A simple fluorescent assay for cyromazine detection in raw milk by using CYR-stabilized G-quadruplex formation

A rapid biosensor for the detection of cyromazine in milk is reported based on a fluorescence quenching result. When an FAM labelled G-rich ssDNA Tcy2 is treated with cyromazine, it can form a G-quadruplex-CYR complex and cause a change in fluorescence. As a result, the presence of cyromazine can be determined by fluorescence quenching. This sensor is selective for the detection of cyromazine in raw milk and has a limit of detection of 0.68 ppb and a detection range from 0 to 200 ppb.


Introduction
Cyromazine (N-cyclopropyl-1,3,5-triazine-2,4,6-triamine, CYR) is a triazine pesticide, which is widely used in the livestock and poultry industry as an insect growth inhibitor for y and maggot control. 1 The overuse of cyromazine has been proved to cause problems of pollution by animal-derived food, and potential problems for environmental and human health, for example, mammary tumors in mice. 2,3 Cyromazine can also be degraded to melamine, an industrial chemical compound which has been illegally added to food, animal feed, and even milk to increase the apparent protein content of products due to its high nitrogen content (66% by mass), and which causes kidney stones and kidney failure in high doses. 4,5 Nowadays, the residues of cyromazine and melamine must be determined according to national standards before the derived food can be put on sale, so the eld testing of cyromazine residues in animal-derived food, even raw milk, has become a hot issue. The U.S. FDA and China have set maximum residue limits (MRLs) for cyromazine in animal-derived food at 0.5 ppm, in contrast with the CAC limit of 1 ppm. Therefore, the development of an accurate, rapid and reliable method for the determination of cyromazine in raw milk is required to ensure food safety.
Nowadays, the main methods for the detection of cyromazine in raw milk and dairy products are as follows: gas chromatography-mass spectrometry (GC-MS) with the limits of quantication (LOQs) ranging from 10 to 100 ppb [6][7][8] and liquid chromatography-mass spectrometry (LC-MS), [9][10][11][12][13][14][15] ultra high performance liquid chromatography-high resolution mass spectrometry (UHPLC-MS/MS), and molecularly imprinted solid-phase extraction-ultra-performance liquid chromatography (MISPE-UPLC) with LOQs ranging from 0.05 ppm to 40 ppm. 16,17 In the national standard of China (GB29704-2013), the residues of cyromazine and melamine are determinated by an ultra performance liquid chromatography-tandem mass spectrometric method (UPLC-MS). Although these methods are highly sensitive, most are either time consuming due to extensive pretreatment, or have a high cost due to the need for expensive instruments. Additionally, some new methods are also being developed for cyromazine and melamine monitoring, such as surface enhanced Raman spectroscopy (SERS), uorescence, spectrophotometric absorption and chemiluminescence. [18][19][20][21] Most of the rapid methods have been reported for melamine detection, [22][23][24][25] and only a few are especially for cyromazine detection, such as the visual colorimetric detection of cyromazine in river water using gold nanoparticles. 21 But there are no clear distinctions between cyromazine and its analogs. Therefore, simple, rapid, low-cost, easily operable methods are required for detecting cyromazine in milk and dairy products with good sensitivity and reliability.
Aptamers are articial nucleic acid ligands that can bind their targets with high affinity and specicity, and they have therefore been widely used as recognition elements in the construction of biosensors. 26 Hydrogen bonding between thymine and amino groups has been reported. With this hydrogen bonding, we have designed an aptamer-modied nanogold probe for melamine colorimetric detection using an ssDNA with 31 T bases, in which melamine can form a melamine-aptamer complex via hydrogen bonding. Thus, the resulting cationic polymer can aggregate the AuNPs and cause a remarkable change in color. 27 A label-free AuNP based visual detection method for cyromazine in cucumbers using an ssDNA with ten T bases has also been reported. 28 But these methods for the detection of cyromazine would be disturbed by analogs of cyromazine, such as melamine. Linear G-rich telomeric DNA strands can fold into G-quadruplex structures in the presence of monovalent cations such as K + or Na + . G-quadruplex structures consist of two or more stacked planar G-tetrads. [29][30][31] Inspired and encouraged by the similar structure and difference between cyromazine and melamine, we reported a sensitive detection system based on uorescence quenching. When an FAM-labelled G-rich ssDNA Tcy2 was treated with cyromazine, it could form a G-quadruplex-CYR complex. Both the structure of the G-quadruplex formed by cyromazine and the occurrence of FRET between FAM and cyromazine can lead to a decrease in uorescence, making the uorescence quenching rate higher.
In this paper, an aptamer with 18 bases is designed to specically combine with cyromazine, forming a G-quadruplex-CYR complex which results in a uorescence quenching phenomenon, and a simple, sensitive and selective assay for cyromazine detection is proposed based on it.

Materials and apparatus
We designed an ssDNA that is supposed to bind with cyromazine to form a G-quadruplex, and a random sequence was also used as a negative control. All the ssDNAs were labeled with 6-carboxyuorescein (FAM) at the 5 0 end and synthesized by Sangon Biotechnology Co, Ltd. (Shanghai, China), and the sequences are as follows: Tcy2 5 0 -FAM-GGTTGGTTGGTTGGTTTT-3 0 (18 bp) Apt1 5 0 -FAM-TTTTTTTTTTTTTTTTTTTTTTTTTTTTTTT-3 0 (31 bp) Apt2 5 0 -FAM-GGGTAGGGCGGGTTGGG-3 0 (17 bp) Random DNA 5 0 -FAM-ATCGACATGTAGCCGATGGC-3 0 (20 bp) An F-4500 uorescence spectrophotometer (Hitachi, Japan) was used to record the uorescence intensity, with a response time of 0.5 s, PMT voltage of 700 V, scan speed of 1200 nm min À1 , excitation wavelength of 480 nm, and excitation and emission slits of 10 nm. A time scan was operated when studying the kinetics of uorescence quenching, with a scan time of 1200 s, excitation wavelength of 480 nm and emission wavelength of 520 nm.
A J-815 CD spectrometer (Jasco, Japan) was employed to characterize the structural changes in the oligonucleotides. The optical chamber (1 cm path length, 1 mL volume) was deoxygenated with dry puried nitrogen (99.99%) before use and kept under nitrogen atmosphere during the experiments. The scans (100 nm min À1 ) from 200 to 320 nm were taken three times at 1 nm intervals, then accumulated and averaged. The background of the buffer solution was subtracted from the CD data.
A thermostatic incubating device (Eppendorf, China) was used to carry out the quenching experiments at various temperatures.
The raw milk was purchased from a local nearby cattle farm. All reagents were of analytical grade. Milli-Q water (18 MV cm) was used in all experiments.

Pretreatment of samples
5 mL of 1% acetic acid was added into a 1.0 mL raw milk sample in a 10 mL centrifuge tube, then incubated for 5 min at room temperature aer mixing well. The sample was centrifuged at 10 000 rpm for 10 min to separate the liquid component from the white opaque precipitation. In the following steps the nal solution volumes of the samples were adjusted with ultrapure water.

Selection of cyromazine-binding ssDNA
Different ssDNAs with the same nal concentration (25 nM) were individually dissolved in Tris-acetate buffer (10 mM, pH 8.0), and then various concentrations of cyromazine were added. A blank sample for each ssDNA was carried out in the absence of cyromazine. This step was done in order to calculate the F 0 . Aer incubation for 2 min at room temperature, the uorescence intensity of each sample was measured and the quenching ratio, (F 0 À F)/F 0 , was calculated, where F 0 stands for the uorescence intensity of FAM in the absence of cyromazine and F for the uorescence intensity of FAM aer the addition of cyromazine. The ssDNA which responded with the largest quenching ratio was chosen for subsequent study.

The kinetic curves of Tcy2 with cyromazine or K +
First the uorescence intensity of each sample containing 25 nM Tcy2 was measured and then various concentrations of cyromazine or K + were added individually. The uorescence intensity of each sample was re-measured at 0, 10, 20, 30, 60, 120, 180, 240, 300, 360 seconds.

Investigation of the quenching mechanism
Quenching experiments at various temperatures (25 C, 35 C, 45 C and 55 C) and different cyromazine concentrations (0 ppb, 6 ppb, 15 ppb, 40 ppb, 100 ppb and 200 ppb) were carried out in Tris-acetate buffer (10 mM, pH 8.0) for 2 min and the uorescence intensities in the absence and presence of cyromazine were recorded as F 0 and F, respectively. A Stern--Volmer plot was generated by plotting F 0 /F against cyromazine concentration.
2.6 Sensitivity and selectivity of the detection of cyromazine 2.5 mL of 5 mM Tcy2 was rstly added into an appropriate volume of Tris-acetate buffer (10 mM, pH 8.0), and then various concentrations (from 0.5 to 200 ppb) of cyromazine were introduced into the above solution. The total volume of the nal solution was xed at 500 mL. Aer incubation for 2 min at room temperature, the uorescence intensity was measured.
To determine the selectivity of the uorescence assay, different veterinary drugs, including levamisole, abamectin, clopidol, amitraz, chloromycetin, thiamphenicol and terramycin, and different analogs of cyromazine and amino acids, such as ammonium hydroxide (NH 3 $H 2 O), urea, melamine, ammeline, ammelide, cyanuric acid, IgG, L-tyrosine, L-lysine, individually added to the sensor solution and the change in the uorescence intensity was monitored.

Principle of cyromazine biosensing using CYR-stabilized G-quadruplex formation
The single-strand aptamer (ssDNA) of cyromazine, labelled with 6-carboxyuorescein (FAM) at the 5 0 end, is referred to as Tcy2. In the absence of cyromazine, ssDNA Tcy2 stays in its random coil conformation and the FAM shows strong uorescence intensity. Whereas in the presence of cyromazine, a signicant decrease in uorescence intensity occurred. This uorescence quenching phenomenon could be illustrated as follows (Scheme 1). In the absence of cyromazine, uorescence of the labelled FAM could be excited by the exciting light, thus exhibiting strong uorescence intensity. When the ssDNA was treated with other veterinary drugs that could not bind with it, it could not fold into the G-quadruplex. Under these conditions, the probability of intermolecular collisions between the uorophore FAM and the drug would be considerably increased if the concentration of the drug was high enough. 32 But, as Forster showed, the rate of uorescence resonance energy transfer (FRET) depends on the inverse sixth power of the distance between the two uorophores (FAM and drug), 33 so the occurrence of FRET could not be guaranteed just by random collisions 34 without the formation of a G-quadruplex. When the ssDNA was treated with other ions that have the ability to induce G-rich oligonucleotides to form a G-quadruplex, such as K + , the ssDNA could fold into the G-quadruplex which could bind some molecules with conjugated aromatic and potentially positively charged moieties. Compared with cyromazine, these molecules were implanted into the stacked G-quartets, not combined with the ssDNA. According to Kotch's report, 35 the K(I)-stabilized G-quadruplex has a K-O distance of 2.80 A, an O-O distance of 4.58 A and a vertical separation of the Gquartets of 3.31 A. In the present study, FAM was xed at the 5 0 end, and the FAM-K distance is approximately the same as that of K-O. Thus, based on the predicted secondary structure of the G-quadruplex formed from Tcy2, the FAM-CYR distance Scheme 1 Schematic representation of the fluorescence assay for cyromazine detection based on the G-quadruplex combined with cyromazine.
is much less than that of K-O. On the basis of Forster's theory, 33 the longer distance could result in a dramatically lower FRET rate than that of cyromazine, hence ensuring the specicity of the cyromazine led uorescence quenching. However, if the added ion was cyromazine, more than one molecule of cyromazine is in range so that the acceptor could get a reasonable energy transfer signal from the donor, 36,37 thus making it possible to bring about the occurrence of FRET which results in uorescence quenching.
For the purpose of obtaining the most sensitive response and gaining a deeper insight into the mechanism, several ssDNAs (Apt1, Apt2) are supposed to bind with cyromazine to form a G-quadruplex according to previous reports 38 and a randomly designed sequence (random DNA) was studied. As shown in Fig. 1, no obvious quenching effect was obtained when the FAM-labelled Random DNA was treated with cyromazine at low or high concentrations. This result indicates that the quenching phenomenon appearing in our study was actually aroused by some reaction between cyromazine and the CYR-binding ssDNA, thus disproving the hypothesis that the cyromazine may cause the uorescence quenching directly.

Interactions between the ssDNA and cyromazine
Circular dichroism (CD) is a commonly used method for the analysis of conformational change in nucleic acid aptamer reactions. Thus, CD provides a way to conrm that some specic structure was formed when cyromazine reacted with the ssDNA Tcy2. The results showed that an enhancement in the negative peak around 240 nm and the positive peak around 260 nm was observed upon the addition of 1 ppm of cyromazine (Fig. 2). As it has been reported that a typical CD spectrum of a "parallel" G-quadruplex structure has a negative peak at around 240 nm and a positive peak near 260 nm, 39-41 the results indicated that a parallel G-quadruplex was formed with the presence of cyromazine.

The kinetic curves of Tcy2 with cyromazine or K +
Time plays an important role in the cyromazine-oligonucleotide reaction, so the uorescence quenching of this sensor system was tested at various reaction times. As shown in Fig. 3, all the FAM-labelled ssDNAs show uorescence intensity at the same level before being treated with molecules. Aer various concentrations of cyromazine were added, different uorescence intensity changes occurred, while K + could not bring about notable changes for a long time. It could be observed that cyromazine caused obvious changes quickly even at low concentration, and the higher the concentration of cyromazine, the more obvious the uorescence quenching. As all the uorescence intensities tended to remain constant aer 120 s, 2 min or a little longer was consequently chosen as the optimum reaction time. Although K + might induce the ssDNA to form a G-quadruplex structure, the FRET between K + and FAM   Paper was much lower than that between cyromazine and FAM, and it hardly changed as time went on.

Investigation of the quenching mechanism
The well-known Stern-Volmer equation, F 0 /F ¼ 1 + K sv [Q], was applied to investigate the quenching mechanism and to determine whether it is a dynamic or static process, where F 0 and F are the uorescence intensities in the absence and presence of cyromazine, respectively, K sv is the Stern-Volmer quenching constant and [Q] is the quencher (i.e. cyromazine) concentration, 42 and the temperature dependence of the quenching was examined. 43 Quenching experiments at various temperatures and different cyromazine concentrations were carried out and a Stern-Volmer plot was generated (Fig. 4). The results show that the value of F 0 /F at each concentration of cyromazine decreased with an increase in temperature and the curvature went down at high cyromazine concentration. So from the results two conclusions could be inferred. First, when the temperature increased, the quenching constant K sv decreased, so the quenching in our study is static quenching. 44 Second, the downward curvature reveals the inaccessibility of the uorophore fraction to the quencher at high cyromazine concentration. This veries our conjecture that the uorescence quenching is primarily caused by the formation of a CYRstabilized G-quadruplex. When the cyromazine concentrations increased to a high level, the excess cyromazine could not induce any more Tcy2 to form a G-quadruplex because the amount of G-quadruplex was xed by the concentration of Tcy2.

Sensitivity and selectivity of the detection of cyromazine
The uorescence signals for different concentrations of cyromazine were measured to explore the sensitivity of this assay. 2517. This sensor has a limit of detection (LOD) of 0.68 ppb, which was calculated from 3 s per slope. 45,46 As the maximum contamination level for cyromazine in raw milk, as dened by the U.S. FDA and China, is 50 ppb or 300 nM, the assay has high sensitivity for the quantitative analysis of cyromazine in raw milk.
The selectivity of this assay for the determination of cyromazine was also investigated. Control experiments were performed using other common veterinary drugs, such as levamisole, abamectin, clopidol, amitraz, chloromycetin, thiamphenicol and terramycin. Eleven competitive compounds with a similar structure or which occur in raw milk, ammonium hydroxide (NH 3 $H 2 O), urea, melamine, ammeline, ammelide, cyanuric acid, IgG, L-tyrosine, L-lysine, L-phenylalanine and L-valine, were also individually added to the sensor. As shown in Fig. 6, we can see that these nonspecic veterinary drugs did not lead to characteristic changes in uorescence even at high   concentrations. Fig. 7 shows that there were no evident changes in uorescence for the mixture of the sensor solution with the eleven compounds, except with cyromazine. Although there is also hydrogen bonding between thymine and the amino group of some of the compounds above, they would not form the same G-quadruplex and cause a remarkable change in uorescence. These results demonstrate that all the other compounds display slight or negligible interference with the detection of cyromazine by this biosensor.

Detection of cyromazine in raw milk samples
Before sample pretreatment, we added different amounts of cyromazine to the raw milk. This proposed method was repeated 3 times for each sample and HPLC-MS was also applied to analyse the cyromazine in the raw milk samples. As is shown in Table 1, the recoveries are between 90% and 120% by this method, while the recoveries are between 98% and 102% by HPLC-MS. This indicates that this method of using a CYRstabilized G-quadruplex as a probe for the rapid detection of cyromazine in raw milk is feasible.

Conclusions
In summary, we have developed a rapid, sensitive and selective uorescence assay for the detection of cyromazine in raw milk by using an FAM-labelled G-rich oligonucleotide Tcy2 as a recognition probe. This assay is based on the formation of a CYR-stabilized G-quadruplex and the uorescence resonance energy transfer (FRET) between cyromazine and FAM. In the absence of cyromazine, ssDNA Tcy2 stays in its random coil conformation, accompanied with strong uorescence intensity, but once cyromazine was added into this system, Tcy2 formed into a G-quadruplex structure and made the FAM closed to cyromazine, enabling the FRET between them, leading to a remarkable uorescence quenching phenomenon. The LOD of this method for the detection of cyromazine in raw milk is as low as 0.68 ppb. Moreover, compared with traditional cyromazine sensors, this assay is facile and convenient without involving expensive sophisticated instruments and long response times. So we hope that this type of detection method, which holds great practicality for real-time and on-site cyromazine detection in environmental monitoring, will be realized for the detection of other molecules in food samples.