are present in the mixture. An NR- was adsorbed on a positively charged SLB only when the molar percentage of DOTAP in the preparation mixture was 50%. In this case, the difference between the ζ measured for the lipid vesicles that were used for the SLB formation and the ζ measured for the NRs was │Δζ(POPC/DOTAP 50:50 – NR-)│≈ 73±2 mV (see Figure S3A). In all other cases, no interaction was observed. NR+, however, was

ACS Paragon Plus Environment

ACS Chemical Neuroscience 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

Figure 1. Mass adsorbed per unit surface area (Δm) on SLBs due to the interaction of NRs with lipids. Δm is plotted versus the molar percentage of charged lipid in the different lipid mixtures employed. The corresponding ζ is plotted as the top x-axis for clarity. Red circles ⧳ represent the mass of positively charged NRs (NR+ ζ= +11 mV) with SD, the blue squares ⧯ represent the same for negatively charged ones (NR- ζ= -24 mV). The sigmoid curves (dotted lines) have been plotted so that the correlation between the data points can be easily identified. adsorbed on the anionic SLB in the presence of a POPS molar content as low as 10%, showing a Δζ of │ Δζ(POPC/POPS 90:10 – NR+) │ = 65±2 mV. Increasing the molar percentage of the anionic lipid did not increase the adsorbed NR+ mass (ca. 400 ng/cm2), which apparently reached a saturation value (Figure S3B). Using a SLB with a molar content of POPS of 25%, after the NR+ adsorption phase (corresponding to a frequency decrease) that lasted ca. 30 min, an increase in the frequency shift was registered, while the dissipation value remained constant. We interpreted this event as the removal of material from the substrate surface (Figure S4). We highlight here that there was no adsorption of either NR- or NR+ (Δm ≈ 0) in the zwitterionic lipid bilayer (0% molar of charged lipids, pure POPC, ζ = -23±1 mV): the absolute value of the difference between the ζ of the lipids and the ζ of the NRs (│Δζ(POPC – NR)│) is 1±1mV for NRs- and -34±1 mV for NR+. A small decrease in the frequency was also observed for the NR+ that interacted with the positive SLB (both 25% and 50% of DOTAP). We interpreted this as the adsorption of NR+ on the bare SiO2 surface, which is in the substrate areas that are not covered by the bilayer, as was previously mentioned. Increasing the DOTAP

ACS Paragon Plus Environment

Page 6 of 30

Page 7 of 30 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

ACS Chemical Neuroscience

molar percentage in the vesicles, the SLB coverage on the sensor’s surface was so small that both NR+ and NR- interacted producing a similar frequency down-shift (data not shown): the NR+ was adsorbed on bare SiO2 and the NR- was adsorbed by the positive membrane. Summarizing the results, we observed an interaction between NRs and fluid state SLBs: not only does the charge of the two objects have to be opposite in sign but a certain │Δζ(SLB – NR)│value has to be overcome, e.g., ≈ 70 mV. For lower differences in ζ, no interaction occurs in the cases in which an electrostatic interaction is to be expected (e.g. in the case of POPC/DOTAP 75:25 mol:mol with NR-, (│Δζ(SLB – NR)│= 37±1 mV and of POPC with NR- (│Δζ(SLB – NR)│=34±1 mV ). QCM-D measurements: phase separated bilayers. Even the presence of a small amount of saturated lipid in the mixture drastically changed the way in which NRs and lipids interacted. The mixtures that interact with the oppositely charged NRs - as in the unsaturated mixture study - were modified. 5, 10 and 20% mole of the unsaturated zwitterionic lipid (i.e., POPC) was substituted with a saturated chains lipid with the same hydrocarbon chain length (i.e., DPPC), and the charged lipid molar fraction was kept constant (10% molar POPS or 50% molar DOTAP for anionic and cationic SLB). DLS measurements showed the same ζ values for unsaturated and saturated lipid mixture vesicles at equal anionic/cationic lipid contents (Table 1). However, the interaction with the NRs was different in each of the two cases. Figure 2 shows the mass of NRs that was adsorbed on the anionic and cationic SLBs, interacting with NR+ and NR- respectively, as a function of the molar percentage of DPPC. In both cases, the amount of NRs that were adsorbed significantly reduced, and quite in the same amount, when DPPC 5 and a 10% molar content were present. No adsorption was registered for a 20% molar content of DPPC. All these data were collected at 22 °C, a temperature at which DPPC is in a solid phase but segregates into a gel phase domain in the fluid lipid matrix that is constituted by POPC and POPS or DOTAP37. These results indicate that not only the charge of the system drives the interaction between the lipid membrane and the NRs, but the membrane composition also plays an important role, since even a small quantity of saturated chain lipids can modify the mechanical properties of the fluid state bilayer creating stiffer zones made of lipids with more compacted hydrophobic chains mimicking the so called lipid rafts in the natural membrane38, thus inhibiting the interaction with the NRs. π–A isotherms. Starting with the QCM experiments results, π–A air-water isotherms were performed on the same lipid mixtures that showed a stable interaction with NRs of opposite sign, i.e. POPC/POPS 90:10 and POPC/DOTAP 50:50 (mol:mol ratios) mixtures with NR+ and NR-, respectively. Mixtures containing 10% molar DPPC

ACS Paragon Plus Environment

ACS Chemical Neuroscience 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

Figure 2. Normalized mass of NRs adsorbed on SLBs (Δm) versus the molar percentage of the saturated DPPC that is present in the lipid mixture. Anionic lipid mixtures (POPC/POPS/DPPC 90-x:10:x) were tested with NR+ (red line), cationic lipid mixtures (POPC/DOTAP/DPPC 50-x:50:x) with NR- (blue line). were also investigated. In Figure 3, the isotherms obtained from the NR subphase are reported and compared to the isotherms in ultrapure water. Four different subsequent isotherms were carried out for each sample. POPC/POPS 90:10 mol:mol had a larger lift-off area per molecule (Amol) than POPC/DOTAP 50:50 mol:mol (120 vs 70 Å2/molecule). The collapse pressures (πcol) were 47 and 49 mN/m, respectively. In the presence of DPPC, the lift-off Amol was similar for both the anionic and the cationic mixtures (≈ 90 Å2/molecule), whereas the πcol were 42 mN/m for the anionic and 35 mN/m for the cationic mixture. In the case of POPC/POPS mixtures, the presence of NR+ in the water subphase led to a left-shift of the isotherm, which was more pronounced for each subsequent compression. The lift-off Amol progressively changed from 120 to 110 Å2/molecule during the subsequent compression cycles, indicating that NR+ strongly interacts with the lipid head-groups. NR+ alone did not cause any increase in the surface pressure (as NR-), as is shown in Figure S5. The leftshifted isotherms and the lower collapse π with respect to the previous curves might indicate a complexation of the serine head-groups with the amine functionalized NRs and a progressive removal of lipid molecules from the surface, as is shown in the sketch (Figure 3A). As in the case of the QCM-D experiments, a strong interaction between NR+ and POPC/POPS 90:10 mol:mol occurred. However, in the case of SLB, the destabilization of the membrane was observed only at a higher POPS content, and this is

ACS Paragon Plus Environment

Page 8 of 30

Page 9 of 30 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

ACS Chemical Neuroscience

associated to the lipid removal from the sensor surface. This could be an indication that 1) the ζ of the SLB system is not exactly the same as that of the monolayer due to possible asymmetries34; 2) the interaction in the QCM-D experiment occurs after the SLB has already formed. The molecular packing could hinder the interaction, considering that the typical surface pressure of a lipid bilayer is πbilayer ≈ 30 mN/m39, and the initial pressure is 0 mN/m in the case of the LMs. For the anionic mixture containing 10% molar of DPPC (Figure 3C), the left-shift of the isotherm is very small but still appreciable, which is in agreement with the QCM-D measurements. In this case, less NR+ adsorbed on the bilayer: NRs are attracted by the oppositely charged lipids towards the surface but the lateral packing between the lipid molecules prevents the insertion of polymer coated NRs between the lipids. For cationic monolayers, the interaction with NR- causes a progressive right-shift of the isotherms upon subsequent compressions, suggesting that the NR- (or of their polymer coating) interacts with the lipids and causes an increase in the area per molecule (Figure 3B) (the lift-off area increases from 70 to 95 Å2/molecule during four subsequent compressions). As in the previous case, this effect is strongly hindered by the presence of DPPC (Figure 3D), and only a negligible right-shift of the isotherm is visible at low surface pressure. As in the QCM-D measurements, the monolayer isotherms were greatly modified in the presence of PEG-amine functionalized NRs in the case of unsaturated lipid mixtures, while very small changes in the mixtures containing DPPC were detected. Compressibility modulus evaluation. The most relevant changes in the investigated parameter took place in the unsaturated lipid mixtures: in the presence of NRs, the monolayer becomes softer for POPC/POPS 90:10 mol:mol LM (a decrease in the compressibility modulus (Cs-1) can be seen in Figure 4A) but it becomes more rigid for POPC/DOTAP 50:50 mol:mol (an increase in the Cs-1 can be seen Figure 4B). The differences in the Cs-1 in the presence of DPPC were practically negligible, as is to be expected (see Figure 4C and D). The softening of the 10% molar POPS mixture confirmed the presence of fewer molecules at the interface. The interaction takes place at a low π and it is represented by a change in the curve trend in the window between 0 and 9 mN/m: instead of the Cs-1 growing in a monotone manner, as it does in the ultrapure water subphase case (in black), it shows a local maximum at 4mN/m followed by a minimum at 9mN/m in the NR+ dispersion subphase (in red). The presence of a local minimum at 9 mN/m is an indication of a possible phase transition in the system; at higher  the trend was similar to the curve obtained on ultrapure water, but at a slightly lower values. As a result of the decreasing number of molecules at the interface, it was not possible to reach a collapse value in the NR+ dispersion subphase.

ACS Paragon Plus Environment

ACS Chemical Neuroscience 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

Figure 3. π –A isotherms of lipid monolayers interacting with NRs. The lipids we used are POPC (zwitterionic, unsaturated), DPPC (zwitterionic, saturated), DOTAP (cationic, unsaturated) and POPS (anionic. unsaturated). A) POPC/POPS 90:10 mol:mol B) POPC/DOTAP 50:50 mol:mol C) POPC/POPS/DPPC 80:10:10 mol:mol:mol and D) POPC/DOTAP/DPPC 40:50:10 mol:mol:mol. Black: lipid mixtures in an ultrapure water subphase. Red, green, blue, cyan represent the first, second, third and fourth compressions, respectively, in the NR dispersion subphase. Adding DPPC to the mixture, the local minimum at 9 mN/m was less pronounced and, at a higher π, the curve was identical to the curve on the ultrapure pure water subphase. This result agrees with the QCM-D measurements, which demonstrated that the presence of DPPC hindered the interaction between the NRs and the lipids. Regarding the POPC/DOTAP 50:50 mol:mol mixture, no considerable differences were registered between the curve obtained in the presence or in the absence of NR- in the subphase when a low surface pressure range (π = 0-13 mN/m) was used (there was a decrease in Cs-1 of ≈ 5 mN/m). However, when higher pressures were reached, Cs-1 in presence of NRs is higher, indicating an increase in the rigidity of the monolayer due to

ACS Paragon Plus Environment

Page 10 of 30

Page 11 of 30 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

ACS Chemical Neuroscience

the insertion of NR- functionalization molecules between the lipids. When DPPC was added to the mixture, values at very low pressures (π = 0-4.5 mN/m) and high pressures (π = 16.5-33 mN/m) remained unchanged. There were small differences in the middle range of pressures of about 8 mN/m, which is comparable with the decrease registered in the unsaturated mixture.

Figure 4. Compressibility modulus Cs-1 versus the surface pressure π. Black: LMs in ultrapure water subphase. Red: LMs in NR dispersion subphase. A) and C) are anionic mixtures with NR+, C) and D) are cationic mixtures with NR-. All lipid compositions are in molar ratio. DLVO simulations. The interaction between charged NPs and lipid bilayers of various compositions was simulated in the DLVO approximation, and an interaction between a flat plate and a particle was considered40. DLVO theory is routinely applied to predict the colloidal stability of NP dispersions by summing the electrical double layer of two surfaces and the Van der Waals interaction. Among the different DLVO approximation methods (constant potential, constant charge, linear superimposition approximation), we chose to use a linearized Poisson-Boltzmann approximation at a

ACS Paragon Plus Environment

ACS Chemical Neuroscience 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

constant potential or constant charge between the flat plate and the particle. The corresponding equations are reported in the methods section. The results are displayed in Figure 5. The interaction energy profiles showed that NR+ (Figure 5A) are attracted by negatively charged bilayers (i.e., POPC/POPS mixtures). The interaction is strongest for POPC/POPS 75:25 mol:mol, which is the membrane with the most negative ζ. Repulsion is predicted between NR+ and POPC/DOTAP. These results are to be expected, and they are in line with the experimental findings (i.e. with the QCM-D results which showed NR accumulation and interaction with LM). However, the model also predicts an attraction to a POPC bilayer, which is different from what was experimentally observed. Therefore, in this case, the repulsive interaction that occurs is higher than the one that was predicted by the model. The same holds for the mirror system, in which NR- interacts with positive membranes (Figure 5B): for both the highly positive lipid system (POPC/DOTAP 50:50 mol:mol) and the most negative lipid system, attraction and repulsion were predicted, respectively, but the model failed to predict the experimental results for the intermediate case (i.e., POPC/DOTAP 75:25 mol:mol). When the potential barrier is low, the data is in accordance with the predictions only using an approximation different from the constant potential one, i.e. in the model of constant charge approximation, which however overestimates the potential barrier40 (Figure S6 and S7). The results indicate that an additional repulsive potential such as a steric interaction and/or hydration force41 plays a role in the short-range NR/lipid interaction. The non-DLVO interactions are enhanced when saturated chains are present in the membranes and in the associated DPPC gel domains, as is clear from the QCM-D and Langmuir results. It should be noted that the ζ of the lipids applied in the model calculation is measured from the vesicle dispersion therefore it might not represent the actual potential of the bilayer, for instance, due to pure geometrical reasons or to an internal rearrangement of the charged lipids, resulting in an asymmetric bilayer. Further equations (1) and (2) hold for spherical particles. Summary and Conclusions The interaction between negative (NR-, ζ=-24 mV) and positive (NR+, ζ=+11 mV) amphiphilic polymer coated CdSe/CdS nanorods and lipid model membranes made of POPC, POPS, DOTAP and DPPC mixed at different molar ratios was studied systematically as a function of the membrane charge and gel phase presence. The majority of literature on this topic 16-18, 42 addresses the interaction of spherical nanoparticles with cellular or model membranes; our investigation offers information on the interplay between nanorods and membranes, of which there has been little to date.

ACS Paragon Plus Environment

Page 12 of 30

Page 13 of 30 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

ACS Chemical Neuroscience

Figure 5. Simulations of the lipid bilayer – nanoparticle interaction in the DLVO approximation at constant potential. All lipids that we used were unsaturated: POPC (zwitterionic), POPS (anionic), DOTAP (cationic). A) The interaction of NR+ (ζ=+11 mV) with the membrane of different compositions is attractive for POPC/POPS 75:35, POPC/POPS 90:10 (mol:mol ratios) and for the zwitterionic POPC; repulsion is obtained for a positive membrane POPC/DOTAP 75:25 mol:mol B) The interaction of NR- (ζ=-24 mV) is attractive in the case of lipid bilayers containing DOTAP and repulsive for a zwitterionic POPC membrane. Supported lipid bilayers and lipid monolayers, either in a fluid phase or in the presence of gel phase domains, were considered as models. The results have indicated that there is a fine interplay between the properties of the PEG-amine complex that was used for the functionalization of the NRs and the lipids.

ACS Paragon Plus Environment

ACS Chemical Neuroscience 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





In fluid phase bilayers the difference in the zeta potential between the bilayer and the nanoparticle drives the interaction. It occurs only when a difference of at least 70 mV exists (POPC/POPS 90:10 mol:mol with positive NRs and POPC/DOTAP 50:50 mol:mol with negative ones). Svedhem et al. 42 performed a similar study on fluid state lipid bilayers with different compositions and superficial charges, observing that the interaction of un-functionalized spherical TiO2 nanoparticles is modulated by the charge of the lipid system. Comparing the results with those obtained with neuronal cultures7, we observed that also in that case (in which only highly negatively charged NPs were interacting with the membrane) the interaction was allowed beyond a zeta potential threshold. Neurons are known to have action potentials, varying their membrane potential from highly negative to positive values. Probably, during this dynamic change and when the membrane potential turns positive, the NPs interact. Here, we calculated this threshold using a model system. According to the DLVO simulations, however, adsorption is also expected for lower values. Therefore, non-DLVO terms play a role in the interaction, which is not purely driven by electrostatic and van der Waals forces. This is further supported by the observation that a molar content of saturated lipids as small as 5% severely decreases the amount of NRs that are adsorbed on the bilayer, although the overall charge is constant. This indicates that the interaction is not exclusively regulated by lipid head-group charges. Granick et al. 43 explained the correlation between the presence of nanoparticles and the change in headgroups in lipid vesicles, inducing gelation in fluid membranes or fluidizing full gel state ones. In our case, the steric hindrance, which occurred due to the nano and micro gel phase domains, might inhibit the interaction, thus preventing adsorption and state modifications. Again, we can speculate on our results as an preliminary explanation of the non-interaction between NPs and glial cells, that can be modeled by our static negatively charged lipid bilayer. In the presence of gel phase domains low or even no interaction was registered, as in the glial cells7. This can explain that a situation of dynamic potential changes is necessary to induce the adsorption on NPs onto the membrane, as it happens during the action potential of the neurons. The lipid monolayer investigations indicate that either a removal of lipids or an accumulation of NRs at the interface may occur, depending on the PEGamine/lipid group interaction. This finding highlights the importance of the polymer coating for fine tuning the interaction, which is in line with literature. Barros-Timmons at al.44 investigated the behavior of lipid monolayers with coated NPs and with the coating alone. They showed that when there is an

ACS Paragon Plus Environment

Page 14 of 30

Page 15 of 30 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

ACS Chemical Neuroscience

interaction with an oppositely charged NP, the functionalization alone would induce the same reaction. In our experiments, removal and accumulation events were confirmed by an evaluation of the monolayer’s compressibility modulus, which showed variations in the rigidity of the system due to the interaction with the NRs, i.e. softening in the anionic monolayer interacting with positive NR and stiffening for cationic monolayers with negative particles. Furthermore, we confirmed the inhibition of these phenomena due to the presence of saturated chain lipids in the mixture. This suggests that other properties, such as the rigidity of the lipid systems, are also important in preventing/allowing NRmembrane interaction. The results of the DLVO simulation fully support this fact, since they demonstrate that the interaction is not explained by only considering the electrostatic and van der Waals forces. Therefore, in order to predict and tune the application of NPs, a fine tuning of the membrane/NP interface is necessary in terms of the overall charge and mechanical properties. Methods Materials. CdSe/CdS NRs with a length of 30 nm and a diameter of 5 nm were synthetized in-house following a published protocol.45 Poly(ethylene glycol) bis(amine) (NH2-PEG-NH2, MW = 2000) was purchased from Rapp Polymere (Tübingen, Germany). Phosphate Buffered Saline (PBS) dry powder, chloroform (≥99.5%), methanol (≥99.8%), Sodium Docecyl Sulfate (SDS), poly(maleic anhydride-alt-1 octadecene) (MW = 30000), N,N-dimethylethylendiamine (DMEDA), N-(3-(dimethylamino)propyl)-N′ethylcarbodiimide hydrochloride (EDC) and N-[1-(2,3-Dioleoyloxy)propyl]-N,N,Ntrimethylammonium chloride (DOTAP) were purchased from Sigma-Aldrich® (Saint Louis, MO, U.S.A.). 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC), 1palmitoyl-2-oleoyl-sn-glycero-3-phospho-L-serine (sodium salt) (POPS) and 1,2dipalmitoyl-sn-glycero-3-phosphocholine (DPPC) were purchased from Avanti Polar Lipids, Inc. (Alabaster, AL, U.S.A). The lipids’ molecular structure can be found in Figure S8 in the Supporting Information. Ultrapure water (18.2 MΩ) was obtained using a Millipore Milli-Q purification system. CdSe/CdS NRs’ surface functionalization. The NRs were water solubilized with poly(maleic anhydride-alt-1 octadecene), using a previously reported polymer coating procedure.46 They were surface-functionalized with different amounts of NH2-PEG-NH2 and a tertiary amine (DMEDA), using an EDC cross-linking reaction scheme47 in order to tune the superficial charge. The sample was prepared in a PBS 1x solution, while the final dilution was made in ultrapure water. The NRs’ inorganic and hydrodynamic sizes were determined by a JEM 1011 Transmission Electron Microscopy (TEM, see Figure S9,

ACS Paragon Plus Environment

ACS Chemical Neuroscience 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

JEOL USA, Inc., Paebody, MA, U.S.A) and by a Dynamic Light Scattering (DLS) set up. The zeta potential (ζ) was measured with a Malvern Instruments Zeta Sizer Nano ZS system (Malvern, U.K.). The concentration of the NRs was determined by elemental analysis through measuring the amount of cadmium in the nanostructures, as in ref 45. Large Unilamellar Vesicle (LUV) preparation. Powdered POPC, POPS, DOTAP and DPPC were dissolved in 2:1 vol:vol chloroform/methanol and stored at –20 °C. Proper volumes of the dissolved lipids were mixed in a round bottom glass vial to obtain anionic (-), cationic (+) or zwitterionic mixtures (+/-). In detail we used: POPC (+/-), POPC/POPS 90:10 (-), POPC/POPS 75:25 (-), POPC/DOTAP 75:25 (+), POPC/DOTAP 50:50 (+) (all molar ratios). These mixtures contain only lipid with unsaturated alkyl chains. In other mixtures, a molar fraction of DPPC with fully saturated alkyl chains was introduced so that we could obtain POPC/POPS/DPPC 85:10:5 (-), POPC/POPS/DPPC 80:10:10 (-), POPC/POPS/DPPC 70:10:20 (-), POPC/DOTAP/DPPC 45:50:5 (+), POPC/DOTAP/DPPC 40:50:10 (+) and POPC/DOTAP/DPPC 30:50:20 (+) molar ratio mixtures. The solvent was left to evaporate under N2 flow; the vials were then left under vacuum overnight to completely eliminate all of the residual solvent. Vials were weighed before and after solvent evaporation in order to know the exact total amount of lipids (in mg). The lipids were then re-suspended in ultrapure water for mixtures containing DOTAP, and in PBS 1x for the rest in order to obtain 1 mg/mL lipid dispersions, an amount that is easy to handle for the vesicle extrusion. Lipids were left to hydrate for 20 min then shortly vortexed before being extruded at least 11 times with an Avanti Mini Extruder using 100 nm filters (Avanti Polar Lipids, Inc.) to form LUV dispersions with a total amount of 1mg/mL of lipid. Extrusion was performed at room temperature for unsaturated lipid mixtures and at 60 °C for mixtures containing DPPC. Each vesicle dispersion then underwent a size and ζ characterization in a 1.4 mM PBS solution (Table 1). Supported lipid bilayers and QCM-D measurements. SLBs were obtained on quartz crystals sensors covered with SiO2 (nominal resonance frequency of 5MHz) in a KSV QCM-Z500 QCM-D tool (Biolin Scientific, Gothenburg, Sweden). Every crystal was cleaned by sonication in a 2% SDS solution for 20 min, then rinsed three times with ultrapure water and dried under N2 flow. To remove any organic contaminants from the surface, the sensors were placed in a UV/Ozone ProCleaner™ chamber (BioForce Nanosciences, Inc., U.S.A.) for 10 min before use. To prepare the SLB, ultrapure water or PBS 1x buffer was injected into the QCM-D measurement chamber until stability was reached. 2mL of the lipid vesicle solution was then added at a concentration of 0.1 mg/mL to form a stable SLB. The temperature for SLB formation was 22 °C for unsaturated mixtures and 60 °C for mixtures containing DPPC. After it reached stability, the temperature of the SLBs containing saturated lipids was set to 22 °C. After SLB

ACS Paragon Plus Environment

Page 16 of 30

Page 17 of 30 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

ACS Chemical Neuroscience

formation, unattached vesicles were rinsed in either ultrapure water or PBS 1x. Frequency and dissipation change measurements were used to confirm the existence of a stable SLB for further investigation with NRs35, 48, 49. NRs interact with SLBs if they are adsorbed on their surface, i.e. if a negative shift in the resonance frequency of the QCM sensor is registered. In the case of POPC/POPS 75:25 mol:mol, POPC/POPS 90:10 mol:mol and POPC, whose vesicles were prepared in PBS 1x buffer, a further rinse with ultrapure water was necessary. After stabilization and a rinsing step, a solution of NRs in ultrapure water (at a concentration of 1 nM for NR- and 5 nM for NR+) was injected into the measurement chamber. After 30 min, a final rinsing step with ultrapure water was carried out in order to remove any non-interactive NRs. Changes in the dissipation value ΔD gave us information about the viscoelastic properties of the adsorbed layer, which is considered rigid for a small value of ΔD. The Sauerbrey equation36 is used for the quantification of the adsorbed mass, which is calculated as a function of the sensor’s properties as follows: Δm = - C

Δf 𝑛

in which Δm is the mass per unit area that is adsorbed on the sensor, C is the coefficient that describes the sensitivity of the instrument to changes in mass, Δf = f-f0 is the shift in frequency and n is the overtone number. It should be noted that the coefficient depends on the crystal properties which, in this case, are C ≈ 17.7 ng∙cm-2∙Hz-1. For every measurement, the 7th overtone was considered, since it is the most stable among the 11 overtones that were investigated. Figure S10 shows the frequency shifts of all the measured overtones for positive and negative NRs adsorbing on POPC/POPS and on POPC/DOTAP SLBs, respectively. Lipid monolayer and surface pressure area isotherms (π-A). 0.3 mg/mL of lipid solutions were prepared in a 2:1 chloroform/methanol mixture. Every solution was stored in the fridge at 4 °C and left at room temperature for a few hours before use. To minimize the risk of changes in the lipid concentration due to the fast evaporation of the solvent, vials closed with Mininert® Valves 15 mm (by Supelco, Bellefonte, PA, U.S.A.) were used. 20 µL of a lipid solution was spread with a Hamilton microsyringe (10 µL, by Sigma-Aldrich®) at the air-water interface of an ultrapure water subphase or a NR dispersion subphase (the NRs were dispersed in ultrapure water at a concentration of 0.5 nM) in a polytetrafluoroethylene (PTFE) custom-made Langmuir trough equipped with two PTFE barriers for symmetrical compression, with an initial open area of 7500 mm2. The trough was fabricated in order to minimize the subphase volume, creating a central well that was 6 mm deep and two lateral wells that were 2 mm deep (Figure S11). After spreading the solution, the solvent was left to evaporate for ca. 15 min. The

ACS Paragon Plus Environment

ACS Chemical Neuroscience 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 30

subphase temperature was 20 ± 2 °C. Compression π-A isotherms were then performed at a compression rate of 10 mm/min, using a platinum Wilhelmy plate to detect the instantaneous surface pressure π, which is equal to π = γ0 - γsurfactant in which γ0 is the subphase surface tension (which is 72 mN/m for water) and γsurfactant is the surface tension of the subphase in the presence of the monolayer50. Compressibility modulus. The evaluation of the compressibility modulus of the lipid monolayer Cs-1 was carried out using the first derivative of π on the molecular area that was occupied by the lipids (A). It is be defined as 50, 51 : ∂π

Cs-1= ―A∂A Cs-1 gives us information about the rigidity of the monolayer. Comparing the results that were obtained with and without the presence of saturated lipids in the mixtures can help us clarify the mechanisms of interaction between the NRs and the lipid models. DLVO model. We used a particle-plate interaction model to describe the interaction between a NR particle and a lipid bilayer, following the approach described by Mikelonis et al40. The electrostatic interactions, which can be described by a screened Coulomb potential, dominate at large separation distance and are sensitive to the electrolyte concentration; the van der Waals interaction dominates when the gap between the surfaces is small. The size of the particle was set to the average hydrodynamic size of the NRs. ζ values for the supported lipid membrane (which were estimated by the corresponding values that were obtained for liposomes in bulk, see Table 1) and NRs (measured, see Table 2) were used. The electrostatic interaction was calculated at a constant potential or constant charge approximation according to the following equations:

{

[

{

[

V𝑅,𝑝𝑓 = 𝜋𝜀0𝜀𝑟𝑎𝑝 2Ψ𝑑𝑓Ψ𝑑𝑝ln

V𝑅,𝑝𝑓 = 𝜋𝜀0𝜀𝑟𝑎𝑝 2Ψ𝑑𝑓Ψ𝑑𝑝ln

Ψ=

4𝑘𝐵𝑇 𝑧𝑒

1 + 𝑒( ― 𝜅𝑠) 1―𝑒

( ― 𝜅𝑠)

1 + 𝑒( ― 𝜅𝑠) 1 ― 𝑒( ― 𝜅𝑠)

}

]

― (Ψ𝑑𝑓2 + Ψ𝑑𝑝2)ln [1 ― 𝑒( ― 2𝜅𝑠)] (1)

]

+ (Ψ𝑑𝑓2 + Ψ𝑑𝑝2)ln [1 ― 𝑒( ― 2𝜅𝑠)] (2)

( ( )

tanh ―1 tanh

}

)

𝑧𝑒𝜉 ⨯ 𝑒 ―𝜅𝑑 4𝑘𝐵𝑇

ACS Paragon Plus Environment

(3)

Page 19 of 30 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

ACS Chemical Neuroscience

𝐴𝑎𝑝

(

14𝑠 V𝐴,𝑝𝑓 = 1+ 6𝑠 𝜆

)

―1

(4)

Equation (1) holds in the case of constant potential approximation. Equation (2) was used to model the constant charge approximation40. Equation (3) was used to convert the measured ζ into an estimated surface potential using the Gouy-Chapman model. In each of the formulae, ε represents permittivity, ap represents the particle size, s represents the separation distance, Ψ represents the surface potential, κ represents the Debye length, and n represents the number concentration. The van der Waals interaction dominates when the gap between the surfaces is small. It is also rather insensitive to the concentration of the electrolytes, and it is obtained by using the Hamaker theory (see equation (4), in which s represents the separation distance, A is the Hamaker constant, and λ is the characteristic wavelength of the interaction). Typically, a value of 100 nm is used for λ52. The value of the Hamaker constant is a source of uncertainty in the van der Waals energy calculation. In literature, the Hamaker constant of the core material is often used, regardless of the outer stabilizing agent53, 54. On the other hand, it is recognized that steric and electrosterically stabilized particles require a different effective Hamaker constant. Here, the Hamaker constant was set to 1.62 10-20 J, and was calculated as (A11A22)1/2 in which A11 is the Hamaker constant value for the PEG-H20 interaction (7.2 10-20 J)55 and A22 is the Hamaker constant for the lipid bilayer interaction (3.65 10-21 J)56. All the calculations have been performed using the Igor Pro, Version 6.1, Wavemetrics platform. Supporting Information A TEM image of NRs, the chemical structures of the employed lipids, the QCM-D curves of the formation of POPC/POPS 90:10 and POPC/DOTAP 50:50 (mol:mol ratios) SLBs, the adsorption of the NRs on bare substrates and bilayers, NR dispersions in ultrapure water π-A isotherms, the frequency shifts of all the measured overtones by QCM-D for positive and negative NRs adsorbing on POPC/POPS and on POPC/DOTAP SLBs, respectively and a comparison between the constant charge and the constant potential of the DLVO models can be all be found in the Supporting Information. Abbreviations Nanorods (NRs); Cadmium Selenium/Cadmium Sulfur (CdSe/CdS); zeta potential (ζ); supported lipid bilayers (SLBs); lipid monolayers (LMs); Derjaguin-Landau-VerweyOverbeek (DLVO); 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC); 1-

ACS Paragon Plus Environment

ACS Chemical Neuroscience 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

palmitoyl-2-oleoyl-sn-glycero-3-phospho-L-serine (sodium salt) (POPS); 1,2-dipalmitoylsn-glycero-3-phosphocholine (DPPC); N-[1-(2,3-Dioleoyloxy)propyl]-N,N,Ntrimethylammonium chloride (DOTAP); large unilamellar vesicle (LUV); dynamic light scattering (DLS); phosphate buffered saline (PBS); sodium docecyl sulfate (SDS); poly(ethylene glycol) bis(amine) (PEG); N,N-dimethylethylendiamine (DMEDA); N-(3(dimethylamino)propyl)-N′-ethylcarbodiimide hydrochloride (EDC); negatively charged NR (NR-); positively charged NR (NR+); surface pressure of the lipid monolayer (π); compressibility modulus of the lipid monolayer (Cs-1). Author Information Corresponding Authors: Barbara Salis, Silvia Dante *E-mail: [email protected], [email protected] Authors Contribution S.D., T.P. and A.D. conceived the experiments, B.S. and G.P. performed the surface functionalization of NRs, B.S. performed QCM-D and pressure-area isotherms experiments, S.D. performed the DLVO simulations, ,. B.S. and S.D. wrote the manuscript and all the authors revised it. Funding Sources This research is partially supported by the Associazione Italiana per la Ricerca sul Cancro, identification N.: IG 20790 to T.P. Acknowledgments We thank Riccardo Carzino for his help with designing the custom-made Langmuir trough, Francesco De Donato for synthesizing semi-conductor NRs and Marco Salerno for critically reading the manuscript. References 1. Gupta, A. K., and Gupta, M. (2005) Synthesis and surface engineering of iron oxide nanoparticles for biomedical applications, Biomaterials 26, 3995-4021. 2. Jinhao Gao, H. G. a. B. X. (2009) Multifunctional Magnetic Nanoparticles: Design, Synthesis, and Biomedical Applications, Acc. Chem. Res. 42 (8). 3. Parveen, S., Misra, R., and Sahoo, S. K. (2012) Nanoparticles: a boon to drug delivery, therapeutics, diagnostics and imaging, Nanomedicine 8, 147-166. 4. Richard, S., Boucher, M., Lalatonne, Y., Mériaux, S., and Motte, L. (2017) Iron oxide nanoparticle surface decorated with cRGD peptides for magnetic resonance imaging of brain tumors, Biochim. Biophys. Acta, Gen. Subj. 1861, 1515-1520. 5. Bothun, G. D., Ganji, N., Khan, I. A., Xi, A., and Bobba, C. (2017) Anionic and Cationic Silver Nanoparticle Binding Restructures Net-Anionic PC/PG Monolayers with Saturated or Unsaturated Lipids, Langmuir 33, 353–360.

ACS Paragon Plus Environment

Page 20 of 30

Page 21 of 30 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

ACS Chemical Neuroscience

6. 7.

8. 9. 10. 11.

12. 13. 14.

15. 16.

17. 18.

Malvindi, M. A., Carbone, L., Quarta, A., Tino, A., Manna, L., Pellegrino, T., and Tortiglione, C. (2008) Rod-Shaped Nanocrystals Elicit Neuronal Activity In Vivo, Small 4, 1747–1755. Dante, S., Petrelli, A., Petrini, E. M., Marotta, R., Maccione, A., Alabastri, A., Quarta, A., De Donato, F., Ravasenga, T., Sathya, A., Cingolani, R., Proietti Zaccaria, R., Berdondini, L., Barberis, A., and Pellegrino, T. (2017) Selective Targeting of Neurons with Inorganic Nanoparticles: Revealing the Crucial Role of Nanoparticle Surface Charge, ACS Nano 11, 6630-6640. Xu, C., Tung, G. A., and Sun, S. (2008) Size and Concentration Effect of Gold Nanoparticles on X-ray Attenuation As Measured on Computed Tomography, Chem. Mater. 20, 4167-4169. Bailey, C. M., Kamaloo, E., Waterman, K. L., Wang, K. F., Nagarajan, R., and Camesano, T. A. (2015) Size dependence of gold nanoparticle interactions with a supported lipid bilayer: A QCM-D study, Biophys. Chem. 203-204, 51-61. Tortiglione, C., Quarta, A., Malvindi, M. A., Tino, A., and Pellegrino, T. (2009) Fluorescent Nanocrystals Reveal Regulated Portals of Entry into and Between the Cells of Hydra, PLoS One 4, e7698. Czogalla, A., Grzybek, M., Jones, W., and Coskun, Ü. (2014) Validity and applicability of membrane model systems for studying interactions of peripheral membrane proteins with lipids, Biochim. Biophys. Acta, Mol. Cell Biol. Lipids 1841, 1049-1059. Chan, Y.-H. M., and Boxer, S. G. (2007) Model Membrane Systems and Their Applications, Curr. Opin. Chem. Biol. 11, 581-587. Braydich-Stolle, L., Hussain, S., Schlager, J. J., and Hofmann, M.-C. (2005) In Vitro Cytotoxicity of Nanoparticles in Mammalian Germline Stem Cells, Toxicol. Sci. 88, 412-419. Negoda, A., Liu, Y., Hou, W.-C., Corredor, C., Moghadam, B. Y., Musolff, C., Li, L., Walker, W., Westerhoff, P., Mason, A. J., Duxbury, P., Posner, J. D., and Worden, R. M. (2013) Engineered nanomaterial interactions with bilayer lipid membranes: screening platforms to assess nanoparticle toxicity, Int. J. Biomed. Nanosci. Nanotechnol. 3, 52-83. Nel, A., Xia, T., Mädler, L., and Li, N. (2006) Toxic Potential of Materials at the Nanolevel, Science 311, 622-627. Atukorale, P. U., Guven, Z. P., Bekdemir, A., Carney, R. P., Van Lehn, R. C., Yun, D. S., Jacob Silva, P. H., Demurtas, D., Yang, Y.-S., Alexander-Katz, A., Stellacci, F., and Irvine, D. J. (2018) Structure–Property Relationships of Amphiphilic Nanoparticles That Penetrate or Fuse Lipid Membranes, Bioconjugate Chem. 29, 1131-1140. Rossi, G., and Monticelli, L. (2016) Gold nanoparticles in model biological membranes: A computational perspective, Biochim. Biophys. Acta, Biomembr. 1858, 2380-2389. Chen, H., Dorrigan, A., Saad, S., Hare, D. J., Cortie, M. B., and Valenzuela, S. M. (2013) In Vivo Study of Spherical Gold Nanoparticles: Inflammatory Effects and Distribution in Mice, PLoS One 8, e58208.

ACS Paragon Plus Environment

ACS Chemical Neuroscience 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

19. 20. 21. 22. 23. 24. 25. 26. 27. 28.

29. 30. 31. 32. 33.

Lins, P. M. P., Marangoni, V. S., Uehara, T. M., Miranda, P. B., Zucolotto, V., and Cancino-Bernardi, J. (2017) Differences in the Aspect Ratio of Gold Nanorods that Induce Defects in Cell Membrane Models, Langmuir 33, 14286-14294. Di Silvio, D., Maccarini, M., Parker, R., Mackie, A., Fragneto, G., and Baldelli Bombelli, F. (2017) The effect of the protein corona on the interaction between nanoparticles and lipid bilayers, J. Colloid Interface Sci. 504, 741-750. Montis, C., Generini, V., Boccalini, G., Bergese, P., Bani, D., and Berti, D. (2018) Model lipid bilayers mimic non-specific interactions of gold nanoparticles with macrophage plasma membranes, J. Colloid Interface Sci. 516, 284-294. Volinsky, R., Kolusheva, S., Berman, A., and Jelinek, R. (2006) Investigations of antimicrobial peptides in planar film systems, Biochim. Biophys. Acta, Biomembr. 1758, 1393-1407. Dietrich, C., Bagatolli, L. A., Volovyk, Z. N., Thompson, N. L., Levi, M., Jacobson, K., and Gratton, E. (2001) Lipid Rafts Reconstituted in Model Membranes, Biophys. J. 80, 1417-1428. Simons, K., and Ikonen, E. (1997) Functional rafts in cell membranes, Nature 387, 569. Seghezza, S., Dante, S., Diaspro, A., and Canale, C. (2015) High resolution nanomechanical characterization of multi-domain model membranes by fast Force Volume, J. Mol. Recognit. 28, 742-750. Seghezza, S., Diaspro, A., Canale, C., and Dante, S. (2014) Cholesterol Drives Aβ(1–42) Interaction with Lipid Rafts in Model Membranes, Langmuir 30, 1393413941. Giocondi, M.-C., Yamamoto, D., Lesniewska, E., Milhiet, P.-E., Ando, T., and Le Grimellec, C. (2010) Surface topography of membrane domains, Biochim. Biophys. Acta, Biomembr. 1798, 703-718. Oropesa-Nuñez, R., Seghezza, S., Dante, S., Diaspro, A., Cascella, R., Cecchi, C., Stefani, M., Chiti, F., and Canale, C. (2016) Interaction of toxic and non-toxic HypF-N oligomers with lipid bilayers investigated at high resolution with atomic force microscopy, Oncotarget 7, 44991-45004. Almonte, L., and Colchero, J. (2017) True non-contact atomic force microscopy imaging of heterogeneous biological samples in liquids: topography and material contrast, Nanoscale 9, 2903-2915. Mohwald, H. (1990) Phospholipid and phospholipid-protein monolayers at the air/water interface, Annu. Rev. Phys. Chem. 41, 441-476. Israelachvili, J. N. (1992) Intermolecular and Surface Forces, 2nd ed., Academic Press: San Diego. Lagace, T. A., and Ridgway, N. D. (2013) The role of phospholipids in the biological activity and structure of the endoplasmic reticulum, Biochim. Biophys. Acta, Mol. Cell Res. 1833, 2499-2510. Simberg, D., Weisman, S., Talmon, Y., and Barenholz, Y. (2004) DOTAP (and Other Cationic Lipids): Chemistry, Biophysics, and Transfection, Crit. Rev. Ther. Drug Carrier Syst. 21, 62.

ACS Paragon Plus Environment

Page 22 of 30

Page 23 of 30 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

ACS Chemical Neuroscience

34. 35. 36. 37. 38. 39.

40. 41. 42.

43. 44.

45.

46.

Wacklin, H. P. (2011) Composition and Asymmetry in Supported Membranes Formed by Vesicle Fusion, Langmuir 27, 7698-7707. Richter, R., Mukhopadhyay, A., and Brisson, A. (2003) Pathways of Lipid Vesicle Deposition on Solid Surfaces: A Combined QCM-D and AFM Study, Biophys. J. 85, 3035-3047. Kankare, J. (2002) Sauerbrey Equation of Quartz Crystal Microbalance in Liquid Medium, Langmuir 18, 7092-7094. Schram, V., Lin, H. N., and Thompson, T. E. (1996) Topology of gel-phase domains and lipid mixing properties in phase-separated two-component phosphatidylcholine bilayers, Biophys. J. 71, 1811-1822. Picas, L., Rico, F., and Scheuring, S. (2012) Direct Measurement of the Mechanical Properties of Lipid Phases in Supported Bilayers, Biophys. J. 102, L01L03. Demel, R. A., Geurts van Kessel, W. S. M., Zwaal, R. F. A., Roelofsen, B., and van Deenen, L. L. M. (1975) Relation between various phospholipase actions on human red cell membranes and the interfacial phospholipid pressure in monolayers, Biochim. Biophys. Acta, Biomembr. 406, 97-107. Mikelonis, A. M., Youn, S., and Lawler, D. F. (2016) DLVO Approximation Methods for Predicting the Attachment of Silver Nanoparticles to Ceramic Membranes, Langmuir 32, 1723-1731. Israelachvili, J. N., and Wennerstroem, H. (1990) Hydration or steric forces between amphiphilic surfaces?, Langmuir 6, 873-876. Zhao, F., Holmberg, J. P., Abbas, Z., Frost, R., Sirkka, T., Kasemo, B., Hassellov, M., and Svedhem, S. (2016) TiO2 nanoparticle interactions with supported lipid membranes - an example of removal of membrane patches, RSC Adv. 6, 9110291110. Wang, B., Zhang, L., Bae, S. C., and Granick, S. (2008) Nanoparticle-induced surface reconstruction of phospholipid membranes, Proc. Natl. Acad. Sci. U. S. A. 105, 18171. Torrano, A. A., Pereira, Â. S., Oliveira, O. N., and Barros-Timmons, A. (2013) Probing the interaction of oppositely charged gold nanoparticles with DPPG and DPPC Langmuir monolayers as cell membrane models, Colloids Surf., B 108, 120126. Carbone, L., Nobile, C., De Giorgi, M., Sala, F. D., Morello, G., Pompa, P., Hytch, M., Snoeck, E., Fiore, A., Franchini, I. R., Nadasan, M., Silvestre, A. F., Chiodo, L., Kudera, S., Cingolani, R., Krahne, R., and Manna, L. (2007) Synthesis and Micrometer-Scale Assembly of Colloidal CdSe/CdS Nanorods Prepared by a Seeded Growth Approach, Nano Lett. 7, 2942-2950. Di Corato, R., Quarta, A., Piacenza, P., Ragusa, A., Figuerola, A., Buonsanti, R., Cingolani, R., Manna, L., and Pellegrino, T. (2008) Water solubilization of hydrophobic nanocrystals by means of poly(maleic anhydride-alt-1-octadecene), J. Mater. Chem. 18, 1991-1996.

ACS Paragon Plus Environment

ACS Chemical Neuroscience 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

47. 48. 49. 50. 51. 52. 53.

54.

55. 56.

Quarta, A., Curcio, A., Kakwere, H., and Pellegrino, T. (2012) Polymer coated inorganic nanoparticles: tailoring the nanocrystal surface for designing nanoprobes with biological implications, Nanoscale 4, 3319-3334. Richter, R. P., and Brisson, A. R. (2005) Following the Formation of Supported Lipid Bilayers on Mica: A Study Combining AFM, QCM-D, and Ellipsometry, Biophys. J. 88, 3422-3433. Richter, R. P., Bérat, R., and Brisson, A. R. (2006) Formation of Solid-Supported Lipid Bilayers:  An Integrated View, Langmuir 22, 3497-3505. J.T. Davies, E. K. R. (1963) Interfacial phenomena, 2nd ed., New York : Academic Press. Vollhardt, D. (1996) Morphology and phase behavior of monolayers, Adv. Colloid Interface Sci. 64, 143-171. Anandarajah, A., and Chen, J. (1995) Single Correction Function for Computing Retarded van der Waals Attraction, J. Colloid Interface Sci. 176, 293-300. Petosa, A. R., Jaisi, D. P., Quevedo, I. R., Elimelech, M., and Tufenkji, N. (2010) Aggregation and Deposition of Engineered Nanomaterials in Aquatic Environments: Role of Physicochemical Interactions, Environ. Sci. Technol. 44, 6532-6549. Song, J. E., Phenrat, T., Marinakos, S., Xiao, Y., Liu, J., Wiesner, M. R., Tilton, R. D., and Lowry, G. V. (2011) Hydrophobic Interactions Increase Attachment of Gum Arabic- and PVP-Coated Ag Nanoparticles to Hydrophobic Surfaces, Environ. Sci. Technol. 45, 5988-5995. Chan, J., Popov, J. J., Kolisnek-Kehl, S., and Leaist, D. G. (2003) Soret Coefficients for Aqueous Polyethylene Glycol Solutions and Some Tests of the Segmental Model of Polymer Thermal Diffusion, J. Solution Chem. 32, 197-214. Robert, P. (2006) Principles of Cellular Engineering, 1st ed.

ACS Paragon Plus Environment

Page 24 of 30

Page 25 of 30 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

ACS Chemical Neuroscience

For Table of Contents Use Only

Polymer Coating and Lipid Phases Regulate Semi-Conductor Nanorods’ Interaction with Neuronal Membranes: a Modeling Approach Barbara Salis1,2*, Giammarino Pugliese3, Teresa Pellegrino3, Alberto Diaspro2,4, Silvia Dante2*

ACS Paragon Plus Environment

ACS Chemical Neuroscience 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

231x176mm (72 x 72 DPI)

ACS Paragon Plus Environment

Page 26 of 30

Page 27 of 30 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

ACS Chemical Neuroscience

288x201mm (300 x 300 DPI)

ACS Paragon Plus Environment

ACS Chemical Neuroscience 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

189x148mm (150 x 150 DPI)

ACS Paragon Plus Environment

Page 28 of 30

Page 29 of 30 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

ACS Chemical Neuroscience

182x143mm (150 x 150 DPI)

ACS Paragon Plus Environment

ACS Chemical Neuroscience 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

125x182mm (150 x 150 DPI)

ACS Paragon Plus Environment

Page 30 of 30