Determination of Aflatoxins in Medicinal Plants by High-Performance Liquid Chromatography–Tandem Mass Spectrometry

– Purpose. The intention of the proposed work is to study the presence of the aflatoxins B 1 , B 2 , G 1 and G 2 in medicinal plants, namely Mucuna pruriens , Delphinium denudatum and Portulaca oleraceae . Methodology. The aflatoxins were extracted, purified by immunoaffinity column chromatography and analysed by high-performance liquid chromatography–tandem quadrupole mass spectrometry with electrospray ionisation (HPLC–MS/MS). Fungal count was carried out in PDA media. Results. A good linear relationship was found for AFB 1 , AFB 2 , AFG 1 and AFG 2 at 1–10 ppb (r>0.9995). The analyte accuracy under three different spiking levels was 86.7–108.1 %, with low per cent relative standard deviations in each case. The aflatoxins can be separated within 5 to7 min using an Agilent XDB C18-column. We found that AFB 1 and AFB 2 were in trace amounts below the detection limit in M . pruriens whilst they were not detected in D. denudatum . P. oleraceae was found to be contaminated with AFB 1 and AFB 2 . AFG 1 and AFG 2 were not detected in M. pruriens, P. oleraceae and were below the detection limit in D . denudatum . This was consistent with very low numbers of fungal colonies observed after 6 hr of incubation. Conclusion. The analytical method developed is simple, precise, accurate, economical and can be effectively used to determine the aflatoxins in medicinal plants and therefore to control the quality of products. The aflatoxin levels in the plant extracts examined were related to the minimal fungal load in the medicinal plants examined.


INTRODUCTION
The plants Mucuna pruriens, Delphinium denudatum and Portulaca oleraceae have the potential for wide therapeutic applications. Mucuna pruriens L (Velvet beans; Fabaceae), a herbaceous forage and food legume, has found widespread usage in the management of several diseases, including diabetes, rheumatoid arthritis, atherosclerosis, male infertility and nervous disorders, and is a good source of 3,4-dihydroxyphenylalanine (DOPA) (1,2). In addition to health benefits the seeds possess potent aphrodisiac, geriatric tonic and vermifuge activities (3). The major constituent of Mucuna seeds was found to be L-dopa (5%), along with minor amounts of methylated and nonmethylated tetrahydroisoquinolines (0.25 %) (4). Delphinium denudatum Wall (Jadwar; Ranunculaceae) is one of the important medicinal drugs used as indigenous medicine in India, and the entire plant is reported to be useful in a variety of ailments (5). The root is used in various medical formulations in Unani and Ayurveda to reduce the withdrawal symptoms in people on de-addiction therapy (6). It is a natural analgesic and is prescribed for relief from toothache (7). Herbalists recommend the roots in the treatment of fungal infections, asthma, cough, jaundice and nervous problems (8). It is considered as a natural blood purifier and prescribed for people who have digestive problems and provides relief from painful wounds and inflammation (9). Chemically D. denudatum contains many active compounds, some of which are alkaloids including delphocurarine, staphisagrine, delphine, condelphine, denudatin, talatizidine, acetylhetero-phylistine and a diterpenoid alkaloid identical to condelphine (10). Portulaca oleracea L. (Purslane; Portulaceae) grows widely in different areas of the world including India. The seeds are regarded as valuable in the treatment of urinary problems, digestive problems and cardiovascular diseases (11). Studies demonstrated a variety of pharmacological activities, including hypoglycemic (12,13), hypocholesterolemic, antioxidant, analgesic, antiinflammatory, skeletal muscle relaxant, smooth muscle relaxant, neuroprotective (14), anxiolytic, sedative, anticonvulsant, antiulcerogenic, bronchodilatory, wound healing, and antifungal effects (15). It contains a variety of bioconstituents, including catecholamines, l-noradrenaline, dopamine, α-amyrin, β-amyrin and portuloside A (16,17,18).
Aflatoxins are a group of structurally related toxic metabolites of mycotoxins produced by Aspergillus flavus and Aspergillus parasticus (19). In the aflatoxin group, about 16 compounds are known, but only aflatoxin B 1 , B 2 , G 1 , G 2 , M 1 and M 2 are routinely monitored (20,21). Aflatoxin B 1 , B 2 , G 1 and G 2 are chemical derivatives of difurancoumarin ( Fig. 1). Aflatoxin B 1 is the prevalent compound in plant samples and it is acutely toxic and the most carcinogenic amongst the aflatoxins (22). Aflatoxins may increase stress susceptibility and compromise growth efficiency. The clinical signs of aflatoxicosis in humans are extremely varied. Signs of acute aflatoxicosis include depression, nervousness, abdominal pain, diarrhea and death (23). The aflatoxins occur widely in food, especially in corn, nuts, peanuts, coconut, fruits, dried fruits and beer (24,25). High moisture and temperature are two main factors that cause the occurrence of mycotoxins at pre-harvest and post harvest stages (26). Fungi are widely distributed as environmental contaminants (27), in fact, under favorable conditions of temperature and humidity, they grow on many commodities including cereals, oil seeds, nuts, herb-teas and spices (28). They have also been found in medicinal herbs that are not fully dried or stored improperly (e.g. at high temperature and humidity) (29,30). Despite many studies on mycotoxins in agricultural products, only a few are concerned with spices and herbal medicines. The determination of aflatoxins in herbs is becoming important for medicinal safety, as more and more people are using medicinal herbs for complementary and alternative therapy and they play an important role in the economy, especially in developing countries. Previously aflatoxins were analyzed in peanut, cocoa, pistachio, chili, ginger (31,32), chocolate, cotton-seed, job's tears, pepper, paprika, feed and wheat samples (33). The current acceptance level of total aflatoxin in maize set by the United States-Food and Drug Administration (USFDA) is 20 μg/kg (34). Health Canada has set acceptance criteria of < 20 µg/kg (ppb) for aflatoxins (B 1 +B 2 +G 1 +G 2 ), and < 5 µg/kg (ppb) for aflatoxin B 1 for evening primrose oil, sugar cane, sugar beets, cottonseed, peanuts, corn, ginseng or any substance derived from these sources (35). Recently, the Codex Alimentarius Commission, the Joint FAO/WHO Food Standards Program, adopted a limit of 15µg/kg for total aflatoxins (36). In 2008, the Codex Alimentarius set a maximum level of 10 µg/kg total aflatoxins in ready-to-eat almonds, hazelnuts, and pistachios (37). In the European Union, the total aflatoxins and AFB 1 levels in human commodities are regulated with maximum residue levels (MRLs) that cannot be greater than 2 and 4 µg/kg, respectively (38). In the year 2002 the European Food Safety Authority (EFSA) set maximum limits in spices (Capsicum spp., Piper spp., Myristicafragrans, Zingiber officinale, Curcuma longa) (AFB 1 5 µg/kg; total AFs 10 µg/kg) (02/472/EC) (39).
Environmentally sound and most widespread methods applicable for quantitative determination of aflatoxin content in different samples are thin-layer chromatography (40,41) and high performance liquid chromatographic (HPLC) (42,43). Aflatoxins can be separated and detected using either normal-or reversed-phase HPLC methods mainly with fluorescence spectrometric detection. The use of two-dimensional thin-layer chromatography (2D-TLC) (44,45) and high performance thin-layer chromatography (HPTLC) for the determination of aflatoxins has been reported (46). An alternative complex media based method is reported to detect the natural fluorescence of aflatoxins released by the growing mycelium (47) or relies on multiplex polymerase chain reaction (PCR) and real time polymerase chain reaction (RT-PCR) detection of genes or their transcripts involved in the aflatoxin biosynthetic pathway (48,49). An enzyme-linked immunosorbent assay (ELISA) for AFB 1 , and levels of total aflatoxins were quantified and confirmed by liquid chromatography combined with tandem mass spectrometry (LC-MS/MS) (50). However, a lot of time and a tedious pre-or post-column derivatization is required with conventional HPLC in order to provide better detection limits of AFB 1 , AFB 2 , AFG 1 and AFG 2 (51). Furthermore, many other structurally similar components to the AFs in the plants can also be retained by the aflatoxin immunoaffinity column, and as a result produce false positive results in many chromatographic techniques.
Despite the numerous uses and the various biologically active chemical constituents reported in M. pruriens, D. denudatum and P. oleracea, no data on aflatoxin detection of these plants are available in the literature. In order to ensure safety and quality we have collectively developed a cost efficient, flexible and rapid method for the quantification of aflatoxins in M. pruriens, D. denudatum and P. oleraceae as these plants are commonly consumed by people without any awareness of their safety. There exist no well-accepted scientific methods for proper determination and quality control of aflatoxins in selected plant materials. The aim of this study was to screen the content of aflatoxins as well as providing a sensitive, accurate and reproducible analytical method for the detection of aflatoxins. Further, the purpose of this study is to provide useful information on aflatoxin levels in M. pruriens, D. denudatum and P. oleraceae and to achieve a confirmatory method for quick determination of AFB 1 , AFB 2 , AFG 1 and AFG 2 by high performance liquid chromatographyelectrospray ionization tandem mass spectrometry (HPLC-MS/MS). The proposed method overcame the drawbacks of HPLC by using tandem MS/MS detection, which can determine the presence of a specific substance more accurately than HPLC and avoids the tedious derivatization process. The method is validated for linearity, recovery, sensitivity, precision and repeatability, and was shown to successfully determine the levels of AFs in plants samples.

Equipment and chemicals
The analysis was performed on an Agilent 1200 system using an Agilent XDB-C18 4.6-50 mm column containing 1.8 mm particles. MS/MS detection was achieved on an Agilent 6410B tandem quadrupole mass spectrometer using an electrospray ion (ESI) source in positive mode. Data were acquired and processed using the Masshunter software (Agilent). Acetonitrile, methanol, ammonium acetate and HPLC grade water were supplied by Merck (Darmstadt, Germany). AFB 1 , AFB 2 , AFG 1 and AFG 2 were purchased from Sigma (St. Louis, MO). All other inorganic chemicals and organic solvents were of reagent grade.

HPLC-MS/MS analysis
Chromatographic analyses were carried out using a binary gradient elution with 70:30 v/v methanol: acetonitrile as eluent A and 5 mM aqueous ammonium acetate as eluent B. The total run time was 5 to 8 min. The injection volume was 5 mL, the flow rate was set at 5 µL/min and the column temperature was maintained at 30 o C. The ionisation source of the MS/MS detector had 4.0 kV capillary voltage, 350 o C source temperature and 9 L/min desolvation gas flow (both gases were nitrogen).  Table 1.

Sample preparation
The samples (5 g) were sonicated in 100 mL 70:30 v/v methanol:water for 45 min and centrifuged (Remi Compufuge-CR30) for 10 min, then 5 mL of the supernatant was drawn, diluted with 30 mL milli-Q water, and passed through the immunoaffinity column at no more than 5 mL/min (the column was previously conditioned with 30 mL distilled water). The column was rinsed with 15 mL double distilled water to remove the matrix components and then dried by passing air through to remove any adherent water. The eluate was diluted to 2 mL with water and passed through a 0.22 mm filter, and the filtrate was injected into the HPLC-MS/MS.

Peak recognition
The AFs in samples were recognized by comparing their retention times and MS spectra (MRM mode) with the standards. The standards used are summarized in Table 1.

Fungal count in the plant materials
All the plant materials were diluted with sterile water and 500 µL of the samples were spread over potato dextrose agar (PDA)-Ox bile media. They were incubated for 5 days at 28 o C and 75 % relative humidity, and the colony forming units (CFU) of the fungal strains were observed.

Selection of mobile phase
The HPLC mobile phase was optimized by considering the retention time, peak response, sensitivity, peak sharpness as well broadening and separation efficiencies. Methanol-water in the combination of 60:40 v/v completely separated the analytes but was found not suitable for the column, while acetonitrile-water 60:30 v/v failed to separate the AFs. Methanol:acetonitrile in a combination of 70:30 v/v with 5.0 mM aqueous ammonium acetate was found the best mobile phase for AF standards separation (Fig. 2).

Optimisation of the mass spectra
Mass spectrometry detection was carried out on positive ionisation mode because this mode gave sharp and sensitive signals. Mass spectrometry detection was optimised by injecting the AF standards (100 ng/mL) under positive ionisation mode. The ion fragments were evaluated by their run time to the most abundant m/z and the ion with the uppermost intensity was selected as the basic ion for quantitation (Fig. 3).

Limits of detection and limits of quantification
The AF analyses showed good results in the MRM mode, the LOD reaching 0.28-1.10 µg/kg and LOQ 0.79-3.34 µg/kg, with a signal-to-noise ratio (S:N) of 3. Linear calibration curve for various concentrations of AFs ranging from 1 ppb-10 ppb (three points) were analysed. The calibration curves were constructed by plotting the response against the concentration. A linear relationship was obtained for each compound, and the correlation coefficients ranged from 0.9992 to 0.9995 ( Table 2). The relative standard deviations (RSDs) of precision of AF mixtures were 7.04-2.17 % (n=6). Excellent retention times for AFB 1 (7.973±0.018), AFB 2 (7.361±0.017), AFG 1 (6.689±0.015) and AFG 2 (5.830±0.026) were observed for corresponding mixtures of AFs (Table 2).

Accuracy
Accuracy analysis was performed by testing samples at three different concentrations 1 ppb, 5 ppb and 10 ppb. All plants were used for the accuracy test to ensure that the method has wide adaptability. The plant sample (5 g) was spiked with high, medium and low levels of AF standards. The accuracy ranged from 86.7% to 108.1% (Table 3), indicating good accuracy of the method.

HPLC-MS/MS Chromatograms (Counts vs
Acquisition time (min)) for AFB 1 , AFB 2 , AFG 1 and AFG 2 were determined for M. pururiens (Fig. 4), D. denudatum (Fig. 5) and P. oleraceae (Fig. 6). The study revealed mass peaks of AFB 1 and AFB 2 in M. pruriens below the detection limit whilst a resultant mass peak of AFB 1 and AFB 2 was observed in P. oleraceae, whereas no mass peak was found in D. denudatum (Figs. 4, 5 and 6). The aflatoxin levels in the plants investigated is summarized in Table 4.

Fungal count
Total fungal colonies observed per gram of M. pruriens, D. denudatum and P. oleraceae were 18×10 2 , 4×10 2 and 20×10 2 respectively in PDA-ox bile media. The total fungal CFU also show that there was very low growth of fungus in these tested plant materials, which is consistent with the low levels of mycotoxins.

DISCUSSION
An LC-MS method was developed which was found to be suitable for the rapid and accurate assessment of aflatoxin levels in selected plant samples. The method was subsequently applied to determine the natural occurrence of AFs in selected widely used commercial medicinal plant samples (Table 1). We found that P. oleraceae was contaminated with AFB 1 (1.675 µg/kg) and AFB 2 (1.335 µg/kg), whilst levels in M. pruriens were below the detection limit for AFB 1 (0.04 µg/kg) and AFB 2 (0.05 µg/kg) and were not detected in D. denudatum. AFG 1 and AFG 2 were absent in M. pruriens and P. oleraceae, whilst a relatively low value (0.07 µg/kg) of AFG 2 was found in D. denudatum (Table 4). Table 2. Linearity, limit of detection, limit of quantification and precision for the analysis of AFB 1 , AFB 2 , AFG 1 (52). This may account for the high incidence and concentrations of AFB 1 and AFB 2 compared with AFG 1 and AFG 2 in the three plants investigated in this study. Rhizomes, roots, seeds and fruits have apparently higher AFs than in flowers, indicating that the herbs rich in starch and oil might be more favorable for fungal growth (53).
In particular, contamination levels of some samples i.e. root of D. denudatum, seeds of M. pruriens and P. oleraceae were notable, as all of these have a content of fat or fatty oil, suggesting that plants and their products rich in fat or fatty acids should be carefully monitored. The plant samples should be dried as soon as possible and stored in moistureproof containers with good aeration to avert the growth of AFs, as inappropriate storage and moisture content in samples encourages augmentation of mycotoxins (54). We conclude from this study that the risk of AF contamination in medicinal herb like M. pruriens, D. denudatum and P. oleraceae appears low but the presence of AFs in these plants needs to be analyzed before processing to confirm the absence of mycotoxins. The consumer must be cautious while consuming these plants for medicinal purposes. The monitoring scheme described could be applied to determine the aflatoxin contamination in a large number of samples in a cost and time effective manner. The data obtained from this monitoring can be used as a basis for risk analysis of aflatoxin contamination, thereby maintaining the aflatoxin contents at the lowest possible levels.