A Photoluminescent Colorimetric Probe of Bovine Serum Albumin-Stabilized Gold Nanoclusters for New Psychoactive Substances: Cathinone Drugs in Seized Street Samples

Screening of illicit drugs for new psychoactive substances—namely cathinone—at crime scenes is in high demand. A dual-emission bovine serum albumin-stabilized gold nanoclusters probe was synthesized and used for quantitation and screening of 4-chloromethcathinone and cathinone analogues in an aqueous solution. The photoluminescent (PL) color of the bovine serum albumin-stabilized Au nanoclusters (BSA-Au NCs) probe solution changed from red to dark blue during the identification of cathinone drugs when excited using a portable ultraviolet light-emitting diodes lamp (365 nm). This probe solution allows the PL color-changing point and limit of detection down to 10.0 and 0.14 mM, respectively, for 4-chloromethcathinone. The phenomenon of PL color-changing of BSA-Au NCs was attributed to its PL band at 650 nm, quenching through an electron transfer mechanism. The probe solution was highly specific to cathinone drugs, over other popular illicit drugs, including heroin, cocaine, ketamine, and methamphetamine. The practicality of this BSA-Au NCs probe was assessed by using it to screen illicit drugs seized by law enforcement officers. All 20 actual cases from street and smuggling samples were validated using this BSA-Au NCs probe solution and then confirmed using gas chromatography–mass spectrometry. The results reveal this BSA-Au NCs probe solution is practical for screening cathinone drugs at crime scenes.


Introduction
Synthesized cathinones, a type of new psychoactive substances (NPS), have emerged in the past 10 years and have caused serious social problems worldwide [1]. Cathinone drug abuse has adverse psychiatric effects, including anxiety, anorexia, addiction, and panic attacks, leading to homicidal combative behavior and deaths, which puts a community in danger [2]. More than 30 synthesized cathinone analogues derived from natural cathinone with main structure of phenyl-ketone group had been reported by 2018. The fast appearance of cathinone drugs on the drug market poses a considerable challenge for law enforcement officers and analytical chemists [3].

Synthesis of BSA-Au NCs
The synthetic process in this study was similar to that described in relevant research, with moderate modifications [26]. An aliquot (1.0 mL) of HAuCl 4 (aq) (10 mM) was mixed with 10 mL of BSA (aq) (30 mg mL −1 ) in a 50-mL centrifugation tube. After stirring for 30 min, 0.5 mL of NaOH (aq) Sensors 2019, 19, 3554 3 of 11 (1 M) was added to the mixture and incubated at 45 • C for 18 h. The color of the solution changed from light yellow to dark yellow, indicating the formation of BSA-Au NCs. With the aim to remove unreactive gold species and to purify the resulting BSA-Au NCs solution, the mixture was purified against ultrapure water through a dialysis membrane (MWCO 3.5 kD) purchased from Spectra Labs (Rancho Dominguez, CA, USA) for 24 h. Next, the purified BSA-Au NCs solution was freeze-dried to form a light pink powder (259 mg), which was then re-dispersed in sodium phosphate buffers (pH 7.0, 100 mM). The as-prepared BSA-Au NCs solution (22.5 mg mL −1 ) was stored at 4 • C in the dark until ready to use.

Characterization of BSA-Au NCs
The BSA-Au NCs light pink powder (0.1 mg) was diluted with ultrapure water (1 mL). The diluted BSA-Au NCs were deposited on 400-mesh carbon-coated Cu grids, and excess solvent was evaporated at an ambient temperature and pressure. Images of BSA-Au NCs were captured using transmission electron microscopy (TEM) purchased from GCE Market (FEI Tecnai-G2-F20, Blackwood, NJ, USA) operated at 200 kV to measure particle sizes. A SpectraMax i3x Multi-Mode Microplate Reader from Molecular Devices (San Jose, CA, USA) was used to record the absorption spectrum and PL spectrum of BSA-Au NCs. A spectrometer (Theta Probe) from VG Scientific (East Grinstead, UK) with Al Kα X-ray radiation was used to measure X-ray photoelectron spectroscopy (XPS) of BSA-Au NCs. A Thermo Scientific spectrometer (Nicolet iS5, Waltham, MA, USA) with a resolution of 4 cm −1 and 64 scans was used to record the Fourier transform infrared (FTIR) spectrum of BSA-Au NCs and BSA to analyze the functional groups on the surface. Fluorescent quantum yields (Φ f ) of the BSA-Au NCs dispersed in a sodium phosphate buffer (pH 7.0, 100 mM) at the excitation and emission wavelengths of 365/460 nm and 365/650 nm, were separately measured using quinine sulfate (Φ f = 0.54) dissolved in 0.1 M H 2 SO 4 as a standard [27].

Detection of 4-CMC and Cathinone Analogues
An aliquot (100 µL) of the as-prepared BSA-Au NCs solution was dispersed in a sodium phosphate buffer (400 µL, 100 mM, pH 7.0) containing various concentrations of 4-CMC to evaluate the sensitivity of BSA-Au NCs for detection of 4-CMC. After mixing for 1 min, the PL spectrum of the BSA-Au NCs solution with and without 4-CMC were recorded. Similarly, a selectivity test was assessed using 25 mM of other illicit drugs, including heroin, cocaine, methamphetamine, and ketamine dissolved in sodium phosphate buffers, separately. Additives such as glucose, sucrose, and fructose were tested for their interference. In addition, cathinone analogues, including butylone, 4 -chloro-α-PPP, dibutylone, ephylone, ethylone, 3-MeOMC, 4-MEAPP, mexedrone, and pentylone were tested to confirm the sensing ability of BSA-Au NCs probe for detection of cathinone drugs. All experiments were performed under ambient conditions (25 • C, 1 atm). To investigate the sensing mechanism, a reduction of 4-CMC to form 1-(4-chlorophenyl)-2-(methylamino) propan-1-ol was synthesized by using a reducing reagent of NaBH 4 (see Supporting Information I and Figure S1) to study analyte-induced PL quenching. In addition, time-resolved absorption and emission spectroscopy (NS010, Pascher Instrument, Lund, Sweden) was performed using a photon-counting system with a 365 nm laser to measure lifetimes of BSA alone and BSA-Au NCs with and without 4-CMC (25 mM), which were dispersed in sodium phosphate buffers (100 mM, pH 7.0).

Interference Study
An aliquot (100 µL) of the as-prepared BSA-Au NCs solution was mixed with 400 µL of the sodium phosphate buffer (100 mM, pH 7.0) in a 2-mL centrifugation tube for use as a BSA-Au NCs probe solution. Various concentrations of 4-CMC mixed with glucose (w/w% from 0% to 20%) were individually prepared for interference study. Each real sample (30 mg) was dissolved in 500 µL of a BSA-Au NCs probe solution. After being mixed for 1 min, PL images of the BSA-Au NCs probe solution were recorded using a smartphone (Galaxy Note 3, Samsung, Seoul, South Korea) and a portable

Theoretical Calculation
The structures of 4-CMC and its reduction form of 1-(4-chlorophenyl)-2-(methylamino) propan-1-ol were optimized by using Gaussian 09 [28]. Then, density-functional theory (DFT) calculation based on B3LYP/6-31G** mode was used to calculate electron distribution and energy levels of the two analytes for their highest occupied molecular orbitals (HOMOs) and lowest unoccupied molecular orbitals (LUMOs). Molecular orbitals were displayed using a Chemcrafe.

Analysis of Actual Samples
Twenty actual cases from street and smuggling samples from different sources were validated by using a BSA-Au NCs probe solution and a portable UV-LED lamp, and then they were confirmed using GC-MS measurement (Agilent 7890A/5975C MSD). The procedure was similar to that used for interference study, but only 2 to 5 mg of unknown samples were sampled for screening.

Characterization of BSA-Au NCs
The TEM image displayed in Figure 1A shows that the as-prepared BSA-Au NCs are uniform and monodispersed spheres, with an average diameter of 3.2 nm (100 counts) [29]. The XPS spectrum presented in Figure 1B reveals the existence of various elements in BSA-Au NCs, including Au, O, N, and C. Binding energies of 285.1, 399.1, and 531.1 eV were attributed to C 1s , N 1s , and O 1s , respectively [30]. In the case of Au, binding energies of 84.3 and 87.6 eV were attributed to 4f 7/2 and 4f 5/2 , respectively, and this further confirmed the formation of BSA-Au NCs [30]. Moreover, the deconvoluted Au XPS spectrum of BSA-Au NCs revealed four peaks obtained at 84.3 and 87.8, 85.7, and 89.6 eV, which were attributed to oxidation states of Au (0) and Au (I), respectively, as shown in Figure 1C [31,32]. The number of gold atoms in the BSA-Au NCs was proposed approximately to be 25, due to its red emission from Au clusters [32][33][34]. As indicated in Figure S2, the FTIR spectrum of BSA-Au NCs was similar to that of BSA, further confirming that Au NCs was stabilized in the BSA [30]. The absorption and PL spectrum of the BSA and BSA-Au NCs are provided in Figure 1D. The BSA-Au NCs had a significant absorption band at a wavelength of 280 nm that was attributable to the tryptophan in BSA [35]. The absorption spectrum of BSA-Au NCs was similar to that of BSA alone, further indicating that BSA was stabilized on the surface of BSA-Au NCs. The PL spectrum of BSA-Au NCs showed two emission bands at 460 nm and 650 nm when excited at 365 nm. The emission bands of BSA-Au NCs at 460 nm and 650 nm were blue and red in color and corresponded to BSA and Au NCs, respectively. Table 1 summarizes the lifetimes of BSA and BSA-Au NCs at PL bands of 460 nm and 650 nm at an excitation wavelength of 365 nm. The lifetimes of BSA-Au NCs at PL bands of 460 nm and 650 nm were 10.5 ns and 1820.7 ns, respectively [36], and the lifetime of BSA at the PL band of 460 nm was 11.3 ns, indicating that the PL band at 460 nm of BSA-Au NCs should be attributed to BSA. Fluorescent quantum yield values for BSA-Au NCs at 460 nm and 650 nm were 1.2% and 1.5%, respectively. The stability of BSA-Au NCs against time and pH was assessed as shown in Figure S3. The results revealed that the PL intensities of BSA-Au NCs for a time of 480 min and pH values in the range of 7-11.5 were stable, demonstrating excellent stability for use. After pH values (7, 9.5, and 11.5) was assessed, a mild pH of 7.0 was selected for further study [37].  where F and F0 are the PL intensities of the BSA-Au NCs at 650 nm with and without 4-CMC, respectively. The LOD was calculated to be 0.14 mM based on 3δ/slope, where δ represented the deviation of (F0−F)/F0 for a blank test and the slope denoted the calibration curve. As shown in Figure  2B, the results of the selectivity test demonstrate that the BSA-Au NCs probe is specific to 4-CMC and cathinone analogues (their structures are displayed in Table S1). In addition, common illicit drugs and additives, including cocaine, heroin, ketamine, methamphetamine, glucose, sucrose, and fructose, did not induce BSA-Au NCs quenching. Figure 3A presents the PL color of the BSA-Au NCs probe solutions with increasing concentrations of 4-CMC. The change in PL color from red to dark blue was visible to the naked eye; thus, it was determined that the changing point was 10.0 mM, indicating that a small amount of the analyte (1 mg) was enough for this probe solution to change PL color. Figure 3B shows that the 4-CMC-induced PL color of BSA-Au NCs probe solution changed mixing with glucose at various concentration levels (w/w% from 0% to 25%). When the concentration  Figure 2A (inset) shows that the relationship (F 0 −F)/F 0 of the BSA-Au NCs is linear for the range of 0.48-7.5 mM of 4-CMC (R 2 = 0.998), where F and F 0 are the PL intensities of the BSA-Au NCs at 650 nm with and without 4-CMC, respectively. The LOD was calculated to be 0.14 mM based on 3δ/slope, where δ represented the deviation of (F 0 −F)/F 0 for a blank test and the slope denoted the calibration curve. As shown in Figure 2B, the results of the selectivity test demonstrate that the BSA-Au NCs probe is specific to 4-CMC and cathinone analogues (their structures are displayed in Table S1). In addition, common illicit drugs and additives, including cocaine, heroin, ketamine, methamphetamine, glucose, sucrose, and fructose, did not induce BSA-Au NCs quenching. Figure 3A presents the PL color of the BSA-Au NCs probe solutions with increasing concentrations of 4-CMC. The change in PL color from red to dark blue was visible to the naked eye; thus, it was determined that the changing point was 10.0 mM, indicating that a small amount of the analyte (1 mg) was enough for this probe solution to change PL color. Figure 3B shows that the 4-CMC-induced PL color of BSA-Au NCs probe solution changed mixing with glucose at various concentration levels (w/w% from 0% to 25%). When the concentration of 4-CMC was higher than 5%, the PL color of BSA-Au NCs was completely blue, indicating high concentration of glucose interference would not change the sensing results of BSA-Au NCs for the screening of 4-CMC.

Sensitivity, Selectivity, and Interference Study
Sensors 2019, 19, x FOR PEER REVIEW 6 of 11 of 4-CMC was higher than 5%, the PL color of BSA-Au NCs was completely blue, indicating high concentration of glucose interference would not change the sensing results of BSA-Au NCs for the screening of 4-CMC.

Sensing Mechanism
A Stern-Volmer plot, as displayed in Figure 4A, was used to study the quenching mechanism of BSA-Au NCs for 4-CMC at emission and excitation wavelengths of 650 nm and 365 nm. The slope (KSV) of the linear plot was determined to be 0.269 (R 2 = 0.996), revealing a dynamic quenching mechanism (fluorescence resonance energy transfer or electron transfer) [38]. In addition, the lifetimes of BSA-Au NCs without and with 25 mM of 4-CMC at emission and excitation wavelengths of 650 nm and 365 nm were determined to be 1820.7 and 963.4 ns, respectively [39]. To confirm fluorescence resonance energy transfer or electron transfer, the PL spectrum of BSA-Au NCs at an excitation wavelength of 365 nm and the absorption spectrum of 4-CMC were plotted, as displayed in Figure 4B. The results revealed no overlap between the PL spectrum of BSA-Au NCs in the range

Sensing Mechanism
A Stern-Volmer plot, as displayed in Figure 4A, was used to study the quenching mechanism of BSA-Au NCs for 4-CMC at emission and excitation wavelengths of 650 nm and 365 nm. The slope (KSV) of the linear plot was determined to be 0.269 (R 2 = 0.996), revealing a dynamic quenching mechanism (fluorescence resonance energy transfer or electron transfer) [38]. In addition, the lifetimes of BSA-Au NCs without and with 25 mM of 4-CMC at emission and excitation wavelengths of 650 nm and 365 nm were determined to be 1820.7 and 963.4 ns, respectively [39]. To confirm fluorescence resonance energy transfer or electron transfer, the PL spectrum of BSA-Au NCs at an excitation wavelength of 365 nm and the absorption spectrum of 4-CMC were plotted, as displayed in Figure 4B. The results revealed no overlap between the PL spectrum of BSA-Au NCs in the range

Sensing Mechanism
A Stern-Volmer plot, as displayed in Figure 4A, was used to study the quenching mechanism of BSA-Au NCs for 4-CMC at emission and excitation wavelengths of 650 nm and 365 nm. The slope (K SV ) of the linear plot was determined to be 0.269 (R 2 = 0.996), revealing a dynamic quenching mechanism (fluorescence resonance energy transfer or electron transfer) [38]. In addition, the lifetimes of BSA-Au NCs without and with 25 mM of 4-CMC at emission and excitation wavelengths of 650 nm and 365 nm were determined to be 1820.7 and 963.4 ns, respectively [39]. To confirm fluorescence resonance energy transfer or electron transfer, the PL spectrum of BSA-Au NCs at an excitation wavelength of 365 nm and the absorption spectrum of 4-CMC were plotted, as displayed in Figure 4B. The results revealed no overlap between the PL spectrum of BSA-Au NCs in the range of 530-730 nm and the absorption spectrum of 4-CMC, ruling out fluorescence resonance energy transfer [40,41]. Thus, an electron transfer occurred from BSA-Au NCs to 4-CMC when excited at 365 nm, mainly due to the phenyl-ketone structure in 4-CMC, which was a good electron acceptor [42]. To prove this, 4-CMC was reduced to form 1-(4-chlorophenyl)-2-(methylamino) propan-1-ol using NaBH 4 as shown in supporting information I; thus, the analyte did not induce PL quenching of BSA-Au NCs. Also, electron distribution and energy levels of HOMO and LUMO for 4-CMC were acquired by DFT calculation. As shown in Figure 4C, the HOMO orbitals are localized on the amino group, and the LUMO orbitals are delocalized on a phenyl-ketone structure. Their HOMO and LUMO energy levels are -5.90 and -1.83 eV, respectively. The DFT calculation results further support the phenyl-ketone group in 4-CMC, which would accept an electron from BSA-Au NCs. On the other hand, HOMO and LUMO energy levels of 1-(4-chlorophenyl)-2-(methylamino) propan-1-ol are-6.51 and -0.43 eV, respectively, indicating the high energy level of LUMO would not allow an electron transfer from BSA-Au NCs. On the basis of the Stern-Volmer plot, the lifetimes, absorption, PL, and experiment results of a reductive form of 4-CMC as well as DFT calculation data, a sensing mechanism of BSA-Au NCs for 4-CMC, is proposed, as displayed in Scheme 1. of 530-730 nm and the absorption spectrum of 4-CMC, ruling out fluorescence resonance energy transfer [40,41]. Thus, an electron transfer occurred from BSA-Au NCs to 4-CMC when excited at 365 nm, mainly due to the phenyl-ketone structure in 4-CMC, which was a good electron acceptor [42].
To prove this, 4-CMC was reduced to form 1-(4-chlorophenyl)-2-(methylamino) propan-1-ol using NaBH4 as shown in supporting information I; thus, the analyte did not induce PL quenching of BSA-Au NCs. Also, electron distribution and energy levels of HOMO and LUMO for 4-CMC were acquired by DFT calculation. As shown in Figure 4C,

Analysis of Actual Samples
Twenty actual cases from street and smuggling samples were validated by using this BSA-Au NCs probe solution and then confirmed using GC-MS measurements, as shown in Figure 5. Actual cases of 2, 3, 11, and 16 screened positive results and were thus confirmed to be ephylone, butylone, a mixture of dibutylone and 4-chloro-N,N-dimethylcathinone (4-chloro-N,N-DMC), and 4-chloro-N,N-DMC, respectively, which all were cathinone drugs. Other cases tested negative and were thus confirmed to be amphetamine, ketamine, methoxetamine, cocaine, 2,5-dimethoxy-4bromophenethylamine (2C-B), 3,4-methylenedioxymethamphetamine (MDMA), ACHMINACA, 5chloro-AKB48, fentanyl, etizolam, and heroin, respectively. All screening results fit the outcomes confirmed by GC-MS measurements. Moreover, it was noted that PL images of BSA-Au NCs probe solutions were captured using a smartphone and a portable UV-LED lamp, indicating that the developed sensor system was practical and convenient for law enforcement officers to use at crime scenes.

Analysis of Actual Samples
Twenty actual cases from street and smuggling samples were validated by using this BSA-Au NCs probe solution and then confirmed using GC-MS measurements, as shown in Figure 5. Actual cases of 2, 3, 11, and 16 screened positive results and were thus confirmed to be ephylone, butylone, a mixture of dibutylone and 4-chloro-N,N-dimethylcathinone (4-chloro-N,N-DMC), and 4-chloro-N,N-DMC, respectively, which all were cathinone drugs. Other cases tested negative and were thus confirmed to be amphetamine, ketamine, methoxetamine, cocaine, 2,5-dimethoxy-4-bromophenethylamine (2C-B), 3,4-methylenedioxymethamphetamine (MDMA), ACHMINACA, 5-chloro-AKB48, fentanyl, etizolam, and heroin, respectively. All screening results fit the outcomes confirmed by GC-MS measurements. Moreover, it was noted that PL images of BSA-Au NCs probe solutions were captured using a smartphone and a portable UV-LED lamp, indicating that the developed sensor system was practical and convenient for law enforcement officers to use at crime scenes.

Analysis of Actual Samples
Twenty actual cases from street and smuggling samples were validated by using this BSA-Au NCs probe solution and then confirmed using GC-MS measurements, as shown in Figure 5. Actual cases of 2, 3, 11, and 16 screened positive results and were thus confirmed to be ephylone, butylone, a mixture of dibutylone and 4-chloro-N,N-dimethylcathinone (4-chloro-N,N-DMC), and 4-chloro-N,N-DMC, respectively, which all were cathinone drugs. Other cases tested negative and were thus confirmed to be amphetamine, ketamine, methoxetamine, cocaine, 2,5-dimethoxy-4bromophenethylamine (2C-B), 3,4-methylenedioxymethamphetamine (MDMA), ACHMINACA, 5chloro-AKB48, fentanyl, etizolam, and heroin, respectively. All screening results fit the outcomes confirmed by GC-MS measurements. Moreover, it was noted that PL images of BSA-Au NCs probe solutions were captured using a smartphone and a portable UV-LED lamp, indicating that the developed sensor system was practical and convenient for law enforcement officers to use at crime scenes.