Antihuman IgG (#I9764), human IgG (#I4506), NHS (N-hydroxysuccinimide, #130672), EDC (1-ethyl-3-[3-dimethylaminopropyl]carbodiimide hydrochloride, #03450), ethanolamine (#398136), BSA (bovine serum albumin, #A7030), dansyl chloride (#39220), FITC (Fluorescein 5(6)-isothiocyanate, #46950), and TEMPO free radical [(2,2,6,6-tetramethyl-piperidin-1-yl)oxyl, #426369] were obtained from Sigma-Aldrich (Helsinki, Finland). The water used in all experiments was deionized and further purified with a Millipore Synergy UV unit (MilliQ-water). All other chemicals were used without any purification steps. An Epson R800 piezoelectric inkjet printer with a CD-printing tray was used without any modifications. Clean unused inkjet cartridges were obtained from MIS Associates, MI, USA.
Preparation of Nanofibrillar Cellulose (NFC) Films and NFC Model Surfaces
The NFC films used in this work were prepared as described elsewhere . In brief, bleached hardwood pulp (birch) was mechanically processed using a Masuko grinder with five passes and then further disintegrated by a M110P fluidizer (Microfluidics corp., Newton, MA, USA) with six passes. The produced NFC-gel was then filtered to remove the excess of water with a filter membrane using 2.5 bar. The film was then rolled five times with a smooth metal rolling pin to further consolidate the structure of the film. Finally the prepared films were dried between clean blotting boards (highly absorbent paper sheets) under external pressing and stored under ambient conditions until used.
NFC model surfaces (thin films) were prepared as described by Ahola et al. . In brief, bleached hardwood pulp (birch) was mechanically treated (five times with Masuko grinder) and then further disintegrated by microfluidization with 20 passes. The individual cellulose nanofibrils were then produced using mechanical stirring and tip ultrasonication (10 min, 25 % amplitude). The resulting NFC suspension was centrifuged at 10,400 rpm for 45 min and colloidal nanofibrils were then collected from the supernatant by pipetting. The collected nanofibrils (0.148 wt% NFC in water) were then spin-coated (Model WS-650SX-6NPP, Laurell Technologies, PA, USA) at 3,000 rpm and 90 s spinning time on QCM SiO2-crystals carrying a thin layer of PEI. The NFC-coated QCM crystals were stored in a desiccator and prior to use in QCM-D measurements they were stabilized overnight in water.
Activation of the NFC Films via TEMPO-Mediated Oxidation and EDC/NHS Treatment
NFC-films were oxidized by using the 2,2,6,6,-tetramethylpipelidine-1-oxyl radical (TEMPO) NaBr–NaClO system as described by Isogai et al. . 0.13 mmol TEMPO and 4.7 mmol NaBr were dissolved in 100 mL water. Then 5.65 mmol NaClO was added in the solution, and the pH was adjusted to 10 by adding 1 M HCl. Next, NFC-film (size 2.5 × 2.5 cm2) was placed in a glass petri dish in which the prepared TEMPO-solution was added. NFC-films were kept in the solution for times varying from 10 to 300 s, and the oxidation reaction was quenched by adding ethanol and washing with a large amount of water to remove the excess of carboxylation chemicals. Finally, carboxylated NFC-films were made amine-reactive via EDC/NHS activation as follows: carboxylated NFC-film was kept in a solution of 0.1 M EDC and 0.4 M NHS (10 mM NAOAc buffer, pH 5) for 20 min, and then rinsed with water. Activated NFC-film was then dried between clean blotting boards to prevent film buckling.
QCM-D In-situ Monitoring of TEMPO-Mediated Oxidation and EDC/NHS Activation of NFC Model Surfaces
Interaction and analyses of the activation of the NFC thin films were carried out using a QCM-D E4 instrument (Biolin Scientific AB, Gothenburg, Sweden) with controlled flow [16, 36]. All measurements were performed using a constant 100 μL/min flow rate, 25 °C temperature, and all experiments were at least performed in duplicate. First, TEMPO-mediated oxidation reaction of NFC thin films was investigated by using QCM-D. To this end, TEMPO-solution (0.13 mmol TEMPO, 4.7 mmol NaBr, and 5.65 mmol NaOCl in water with fixed pH of 10) was allowed to flow over the QCM sensors carrying the NFC thin films for 2 min. The oxidation reaction was then quenched by adding ethanol in the oxidation solution (10 % v/v), and the constant flow was continued for 2 min to stop the carboxylation reaction. Next, the carboxylated NFC thin films were rinsed with water for 20 min to remove excess carboxylation chemicals. Finally, the surfaces were stabilized in 10 mM NaOAc buffer at pH 5. The carboxylated NFC surfaces were made amine reactive by flowing a solution of 0.1 M EDC with 0.4 M NHS (10 mM NaOAc buffer, pH 5) for 20 min. The excess of chemicals were then rinsed out with 10 mM NaOAc buffer. The activated amine NFC thin films were finally dried using a nitrogen gas, and stored in a desiccator, if not immediately used.
The measured shift in frequency obtained in QCM experiments was fitted to the Johannsmann’s model  to quantify changes in mass of the QCM sensors carrying the NFC thin films. The iterative process in the model used the third, fifth, and seventh frequency overtones of the QCM-D crystal.
Conjugation of BSA and Antihuman IgG on Activated NFC Surfaces Monitored In-situ by QCM-D
The reactivity of activated NFC thin films was verified using QCM-D as follows. The pre-activated NFC surfaces were first stabilized in 10 mM NaOAc buffer at pH 5 for 1 h, and then 100 μg/mL of BSA solution (10 mM NaOAc buffer, pH 5) was introduced for 30 min. The pH conditions were selected, because BSA adsorption is maximized at its isoelectric point (pH 5). The conjugation was confirmed by a sequential rinsing sequence NaOAc buffer (pH 5), 10 mM NaCl (pH 10), and again with NaOAc buffer (pH 5). It should be noted that the rinsing with alkaline buffer (10 mM NaCl, pH 10) resulted in an electrostatic repulsion between BSA and the activated NFC substrate which in turn removed electrostatically bound BSA. Two reference experiments were performed identically by using unmodified and TEMPO-oxidized NFC surfaces; this was done in order to rule out the contribution of interactions between BSA and the substrate different than covalent bonding.
The conjugation of antihuman IgG on activated NFC surfaces was also verified using QCM-D by flowing 100 μg/mL of antihuman IgG solution (10 mM NaOAc, pH 5) for 30 min followed by buffer rinsing (10 mM NaOAc, pH 5). The excess of NHS-esters was then removed by rinsing with 0.1 M ethanolamine solution at pH 8.5 for 15 min. Non-specific binding of proteins was prevented by using a 15 min Superblock treatment. Superblock is a protein solution that fills the free spaces between the antibodies, thus preventing the non-specific binding of antigens on the surface of biointerface. The detection sensitivity of prepared biointerfaces with conjugated antihuman IgG was verified by adsorbing 10 μg/mL hIgG solution (10 mM phosphate buffer, pH 7.4) for 10 min. As a reference NFC films without antihuman IgG were tested. In addition, two reference experiments were performed identically by adsorbing antihuman IgG and human IgG on unmodified and TEMPO-oxidized NFC surfaces; this was done in order to rule out the contribution of interactions between antihuman IgG and the substrate different than covalent bonding as well as to explore the non-specific binding of proteins.
Inkjet Printing and Physical Adsorption of Antihuman IgG on Activated NFC-Films
Inkjet printing of antibodies on activated NFC-films was demonstrated using EPSON R800 inkjet printer. Fluorescence (dansyl or FITC-probe)-stained antihuman IgG (1 mg/mL) in 10 mM phosphate buffer (pH 7.4) was printed on the activated NFC films (size 2 × 4 cm2) using a CD printing tray with an EPSON print-CD software. The fluorescence staining of antihuman IgG was done following the procedures described elsewhere [37, 13]. Stained antibody solutions were purified with Amicon Ultra centrifugal filter tubes (Mw cutoff of 30 kD) using three cycles of 10 mM phosphate buffer (pH 7.4) and a total rinsing volume six times larger than the original sample volume. The stained antihuman IgG was diluted to the final concentration using 10 mM phosphate buffer (pH 7.4) followed by an immediate printing on activated NFC films. The printing conditions were selected using a CMYK color profile which allows printing from a selected cartridge. All other cartridges were filled with 10 mM phosphate buffer (pH 7.4). Printed NFC films were then treated with 0.1 M ethanolamine at pH 8.5 for 15 min, and dried at ambient conditions. The fluorescence of NFC films was investigated using UV-light (366 nm, Camag, Berlin, Germany), CLSM and AFM.
The physical adsorption of antihuman IgG on the activated NFC films was also examined. FITC-stained or unstained antihuman IgG (100 μg/mL) in 10 mM phosphate buffer (pH 7.4) was adsorbed on unmodified and activated NFC for 20 min. Corresponding films were then treated with 0.1 M ethanolamine at pH 8.5 for 15 min, rinsed with 10 mM phosphate buffer (pH 7.4), and 10 mM NaCl (pH 10) to remove electrostatically bound antibodies. Finally, the surfaces were rinsed with Milli-Q water and dried at ambient conditions. Adsorbed FITC-stained antihuman IgG on unmodified and activated NFC films was imaged using CLSM.
Contact Angle Measurements
Changes in the surface hydrophilicity of TEMPO-oxidized NFC films were monitored using a contact angle goniometer CAM 200 (KSV instruments Ltd, Helsinki, Finland). Measurements were performed at room temperature with water as a probe liquid. A droplet volume of 6.5 μL and a recording time of 120 s were used to measure the time dependency of the contact angle. Contact angles were measured on three different locations on each sample.
The increase in negatively charged groups (carboxyls) in TEMPO-oxidized NFC-films was measured using a conductometric titrator 751 GPD Titrino (Metrohm AG, Herisau, Switzerland) following SCAN-CM 65:02 standard method. NFC-films (size 2.5 × 2.5 cm2) were acid washed with 0.01 M HCl for 1 h, and then disintegrated in water with a blade type homogenizer Polytron PT 2000 (Kinematica Inc., NY, USA). The conductometric titration was performed by adding 0.02 mL of 0.1 M NaOH using 30 s intervals. The amount of weak acid (carboxyl) groups was calculated as described in the standard method (SCAN-CM 65:02).
X-Ray Photoelectron Spectroscopy
The surface chemistry of the topmost 10 nm of the NFC films was examined using a Kratos Analytical AXIS 165 electron spectrometer with a monochromatic Al Kα X-ray source at 100 W and a neutralizer. The XPS experiments were performed on the dry films, which were pre-evacuated overnight. At least three different spots of each sample were scanned. Spectra were collected at an electron take-off angle of 90° from sample areas less than 1 mm in diameter. Elemental surface compositions were determined from low-resolution measurements (80 eV pass energy and 1 eV step), while the surface chemistry was probed with high resolution measurements (20 eV pass energy and 0.1 eV step). The carbon C1s high-resolution spectra were curve fitted using parameters defined for cellulosic materials  and all binding energies were referenced to the aliphatic carbon component of the C1s signal at 285.0 eV . According to the in situ reference (100 % cellulose ash free filter paper), measured along with each sample batch, the conditions in UHV remained satisfactory during the XPS experiments .
Macroscale and Nanoscale Topography: CLSM and AFM
The macro-scale topography of NFC films was analyzed with a Leica TCS SP2 CLSM (Leica microsystems CMS GmbH, Manheim, Germany). The image was obtained using a reflection image mode with an excitation wavelength of 488 nm and a detection wavelength range of 490–550 nm. The image size was 750 × 750 μm2. The intensity images were scanned using an averaging mode and constant imaging conditions (laser power was 848 V in all measurements). Intensity of the fluorescence images was measured using analysis tool of Photoshop (Adobe). The intensity was measured from the unmodified raw images. 3D-image was obtained from 60 optical sections of the NFC film by using topography. No sample pretreatment was done except placing the sample between two clean microscopy glasses. The roughness profiles of the NFC films were calculated using the obtained topological 3-D image together with Leica confocal microscope software.
Nano-scale topological changes on the NFC films were investigated using Nanoscope IIIa Multimode scanning probe microscope (Digital Instruments, Inc., Santa Barbara, CA). The images were scanned using the silicon cantilevers (Ultrasharp μmasch, Tallinn, Estonia). At least three different locations on each sample were scanned with image sizes of 5 × 5 and 1 × 1 μm2. No image processing was done except image flattening. Roughness profiles were calculated using 5 × 5 μm2 images.