Illicit drugs consumption, such as cocaine and opiates, is a worldwide steadily increasing problem. Their fast and reliable detection is critical for both the perspective of medical intervention and law enforcement [1].

Cocaine is a powerfully addictive drug causing severe side effects mostly through the blockade of catecholamine reuptake [2]. Currently, there are many techniques engaged in cocaine quantification, such as spectrometry and chromatography [3]. The majority of them are costly, bulky, or complicated. On the other hand, sniffer dogs commonly use at the airports can work only for limited time periods and require trainers contributing to their expense [4].

A biosensor is a handheld user-friendly alternative to other more complex detection systems. It is a device that comprises a transducer responsible for the sensitivity, and a receptor that ensures the selectivity of the detection [5]. Lately, our focus has been devoted to the implementation of a promising new class of recognition elements known as aptamers or “chemical antibody” [6] to the optical biosensor. Aptamers are long oligonucleotides that are able to bind target molecules with high affinity and specificity [7,8]. They are similar to antibodies in their binding ability; however, aptamer production is easier and more cost-efficient compared to traditional methods for obtaining antibodies [6].

As reported in the literature, cocaine aptasensors usually rely on electrochemical or optical (fluorescent, colorimetric, or chemiluminescent) transducers [9]. Among the last group, we should emphasize the optical fiber aptasenors, which recently have been vigorously developed, like fluorescent cocaine aptasensors based on evanescent wave fiber [10,11]. Besides aptasensing, cocaine can be fluorescently detected by an optical fiber combined with other receptors, such as antibodies [12–14] or molecularly imprinted polymers [15–17]. However, the major disadvantage associated with fluorescent techniques is the requirement of prior labeling, which is complex, more expensive, and affects the receptor stability and selectivity [18].

Herein, we present a proof-of-concept of a label-free cocaine detection system using an LPFG probe. The sensor measures the change of the refractive index (RI) of probe surroundings. We combined the structure-switching cocaine aptamer MN6 with the ultrasensitive optical LPFG transducer. This platform did not require any enhancement of its sensitivity resulting in a straightforward device.

The LPFG is produced by imprinting a periodic variation of the RI into the core, the cladding of the optical fiber, or both [19]. The resonance wavelength of the LPFG is determined by effective RI and is affected by physical perturbations like temperature, strain pressure, as well as variation in the RI of the external medium. This last feature has been used as the basis for signal transduction and biosensor development [20–22].

LPFG combined with an aptamer has been applied for recognition of more complex molecules which provoke a significant change of RI, e.g., enzyme thrombin [23] and bacterial membrane protein [24].

To date, cocaine detection based on LPFG platforms has not been reported to the best of our knowledge. Nevertheless, LPFG-aptasensors have been applied for a small molecule detection such as ATP, using a signaling structure-switch aptamer, which releases DNA upon binding the target causing a large change of the RI [25].

Here, as a bioreceptor, we used a flexible aptamer MN6, which binds to cocaine molecules and folds capturing the target inside of the oligonucleotides chain [26,27] (Fig. 1). By using a sensitive LPFG, we were able to observe and track subtle RI perturbations caused by the change of the aptamers layer density after their conformational switch. This RI variation was shown as a shift in the transmission spectrum of the LPFG.

 

Fig. 1. Scheme representing the LPFG probe functionalized with the cocaine aptamer MN6 and its conformational change upon binding cocaine.

Download Full Size | PDF

The LPFG probe was inscribed in the core of a single mode optical fiber (Corning SMF-28) by UV exposition through an amplitude chromium mask ($\lambda =226.8\mathrm{\mu m}$) using a Pulse Master 840 high-power KrF excimer laser ($\lambda =248\mathrm{nm}$, Lumonics Lasers: Pulse Master-840) [28]. The LPFG sensitivity towards RI in a dual resonance regime was obtained by slowly etching the cladding with hydrofluoric acid. This etching tunes the resonance wavelength close to the dispersion turning point, where the sensitivity of the LPFG is the highest [29]. The spectra were measured using a Yokogawa AQ6370C optical spectrum analyzer in the range of 1100–1700 nm, with an NKT Photonics SuperK COMPACT supercontinuum white light source. The RI measurements of the liquids applied for LPFGs calibration towards RI were done using a digital refractometer VEE GEE PDX-95. The temperature was monitored continuously using a HP34970A Data Acquisition Unit and did not change within the time of the study. The tension was fixed using a piezo tension controller (Strain Gauge Reader TSG001 ThorLabs, Ultra-High Resolution Motion System NanoPZ, Newport) and by gluing the fiber to a silica flow-cell to prevent solvent evaporation (to elude false-positive signals). Moreover, the stability of the receptor attachment to the sensor surface was ensured by the covalent receptor immobilization via a thiourea bond presented in Fig. 1.

Before aptamer immobilization, the LPFG sensitivity towards RI was determined as a slope of the resonance wavelength and RI indices of a reference liquid and was equal to 6000 RIU nm-1. The platform was washed with a mixture of hydrochloric acid and methanol ($1:1$, v/v) for 30 min and then exposed to (3-aminopropyl)triethoxysilane (APTES, 99% Sigma-Aldrich) and triethylamine (catalyst, $\ge 99\%$ Sigma-Aldrich) vapors in order to coat the surface with amine groups [28,30]. Next, the surface was activated for the receptor immobilization, with a homobifunctional cross-linker, 1,4-phenylene diisothiocyanate (PDITC) [31]. For this, the LPFG was incubated in 100 mM PDITC (anhydrous, 99.8%, Sigma-Aldrich) and 100 mM 4-(dimethylamino)pyridine in anhydrous dimethylformamide for 2 h. The 5′-amino modified aptamer MN6 (2 μM, Alpha DNA) in phosphate buffered saline $\mathrm{pH}=7.4$ (PBS, P4417 Sigma-Aldrich) was preheated at 90°C for 5 min to unfold the DNA structure, immediately cooled down on the ice, and incubated with the sensor for 1 h. Subsequently, the LPFG was washed with PBS out from weakly bonded aptamers before incubating with 50 mM glycine in carbonate buffer $\mathrm{pH}=9.6$ for 30 min to block unreacted surface groups (Fig. 1).

The obtained biosensor was then exposed to cocaine hydrochloride (Sigma-Aldrich) in PBS for 30 min. During this process, the transmission spectrum was tracked. After the incubation, the LPFG was washed with PBS, and the spectrum was recorded once again. The procedure was repeated with increasing concentrations of cocaine ranging from 25 to 100 μM.

The fitting curve to the data was required to find real minima and to show slight changes in the spectrum. Raw data and curve fitted to the observed double wavelength resonance peaks are presented in Fig. 2.

 

Fig. 2. Transmission spectrum recorded for 75 μM cocaine in PBS, raw data (red dots) and the fitted curve (blue line).

Download Full Size | PDF

While exposing the biosensor to the cocaine solution, we continuously monitored the transmission spectra. As an example, Fig. 3 presents a 4.2 nm wavelength shift between the first (black solid curve) and the last measurement (25th, red dashed curve) recorded for 75 μM cocaine. What we have noticed, changes occurred during the incubation might indicate the interaction of the aptamer MN6 with cocaine molecules at the biosensor surface. Moreover, the observed spectral shift depends on the applied cocaine concentration (Fig. 4) and linearly increasing up to 75 μM cocaine while dropping rapidly at 100 μM, suggesting saturation of the receptor.

 

Fig. 3. Transmission spectra recorded for 75 μM cocaine in PBS: the first (black solid curve) and the last measurement (25th, red dashed curve), and the difference between their resonance minima, 4.2 nm.

Download Full Size | PDF

 

Fig. 4. Resonance wavelength shift observed between the 1st and 25th measurements recorded at the biosensor immersed in cocaine in PBS (25, 50, 75, and 100 μM).

Download Full Size | PDF

Moreover, after the LPFG incubation in cocaine solution, the sensor was washed with PBS and then several spectra were recorded in a fresh PBS without any significant changes of the minimum wavelength ($\sigma \le 0.21\mathrm{nm}$, $n=10$). The stable signal confirms that the previous spectral shift has its origin in the receptor-target interaction.

We attempted detecting four concentrations of cocaine in the range of 25 to 100 μM (Fig. 5). The spectrum recorded in PBS after the functionalization process (Fig. 5, black solid curve) shows that LPFG works in the dispersion turning point, and it was set as the reference spectrum. Upon the interaction with cocaine, the attenuation of the single resonance increased and started splitting into a double resonance (Fig. 5, red dashed dotted curve). Further exposition to higher concentrations of cocaine produced a proportional expansion in the double resonance (Fig. 5. Red dashed dotted, purple long dashed, green dashed, blue dotted curves), which is correlated to a change in the RI at the biosensor surface. We ensured that the shift did not raise from the change of the solution RI, as all measurements were taken in fresh PBS after extensive washing of the sensor.

 

Fig. 5. Transmission spectra recorded in PBS after the LPFG functionalization process (black solid reference curve) and for the detection of 25, 50, 75, 100 μM cocaine (red dashed–dotted, purple long dashed, green dashed, blue dotted curves, respectively).

Download Full Size | PDF

Based on the recorded spectra (Fig. 5), a calibration curve for cocaine was generated (Fig. 6). The curve represents the relationship between the wavelength of the left resonance minimum shift and cocaine concentration. Several measurements were taken during each step of the experiment to confirm the stability of the signal ($\sigma \le 0.15\mathrm{nm}$, $n=5$). Next, we calculated the average of five resonance wavelength minima measurements for every cocaine concentration. The difference between the averages of the resonances for the reference and 25 μM cocaine solution (${\lambda }_{\mathrm{R}\text{Reference}}-{\lambda }_{\mathrm{R}\text{Cocaine}}=\mathrm{\Delta }{\lambda }_{\mathrm{R}\text{Ref-Coc}}$) equals to $\mathrm{\Delta }{\lambda }_{\mathrm{R}\text{Ref-Coc}}=2.7\mathrm{nm}$ ($\mathrm{RSD}\%=1.72\%$). Thus, 25 μM is the lowest cocaine concentration causing a measurable shift in the spectrum. We observed the linear relationship for the first three cocaine concentrations with a correlation of determination equal to $R^2=0.998$ (Fig. 6). Whereas, the sensor exposition to 100 μM cocaine caused flattening the signal correlated with the lower cocaine accumulation at the sensor surface visible in Fig. 4. This indicates that the receptor has been saturated by cocaine during the detection process, which is typical behavior for a biosensor containing a receptor with limited binding sites [22,32]. Although the obtained detection limit 25 μM is higher than other described cocaine aptasensor, they are working based on a different phenomenon, e.g., fluorescence 10.5 μM [11], chemiluminescence 1 nM [33], label-free fluorescence 200 nM [34], label-free acoustic wave sensing 0.9 μM [35].

 

Fig. 6. Calibration curve for cocaine detection at LPFG modified with aptamer MN6 and its linear range (insert). Measurements were taken in PBS.

Download Full Size | PDF

To conclude, for the very first time, to the best of our knowledge, we presented a cocaine biosensor which links a highly sensitive optical probe, LPFG, with a new class of receptors, structure-switching aptamers. The biosensor works based on the change of the RI of the LPFG surface triggered by the interactions between aptamers and cocaine molecules. The receptors behavior is reflected in the transmission spectrum and can be tracked by the resonance minima shifts proportional to the cocaine concentration in the range of 25 to 100 μM. By applying the ultrasensitive transducer, such as the LPFG, we achieved a promising and simple device which parameters, such as detection limit, can be improved by further research.