“Hearing Loss” in QCM Measurement of Protein Adsorption to Protein

Mar 9, 2017 - Accurate quantification of nonspecific protein adsorption on biomaterial surfaces is essential for evaluation of their antifouling prope...
0 downloads 8 Views 5MB Size
Download PDF
Article pubs.acs.org/ac

“Hearing Loss” in QCM Measurement of Protein Adsorption to Protein Resistant Polymer Brush Layers Yafei Luan,† Dan Li,*,† Ting Wei,† Mengmeng Wang,† Zengchao Tang,† John L. Brash,†,‡ and Hong Chen*,† †

State and Local Joint Engineering Laboratory for Novel Functional Polymeric Materials, College of Chemistry, Chemical Engineering and Materials Science, Soochow University, 199 Ren’ai Road, Suzhou 215123, People’s Republic of China ‡ School of Biomedical Engineering and Department of Chemical Engineering, McMaster University, Hamilton, Ontario, Canada S Supporting Information *

ABSTRACT: Accurate quantification of nonspecific protein adsorption on biomaterial surfaces is essential for evaluation of their antifouling properties. The quartz crystal microbalance (QCM) is an acoustic sensor widely used for the measurement of protein adsorption. However, although the QCM is highly sensitive, it does have performance limitations when working with surfaces modified with thick viscous layers. In the case of polymer brush surfaces, factors such as the thickness and viscosity of the brush may bring such limitations. In the present work, three types of antifouling molecules were used to explore the applicability of QCM for the evaluation of the protein resistance of hydrophilic polymer brush surfaces. Adsorption was also measured by surface plasmon resonance (SPR) as a reference. It was shown that the detection of adsorbed protein requires that protein be located within a critical distance from the QCM chip surface, determined by the viscosity of polymer brush. For larger proteins like fibrinogen, adsorption is expected to occur mainly “on top” of the polymer brush, and brush thickness determines whether protein is located in the “detectable zone”. For smaller proteins like lysozyme, adsorption is expected to occur mainly at the chip surface and within the polymer brush layer and to be detectable by QCM. However, the quantity of adsorbed lysozyme may be underestimated when secondary adsorption also occurred. It is concluded that QCM data suggesting very low protein adsorption on polymer brush surfaces should take account of these considerations and should be treated generally with caution. showed that only ∼20 ng/cm2 of adsorbed fibrinogen (∼3% of monolayer coverage) was sufficient to support platelet adhesion and spreading on methyl chlorosilane surfaces.18 It is possible that even lower adsorbed quantities are problematic. Thus, methods with extra-low detection limits are required for the accurate quantification of nonspecific protein adsorption. The quartz crystal microbalance (QCM), with a detection limit of 1.8 ng/cm2, is an acoustic sensor commonly used to quantify low level protein adsorption.19,20 In the QCM measurement, a shear deformation oscillation is excited in a piezoelectric crystal by an alternating current electric field. Adsorption on the surface of the crystal decreases the resonant frequency of the oscillation in a manner related to the mass of adsorbate. Therefore, by measuring the frequency shift, the mass of protein adsorption can be determined.21 Many researchers have used the QCM technique to show that some hydrophilic polymer-modified surfaces have good protein resistance, and several claims of ultralow or “zero” adsorption have been made.22−24 However, unlike protein adsorption occurring directly on the QCM chip surface or on a

P

rotein−surface interactions are well-known to play an important role in various biological phenomena and to determine the biofunctionality of materials used in tissue engineering,1,2 biosensors,3,4 and blood contacting devices.5,6 Nonspecific protein adsorption is undesirable and is generally the first in a series of interactions leading to device failure. For example, when a biomaterial comes in contact with blood it rapidly becomes covered with a layer of proteins followed by initiation of blood coagulation, complement activation, inflammatory responses, and platelet activation.7−9 To prevent or reduce these effects, biomaterial surfaces may be chemically treated to minimize nonspecific protein adsorption.10−12 Hydrophilic, bioinert polymers are commonly used as surface modifiers to create protein resistant surfaces.13 The water of hydration associated15 with such polymer brush surfaces is believed to act as a barrier to approaching proteins. The flexible polymer chains may also resist protein adsorption by steric exclusion mechanisms.14,15 Although hydrophilic polymers such as polyethylene glycol (PEG)16 and certain zwitterionic polymers,17 have been demonstrated to be strongly protein-resistant, no surface has yet been designed showing “zero” adsorption. However, even extremely small quantities of adsorbed proteins may cause significant adverse biological responses. For example, Park et al. © XXXX American Chemical Society

Received: January 17, 2017 Accepted: March 9, 2017 Published: March 9, 2017 A

DOI: 10.1021/acs.analchem.7b00198 Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry

described,26 Figure S1) in ethanol at a concentration of 1.0 mM for 24 h. The initiator-functionalized chips were rinsed with ethanol and then dried under nitrogen. For surface polymerization of OEGMA by SI-ATRP, a reaction solution, consisting of OEGMA (0.95 g, 2 mmol), dipyridyl (5 mg, 0.032 mmol), CuCl2 (2.72 mg, 0.016 mmol), ascorbic acid (2.82 mg, 0.016 mmol), methanol (2 mL), and H2O (2 mL), was deoxygenated by ventilating under argon for 30 min. Then the chips were immersed into the prepared solution and the polymerization proceeded for a designated time in a glovebox purged with dry N2 gas at room temperature. The reaction was stopped by rinsing the samples with deionized water and methanol thoroughly. The SI-ATRP of HEMA was carried out following the same procedures as for OEGMA. Characterization of Modified Surfaces. Static water contact angles were measured by the sessile drop method with an SL200C optical contact angle meter (Solon Information Technology Co., Ltd.). Fourier transform infrared spectra (FTIR) of the polymer brushes were obtained on a Nicolet 6700 FTIR spectrometer (U.S. Thermo Fisher Scientific Inc.). The dry thickness of the polymer brushes on the chips was measured by ellipsometry (M-88, J. A. Woollam Co., Inc.) using a detection angle of 70°. The analysis software provided by the instrument manufacturer based on the Cauchy-Extended model was employed to estimate the refractive indices of the modified surfaces. The wet thickness of the polymer brushes in PBS were studied with a Multi-Mode Nanoscope 8 atomic force microscope (AFM, Bruker Co., Santa Barbara, CA) with an NP-10 nonconductive silicon nitride tip (Bruker, nominal spring constant 0.58 N/m, resonance frequency 40−75 kHz) in quantitative nanomechanical mapping mode. The scanning areas were 20 × 20 μm. Before measurements, the sensitivity of the cantilever was calibrated on a hard surface. The thicknesses were determined by step-height analysis. In the step-height measurement, a scratch was made on the film, and then the AFM probe was carefully located at the edge of the scratch before scanning. After the thickness measurement in air, PBS was added to the surface and an O-ring was used to retain the liquid in place. The system was kept still for 0.5 h prior to scanning. Protein Adsorption Measured by SPR. Protein adsorption measurements were carried out on a Reichert SR7000DC SPR instrument. The optical parameters and SPR sensor chips were the same as those described previously.18 Protein adsorption was quantified by measuring the change in response units (RU) after the injection of protein. An initial baseline was established by flowing PBS (pH = 7.4) through the instrument for about 5 min. Freshly prepared fibrinogen or lysozyme solution (0.1 mg/mL in PBS) was then flowed through the system for about 37 min, followed by PBS buffer to establish a final baseline. Measurements were carried out at a temperature of 25 °C and a flow rate of 10 μL/min. The RU change was converted to mass of adsorbed protein using the calibration factor 1 RU ≡ 1 pg/mm2 adsorbed protein.27 Protein Adsorption Measured by QCM-D. QCM measurements were conducted on a Q-Sense-E4 instrument (Q-Sense, Sweden) with dissipation (QCM-D, Resonant Probes GmbH, Goslar, Germany). QCM-D chips (AT cut, 5 MHz, 14 mm diameter) modified with SAM-OEG or polymer brush as described above were placed in the fluid chambers. An initial baseline was established by flowing PBS for 30 min. Freshly prepared fibrinogen or lysozyme solution (0.1 mg/mL in PBS) was flowed for 3 h to attain the adsorption plateau.

chip surface modified with a thin, rigid layer of small molecules, adsorption on thicker polymer brush layers may occur at a significant distance from the chip surface, for example, on the top or in the interior of the brush layer. It should be noted that the distance of separation of adsorbed proteins from the base chip surface may influence the measurement of adsorption based on the working principle of QCM. In previous work we showed that, on polymer brush surfaces, protein adsorption, as measured by QCM, was near zero, whereas significant adsorption was detected by SPR and 125I-radiolabeling. These observations raise questions regarding protein adsorption data for such surfaces, as obtained by QCM. In the present study, factors including the thickness and viscosity of the polymer brush and the amount and size of the adsorbed protein were investigated. Adsorption of proteins was measured on QCM chip surfaces modified with poly(hydroxyethyl methacrylate) (PHEMA), poly(oligoethylene glycol methacrylate) (POEGMA), and self-assembled monolayers (SAM) with terminal oligoethylene glycol sequences (SAM-OEG). Adsorption was also measured by surface plasmon resonance (SPR). SPR measures the dry protein, whereas QCM measures the hydrated protein quantity. Moreover, SPR with a sampling depth of about 200 nm is not influenced by the distance of the adsorbate from the base surface in the range of distances encountered in the polymer brushes.

■

EXPERIMENTAL SECTION Materials. Triethylene glycol mono-11-mercaptoundecyl ether (TEGMUE, 95%) and 2,2-bipyridyl were from SigmaAldrich Chemical Co. and used without further purification. LAscorbic acid, copper(II) chloride dihydrate, acetone, ethanol, and all other solvents were from Shanghai Chemical Reagent Co. and used as received. α-Bromoisobutyryl bromide (BIBB, 98%) from Fluka was used as received. Poly(ethylene glycol) methyl ether methacrylate was from Sigma-Aldrich. Deionized water purified with a Millipore water purification system to give a minimum resistivity of 18.2 MΩ·cm was used in all experiments. Nitrogen gas was of high-purity grade. Fibrinogen (Mw = 341 kDa, pI = 5.5, plasminogen-free) was from Calbiochem (La Jolla, CA, U.S.A.). Lysozyme (Lyz, Mw = 14 kDa, pI = 11.1) was from Beijing Solarbio Science and Technology Co., Ltd. Preparation of Self-Assembled Monolayers (SAMs). QCM and SPR sensor chips were washed as described briefly. QCM chips were washed with acetone, treated for 30 min in an ozone plasma, washed with deionized water and ethanol, cleaned in a mixture of ammonia, hydrogen peroxide, and deionized water (NH3·H2O/H2O2/H2O = 1:1:5 v/v/v) for 10 min at 75 °C, and then rinsed with deionized water and dried under nitrogen. SPR chips were treated similarly but without the ammonia, hydrogen peroxide, and deionized water cleaning step. SAMs were formed on clean gold chips by immersion in solutions of TEGMUE in ethanol at a total concentration of 1.0 mM for 24 h. The SAM-OEG modified chips were rinsed with ethanol to remove physisorbed thiol and then dried with nitrogen. Preparation of POEGMA and PHEMA Polymer Brush Layers. POEGMA and PHEMA polymer brush layers with different thickness were prepared by surface-initiated atom transfer radical polymerization (SI-ATRP), as reported previously.25 Briefly, the clean chips were immersed in solution of ω-mercaptoundecyl bromoisobutyrate (synthesized as B

DOI: 10.1021/acs.analchem.7b00198 Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry PBS was then flowed through the system to remove loosely attached protein and establish a final stable baseline. The viscosity of polymer brushes was measured using QCM-D as reported previously.28 Briefly, the baseline was established by flowing air for 30 min. PBS were flowed for 1 h to infiltrate polymer brush until saturation. All QCM-D experiments were carried out at a temperature of 25 °C and a flow rate of 10 μL/ min. The differences in frequency, ΔF, and dissipation factor, ΔD, between the two baselines were attributed to protein adsorption. The mass of protein adsorbed was obtained by fitting ΔF and ΔD to the Voigt model21 using Q-tools software package v.3.0.15.553 (Biolin Sci, AB). Specific input parameters that provided the best data fit were layer density range 1000− 1250 kg·m−3; fluid density 1000 kg·m−3; layer viscosity range 10−4 to 10−2 kg·ms−1; layer shear modulus range 102−109 Pa; and layer thickness range 10−10 to 10−6 m. The third, fifth, and seventh overtones were used for all of the data fits.

Successful grafting of PHEMA and POEGMA brushes on gold surface was confirmed by small angle grazing angle Fourier transform infrared (SAGA-FTIR) spectra (Figure S2). Due to the mechanism of ATRP, polymer brush layers of designated thickness can be obtained by varying the polymerization time. Thickness data for the SAM-OEG layer and the polymer brush layers determined by ellipsometry are shown in Table 1. It can be seen that the thickness of the SAM-OEG layer was ∼1.7 nm. POEGMA surfaces were prepared with thicknesses of ∼10 and 20 nm, and PHEMA surfaces were prepared with thicknesses of ∼20, 40, and 90 nm. Protein Adsorption: SAM versus polymer brushes. SPR and QCM-D were used to quantify the adsorption of fibrinogen on gold surfaces modified with POEGMA brush layers and SAM-OEG. As shown in Figure 1a, strong SPR signals were obtained and adsorption on the modified surfaces was readily measurable even for adsorbed quantities of 0.9 and 4.7 ng/cm2 on SAM-OEG and POEGMA-20 nm, respectively. These low levels indicate excellent protein resistance, presumably due to the PEG-like structures of SAM-OEG and POEGMA. In the QCM-D experiments (Figure 1b), a change in frequency (ΔF) of ∼2.1 Hz was seen on the SAM-OEG surface, and adsorption was accompanied by a change in dissipation (ΔD, 7.5 × 10−7, Figure S3) due to the “soft” nature of the protein. Hence, the Voigt model was used to fit ΔF and ΔD, and the adsorbed quantity was estimated to be 19.5 ng/ cm2. ΔF for the POEGMA-20 nm surface was less than 1 Hz and ΔD was negligible. A ΔF of this magnitude is generally considered as “background” or “noise”. In general, the mass of adsorbed protein measured by QCM is taken as the “wet mass”, equal to the mass of adsorbed protein plus associated water, while the mass of adsorbed protein measured by the optically based SPR method is considered as the “dry mass”.29 Therefore, for a given sample, the mass measured by QCMD should be greater than that measured by SPR. From the data in Figure 1b, the adsorption of fibrinogen measured by QCM-D on the SAM-OEG surface was much greater than by SPR: ∼19.5 versus ∼0.93 ng/cm2. In contrast, adsorption on the POEGMA-modified surface was undetectable by QCM-D measurement whereas the SPR signal on this surface was stronger than that on the SAM-OEG. Based on the working principle of QCM, it seemed possible that the apparent failure to detect protein adsorption on the POEGMA surface might be due to the particular physical

■

RESULTS AND DISCUSSION SAM-OEG, PHEMA, and POEGMA are representative of three types of widely used antifouling surface modifiers and are easily grafted to surfaces in a well-controlled manner. In this work, SAM-OEG, PHEMA, and POEGMA were grafted to gold to obtain hydrophilic, antifouling surfaces. Surface Characterization. As shown in Table 1, the water contact angle decreased from 68° on unmodified gold to 34°, Table 1. Water Contact Angles and Layer Thicknesses (Dry State) of SAM-OEG, POEGMA, and PHEMA Polymer Brush Layers on Golda sample

water contact angle (deg)

thickness (nm)

± ± ± ± ± ±

1.7 ± 0.4 20.1 ± 1.4 22.4 ± 0.9 40.1 ± 0.4 90.7 ± 1.5

Au SAM-OEG POEGMA-20 nm PHEMA-20 nm PHEMA-40 nm PHEMA-90 nm a

77 37 36 44 43 43

2 1 2 1 1 2

Data are mean ± SD, n = 3.

37°, and 42°, respectively, after modification with SAM-OEG, POEGMA, and PHEMA polymer brushes. These decreases are likely due to the high content of hydroxyl groups or -CH2CH2O- residues on the modified surfaces.

Figure 1. Fibrinogen adsorption (0.1 mg/mL) on SAM-OEG layers and POEGMA polymer brushes of 20 nm thickness measured by (a) SPR and (b) QCM-D. Three independent experiments were conducted in each case. C

DOI: 10.1021/acs.analchem.7b00198 Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry

Figure 2. Investigation of the wet thickness of POEGMA and PHEMA brushes using the AFM step-height method and the viscosity of the two polymer brushes using QCM-D: AFM images of scratch on POEGMA-20 nm (a) and on PHEMA-20 nm (b) in PBS; step height of POEGMA-20 nm (c) and PHEMA-20 nm (d) measured on (a) and (b), respectively; viscosity of POEGMA-20 nm (e) and PHEMA-20 nm (f) estimated from Voigt model as a function of time during PBS buffer infiltration.

cannot cause a frequency shift in the base chip directly; rather, the adsorbed mass effect on frequency is transmitted through the polymer brush and may be modulated due to the viscoelastic character of the polymer layer, which may, in turn, affect the ability of the QCM to detect adsorbed protein. Factors such as the thickness and viscosity of the polymer brush, the size of the protein, and the adsorbed mass may play a role. Investigations of these factors are now reported. Influence of Polymer Brush Viscosity on Detection of Adsorbed Protein by QCM-D. To investigate the influence of polymer brush viscosity on the detection of adsorbed protein by QCM-D, two hydrophilic polymer brushes (POEGMA-20 nm and PHEMA-20 nm) were compared. The wet thickness of the two polymer brushes was determined by AFM using the scratch-step-height method. As shown in Figure 2, POEGMA20 nm and PHEMA-20 nm brushes have similar hydrated states with wet thickness of 35 and 34 nm, respectively. The two

properties of the polymer brush layer. In QCM measurements, a shear deformation oscillation is excited in the piezoelectric crystal by an alternating current electric field. A small quantity of mass deposited on the surface of the crystal causes a decrease in the resonant frequency of the oscillation, and by measuring the frequency shift, the mass of adsorbed protein can be determined. The SAM-OEG layer is thinner and thus is expected to be more rigid than the polymer brush layers (Table 1). Adsorption on the SAM-OEG is close to the situation of adsorption on the QCM chip surface, and in this situation the sensitivity is sufficient that ultralow amounts of adsorbed proteins are detected. For a relatively thick and dense polymer brush layer, fibrinogen, as a large protein (Mw 341 kDa, 450 × 90 × 90 Å3), is expected to adsorb mainly “on top” of the layer (termed secondary adsorption). In this situation, the protein and the chip are separated by a relatively soft polymer layer of non-negligible thickness. Therefore, the adsorbed fibrinogen D

DOI: 10.1021/acs.analchem.7b00198 Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry

Figure 3. Fibrinogen adsorption on POEGMA and PHEMA polymer brushes of 20 nm thickness measured by (a) SPR and (b) QCM-D. Three independent experiments were conducted in each case.

Figure 4. (a) Schematic illustration of the attenuation of mass effect through the polymer brush. Fn and Fn′ are the frequencies of oscillation of each sublayer before and after protein adsorption, respectively. (b) Schematic illustration of water liberation from polymer brush upon protein adsorption.

example, in response to shear stress (Figure 4a). Friction between sublayers decreases, that is, slipping increases as viscosity decreases. As protein is adsorbed on top of the polymer brush, the oscillation of the topmost sublayer will be modulated “first”. The effect will be sensed ultimately at the chip surface after transmission through the layers and will be gradually attenuated during transmission due to slipping between the layers. If the viscosity is low enough, the effect of the adsorbed protein mass may be completely eliminated before the chip surface is reached. The POEGMA brush with smaller viscosity may dissipate most of the mass effect over the 35 nm (wet) transmission distance through the layer so that the chip cannot sense the adsorbed protein. Influence of Polymer Brush Thickness on QCM-D Sensitivity. Assuming that the mass effect of the adsorbate is attenuated by the polymer brush, it is expected that the position of the protein within the brush, specifically the distance of the protein from the chip surface, should influence the magnitude of attenuation. Fibrinogen is a large protein that is expected to adsorb mainly on top of the polymer brush. In this case, the distance of the protein from the chip surface is equal to the

polymer brushes were grown from the same initiator-functionalized surface and, thus, can be considered to have the same graft density. Therefore, the effect of the difference in thickness and polymer density can be excluded in this case. Viscosity estimates for the POEGMA-20 nm and PHEMA-20 nm brushes in the hydrated state were obtained via Voigt model calculations from data on the variation of frequency and dissipation factor upon the infiltration of PBS buffer into the dry polymer brushes.30 As shown in Figure 2, the viscosity of POEGMA-20 nm (3.4 × 10−3 kg/ms) was smaller than that of PHEMA-20 nm (3.8 × 10−3 kg/ms), presumably due to the greater mobility and degree of hydration of POEGMA compared to PHEMA. From the data in Figure 3, the masses of fibrinogen adsorbed on POEGMA-20 nm and PHEMA-20 nm measured by SPR were 4.16 and 31.2 ng/cm2, respectively. In contrast, as measured by QCM-D, adsorption on PHEMA20 nm was 567.2 ng/cm2, while adsorption on POEGMA-20 nm was apparently zero. It is suggested that this discrepancy may be due to viscosity effects. The viscous polymer brush layer may be viewed as a system of many sublayers that can slip with respect to one another, for E

DOI: 10.1021/acs.analchem.7b00198 Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry

Figure 5. Fibrinogen adsorption on PHEMA polymer brushes of approximate thickness 20, 40, and 90 nm measured by (a) SPR and (b) QCM-D. Three independent experiments were conducted in each case.

Figure 6. Fibrinogen adsorption on POEGMA-NHS-20 nm and PHEMA-NHS-90 nm surfaces measured by (a) SPR and (b) QCM-D. Three independent experiments were conducted in each case.

brush thickness. An investigation of the influence of brush thickness on the mass of adsorbed fibrinogen measured by QCM-D was carried out using PHEMA brush surfaces of varying thickness. Figure 5a shows fibrinogen adsorption data obtained using SPR. It is seen that the adsorbed mass decreased with increasing thickness of the brush, reflecting the strong protein resistance of PHEMA as has been found by others.31 It has also been reported that beyond a certain “limiting” thickness, adsorption does not decrease further.32 A similar effect is seen in Figure 5a. The quantities adsorbed to the PHEMA-40 nm and PHEMA-90 nm surfaces were 9.64 and 6.15 ng/cm2, respectively, sufficiently close to suggest that the maximum or “limiting” reduction in fibrinogen adsorption as a function of PHEMA thickness was achieved at a thickness of ∼40 nm. Adsorption data from QCM-D measurements are shown in Figure 5b. For the PHEMA-20 nm and PHEMA-40 nm surfaces the trend was similar to that from SPR in that adsorption was much lower on the PHEMA-20 nm. However, the adsorbed quantities on the PHEMA-40 nm and PHEMA-90 nm surfaces were different, with zero adsorption on the PHEMA-90 nm surface. This contradicts the SPR results, which showed finite adsorption on both surfaces and essentially no difference between them. It is suggested that this discrepancy is

due to the difference in attenuation of the effect of the adsorbed mass for the two PHEMA brushes of different thickness. With fibrinogen assumed to be adsorbed on top of the brush, the distance of 90 nm through the brush to the chip surface may be beyond the limit where any effect of adsorbed mass can be transmitted to the chip, suggesting that adsorption on top of a polymer brush cannot be sensed by QCM-D if the brush thickness is beyond a certain critical value. This critical thickness should decrease as the viscosity of the polymer brush decreases. The failure to detect protein adsorption on the longer polymer brushes also might be due to another characteristic of QCM measurement: water infiltration and release can also cause frequency fluctuation. Protein adsorption may cause the polymer brushes to collapse and thereby release water (Figure 4b). Longer polymer brushes have more freedom to change conformation and release water upon the addition of protein. If the mass of water liberated is equal to or larger than the mass of protein adsorbed, QCM would show no frequency shift or even a positive frequency shift. This explains why the frequency curve of PHEMA-90 nm goes a little above the baseline upon fibrinogen adsorption (Figure 5b). Influence of the Mass of Adsorbed Protein on Quantification by QCM-D. According to the analysis given F

DOI: 10.1021/acs.analchem.7b00198 Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry

Figure 7. Lysozyme adsorption on PHEMA brush layers of 20, 40, and 90 nm thickness measured by (a) SPR and (b) QCM-D. Three independent experiments were conducted in each case.

protein, the more likely it is to be able to penetrate the brush and adsorb directly on the substrate (e.g., the QCM chip surface), termed primary adsorption, or within the layer, termed ternary adsorption. Adsorption “on top” of the layer, termed secondary adsorption, is more likely to occur for larger proteins.34 To investigate more closely the influence of protein dimensions on QCM measurements, the adsorption of lysozyme (Lyz), a small protein (Mw = 14 kDa, 30 × 30 × 40 Å3), on PHEMA brush layers of varying thickness was measured as described above. As shown in Figure 7a, the quantities of lysozyme adsorbed, as measured by SPR, were all quite low, and as observed for fibrinogen, adsorption decreased with increasing brush thickness. The quantities adsorbed from 0.1 mg/mL lysozyme solution were 13.45 ng/cm2 on PHEMA-20 nm, 3.26 ng/cm2 on PHEMA-40 nm, and 3.94 ng/cm2 on PHEMA-90 nm, the latter two being very similar. Adsorption of lysosome could be detected by QCM-D on all three PHEMA surfaces, including PHEMA-90 nm (Figure 7b), and the layer thickness dependence was the same as for SPR. As shown above, exposure of the PHEMA-90 nm surface to fibrinogen solution did not give rise to any frequency change in QCM-D, although as measured by SPR fibrinogen adsorption (6.15 ng/cm2) was greater than lysozyme adsorption (3.94 ng/ cm2). It is expected that lysozyme can access the interior of the brush and locate more closely to the chip surface due to its small size, thus, explaining why lysozyme was detected by QCM-D. However, the total quantity of lysozyme adsorbed may be underestimated if a portion is adsorbed on top of the PHEMA-90 nm brush (Figure 8).

above, the sensitivity of QCM for detection of protein adsorbed on a polymer brush layer is closely related to the attenuation of the mass effect in the layer and the water liberation caused by protein adsorption. The viscosity and thickness of the polymer layer may come into play. However, it should be noted that on the surfaces (POEGMA-20 nm and PHEMA-90 nm) that showed no QCM signal for fibrinogen adsorption, the adsorbed quantities measured by SPR were very low, thus raising the question whether the magnitude of the adsorbed quantity influences the QCM measurement. Therefore, it was of interest to investigate the QCM response when the protein binding capacities of these surfaces were increased. To this end, POEGMA-20 nm and PHEMA-90 nm surfaces were treated with N,N′-disuccinimidyl carbonate (DSC) to form N-hydroxysuccinimide (NHS) groups by reaction with hydroxyl groups on the polymer side chains.33 NHS groups can react readily with amino groups in proteins. The NHS-modified surfaces are therefore expected to show increased protein uptake compared to the unmodified ones. As seen from SPR measurements (Figure 6a), fibrinogen adsorption on the POEGMA-NHS-20 nm (8.28 ng/cm2) and PHEMA-NHS-90 nm (21.33 ng/cm2) surfaces was significantly higher than on the POEGMA-20 nm (4.16 ng/cm2) and PHEMA-90 nm (6.15 ng/cm2) surfaces. With the higher adsorbed protein mass, QCM-D measurements gave apparent frequency decreases of 5.87 and 3.96 Hz (masses of 169.8 and 127.6 ng/cm2), respectively, on the POEGMA-NHS-20 nm and PHEMA-NHS-90 nm surfaces. These results suggest that fibrinogen adsorption can be detected by QCM-D when the mass of adsorbed protein is greater than a certain threshold value, even though the “nominal” thickness (i.e., before adsorption) of the polymer brush layer is greater than the “critical thickness”. This effect may be due, at least in part, to water being “squeezed” from the polymer layer by protein of sufficient mass on top of the layer. It is also possible that crosslinking of the polymer brush by DSC increases the viscosity of the polymer brush layer and thus increase the “critical thickness” for sensing protein adsorption. As a result, the protein adsorbed on the polymer brush can be “detected” by QCM-D under this condition. Influence of Protein Size on QCM-D Measurements. The interactions of proteins with polymer brush layers also depend on the protein size.34 In general, the smaller the

■

CONCLUSIONS Three types of antifouling surface were used to explore the applicability of QCM for evaluation of the protein resistance of hydrophilic polymer brush surfaces. By comparing the quantities of proteins adsorbed as determined by SPR and QCM, it was found that the viscous nature of the hydrophilic polymer brushes may cause a “hearing loss” in QCM measurement of protein adsorption. For a given polymer, there appears to be a critical brush thickness beyond which the QCM chip cannot sense protein adsorbed on top of the polymer brush. The “critical thickness” increases as the viscosity of the polymer brush increases. However, a high quantity of G

DOI: 10.1021/acs.analchem.7b00198 Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry

Cardiovascular Surgery and the Natural Sciences and Engineering Research Council of Canada.

■

Figure 8. Schematic illustration of the adsorption of a small protein on a polymer brush surface.

adsorbed protein on top of the layer may squeeze out water and change the viscosity of the polymer brush. In this case, the adsorbed protein may come within the “detectable zone” even though the “nominal” thickness (i.e., before adsorption) of the polymer brush layer is greater than the critical value. Detection failure on longer polymer brushes may also due to water liberation upon protein adsorption, which causes opposite frequency shift. Furthermore, the QCM-D measurement may be affected by the size of the protein. In contrast to fibrinogen, the adsorption of a small protein (e.g., lysozyme) may occur within and at the “bottom” of the polymer brush layer (ternary and primary adsorption, respectively) and can be sensed by QCM-D more easily. However, the absolute quantity of protein adsorbed may be underestimated when secondary adsorption, on top of the layer, occurs.

■

ASSOCIATED CONTENT

* Supporting Information S

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.analchem.7b00198. NMR spectra of SAM-initiator. SAGA-FTIR spectra of POEGMA and PHEMA brush. ΔD induced by fibrinogen adsorbed on SAM-OEG. The frequency and dissipation variations of Fg or Lyz from 3, 5, and 7 overtones as the function of time on all the samples. The amount of protein adsorption measured by SPR and QCM-D (PDF).

■

REFERENCES

(1) Woo, K. M.; Chen, V. J.; Ma, P. X. J. Biomed. Mater. Res. 2003, 67A, 531−537. (2) Boontheekul, T.; Mooney, D. J. Curr. Opin. Biotechnol. 2003, 14, 559−565. (3) Maehashi, K.; Katsura, T.; Kerman, K.; Takamura, Y.; Matsumoto, K.; Tamiya, E. Anal. Chem. 2007, 79, 782−787. (4) Rowe, C. A.; Tender, L. M.; Feldstein, M. J.; Golden, J. P.; Scruggs, S. B.; MacCraith, B. D.; Cras, J. J.; Ligler, F. S. Anal. Chem. 1999, 71, 3846−3852. (5) Liu, X.; Yuan, L.; Li, D.; Tang, Z.; Wang, Y.; Chen, G.; Chen, H.; Brash, J. L. J. Mater. Chem. B 2014, 2, 5718−5738. (6) Tang, Z.; Liu, X.; Luan, Y.; Liu, W.; Wu, Z.; Li, D.; Chen, H. Polym. Chem. 2013, 4, 5597−5602. (7) McClung, W. G.; Clapper, D. L.; Anderson, A. B.; Babcock, D. E.; Brash, J. L. J. Biomed. Mater. Res. 2003, 66A, 795−801. (8) Nagai, H.; Handa, M.; Kawai, Y.; Watanabe, K.; Ikeda, Y. Thromb. Res. 1993, 71, 467−477. (9) Tseng, Y.-C.; Park, K. J. J. Biomed. Mater. Res. 1992, 26, 373−91. (10) Chen, H.; Yuan, L.; Song, W.; Wu, Z.; Li, D. Prog. Polym. Sci. 2008, 33, 1059−1087. (11) Banerjee, I.; Pangule, R. C.; Kane, R. S. Adv. Mater. 2011, 23, 690−718. (12) Currie, E. P. K; Norde, W.; Stuart, M. A. C. Adv. Colloid Interface Sci. 2003, 100−102, 205−265. (13) Halperin, A. Langmuir 1999, 15, 2525−2533. (14) Morra, M. J. Biomater. Sci., Polym. Ed. 2000, 11, 547−569. (15) Vermette, P.; Meagher, L. Colloids Surf., B 2003, 28, 153−198. (16) Benhabbour, S. R.; Sheardown, H.; Adronov, A. Macromolecules 2008, 41, 4817−4823. (17) Ladd, J.; Zhang, Z.; Chen, S.; Hower, J. C.; Jiang, S. Biomacromolecules 2008, 9, 1357−1361. (18) Park, K.; Mao, F. W.; Park, H. J. Biomed. Mater. Res. 1991, 25, 407−420. (19) Luan, Y.; Li, D.; Wang, Y.; Liu, X.; Brash, J. L.; Chen, H. Langmuir 2014, 30, 1029−1035. (20) Wei, Q.; Becherer, T.; Mutihac, R.; Noeske, P. M.; Paulus, F.; Haag, R.; Grunwald, I. Biomacromolecules 2014, 15, 3061−3071. (21) Reviakine, I.; Johannsmann, D.; Richter, R. P. Anal. Chem. 2011, 83, 8838−8848. (22) Zhao, H.; Zhu, B.; Luo, S. C.; Lin, H. A.; Nakao, A.; Yamashita, Y.; Yu, H. H. ACS Appl. Mater. Interfaces 2013, 5, 4536−4543. (23) Zhou, G.; Ma, C.; Zhang, G. Polym. Chem. 2011, 2, 1409−1414. (24) Zhang, Y.; Islam, N.; Carbonell, R. G.; Rojas, O. J. Anal. Chem. 2013, 85, 1106−1113. (25) Tang, Z.; Luan, Y.; Li, D.; Du, H.; Haddleton, D. M.; Chen, H. Chem. Commun. 2015, 51, 14263−14266. (26) Ma, H.; Wells, M.; Beebe, T. P.; Chilkoti, A. Adv. Funct. Mater. 2006, 16, 640−648. (27) Li, X.; Wang, M. M.; Wang, L.; Shi, X. J.; Ya, J. X.; Song, B.; Chen, H. Langmuir 2013, 29, 1122−1128. (28) Kitano, K.; Inoue, Y.; Matsuno, R.; Takai, M.; Ishihara, K. Colloids Surf., B 2009, 74, 350−357. (29) Jung, L. S.; Campbell, C. T.; Chinowsky, T. M.; Mar, M. N.; Yee, S. S. Langmuir 1998, 14, 5636−5648. (30) Voinova, M. V.; Rodahl, M.; Jonson, M.; Kasemo, B. Phys. Scr. 1999, 59, 391−396. (31) Martinsa, M. C. L.; Wang, D.; Jia, J.; Feng, L.; Barbosaa, M. A. Biomaterials 2003, 24, 2067−2076. (32) Zhao, C.; Li, L.; Wang, Q.; Yu, Q.; Zheng, J. Langmuir 2011, 27, 4906−4913. (33) Wu, Z.; Chen, H.; Huang, H.; Zhao, T.; Liu, X.; Li, D.; Yu, Q. Macromol. Biosci. 2009, 9, 1165−1168. (34) Halperin, A. Langmuir 1999, 15, 2525−2533.

AUTHOR INFORMATION

Corresponding Authors

*E-mail: [email protected]. *E-mail: [email protected]. Fax: +86-512-65880583. Tel.: +86-512-65880827. ORCID

Hong Chen: 0000-0001-7799-4961 Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This work was supported by the National Natural Science Foundation of China (21334004 and 21574093), the Priority Academic Program Development of Jiangsu Higher Education Institutions (PAPD), Jiangsu Clinical Research Center for H

DOI: 10.1021/acs.analchem.7b00198 Anal. Chem. XXXX, XXX, XXX−XXX