HPTLC-Densitometry Screening And Mass Identification Of Fluorescent Whitening Agents Contamination in Cereal Flour

Contact: Audrey Hu Whatsapp/hp: 0086 13880143964 Email: audrey.hu@wecistanche.com

Yisheng Chen1,2 & Caihong Huang1,2 & Xueming Xu

Abstract

In recent years, there was increasing concerns on food safety issues associated with fluorescent whitening agents from packing materials. Here, we reported the development of a facile and reliable method suitable for high-throughput quantification and identification of two common fluorescent whitening agents (FWA 184 and FWA 367) in cereal products (wheat and rice flour), platformed by high-performance thin-layer chromatography. First, the sample preparation and cleanup were rapidly accomplished via an optimized solid-liquid extraction. The extracts and reference standard were simultaneously separated on silica gel plates, using the mixture of toluene and ethyl acetate (10/0.3, v/v) as the mobile phase. Then, rapid quantification was performed with densitometry in fluorescent mode (365 </ 400 nm, mercury lamp), reaching a limit of detection of 18–21 μg/kg; meanwhile, good linearity (R2 = 0.9999) of quantification can be achieved within a broad range (100–2000 pg/band). With real cereal samples, the quantitative measurement was validated, showing good spike-recovery rates (88.0–108.4%). Furthermore, the conclusive identification of the targeted band was directly realized by in situ mass detection, providing compound-specific signals (m/z 363.09 and 431.18, for FWA 184 and FWA 369, respectively) as the fingerprint evidence. The proof-of-concept result of this work demonstrated that HPTLC as a versatile analytical platform can achieve an ideal balance among throughput, simplicity, and detectability, therefore, particularly suitable for screen-oriented food analysis.

Keywords HPTLC. Densitometry. Fluorescent whitening agents

Cistanche has a whitening effect on the skin

Introduction

Recent years had witnessed the booming of food package materials. On the other hand, food safety issues associated with package material attracted worldwide attention as well. Among hazardous chemicals possibly originating from these materials, fluorescent whitening agents (FWAs) gained particularly high interest. Molecules of FWAs were capable of absorbing radiation and simultaneously emitting visible lights. Such transformation can offset the yellow color of substances, resulting in intensified white impression. As such, materials incorporated with FWAs were intensively considered for food packages. Especially for cereal products, such a package was favorably used to increase the attractiveness and acceptance of consumers. For instance, abnormally high levels of FWAshad been found in the paper board of popcorn and instant noodles (Jiang et al. 2015). As the result, FWAs migrated into the cereal food and resulted in contamination, posing an emerging challenge for public health.

Decades ago, the remarkably high stability of FWAs had provoked intensive concerns about their potential toxicity. Though the real impact of FWAs on human health was still not conclusive yet, their developmental toxicity and alteration effect of gene expression in model animals had been experimentally evidenced (Belliveau et al. 1990; Jung et al. 2012). With regard to these potential risks, food safety authorities around the world had stipulated restrictions and limits for migratable FWAs in food contact materials.

With the aim to implement these legislative standards, a large array of analytical methods, majorly based UVfluorescence inspection or column chromatography (HPLC and Capillary electrophoresis), had been proposed for the detection of FWAs (Jiang et al. 2015; Wu et al. 2018). But from the view of practice, all these existing methods were not adapted to the principle of cost-effective and reliable screening. For instance, the UV-fluorescence inspection method greatly suffered from poor selectivity. Meanwhile, analysis based on column chromatographic systems displayed a remarkable shortage in simplicity and cost-efficiency. On the contrary, high-performance thin-layer chromatography (HPTLC) was opening a new horizon of food screening methodology. Lots of screening-oriented methods based on HPTLC had been developed and applied in the analysis of various food components, showing unparalleled flexibility while ideally filling up the gap between the aspect of simplicity, specificity, and reliability (Galarce-Bustos et al. 2019; Li et al. 2018; Mikropoulou et al. 2019; Pedan et al. 2018; Premarathne et al. 2018; Stanek et al. 2019; Sun et al. 2018). In this way, many samples (up to 40) can be analyzed simultaneously with almost unlimited matrix tolerance, because each of the plates was disposable (Li et al. 2019). This implied that the demand for sample cleanup and instrument maintenance can be remarkably saved. As enforcement to visual inspection, densitometry enabled precise and accurate measuring of the separation results on HPTLC plate (Lebot et al. 2020). With the fast development of hyphenation technologies, the most featuring merit of HPTLC was that it can be used as a platform efficiently assembling the usability of powerful analytical tools, like surface-enhanced Raman spectroscopy (Kang et al. 2019; Qu et al. 2018; Wang et al. 2018a, b), biosensor (AgatonovicKustrin et al. 2020; Chen et al. 2020a, b; Choma and Grzelak 2011; Galarce-Bustos et al. 2019), image/chemometrics analysis (Rejšek et al. 2016; Xu and Liu 2021; Xu et al. 2019), and mass detection (Wang et al. 2018a, b). As such, HPTLC had been a versatile toolbox particularly suitable for screening tasks.

Against these backgrounds, the objective of this work was to develop a facile and sensitive screening method for FWAs in different cereal flour samples, based on the combination of HPTLC, fluorescent densitometry (FLD), and mass spectrometry (MS). After method development, its performance was validated both quantitatively and qualitatively, substantially evidencing its real superiority as a facile tool for screening.

Materials and Methods

Materials

Standards of FWA184 (2,5-bis(5′-tert-butyl-2-benzoxazolyl) thiophene, CAS 7128-64-5, > 99% purity) and FWA367 (1,4-bis(2-benzoxazolyl) napthalene, CAS 5089-22-5, > 98% purity) were purchased from Aladdin (Shanghai, China). Organic solvents and chemicals of analytical grade were from Sigma Aldrich (Shanghai, China). Silica gel plates 60 F254 (analytical grade, layer thickness 0.1 mm, 10 × 20 cm size, serial No. 1.05729.0001) were from Merck (Darmstadt, Germany). Six blank cereal samples including three wheat flour and three rice flour samples were purchased in the local supermarket. Apart from that, samples of wheat and rice flour with confirmed FWA contamination were provided by Shenzhen Customs Food Center.

cistanche tubulosa extract

Standard Solutions

Stock solutions (0.01 mg/mL) of FWA 184 and FWA 367 were prepared in ethyl acetate, separately or in the mixture, and stored in a refrigerator (4 °C). The working solutions for establishing calibration curves were prepared by further diluting stock solutions to 0.0001 mg/mL.

Sample Preparation

For extracting analytes from cereal samples, ethyl acetate, acetone, and their mixtures (mixing ratio, v/v, 10/0, 6/4, 4/6, and 0/10) were tested as the extraction solvent. Briefly, a 5 g cereal flour sample was mixed with 20 mL extraction solvent; if necessary, 100 or 50 μL stock solutions (0.01 mg/mL) of FWA standards were spiked into the extraction mixture of blank samples, resulting in 200 or 100 μg/kg artificial contamination. Following treatment of 30 °C ultrasonic bath for 5 min, the suspension was separated by centrifugation for 5 min at 4 °C. After that, 1 mL supernatant was carefully pipetted into a 5 mL syringe and filtered through a 0.45-μm nylon membrane to remove solid particles, to avoid needle blockage.

HPTLC Steps

Sample extracts and reference standards were sprayed as 6- mm bands on HPTLC plates by a Linomat 5 semiautomatic sampler (CAMAG, Switzerland) equipped with a 100-μL syringe, at the delivery speed at 100 nL/s and 0.2 μL pre-dosage. The band array was initially 15 mm from the left edge and 10 mm from the bottom, with the automatically calculated interval from each other. The application volumes were 10 μL for blank and spiked sample extracts; for establishing the calibration curve, 1, 5, 10, 15, or 20 μL working solution (0.0001 mg/mL) was applied, resulting in gradient concentrations at 100, 500, 1000, 1500, and 2000 pg/band, respectively. Between each sampling, the syringe was manually rinsed with pure ethyl acetate twice between each application to prevent cross-contamination. After application, the plated was heated by a TLC-heater III (CAMAG, Switzerland) at 60 °C for 2 min to remove solvent residue in application bands. Chromatography was automatically performed with ADC-2 (CAMAG, Switzerland), in order to realize reproducible separation. First, both trenches of the developing chamber were filled with stationary phase consisting of 10 mL toluene and 0.3 mL ethyl acetate. Before dipping the plate bottom into the mobile phase, there was 5 min dryness control by bubbling through saturated MgCl2 aqueous solution to a final relative humidity at 33%, 10 min tank saturation, and 10 min plate precondition. The migration distance of the solvent front line was fixed at 50 mm.

Plate Documentation and Densitometry

After development, digital pictures of the developed plate were documented by a DD70 imaging system (Biostep, Germany) integrated with a Sony EOS700D digital camera lighted by 366 nm UV lamps. Then, the separation results were densitometrically evaluated by a TLC scanner 3 (CAMAG, Switzerland). For locating the working wavelength of densitometry, the excitation spectra of the analyte deposited on the silica gel plate were continuously profiled from 220–400 nm using the D2 & W lamp, with K400 optical filter. Quantitative scanning was performed with general settings: fluorescent mode, mercury lamp, excitation wavelength 365 nm, K400 optical filter, 3.00 × 0.30 mm micro-slit dimension, scanning speed 100 μm/s, and data resolution 100 μm/step. The instrumental operation and data processing were controlled by the winCATS software version 1.4.4.

benefit of cistanche tubulosa extract

HPTLC-MS

Following FLD measurements, the positive bands were further identified by their MS fingerprints, mediated by a TLC-MS interface (CAMAG). Driven by a quaternary pump, the targeted bands visualized under 366 nm illumination was eluted with a flow of acetonitrile containing 0.1% formic acid, at the rate of 0.2 mL/min for 60 s. Directly, the eluent was injected into an electrospray ionization source and simultaneously analyzed by a triple quadrupole mass spectrometer (Quattro Premier XE, Waters). Full-scan MS data acquisition was carried out in ESI+ mode with the following settings: the capillary voltages are 3.5 kV, cone voltages are 50 V, ion source temperature 100 °C, desolvation temperature 400 °C, desolvation gas flow rate 700 L/h, and cone gas flow rate 50 L/h. The spectra were recorded in the ranges of 50–1000 m/z.

Results and Discussion

Chromatography Optimization

Concerning the low polarity of the targeted compounds, the optimization of the mobile phase was started with trial tests with different hydrophobic solvents (including acetone, toluene, hexane, and petroleum ether) and their mixtures. From the initial screening, toluene alone was found to give enough resolution of the targeted compounds from interfering sample matrices that mostly remained at the lower part of the track. To further increase the resolution of analytes, different amounts (1–10 %) of ethyl acetate were added. From comparison, the attributed to the fragile equilibrium between the mobile phase (liquid), chamber atmosphere (gas), and stationary phase (solid). In another word, the phase equilibrium would be broken at once when the plate was moved manually. This strongly evidenced the importance of instrumentation for the repeatability of HPTLC analysis. In addition, the applicability of the optimized separation conditions was further assessed by analyzing artificially contaminated cereal samples, From Fig. 1a, it can be concluded that the presence of sample matrices did not lead to significant variation of chromatographic results. Therefore, toluene/ethyl acetate (10/0.3, v/v) was fixed as the mobile phase throughout this work.

Visual Screening and FLD Quantification

Image acquisition of the whole plate was one of the featuring advantages of HPTLC, with which a semi-quantitative profile of separation can be easily obtained by visual inspection. Here, the intensive native fluorescence of analytes allowed the direct reading of the chromatographic results under 366 nm irradiation, with sensitivity down to 100 pg/band. Thereafter, only horizontal comparison between tracks could lead to a straightforward estimation of the sample contamination degree, which was highly favored in screening tasks. Further, densitometry in the fluorescence mode was employed to perform the scanning of tracks. This step was also very efficient, which can be finished in 2–3 min. With the intention to identify optimal excitation wavelength for densitometry, the excitation spectrum (220–400 nm) of analyte deposited on the silica gel layer was investigated. As presented in Fig. 1b, spectra profiles of analyte in different conditions showed high similarity, exhibiting the most intensive emission at 365 nm excitation. Therefore, the light of the mercury lamp at 365 nm in combination with a 400-nm edge filter (K400) was used. As shown in Fig. 1d, the applicability of these optical parameters is evidenced by clear signal peaks of analyte bands even in the presence of co-extracted sample matrices.

Optimization and Validation of Quantitation

Linearity and Sensitivity

Quantitative capacity of the established densitometric detection was first validated in terms of linearity and sensitivity. To do so, a calibration graph based on 5 levels within the concentrations from 100 to 2000 pg/band was established. As summarized in Table 1, results obtained from HPTLC-FLD showed good linearity with satisfactory coefficients of correlation (R2 = 0.9999). Based on the calibration curve, method sensitivity including limit of detection (LOD) and quantitation (LOQ) were calculated according to the DIN 32645 method with 95% statistic confidence (Deutsches Institut für Normung 2013). From the calculation, the LOD expressed as pg/band was 45 and 52 for FWA 184 and FWA 367, respectively. Fixing the application volume at 10 μL, such detectability was equal to 18 and 21 μg/kg, respectively, which were about 30-folds below the legislative tolerance limit (600 μg/kg) stipulated by EU (European Union 2002).

Recovery and Precision

Before assessing the accuracy of the developed HPTLC-FLD by determining the recovery rates of analytes spiked into cereal samples, a proper extraction method should be optimized. For FWA extraction in cereal flour, Wu et al. suggest using mixture of dichloromethane and acetone (3/2, v/v) (Wu et al. 2018). Concerning the environmental and ccarcinogenic toxicity of dichloromethane, mixtures of ethyl acetate and acetone were tested alternatively as the extraction solvent here. As exemplarily presented in Fig. 2, it was obvious that pure ethyl acetate was the best choice for extracting the targeted compound from both cereal samples, comparable to those by the method of Wu et al. (2018). To guarantee its applicability, the optimized extraction method was extended to 6 cereal samples including 3 wheat flour and 3 rice flour were further tested for the spike-recovery experiment. As summarized in Table 2, the calculated recovery rates were within the range of 88.7–108.4%. Meanwhile, reproducibility of the quantitation in terms of relative standard deviation (%RSD) of triplicates was found to be less than 10.5%. Moreover, it was noteworthy that the obtained data was insignificantly dependent on used samples and spiking levels, demonstrating that the developed quantitative analysis could be an efficient screening tool for FWAs in cereal samples, with acceptable accuracy and precision.

Determination FWAs in Packaged Cereal Samples

To further evaluate the suitability of the optimized extraction conditions and HPTLC-FLD detection, the method was exemplarily applied to quantify two cereal samples of which the package contains FWAs. Simply from visual inspection on the plate image shown in Fig. 3a, it was very easy to reach the conclusion that both cereal samples had been contaminated with FWA residues. Then, quantification was carried out by FLD scanning. As displayed in Fig. 3b, good selectivity of the detection can be observed in the obtained dendrogram. As summarized in Table 3, the residues of analytes in both cereal products were within the μg/kg level (maximally 262.8 μg/kg), which were significantly below their specific migration limits (600 μg/kg). Even so, the potential negative impacts of such contamination remained alarming as well, since the adverse effects caused by chronic exposure were not sufficient yet.

benefit of cistanche tubulosa extract

Molecular Identification by MS Fingerprints

To achieve chemical identification of the bands resulting from HPTLC development, in situ MS detection was incorporated in the screening approach as a confirmative tool, facilitated by an elution-head-based TLC-MS interface. Compounds within the bands of interest were washed off from the silica gel layer and directly measured by electrospray MS. Table 4 summarized the diagnosis signal of analytes recorded from the silica gel layer. It was clear that the 363.1 m/z were the most abundant ion signal for FWA 184, while 431.2 m/z for FWA 367. Agreeing well with theoretical molecular ions in the protonated form, these specific signals were straightforwardly self-evident and therefore can be used as the fingerprint evidence for confirmation. Further, the applicability of this method was exemplarily validated with positive founds in real samples. As shown in Fig. 4, the obtained signals displayed good similarity to the fingerprint of the analyte, demonstrating that the suspected bands can be unambiguously assigned to targeted compounds.

benefit of cistanche tubulosa extract

Conclusions

In this work, HPTLC was successfully employed as a flexible platform efficiently assembling separation and spectroscopic analysis, for facile quantification and confirmation of FWA contamination in cereal samples. With an optimized extraction method, analytes in wheat and rice flour samples were selectively and accurately quantified by HPTLC-FLD, with good linearity and LODs significantly below the tolerance limit. In addition, the separated results staying on the plating layer were further linked to in situ MS analysis, enabling rapid molecule identification of positive bands in separated tracks, therefore unambiguously ensuring the reliability of screening results. Facing the challenges raised by the reality of food screening analysis, the importance of HPTLC was increasingly acknowledged. The work demonstrated that HPTLC as a versatile platform integrating quantitative and qualitative detection in a superiorly cost-efficient manner can be a powerful tool in food analytical chemistry.

benefit of cistanche tubulosa extract supplements