Rapid detection of cocaine using aptamer-based biosensor on an evanescent wave fibre platform

The rapid detection of cocaine has received considerable attention because of the instantaneous and adverse effects of cocaine overdose on human health. Aptamer-based biosensors for cocaine detection have been well established for research and application. However, reducing the analytic duration without deteriorating the sensitivity still remains as a challenge. Here, we proposed an aptamer-based evanescent wave fibre (EWF) biosensor to rapidly detect cocaine in a wide working range. At first, the aptamers were conjugated to complementary DNA with fluorescence tag and such conjugants were then immobilized on magnetic beads. After cocaine was introduced to compete against the aptamer-DNA conjugants, the released DNA in supernatant was detected on the EWF platform. The dynamic curves of EWF signals could be interpreted by the first-order kinetics and saturation model. The semi-log calibration curve covered a working range of 10–5000 µM of cocaine, and the limit of detection was approximately 10.5 µM. The duration of the full procedure was 990 s (16.5 min), and the detection interval was 390 s (6.5 min). The specified detection of cocaine was confirmed from four typical pharmaceutic agents. The analysis was repeated for 50 cycles without significant loss of sensitivity. Therefore, the aptamer-based EWF biosensor is a feasible solution to rapidly detect cocaine.

YQ, 0000-0001-5423-6869 The rapid detection of cocaine has received considerable attention because of the instantaneous and adverse effects of cocaine overdose on human health. Aptamer-based biosensors for cocaine detection have been well established for research and application. However, reducing the analytic duration without deteriorating the sensitivity still remains as a challenge. Here, we proposed an aptamer-based evanescent wave fibre (EWF) biosensor to rapidly detect cocaine in a wide working range. At first, the aptamers were conjugated to complementary DNA with fluorescence tag and such conjugants were then immobilized on magnetic beads. After cocaine was introduced to compete against the aptamer-DNA conjugants, the released DNA in supernatant was detected on the EWF platform. The dynamic curves of EWF signals could be interpreted by the first-order kinetics and saturation model. The semi-log calibration curve covered a working range of 10 -5000 mM of cocaine, and the limit of detection was approximately 10.5 mM. The duration of the full procedure was 990 s (16.5 min), and the detection interval was 390 s (6.5 min). The specified detection of cocaine was confirmed from four typical pharmaceutic agents. The analysis was repeated for 50 cycles without significant loss of sensitivity. Therefore, the aptamer-based EWF biosensor is a feasible solution to rapidly detect cocaine.

Introduction
In situ analysis of cocaine has received considerable attention because of the instantaneous and adverse effects of cocaine

Conjugation of aptamer-FSP
Buffer 1 (hybridization buffer) was used for the solid surface cleaning and aptamer-FSP conjugation, and buffer 2 (reaction buffer) was prepared for the cocaine competition and fibre analysis [19] (electronic supplementary material, M1). The magnetic beads (MB) with diameter of 1 mm were purchased from a supplier abroad (New England Biolabs Inc, MA, USA). The surface of MB was functionalized with biotin-streptavidin for aptamer immobilization. The conjugation of aptamer and FSP on the surface of MB was based on the hybridization method in previous studies [28,35]. Firstly, 500 ml of the MB was washed intensively with buffer 1 five times and stored in a 1.5 ml centrifugal tube after water evacuation. Secondly, 25 ml of the aptamer (4 mM) was mixed with 25 ml of FSP (4 mM). The mixture was warmed at 958C for 5 min in a water bath and then stored at ambient temperature to assist the conjugation. Thirdly, 50 ml of the aptamer-FSP conjugant was added to the tube with clean MB. Buffer 2 was used to dilute the conjugant into 500 ml of solution, which was further agitated in a thermoshaker for 30 min to ensure surface stabilization. Finally, the MB was separated from the supernatant, washed five times and stored in 500 ml of buffer 2 for the next step of competition. The activity of MB-aptamer would not be obviously changed after storage at 48C for at least two weeks [4]. By assuming complete conjugation between the aptamer and FSP, the concentration of FSP was 0.2 mM in the MB solution.

Fabrication of optic fibres
The functionalized optic fibre was fabricated as previously described in the literature [30,31]. Firstly, a multi-mode quartz optic fibre with diameter 600 mm was immersed into piranha solution (30% hydrogen peroxide in concentrated sulfuric acid in v/v ¼ 1 : 3, extreme CAUTION should be used when working with this dangerous mixture) for 1 h to clean and hydroxylate the fibre surface. After that, the optic fibre was treated by 2% (v/v) of 3-aminopropyl triethoxysilane in toluene for 1 h to add amino groups for further reaction. Later, the fibre was introduced to 4% (v/v) glutaraldehyde in pure water to add aldehyde groups to amino groups. Finally, the fibre was immersed into 250 ml of AAP solution (0.25 mM) at 48C overnight for immobilizing AAP layer on the fibre surface. The fabricated optic fibre can be re-used by simple regeneration using buffer 3 and buffer 4 [20]. Buffer 3 (regeneration buffer) was prepared for the dissociation of FSP-AAP affinity on the fibre surface, and buffer 4 (washing buffer) was used for cleaning afterwards for next cycle of analysis (electronic supplementary material, M1).

Cocaine detection by EWF biosensor
The gradient concentrations of cocaine solutions (0, 10, 25, 50, 100, 250, 500, 1000, 2500 and 5000 mM) were prepared for detection on the EWF biosensor. In each centrifugal tube, 450 ml of cocaine solution was mixed with 50 ml of MB that were immobilized by aptamer-FSP conjugants. After diluting the aptamer-FSP conjugant in MB (0.2 mM) 10 times, the theoretical high limit was 20 nM for FSP to be possibly released to the supernatant. The mixture was agitated in a shaker for 10 min to ensure the complete interaction of cocaine and aptamer [20]. After that, the supernatant was collected in the tube by magnetic separation for 1 min. Part of the supernatant was analysed in a fluorescence spectrometer (F7000, Hitach Corp., Tokyo, Japan) to quantify the released FSP. The supernatant was then injected into the EWF biosensor for analysis with interval of 300 s. After one cycle of detection, the optic fibre was regenerated using buffer 3 and cleansed with buffer 4 for the next cycle, as explained in §2.3. The dynamic signals of fluorescence intensity were recorded and analysed for cocaine quantification.
The fluorescence signals at 300 s were used to make the calibration between the cocaine concentrations and the fluorescence intensities.

Data interpretation
To confirm the competition effects between aptamer-MB and cocaine, the supernatant after the competition was directly introduced to a spectrofluorometer to detect the free FSP inside. Assuming that the total active sites of aptamers were limited on MB and constant for cocaine competition, the free FSP would follow the saturation model described as follows: where f donates the released free FSP, f m is the maximum released FSP, c is the concentration of cocaine, and K c is the half saturation constant. The variables are all expressed in milligram per litre. The parameters f m and K c can be estimated by using Lineweaver -Burk form of linear fitting. During cocaine detection, the fluorescence intensity in terms of EWF biosensor voltage signal was governed by the first-order kinetics as previously described [28,30].
where I t represents the fluorescence intensity at time t, I 0 and I m are the initial and maximum values of fluorescence intensity, respectively, and k is the rate constant (/s). The parameters I m and k can be determined by the nonlinear fitting of the dynamic curves in Matlab (R2014b, Mathworks, USA). An example was given in electronic supplementary material, figure S2 and Code 1.

Selectivity and reproducibility
The specificity of cocaine detection was verified from four pharmaceutic agents including neomycin, sulfadimethoxine, ampicillin and kanamycin, which are among the frequently used antibiotics [36]. The four agents were prepared in 2000 mM in pure water for the full analytical procedure, whereas 100 mM of cocaine solution was used for comparison. The difference in concentrations made the four agents strong interferences to cocaine detection. The experiments were duplicated to control the data quality. The reproducibility of cocaine detection was examined by repeating the detection of 250 mM cocaine 50 times, which followed the full procedure including competition, detection and fibre regeneration.

Scheme of protocol
The concept of competitive affinity process for cocaine is shown in figure 1. Firstly, the aptamers were conjugated with Cy3-labelled FSP and further immobilized on the MB surface. Secondly, the MB was mixed with cocaine for aptamer competition and FSP was released from the aptamer-FSP conjugants to the bulk solution. Thirdly, the released FSP was introduced to the surface of an optic fibre for the affinity with the AAP. Finally, the amount of FSP-AAP hybrid was quantified by the fluorescence intensity of the evanescent wave. Fluorescence excitation, data recording, and processing were automatically achieved on the integrated EWF biosensor as developed previously [29][30][31]. Briefly, a laser beam with a wavelength of 535 nm was generated by a pulsed diode and introduced to the optic fibre via total internal reflection, which stimulated evanescent waves at the fibre surface.

Calibration curve and LOD analysis
The dynamic signals of the EWF biosensor for the detection of cocaine are shown in figure 3a. The EWF signal increased with the analytical time after initializing the detection, due to the dynamic affinity between released FSP and immobilized AAP on the fibre surface. After continuous reaction for 300 s, the detection was ceased for fibre surface regeneration using buffers 3 and 4. Thus, the signal dropped to the level lower than control sample by using buffer 3, and reached the baseline by additional cleaning with buffer 4. The analysis was cycled in detection and regeneration for all cocaine samples in sequence.
The signals at the end of detection (300 s) were used to derive the calibration curve as shown in figure 3b. The semi-log equation explained the dose -response effects over the full working range from 10 to 5000 mM, with R 2 at 0.985 in the form of normal logarithm. The LOD of cocaine was estimated to be 10.5 mM by three times the standard deviation of the triplicate blank samples, which had an average signal of 67.4 + 0.5 mV. The increasing signal of the blank sample in figure 3a implied possible interferences from the sample matrix, the instrumental noise, transducer bias and the environmental variation. Thus, reducing the signal background is necessary to improve the LOD of EWF biosensor, which should be the further task. Currently, periodic calibration of EWF biosensor is recommended to control the data quality.

Experimental performance of full analytical procedure
One advantage of our protocol is the rapid detection within a very short period (990 s), which consists of 10 min of competition, 5 min of detection and 1.5 min of regeneration. Table 2 lists the durations and LODs of aptamer-based cocaine biosensors in the recent literature. Typical experimental duration was almost 1-2 h for a full procedure. A recent interesting study using FRET enhancement achieved analytical duration within 1000 s [24]; however, an expensive spectrofluorometer with fine temperature control (378C) was required. Our previous study achieved 450 s for cocaine analysis using split-aptamer on EWF biosensor [20], but skilled operation was required to strictly follow the analytical procedure. The capability of rapid and simple detection with acceptable sensitivity favours our protocol, which allows the use of inexpensive instruments and less human labour.
The LOD of cocaine in this study was approximately 10.5 mM, which is comparable to the reported biosensors of cocaine, e.g. 10 mM by electronics [10] and 10 mM by fluorescence [11] and colorimetry [12]. Usually instrumental analyses with sample extraction own better LOD than biosensors, e.g. 4.9 nM through LC-MS/MS [3] and 19.8 nM via GC-MS [2]. Thus, it is necessary to reduce the LOD of biosensors by various signal amplification techniques, e.g. 0.2 mM by fluorescence quenching [20] and 1.3 nM by enzymatic amplification [14]. LODs lower than 1 nM cocaine are also available by NP enhancement, e.g. 1 nM by graphene-gold NP [21], 0.48 nM by gold NP [18] and 0.29 nM by goldsilica NP [39]. However, expensive materials and advanced instruments are required for the above signal amplification. The previous all-fibre biosensor can detect 0.165 mM of cocaine in serum [20], but the working range is limited below 200 mM and skilled operation is required. Considering that an LOD of 10.5 mM is feasible for the scenarios of rapid screening, the current sensitivity was acceptable by considering its simplicity and rapid operation. Anyway, it is necessary to reduce the background signal in further study, as well as to examine effects of matrices in actual samples by the standard method with recovery rates.

Selectivity and reproducibility
Selectivity is one of the key indicators in the applicability of developed aptasensor. Figure 4a shows the different signals between cocaine and the four other pharmaceutics. A control sample was provided using buffer 2 only. The signal of pharmaceuticals at 300 s on the EWF biosensor was subtracted by that of the control sample. All four pharmaceutics at 2 mM had negligible responses but cocaine sample (0.1 mM) generated an obvious and positive signal (70 mV). The result supported highly specific interaction of aptamers and cocaine molecules [20,27]. Quinine, which was once used to treat malaria and is still used in tonic water, compete the aptamer with strong affinity [40], thus quinine use should be excluded before the detection.   aptamer and optic fibre, fluorescence 990 s 10.5 mM this study a Sample pretreatment is excluded from the duration. b LOD is the abbreviation of limit of detection and the values are uniformed in micromolar. c FRET is the abbreviation of fluorescence resonance energy transfer.