An ultrasensitive nanobiohybrid platform for glucose electrochemical biosensing based on ferrocenyl iminopropyl-modified silica nanoparticles

Abstract

A novel nanobiohybrid platform based on ferrocenyl iminopropyl-modified silica nanoparticle conjugates (fap-SiNPs), entrapped in glucose oxidase (GOx) and bovine serum albumin cross-linked with glutaraldehyde, was developed. The as-synthesized nanobiohybrid composite was cast onto a p-aminothiophenol-modified gold electrode surface to form the platform for glucose electrochemical biosensing. Under optimal conditions, the device was able to detect glucose with a detection limit of 0.68 nM and a linear range from 1 nM to 100 μM. The proximity of these three components: fap-SiNPs, aminothiophenol and GOx enhanced the electron-transfer between film and electrode. The applicability of the platform was tested on glucose detection in saliva. Furthermore, using HRP instead of GOx, the platform can be used to detect 1 pM of H₂O₂.

---

Introduction

According to a recent WHO report,¹ approximately 422 million people around the world are suffering from diabetes, which is a highly extensive disease related to metabolic disorders. For non-diabetic persons, blood glucose ranges from 3.6 to 7.5 mM (65–135 mg dL⁻¹), while it is, for diabetic patients, between 1.1 and 20.8 mM (20–350 mg dL⁻¹).² The glucose concentration in saliva is much lower, usually ranging from 30 to 80 μM, which is beneath the detection limits of most commercially available blood glucose sensors. In the last few years, intense research efforts have been devoted to finding a correlation between glucose levels in blood and saliva.³ The aim is to develop biosensors for monitoring glucose levels in saliva which, in fine, will result in the elimination of the pain and discomfort associated with blood sampling.

Glucose oxidase-based amperometric biosensors rely on the immobilization of the enzyme on the surface of various electrodes. However, it is hard to achieve direct electron-transfer (ET) of GOx at bare and conventional electrodes because the enzyme redox-active site is deeply buried inside the protein shell.⁴ This point implies the application of high positive potentials. Furthermore, the sensitivity of these biosensors can be significantly improved by the addition of mediators in the matrices.⁵ However, leakage has been a main problem for the entrapment of mediators due to their low molecular weight in these matrices. This problem can be circumvented by covalently linking the mediator with nanomaterials or with high molecular weight compounds before immobilization.⁶ Enzyme immobilization on an inorganic solid was extensively investigated, this method improves the stability of enzymes under extreme conditions.⁷ Other studies have reported that silica-based organic–inorganic hybrids are adopted as attractive composite materials for the fabrication of biosensors.⁸

Various kinds of reagentless enzyme-based biosensors were also constructed by co-immobilizing silica nanoparticles and redox mediators in a matrix of the biopolymer, bovine serum albumin,⁹ which facilitated ET processes. Herein, we report the design of a nanobiohybrid platform for glucose detection and the study of its analytical performances. The nanobiohybrid platform was designed by cross-linking glucose oxidase and bovine serum albumin with glutaraldehyde, entrapping ferrocenyl-iminopropyl-modified silica nanoparticle conjugate (fap-SiNPs), and casting it on a gold electrode, which allowed rapid ET and reproducible measurement of glucose. When HRP is used instead of GOx, the platform can be used to sensitively detect hydrogen peroxide.

Experimental

Materials and reagents

All reagents: glucose oxidase (GOx) from aspergillus niger (type X–S, lyophilized powder, 117 kU g⁻¹), Horseradish peroxidase (HRP), bovine serum albumin (BSA), tetraethoxysilane (TEOS, 99.9%), 3-aminopropyltriethoxysilane (APTES, 99%), ferrocene carboxyaldehyde (98%), ammonium hydroxide (aqueous NH₃, 25–28% min), absolute ethanol, hydrogen peroxide 30%, 4-aminothiophenol (4-ATP), phosphate-buffered saline (PBS) tablets and glutaraldehyde (GL) were purchased from Sigma-Aldrich (France) and used as received. Ultrapure water (>18 MΩ cm) obtained using a milli-Q system was used to prepare all the solutions. PBS solutions (pH = 7.4) were prepared by dissolving PBS tablets in deionized water according to manufacturer recommendations.

Gold substrates were fabricated through the French National Nanofabrication Network RENATECH by LAAS, CNRS Toulouse. They were fabricated using standard silicon technologies. (100)-oriented, p-type (3–5 Ω cm) silicon wafers were thermally oxidized to grow an 800 nm-thick silicon oxide layer. Then, a 30 nm-thick titanium layer and a 300 nm thick gold top layer were successively deposited by evaporation under vacuum. The gold electrodes were cut into wafers as 1.2 × 1.2 cm² square plates.

Electrochemical measurements

The electrochemical behavior of the biosensor was investigated by cyclic voltammetry (CV), square wave voltammetry (SWV) (optimal parameters: initial potential = −0.8 V, end potential = 0.8 V, step potential = 1 mV, amplitude = 20 mV, frequency = 10 Hz). The experiments were carried out in aqueous solutions using a conventional three-electrode system with the modified or bare gold substrate as the working electrode (surface 0.19 cm²), a platinum plate as the auxiliary electrode (surface 0.38 cm²). A saturated calomel electrode (SCE) was employed as a reference electrode.

A Voltalab80 impedance analyzer (Radiometer Analytical, France) was used for electrochemical measurements. An equimolar ferrocyanide/ferricyanide ([Fe(CN)₆]^4/3−) mixture (5 mM) in N₂-purged PBS solution was used for voltammetry studies in aqueous media. All measurements were performed in a 5 mL glass cell at room temperature (RT).

Atomic force microscope (AFM) images were obtained with a SPM controller and a Park Systems XE-70 atomic force microscope (Park Systems, Japan). The images were collected in contact mode.

Transmission electron microscopy (TEM) analysis was performed using a JEOL JEM-2100 microscope. Scanning electron microscopy (SEM) images were obtained by using field-emission scanning electron microscopy (SEM) using a JEOL JSM-6335F microscope. UV-visible spectra of the various types of silica were recorded using Unico SpectroQuest 2802 UV-vis spectrophotometer as a suspension in absolute ethanol.

Synthesis of the ferrocenyl-modified silica nanoparticles

Aminopropyl-functionalized silica nanoparticles (ap-SiNPs) were prepared using the Stöber's modified method.¹⁰ Briefly, 4 mL of TEOS and 4 mL of ammonium hydroxide were added to 47 mL of ethanol under stirring for 24 h. After purification the resulting silica colloidal dispersion was further functionalized with APTES by quickly adding 0.3 mL of APTES under vigorous stirring, and then kept overnight. The nanoparticles were purified through centrifugation and redispersion into ethanol.

The procedure to prepare the ferrocenyliminopropyl-silica nanoparticles is summarized in Scheme 1. The fap-SiNPs were prepared from the condensation of ferrocene carboxaldehyde and aminopropyl-modified silica nanoparticles. Briefly, ferrocene carboxaldehyde was added to a dispersion of fap-SiNPs into ethanol and the mixture was stirred for 24 h at RT. The resulting fap-SiNPs solid were collected by filtration and washed several times with ethanol until no ferrocene leaked from the as-prepared reddish solid. The covalent linkage between ferrocene carboxyaldehyde and the amine groups was verified with ninhydrin test, giving a purple coloration with the ap-SiNPs, which is no observed with the fap-SiNPs (data not shown).

Scheme 1 Route for the preparation of the ferrocenyliminopropyl-silica nanoparticles.

Preparation of the biosensor

The gold electrode surface was cleaned in acetone under sonication for 10 min, dried under a nitrogen flow then was dipped for 2 min into a piranha solution 4 : 1 (v/v) (98% H₂SO₄/30% H₂O₂) and then rinsed with ultrapure water and dried under nitrogen flow. The as-cleaned electrode was dipped for 24 h into 4-ATP ethanolic solution (1 mM) shielded from light by aluminum foil. Afterwards, a solution was prepared by dissolving 3 mg of GOx, 3 mg of BSA and 0.7 mg of fap-SiNPs in PBS. The above mixture was applied to the ATP-modified gold electrode surface. 1 μL of an aqueous 2.5% glutaraldehyde (GL) solution was used to cross-link the proteins to the amino-terminated gold surface, thus entrapping fap-SiNPs. The electrode was further dried for 2 h in air. Finally, the electrode was washed with water to remove non-immobilized biomaterials. The prepared electrodes were stored in dry conditions at 4 °C when not in use. The same protocol was used to prepare the HRP-based biosensor for hydrogen peroxide detection.

Results and discussion

Characterization of the fabricated biosensor

The initial silica nanoparticles were prepared in a straightforward manner using a Ströber's modified method. The surface morphology and particle size of SiNPs were assessed using SEM and TEM microscopy. Fig. 1A and B show that the particles are spherical with an average diameter of ca. 300 nm. The morphology and particle size were preserved after the sequential modification with aminopropylsilane and the ferrocene derivative (images not shown). Furthermore, an EDX analysis of fap-SiNPs revealed the existence of carbon, iron, silicon and oxygen (Fig. 1C), which confirms the presence of the ferrocene, thus proving that the two modification steps occurred successfully.

Fig. 1 (A) SEM, (B) TEM images of the SiNPs, (C) EDX analysis diagram of fap-SiNPs, (D) AFM images of 4-ATP-modified gold electrode and (E) AFM of the SAM-modified surface after addition of fap-SiNPs/BSA/GL.

Furthermore, we monitored the film formation through AFM observations. A typical image of 4-ATP modified gold surface shows a faint texturing of the bare gold surface after incubation with 4-ATP. The 4-ATP forms a very densely packed flat layer covering the gold surface (Fig. 1D). After immobilization of the nanobiohybrid composite on the gold electrode surface, the AFM image shows the presence of spherical objects with a diameter of around 200 nm, revealing the presence of the fap-SiNPs (Fig. 1E).

Electrochemical characterization of the biosensor

In presence of [Fe(CN)₆]^4/3− redox probe, the bare gold electrode displays a reversible voltammogram with two sharp peaks, with a peak-to-peak potential separation of ΔE = 177 mV, corresponding to a monoelectronic electron-transfer reaction from and to the redox probe, thus showing a clean gold surface (curve i, Fig. 2A). After the modification with 4-ATP, almost no faradic currents can be observed which indicates that the 4-ATP-based SAM layer is blocking the ET from the solution to the electrode surface (curve ii, Fig. 2A). The resistance of the electron-transfer occurring at the electrode surface is examined using electrochemical impedance spectroscopy. Fig. 2B shows the Nyquist plots of the bare (i) and modified gold electrode (ii) after immobilization of the 4-aminothiophenol. The immobilization of the latter induces an increase in the semi-circular diameter indicating an increase of the electron-transfer resistance. The increase may be due to the formation of an insulating layer which acts as a barrier that decreases the electron transfer rate of the [Fe(CN)₆]^4/3− redox probe.

Fig. 2 (A) CV and (B) EIS recorded with bare (curve i) and 4-ATP-modified (curve ii) gold electrodes in presence of 5 mM [Fe(CN)₆]^4/3− redox probe.

Electrochemical behaviour of the Fc-Si-modified electrode

CV of the fap-SiNP-modified electrode exhibited a pair of redox peaks located at 0.23 and 0.11 V corresponding respectively to oxidation and reduction of the ferrocene moiety (Fig. 3A). The peak-to-peak potential separation (ΔE_p) is ca. 120 mV at 100 mV s⁻¹, denoting a quasi-reversible behavior of the ferrocene-tethered system. Moreover, an increase of the scan rate shows that both anodic and cathodic currents increase proportionally. Furthermore, the anodic peak current increases linearly with the increase of the scan rate in the scan rate range of 10 to 100 mV s⁻¹ (Fig. 3B), suggesting that the electrochemical reaction occurring on the fap-SiNP-modified electrode is a surface-confined process. Since the ferrocene is confined to the surface, we can expect good electronic communication between the GOx redox cofactor (FAD) and ferrocene. The surface coverage can be obtained from the following equation:

Fig. 3 (A) CVs of fap-SiNPs-modified electrode in 0.1 M PBS (pH 7.4) at 10, 25, 50, 75 and 100 mV s⁻¹ (from internal to external), (B) plot of anodic peak current vs. scan rate, (C) SWVs of bioelectrode in 0.1 M PBS containing different concentrations of glucose. Inset: CVs of the bioelectrode after the addition of 1 μM and 10 mM of glucose showing the electrocatalytic effect and (D) the plots of the electrocatalytic peak current as a function of glucose concentration RSD = 6.5% (n = 4).

Using eqn (1) and from the linear anodic part of the i_p = f(v) diagram, which relates to the reversible process with adsorbed ferrocene species, where i_p, A and Γ are peak current, electrode surface area and the surface coverage of the redox species, respectively, the total surface coverage of the immobilized ferrocene species is calculated to be about 2.886 μmol cm⁻², considering the average of anodic currents. This value is eight times higher than a regular ferrocene-based SAM on gold electrode surface, which will be beneficial for the sensor sensitivity.¹³

Electrochemical response of the bioelectrode to glucose

Since the modified electrode exhibits stable and reversible electrochemical features, it can be used as an ET mediator to shuttle electrons between the enzyme and the electrode surface. GOx was selected as a model enzyme to evaluate the electrocatalytic ability of the bioelectrode. In the presence of glucose, CV of the bioelectrode shows an important increase of the anodic current while the cathodic one has almost disappeared (Fig. 3C). Such behavior is ascribed to a marked catalytic effect where the oxidized ferrocene formed at the electrode, during the ongoing sweep, is immediately reduced by an electron from the enzyme redox cofactor (i.e. FADH₂) before the forward sweep. Two ferrocene units are usually necessary to regenerate the oxidized cofactor of the enzyme, in accordance with eqn (2)–(4).

GOx(FAD) + glucose → GOx(FADH₂) + glucolactone (2)

GOx(FADH₂) + 2Fc⁺ → GOx(FAD) + 2Fc (3)

2Fc (at the electrode) → 2Fc⁺ (at the electrode) (4)

SWV experiments were performed using a modified electrode containing various concentrations of glucose (Fig. 3C).

The linearity range of the biosensor for glucose detection can be estimated from the calibration curve displayed in Fig. 3D. The biosensor shows quite a large range in the very low glucose concentrations comprised between 1 nM and 100 μM with high sensitivity (S) estimated at 1.52 μA cm⁻² μM⁻¹. Furthermore, the limits of detection (LoD) and quantification (LoQ) were calculated respectively as 3σ/S and 10σ/S (σ being the standard deviation on blank and S the sensitivity) and were found to be 0.68 and 2.1 nM, respectively.

The curvature from the initial straight showed the characteristics of Michaelis–Menten kinetics. The apparent Michaelis–Menten constant (K_M), which is a characteristic of the enzyme-substrate kinetics for the biosensor, was calculated to be 0.122 μM using the Lineweaver–Burk equation. This K_M value is much lower than those found for native GOx in solution (27.0 mM)¹¹ or for immobilized GOx in sol–gel organic–inorganic material (20 mM) (Table 1A).¹² The low value indicates that the immobilized enzyme retained its high activity and the diffusion barrier is also low.

Table 1 (A) Performance of bioelectrodes using GOx in various configurations (GCE: glassy carbon electrode; MGF: mesocellular graphene foam; AuNPs: gold nanoparticles; PAN: polyaniline; rGO: reduced graphene oxide; GQDs: graphene quantum dots) and (B) summary of performance of various biosensors using ferrocene as a mediator for the detection of levels of glucose

Biosensors	Linear range (mM)	LoD (μM)	KM (mM)	S (μA cm^−2 mM^−1)	Ref.
(A)
GOx/MGF/GCE	1–12	250	3.2	2.87	14
GOx/rGO-AuNPs/GCE	0.02–2.26	4.1	0.038	3.844	15
GOx/rGO/PAN/AuNPs/GCE	0.004–1.12	0.6	0.6	—	16
GOx/GQDs/GCE	0.005–1.27	1.73	0.76	—	17
(B)
CS–Fc/GO/GOx	0.02–6.78	7.6	2.1	10	18
GOx/MWNTs/CS–Fc	0.02–5.36	6.5	6.67	21.94	19
Fc-PNW	0.01–10	5	—	21.9	20
ASNPs/GOx	—	9	—	5.11	21
GOx/fap-SiNPs/Au	10^−6 to 0.1	0.67 × 10^−3	0.122 × 10^−3	1.52 × 10^−3	This work

---

Interference study and comparison with literature

Fig. 4 shows the amperometric responses to glucose (10 μM) and to ascorbic acid and uric acid added in great excess (viz. 1 mM). These last compounds induce little response; hence, the bioelectrode is endowed with good selectivity for glucose.

Fig. 4 Histogram of the bioelectrode response to glucose (Glu) (10 μM) and for ascorbic acid (AA) and uric acid (UA) (1 mM) in 0.1 M PBS (n = 3).

Moreover, Table 1B compares the as-prepared bioelectrode with various biosensors using ferrocene and glucose oxidase in their components. The present electrode shows better performance and a lower limit of detection than those of the previously published biosensors.

Reproducibility and stability

The reproducibility of the biosensor was examined using a 1 μM concentration of glucose and the relative standard deviation for four consecutive runs with five independently prepared electrodes was less than 7%, which indicated a good reproducibility of the biosensor preparation. The biosensor was used daily for 15 days and stored in 0.1 M PBS with pH 7.0 at 4 °C. The results showed that the glucose biosensor retained 80% of its initial activity. These results indicated an acceptable stability.

Real sample analysis

To verify the practicability of the proposed method for glucose analysis, the biosensor was applied to the determination of glucose in human saliva. The quantitative determination of glucose in human saliva samples was carried out using the standard addition method, and the results are summarized in Table 2. These concentration values in saliva were also found using HPLC/MS/MS. The good recovery ratios of the spiked samples indicate that the as-developed sensor can be used for the detection of low concentrations of glucose in human fluids.

Table 2 Determination of glucose concentration in human saliva samples (n = 2)

Samples	Detected/μM	Added/μM	Found/μM	Recovery%
1	91.0	100.0	190.8	99.8
2	63.2	100.0	164.0	100.8

---

Applicability of the platform

To extend the scope of the developed platform, we selected the HRP redox enzyme to design a new bioelectrode for the detection of hydrogen peroxide. As an electron mediator, the ferrocene is expected to regenerate the reduced form of the HRP after it transfers two electrons to reduce H₂O₂ (eqn (5) and (6)). Ferrocene is subsequently reduced at the electrode.

HRP(red) + H₂O₂ + 2H⁺ → HRP(ox) + 2H₂O (5)

HRP(ox) + 2Fc → HRP(red) + 2Fc⁺ (6)

The designed bioelectrode is sensitive to extremely low concentrations of hydrogen peroxide and can detect up to 1 pM (Fig. 5). This sensor may find many applications in the monitoring of biological reactions.

Fig. 5 (A) SWVs of the HRP-based bioelectrode in 0.1 M PBS containing different concentrations of H₂O₂ and (B) current–density vs. concentration plot for low concentrations of hydrogen peroxide.

Conclusion

In this study, we have demonstrated that this nanobiohybrid based on ferrocene-modified silica nanoparticles can be used as a scaffold to build a biosensor for the sensitive determination of glucose and hydrogen peroxide levels. The high catalytic activity for glucose or HRP electrocatalysis can be attributed to the high surface area of the silica loaded with a large amount of ferrocene. The biosensor is able to detect very low concentrations of glucose over five decades (starting from 1 nM) and was successfully applied to the determination of glucose levels in human saliva samples. Furthermore, the bioelectrode shows enhanced selectivity toward the usual interfering agents present in human fluids such as ascorbic and uric acids. Applicability of the platform was also verified by using HRP to detect 1 pM of H₂O₂.

Acknowledgements

The authors wish to acknowledge the financial support granted to M. Saadaoui by LR99ES15. Thanks also go to Ing. Ghazi Jomaa and Dr Rihab Sahli (INRAP, Tunis, Tunisia) for their valuable help with the acquisition of AFM images.

References

1. Global Report on Diabetes, World Health Organization, Geneva, Switzerland, 2016, ISBN: 9789241565257.
2. D. Ye, G. Liang, H. Li, J. Luo, S. Zhang, H. Chen and J. Kong, Talanta, 2013, 15, 223.
3. A. Soni and S. K. Jha, Biosens. Bioelectron., 2015, 15, 763; C. Liu, Y. Sheng, Y. Sun, J. Feng, S. Wang, J. Zhang, J. Xu and D. Jiang, Biosens. Bioelectron., 2015, 70, 455; I. Al-Ogaidi, H. Gou, A. K. A. Al-kazaz, Z. P. Aguilar, A. K. Melconian, P. Zeng and N. Wu, Anal. Chim. Acta, 2014, 811, 76.
4. S. Liu and H. Ju, Biosens. Bioelectron., 2003, 19, 177.
5. A. E. G. Cass, G. Davis, G. D. Francis, H. A. O. Hill, W. J. Aston, I. J. Higgins, E. V. Plotkin, L. D. L. Scott and A. P. F. Turner, Anal. Chem., 1984, 56, 6671.
6. S. Sharma and S. Kanwar, Sci. World J., 2014, 625258; S. Gupta, A. Bhattacharya and C. N. Murthy, Biocatal. Agric. Biotechnol., 2013, 2, 171–190.
7. O. Nadzhafova, M. Etienne and A. Walcarius, Electrochem. Commun., 2007, 9, 1189; H. Takahashi, B. Li, T. Sasaki, C. Miyazaki, T. Kajino and S. Inagaki, Microporous Mesoporous Mater., 2001, 44, 755–762.
8. K. M. Manesh, P. Santhosh, S. Uthayakumar, A. I. Gopalan and K.-P. Lee, Biosens. Bioelectron., 2010, 25, 1579.
9. K. R. Sharma, J. Biotechnol. Biomater., 2012, 2, 1.
10. V. Mahalingam, S. Onclin, M. Peter, B. J. Ravoo, J. Huskens and D. B. Rein-Houdt, Langmuir, 2004, 20, 11756; T. Schiestel, H. Runner and G. E. M. Tovar, J. Nanosci. Nanotechnol., 2004, 4, 504; W. Stöber, A. Fink and E. J. Bohn, J. Colloid Interface Sci., 1968, 26, 62.
11. M. J. Rogers and K. G. Brandt, Biochemistry, 1971, 10, 4624.
12. B. Q. Wang, B. Li, Q. Deng and S. Dong, Anal. Chem., 1998, 70, 3170.
13. R. Sahli, C. Fave, N. Raouafi, K. Boujlel, B. Schöllhorn and B. Limoges, Langmuir, 2013, 29, 5360.
14. Y. Wang, H. Li and J. Kong, Sens. Actuators, B, 2014, 193, 708.
15. X. Cao, Y. Ye, Y. Li, X. Xu, J. Yu and S. Liu, J. Electroanal. Chem., 2013, 697, 10.
16. Q. Xu, S.-X. Gu, L. Jin, Y.-e. Zhou, Z. Yang and W. Wang, Sens. Actuators, B, 2014, 190, 562.
17. H. Ramzi and R. Mohammad-Rezaei, Biosens. Bioelectron., 2013, 41, 498.
18. J.-D. Qiu, J. Huang and R.-P. Liang, Sens. Actuators, B, 2011, 160, 287.
19. R. P. Liang, L. X. Fan, R. Wang and J. D. Qiu, Electroanalysis, 2009, 21, 1685.
20. J. Wang, D. Li, M. Yang and Y. Zhang, Anal. Methods, 2014, 6, 7161.
21. Y. Sun, F. Yan, W. Yang and C. Sun, Biomaterials, 2006, 27, 4042.

---