QCM Analysis for Two-Step Adsorption of Albumin and Fibronectin on Zirconia Surface

The adsorption of proteins on the dental implant surface is the first step in the key role of osseointegration. Many types of proteins exist in the living body and compete for adsorption on the material surface. As an implant material, partially stabilized zirconia (ZrO2) is currently an attractive alternative to titanium to overcome the shortcomings of titanium implants. In this study, we investigated the two-step adsorption of fibronectin (Fn) and bovine serum albumin (Alb) on the ZrO2 surface using a 27-MHz quartz crystal microbalance (QCM) method. A ZrO2 sensor was employed for the QCM measurements. Two-step adsorptions were performed as follows. (1) Fn-Alb series: first, the Fn solution was injected into the phosphate-buffered saline (PBS) solution, followed by the Alb solution. (2) Alb-Fn series: first, the Alb solution was injected, followed by the Fn solution. The decrease in frequency was monitored for 60 minutes after each protein injection. The adsorbed amounts of Fn or Alb were calculated by observing the decrease in frequency, and the apparent reaction rate, kobs, was obtained through the curve fitting of frequency shift against the adsorption time. No significant difference was observed in the adsorbed amounts of Fn and Alb between the Fn-Alb and Alb-Fn series (
 
 P
 >
 0.05
 
 ). The kobs, rate of protein adsorption, in the second step was significantly slower than that in the first step for both Fn and Alb adsorption (
 
 P
 <
 0.05
 
 ). There was no clear correlation between the amount of protein adsorbed on the ZrO2 sensor and the surface topography. It was concluded that the amount of protein adsorbed on the ZrO2 surface was not influenced by the two-step adsorption series, but the adsorption rate of proteins in the second step was affected by the first-step protein adsorption.


Introduction
Early new bone formation and subsequent osseointegration formation is one of the most important factors in the success of dental implant treatment [1]. Trindade et al. reported that osseointegration is a multistage process, in which the first step is the attachment or adsorption of proteins on the surface of dental implant materials [2]. Kasemo et al. reported that adsorbed proteins affect cell adhesion. Inflammation begins with bleeding during the wound healing process. Subsequently, mesenchymal stem cells and osteoblasts induce the differentiation and promotion of bone formation [3,4]. It is suggested that the presence of adhesive proteins and/or bone formationrelated proteins on the material surface controls new bone formation. e study of the adsorption behaviors of these proteins on the implant surface is therefore important to understand the mechanism of osseointegration.
In the last decade, dental zirconia (ZrO 2 ) implants have been introduced in dental clinics. e zirconia used for dental implants is tetragonal zirconia stabilized by the addition of approximately 3-4% yttria (Y-TZP) [5,6]. Y-TZP, a white ceramic material, has superior mechanical properties, such as bending strength and high fracture toughness, owing to the stress-induced transformation mechanism. erefore, it can overcome the shortcomings of titanium, such as dark grayish color and/ or metal sensitivity [7][8][9]. Various surface modifications in zirconia implants have been reported to achieve tight and fast bone bonding and the formation of osseointegration [10][11][12][13].
Quartz crystal microbalance (QCM) is a sensing device used to monitor protein adsorption behavior on the material surface [14]. e QCM technique is a straightforward method used to detect the adsorption or desorption of proteins on a material surface by measuring the differences in the oscillating frequency of the quartz sensor cell. e adsorption of protein on the surface of an oscillating quartz crystal causes a decrease in the oscillation frequency in relation to the amount of protein bound to the crystal surface. e amount of protein adsorbed on the biomaterial can be estimated using the Sauerbrey equation [15].
Numerous studies on protein adsorption using the QCM method have been reported [16][17][18]. We investigated the protein adsorption on implant materials using the QCM method. For example, Kusakawa et al. evaluated the protein adsorption on ZrO 2 or Ti using QCM. ey found that the protein adsorption amounts of fibronectin and albumin on Ti were larger than those on ZrO 2 [19]. Yoshida and Hayakawa used the QCM method to investigate lactoferrin adsorption on Ti, stainless steel, ZrO 2 , and polymethyl methacrylate surfaces. Ti and stainless steel showed a greater amount of lactoferrin adsorption than ZrO 2 and polymethyl methacrylate [20].
Many types of proteins exist in the living body, which compete for adsorption on the material surface. It is presumed that the preadsorption of proteins influences the continuous adsorption of other proteins. Pegueroles et al. investigated the competition of protein adsorption on the TiO 2 surface in two-step adsorption experiments using the QCM method [21]. ey used fibrinogen (Fbg), fibronectin (Fn), and albumin (Alb) as proteins and found that two-step adsorption on the TiO 2 surface was influenced by the differences between the proteins. Imagami et al. reported that preadsorbed proteins influence the adsorption of proteins during the second step by monitoring the two-step adsorption of Fn and Alb on Ti in the QCM analysis method [22]. ere are no reports of two-step adsorption on ZrO 2 surface using the QCM method.
In this study, we aimed to investigate the adsorption of different proteins on the ZrO 2 surface in a two-step adsorption system using the QCM method. Fn, a cell-adhesive protein, and Alb, a cell adhesion inhibiting protein, were evaluated as proteins, and sequential adsorption was conducted in the QCM sensor cell. Figure 1 shows the QCM device and the mounted sensor used in this study. A 27-MHz QCM (AFFINIX QNμ, ULVAC, Inc., Kanagawa, Japan) with 550 μL sensor cells was used. e temperature was maintained at 25 ± 1°C using a temperature control system, and the solution in the cells was stirred during the measurements. e AT-cut quartz crystal was sandwiched between gold electrodes on both sides. e surface area of the Au electrode was 4.9 mm 2 . ZrO 2 sensors were used in the study. is sensor was prepared by sputter coating on the Au electrode using sputtering deposition equipment (CS200, ULVAC, Inc. Kanagawa, Japan). e zirconium target was sputtered at 0.5 Pa in an oxygen chamber for 30 minutes. e film thickness of sputtered zirconia has been reported to have a thickness of 115 nm using a profilometer in our previous study [23]. e ZrO 2 sensor was assembled in a sensor cell and irradiated with ultraviolet radiation (BioForce Nanosciences Holdings Inc., VA, US) for 20 minutes before QCM measurement. UV irradiation of 15 mW/cm 2 was emitted perpendicularly to the sensors at a distance of 20 mm at λ � 254 nm. After irradiation, the ZrO 2 sensor cell was mounted on the cell socket of the QCM apparatus, as shown in Figure 1. UV irradiation on ZrO 2 surface removes the contamination and makes the surface fresh [24].

EPMA Mapping Analysis of ZrO 2 Sensors.
e ZrO 2 sensor surfaces after ultraviolet irradiation were evaluated by electron probe microanalysis (EPMA, JXA-8900R, JEOL Ltd., Tokyo, Japan) at an accelerating voltage of 20 kV by detecting the X-ray intensities of Zr-Kα. e presence of Zr was confirmed by elementary mapping.

AFM Observation of
Sensor. An atomic force microscope (AFM; Nanosurf Easyscan 2, Nanosurf, AG, Switzerland) was used to identify the surface conditions and surface roughness of the ZrO 2 sensor before and after protein adsorption. Fn or Alb was adsorbed on the ZrO 2 sensor, as described above. After rinsing the excess amount of adsorbed Fn or Alb with double-distilled water for 20 seconds, each surface was observed. e AFM images were obtained for an area of 5 × 5 μm 2 . e three-dimensional arithmetic height (Sa) was calculated as a surface roughness parameter. All measurements were performed five times for each sample. Tapping mode silicon probes (Tap190Al-G, force contact 48 N/m, Budget sensors, Sofia, Bulgaria) with resonance frequencies of approximately 190 kHz were used to perform AFM imaging.

Contact Angle Measurement of ZrO 2 Sensors.
e contact angles of the ZrO 2 sensor before and after protein adsorption with respect to double-distilled water were measured using a contact angle meter (DMe-201, Kyowa Interface Science Co. Ltd., Tokyo, Japan). ZrO 2 sensor was measured after ultraviolet irradiation of each sensor. Fn or Alb was adsorbed on the ZrO 2 sensor, as described above. After rinsing the excess amount of adsorbed Fn or Alb with double-distilled water for 20 seconds, each surface was observed. e water drop volume was maintained at 0.5 μL, and the contact angle was the measured value 3 seconds after dripping. Measurements were performed five times at the same room temperature and humidity.

QCM Measurement through Two-Step Adsorption of
Proteins. Human plasma fibronectin (Fn, Harbor Bio-Products, MA, USA) and bovine serum albumin (Alb, Wako Pure Chemical Inds. Ltd., Osaka, Japan) were dissolved in a phosphate-buffered saline (PBS) solution (pH 7.4) with a concentration of 0.5 mg/mL. Fn used in the present study is a dimer of 220 kDa glycoprotein. Dimers are connected by a pair of C-terminal S-S bonds. Fn is well-known cell-adhesive protein. Integrin which is a transmembrane receptor binds to RGD (Arg-Gly-Asp) motif of Fn. Molecular weight of Alb in the present study is approximately 6-7 kDa. Alb is a small protein compared to other proteins and exhibits cyclic threedimensional structure. Serum Alb is major component in serum. Alb is known as cell-adhesion inhibiting protein as mentioned.
A typical frequency-decrease curve is shown in Figure 2. A two-stage decrease in frequency was observed based on the two-step adsorption of proteins. e sensor cell with the ZrO 2 sensor was filled with 500 μL of PBS. After stabilization of the frequency shift, two-step adsorption of Fn and Alb was monitored in the following two series.
(1) Fn-Alb series: in the first step, 5 μL of Fn solution was injected into the PBS solution in the cell. e decrease in frequency due to Fn adsorption on the ZrO 2 sensor was monitored for 60 minutes after Fn injection. In the second step, 5 μL of Alb solution was injected on the Fn preadsorbed sensor into the PBS solution in the cell, and the decrease in frequency was monitored for an additional 60 minutes. (2) Alb-Fn series: first, 5 μL of Alb solution was injected.
After 60 minutes of Alb injection, 5 μL of Fn solution was injected on the Alb preadsorbed sensor in the second step. e decrease in frequency was monitored for an additional 60 minutes.
e amount of Fn and Alb adsorbed on the ZrO 2 sensor surface (Δm) was calculated 60 minutes after each protein injection using Sauerbrey's equation [15]: where ΔF is the measured frequency shift (Hz), Δm is the mass change (g), F 0 is the fundamental frequency of the quartz crystal (27 × 10 6 Hz), A is the electrode area (0.049 cm 2 ), ρ q is the density of quartz (2.65 g/cm 3 ), and μ q is the shear modulus of quartz (2.95 × 10 11 dyn/cm 2 ). At 27 MHz, a decrease of 1 Hz in frequency corresponded to a mass change of approximately 0.62 ng/cm 2 . rough the curve fitting of ΔF curve against the adsorption time, the apparent reaction rate, k obs , was obtained using the following equation: Five runs of QCM measurements for two series were performed.
2.6. Statistical Analysis. Significant differences were determined using SPSS version 25 (IBM Japan, Ltd., Tokyo, Japan). e surface roughness data were evaluated by conducting one-way analysis of variance (ANOVA) and the Bonferroni test for multiple comparisons between the means. e data of adsorbed amounts at each step and k obs at the same time after the injection, i.e., at 5 minutes in the Fn-Alb series or in the Alb-Fn series, were analyzed using a nonpaired t-test. e statistical significance was set at P < 0.05, and data were expressed as mean ± standard deviation (SD). Figure 3 show the elementary distribution of Zr on the ZrO 2 sensor. Homogeneous sputter coating of Zr on each sensor surface was confirmed.

AFM Analysis and Surface Roughness and Contact Angles of ZrO 2 Sensor.
e AFM images of the native ZrO 2 sensor and the Fn-and Alb-adsorbed ZrO 2 sensor surfaces after rinsing with double-distilled water are shown in Figure 4. Small particles (0.2 μm or less) formed through the sputtering process were uniformly recognized on the native ZrO 2 sensor. After rinsing, some amounts of Alb and Fn were observed on the Fn-and Alb-adsorbed ZrO 2 sensors. It appeared that more adsorbed proteins were observed on the Fn-adsorbed sensor than on the Alb-adsorbed sensor.   Table 1. Fn-adsorbed ZrO 2 showed a significantly greater surface roughness than native ZrO 2 (P < 0.05). No significant difference in surface roughness was observed between Fn-adsorbed and Albadsorbed ZrO 2 (P > 0.05). Fn-and Alb-adsorbed ZrO 2 sensor showed significantly greater contact angles than ZrO 2 (P < 0.05), and Fn-adsorbed ZrO 2 showed a significantly greater contact angle than Alb-adsorbed ZrO 2 (P < 0.05).

QCM Measurements through Two-
Step Adsorption of Fn and Alb. Figure 5 shows the ΔF curves for the Fn-Alb and Alb-Fn series to the ZrO 2 sensor using QCM measurements. e decrease in ΔF immediately after the Fn or Alb injection can be monitored, which corresponds to an increase in the adsorption of Fn and Alb. e Fn injection produced a greater decrease in frequency than the Alb injection in both the first and second steps. Eventually, there were no distinct differences in the decrease in frequency between the Fn-Alb and Alb-Fn series, 120 minutes after protein injection. Figure 6 shows the adsorbed amounts of Fn and Alb on the ZrO 2 sensor in different two-step adsorption methods, which were estimated using the Sauerbrey equation [15]. e adsorbed amount of Fn is significantly higher than that of Alb in both the Fn-Alb and Alb-Fn series (P < 0.05). No significant differences were observed for Fn and Alb adsorption between the first and second adsorption steps (P > 0.05). Moreover, there were no significant differences in total adsorbed protein amounts between the Fn-Alb and Alb-Fn series (P > 0.05). Table 2 lists the k obs values for the adsorption of Fn and Alb on ZrO 2 sensor obtained 5 minutes after injection. A larger value of k obs indicates a relatively rapid adsorption rate. In both series, Fn shows significantly greater k obs than Alb (P < 0.05), and the k obs of protein adsorption in the first step is significantly greater than k obs of protein adsorption in the second step (P < 0.05).

Discussion
In this study, we evaluated the two-step adsorption behavior of fibronectin and albumin on a ZrO 2 sensor using a 27-MHz QCM. Two types of proteins, Fn and Alb, were used as a mixed model to simulate body fluid. is study revealed that the preadsorption of proteins in the first step influenced protein adsorption in the second step.
e ZrO 2 sensor used in this study was prepared using sputter coating. In the sputtering process, atoms or molecules are ejected from the target material through plasma or gas bombardment. e ejected atoms or molecules are then deposited on the substrate. In this study, zirconium was used as the target for the ZrO 2 sensor. Kusakawa et al. prepared a ZrO 2 sensor under the same conditions [19]. We confirmed the homogeneous sputter coating of Zr on a sensor surface in this study. First, we analyzed the surface topographies of the ZrO 2 sensor to check its surface conditions. In addition, the surface conditions after the first step of protein adsorption were analyzed. e adsorption amount of Fn was higher than that of Alb during the first adsorption. is result was the same as the previous result. Our previous studies also reported that the adsorption of Fn was greater than that of Alb on Ti or ZrO 2 [19,25]. Interestingly, the Fn adsorption amount was significantly larger in the first step as well as in the second step.
Our previous studies on two-step adsorption to Ti showed that the order of protein adsorption influenced the total adsorption amount [22]. e adsorption amounts in the Alb-Fn series were significantly greater than those in the Fn-Alb series. In particular, Fn adsorbed to Alb-adsorbed Ti was more than that absorbed to native Ti. However, the present two-step adsorption study on ZrO 2 showed that there were no differences in the total adsorption amounts between the two adsorption series. e adsorption of Fn on  Advances in Materials Science and Engineering Alb-adsorbed ZrO 2 and that of Alb on Fn-adsorbed ZrO 2 was almost the same as that on the native ZrO 2 surface. It is well known that the surface roughness of a material influences protein adsorption. More proteins are adsorbed on a rougher surface [26]. However, there was no clear relationship between surface roughness and adsorbed amounts of proteins in the present study. e rough surfaces at the second step of adsorption did not increase the amount of protein adsorption compared to the first step of adsorption.
Hydrophilicity is another important factor in protein adsorption. Wei et al. investigated protein adsorption on a surface with a wide range of wettability [27]. ey prepared this surface through the combination of hexamethyldisiloxane film coating and O 2 -plasma treatment, where the contact angle of water ranged from 0°to 106°. ey reported that Fn was adsorbed on both hydrophilic and hydrophobic surfaces. On the contrary, Alb showed greater adsorption on the hydrophobic surface. UV irradiation produced a hydrophilic ZrO 2 surface as a previous report [19]. Proteinadsorbed surfaces in this study are more hydrophobic than UV-irradiated native surfaces. It was predicted that more amount of Fn adsorbed surface to UV-irradiated native surfaces than Alb-adsorbed surface and also more amount of   Step1: Fn or Alb

Advances in Materials Science and Engineering
Alb adsorbed to Fn-adsorbed surface than UV-irradiated native surface. However, there were no distinct differences in the amounts of protein adsorption between the first and second adsorptions. Electrostatic interaction is also an important factor in protein adsorption. e isoelectric points (IP) of Fn and Alb are very similar at 5.0-6.0 and 4.7-4.9, respectively [20,28,29]. Fn and Alb were negatively charged under the present condition of pH � 7.4. Yoshida and Hayakawa [20] reported that the zeta potentials of ZrO 2 after UV irradiation at pH � 7.4 were approximately −43 mV by the streaming potential method. ese results indicated that the surface of ZrO 2 is also negatively charged at the present conditions. us, electric repulsion occurred between each protein, Fn or Alb, and UV-irradiated native surface. It is suggested that the adsorption behavior of Fn to UV-irradiated native surface at the first step of adsorption was the results of the positive effect in adsorption due to the hydrophilic interaction and the negative results due to the electrostatic repulsion. Rezwan et al. investigated the adsorption of BSA to alumina, silica, titania, and ZrO 2 particles [30]. ey reported that adsorbed amounts of BSA to alumina, silica, and titania are highly correlated with the zeta potential, but not on the zirconia. ey also suggested that hydrophobic effect plays a more important role than electro static interaction for zirconia adsorption. However, it is suggested that another factor besides surface roughness, hydrophilicity, and electrostatic interaction influence the two-step adsorption behaviors in the present study.
Pegueroles et al. recorded an increase in frequency of Alb adsorption on Fn preadsorbed surfaces [21]. ey suggested that Alb molecules displaced some of the Fn molecules through displacement interactions due to the Vroman effect [31], in which proteins with stronger binding affinities competitively displace earlier adsorbed higher-concentration proteins. ey speculated that Alb has a higher affinity for the TiO 2 surface than Fn. Imagami et al. also suggested that the difference between affinities of Fn and Alb to Ti may influence the second protein adsorption behavior [22]. For the Ti surface, Alb showed a higher affinity than Fn. In the present study, more remaining molecules of Fn were observed on the ZrO 2 surface after rinsing the Fn-adsorbed surface due to the higher affinity of Fn for the ZrO 2 surface than Alb. Protein affinity to the ZrO 2 surface was reversed to the Ti surface. It is presumed that this difference between protein affinities of Ti and ZrO 2 may influence the protein adsorption behaviors. However, a clear mechanism has yet to be discovered. Other factors, such as conformational and electrical changes of preadsorbed proteins, contribute to the adsorption of protein at the second step. A more detailed surface analysis after the first-step adsorption is required. e results of k obs indicated that the protein-adsorbed surface delayed the second-step adsorption rate. is result is the same as that for the Ti surface as previously reported [22]. It was concluded that protein preadsorbed surfaces induce rapid adsorption of proteins at the second step. More surface characterizations after protein adsorption should be needed. It is suggested that surface roughness or hydrophilicity may influence the adsorption rate of proteins.
ere are various types of proteins in living tissues, and their adsorption behaviors are more complex. In body fluids, albumin concentration is higher than fibronectin, and the difference in the concentration of proteins also influences the adsorption behavior. Moreover, competitive or step-wise protein adsorption to the implant material may occur. e analysis of competitive adsorption is more complex and difficult. So, we intended to analyze step-wise protein adsorption at first. As a series of step-wise protein adsorption experiment, we designed the simplest experiments model of step-wise adsorption. Same concentrations of proteins are used for two-step adsorption in this preliminary experiment. In the future, we will further investigate the two-or threestep adsorption behaviors of proteins of different concentrations to obtain the information about protein adsorption on implant materials in vivo. Other proteins such as osteocalcin and osteopontin play a key role in bone formation and osseointegration. e two-or three-step adsorption behaviors of proteins with different concentrations will be further investigated to obtain information about protein adsorption on implant materials.

Conclusion
We investigated the two-step adsorption of Fn and Alb on ZrO 2 as an implant material using the QCM method. ere were no significant differences in the adsorbed amounts of   Advances in Materials Science and Engineering Fn and Alb between the different two-step adsorption series (P > 0.05). In contrast, the adsorption rate in the second step of adsorption was significantly slower than that in the first step of adsorption for both Fn and Alb adsorption (P < 0.05). ere was no clear correlation between the amount of protein adsorbed by the ZrO 2 sensor and surface topography.
It was concluded that the amount of protein adsorbed on the ZrO 2 surface was not influenced by the two-step adsorption series, but the adsorption rate of proteins in the second step was affected by the first-step protein adsorption.

Data Availability
e data used to support the findings of this study are included within the article.

Conflicts of Interest
e authors declare that there are no conflicts of interest associated with this study.