Multi-purpose electrochemical biosensor based on a “green” homobifunctional cross-linker coupled with PAMAM dendrimer grafted p-MWCNTs as a platform: application to detect α2,3-sialylated glycans and α2,6-sialylated glycans in human serum

Abstract

Sialylated glycans are crucial molecular targets for cancer diagnosis and clinical research. α2,3-Sialylated glycans and α2,6-sialylated glycans are the predominant sialic acids found in nature. Different expression of the quantity of glycans can result in development of different disease. However, there are no ideal methods for discriminating α2,3-sialylated glycans and α2,6-sialylated glycans. In this work, a multi-purpose biosensor is fabricated for sensitive detection of α2,3-sialylated glycans and α2,6-sialylated glycans. To improve the sensitivity of the biosensor, p-MWCNTs were integrated with PAMAM, as PAMAM has highly branched and abundant amino groups, providing a large available surface area for linking with other substances. To achieve distinguishable recognition, Maackia amurensis lectin (MAL) and Sambucus nigra agglutinin (SNA) were included. To facilitate the lectin fixation, PDITC, a kind of green homobifunctional cross-linker, was selected. Under optimized detection conditions, the linear range of detection for α2,3-sialylated glycans is 10 fg mL⁻¹ to 50 ng mL⁻¹ with a lower detection limit of 3 fg mL⁻¹, and the linear range of detection for α2,6-sialylated glycans is 10 fg mL⁻¹ to 50 ng mL⁻¹ with a detection limit of 3 fg mL⁻¹. This work not only provides a method for distinguishing detection of α2,3-sialylated glycans and α2,6-sialylated glycans, but also provides a reference for future clinical testing.

---

1. Introduction

Sialic acids (Sia), derivatives of N-acetylneuraminic acid, are important components of glycoproteins or glycolipids.¹ Sia are involved in cell adhesion, contact inhibition, cell transformation, metastasis, proliferation, and tumor antigens.² When a cell undergoes pathological changes, the Sia on the cell surface membrane are secreted or shed, thus elevating the blood content of Sia. Recent studies have reported significant elevation of serum Sia levels in cancer patients compared with controls.³ This suggests that the expression level of glycans could be a target or clinical biomarker for diagnosis of various cancers.⁴

α2,3-Sialylated glycans and α2,6-sialylated glycans are the predominant Sia found in nature,^5,6 making them molecular targets for cancer diagnosis. Different patterns of expression of glycans are seen in development of different diseases, for example, expression of α2,3-sialylated glycans is increased in patients with gastric cancer and prostate cancer,⁷ and high expression of α2,6-sialylated glycans has been found in liver cancer and colon cancer.⁸ However, as yet, no methods are available for discrimination between α2,3-sialylated glycans and α2,6-sialylated glycans. Therefore, a simple and sensitive method for discrimination between α2,3-sialylated glycans and α2,6-sialylated glycans is required.

The remarkable chemical diversity of sialylated glycans (α2,3 and α2,6) results in multiple enzymatic mechanisms,^1,9 which can be recognized by specific lectins. Certain Sia-binding lectins have been proven to be powerful tools for detection of Sia-specific glycoconjugates,¹⁰ for example, Limax flavus agglutinin and wheat germ agglutinin have been used to detect sialylated glycoconjugates.^10–12 Sambucus nigra agglutinin (SNA) appears to be an ideal tool for recognizing α2,6-sialylated glycans, and Maackia amurensis lectin (MAL) an ideal tool for recognizing α2,3-sialylated glycans.¹³

Accurate analysis of biomarker molecules for early detection, diagnosis, and treatment of cancer is essential.¹⁴ To this end, electrochemical biosensors based on various types of nanomaterials are fast, simple, and offer high sensitivity. In development of electrochemical biosensors, considerable efforts have been devoted to amplifying and immobilizing bio-components.¹⁵ It has been found that multiwalled carbon nanotubes (MWCNTs)-modified electrodes have better electrochemical behavior than those incorporating single-walled carbon-nanotubes (SWCNTs).¹⁶ In addition, when chemical modification occurs on the surface of MWCNTs, the outer cylinder acts as a protective sheath so that the electronic properties of the inner tube are protected.¹⁷ However, the MWCNTs have poor biocompatibility and dispersion. In this study we used carboxyl-functionalized multiwalled carbon nanotubes (p-MWCNTs) of shorter size,¹⁶ which overcome this poor biocompatibility and dispersion.

Although carboxyl functionalization improves the dispersion of MWCNTs, there are certain limits in their practical application, so polyamidoamine (PAMAM) were used to further improve the dispersion. PAMAM dendrimers are highly branched and monodispersed macromolecules with well-defined three-dimensional and globular structure.¹⁷ The number of peripheral amine groups of PAMAM (G5.0) reaches 64, being suitable for combination with p-MWCNTs, and when PAMAM is integrated with the p-MWCNTs, the stability of the biosensor is improved.

1,4-Phenylene diisothiocyanate (PDITC) is a kind of green homobifunctional cross-linker that can be used to immobilize the SNA and MAL. PDITC was chosen for use here as it is stable, flexible, and has low toxicity.¹⁸ Also, it has better conductive properties than glutaraldehyde.¹⁹ In addition, it is thought that PAMAM may enhance immobilization by PDITC.

We report, for the first time, an ultrasensitive and multi-purpose biosensor for detection of α2,3-sialylation and α2,6-sialylation. The biosensor incorporates p-MWCNTs with large surface area and possessing certain carboxyl groups, which were successfully conjugated with PAMAM. PDITC was used as a homobifunctional cross agent to link between the PAMAM and lectin. The multi-purpose biosensor exhibited excellent electrochemical response to specific detection of α2,6-sialylated and α2,3-sialylated glycans.

2. Experimental

2.1 Material and reagents

Neu5Acα(2–6)Gal β MP glycoside and Neu5Acα(2–3)Gal β MP glycoside were purchased from Tokyo Chemical Industry (Japan, http://www.TCIchemicals.com, 90.0%). Sambucus nigra agglutinin (SNA) and Maackia amurensis lectin (MAL) were purchased from Vector Labs (http://www.vectorlabs.com). 1,4-Phenylene diisothiocyanate (C₆H₄(NCS)₂), polyamidoamine PAMAM dendrimer, bovine serum albumin (BSA, 96–99%), N-hydroxysuccinimide (NHS), and N-(3-dimethylaminopropyl)-N-ethylcarbodiimide hydrochloride (EDC) were purchased from Sigma (St. Louis, USA, http://www.sigmaaldrich.com). Dopamine (DA), ascorbic acid (AA), L-cysteine, and glucose were obtained from Aladdin (China, http://www.aladdin-e.com). The other chemicals were of analytical grade, and all solutions were prepared in ultrapure water.

2.2 Apparatus and measurements

The morphology of the pretreated multiwalled carbon nanotubes (p-MWCNTs) and the pretreated multiwalled carbon nanotube–polyamidoamine dendrimer (p-MWCNTs–PAMAM) microspheres were characterized by transmission electron microscopy (TEM, Hitachi-7500) and scanning electron microscopy (SEM, Hitachi-7500518). Atomic force microscope (AFM) images were obtained with a Bruker Dimension Icon microscope (USA). Differential scanning calorimetry (DSC) was performed using the STA449 F3 Jupiter. Electrochemical experiments, electrochemical impedance spectroscopy (EIS), differential pulse voltammetry (DPV), and cyclic voltammetry (CV) were carried out using an electrochemical workstation (CHI660E) and a three-electrode system (Shanghai Chenhua Apparatus Corporation, China) composed of a glassy carbon electrode (GCE, 4 mm in diameter) as a working electrode, platinum as the counter electrode, and a saturated calomel electrode (SCE) as the reference electrode. 5 mM K₃[Fe(CN)₆]/K₄[Fe(CN)₆] (1 : 1) and 0.1 M KCl were used for all of the electrochemical measurements.

2.3 Preparation of pretreated multi-walled carbon nanotubes

Pretreated multi-walled carbon nanotubes were prepared according to methods from previous reports^20,21 with slight changes. Briefly, 10 mg of MWCNTs (in solid form) were refluxed in a H₂SO₄–HNO₃ mixture (3 : 2 by volume) for 24 h and dried under vacuum until a white solid residue was obtained. Distilled water was used for washing several times until the pH of the solution reached 7.0. The resultant was dried in an oven to obtain carboxyl group functionalized MWCNTs.

2.4 Preparation of the p-MWCNTs–PAMAM complex

The p-MWCNTs–PAMAM nanocomposite was prepared according to the literature with slight modifications.²² 2 mg of MWCNTs were added to ultrapure water under ultrasonication for 2 h. The nanocomposite was subsequently mixed with 50 μL of NHS (50 mg mL⁻¹) and 50 μL of EDC (50 mg mL⁻¹) with stirring for 30 min. Then 50 μL PAMAM was added to the resulting mixture and stirred for 5 h to obtain p-MWCNTs–PAMAM (the process is shown in Scheme 1A).

Scheme 1 Schematic representation of the electrochemical biosensor.

2.5 Fabrication of the biosensor

The biosensor was prepared following the steps outlined in Scheme 1B. Prior to use, the glassy carbon electrode (GCE) was polished carefully with Al₂O₃ powder of 0.3 and 0.05 μm to a mirror-like surface. Then, 6 μL of the as-prepared p-MWCNTs–PAMAM were dropped onto the electrode surface and dried at room temperature. Next, 6 μL of PDITC solution was added to the p-MWCNTs–PAMAM/GCE and incubated for 1 h at room temperature. After washing with distilled water, 6 μL of SNA and MAL were coated onto the electrode surface and incubated for 2 h to yield MAL/PDITC/p-MWCNTs–PAMAM/GCE. 6 μL of 1 wt% BSA were added to the modified electrode for 1 h, to block the nonspecific binding sites. Subsequently, 6 μL of a standard α2,6-sialic acid solution and standard α2,3-sialic acid solution were dropped onto the electrode surface and left standing for 2 h. After each step, the modified electrode was cleaned thoroughly with ultrapure water. Thus, a biosensor for determination of α2,3-sialylated glycans and α2,6-sialylated glycans was produced.

3. Results and discussion

3.1 Characterization of prepared material

Fig. 1 shows TEM images of the p-MWCNTs and p-MWCNTs–PAMAM. Fig. 1A reveals the tubular structure of the p-MWCNTs. In Fig. 1B and C, a lot of scattered and small black dots (PAMAM) can be seen connecting with the p-MWCNTs. The length of the p-MWCNTs is approximately 0.5 μm (Fig. 1C), which is shorter than the length of normal multi-walled carbon nanotubes (5–20 μm).²³ This suggests that the p-MWCNTs–PAMAM composite was successfully prepared. Differential scanning calorimetry (DSC) and UV-Vis spectroscopy were used to obtain more evidence of successful preparation of the composite. Fig. S2† shows that the p-MWCNTs cause no significant change in the value of DSC. The PAMAM have an exothermic peak at 290–300 °C. However, this peak is not visible after preparation of p-MWCNTs–PAMAM. Meanwhile, as can be seen in Fig. 1D, the pure p-MWCNTs solution displayed a characteristic band at 300 nm (red line),^24,25 whereas for the PAMAM, the characteristic absorption is at 280 nm (blue line). Typical absorption peaks of an aromatic system at 250–300 nm are similar to those of the polycyclic aromatic hydrocarbons.^26,27 After the p-MWCNTs connected with PAMAM, an absorption band was seen at 420 nm (black line). Taken together, all of these results indicate successful synthesis of the nanocomposites. When the p-MWCNTs solution and the p-MWCNTs–PAMAM solution were stored for 7 days, the p-MWCNTs solution created an inhomogeneous mixture and the p-MWCNTs–PAMAM solution was homogeneous (Fig. 1E and F). Clearly, the stability of the solution was improved when PAMAM was combined with p-MWCNTs.

Fig. 1 TEM images of p-MWCNTs (A) and p-MWCNTs–PAMAM (B and C). UV-Vis (D) spectroscopy of p-MWCNTs (black), PAMAM (blue), and p-MWCNTs–PAMAM (red). (E and F) Photographs of p-MWCNTs solution (E1, F1) and p-MWCNTs–PAMAM (E2, F2).

To obtain more evidence to confirm the successful immobilization of PDITC and MAL on the nanocomposite film surface, atomic force microscopy (AFM) was used to investigate the bioconjugation process. Fig. 2 shows three-dimensional images of p-MWCNTs–PAMAM (A), PDITC/p-MWCNTs–PAMAM (B), and MAL/PDITC/p-MWCNTs–PAMAM (C). The greatest peak height of the p-MWCNTs–PAMAM was 21.92 nm (Fig. 2A). The average roughness (R_a) was 3.71 nm. After immobilization of PDITC on the p-MWCNTs–PAMAM, the peak height was approximately 16.40 nm and the R_a was 2.73 nm, resulting from the PDITC aggregates filling the gaps on the nanocomposite (Fig. 2B). When the MAL linked with PDITC, the greatest peak height was 11.40 nm and R_a was 1.62 nm. The surface roughness of MAL/PDITC/p-MWCNTs–PAMAM is further smoothed compared with PDITC/p-MWCNTs–PAMAM (Fig. 2C), because the MAL covers the gaps in the electrode surface.²⁸ All of these results suggest successful fabrication of the biosensor. The roughness of the SNA/PDITC/p-MWCNTs–PAMAM biosensor was similar to that of the MAL/PDITC/p-MWCNTs–PAMAM biosensor.

Fig. 2 AFM images of (a) three-dimensional and (b) cross-sectional graphs of (A) p-MWCNTs–PAMAM, (B) PDITC/p-MWCNTs–PAMAM, and (C) MAL/PDITC/p-MWCNTs–PAMAM.

3.2 Electrochemical characterization of the stepwise-modified electrode

Electrochemical impedance spectroscopy (EIS) is an effective method for studying electron-transfer resistance at the electrode surface and the interface properties of surface-modified electrodes.²⁹ The impedance spectra include a semicircular portion and a linear portion. The linear portion at lower frequencies corresponds to the diffusion-limited process. The semicircular portion at higher frequencies represents the electron transfer limited process. The diameter of the semicircle equals the electron transfer resistance (R_et).²⁹ As shown in Fig. 3B, the bare GCE presents a small semicircle at high frequency (curve a) of R_et value of 307.6 Ω. After the electrode was modified with p-MWCNTs–PAMAM, the resistance was much less than that of the bare GCE (curve b, R_et = 182.2 Ω), suggesting that the p-MWCNTs–PAMAM favor interfacial and electron transfer. The resistance increased after incubation with PDITC (curve c, 372.0 Ω), which indicates that the PDITC was linked with p-MWCNTs–PAMAM. The resistance continued to increase after incubation with MAL (curve d, 507.7 Ω), indicating that MAL was immobilized on the modified electrode successfully, because of the MAL blocking electron transfer.

Fig. 3 Typical CV (A) and (B) studies of K₃[Fe(CN)₆]/K₄[Fe(CN)₆] (5 mM) for (a) a bare electrode, (b) p-MWCNTs–PAMAM/GCE, (c) PDITC/p-MWCNTs–PAMAM/GCE, (d) MAL/PDITC/p-MWCNTs–PAMAM/GCE, (e) blocking with 1% BSA, and (f) specific recognition with 5 ng mL⁻¹ Neu5Acα(2–6)Gal β MP glycoside. (C) Cyclic voltammograms of the electrochemical biosensor at different scan rates ranging from 20 to 250 mV s⁻¹ in [Fe(CN)₆]^3−/4− (1 : 1) and 0.1 M KCl. (D) The linear relation of current versus the square root of the scan rates. (E) DPV responses of the proposed biosensor after incubation with different concentrations of Neu5Acα(2–6)Gal β MP glycoside. (F) Calibration curve of the biosensor toward different concentration of Neu5Acα(2–6)Gal β MP glycoside (n = 5).

After BSA blocked the electrode, the peak currents continued to decrease because of electron hindrance of BSA, with the R_et enlarged (curve e, 909.9 Ω). Subsequently, the R_et of curve f increased (1032.0 Ω) because the Neu5Acα(2–3)Gal β MP glycoside was captured by MAL. The inset in Fig. 3B (top right corner) is the equivalent circuit used to fit the impedance spectra. This circuit includes the constant phase element CPE related to the double layer capacitance, C_dl, the electrolyte solution, R_S, the Warburg impedance, Z_W, which causes diffusion of the redox probe ions to the electrode interface from the bulk of the electrolyte, and the resistance of the electron transfer, R_et. The data obtained after fitting the resistance spectra are recorded in Table 1. The data of the CV are in agreement with the result from EIS, suggesting successful fabrication of the biosensor.

Table 1 Simulation parameters of the equivalent circuit components

Electrode	RS (Ω cm^2)	Ret (Ω cm^2)	Cdl (μF cm^2)	n	10^3 × ZW (Ω cm^2)
GCE	12.22	21.36	72.02	0.84	0.19
p-MWCNTs–PAMAM/GCE	5.08	10.71	23.40	0.84	0.18
PDITC/p-MWCNTs–PAMAM/GCE	5.95	26.81	57.32	0.86	0.64
MAL/PDITC/p-MWCNTs–PAMAM/GCE	4.86	33.80	10.82	0.89	0.39
Blocking with 1% BSA	5.98	75.53	27.12	0.86	1.50
50 ng mL^−1 Neu5Acα(2–3)Gal β MP glycoside	6.69	84.91	34.90	0.84	0.51

---

The influence of the potential scan rate was investigated on the peak current response about the PDITC/p-MWCNTs–PAMAM electrode. The CV of PDITC/p-MWCNTs–PAMAM in 5 mM of K₃[Fe(CN)₆]/K₄[Fe(CN)₆] at different scan rates are shown in Fig. 3C. The redox peaks current increased gradually with increasing scan rate in the range of 20–250 mV s⁻¹. Fig. 3D shows a good linear relationship between the peak current and the square root of scan rate. This result indicates that the redox reaction on the electrode surface is a diffusion-controlled process, which is consistent with previous reports.³⁰ According to the Randles–Sevcik equation^31,32 I = 2.69 × 10⁵ × A × D^1/2n^3/2v^1/2C, in which n is the number of electrons transferred in the redox reaction (n = 1), A is the electrode area, C is the concentration of the reactant (at 25 °C, D = 6.70 × 10⁻⁶ cm² s⁻¹), (5 mM K₃[Fe(CN)₆]/K₄[Fe(CN)₆]), D is the diffusion coefficient, I refers to the redox peak current, and v is the scan rate of the CV measurement, the surface area of the PDITC/p-MWCNTs–PAMAM modified electrode was calculated to be 30.84 mm².

3.3 Optimization of experimental conditions

Electrochemical performance of biosensors can be influenced by experimental conditions. Therefore, the effects of the volume of p-MWCNTs–PAMAM hybrid nanocomposite, the incubation time of PDITC, the concentration of MAL and SNA, the incubation time of MAL, the incubation time of SNA, and the length of recognition time were investigated.

As shown in Fig. 4A, the current response increased rapidly with increasing PAMAM volumes from 20 μL to 50 μL. At PAMAM volumes exceeding 50 μL, a decrease was observed in the current change, possibly because the excess PAMAM failed to conjugate with the MWCNTs. Thus, 50 μL of PAMAM was used as the optimal condition.

Fig. 4 Effects of (A) volume of PAMAM, (B) volume of p-MWCNTs–PAMAM, (C) incubation time of PDITC, (D) incubation time of SNA, (E) concentration of SNA, and (F) recognition time of α2,6-sialylated glycans.

The volume of p-MWCNTs–PAMAM was tested in the range 2–10 μL to provide the best platform for incubation with PDITC. The current change increased up to a volume of 6 μL, then remained stable (Fig. 4B). Therefore, 6 μL was determined to be the optimal volume of p-MWCNTs–PAMAM.

The incubation time of PDITC is a critical factor in preparation of the biosensor. PDITC plays a key role in providing a functional group for MAL immobilization. As shown in Fig. 4C, the current change increased sharply from 30 to 60 min, reaching a maximum at 60 min. Any further increase in incubation time had no significant change on current, therefore, 60 min was selected as the optimal incubation time.

To optimize the reaction time, the assembly time of analysis on SNA was investigated (Fig. 4D). A period of time is required for the lectin to bind the sialylated glycans and form a complex. Gradually increasing the reaction time results in a decrease in signal, which implies that SNA were immobilized on binding sites and hindered electron transfer. In our work, the SNA observed response occurred from 30 to 180 min, with the current change initially increasing then stabilizing, the highest peak being at 150 min. Thus, 150 min was selected as the optimal incubation time of SNA for this work.

The concentration of SNA is also an important experimental factor. When the SNA fixed on the electrode surface is maximized, the biosensor has high sensitivity. A concentration range was investigated from 0.5 mg mL⁻¹ to 2.5 mg mL⁻¹ (Fig. 4E). Changes of current were observed until 2.0 mg mL⁻¹, then it stabilized. Therefore, 2.0 mg mL⁻¹ was selected as the optimal concentration for the experiment.

The incubation time of Neu5Acα(2–6)Gal β MP glycoside is an important parameter in which glycoside is recognized by SNA and a complex is formed. To explore the effect of the incubation time on amperometric response, incubation times of the fabricated biosensor were chosen from 60 to 180 min (Fig. 4F). With increasing time, more glycoside was captured by lectin and the current change was increased. With the incubation time from 60 min to 120 min, the current change gradually increased, then stabilized as the amount of Neu5Acα(2–6)Gal β MP glycoside captured on the sensor surface reached a maximum value. Therefore, the optimal incubation time of SNA was 150 min.

The MAL experimental conditions found to give best results are as follows (Fig. S3A–C†): (d) MAL incubation time of 2 h; (e) MAL concentration of 2.0 mg mL⁻¹; (f) recognition time of 2 h of the α2,3-sialylated glycans.

3.4 Performance of the biosensor

Calibration plots for detection of Neu5Acα(2–3)Gal β MP glycoside and Neu5Acα(2–6)Gal β MP glycoside were determined under the optimal experimental conditions. As shown in Fig. 3E and F, the prepared electrochemical biosensor was used to detect Neu5Acα(2–3)Gal β MP glycoside and Neu5Acα(2–6)Gal β MP glycoside by DPV. With increasing concentration of glycoside, formation of glycan–lectin complexes increased and the spread of the redox probe was gradually blocked. This indicates that the change in current response is related to the amount of glycoside captured at the electrode surface. The peak current response decreased as the concentrations of Neu5Acα(2–6)Gal β MP glycoside increased. The calibration plot has a good linear relationship in the range of 10 fg mL⁻¹ to 50 ng mL⁻¹ with a detection limit of 3 fg mL⁻¹ (S/N = 3). As shown in Fig. S2,† for accuracy of the detection, the equation of the calibration plot was separated into two parts. For values below 10 pg mL⁻¹, the equation used was Δcurrent (μA) = 65.71 + 18.53 log C, R² = 0.997. Over 10 pg mL⁻¹, the equation used was Δcurrent (μA) = 79.72 + 3.34 log C, R² = 0.986. As shown in Fig. 3F, below values of 10 pg mL⁻¹, the equation used was Δcurrent (μA) = 59.00 + 20.95 log C, R² = 0.988. Over 10 pg mL⁻¹, the equation used was Δcurrent (μA) = 79.66 + 3.39 log C, R² = 0.980. Therefore, a novel and specific biosensor was designed to sensitively detect the Neu5Acα(2–3)Gal β MP glycoside and Neu5Acα(2–6)Gal β MP glycoside within the analytical concentration range.

3.5 Selectivity, stability, and reproducibility of the electrochemical biosensor

To evaluate the selectivity of the electrochemical biosensor, we explored the effect of other molecules on the system including Neu5Acα(2–3)Gal β MP glycoside, ascorbic acid (AA), L-cysteine (L-cys), glucose, uric acid, and dopamine (DA). These studies confirmed that the change of current resulted from specific recognition between glycoside and lectin (Fig. 5A). The current change caused by other molecules was less than 10%, suggesting that the fabricated biosensor exhibited good selectivity and that the biosensor can distinguish detection of α2,3-sialylated glycans and α2,6-sialylated glycans.

Fig. 5 (A) α2,6-Sialylated glycans (50 ng mL⁻¹), α2,3-sialylated glycans (500 ng mL⁻¹), glucose (500 ng mL⁻¹), ascorbic acid (500 ng mL⁻¹), uric acid (500 ng mL⁻¹), L-cysteine (500 ng mL⁻¹), and dopamine (500 ng mL⁻¹). The stability of the electrochemical biosensor after 0 days (a), 7 days (b), 14 days (c), and 21 days (d) (B). The reproducibility of five different electrodes modified with 50 ng mL⁻¹ α2,6-sialylated glycans (C).

The MAL/PDITC/p-MWCNTs–PAMAM/GCE modified electrode was stored at 4 °C. The observed current change was retained at 95.21%, 91.92%, and 89.42% compared with the original current after 7, 14, and 21 days, respectively (Fig. 5B). Possible reasons for this observation are that p-MWCNTs–PAMAM has good biocompatibility and stability, and MAL has strong specific recognition ability.

The reproducibility of the electrochemical biosensor was evaluated. The intra-assay precision of the biosensor was estimated by analyzing five prepared electrodes for detection of Neu5Acα(2–6)Gal β. The relative standard deviation (RSD) of the measurements for the five electrodes was 3.31% at a Neu5Acα(2–6)Gal β MP glycoside concentration of 50 ng mL⁻¹, indicating acceptable precision and good reproducibility (Fig. 5C).

3.6 Application analysis in human serum samples

To examine the applicability of the present electrochemical biosensor on real samples, determination of Neu5Acα(2–6)Gal β MP glycoside in human serum was performed using DPV analysis. Diluted human serum samples spiked with 1 ng mL⁻¹, 5 ng mL⁻¹, and 40 ng mL⁻¹ Neu5Acα(2–6)Gal β MP glycoside, were applied to the electrochemical biosensor (Table 2). The application analysis of the Neu5Acα(2–3)Gal β MP glycoside is detailed in the ESI (Table S1†).

Table 2 Recovery of serum samples for the electrochemical biosensor using amperometric i–t curve analysis (n = 3)

Samples	Canalyte	Added	Found	Recovery	RSD (%)
Sample-1	NT	0.5 pg mL^−1	0.51 pg mL^−1	101.72	4.29
Sample-2	NT	5 ng mL^−1	5.21 ng mL^−1	104.25	3.27
Sample-3	NT	40 ng mL^−1	39.32 ng mL^−1	98.32	1.64

---

4. Conclusions

We report an ultrasensitive and specific electrochemical biosensor based on PDITC cross-linked with p-MWCNTs–PAMAM, able to distinguish detection of α2,3-sialylated glycans and α2,6-sialylated glycans in human serum and detection of α2,3-sialylated glycans for the first time. The developed electrochemical biosensor has high specificity and a broad linear range with a low detection limit, indicating resilience to endogenous interferences in human serum. The biosensor also has good reproducibility and high sensitivity, with potential as a rapid analytical tool for diagnostic and clinical research. The p-MWCNTs–PAMAM biosensor could have wide applications in routine detection of other proteins or glycans, which will require the use of other affinity binding pairs.

Acknowledgements

This study was supported financially by the National Natural Science Foundation of China (81370403, 21205146).

References

1. S. Kim, D. B. Oh, H. A. Kang and O. Kwon, Appl. Microbiol. Biotechnol., 2011, 91, 1–15.
2. A. Varki, Nature, 2007, 446, 1023–1029.
3. R. J. Winzler, Methods Biochem. Anal., 1955, 2, 279–311.
4. Y. K. Matsuno, T. Saito, M. Gotoh, H. Narimatsu and A. Kameyama, Anal. Chem., 2009, 81, 3816–3823.
5. D. B. Werz, R. Ranzinger, S. Herget, A. Adibekian, C. W. von der Lieth and P. H. Seeberger, ACS Chem. Biol., 2007, 2, 685–691.
6. R. J. Fair, H. S. Hahm and P. H. Seeberger, Chem. Commun., 2015, 51, 6183–6185.
7. C. Ohyama, M. Hosono, K. Nitta, M. Oh-eda, K. Yoshikawa, T. Habuchi, Y. Arai and M. Fukuda, Glycobiology, 2004, 14, 671–679.
8. T. Sata, J. Roth, C. Zuber, B. Stamm and P. U. Heitz, Am. J. Pathol., 1991, 139, 1435–1448.
9. T. J. Boltje, T. Buskas and G. J. Boons, Nat. Chem., 2009, 1, 611–622.
10. L. Gao, J. He, W. Xu, J. Zhang, J. Hui, Y. Guo, W. Li and C. Yu, Biosens. Bioelectron., 2014, 62, 79–83.
11. J. J. Park, J. Y. Yi, Y. B. Jin, Y. J. Lee, J. S. Lee, Y. S. Lee, Y. G. Ko and M. Lee, Biochem. Pharmacol., 2012, 83, 849–857.
12. N. Shibuya, I. J. Goldstein, W. F. Broekaert, M. Nsimba-Lubaki, B. Peeters and W. J. Peumans, Arch. Biochem. Biophys., 1987, 254, 1–8.
13. C. Geisler and D. L. Jarvis, Glycobiology, 2011, 21, 988–993.
14. X. Liu, Q. Dai, L. Austin, J. Coutts, G. Knowles, J. Zou, H. Chen and Q. Huo, J. Am. Chem. Soc., 2008, 130, 2780–2782.
15. G. Yuan, C. Yu, C. Xia, L. Gao, W. Xu, W. Li and J. He, Biosens. Bioelectron., 2015, 72, 237–246.
16. Y. Liang, H. Wang, P. Diao, W. Chang, G. Hong, Y. Li, M. Gong, L. Xie, J. Zhou, J. Wang, T. Z. Regier, F. Wei and H. Dai, J. Am. Chem. Soc., 2012, 134, 15849–15857.
17. D. M. Kim, M. A. Rahman, M. H. Do, C. Ban and Y. B. Shim, Biosens. Bioelectron., 2010, 25, 1781–1788.
18. R. Elshafey, C. Tlili, A. Abulrob, A. C. Tavares and M. Zourob, Biosens. Bioelectron., 2013, 39, 220–225.
19. Y. Zhao, J. He, G. Yuan, C. Xia, Y. Li and C. Yu, RSC Adv., 2015, 5, 88392–88400.
20. C. Zhu, S. Guo and S. Dong, Adv. Mater., 2012, 24, 2326–2331.
21. W. Xu, J. He, L. Gao, J. Zhang and C. Yu, Microchim. Acta, 2015, 182, 2115–2122.
22. J. Zhang, J. He, W. Xu, L. Gao, Y. Guo, W. Li and C. Yu, Electrochim. Acta, 2015, 156, 45–52.
23. W. Qiu, H. Xu, S. Takalkar, A. S. Gurung, B. Liu, Y. Zheng, Z. Guo, M. Baloda, K. Baryeh and G. Liu, Biosens. Bioelectron., 2015, 64, 367–372.
24. X. Zhang, W. Lu, J. Shen, Y. Jiang, E. Han, X. Dong and J. Huang, Biosens. Bioelectron., 2015, 74, 291–298.
25. Y. Yao, X. Bai and K.-K. Shiu, Nanomaterials, 2012, 2, 428–444.
26. Q. Gao, J. Han and Z. Ma, Biosens. Bioelectron., 2013, 49, 323–328.
27. S. Y. Xie, R. B. Huang and L. S. Zheng, J. Chromatogr. A, 1999, 864, 173–177.
28. Y. Li, L. Fang, P. Cheng, J. Deng, L. Jiang, H. Huang and J. Zheng, Biosens. Bioelectron., 2013, 49, 485–491.
29. Y. Li, J. He, Y. Niu and C. Yu, Biosens. Bioelectron., 2015, 74, 953–959.
30. X. Lu, Y. Li, X. Zhang, J. Du, X. Zhou, Z. Xue and X. Liu, Anal. Chim. Acta, 2012, 711, 40–45.
31. C. I. Muller and C. Lambert, Langmuir, 2011, 27, 5029–5039.
32. S. Xie, J. Zhang, Y. Yuan, Y. Chai and R. Yuan, Chem. Commun., 2015, 51, 3387–3390.

---

Footnotes

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra03570a

‡ Junlin He, Yuliang Li and Yazhen Niu contributed equally to this work.

---