3.1. Principle of Preparation GOx/PQ/Egg White Protein MPs
Encapsulating egg white proteins with a CaCO
3 template involved coprecipitation, where CaCO
3 MPs were formed simultaneously with egg white proteins. In this strategy, egg white proteins were initially dissolved in an aqueous solution containing calcium ions (Ca
2+). Following that, carbonate ions (CO
32−) were added to bind with Ca
2+ and to form solid CaCO
3 particles. These solid particles acted as a trap, encapsulating the egg white protein molecules within them (
Figure 1(Aa)). After the egg white proteins were entrapped inside the CaCO
3 particles, the loaded particles were crosslinked using glutaraldehyde (
Figure 1(Ab)). This process involved treating proteins, mainly ovalbumin, with glutaraldehyde, which formed strong covalent bonds between the amino groups of the protein and the aldehyde groups of glutaraldehyde, resulting in a cross-linked network. The protein in egg white underwent a conformational transition, changing from a water-soluble form to a water-insoluble form. The crosslinked ovalbumin could serve as a scaffold for the subsequent immobilization of enzymes. Subsequently, the sacrificial CaCO
3 MPs were dissolved using an acidic solution, resulting in the formation of egg white protein MPs that maintained the spherical shape of the original CaCO
3 particles (
Figure 1(Ac)). During this step, the CaCO
3 particles broke down, and Ca
2+ ions were released along with carbon dioxide. This process is widely employed in various applications, including material synthesis, because an acid solution can selectively dissolve and remove CaCO
3 without affecting other materials or surfaces. It is an inexpensive and easy-to-implement method [
21].
To showcase an application of the food-based protein MPs, a glucose biosensor was presented as an example. The subsequent functionalization of essential components, including redox molecules and GOx, into the egg white protein MPs, was demonstrated.
After achieving crosslinked egg white protein MPs, the particles were incubated with PQ (
Figure 1(Ad)). PQ is an organic compound that serves as a redox mediator, facilitating electron transfer between the electrode surface and the enzyme. It consists of a phenanthrene ring fused with a quinone moiety [
22]. Quinones belong to a class of organic compounds characterized by a conjugated cyclic system with two carbonyl groups [
23,
24]. They can undergo reversible reductions, both one-electron and two-electron-based, resulting in the generation of a semiquinone radical or a quinone, respectively [
25]. The mediated function of a quinone compound plays a critical role in enhancing the efficiency and sensitivity of biosensing applications. It facilitates the transfer of electrons involved in electrochemical reactions. Moreover, a mediator-based glucose biosensor improves electron transfer, reduces interference effects, and mitigates the impact of oxygen compared with the first generation of enzymatic glucose biosensors [
26]. Therefore, our sensor utilized the second generation of enzymatic glucose sensors with PQ as the mediator.
To create glucose-sensitive MPs, we immobilized PQ/egg white protein MPs with GOx, an enzyme that catalyzes the oxidation of glucose (
Figure 1(Ae)). This immobilization technique improved the stability of the enzyme, making it well-suited for glucose biosensors. GOx is composed of two identical 80 kDa subunits and is associated with two non-covalently bound flavin adenine dinucleotides. The flavin adenine dinucleotides coenzyme serves as an electron carrier and facilitates the oxidation of D-glucose. This reaction yields D-gluconolactone and H
2O
2 as products, while the mediator PQ facilitates electron transfer between the enzyme and the electrode surface (
Figure 1B). Once the PQ/GOx/egg white protein MPs were obtained, they were ready to be applied in glucose biosensor applications.
3.2. Characterization and Morphology
Optical microscopy was employed to examine the morphologies of GOx/PQ/egg white protein MPs once the CaCO
3 template was eliminated. This method enabled visual observation and analysis of the physical structures and shapes of the MPs.
Figure 2A shows that egg white protein MPs exhibit a uniform spherical shape with a narrow size distribution, characterized by an average diameter of 5 µm. Additionally, the corresponding CaCO
3 MPs loaded with egg white proteins before template removal demonstrate a similar size distribution, with an average diameter of 5 µm (
Figure S1).
The morphologies of egg white protein MPs loaded with PQ and GOx were examined using scanning electron microscopy (SEM) (
Figure 2B). The GOx/PQ/egg white protein MPs exhibited a spherical structure. During the coprecipitation process involving CaCl
2 and Na
2CO
3, nucleation led to the formation of granules, which subsequently assembled into interconnected spherical structures with pores. This arrangement facilitated the encapsulation of many molecules of egg white proteins within the template. The uniform shape and small size of these spherical MPs offer advantages such as a high surface-to-volume ratio, enabling enhanced contact with target substances and electrode surfaces. Additionally, their homogeneity ensures consistent performance.
By studying changes in the fluorescence emission spectrum with excitation wavelength at 300 nm, it is possible to examine molecular interactions such as binding to a target molecule or undergoing conformational changes (
Figure 2C). We investigated the fluorescence emission spectra of raw egg white, egg white protein MPs, PQ/egg white protein MPs, and GOx/PQ/egg white protein MPs, under an emission wavelength range of 315–550 nm. The fluorescence emission spectra reveal that crosslinking with glutaraldehyde did not significantly alter the spectral shape of fluorescence emitted from egg white proteins (
Figure 2(Ca)), compared with crosslinked egg white protein MPs (
Figure 2(Cb)).
Figure 2(Cc) shows the shift in the PQ/egg white protein MP signal to a lower wavelength position compared with the peak signal of PQ alone (
Figure S2). Note that a peak on the shoulder around 380 nm could be observed in
Figure 2(Cc), while a peak at 420 nm was observed in free PQ (
Figure S2). This observation suggests that PQ molecules, located within the protein scaffold, may be sensitive to the environment provided by the egg white protein MPs [
27], resulting in the observed hypsochromic shift. It also indicates that the PQ molecules were securely loaded within the 3D crosslinked structure of the egg white proteins. The emission peak of the GOx/PQ/egg white protein MPs was observed to undergo almost no shift (
Figure 2(Cd)) when compared with the spectrum of PQ/egg white protein MPs, likely because of the dominance of fluorescence emission from GOx that overwhelms the fluorescence emission from PQ/egg white protein MPs.
Furthermore, we employed Fourier transform infrared spectroscopy (FTIR) to confirm the formation of GOx/PQ/egg white protein MPs, which provided information about functional groups and various bonds. Specifically, the IR spectra of egg white proteins prior to crosslinking with glutaraldehyde showed a broad peak at 3278 cm
−1, corresponding to O-H stretching and N-H stretching, which are characteristic functional groups of proteins (
Figure 2(Da)). Moreover, it showed a peak of C=O stretching (at around 1633 cm
−1) and N-H bending (at around 1548 cm
−1). After the crosslinking step using glutaraldehyde, the intensity of these two peaks decreased (at 1635 and 3346 cm
−1) (
Figure 2(Db)). Additionally, the C-H stretching (at around 2933–2959 cm
−1), N-H stretching (at around 2874 cm
–1), and C-H bending (at around 1380–1465 cm
–1) were prominent when crosslinking with egg white proteins. This information suggested a successful crosslinking reaction between the egg white proteins and glutaraldehyde. The IR spectrum of PQ/egg white protein MPs showed a O-H stretching peak at around 3323 cm
−1 (
Figure 2(Dc)). Furthermore, these PQ/egg white protein MPs showed increasing intensity of the C-H stretching peak at around 2873–2958 cm
−1. Moreover, the C-H bending peak (at around 1380–1465 cm
−1) was still present with a slight increase in intensity. This suggests successful loading of PQ into the egg white protein MPs. After functionalizing the GOx (prepared in egg white proteins) in PQ/egg white protein MPs, the FTIR analysis revealed two distinct peaks: a O-H stretching peak and C=O stretching peaks at around 3280 and 1640 cm
−1, respectively (
Figure 2(Dd)). These two peaks were more prominent due to the addition of GOx along with additional egg white proteins. Therefore, the addition of GOx and egg white proteins enhanced the intensity of the O-H stretching and C=O stretching peaks, attributable to the functional groups present in GOx and egg white proteins. Additionally, this large addition obscured other peaks associated with those particular functional groups.
3.3. Electrochemical Characterizations
To assess the electrochemical performance of the egg white protein MPs and PQ-containing MPs, CV was employed. Three distinct working electrodes were investigated: (a) a bare screen-printed electrode, (b) an electrode coated with egg white protein MPs, and (c) an electrode coated with PQ/egg white protein MPs. The CV curve in
Figure 3A, obtained at a scan rate of 10 mV s
−1, reveals the background current of the screen-printed electrode when evaluating in 0.1 M PBS at pH 7.0, serving as the supporting electrolyte (indicated by (a)). This indicates the stability of the screen-printed electrode within a potential window ranging from −0.40 to 0.10 V. The observed increase in cathodic current may be attributed to the oxygen reduction reaction occurring at a higher cathodic potential (initiating around −0.20 V).
The impact of loading PQ molecules in MPs on the electrochemical properties was investigated. The PQ is a redox mediator as it can undergo reversible oxidation and reduction reactions, facilitating electron transfer during electrochemical processes [
22]. The inclusion of PQ in egg white protein MPs resulted in the emergence of distinctive redox peaks in the voltammogram, corresponding to the oxidation and reduction of PQ (represented by the blue line, curve c in
Figure 3A), in contrast to the screen-printed electrode and egg white-protein-MP-modified electrodes (curves a and b). The oxidation and reductions peak were observed at −0.14 V and −0.20 V. This notable increase in peak currents could be attributed to the redox activity of PQ. The well-defined shapes of the redox peaks in
Figure 3A, curve c, indicate the two-electron reduction and oxidation process of PQ at the electrode modified with the PQ-containing MPs [
28]. Note that PQ molecules contain high aromaticity, leading to π–π stacking interactions between PQ and the screen-printed carbon nanomaterials, particularly at sp
2 carbon in CNTs, ensuring a strong and effective interaction. This suggests the successful immobilization of PQ on egg white protein MPs and the redox performance of the modified electrode using the synthesized MPs.
Furthermore, we estimated the capacitance of the electrodes by considering the cyclic voltammogram (CVs). For more details, refer to
Equation (S1), Equation (S2), and Supporting Note S1. This evaluation enables us to determine the amount of charge involved in the electrochemical reaction. The specific capacitance values of screen-printed CNT-modified electrode, egg white protein MPs, and PQ-containing MPs were 0.2, 0.2, and 1.5 mF cm
−2, respectively. This result demonstrates that the attachment of PQ enhances the charge storage capabilities through successful reversible redox reactions on the electrode.
Figure 3B illustrates the CVs of screen-printed electrode coated with PQ-containing MPs, with a scan rate ranging from 2.5 to 200 mV s
−1 in 0.1 M PBS pH 7.0. These plots reveal the presence of quasi-reversible PQ redox couples, and the anodic/cathodic peak current ratio (I
c/I
a ratio) approaches unity, ranging from 1.0 to 1.1. The CV shapes remain well preserved, indicating that the screen-printed electrode coated with PQ-containing MPs exhibits good rate performance. Furthermore, as depicted in
Figure 3C, the resulting anodic and cathodic peak currents show a proportional relationship to the square root of the scan rate. This behavior aligns with the Randles–Sevcik equation [
29] and suggests that the processes occurring at the PQ/egg white-protein-MP-modified electrode are diffusion-controlled, consistent with previous studies conducted on a PQ-modified electrode [
30].
Additionally, we studied the effect of scan rate on the surface capacitance by considering the area under the cyclic voltammogram, with PQ serving as the redox-active molecules attached to the electrode (
Figure S3A,B). We observed a decrease in capacitance from 2.64 to 0.56 mF cm
−2 when the scan rate increased from 2.5 to 200 mV s
−1. Faster scan rates during CV resulted in a decrease in capacitance because of the limited time available for redox reactions to occur at the electrode surface.
Additionally, we measured CVs using the bare electrode, egg white protein MPs, and PQ-containing MPs in 0.1 M KCl, with and without an additional external redox. In the KCl electrolyte, the PQ-containing MPs exhibited a higher redox peak current compared with the bare electrode and egg white protein MPs. This difference is due to the presence of PQ as a redox mediator (see inset
Figure S6A). When the solution contained an external redox probe, we observed a relatively smaller redox peak current when using the electrode modified with egg white protein MPs and PQ-containing MPs, compared with the bare electrode (
Figure S6A). This decrease can be attributed to the insulating nature of egg white protein and the GOx enzyme, which restricts the accessibility of the external redox probe in reaching the electrode. Furthermore, we compared the active surface area of the pristine commercial conductive ink with a screen-printed CNT-modified electrode (see
Figure S6B and Supporting Note S2). The redox peak of the screen-printed electrode was higher than that of the commercial conductive electrode. The active area of the screen-printed electrode was 0.12 cm
−2, while that of the commercial electrode was 0.09 cm
−2. This indicates that the modification can increase the active area by approximately 25% compared with the pristine electrode. Additionally, the EIS measurements showed that our ink composed of CNT exhibited a lower resistance value than the commercial ink (see
Figure S7). This enhancement can be attributed to improved electron transfer.
To evaluate the bioelectrocatalytic characteristics of the GOx/PQ-based MPs, while minimizing the contribution of capacitive current resulting from coated materials and the large electroactive area of the CNT-based electrode, the LSV technique was performed at a low scan rate of 5 mV s
−1. The LSV technique was used to vary the potential of the screen-printed electrode coated with GOx/PQ-based MPs linearly with time while measuring the resulting current. The voltammograms shown in
Figure 3D (dashed line) display an anodic peak around 0.15 V, indicating the oxidation of PQ molecules in the absence of glucose. This voltammogram reflects the oxidation behavior of PQ. In the presence of glucose, the oxidation peak current increased (
Figure 3D, solid lines). The anodic peak observed during LSV was higher compared with the absence of glucose, indicating the effective catalytic action of GOx, functionalized in the synthesized MPs. The presence of glucose led to an increased anodic peak in the LSV, enabling quantitative measurement of the glucose levels. Amperometry was further employed to demonstrate glucose biosensors using these GOx/PQ-based MPs (
Section 3.4 Amperometry). This observation confirms the successful electrochemical behavior of a screen-printed electrode coated with GOx/PQ-based MPs upon the addition of glucose. This is a result of effective enzyme immobilization and PQ mediation within the microspheres.
In addition to DC techniques, EIS is one of the techniques that has been widely used, particularly in chemical and biosensing applications [
31]. In this study, EIS was conducted to evaluate the bioelectrochemical performance of the synthesized GOx-based MPs towards changes in glucose concentration. This evaluation was performed by comparing a bare electrode (
Figure 4A) to a screen-printing electrode coated with GOx/PQ/egg white protein MPs (
Figure 4B). Phase angle plots, as shown in
Figure 4(Aa,Ba), were observed across a frequency range of 10
0 to 10
4 Hz and with varying glucose concentrations ranging from 0.5 to 10 mM. The phase angle values were found to be around 50–75° over a frequency range of 10
0 to 5 × 10
1 Hz, which is lower than the expected 90° for ideal capacitive behavior [
32].
Figure 4(Ab) displays a Bode plot including the total impedance (
) as a function of the logarithm of the frequency depicting different glucose concentrations. The slope of the bare screen-printed electrode was determined to be −0.8 based on data at a frequency of 10
0–10
2 Hz. Additionally, the inset calibration graph in
Figure 4(Ab) illustrates the relationship between
at a frequency of 10
4 Hz and glucose concentration. The inset plot shows that the impedance signal obtained from the bare electrode does not depend on changes in glucose concentration due to the absence of glucose-sensitive particles.
When comparing the performance of a screen-printed CNT-modified electrode to an electrode modified with GOx/PQ/egg white protein MPs, the impedance signal increased with increasing glucose concentration from 0.5 to 10 mM. The slopes of the
vs.
graph obtained for screen-printing electrodes coated with GOx/PQ-based MPs were determined to be −0.7 (over a range of 10
0 to 10
2 Hz) (
Figure 4(Bb)), indicating pseudocapacitive behavior in the mid-frequency region. At high-frequency regions (over around 300 Hz), the slopes approached zero, indicating resistive behavior at those frequencies [
32]. Furthermore, the calibration graph (shown in the inset in
Figure 4(Bb)), which depicts the correlation between the logarithm of impedance (log
Z) recorded at a high frequency of 10
4 Hz and glucose concentration, exhibited an upward trend in the impedance signal as the glucose concentration increased. When using GOx/PQ/egg white protein MPs, the sensitivity for glucose detection (measured in mM
−1 unit) was enhanced by approximately 43 times compared with the bare electrode. These results confirm the glucose-sensitive behavior of the synthesized GOx-based protein MPs. Additionally, these findings highlight the beneficial effects of electrode modification in improving the electrochemical performance of glucose sensing applications. Additionally, we compared the Bode plot generated from experimental data with the fit of the EIS data (
Figure S4). The Randles circuit was applied to simulate the EIS data (
Figure S5). The circuit consists of a resistor (
Rel) representing the electrolyte resistance, a constant phase element (
CPEdl), or a capacitor (
C) for the double layer for non-ideal behavior at the electrode–electrolyte interface, and a Warburg element (
WD) describing the semi-infinite diffusion of the species in the electrolyte towards the electrode surface.
3.4. Amperometry
Investigating the electrochemical enzyme kinetics of redox-mediated GOx reactions is crucial as it helps determine the enzyme’s affinity for glucose and its maximum catalytic efficiency. In our study using a heterogeneous system, we employed an amperometric technique to examine the interaction between GOx and glucose (substrate) and to measure the rate of electron generation over time (
Figure 5A). By incrementally adding glucose concentrations ranging from 0.125 to 40 mM while maintaining a potential of 0.20 V vs. Ag/AgCl, we correlated the current responses with the glucose concentration (
Figure 5B). Moreover, the calibration in the inset of
Figure 5B shows a linear range of glucose concentration from 0.125 to 10 mM with a sensitivity of 0.008 µA mM
−1. The Michaelis–Menten constant (
) serves as an important indicator of enzyme-substrate kinetics. The results showed that
was 4.6 mM, indicating that the immobilized GOx exhibited high enzymatic activity and that the proposed electrode had a strong affinity for glucose. The current responses exhibited the characteristic features of the Michaelis–Menten kinetic mechanism (Equation (1)). In enzymes that follow the Michaelis–Menten mechanism, increasing substrate concentrations in the initial phase led to a rapid increase in the current, followed by a gradual increase as the enzyme approaches its maximum activity. At high substrate concentrations, the enzyme becomes saturated, and the maximum achieved current reflects the enzyme–substrate complex being fully formed. Thus, the resulting curve represents the kinetic parameters defining the upper and lower boundaries of substrate concentration. Therefore,
Km serves as an important parameter reflecting the enzyme–substrate affinity and influences the reaction velocity at various glucose concentrations.
where
is the steady-state current after the addition of glucose,
is the maximum current obtained from saturated glucose concentrations,
is the glucose concentration, and
is the Michaelis–Menten constant.
To validate our results and to obtain a deeper comprehension of the enzymatic properties, we compared the
Km value calculated for our biosensor with values reported in the literature for similar enzyme–substrate systems on amperometric biosensors. For example, in the comparison of
Km values, the encapsulation of GOx within spindle-like copper hydroxysulfate nanocrystals facilitated the precipitation Cu
2+ ions and biomimetic mineralization of brochantite [
33]. The
Km of the free enzyme was compared with that of GOx@copper hydroxysulfate nanocrystals, yielding
Km values of 19.01 mM and 14.39 mM, respectively. Another example, a CNT-chitosan-nanowire electrode was immobilized with GOx, resulting in tight immobilization within the matrix through adsorption [
34]. The enzyme electrode exhibited a low
Km value of 7.1 mM, which can be attributed to synergistic augmentation from CNT, chitosan, and the GOx matrix. The
Km value of these papers were higher than that in our work, which obtained a weaker binding affinity between the enzyme and substrate, requiring higher substrate concentrations to reach half of the maximum reaction velocity. Additionally, a glucose sensor was developed using zinc oxide nanowires immobilized with GOx, which led to a
Km value of 4.1 mM [
35]. The observation of similar
Km values with our work in different enzyme–substrate systems suggests comparable affinities between immobilized GOx and glucose.
Investigating the selectivity of the new biosensor for glucose is crucial due to the presence of various physiologically significant interferences in real samples, such as a saliva sample that contains glucose, uric acid, lactate, ascorbic acid, and creatinine. To determine the biosensor’s selectivity,
Figure 5C,D were used to assess the impact of different potential interferences on the response to 2 mM uric acid [
36], 12.5 mM lactate [
36], 0.5 mM creatinine [
37], and 100 mM ascorbic acid [
38]. These used concentrations were around five times as high as those informed in the literature. The selectivity test results indicate that the presence of coexisting compounds has little effect on the glucose response. For this experiment, the GOx/PQ/egg white protein MPs, based on a screen-printed electrode, could measure glucose using amperometry techniques at a low applied voltage in electrochemical analysis, resulting in reduced electrical input and power consumption [
39]. Utilizing low applied voltages can help prevent electrode degradation and enhance electrode stability, as high voltages may induce side reactions [
40]. Additionally, low voltages can enable better control over the desired reaction and enhance process selectivity, as certain reactions may exhibit different mechanisms or pathways at high voltages, leading to unwanted by products or decreased selectivity due to interference effect from substances, e.g., ascorbic acid and uric acid [
41]. Furthermore, functionalizing enzymes in MP environments allows for achieving a high level of specificity [
42]. When enzymes are immobilized or encapsulated within MPs, the enzymatic reactions occur in a controlled and confined environment. This microenvironment enhances the enzyme’s specificity by facilitating substrate recognition and catalytic activity. The MP environment provides protection to the enzymes from harsh conditions or interference from other molecules, thereby increasing their stability and preserving their specificity. Therefore, the use of MP environments offers significant advantages in attaining and maintaining high enzyme specificity across glucose biosensor applications.