Application of sulfur and nitrogen doped carbon quantum dots as sensitive fluorescent nanosensors for the determination of saxagliptin and gliclazide

In this study, highly fluorescent sulfur and nitrogen doped carbon quantum dots (S,N-CQDs) were used as fluorescent nanosensors for direct spectrofluorimetric estimation of each of gliclazide (GLZ) and saxagliptin (SXG) without any pre-derivatization steps for the first time. S,N-CQDs were synthesized employing a simple hydrothermal technique using citric acid and thiosemicarbazide. The produced S,N-CQDs were characterized using different techniques including fluorescence emission spectroscopy, UV spectrophotometry, high-resolution transmission electron microscopy and FT-IR spectroscopy. Following excitation at 360 nm, S,N-CQDs exhibited a strong emission peak at 430 nm. The native fluorescence of S,N-CQDs was quantitatively enhanced by addition of increased concentrations of the studied drugs. The fluorescence enhancement of S,N-CQDs and the concentrations of the studied drugs revealed a wide linear relationship in the range of 30.0–500.0 µM and 75.0–600.0 µM with limits of detection of 5.0 and 10.15 µM for GLZ and SXG, respectively. The proposed method was efficiently used for determination of cited drugs in their commercial tablets with % recoveries ranging from 98.6% to 101.2% and low % relative standard deviation values (less than 2%). The mechanism of interaction between S,N-CQDs and the two drugs was studied. Validation of the proposed method was carried out in accordance with International Conference on Harmonization (ICH) guidelines.


Introduction
Gliclazide (GLZ) (N-[[(hexahydrocylopenta[c]pyrrol-2(1H)-yl)amino]carbony]-4-methylbenzene sulfonamide) (figure 1a) is a second generation sulfonylurea used for the control of type 2 diabetes mellitus (DM) by increasing insulin amount secreted by the pancreatic cells [1]. Previous reports showed that GLZ slows down the progression of diabetic retinopathy. It also has low incidence of hypoglycaemia and good tolerability. Consequently, it is considered as the drug of choice for long-term management of non-insulindependent DM patients [1,2].
Carbon quantum dots (CDs) are considered as a novel type of fluorescence nanomaterials that range in size from 2.0 to 10.0 nm. They are biocompatible, non-toxic, easily synthesized from cheap starting materials and chemically stable with good photoluminescence properties. Furthermore, they can be easily functionalized using a variety of species and they are highly water soluble [26][27][28][29]. For preparation of fluorescent CDs, various approaches have been reported like chemical oxidation [30], microwave-assisted [31], carbonizing organic methods [32] and hydrothermal synthesis [33]. Synthesis of doped CDs is usually carried out by inserting heteroatoms (such as nitrogen (N), phosphorus (P), sulfur (S), fluorine (F), boron (B)) into the CDs' general structure to enhance their photoluminescence properties [26,[34][35][36]. From literature, CDs were efficiently applied as sensitive probes for the determination of many pharmaceutical compounds in different matrices [26,37,38].
In the present study, a green, simple and economic method was adopted for preparation of S,N-CQDs applying a hydrothermal synthetic approach using thiosemicarbazide (TS) as nitrogen and sulfur source and citric acid (CA) as carbon source producing fluorescence probes for the estimation of each of GLZ and SXG [26]

Instruments and software
UV-1601 PC spectrophotometer was used to carry out UV spectrophotometric measurements (Shimadzu, Kyoto, Japan) using a 1 cm quartz cell. Agilent Technologies' Cary Eclipse fluorescence spectrophotometer was used for fluorescence measurements (Santa Clara, USA). The slit width was adjusted to 5 nm and the instrument was set to 750 V mode. Thermo-Fisher Scientific Nicolet-iS10 FT-IR spectrometer was used to obtain the FT-IR spectra (Waltham, MA, USA). It had a 4000-1000 cm −1 deuterated triglycine sulfate (DTGS) detector and a Ge/KBr beam splitter. The measurements were taken in 32 scans with a resolution of 4 cm −1 . A JEM-2100 high-resolution transmission electron microscope (HRTEM) (JEOL, Tokyo) working at 200 kV was used to investigate morphology of S,N-CQDs. pH-meter (Consort, NVP-901, Belgium) was also used.

Standard stock solutions
GLZ and SXG are relatively insoluble in water so their standard stock solutions (1.0 mM) were prepared in methanol and different concentrations were obtained by serial dilution with double distilled water as appropriate. The prepared solutions were stable for about 10 days when stored at 4°C.

Synthesis of S,N-CQDs
Synthesis of S,N-CQDs was performed by applying a hydrothermal methodology that was recently reported by Magdy et al. [26]. S,N-CQDs were prepared through mixing of 0.52 g CA and 0.68 g TS with 20 ml of double distilled water, and ultrasonication was carried out for 20 min. The mixture was refluxed at 160°C for 12 h until highly fluorescent S,N-CQDs (dark orange colour) were formed, then cooled and kept in the refrigerator for further use.  royalsocietypublishing.org/journal/rsos R. Soc. Open Sci. 9: 220285 2.6. Quantum yield measurements

Fluorescence emission spectroscopy
The following equation [40,41] was used to calculate the quantum yield of S,N-CQDs: where Φ is the quantum yield, F represents the integrated measured emission intensity, η is the solvent refractive index and A is the absorbance. The standard used was quinine sulfate (QS). It was dissolved in 0.1 M H 2 SO 4 (QY: 0.54 at 350 nm). In the aqueous solutions η S,N-CQDs /η st equals to 1.

Analysis of GLZ and SXG in their tablets
Ten tablets of each of Diamicron or Formigliptin were separately weighed and homogeneously ground. An accurately weighed quantity of the powder corresponding to 30 mg of GLZ or 5 mg of SXG was transferred into a measuring flask (100 ml), followed by addition of 40 ml of methanol. Sonication for 20 min, dilution with methanol to the mark then filtration were performed. Suitable aliquots were transferred from the filtrate into 10 ml measuring flasks, and then the procedure described in §2.5 was performed. The nominal content of tablets was calculated using the corresponding regression equation.

Characterization of S,N-CQDs
A facile approach was applied in this study to prepare highly fluorescent S,N-CQDs. The adopted approach is based on the hydrothermal treatment of TS as nitrogen and sulfur source and CA as a carbon source [26]. Under UV light, the S,N-CQDs solution exhibited strong blue fluorescence with a long-lasting homogeneous phase and no apparent precipitation for around 14 days in the refrigerator. Spectrofluorimetry, UV absorption spectroscopy, FT-IR and HRTEM were used to characterize S,N-CQDs. Figure 2A shows the UV absorption spectra of S,N-CQDs, CA and TS. S,N-CQDs had a clear UV absorption band at a maximum wavelength of 330 nm [26,42,43]. Figure 2B also shows the S,N-CQDs fluorescence emission and excitation spectra in aqueous solution. The optimum excitation and emission wavelengths were found to be 360 and 430 nm, respectively. When the excitation wavelength was changed from 340 to 380 nm, the fluorescence spectra of S,N-CQDs shift, and the highest fluorescence intensity was found at 360 nm (figure 2C).
The size and surface morphological characteristics of S,N-CQDs were studied by HRTEM. The samples were placed onto carbon-coated Cu-grid (200 mesh) and investigated using HRTEM at a voltage of 200 kV. As presented in figure 3a, S,N-CQDs are spherical in shape and range in size from 8 to 20 nm.
To study the main surface functional groups of S,N-CQDs, FT-IR analysis was used. As shown in figure 3b, several distinct vibrational modes were observed. The stretching modes of O-H/N-H groups are represented by broad bands at 3500-3100 cm −1 . C-N vibration is responsible for the band at 2065 cm −1 . At 1700 cm -1 , the C=O of the carboxylic acid group is displayed. C=S and C=C are responsible for the stretching maxima of 1233 and 1621 cm −1 , respectively. A vibration peak was also observed at 578 cm −1 for C-H bond [26,44,45].
The quantum yield of S,N-CQDs was also studied as mentioned in §2.6, and they showed a high quantum yield (58.5%) using QS as a reference.

Interaction mechanism between S,N-CQD and GLZ or SXG
S,N-CQDs emission fluorescence spectra in the presence of various GLZ and SXG concentrations are shown in figures 4 and 5, respectively. By increasing the concentrations of the studied drugs, S,N-CQDs fluorescence intensity was quantitatively enhanced. The native fluorescence of S,N-CQDs was enhanced by 37% and 70% by addition of GLZ (500.0 µM) and SXG (600.0 µM), respectively. The fluorescence enhancement may be attributed to the interaction of each of GLZ and SXG with S,N-CQDs. However, GLZ and SXG cannot form any aggregates with S,N-CQDs, otherwise, the royalsocietypublishing.org/journal/rsos R. Soc. Open Sci. 9: 220285 interaction of GLZ or SXG with S,N-CQDs would lead to decrease of surface defects [46][47][48]. Moreover, the enhancement of fluorescence intensities may be due to saturation of the dangling bonds at the surface of QDs which effectively removes the QDs surface defects. Due to efficient blocking of non-radiative electron/hole recombination on the surface of QDs, the removal of local trap positions leads to formation of more radiative centres. It seems that the increasing concentrations of GLZ and SXG in the solution get attached to the QDs surface and thereby correct the defective energy levels. This removal of defect levels improves the exciton emissions and subsequently enhances the fluorescence emission intensities of the QDs [46][47][48]. royalsocietypublishing.org/journal/rsos R. Soc. Open Sci. 9: 220285

Inner filter effect of GLZ and SXG
The inner filter effect of the studied drugs was carefully studied to confirm that the enhancement of the fluorescence intensity of S,N-CQDs is due to the interaction between QDs and drugs and not due to the native fluorescence of the studied drugs. It was found that 500.0 µM of GLZ and 600.0 µM of SXG

Optimization of factors affecting interaction of GLZ and SXG with S,N-CQDs
To achieve the maximum sensitivity of the method for GLZ and SXG determination, different factors influencing the fluorescence intensities were studied; including pH of the medium, incubation time and temperature.  royalsocietypublishing.org/journal/rsos R. Soc. Open Sci. 9: 220285

Effect of pH
To investigate the impact of pH, the experiments were carried out over pH range from 3.5 to 12 using acetate and borate buffers. It was found that the maximum S,N-CQDs emission intensities were achieved at pH 7 and 11 for GLZ and SXG, respectively (figure 7A). Accordingly, the volume of borate buffer was examined from 0.5 to 4 ml, and it was found that the optimum volume is 1 ml for both drugs.

Incubation time
The incubation time of S,N-CQDs with the investigated drugs was studied by recording the emission fluorescence spectra at time intervals ranging from 1 min to 1 h. The obtained outcomes displayed a fast increasing response after mixing the cited drugs and QDs that reached a constant value after 10 min. The fluorescence signals remained stable for 1 h, giving the suggested method an additional advantage ( figure 7B).

Effect of temperature
Over a temperature range of 20-60°C, the influence of temperature on signal enhancing effect of GLZ and SXG was studied. The obtained results demonstrated that the maximum response was achieved at 40°C and 25°C for GLZ and SXG, respectively (figure 7C ).

Method validation
The developed method was validated according to ICHQ2(R1) guidelines [49]. where F 0 and F are the QDs fluorescence intensities in absence and presence of cited drugs, respectively, C is the concentration in μM. Values of correlation coefficients (r) approximating unity indicate the acceptable linearity of the developed procedure (table 1).

LOD and LOQ
The following equations were used to calculate limit of quantitation (LOQ) and limit of detection (LOD) values: LOD = 3.3 S a /b, LOQ = 10 S a /b, where S a is the standard deviation of y-intercept, b is the slope. The obtained results are abridged in table 1.

Precision and accuracy
Three different concentrations of the cited drugs and three replicates of each concentration were used to test the intra-day and inter-day precisions. The obtained results showed small % relative standard deviation (RSD) values (less than 2%), indicating an acceptable precision of the developed approach (table 2). The accuracy and precision were demonstrated by statistical comparison of the obtained results with those given by comparison methods for both drugs [2,23], showing insignificant difference between them as presented by t and F values, respectively [50] (table 3).

Selectivity
The proposed method's selectivity was proved by its capacity to estimate the cited drugs in presence of other antidiabetic drugs including metformin, dapagliflozin, empagliflozin, canagliflozin, alogliptin, omarigliptin and glebenclamide. The tolerance limits of these drugs were calculated as the concentration that results in 2% relative error [51] and the obtained results are summarized in table 5. In addition, the method selectivity was confirmed by its capacity to analyse the drugs in their tablets with low % RSD values (less than 2%) and high % recoveries (98.64-101.21%); demonstrating no interference from common excipients (table 6).

Analysis of GLZ and SXG in their tablets
Due to the high selectivity and reproducibility of the developed method, it was efficiently used to estimate the cited drugs in their commercial tablets with low % RSD values and high % recoveries. Statistical analysis of obtained results with those given by comparison methods for GLZ and SXG [2,23], showed insignificant difference between them regarding precision and accuracy as indicated by the Fand t-values, respectively [50] (table 6). The values between parentheses are the tabulated t-and F-values at p = 0.05 [50].

Conclusion
In this study, a facile hydrothermal approach was used for synthesis of S,N-CQDs using CA and TS. The prepared S,N-CQDs were characterized using various techniques. S,N-CQDs were used as fluorescent nanosensors for estimation of each of GLZ and SXG depending on their enhancement effect on S,N-CQDs fluorescence intensities. The developed method showed good selectivity for estimation of GLZ and SXG in their tablets and in presence of other antidiabetic drugs. The proposed method is rapid, simple and cost-effective without the need for sophisticated instruments or prior derivatization of the studied drugs. Full validation of the proposed method was performed in accordance with ICH recommendations. All authors gave final approval for publication and agreed to be held accountable for the work performed therein.