Application of the Reduced Graphene Oxide–Multiwalled Carbon Nanotubes Composite for Development of the Electrochemical Aptasensor for Oxytetracycline Detection ^†

Abstract

Excessive use of oxytetracycline (OTC) in veterinary medicine has increased the presence of antibiotics in food, which accelerates the development of antimicrobial resistance. We report the development of a highly sensitive electrochemical aptasensor for OTC detection, based on a glassy carbon electrode (GCE) modified with reduced graphene oxide (rGO) and multiwalled carbon nanotubes (MWCNTs) nanocomposite. DNA aptamers specific to OTC were covalently attached to the nanocomposite surface via carbodiimide chemistry. Differential pulse voltammetry (DPV) showed a decrease in peak current due to the binding of OTC to the aptamers. The sensor exhibited a limit of detection (LOD) of 1.72 ng/mL, which is below the maximum residue limit (MRL) for OTC (100 ng/mL) established by European Union. The sensor has been tested on a spiked milk sample.

1. Introduction

Excessive application of oxytetracycline (OTC), a broad-spectrum antibiotic of the tetracycline group in veterinary medicine, has increased the background contamination of food products and may be a leading cause behind decreased antibiotic efficiency. For instance, the contamination of milk and milk products with OTC leads to increased risks of the development of the antimicrobial resistance (AMR) of bacteria, which are harmful for consumers. To mitigate potential issues, the European Union established the maximum residue limit (MRL), restricting OTC in milk, as 100 µg/kg (~100 ng/mL) [1]. Antibiotics decompose slowly in the environment and tend to accumulate rapidly in organisms of the ecosystem, posing direct or indirect risks to human health, including a risk of the growth of AMR [2]. To control food quality, the development of cost-efficient, high-sensitivity methods for OTC detection in food products, including milk, which should be integrated into the food production/processing chain, is urgently needed.

High-performance liquid chromatography (HPLC) can be used to detect the presence of multiple antibiotics simultaneously. However, it is time-consuming, and its prohibitively high cost makes it difficult for its implementation as a standard analytical tool at every milk production facility [3]. Similar problems make gas chromatography–mass spectrometry (GC-MS) [4] and liquid chromatography–mass spectroscopy (LC-MS) unsuitable for routine or on-site antibiotic testing, despite their excellent sensitivity [5]. Enzyme-linked immunosorbent assay (ELISA) is commonly used for rapid screening of antibiotics in milk and meat, but their cross-reactivity and limited specificity often result in requiring additional chromatographic tests [6].

As an alternative to the conventional high-cost methods, a new line of low-cost biosensors is emerging. They build on the surface of a transducer that converts the chemical signal (affinity interactions) into electrical, optical, or acoustic (mass sensitive) signals, and an analytical instrument (potentiostats, spectrometers, or network analyzers). The sensing surface of such a sensor is modified with immobilized receptors (antibodies or nucleic acid aptamers), which ensure selective interaction with antibiotics [6]. The advantage of aptamers over antibodies is the possibility of their in vitro structural/compositional optimization using combinatorial chemistry approaches, including SELEX (Systematic evolution of ligands by exponential enrichment) [7]. In this method, the library (around 10¹³–10¹⁵ molecules) of the random sequences of the synthetic single-stranded DNA (ssDNA) (typically 30–60 nucleotides) is allowed to interact with the target of interest. ssDNA sequences bind to the target with various affinities. The complexes of the ssDNA-target are then dissociated. The ssDNA sequences are multiplied via polymerase chain reaction (PCR) and used in the next cycle of selection. After several cycles (around ten), the sequence with the highest affinity to the target is obtained. The SELEX was used for the development of DNA aptamers that specifically bind OTC. The SELEX protocol was further modified to include magnetic beads (Flu-Mag-SELEX), Toggle-SELEX, graphene oxide-based SELEX, and others [8]. For example, Niazi et al. [9] selected the aptamers for a tetracycline group of antibiotics using a random sequence of 10¹⁵ ssDNA, using Toggle-SELEX combined with Flu-Mag-SELEX. The magnetic beads coated with antibiotics allowed the development of DNA aptamers with high affinity to tetracycline or OTC after four rounds with the dissociation constant, K_D = 9.6 nM (4.41 ng/mL). Aptamers are more stable than antibodies and can be synthesized with high purity and yield. Aptamers can also be chemically modified to increase their stability or functionality for surface immobilization. The increased interest in DNA-aptamer-based biosensors (aptasensors) for antibiotic detection was discussed in several comprehensive reviews [6,10,11]. The sensitivity of these sensors is usually better than the MRL. Current sensor development trends indicate growing interest in low-cost electrochemical methods which integrate nanocomposite materials. This is related to the improved electrical properties, increased surface area, presence of COOH functional groups, and possibility to increase surface concentration of aptamers [12].

Here, we report the development of an electrochemical aptasensor based on carbon nanocomposites for the sensitive detection of OTC in milk. By utilizing a glassy carbon electrode (GCE) modified with reduced graphene oxide–multiwalled carbon nanotubes (rGO-MWCNTs) composites, we achieved enhanced surface area and improved electron transfer properties. The aptasensor was evaluated using differential pulse voltammetry (DPV), providing a sensitive and straightforward approach for OTC detection.

2. Materials and Methods

2.1. Chemicals

The DNA aptamers, (APT) specific to oxytetracycline (OTC) modified at 5′ end by amino group (5′-amino-(CH₂)₆-TTT TTT TTT GGG GGC ACA CAT GTA GGT GCT GTC CAG GTG TGG TTG TGG T-3′), were obtained from GeneCust (Boynes, France). The aptamer reported by Niazi et al. [9] was further modified to include an additional dT₉ sequence to provide better conformational flexibility of the aptamers. The aptamers have been dissolved in TE buffer (10 mM Tris-HCl, 1 mM EDTA, pH 8) in a concentration of 0.1 mM. Graphene oxide (GO), multiwalled carbon nanotubes (MWCNTs), OTC, tetracycline (TET), kanamycin (KAN), penicillin G (PEN), chloramphenicol (CAP), 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide (EDC), N-hydroxysuccinimide (NHS), ethanolamine, Tris–HCl, ethylenediaminetetraacetic acid (EDTA), Tween 20, K₃[Fe(CN)₆], and K₄[Fe(CN)₆] have been purchased from Sigma-Aldrich (Darmstadt, Germany). NaCl, KCl, HCl, NaOH, MgCl_2, and CaCl₂ were from Slavus (Bratislava, Slovakia). The experiments were performed in Tris-HCl (20 mM, pH 7.6, contained 100 mM NaCl, 2 mM MgCl₂, 5 mM KCl, and 1 mM CaCl₂, 0.02% Tween 20). Deionized (DI) water with a resistance of 18 MΩ·cm, prepared by Purelab Classic UV (Elga, High Wycombe, UK), was used for the preparation of all aqueous solutions. The pH of the solution was adjusted by 0.1 M HCl or 0.1 M NaOH using the pH meter LE438 Mettler-Toledo (Greifensee, Switzerland).

2.2. Preparation of the Biosensor and Electrochemical Measurements

The 2 mm diameter glassy carbon electrodes (GCE) (CH Instruments, Austin, TX, USA) were used for biosensor preparation. The bare GCE was polished with 1.0 μm and 0.3 μm alumina slurry (CH Instruments) to obtain a mirror surface, then ultrasonicated for 3 min in DI water and dried under nitrogen gas. The GCE was electrochemically cleaned in 0.5 M H₂SO₄ aqueous solution by using cyclic voltammetry (CV) in the voltage range between 0.2 and 1.5 V. The Ag/AgCl reference electrode underwent a scan rate of 100 mV/s until the characteristic cyclic voltammogram was obtained. The electrodeposition was performed from 1 mg/mL GO-MWCNTs at 1.6 V for 30 min, forming a thin film on top of the GCE. Electrochemical reduction in GO was performed by CV, in a potential range between −0.2 to 1.0 V, with a scan rate of 100 mV/s, in 0.1 M KH₂PO₄ solution. After 60 cycles, the GO was reduced and rGO-MWCNTs on the working electrode were obtained. After washing the rGO-MWCNTs with DI, the COOH-functionalized MWCNTs were activated with the incubation of 100 µL of 20 mM EDC/50 mM NHS for 120 min. Then, 100 μL of 5 μM amino-modified OTC aptamers were added on the surface of the nanofabricated GCE. In addition, 100 μL of 0.1 M ethanolamine (ETA) was added at the sensing surface, blocking the free COO sites. Then, OTC was tested at different concentrations of 0.3, 0.5, 1, 5, 10, and 30 ng/mL. Figure 1 shows the scheme of sensor preparation. It has been adopted from paper by Jia et al. [13] with a similar design but in which detection of Salmonella was reported.

The biosensor was tested in milk spiked by OTC. For this purpose, the 1.5% fat milk from the local supermarket was used. The milk was thermostated at 40 °C for 30 min, diluted with ethanol in a 1:3 v/v ratio and centrifuged at 5000 rpm for 5 min. The supernatant, which is mainly ethanol and water mixture, were separated from the sediment, diluted 10-fold with the OTC binding buffer, and spiked by various concentrations of OTC [14].

Differential pulse voltammetry (DPV) was performed from −0.2 to 0.6 V in a 5 mM [Fe(CN)₆]^3−/4− solution at a scan rate of 100 mV/s. All electrochemical measurements were performed using CHI 410 potentiostat (CH Instruments) or µAutolab III/Fra2 (Metrohm, Herisau, Switzerland) in the three-electrode configuration consisting of the GCE working electrode, Ag/AgCl reference, and the platinum counter electrode (CH Instruments). All the experiments were performed at room temperature (~22 °C).

3. Results and Discussion

In the first series of experiments, we studied the DPV of the modified GCE in the presence of OTC. Figure 2a shows DPV of the bare, GO/MWCNTs and rGO/MWCNTs modified GCE in the presence of 5 mM [Fe(CN)₆]^3−/4− redox probe. A typical current peak at approx. 300 mV vs. Ag/AgCl reference electrode for bare GCE corresponds to reduction in [Fe(CN)₆]⁴⁻. The value of current decreased after electrodeposition of the GO-MWCNTs on the surface of GCE, due to blocking of the electron transfer to the electrode surface from the redox probe caused by the negatively charged GO in GO/MWCNTs composite. The reduction in GO (rGO/MWCNTs) increased the current due to elimination of the negatively charged groups of GO. The immobilization of the aptamers, and covering the bare COO⁻ sites by ethanolamine (ETA), reduced the amplitude of the current by several µA (Figure 2b) due to increase in the negative change in the electrode surface caused by negatively charged aptamers. Then, the aptasensor was incubated for 30 min in a series of increasing concentrations, from 0.3 to 30 ng/mL OTC, followed by gentle rinsing with a binding buffer. Blocking the charge transfer through the sensing surface by OTC leads to a decrease in current amplitude and its slight shift toward higher potential. The isoelectric point of OTC is 5.2 [15]. Therefore, at pH 7.6, the OTC is negatively charged. This also contributes to the decrease in the current amplitude and modify the electrode potential. Thus, a more positive voltage is required for reduction in [Fe(CN)₆]⁴⁻.

Based on the results presented in Figure 2b, we constructed the plot of the normalized changes in the current ΔI/I₀ vs. concentration of OTC. Here, ΔI = I − I₀, where I is the current amplitude at a certain concentration of OTC and I₀ is the current for the bare aptasensor without OTC. As can be seen from Figure 3a, the value ΔI/I₀ decreases with increasing OTC concentration. We find that the Langmuir model does not describe the observed changes, whereas the Freundlich adsorption isotherm, ΔI/I₀ = K_FC^1/n, where K_F and n are the constants and C is the concentration of OTC, describes the observed adsorption well. It suggests the heterogeneous nature of the surfaces, and it can be interpreted that the binding of OTC to the aptamers is combined with physical adsorption [16]. We estimated the value of the Freundlich adsorption constant K_F to be −(0.25 ± 0.01) (mL/ng)^1/n and n = 5.18 ± 0.47, R² = 0.97.

A linear dependence of the relative changes in the current ΔI/I₀ on the Log[OTC] (note logarithmic scale of concentration axes) can be seen on a calibration plot ΔI/I₀ vs. Log[OTC] (Figure 3b). Using this plot, we estimated the limit of detection (LOD) using a formula LOD = 10^a, where a = 3.3σ/b, σ is the standard deviation of the intercept, and b is the slope of the linear regression line [17,18]. The LOD = 1.72 ng/mL is about 58 times lower than the MRL established for OTC by the EU (100 ng/mL) [1]. This is also better than LOD reported for the electrochemical aptasensor based on a gold screen-printed electrode (LOD = 14 ng/mL) [19] and MWCNTs, gold nanoparticles (AuNPs), rGO, and a chitosan sensor (LOD = 4.2 ng/mL) [2]. Better sensitivity (LOD = 0.23 ng/mL) was achieved using GCE grafted by diazonium salts with covalently attached aptamers [20]. The high sensitivity of impedimetric aptasensor using covalent organic framework nanospheres, with a LOD of 7.4 fg/mL, has been reported by Yuan et al. [21]. However, the sensor response saturated at 10 pg/mL OTC, which limits its practical use. A comprehensive review of electrochemical aptasensors for OTC detection is available elsewhere [6,10]. The review of most recent publications (2023–2025) on electrochemical aptasensors for OTC detection is presented in

Table 1. The comparison of the properties of the electrochemical aptasensors for detection oxytetracycline published between 2023–2025.

Sensing Surface	Detection Method	Dynamic Range ng/mL	LOD, ng/mL	Sample Type/ Recovery %	Reference
SPE Au electrode	DPV	1–500	14 (PBS)/10 (Milk)	Milk/124	[19]
Au/COF nanospheres	EIS	10−5–10−2	7.4 × 10−6	Milk/95.7–107	[21]
Au/HOF	CV, EIS	(0.01–0.9) × 10−3	9.4 × 10−4	-	[22]
FTO/Zn-defective CdS/ZnS heterostructure	PEC EC	0.01–50	1.86 × 10−3 3.08 × 10−3	-	[23]
Au/Kapton film	DPV	1.5–3	5	Milk/95	[24]
GCE/MOF(Fe)@PDA/SPE	DPV/ CA	4.6–4600	3.17/2.56	Tap water/98.2/99.4	[25]
SPCE/AuNPs/ErGO/Cu-MOF	CV	0.1–105	0.03	Milk/87–110	[26]
Au/Kapton film	DPV	1.4–2.7	7	Milk/107–110	[27]
GCE/rGO/MWCNTs	DPV	1–30	1.72	Milk/109	This work

AuNPs: gold nanoparticles; CA: chronoamperometry; COF: covalent organic framework; CV: cyclic voltammetry; EC: electrochemical; EIS: electrochemical impedance spectroscopy; ErGO: electrochemically reduced graphene oxide; FTO: fluorine doped tin oxide; GCE: glassy carbon electrode; HOF: hydrogen organic frameworks; MOF: metal–organic frameworks; PDA: polydopamine; PEC: photoelectrochemical; SPCE: screen-printed carbon electrode; and SPE: screen-printed electrode.

.

As can be seen from Table 1, the highest sensitivity for OTC detection was obtained from the nanocomposites based on covalent organic frameworks (COF) [21]. However, even for sensing surfaces based on gold layers covered by Kapton film [24,27], the LOD was below the MRL. The LOD obtained in our work is better than those based on MOF(Fe); however, the dynamic range should be improved. Still, the sensitivity of all sensors presented is sufficient for practical applications, even for diluted food samples. Most of the sensors have been validated in milk with sufficient recovery. Typically, 10x diluted food samples were analyzed. Further effort should, therefore, be focused on the design of the sensing surface for the improvement of its antifouling properties [17,28]. This allows the application of the sensor directly in raw milk.

We also analyzed the selectivity of the sensor response in presence of 10 ng/mL of other antibiotics. Figure 4 shows the sensor response for OTC, chloramphenicol (CAP), kanamycin (KAN), penicillin G (PEN), and tetracycline (TET). While a higher response for OTC is evident, this is not OTC-specific, suggesting some nonspecific binding of CAP, KAN, PEN, and TET to the aptamers.

We performed experiments on the OTC charge transfer detection in spiked milk samples. Figure 5 shows the sensor response in PBS and in milk samples spiked by 1, 10, and 30 ng/mL OTC. The changes in the current amplitude measured are higher for OTC spiked milk compared with those in PBS. This can be explained as the influence of other milk components (caseins or lipids) on the sensor response [19]. The best recovery of 109% and the lowest relative standard deviation (RSD = 10.5%) was observed for 30 ng/mL, the highest OTC concentration tested. This means that the accuracy of measurements is improved at a higher OTC concentration, which is associated with the higher current amplitude changes. Considering rather high sensitivity and relatively large MRL for OTC detection, the sensor can be used in substantially diluted milk samples, which also helps to reduce the effect of the milk components.

4. Conclusions

We developed a highly sensitive electrochemical aptamer-based biosensor for detection oxytetracycline using GCE with electrodeposited and reduced graphene oxide–multiwalled carbon nanotubes (rGO/MWCNTs) composite. The sensor demonstrated high sensitivity in OTC detection with the LOD of 1.72 ng/mL. The binding of OTC to the sensor surface can be characterized by the Freundlich adsorption isotherm. This suggests the heterogeneous nature of the surfaces, and it can be interpreted that the binding of OTC to the aptamers is combined with physical adsorption. The sensor was tested in spiked milk sample with 109% recovery. We speculate that the observed effect can arise from the nonspecific adsorption of milk compounds, such as proteins and lipids, to the sensing surface. Further experiments are needed to improve the antifouling properties of the sensor as well as to increase the sensor’s dynamic range.