A Universal Photochemical Method to Prepare Carbohydrate Sensors Based on Perfluorophenylazide Modified Polydopamine for Study of Carbohydrate-Lectin Interactions by QCM Biosensor

A universal photochemical method to prepare carbohydrate sensors based on perfluorophenylazide (PFPA) modified polydopamine (PDA) for the study of carbohydrate–lectin interactions by a quartz crystal microbalance (QCM) biosensor was developed. The PFPA was immobilized on PDA-coated gold sensors via Schiff base reactions. Upon light irradiation, the underivatized carbohydrates were inserted into the sensor surface, including mannose, galactose, fucose and N-acetylglucosamine (GlcNAc). Carbohydrate sensors were evaluated for the binding to a series of plant lectins. A kinetic study of the interactions between mannose and concanavalin A (Con A), fucose and Ulex europaeus agglutinin I (UEA-I) were performed. This method can eliminate the tedious modification of carbohydrates, improve the experimental efficiency, and reduce the experimental cost, which is of great significance for the development of QCM biosensors and the study of biomolecular interactions.


Introduction
Carbohydrates play important roles in cell communication, apoptosis, immune response and infection by interacting with proteins and other biomolecules [1][2][3]. The quartz crystal microbalance (QCM) is an ultrasensitive weighing device, consisting of a thin disk of single crystal quartz, with metal electrodes deposited on each side of the disk. The crystal can be made to oscillate at its resonant frequency, f, when connected to an external driving oscillator circuit [4]. Biosensors based on QCM have proven to be an efficient and powerful instrument for label-free detection of biomolecular interactions in real-time, permitting the determination of the kinetics and affinity parameters of the interaction, including the dissociation rate constant, association rate constant, and affinity constant. It can be used for analysis of interactions between protein and carbohydrate with carbohydrate sensors, protein sensors, and cell sensors [5][6][7][8][9][10][11][12]. Therefore, as an efficient analysis tool, QCM sensor technology has attracted widespread attention, especially carbohydrate sensors for carbohydrate-protein interactions [13,14]. Lectins are a heterogeneous group of proteins, have at least one non-catalytic domain that recognize selectively and bind reversibly to specific free carbohydrates or glycans presenting on glycoproteins Spectrometer (Karlsruhe, Germany) in CDCl 3 . SEM-EDX were obtained using the S-4800 instrument (Hitachi Ltd.) with an accelerating voltage of 10.0 kV. The photochemistry in this work was undertaken with a Rayonet PRP-100 photochemical reactor (Branford, CT, USA). The experiments were carried out using an Attana QCM Cell-A200 (Stockholm, Sweden) ( Figure 1). An Attana gold sensor (Stockholm, Sweden) was used for further modifications to obtain a carbohydrate sensor. to a series of plant lectins, which include wheat germ agglutinin (WGA), soybean agglutinin (SBA), peanut agglutinin (PNA), UEA-I and Con A. In addition, two kinds of carbohydrates and their corresponding lectins were selected for kinetic study. This research provides a universal method for the preparation of carbohydrate sensors based on PFPA modified PDA sensors by photochemical covalent immobilization of underivatized carbohydrates, which can be used for the analysis of carbohydrate-lectin interactions by QCM biosensors.

Materials
2,2'-(ethylenedioxy)bis(ethylamine), dopamine hydrochloride and D-mannose were obtained from Xianding Biotechnology Co. (Shanghai, China). Methyl pentafluorobenzoate and aminoacetic acid was obtained from Shanghai Aladdin Bio-Chem Technology Co. (Shanghai, China). N-acetyl-Dglucosamine and L-fucose were obtained from Shanghai Macklin Biochemical Co. (Shanghai, China). Bovine serum albumin (BSA) and D-galactose were obtained from Energy Chemical (Shanghai, China). SBA, WGA, PNA, and UEA-I were obtained from Vector Laboratories (Burlingame, CA, USA). Con A, concanavalin A fluorescein conjugate (FITC-Con A) and tris(hydroxymethyl)aminomethane (Tris) were obtained from Sigma-Aldrich (Shanghai, China). 13 C, 19 F and 1 H NMR spectra were recorded by using a Bruker AVANCE III HD 500MHz Spectrometer (Karlsruhe, Germany) in CDCl3. SEM-EDX were obtained using the S-4800 instrument (Hitachi Ltd.) with an accelerating voltage of 10.0 kV. The photochemistry in this work was undertaken with a Rayonet PRP-100 photochemical reactor (Branford, CT, USA). The experiments were carried out using an Attana QCM Cell-A200 (Stockholm, Sweden) ( Figure 1). An Attana gold sensor (Stockholm, Sweden) was used for further modifications to obtain a carbohydrate sensor.

Sensors Cleaning
QCM gold sensors were placed in a 50 mL beaker and immersed with 1% NaClO solution. The beaker was shaken slowly to clean the sensors for 2-3 min, and then the sensors were cleaned with 1% NaClO solution 2-3 times until the surface of the sensors were bright yellow. Furthermore, the sensors were washed with Milli-Q water 4-5 times and dried with nitrogen stream.

Sensors Cleaning
QCM gold sensors were placed in a 50 mL beaker and immersed with 1% NaClO solution. The beaker was shaken slowly to clean the sensors for 2-3 min, and then the sensors were cleaned with 1% NaClO solution 2-3 times until the surface of the sensors were bright yellow. Furthermore, the sensors were washed with Milli-Q water 4-5 times and dried with nitrogen stream. The sensors mentioned above were cleaned for 1 h in 30% H 2 O 2 (0.4 mL) and H 2 SO 4 (1.2 mL) at 50 • C. At the same time, the beaker was shaken continuously during this process. The sensors were washed with a large amount of Milli-Q water until neutral, then washed with ethanol three times and dried with nitrogen stream.

Modification of Polydopamine on the Sensor Surfaces
50 µL dopamine solution (2 mg/mL) in 10 mM Tris buffer (pH 8.5) was added to the gold surfaces at room temperature for 1.5 h. After incubation, the sensor surfaces were washed with Milli-Q water and dried with nitrogen stream to obtain the polydopamine-modified QCM sensors.

Synthesis of Amino-PFPA
Compound 7 was synthesized via the previously reported method, which was shown in the Supporting Information.

Preparation of the PFPA-modified Sensors
Compound 7 (5 mg) was dissolved in 100 µL PBS of pH 8.5. The solution (50 µL) was added to the surfaces of the modified PDA-coated sensors, and incubated in a wet box for 4 h, then the sensor surfaces were further rinsed with PBS in order to remove compound 7, which was unimmobilized on the surfaces.

Preparation of Carbohydrate Sensors
50 µL carbohydrate solution (100 mM) was added to the PFPA-modified sensors, after which the photochemical fabrication of PFPA was carried out under ultraviolet light of 300 nm for 10 min. The sensor surfaces were washed three times with PBS to remove uninserted carbohydrates and dried with nitrogen stream to obtain the carbohydrate sensors.

SEM-EDX Measurement
The samples were characterized by SEM-EDX to analyze the surface chemical composition of sensor surfaces after drying overnight.

QCM Analysis of Carbohydrate-Lectin Interactions
The sensors were docked into the QCM Cell-A200 and equilibrated under a continuous flow rate of running buffer (100 µL/min). Subsequently, the flow rate was reduced to 25 µL/min to study interactions between carbohydrate and lectin. To block the areas which were not react with carbohydrate, BSA (100 µg/mL) was subsequently injected over the carbohydrate sensor surfaces until the surfaces were saturated. Then, 100 µg/mL lectins were injected into the surface of the sensors. The resonant frequency of the quartz crystal could be recorded with the Attester software in real time, just as the frequency shift (∆f ) coupled to the association or dissociation. The interactions of carbohydrate and lectin were monitored for an association period of 84 s and a dissociation period of 300 s. The evaluation software of the Attana QCM Cell-A200 was used to record data. Finally, the bound lectin was removed by injecting 10 mM glycine of pH 1.5 to regenerate the sensors. Attester Evaluation software was used to analyze the data.

Results and Discussion
The QCM gold sensor surfaces were prepared employing a three-step process. As shown in Figure 2, firstly, in alkaline environment, the gold surfaces were fabricated with PDA via dopamine self-polymerization. After coating of the PDA, the PFPA with an amino group was immobilized on the surfaces of the PDA-coated sensor by Schiff-base reactions. Finally, PFPA of photochemical fabrication was used to immobilize underivatized carbohydrates on the sensor surfaces under an ultraviolet light of 300 nm. In this work, the carbohydrate-lectin interactions were monitored by QCM biosensor. Setting the rate of flow to 25 µL/min of running buffer (10 mM PBS of pH 7.4), the interaction between carbohydrates and different lectins was detected. In addition, the kinetic properties of the interactions between carbohydrate and lectin were further performed, where we selected two kinds of carbohydrates for the kinetic study. The interactions between carbohydrate and lectin can be monitored by QCM biosensor, permitting the determination of the affinity and the kinetics parameters of the interactions. interaction between carbohydrates and different lectins was detected. In addition, the kinetic properties of the interactions between carbohydrate and lectin were further performed, where we selected two kinds of carbohydrates for the kinetic study. The interactions between carbohydrate and lectin can be monitored by QCM biosensor, permitting the determination of the affinity and the kinetics parameters of the interactions.

SEM-EDX Analysis of the Sensor Surfaces
As shown in Figure 3, the chemical composition of these sensor surfaces can be analyzed by SEM-EDX [36]. Based upon the spectra analysis of SEM-EDX, the elemental peak of C, N and O appeared on the PDA-coated gold sensors, compared with the element composition of gold sensors ( Figure 3a). The spectrum of SEM-EDX indicated the evident signal of the PDA's atomic composition ( Figure 3b). As shown in Figure S6 of Supporting Information, the region of peaks below 1 keV for Figure 3a,b were enlarged. Compared with the sensor surfaces coated with PDA, the SEM-EDX spectral signal of the chemical composition F appeared on the sensor surfaces after immobilizing compound 7 (Figure 3c), which indicated that the PFPA was connected to the surfaces of the sensor successfully. As shown in Figure 3d, the spectral signal of the chemical composition C and O improved markedly compared to the PFPA-modified sensor, which indicated the immobilization of the carbohydrate on the sensor surfaces.

SEM-EDX Analysis of the Sensor Surfaces
As shown in Figure 3, the chemical composition of these sensor surfaces can be analyzed by SEM-EDX [36]. Based upon the spectra analysis of SEM-EDX, the elemental peak of C, N and O appeared on the PDA-coated gold sensors, compared with the element composition of gold sensors ( Figure 3a). The spectrum of SEM-EDX indicated the evident signal of the PDA's atomic composition ( Figure 3b). As shown in Figure S6 of Supporting Information, the region of peaks below 1 keV for Figure 3a,b were enlarged. Compared with the sensor surfaces coated with PDA, the SEM-EDX spectral signal of the chemical composition F appeared on the sensor surfaces after immobilizing compound 7 (Figure 3c), which indicated that the PFPA was connected to the surfaces of the sensor successfully. As shown in Figure 3d, the spectral signal of the chemical composition C and O improved markedly compared to the PFPA-modified sensor, which indicated the immobilization of the carbohydrate on the sensor surfaces. Polymers 2019, 11, x FOR PEER REVIEW 6 of 13

FITC-Con A Immobilization on the Carbohydrate Sensor Surface and Regeneration Result
To verify the interactions between carbohydrate and lectin, FITC-Con A with green fluorescence was used to observe whether the carbohydrate on the sensor surface binding with lectin successfully. When the baseline was stable (frequency shift less than 0.2 Hz/min), BSA was injected into the carbohydrate sensor surface to block the area where PFPA did not react with carbohydrate until the sensor surface was saturated. Then, 100 μg/mL FITC-Con A was injected and observed by a fluorescence microscope to prove interactions between carbohydrate and lectin. As shown in Figure  4A, the binding response of interactions between mannose and FITC-Con A was recorded, resulting in a frequency shift about 20 Hz. Before and after the injection of the FITC-ConA, a fluorescence microscopic evaluation was performed. Before injection, only a black field was observed, while a green fluorescence could be observed after injection of FITC-ConA onto the carbohydrate sensor surface. Furthermore, regeneration of the carbohydrate sensor surface was carried out, where the bounded Con A can be removed by injecting 10 mM glycine of pH 1.5 ( Figure 4B). As can be seen, the baseline reverted to its original state, which indicates that the carbohydrate sensor surface was regenerated successfully.

FITC-Con A Immobilization on the Carbohydrate Sensor Surface and Regeneration Result
To verify the interactions between carbohydrate and lectin, FITC-Con A with green fluorescence was used to observe whether the carbohydrate on the sensor surface binding with lectin successfully. When the baseline was stable (frequency shift less than 0.2 Hz/min), BSA was injected into the carbohydrate sensor surface to block the area where PFPA did not react with carbohydrate until the sensor surface was saturated. Then, 100 µg/mL FITC-Con A was injected and observed by a fluorescence microscope to prove interactions between carbohydrate and lectin. As shown in Figure 4A, the binding response of interactions between mannose and FITC-Con A was recorded, resulting in a frequency shift about 20 Hz. Before and after the injection of the FITC-ConA, a fluorescence microscopic evaluation was performed. Before injection, only a black field was observed, while a green fluorescence could be observed after injection of FITC-ConA onto the carbohydrate sensor surface. Furthermore, regeneration of the carbohydrate sensor surface was carried out, where the bounded Con A can be removed by injecting 10 mM glycine of pH 1.5 ( Figure 4B). As can be seen, the baseline reverted to its original state, which indicates that the carbohydrate sensor surface was regenerated successfully.
The prepared carbohydrate sensors were inserted in the Attana QCM Cell-A200 biosensor. When stable baseline was achieved at 25 μL/min of the PBS running buffer, we blocked the areas which did not react with carbohydrate by BSA. The BSA was repeatedly injected to the carbohydrate sensor surfaces until the surfaces were saturated. Then a series of lectins of 100 μg/mL WGA, SBA, PNA, UEA-I, and Con A were injected to detect the interactions of carbohydrate-lectin, respectively. After each injection of the lectin, 10 mM glycine (pH 1.5) was injected to the surface for regeneration, and the binding lectin was removed after the cycle. We used Attester software to realize real-time monitoring of carbohydrate with lectin interactions and regeneration processes. Figure 5 shows frequency change of five different lectins (WGA, SBA, PNA, UEA-I, Con A) binding to GlcNAcmodified sensor chip surface. As can be seen, the binding of WGA is the highest, at about 20 Hz, indicating that other lectins are basically not bound except for non-specific adsorption.
The prepared carbohydrate sensors were inserted in the Attana QCM Cell-A200 biosensor. When stable baseline was achieved at 25 µL/min of the PBS running buffer, we blocked the areas which did not react with carbohydrate by BSA. The BSA was repeatedly injected to the carbohydrate sensor surfaces until the surfaces were saturated. Then a series of lectins of 100 µg/mL WGA, SBA, PNA, UEA-I, and Con A were injected to detect the interactions of carbohydrate-lectin, respectively. After each injection of the lectin, 10 mM glycine (pH 1.5) was injected to the surface for regeneration, and the binding lectin was removed after the cycle. We used Attester software to realize real-time monitoring of carbohydrate with lectin interactions and regeneration processes. Figure 5 shows frequency change of five different lectins (WGA, SBA, PNA, UEA-I, Con A) binding to GlcNAc-modified sensor chip surface. As can be seen, the binding of WGA is the highest, at about 20 Hz, indicating that other lectins are basically not bound except for non-specific adsorption.
The binding results between the other three carbohydrates and five different lectins were represented in the Figures 6-8, respectively. As can be seen in Figure 6, Con A has the highest binding with mannose, while the other lectins were basically not bound except that a small amount of non-specific adsorption was detected. As can be seen in Figure 7, UEA-I has the highest binding with fucose, while the other lectins were basically not bound except that a small amount of non-specific adsorption was detected. As can be seen in Figure 8, SBA has the highest binding with galactose, followed by PNA. This is because both SBA and PNA can bind to galactose, but the binding strengths are different, while the other lectins are not bound except for a small amount of non-specific adsorption. The experimental results show that the interactions between carbohydrate and lectin are specific, which are consistent with the results obtained by other methods, indicating the universality of this method.
After each injection of the lectin, 10 mM glycine (pH 1.5) was injected to the surface for regeneration, and the binding lectin was removed after the cycle. We used Attester software to realize real-time monitoring of carbohydrate with lectin interactions and regeneration processes. Figure 5 shows frequency change of five different lectins (WGA, SBA, PNA, UEA-I, Con A) binding to GlcNAcmodified sensor chip surface. As can be seen, the binding of WGA is the highest, at about 20 Hz, indicating that other lectins are basically not bound except for non-specific adsorption.  The binding results between the other three carbohydrates and five different lectins were represented in the Figures 6-8, respectively. As can be seen in Figure 6, Con A has the highest binding with mannose, while the other lectins were basically not bound except that a small amount of nonspecific adsorption was detected. As can be seen in Figure 7, UEA-I has the highest binding with fucose, while the other lectins were basically not bound except that a small amount of non-specific adsorption was detected. As can be seen in Figure 8, SBA has the highest binding with galactose, followed by PNA. This is because both SBA and PNA can bind to galactose, but the binding strengths are different, while the other lectins are not bound except for a small amount of non-specific adsorption. The experimental results show that the interactions between carbohydrate and lectin are specific, which are consistent with the results obtained by other methods, indicating the universality of this method.   The binding results between the other three carbohydrates and five different lectins were represented in the Figures 6-8, respectively. As can be seen in Figure 6, Con A has the highest binding with mannose, while the other lectins were basically not bound except that a small amount of nonspecific adsorption was detected. As can be seen in Figure 7, UEA-I has the highest binding with fucose, while the other lectins were basically not bound except that a small amount of non-specific adsorption was detected. As can be seen in Figure 8, SBA has the highest binding with galactose, followed by PNA. This is because both SBA and PNA can bind to galactose, but the binding strengths are different, while the other lectins are not bound except for a small amount of non-specific adsorption. The experimental results show that the interactions between carbohydrate and lectin are specific, which are consistent with the results obtained by other methods, indicating the universality of this method.   Experiments were repeated three times, and the error bar gives a general idea of how precise a measurement is. Mannose, fucose, galactose and GlcNAc were used to study specific binding with lectins under the same concentration (100 μg/mL). As shown in Figure 9, the error bar results showed that each carbohydrate and its corresponding lectin can bind specifically, illustrating the feasibility of this method.

Kinetic and Affinity Studies
We used an Attana QCM Cell-A200 biosensor to perform in situ real-time measurements of the binding kinetics of the interactions between lectin and carbohydrate sensor. To study the kinetic properties of the interactions between UEA-I and fucose, a diluted solution (25, 50, 75 and 100 μg/mL) of UEA-I in the running buffer was injected to the surface. The binding was monitored in real time (the binding responses are shown in Figure 10). The overall fit was performed using a theoretical 1:1 interaction model with limitations of mass transport and employing Attana QCM Cell-A200 instrument evaluation software. Furthermore, the sensor surface was treated with 25-100 μg/mL of UEA-I, then the instrument recorded the response (black line). The Attana evaluation software generated a theoretical 1:1 fit that was covered (red line). We obtained constants of interactions between UEA-I and fucose, such as the association rate constant (Kon = 1.66 × 10 4 M −1 s −1 ), the dissociation rate constant (Koff = 1.27 × 10 −3 s −1 ) and affinity constant (KD = 76.5 nM). Experiments were repeated three times, and the error bar gives a general idea of how precise a measurement is. Mannose, fucose, galactose and GlcNAc were used to study specific binding with lectins under the same concentration (100 µg/mL). As shown in Figure 9, the error bar results showed that each carbohydrate and its corresponding lectin can bind specifically, illustrating the feasibility of this method. Experiments were repeated three times, and the error bar gives a general idea of how precise a measurement is. Mannose, fucose, galactose and GlcNAc were used to study specific binding with lectins under the same concentration (100 μg/mL). As shown in Figure 9, the error bar results showed that each carbohydrate and its corresponding lectin can bind specifically, illustrating the feasibility of this method.

Kinetic and Affinity Studies
We used an Attana QCM Cell-A200 biosensor to perform in situ real-time measurements of the binding kinetics of the interactions between lectin and carbohydrate sensor. To study the kinetic properties of the interactions between UEA-I and fucose, a diluted solution (25, 50, 75 and 100 μg/mL) of UEA-I in the running buffer was injected to the surface. The binding was monitored in real time (the binding responses are shown in Figure 10). The overall fit was performed using a theoretical 1:1 interaction model with limitations of mass transport and employing Attana QCM Cell-A200 instrument evaluation software. Furthermore, the sensor surface was treated with 25-100 μg/mL of UEA-I, then the instrument recorded the response (black line). The Attana evaluation software generated a theoretical 1:1 fit that was covered (red line). We obtained constants of interactions between UEA-I and fucose, such as the association rate constant (Kon = 1.66 × 10 4 M −1 s −1 ), the dissociation rate constant (Koff = 1.27 × 10 −3 s −1 ) and affinity constant (KD = 76.5 nM).

Kinetic and Affinity Studies
We used an Attana QCM Cell-A200 biosensor to perform in situ real-time measurements of the binding kinetics of the interactions between lectin and carbohydrate sensor. To study the kinetic properties of the interactions between UEA-I and fucose, a diluted solution (25,50,75 and 100 µg/mL) of UEA-I in the running buffer was injected to the surface. The binding was monitored in real time (the binding responses are shown in Figure 10). The overall fit was performed using a theoretical 1:1 interaction model with limitations of mass transport and employing Attana QCM Cell-A200 instrument evaluation software. Furthermore, the sensor surface was treated with 25-100 µg/mL of UEA-I, then the instrument recorded the response (black line). The Attana evaluation software generated a theoretical 1:1 fit that was covered (red line). We obtained constants of interactions between UEA-I and fucose, such as the association rate constant (K on = 1.66 × 10 4 M −1 s −1 ), the dissociation rate constant (K off = 1.27 × 10 −3 s −1 ) and affinity constant (K D = 76.5 nM). Similarly, in order to obtain the kinetic information of the interaction between Con A and mannose, a diluted Con A solution (6.25, 12.5, 20 and 40 μg/mL) was injected to the surface of the sensor. A global data fit was performed using Attana QCM Cell-A200 evaluation software, based on a theoretical 1:1 interaction model with a mass transfer limitation. As shown in Figure 11, the black line indicates the response when we injected 6.25-40 μg/mL of Con A to the surface of the sensor. It covers the theoretical 1:1 fit obtained by Attana evaluation software (red line). We obtained the constants of interactions between Con A and mannose, which are the association rate constant (Kon = 2.71 × 10 4 M −1 s −1 ), the dissociation rate constant (Koff = 1.05 × 10 −3 s −1 ) and affinity constant (KD = 38.8 nM).

Conclusions
In this study, a universal method for photochemical covalent fabrication of the underivatized carbohydrates to prepare the carbohydrate sensors based on PFPA modified PDA sensors for realtime detection of carbohydrate-lectin interactions by QCM biosensor was developed. The photochemical fabrication of PFPA was able to prepare carbohydrate sensors by directly inserting the underivatized carbohydrates on the sensor surfaces. The SEM-EDX energy spectrum indicated that the PFPA was successfully connected to the sensor surface. Four different underivatized carbohydrates were immobilized on the PFPA modified PDA surfaces by photochemical fabrication to prepare the carbohydrate sensors for further detecting their interactions with five different lectins (WGA, SBA, PNA, UEA-I, Con A). The experimental results show that the carbohydrate-lectin interactions are specific, which are consistent with the results obtained by other methods. A kinetic study of the interactions between mannose and Con A, fucose and UEA-I were performed. Our work provides an efficient way for the fabrication of a QCM carbohydrate sensor by underivatized Similarly, in order to obtain the kinetic information of the interaction between Con A and mannose, a diluted Con A solution (6.25, 12.5, 20 and 40 µg/mL) was injected to the surface of the sensor. A global data fit was performed using Attana QCM Cell-A200 evaluation software, based on a theoretical 1:1 interaction model with a mass transfer limitation. As shown in Figure 11, the black line indicates the response when we injected 6.25-40 µg/mL of Con A to the surface of the sensor. It covers the theoretical 1:1 fit obtained by Attana evaluation software (red line). We obtained the constants of interactions between Con A and mannose, which are the association rate constant (K on = 2.71 × 10 4 M −1 s −1 ), the dissociation rate constant (K off = 1.05 × 10 −3 s −1 ) and affinity constant (K D = 38.8 nM). Similarly, in order to obtain the kinetic information of the interaction between Con A and mannose, a diluted Con A solution (6.25, 12.5, 20 and 40 μg/mL) was injected to the surface of the sensor. A global data fit was performed using Attana QCM Cell-A200 evaluation software, based on a theoretical 1:1 interaction model with a mass transfer limitation. As shown in Figure 11, the black line indicates the response when we injected 6.25-40 μg/mL of Con A to the surface of the sensor. It covers the theoretical 1:1 fit obtained by Attana evaluation software (red line). We obtained the constants of interactions between Con A and mannose, which are the association rate constant (Kon = 2.71 × 10 4 M −1 s −1 ), the dissociation rate constant (Koff = 1.05 × 10 −3 s −1 ) and affinity constant (KD = 38.8 nM).

Conclusions
In this study, a universal method for photochemical covalent fabrication of the underivatized carbohydrates to prepare the carbohydrate sensors based on PFPA modified PDA sensors for realtime detection of carbohydrate-lectin interactions by QCM biosensor was developed. The photochemical fabrication of PFPA was able to prepare carbohydrate sensors by directly inserting the underivatized carbohydrates on the sensor surfaces. The SEM-EDX energy spectrum indicated that the PFPA was successfully connected to the sensor surface. Four different underivatized carbohydrates were immobilized on the PFPA modified PDA surfaces by photochemical fabrication to prepare the carbohydrate sensors for further detecting their interactions with five different lectins (WGA, SBA, PNA, UEA-I, Con A). The experimental results show that the carbohydrate-lectin interactions are specific, which are consistent with the results obtained by other methods. A kinetic study of the interactions between mannose and Con A, fucose and UEA-I were performed. Our work provides an efficient way for the fabrication of a QCM carbohydrate sensor by underivatized Figure 11. Kinetic evaluation of the interaction between mannose and Con A.

Conclusions
In this study, a universal method for photochemical covalent fabrication of the underivatized carbohydrates to prepare the carbohydrate sensors based on PFPA modified PDA sensors for real-time detection of carbohydrate-lectin interactions by QCM biosensor was developed. The photochemical fabrication of PFPA was able to prepare carbohydrate sensors by directly inserting the underivatized carbohydrates on the sensor surfaces. The SEM-EDX energy spectrum indicated that the PFPA was successfully connected to the sensor surface. Four different underivatized carbohydrates were immobilized on the PFPA modified PDA surfaces by photochemical fabrication to prepare the carbohydrate sensors for further detecting their interactions with five different lectins (WGA, SBA, PNA, UEA-I, Con A). The experimental results show that the carbohydrate-lectin interactions are specific, which are consistent with the results obtained by other methods. A kinetic study of the interactions between mannose and Con A, fucose and UEA-I were performed. Our work provides an efficient way for the fabrication of a QCM carbohydrate sensor by underivatized carbohydrates based on PFPA modified PDA surfaces, avoiding a tedious modification of carbohydrates, which makes the QCM biosensor analysis more efficient.
Supplementary Materials: The following are available online at http://www.mdpi.com/2073-4360/11/6/1023/s1. Author Contributions: L.G. performed experiments, analyzed data and assisted in drafting the manuscript; S.C. analyzed data and assisted in drafting the manuscript; P.H. and X.L. performed the experiments; Q.S. performed a small part of the experimental work; C.W. participated in the design of the paper and revised the manuscript; Y.P. assisted with the discussion and revision of the manuscript; Z.P. supervised the work presented in this paper and revised the manuscript.