Biomaterials xxx (2014) 1e9

Contents lists available at ScienceDirect

Biomaterials journal homepage: www.elsevier.com/locate/biomaterials

Functional surface engineering by nucleotide-modulated potassium channel insertion into polymer membranes attached to solid supports Justyna Ł. Kowal a, Julia K. Kowal b, Dalin Wu a, Henning Stahlberg b, Cornelia G. Palivan a, Wolfgang P. Meier a, * a b

Chemistry Department, University of Basel, Klingelbergstrasse 80, 4056 Basel, Switzerland Center for Cellular Imaging and NanoAnalytics (C-CINA), Biozentrum, University of Basel, Mattenstrasse 26, 4058 Basel, Switzerland

a r t i c l e i n f o

a b s t r a c t

Article history: Received 4 April 2014 Accepted 16 May 2014 Available online xxx

Planar solid-supported membranes based on amphiphilic block copolymers represent promising systems for the artificial creation of structural surfaces. Here we introduce a method for engineering functional planar solid-supported membranes through insertion of active biomolecules. We show that membranes based on poly(dimethylsiloxane)-block-poly(2-methyl-2-oxazoline) (PDMS-b-PMOXA) amphiphilic diblock copolymers, which mimic natural membranes, are suitable for hosting biomolecules. Our strategy allows preparation of large-area, well-ordered polymer bilayers via LangmuireBlodgett and LangmuireSchaefer transfers, and insertion of biomolecules by using Bio-Beads. We demonstrate that a model membrane protein, the potassium channel from the bacterium Mesorhizobium loti, remains functional after insertion into the planar solid-supported polymer membrane. This approach can be easily extended to generate a platform of functional solid-supported membranes by insertion of different hydrophobic biomolecules, and employing different types of solid substrates for desired applications. © 2014 Elsevier Ltd. All rights reserved.

Keywords: Functional surfaces Planar solid-supported polymer membranes Membrane protein Biomolecule insertion Amphiphilic block copolymers

1. Introduction Biological membranes are complex structures, consisting of phospholipids, proteins, and oligosaccharides, where a variety of processes, such as active and passive transport across the membrane or molecular recognition interactions, occur simultaneously. Because of its complexity, it has not yet been possible to reconstruct a cell membrane, and simplified membrane models, mainly based on phospholipids, have been developed to investigate selected processes that take place in membranes. Phospholipids have been preferred because they are components of the natural membranes and can be used for in vivo applications [1]. However, as they are not sufficiently stable for long-term experiments, other models based on amphiphilic block copolymers have been introduced [2]. These are an improved alternative to phospholipids, because they selfassemble in specific aqueous conditions, and have increased mechanical stability [3]. In addition, their structure and properties can be designed through copolymer chemical engineering [4]. Vesicles, which find applications, e.g. in drug delivery [5], or as nanoreactors [6], are frequently studied membrane models.

* Corresponding author. Fax: þ41 (0) 61 267 38 55. E-mail address: [email protected] (W.P. Meier).

Because of their hollow structure, the bilayer is hydrated on both sides, which make them similar to natural membranes. Appropriate insertion of membrane proteins such as Complex I in a desired position has allowed the development of nanodevices to conduct electron transfer from the environment to a specific location inside the membrane [7]. Additionally, successful insertion of channel proteins in the membrane of polymer vesicles loaded with catalysts (enzymes, proteins, mimics) has allowed the exchange of molecules with the environment, and supported in situ functionality of these nanoreactors. Another membrane model is a freestanding membrane, which has been used for investigations of protein insertion mechanisms [8]. Such planar membranes are accessible from both sides, and act as perfect insulators; hence insertion of a protein can be detected by a change in electrical properties of the system [9]. While there are examples of freestanding polymer membranes, which have successfully hosted membrane proteins, the disadvantages of this system, i.e. a low stability and difficult to handle, make it of low technological interest. Planar solid-supported membranes represent a step further as membrane models in terms of increased stability and preserved fluidity [10]. Therefore, they are more appropriate for insertion of membrane proteins and the study of protein functions or biomoleculeesurface interactions. These membranes can be used for

http://dx.doi.org/10.1016/j.biomaterials.2014.05.043 0142-9612/© 2014 Elsevier Ltd. All rights reserved.

Please cite this article in press as: Kowal JŁ, et al., Functional surface engineering by nucleotide-modulated potassium channel insertion into polymer membranes attached to solid supports, Biomaterials (2014), http://dx.doi.org/10.1016/j.biomaterials.2014.05.043

2

J.Ł. Kowal et al. / Biomaterials xxx (2014) 1e9

investigating multivalent ligand-receptor binding, biomimetic sensing, or drug screening [11,12]. Lipid solid-supported membranes have been broadly investigated, and numerous examples of adsorption of biomolecules into these membranes have been reported, e.g. incorporation of ATPase [13], cytochrome c oxidase [14], a-hemolysin (a-He) [15], or outer membrane proteins (OmpF and OmpA) [16]. However polymer membranes are more appropriate for applications, because they surpass lipid membranes in the respect of stability, i.e. their structures can persist for a few hours even after drying [17]. Lipids form rigid monolayers in a liquid condensed state (Fig. S1, Supplementary Information) while the polymer films are in a liquid expanded state at the airewater interface [18,19]. This indicates that an appropriate selection of the polymer blocks leads to a resulting membrane more flexible than lipid membranes and thus better suited for biomolecule insertion, as previously reported by insertion of Complex I in the polymer membrane of vesicles [7]. The drawback of solid-supported membranes is the risk of protein denaturation, as the result of direct contact between transmembrane protein and substrate, but this problem can be overcome by separating the solid substrate from the membrane with a spacer [20]. Dorn et al. were the first to investigate interactions between a planar solid-supported polymer membrane, formed by poly(butadiene)-block-poly(ethylene oxide) (PB-b-PEO) block copolymer, and a polypeptide polymyxin B [21]. However the peptide was adsorbed into the polymer membrane only temporarily, and slowly diffused back into the solution. A further step was realized by incorporating water soluble a-He into solid-supported PB-b-PEO membrane when the membrane was destabilized by an electrical current [22]. This method allowed a permanent and functional insertion of the protein, as proved by a flow of ions through the membrane until Donnan equilibrium was reached. However, except for this example of a-He insertion supported by membrane destabilization through application of electric current, there have been no other methods reported for insertion of membrane proteins in solid-supported polymer membranes. This is explained by the complex scenario and requirements that are necessary for functional insertion of a membrane protein: i. a homogeneous and stable membrane, ii. a membrane with sufficient fluidity to host a protein, and iii. the presence of a spacer between the substrate and membrane (e.g. polymer layer or a water reservoir) to prevent protein denaturation, which can occur as the result of interactions with the hard support. Here we present a new approach for incorporating membrane proteins into large area, solid-supported polymer membranes without using an electrical current to destabilize the membrane (Fig. 1). In this approach, well-organized polymer bilayers have been prepared by film transfer techniques. We selected poly(dimethylsiloxane)-block-poly(2-methyl-2-oxazoline) (PDMS-bPMOXA) diblock copolymer as a good candidate for forming an artificial membrane, as its hydrophobic PDMS block provides the necessary flexibility for the membrane [23], while the water soluble PMOXA block acts as a buffer to prevent protein interactions with the solid substrate. In order to characterize the polymer membrane, a combination of surface analysis techniques were used, i.e. atomic force microscopy (AFM), ellipsometry, contact angle measurements, and attenuated total reflection Fourier transform infrared spectroscopy (ATR-FTIR). To perform a functional insertion of the membrane protein into the polymer membrane, both the protein and polymer membrane have to be destabilized. For this purpose, we used Bio-Beads which are non-polar, polystyrene beads capable of adsorbing organic materials from aqueous solutions, and thus providing a reproducible and simple way to obtain polymer vesicles [7] and liposomes with inserted proteins [24]. However Bio-Beads

Fig. 1. Schematic representation of membrane protein insertion into solid-supported polymer membrane with usage of Bio-Beads.

have never been used previously for protein incorporation into planar solid-supported polymer membranes. The critical point of the whole procedure is the preparation of defect-free homogeneous membranes, because each defect influences the conductance measurement. It is also of high importance that protein destabilization is gentle, because too rapid detergent removal can induce protein precipitation. Our approach is an easy and straightforward way to engineer functional surfaces by protein insertion into solid-supported polymer membranes. Besides its simplicity, the advantage of this method is the possibility of employing different solid substrates of unrestricted sizes, which represents a key point for development of large functional surfaces for technological, medical or environmental applications. In addition, this approach can be easily extended in terms of functionality by insertion of other hydrophobic proteins and biomolecules. 2. Materials and methods 2.1. Polymer synthesis and characterization We used PDMS-b-PMOXA diblock copolymers with an aldehyde end-group. The polymer with a molar mass of 5735 g mol1 was composed of 65 PDMS units and 12 PMOXA units. It was synthesized according to the procedure described by Egli et al. [25]. The molecular mass of the polymer and length of the blocks were calculated from the 1H NMR spectra. GPC data showed PDI of the PDMS-b-PMOXA diblock copolymer to be 1.67. Oxidation of the hydroxyl end-group was performed using Dess-Martin periodinane (DMP, Aldrich) [26]. PDMS-b-PMOXA-OH (200 mg) and DMP (17 mg, 40 mmol) were added to a two-neck round bottom flask, closed, and degassed. Then anhydrous dichloromethane (10 ml) was introduced under a stream of argon, and the reaction mixture was stirred for 24 h at room temperature. The modified polymer was purified by dialysis (Spectrapor® with MWCO 3500 Da) in ethanol for 18 h. 2.2. MloK1 expression and purification Full-length, cyclic nucleotide-modulated potassium channel MloK1 was expressed and purified to homogeneity as described in Ref. [27]. Briefly, Escherichia coli cells containing His-tagged MloK1 construct were grown in LB medium at 37  C.

Please cite this article in press as: Kowal JŁ, et al., Functional surface engineering by nucleotide-modulated potassium channel insertion into polymer membranes attached to solid supports, Biomaterials (2014), http://dx.doi.org/10.1016/j.biomaterials.2014.05.043

J.Ł. Kowal et al. / Biomaterials xxx (2014) 1e9 Protein expression was induced with anhydrotetracycline (0.2 mg/ml) for 2 h at OD600 of 0.7. Bacterial cells were then centrifuged and disrupted by sonication. The membrane fraction was isolated by ultracentrifugation and solubilized for 2.5 h at 4  C in buffer containing 1.2% n-decyl-b-D-maltopyranoside (DM; Anatrace), 295 mM NaCl, 5 mM KCl, 20 mM TriseHCl pH 8.0, 10% glycerol, 1 mM PMSF, 0.2 mM cAMP (Fluka). Insoluble material was removed by ultracentrifugation and extracted MloK1 was purified by Co2þ-affinity chromatography in buffer containing 295 mM NaCl, 5 mM KCl, 20 mM TriseHCl pH 8.0, 10% glycerol, 1 mM PMSF, 40/500 mM (wash/ elution) imidazole, 0.2% DM, 0.2 mM cAMP.

2.3. Surface pressure e area isotherms Polymer monolayers at the airewater interface were investigated on a KSV Inc. (Finland) Langmuir Teflon® trough with the area of 420 cm2 and equipped with two symmetric, hydrophilic Delrin® barriers and a Wilhelmy plate made of ashless filter paper. Before each use the trough was cleaned with high purity chloroform and ethanol, whereas barriers were cleaned with ethanol and rinsed with doubledistilled water. All experiments were performed at 20  C using double-distilled water as the subphase. A polymer solution in chloroform (1 mg ml1) was spread drop-wise on the pure water and the solvent was allowed to evaporate over 15 min. The monolayer compression was performed with barriers speed of 10 mm min1.

2.4. Substrate preparation As substrates we selected silicon dioxide wafers (SiO2, Si-Mat Silicon Materials, €ser, Germany), and gold substrates (Ssens, Germany), glass cover slips (Menzel-Gla Netherlands). The cleaning procedure for these surfaces was the same: the slides were cleaned in an ultrasonic bath in chloroform three times for 15 min, rinsed with ethanol, and then activated in a UV/ozone chamber for 15 min. Freshly cleaned silicon dioxide and glass substrates were modified with 3aminopropyltriethoxysilane (99%, APTES, Acros Organics) in a nitrogen atmosphere. The substrates were incubated in 5% (v/v) solution of toluene (anhydrous, 99.9%) and APTES for 3 h, then rinsed thoroughly with toluene and ethanol, and dried in a stream of nitrogen. Gold substrates were modified with 11-amino-1-undecanethiol hydrochloride (ADT, SigmaeAldrich). Au slides were incubated overnight in 0.5 mM solution of ADT in ethanol containing 3% (v/v) triethylamine, then rinsed with ethanol, and dried with stream of nitrogen [28].

2.5. LangmuireBlodgett (LB) and LangmuireSchaefer (LS) transfers LB and LS transfers of PDMS-b-PMOXA diblock copolymer were carried out on a Mini-trough (KSV Instruments, Finland) with the surface area of 242 cm2. In order to perform the LB transfer, the silica wafer was immersed into the subphase prior to film spreading. The polymer film was compressed to a surface pressure of 37 mN m1 and left for 10 min to equilibrate. Then the film was transferred with a constant dipper speed of 0.5 mm min1. Such prepared monolayers were used for the transfer of the second layer with the LS technique. In this approach, the polymer film was first compressed at the airewater interface with a surface pressure of 37 mN m1, then after 10 min of stabilizing the slide was placed horizontally above the floating film and pressed through the interface into the water with a constant dipper speed of 50 mm min1. The water surface was precisely cleaned so that the wafer with bilayer could be removed from the reservoir and placed in a vessel filled with ultra-pure water.

2.6. Protein incorporation MloK1 was dissolved in Tris buffer (20 mM Tris pH 7.6, 20 mM KCl, 200 mM adenosine 30 ,50 -cyclic monophosphate (cAMP), 0.4% n-decyl-b-D-maltopyranoside (DM, Affymetrix), 0.02% NaN3). All experiments were performed in Tris buffer expect for those with labeled protein in which Bicine buffer (20 mM Bicine pH 7.6, 20 mM KCl, 200 mM cAMP, 0.4% DM) was used. To incorporate the protein, the substrate with tethered polymer bilayer was placed in buffer, then Bio-Beads SM-2 (Bio-Rad, Richmond, California) and protein solution (final protein concentration: 14.0 mg ml1) were added. The vessel was sealed and kept for 3 h at room temperature. Then the substrates were thoroughly rinsed with detergent-free buffer (20 mM Tris pH 7.6, 20 mM KCl, 200 mM cAMP, 0.02% NaN3).

2.7. Protein labeling MloK1 was labeled at its primary amines with the fluorescent dye DyLight 488 Amine-Reactive (Thermo Fisher Scientific). Tris buffer was not used for labeling, because it is a primary amine which will compete with the protein for reaction with the dye, and instead Bicine (20 mM Bicine pH 7.6, 20 mM KCl, 200 mM cAMP, 0.4% DM), which is a tertiary amine, was used as the buffer. DyLight 488 (1 mg) was dissolved in DMSO (100 ml), then the dye solution (10 ml) was added to MloK1 (100 ml, 2.7 mg ml1) dropwise, and stirred for 2 h. The protein was purified from free dye by dialysis (MWCO 10 kDa) against Bicine buffer for 5 days at 6  C.

3

2.8. Characterization methods 1 H NMR spectra were recorded with a Bruker DPX-400 spectrometer using deuterated chloroform (99.8% CDCl3, 0.05% TMS) as solvent, and analyzed with MestReNova 6.1.1 software. The molecular mass of the polymer and length of the blocks were calculated from the 1H NMR spectra. Atomic force microscopy (AFM) images were recorded with an Agilent 5100 AFM/ SPM microscope (PicoLe System, Molecular Imaging). All the measurements were carried out using silicon nitride cantilevers (PNP-TR, NanoWorld AG) with a nominal spring constant of 0.32 N m1 in the contact mode in the Bicine buffer (20 mM Bicine pH 7.6, 20 mM KCl, 200 mM cAMP). Ellipsometric measurements were carried out on an EP3 SW imaging ellips€ ttingen, Germany) with Nd-YAG laser at ometer (Nanofilm Technologie GmbH, Go 532 nm. Measurements (one every 2 ) were performed in air for angles of incidence ranging from 55 to 75 . For the silica substrates, the thickness of the layer was estimated by a model which included the silicon dioxide thickness (~2 nm). Refractive index values used for modeling were: nAPTES ¼ 1.465, nADT ¼ 1.53, and npolymer ¼ 1.5. Static contact angle measurements were performed with a contact angle goniometer, CAM 100 (KSV Instruments, Finland) based on a CDD camera with 50 mm optics. Droplets of ultrapure water were placed on the substrates with a microsyringe, and the contact angle was automatically calculated by fitting the curve with the YoungeLaplace equation. Each sample was measured at least 10 times and the average value was calculated. Attenuated total reflection Fourier transform infrared spectroscopy (ATR-FTIR) measurements were performed on a Platinum ATR ALPHA (Bruker, Germany) spectrometer with a single reflection diamond ATR sampling module. All spectra were recorded with a resolution of 2 cm1 in the range 400e4000 cm1, with 128 acquisition scans. Confocal laser scanning microscopy (CLSM) measurements were performed on a Zeiss LSM 510-META/Confocor2 (Germany), in LSM mode with an Ar laser (488 nm) and a 40 water-immersion objective (Zeiss C-Apochromat 40, NA 1.2) with pinhole adjusted to 70 mm. An Ar laser was used as excitation source with excitation transmission at 488 nm set for 4%. Samples for CLSM measurements were prepared on glass cover slips. Before performing the measurements, a small volume of Bicine buffer (20 mM Bicine pH 7.6, 20 mM KCl, 200 mM cAMP) was placed on a cleaned microscope slide and covered with a glass cover slip so that the polymer membrane was enclosed between two slides. Measurements were performed at room temperature, and after adjusting for a sharp image, the sample was scanned randomly throughout the surface. Electrical conductance measurements were performed with a source-meter Keithley 2636A (Keithley International, Germany). To carry out these measurements we built an electric circuit, and used samples prepared on conductive substrates, i.e. gold slides. The gold substrate with polymer membrane was covered with a PDMS liquid chamber (which had a small vertical hole) in order to always have the same measurement area and constant buffer volume. A gold wire was attached with a silver paint to the sample so that the gold substrate was connected to the circuit [22]. The paint was left for 30 min to dry but the membrane was still hydrated, and the liquid chamber was then filled with buffer (20 mM Tris pH 7.6, 20 mM KCl, 200 mM cAMP, 0.02% NaN3) and left for 15 min to stabilize. From the top, the liquid chamber was closed with an electrode. A constant voltage of 40 mV was applied to the system, and the current was measured. All devices were controlled by self-made LabView software.

3. Results and discussion 3.1. Strategy for creation of functional surfaces PDMS65-b-PMOXA12 diblock copolymer with aldehyde endgroups was used to support a covalent attachment of the polymer film through formation of an imine bond with an amino modified substrate (Fig. 2). All substrates were modified with linkers, which function as spacers between the polymer film and solid support, and additionally increase the polymer membrane stability through formation of a covalent bond with the first membrane interface. Silicon dioxide and glass substrates were modified with 3aminopropyltriethoxysilane (APTES), while gold substrates were modified with 11-amino-1-undecanethiol hydrochloride (ADT). These linkers are known to assemble on the surfaces and to form films with exposed amino groups [28,29]. Formation of a weak imine bond is advantageous, since it provides the membrane with increased stability, while preserving its fluidity, because not all polymer chains attach permanently to the substrate. In addition, fluidity of the membrane is provided by non-covalent interactions of the second membrane layer.

Please cite this article in press as: Kowal JŁ, et al., Functional surface engineering by nucleotide-modulated potassium channel insertion into polymer membranes attached to solid supports, Biomaterials (2014), http://dx.doi.org/10.1016/j.biomaterials.2014.05.043

4

J.Ł. Kowal et al. / Biomaterials xxx (2014) 1e9

Fig. 2. Schematic representation of solid-supported polymer membrane of PDMS-b-PMOXA applied for protein incorporation.

In order to covalently attach the polymer to the amino-modified substrate, the polymer has to bear an aldehyde-end group. For this purpose the hydroxyl end-group of the polymer was oxidized using Dess-Martin periodinane (DMP), a selective oxidant for primary alcohols [26]. As a model membrane protein, we selected a cyclic nucleotide-modulated potassium channel from Mesorhizobium loti (MloK1) [27,30]. MloK1 is interesting to be used as model protein because of structure similarity to eukaryotic cyclic nucleotidemodulated ion channels, which are well known for signal transduction in eukaryotes [31e33]. The MloK1 monomer consists of six transmembrane a-helices and an N-terminal cytoplasmic cyclic nucleotide binding domain (CNBD). MloK1 forms a tetrameric complex with a molecular mass of approximately 148 kDa, a height of 10 nm, and a width of 8.5 nm [30]. The tetrameric full-length channel is composed of: i. a transmembrane part containing the central pore and putative voltage sensing domains, and ii. the cytosolic part comprising four CNBDs. When the latter bind 30 ,50 cyclic adenosine monophosphate (cAMP) or cyclic guanosine monophosphate (cGMP), they induce conformational changes that activate the channel [34]. Comparison of the two structures of MloK1 reconstituted in a lipid bilayer revealed structural changes of the channel upon ligand binding, including vertical movement of the CNBDs towards the membrane to bring them into direct contact with the putative voltage sensor domains [27]. As common for membrane proteins, MloK1 is insoluble in water because of its hydrophobic transmembrane region. Use of a detergent (n-decyl-b-D-maltopyranoside, DM) is required to dissolve it in a buffer solution. In aqueous solution the amphiphilic detergent surrounds the protein which keeps it in its native structure. To incorporate the detergent-solubilized protein into a membrane, we removed the detergent by adding Bio-Beads (pore diameter of approximately 90 Å) to a protein-copolymer-detergent mixture [35]. Addition of Bio-Beads allows efficient detergent removal from the solution without affecting the protein [36,37]. This drives the insertion of the protein into the amphiphilic copolymer membrane. In order to preserve its structure, the protein is forced to reconstitute in the polymer bilayer, which mimics a natural membrane. 3.2. Formation of polymer films at the airewater interface To perform a successful transfer of the polymer to a solid substrate, the polymer has to form a homogeneous and stable film at the airewater interface. To prepare such film, a chloroform solution of the polymer was spread dropwise on the water surface. After

evaporation of the solvent, the polymer was compressed by movable barriers to a required surface pressure. The behavior of the aldehyde terminated PDMS-b-PMOXA diblock copolymer at the airewater interface was evaluated by measurements of surface pressure-molecular area isotherms performed on a Langmuir trough. A Brewster angle microscope (BAM) gave information about film morphology during film compression. The film formed by PDMS65-b-PMOXA12 diblock copolymer had a collapse point at 49 mN m1. Typical surface pressure-area isotherms for PDMS-b-PMOXA diblock copolymers with aldehyde end-groups were reproducible and reversible with a characteristic plateau at a molecular area of 400e1000 Å2 (Fig. S2, Supplementary Information). Such plateaus have also been observed for PMOXA-bPDMS-b-PMOXA triblock copolymers, and are explained by polymer rearrangement upon compression [18]. After spreading on the airewater interface, polymer chains stay in a relaxed state with the hydrophilic blocks (PMOXA) in water, and the hydrophobic blocks (PDMS) directed toward air. In the plateau area, PMOXA blocks are desorbing to form a brush in water, while PDMS blocks are starting the formation of a homogenous film (Fig. S3, Supplementary Information) [38]. The stability of the polymer film at the airewater interface is an essential factor for successful transfer of the monolayer to the solid support. At a surface pressure of 37 mN m1, the polymer films were stable for >1 h. Additionally, BAM indicated that the film stayed smooth during the whole compression (Fig. S5, Supplementary Information). The fluidity of the membrane is crucial for polymer films to cope with insertion of biomolecules. The compressibility modulus, which is a number describing the physical state of Langmuir films, can be calculated from the surface pressure-area isotherms Cs1 ¼ A(vp/vA)T, where A is the mean molecular area (Å2/molecule), p is the surface pressure (mN mol1), and T is the temperature ( C). The value of the compressibility modulus calculated for a PDMS65-b-PMOXA12 diblock copolymer film with oxidized endgroup (up to 45 mN m1, Fig. S2, Supplementary Information) indicates that the film was in a liquid expanded state during transfer to the substrate [39], and because of its preserved fluidity, the film was appropriate for protein incorporation. 3.3. Functionalization of the solid support with amino groups We selected three different substrates (silicon dioxide, glass, and gold) with different properties that allow a specific characterization

Please cite this article in press as: Kowal JŁ, et al., Functional surface engineering by nucleotide-modulated potassium channel insertion into polymer membranes attached to solid supports, Biomaterials (2014), http://dx.doi.org/10.1016/j.biomaterials.2014.05.043

J.Ł. Kowal et al. / Biomaterials xxx (2014) 1e9

and to modulate their interactions with the polymer film. Silicon wafers and glass are widely used as biomaterials and supports for investigation of protein adsorption on modified surfaces [40,41]. With similar properties and reactivity to silicon-, glass- substrates can be investigated by confocal laser scanning microscopy (CLSM) where transparent surfaces are essential. We employed gold substrates for conductance measurements, where silicon wafers are not appropriate. Au substrates find multiple applications, e.g. in imaging of biomolecules and biosensing, due to their optical and electrical properties, and biocompatibility [42]. Silicon and glass substrates were modified with APTES so that the polymer monolayer could bind covalently to the substrate via formation of a Schiff base between aldehyde end-groups of the polymer and the amino functionalized surface [43]. To avoid formation of multilayers, functionalization of the substrate was performed in water- and oxygen-free conditions with a short reaction time [29,44]. The thickness of the resulting APTES film determined by ellipsometry was approximately 0.9 ± 0.1 nm, in agreement with the theoretical value of 0.8 nm [29]. After silanization, a change in hydrophobicity of SiO2 substrates was observed: the contact angle increased from 35 for a bare silicon wafer to 66 for the silanized surface. AFM measurements showed that the layer was homogeneous (Fig. S6, Supplementary Information). Silanization was also investigated by ATR-FTIR spectroscopy. As reference we measured a bare silicon slide for which no absorption bands were detected. Appearance of peaks characteristic of APTES in the spectrum of APTES-modified silicon substrate indicates successful functionalization (Fig. S7a, Supplementary Information). The observed peaks from functional groups of APTES are: SieO at 615 cm1, 740 cm1, and 1106 cm1, and CeH from alkyl groups at 2675 cm1. Gold substrates were modified with an ADT linker bearing amino and thiol end-groups, which is known to form selfassembled monolayers on gold substrates [45]. The contact angle decreased from 103 for bare gold to 76 for the ADT modified substrate, thus showing that the linker was successfully attached to the substrate. By ellipsometry, the thickness of the ADT layer was evaluated as 3.3 ± 0.2 nm, and AFM showed the film to be smooth (Fig. S6, Supplementary Information). Functionalization of Au surfaces was also established by ATR-FTIR. Thus, whereas a blank gold substrate did not show any absorption bands, the spectrum of ADT showed: a peak at 1046 cm1 corresponding to the CeN group, and peaks at 2897 cm1 and 2985 cm1 corresponding to the CeH stretching modes (Fig. S7b, Supplementary Information). 3.4. Preparation of well-organized solid-supported membranes Incorporation of biomolecules into the solid-supported membranes requires the formation of homogeneous and smooth films. To address these requirements we prepared the polymer membranes by LangmuireBlodget (LB) and LangmuireSchaefer (LS) techniques, which are known to provide transfers of highly ordered and defect-free monolayers. The transfers were performed at a surface pressure of 37 mN m1. At this surface pressure the isotherm forms a steep slope and is close to the collapse point (Pcp ¼ 49 mN m1); thus the polymer film forms a densely packed brush. Additionally, the film is elastic due to the compressibility modulus calculated for this surface pressure (Cs1 ¼ 45 mN m1), and the polymer is in the liquid expanded phase. To prepare the first layer of the membrane we used amino modified substrates and applied the LB technique; freshly transferred films on silicon dioxide were characterized by contact angle, ellipsometry and AFM. The contact angle of the polymer monolayer on silicon surface was different from that of bare and silanized Si slides: the surface was more hydrophobic with a contact angle of

5

80 , due to the presence of PDMS blocks directed upwards. By taking into account the thickness of the SiO2 and APTES layers, the thickness of the polymer film was calculated as 6.5 ± 0.5 nm. The monolayer was homogeneous and smooth, as can be seen by AFM (Fig. 3a); the roughness of the surface did not exceed 0.2 nm. Bright structures visible on the surface correspond to impurities from water, which was adsorbed on the polymer film during transfer, and not removed by rinsing. Scratching experiments showed that the polymer film could not be removed from the substrate, which confirms covalent attachment. Scratching produced a trough with a depth of 2 nm, which was due to the removal of non-covalently bound polymer chains (Fig. 3c) [17]. The second layer of the polymer membrane was transferred using the LS technique. As this upper layer attaches to the lower one by non-covalent interactions, the membrane stays fluid, which is an essential condition for protein incorporation. The average thickness of the dry membrane was 11.3 ± 0.5 nm, as determined by ellipsometry. AFM images show a homogeneous surface, with a roughness of approximately 0.6 nm, which is higher than the roughness of the polymer monolayer due to the membrane fluidity (Fig. 4). In order to prove covalent attachment of the polymer to amino modified substrates we used ATR-FTIR spectroscopy for membranes prepared on both silicon and gold slides. The solidsupported membrane was measured by ATR-FTIR, then thoroughly rinsed with ethanol, which is a good solvent for PDMS-bPMOXA diblock copolymer, and measured again. The characteristic peaks for polymer, present in the spectra after rinsing of sample indicate that the polymer did not detach from the surface (Fig. 5), and was thus tethered covalently. Bare silicon or gold slides were used as the background. The appearance of the peaks characteristic of the polymer indicate the presence of the polymer on the substrate. The peaks at 2958 cm1 (on the spectrum of polymer membrane prepared on SiO2 substrate), and at 2963 cm1 (on the spectrum of polymer membrane prepared on Au substrate), are associated with CeH bond from alkyl groups, those at 1634 cm1 (SiO2 substrate) and 1645 cm1 (Au substrate) are from the amide group, the peaks between 1265 cm1 and 1050 cm1 correspond to SieOeSi stretching, and the peaks at 820 cm1 are assigned to SieCH3 group. 3.5. Incorporation of the membrane protein into the polymer bilayer In order to observe protein adsorbed to the membrane we prepared a polymer bilayer on the glass substrate, and used MloK1 labeled with a fluorescent dye (DyLight 488 e attachment to the primary amine of MloK1). We performed CLSM measurements, because it enables the observation of fluorescent objects on transparent substrates. Incubation of the labeled protein with the solid-supported polymer membrane and Bio-Beads results in a surface covered with green (in web version) spots of identical size, which correspond to protein (Fig. 6a). When protein insertion was performed without Bio-Beads in the solution, no fluorescent features attached to the surface could be observed (Fig. 6b). Addition of the detergent-solubilized protein to the solution, in which the solidsupported membrane was present, was not sufficient for protein insertion. To show that MloK1 was incorporated into the membrane and not just attached to the polymer surface, we performed an insertion experiment of: i. protein in the presence of the silanized substrate (without polymer membrane) and Bio-Beads, and ii. dye in the presence of polymer membrane and Bio-Beads. No protein attachment took place on silanized substrate (Fig. 6c), which demonstrates that the labeled protein inserts into the

Please cite this article in press as: Kowal JŁ, et al., Functional surface engineering by nucleotide-modulated potassium channel insertion into polymer membranes attached to solid supports, Biomaterials (2014), http://dx.doi.org/10.1016/j.biomaterials.2014.05.043

6

J.Ł. Kowal et al. / Biomaterials xxx (2014) 1e9

a

8.9 nm

0.6 0.4

6.0 5.0 4.0 3.0

Height [nm]

7.0

2.0 1.0

0.2 0.0 -0.2 -0.4

0.0

0

-1.1 30.0 nm

c

b

8.0

25.0 20.0

3 2

1

4

3 2 Distance [μm]

5

d

5.0 -0.0 -5.0 -10.0 -15.0 -20.0 -25.7

Height [nm]

15.0 10.0

1 0 -1 -2 0

1

2

3 4 5 6 Distance [μm]

7

8

9 10

Fig. 3. AFM image of (a) polymer monolayer and (b) the corresponding cross-section. (c) shows an image of scratched monolayer and (d) the corresponding cross-section. Scale bars: 2 mm in (a), 4 mm in (c).

polymer bilayer. Also, incubation of the polymer membrane with Bio-Beads and fluorescent dye resulted in a CLSM micrograph without any fluorescent features on the surface (Fig. 6d), which indicates that the dye is not incorporated into the polymer membrane. CLSM micrographs confirmed that MloK1 adsorbs to the substrate only when polymer membrane and Bio-Beads are present in the system. The functionality of the protein inserted into the polymer membrane was evaluated by measuring the electrical conductance (with the setup presented in Fig. 7). We measured a current across the membrane as a function of time for a constant applied voltage of 40 mV. The conductance was calculated as G ¼ I/V, where I is an electrical current, and V voltage. Due to the fact that each defect,

e.g. inhomogeneity of membrane, and surface contamination, influences the final result, conductance measurements have high inherent errors. To obtain accurate results all samples (blank gold, polymer membrane, and polymer membrane with incorporated protein) were prepared separately at least 10 times and the average values of their conductance calculated (Fig. 8). Bare gold substrates gave high conductance (191.8 ± 31.1 nS), and as expected, formation of the polymer membrane resulted in a strong decrease in conductance to 25.1 ± 8.5 nS, which corresponds to a resistance (1/G) value of 40 MU cm2. In order to investigate the influence of Bio-Beads on membrane stability, we performed electrical measurements of the polymer membrane after 3 h of incubation with Bio-Beads in Tris buffer. Conductance increased to

Fig. 4. AFM image of a polymer bilayer and the corresponding cross-section. Scale bar: 2 mm.

Please cite this article in press as: Kowal JŁ, et al., Functional surface engineering by nucleotide-modulated potassium channel insertion into polymer membranes attached to solid supports, Biomaterials (2014), http://dx.doi.org/10.1016/j.biomaterials.2014.05.043

J.Ł. Kowal et al. / Biomaterials xxx (2014) 1e9

7

Fig. 5. ATR-IR spectra of polymer bilayer on silicon dioxide (a) and gold (b) substrates before (black line) and after (red line) rinsing with ethanol. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

33.1 ± 4.3 nS, which is close to the conductance of the bilayer, and indicates that Bio-Beads do not affect the membrane structure. A bilayer incubated for 3 h with the protein, but without Bio-Beads, exhibited a conductance at the same level as that of the intact membrane (35.0 ± 5.5 nS), which indicates that the protein was not inserted into the polymer bilayer. Conductance increased to 70.9 ± 22.5 nS, only when protein was incubated with polymer membrane and Bio-Beads, which indicates a successful insertion of the protein into the membrane [15,46]. The same measurements were performed for solid-supported membrane prepared with lipid (1, 2-diphytanoyl-sn-glycero-3phosphocholine, DPPC). Lipid membrane presented a conductance of 312.5 ± 13.7 nS, while after insertion of MloK1 it increased to 422.5 ± 17.0 nS, which indicates a similar behavior to the polymer membrane (Fig. S9, Supplementary Information). As expected,

Fig. 6. CLSM micrographs of: (a) polymer membrane after MloK1 incorporation using Bio-Beads, (b) polymer membrane after incubation with protein, without Bio-Beads, (c) silanized substrate after incubation with protein and Bio-Beads, and (d) polymer membrane after incubation with dye and Bio-Beads. Scale bars: 50 mm.

the lipid membrane has a lower resistance (3.2 MU cm2) than polymer membrane, because of smaller molecule size, and thus lower membrane thickness [9], known to induce a lower mechanic stability than the polymer membrane. The change of conductance after protein insertion was higher for lipid membranes than for polymer membranes, which corresponds to higher number of inserted MloK1 [22]. However in development of robust proteinpolymer membrane systems, a balance between the number of inserted proteins and membrane stability is necessary. As the solid-supported membranes are asymmetric, during the conductance measurement ions accumulate in the free, hydrophilic space between membrane and substrate. The increase in conductance for the protein-incorporated polymer membrane is the result of greater accumulation of ions in the inner part of the system due to selective permeability of MloK1 for potassium ions [22]. Because the amino-linker on the substrate forms a film, ions accumulate rapidly in the spaces of a limited volume which exist due to the: i. polymer chains of the first membrane layer (that are not covalently bound), and ii. fluidity of the polymer membrane, which being in the liquid expanded state, does not form a rigid brush. Because detailed current measurements for MloK1 have not been performed previously, we compared our results with data obtained for a protein with a similar structure to MloK1, i.e. the

Fig. 7. Schematic representation of the setup used for conductance measurements through polymer membrane (S-M e source-meter, PDMS e poly(2-methyl-2oxazoline) stamp).

Please cite this article in press as: Kowal JŁ, et al., Functional surface engineering by nucleotide-modulated potassium channel insertion into polymer membranes attached to solid supports, Biomaterials (2014), http://dx.doi.org/10.1016/j.biomaterials.2014.05.043

8

J.Ł. Kowal et al. / Biomaterials xxx (2014) 1e9

Fig. 8. (a) Time course for conductance of solid-supported polymer bilayer (black line) and solid-supported polymer bilayer with incorporated MloK1 (red line). (b) Conductance measured at a constant applied voltage of 40 mV (Au e gold substrate, BB e Bio-Beads). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

cyclic nucleotide-gated Kþ channel from Magnetospirillum magnetotacticum (MmaK), which when expressed in E. coli cells, has an unitary conductance of 10.8 pS at given experimental conditions [47]. Assuming that MloK1 has a similar input into membrane conductance, and taking into account the conductance increase of our sample after protein insertion (46 nS for measured gold surface area of 7 mm2), we estimate the density of incorporated protein to be approximately 610 MloK1 per mm2. The activity of the potassium channel was established by conductance measurements performed in the presence and absence of cAMP, which is known to modulate its functionality [48,49]. cAMP did not influence the conductance of the solidsupported polymer membrane (Fig. 9). The conductance of the polymer membrane with incorporated MloK1 increased significantly in the presence of cAMP, up to 70.9 ± 22.5 nS, which indicates an open channel of the protein. In contrast, the conductance was 39.5 ± 7.5 nS when the polymer membrane with incorporated MloK1 was measured in cAMP-free buffer. The

decrease of the conductance of the MloK1-containing polymer membrane in the absence of cAMP indicates closure of the channel in the protein. 4. Conclusions We have introduced a new method for engineering functional surfaces by protein insertion into solid-supported polymer membranes. The controlled addition of Bio-Beads is used to destabilize a membrane protein and to act as a driving force for its insertion into the polymer membrane attached to various solid substrates. Membranes based on the PDMS-b-PMOXA diblock copolymer were prepared by LB and LS transfer techniques, which resulted in highly ordered, defect-free polymer films. The stability of the membrane was improved by covalent attachment of the first layer to an aminomodified substrate. An amino-linker was introduced as a spacer between the solid surface and membrane to preserve fluidity and to prevent the protein from denaturation. The polymer bilayer was smooth, homogeneous, and stable, as established by AFM, ellipsometry, and ATR-FTIR measurements. Measurements of the electrical properties of membranes showed that a potassium channel MloK1 was successfully inserted into polymer membrane. The advantage of this method for engineering functional surfaces is that it can be achieved on solid substrates of unrestricted size, which represents an advance for technological applications. In addition, biofunctional membranes prepared in this way mimic natural membranes, and by changing the membrane protein, different membrane functionalities can be achieved. Acknowledgments We thank the Swiss National Science Foundation and the University of Basel for financial support. We acknowledge Prof. Chris€nenberger from University of Basel for the access to the tian Scho electric measurements. JK thanks Dr. Wangyang Fu, Mathis Wipf, and Raphael Wagner for experimental assistance. We thank Dr. B.A. Goodman for useful discussions and reading the manuscript. Appendix A. Supplementary data

Fig. 9. Conductance measurements performed in buffer with and without cAMP.

Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.biomaterials.2014.05.043.

Please cite this article in press as: Kowal JŁ, et al., Functional surface engineering by nucleotide-modulated potassium channel insertion into polymer membranes attached to solid supports, Biomaterials (2014), http://dx.doi.org/10.1016/j.biomaterials.2014.05.043

J.Ł. Kowal et al. / Biomaterials xxx (2014) 1e9

References [1] Torchilin VP. Recent advances with liposomes as pharmaceutical carriers. Nat Rev Drug Discov 2005;4:145e60. [2] Taubert A, Napoli A, Meier W. Self-assembly of reactive amphiphilic block copolymers as mimetics for biological membranes. Curr Opin Chem Biol 2004;8:598e603. [3] Discher DE, Eisenberg A. Polymer vesicles. Science 2002;297:967e73. [4] Kita-Tokarczyk K, Grumelard J, Haefele T, Meier W. Block copolymer vesiclesdusing concepts from polymer chemistry to mimic biomembranes. Polymer 2005;46:3540e63. [5] Cabane E, Zhang X, Langowska K, Palivan C, Meier W. Stimuli-responsive polymers and their applications in nanomedicine. Biointerphases 2012;7: 1e27. [6] Tanner P, Baumann P, Enea R, Onaca O, Palivan C, Meier W. Polymeric vesicles: from drug carriers to nanoreactors and artificial organelles. Acc Chem Res 2011;44:1039e49. [7] Graff A, Fraysse-Ailhas C, Palivan CG, Grzelakowski M, Friedrich T, Vebert C, et al. Amphiphilic copolymer membranes promote NADH: ubiquinone oxidoreductase activity: towards an electron-transfer nanodevice. Macromol Chem Phys 2010;211:229e38. [8] Ries RS, Choi H, Blunck R, Bezanilla F, Heath JR. Black lipid membranes: visualizing the structure, dynamics, and substrate dependence of membranes. J Phys Chem B 2004;108:16040e9. [9] Wong D, Jeon TJ, Schmidt J. Single molecule measurements of channel proteins incorporated into biomimetic polymer membranes. Nanotechnology 2006;17:3710e7. [10] Sackmann E. Supported membranes: scientific and practical applications. Science 1996;271:43e8. [11] Bally M, Bailey K, Sugihara K, Grieshaber D, Voros J, Stadler B. Liposome and lipid bilayer arrays towards biosensing applications. Small 2010;6: 2481e97. [12] Reimhult E, Kumar K. Membrane biosensor platforms using nano- and microporous supports. Trends Biotechnol 2008;26:82e9. [13] Naumann R, Jonczyk A, Hampel C, Ringsdorf H, Knoll W, Bunjes N, et al. Coupling of proton translocation through ATPase incorporated into supported lipid bilayers to an electrochemical process. Bioelectrochem Bioener 1997;42: 241e7. [14] Naumann R, Schmidt EK, Jonczyk A, Fendler K, Kadenbach B, Liebermann T, et al. The peptide-tethered lipid membrane as a biomimetic system to incorporate cytochrome c oxidase in a functionally active form. Biosens Bioelectron 1999;14:651e62. €per I, Naumann R, Sinner E-K. Tethered bimolecular lipid mem[15] Knoll W, Ko branesda novel model membrane platform. Electrochim Acta 2008;53: 6680e9. €hnig F, Sackmann E. Impedance spectroscopy of porin [16] Gritsch S, Nollert P, Ja and gramicidin pores reconstituted into supported lipid bilayers on indiumetin-oxide electrodes. Langmuir 1998;14:3118e25. [17] Belegrinou S, Dorn J, Kreiter M, Kita-Tokarczyk K, Sinner E-K, Meier W. Biomimetic supported membranes from amphiphilic block copolymers. Soft Matter 2010;6:179e86. [18] Haefele T, Kita-Tokarczyk K, Meier W. Phase behavior of mixed Langmuir monolayers from amphiphilic block copolymers and an antimicrobial peptide. Langmuir 2006;22:1164e72. [19] Kita-Tokarczyk K, Itel F, Grzelakowski M, Egli S, Rossbach P, Meier W. Monolayer interactions between lipids and amphiphilic block copolymers. Langmuir 2009;25:9847e56. [20] Castellana ET, Cremer PS. Solid supported lipid bilayers: from biophysical studies to sensor design. Surf Sci Rep 2006;61:429e44. [21] Dorn J, Belegrinou S, Kreiter M, Sinner E-K, Meier W. Planar block copolymer membranes by vesicle spreading. Macromol Biosci 2011;11:514e25. [22] Zhang XY, Fu WY, Palivan CG, Meier W. Natural channel protein inserts and functions in a completely artificial, solid-supported bilayer membrane. Sci Rep-UK 2013;3:2196. [23] Nardin C, Meier WG. Polymerizable amphiphilic block copolymers: from nanostructured hydrogels to nanoreactors and ultrathin films. Chimia 2001;55:142e6. vy D. Reconstitution of membrane proteins into liposomes. In: [24] Rigaud J-L, Le Nejat D, editor. Methods in enzymology. Academic Press; 2003. pp. 65e86.

9

[25] Egli S, Nussbaumer MG, Balasubramanian V, Chami M, Bruns N, Palivan C, et al. Biocompatible functionalization of polymersome surfaces: a new approach to surface immobilization and cell targeting using polymersomes. J Am Chem Soc 2011;133:4476e83. [26] Dess DB, Martin JC. Readily accessible 12-I-5 oxidant for the conversion of primary and secondary alcohols to aldehydes and ketones. J Org Chem 1983;48:4155e6. [27] Kowal J, Chami M, Baumgartner P, Arheit M, Chiu P-L, Rangl M, et al. Ligandinduced structural changes in the cyclic nucleotide-modulated potassium channel MloK1. Nat Commun 2014;5:3106. [28] Wang H, Chen S, Li L, Jiang S. Improved method for the preparation of carboxylic acid and amine terminated self-assembled monolayers of alkanethiolates. Langmuir 2005;21:2633e6. €hmler J, Ploux L, Ball V, Anselme K, Ponche A. Necessity of a thorough [29] Bo characterization of functionalized silicon wafers before biointerface studies. J Phys Chem C 2011;115:11102e11. [30] Chiu P-L, Pagel MD, Evans J, Chou H-T, Zeng X, Gipson B, et al. The structure of the prokaryotic cyclic nucleotide-modulated potassium channel MloK1 at 16 Å resolution. Structure 2007;15:1053e64. [31] Kaupp UB, Seifert R. Molecular diversity of pacemaker ion channels. Annu Rev Physiol 2001;63:235e57. [32] Kaupp UB, Seifert R. Cyclic nucleotide-gated ion channels. Physiol Rev 2002;82:769e824. [33] Robinson RB. Hyperpolarization-activated cation currents: from molecules to physiological function. Annu Rev Physiol 2003;65:453e80. [34] Craven KB, Zagotta WN. CNG and HCN channels: two peas, one pod. Annu Rev Physiol 2006;68:375e401. [35] DeGrip WJ, VanOostrum J, Bovee-Geurts PHM. Selective detergent-extraction from mixed detergent/lipid/protein micelles, using cyclodextrin inclusion compounds: a novel generic approach for the preparation of proteoliposomes. Biochem J 1998;330:667e74. [36] Knol J, Sjollema K, Poolman B. Detergent-mediated reconstitution of membrane proteins. Biochemistry-U S 1998;37:16410e5. [37] Ritchie TK, Grinkova YV, Bayburt TH, Denisov IG, Zolnerciks JK, Atkins WM, et al. Chapter 11 Reconstitution of membrane proteins in phospholipid bilayer nanodiscs. In: Nejat D, editor. Methods in enzymology. Academic Press; 2009. pp. 211e31. [38] Wesemann A, Ahrens H, Steitz R, Forster S, Helm CA. Internal interface of a compressed PEE-PEO diblock copolymer monolayer. Langmuir 2003;19: 709e16. [39] Harkins WD. The physical chemistry of surface films. Reinhold; 1952. pp. 107e11. [40] Lee SW, Laibinis PE. Protein-resistant coatings for glass and metal oxide surfaces derived from oligo(ethylene glycol)-terminated alkyltrichlorosilanes. Biomaterials 1998;19:1669e75. [41] Ma H, Li D, Sheng X, Zhao B, Chilkoti A. Protein-resistant polymer coatings on silicon oxide by surface-initiated atom transfer radical polymerization. Langmuir 2006;22:3751e6. [42] Jiang XM, Wang LM, Wang J, Chen CY. Gold nanomaterials: preparation, chemical modification, biomedical applications and potential risk assessment. Appl Biochem Biotechnol 2012;166:1533e51. €msdorf A, Fo € rster S. Covalent attachment of [43] Domes S, Filiz V, Nitsche J, Fro polymersomes to surfaces. Langmuir 2010;26:6927e31. [44] Pasternack RM, Rivillon Amy S, Chabal YJ. Attachment of 3-(aminopropyl) triethoxysilane on silicon oxide surfaces: dependence on solution temperature. Langmuir 2008;24:12963e71. [45] Love JC, Estroff LA, Kriebel JK, Nuzzo RG, Whitesides GM. Self-assembled monolayers of thiolates on metals as a form of nanotechnology. Chem Rev 2005;105:1103e69. [46] Shenoy DK, Barger WR, Singh A, Panchal RG, Misakian M, Stanford VM, et al. Functional reconstitution of protein ion channels into planar polymerizable phospholipid membranes. Nano Lett 2005;5:1181e5. [47] Kuo MMC, Saimi Y, Kung C, Choe S. Patch clamp and phenotypic analyses of a prokaryotic cyclic nucleotide-gated Kþ channel using Escherichia coli as a host. J Biol Chem 2007;282:24294e301. [48] Mari SA, Pessoa J, Altieri S, Hensen U, Thomas L, Morais-Cabral JH, et al. Gating of the MlotiK1 potassium channel involves large rearrangements of the cyclic nucleotide-binding domains. Proc Natl Acad Sci U S A 2011;108:20802e7. [49] Nimigean CM, Shane T, Miller C. A cyclic nucleotide modulated prokaryotic Kþ channel. J Gen Physiol 2004;124:203e10.

Please cite this article in press as: Kowal JŁ, et al., Functional surface engineering by nucleotide-modulated potassium channel insertion into polymer membranes attached to solid supports, Biomaterials (2014), http://dx.doi.org/10.1016/j.biomaterials.2014.05.043