Application of Liquid Chromatography / Ion Trap Mass Spectrometry Technique to Determine Ergot Alkaloids in Grain Products

Ergot alkaloids are mainly produced by Claviceps purpurea parasitic fungi during the plant fl owering period (1). Ergot grows on cereals such as rye, triticale, wheat and barley, where its sclerotia replace grains in the plant ears. Total content of alkaloids in ergot sclerotia may vary from 0.01 to 0.5 %. Sclerotia harvested together with cereal may contaminate food and feed. The consumption of food contaminated with ergot toxins may cause numerous diseases (1,2).


Introduction
Ergot alkaloids are mainly produced by Claviceps purpurea parasitic fungi during the plant fl owering period (1).Ergot grows on cereals such as rye, triticale, wheat and barley, where its sclerotia replace grains in the plant ears.Total content of alkaloids in ergot sclerotia may vary from 0.01 to 0.5 %.Sclerotia harvested together with cereal may contaminate food and feed.The consumption of food contaminated with ergot toxins may cause numerous diseases (1,2).
Poisoning with ergot alkaloid-contaminated food; i.e. ergotoxicosis, was in medieval Europe commonly referred to as St. Anthony's fi re (3)(4)(5).Ergot alkaloids may cause adrenergic cramps, swelling and red marks; in extreme cases limb necrosis may lead to death.Chronic ex-posure to ergot alkaloids may disturb the central nervous system, inducing dizziness, pins and needles, nausea, limb seizure, paralysis, psychoses, dementia and even death (5,6).Clinical symptoms in animals may include gangrene, miscarriage, convulsions, loss of lactation, oversensitivity and/or ataxia, i.e. disorder of muscle movement coordination (3).
In nature, ergot alkaloids are organic bases.The values of acid dissociation constant, pK a , for ergopeptines range from 5.5 (ergocristine) to 6.0 (ergometrine), while those for ergopeptinines range from 4.8 (ergocorninine) to 6.2 (ergometrinine).Natural ergopeptines are always accompanied by ergopeptinines.Ergopeptines may transform into ergopeptinines during long-term storage (particularly if storage conditions are improper) or during the extraction of ergot alkaloids from cereals.Diff erent epimers may have diff erent biological and physicochemical properties.Ergopeptines are more active biologically than ergopeptinines (2).
Currently, natural levels of ergot alkaloids in cereals are rarely reported (3).Nevertheless, some studies in Germany showed that the Claviceps purpurea rate of occurrence in cereals increased in recent years, which might be a ributed to the common cultivation of monocultures and mixed varieties of cereals (2).Despite effi cient grain cleaning procedures, ergot alkaloids can still be found in grain, particularly in rye-based cereals (3,8).Poisoning in humans with ergot alkaloids has been practically eliminated even if measurable contents of the toxins may be found in foodstuff s.However, alkaloids remain an important veterinary issue, particularly in feeding of ca le, horses, sheep, pigs and chickens (3).
No maximum permissible levels (MPLs) of ergot alkaloids in grain and/or cereals have been set in the EU, whereas regulations in Switzerland specify that the level in cereals is 100 µg/kg (1).In fact, it is diffi cult to set the MPLs because many structurally similar forms of ergot alkaloids may coexist (9).Furthermore, C. purpurea sclerotia content in grain to be used in an unground form in a fodder is limited by the 2002/32/EC European Directive to 1000 mg/kg (2).In a Recommendation issued on March 15, 2012, the European Commission recommended monitoring at least six ergot alkaloids (ergometrine, ergocornine, ergocristine, ergokryptine, ergosine and ergotamine) and their epimers in grain and cereals that are intended for human consumption or animal feeding (Recommendation (EC) No 2012/154/EU) (10).
Suitable analytical methods are required to acquire data on the occurrence of ergot alkaloids necessary to assess human exposure and to suggest respective MPL regulations.High-performance liquid chromatography (HPLC)--based methods are the most popular instrumental methods among numerous ones that are used.Fluorescence detectors (FLD) are most commonly used in analyses of ergot alkaloids (11)(12)(13)(14); however, mass spectrometers (MS) are becoming increasingly popular (3,9,(15)(16)(17).A comprehensive review of analytical procedures to determine ergot alkaloids was published by Krska and Crews (18).
Various mobile phases at isocratic or gradient fl ow are used to separate alkaloids in chromatographic columns.The most common phase is composed of an acetonitrile (ACN)/water mixture, which is sometimes slightly acidifi ed to improve the ionisation conditions in the electrospray ion (ESI) source of the mass spectrometer.In most qualitative and quantitative methods, the alkaloids are extracted using a mixture of polar and nonpolar solvents in an alkaline medium or a polar solvent in an acid medium.The clean-up procedures involve solid phase extraction (SPE) cartridges, which are fi lled with a cation--exchange resin in most cases (2).Neutral alumina-based beds were applied to clean up ergot alkaloids in grains as reported by Müller et al. (14).
Most available data in the literature on ergot alkaloid determination were obtained using mass spectrometers with triple quadruple analysers; few authors used the ion trap MS/MS, although it may show high sensitivity.However, this detector is also vulnerable to matrix eff ects: too many trapped ions degrade the trap performance because of their interactions (the so-called space charge eff ect).As a result, instrument sensitivity and/or repeatability may signifi cantly decrease.Thus, the to-be-analysed samples must be cleaned up very carefully or surplus ions from the matrix residues will 'clog' the instrument.
In this study, the analytical procedure for determination of six major ergot alkaloids and their epimers in rye-

Research material
The research material was composed of 65 rye and rye-based cereal samples.The types of tested samples were randomly selected.The rye grain samples (N=18) were obtained from grain mills located in diff erent regions of Poland.The rye samples (1 kg) were taken from grain elevators and ground using a mechanical impact mill (model WŻ-1S, ZBPP, Bydgoszcz, Poland).From the ground samples, laboratory subsamples were taken for testing.Other tested products, i.e. rye fl our (N=34), rye bran (N=12) and rye fl akes (N=1) came from diff erent suppliers and were bought in local supermarkets.

Extraction and purifi cation of ergot alkaloids
A mass of 5 g of a well ground sample was homogenised with 50 mL of 84:16 CAN/ammonium carbonate (200 mg/L) (by volume) solution for 2 min.The entire mixture was transferred into a 50-mL polypropylene tube and centrifuged for 10 min at 10 730×g.Subsequently, 20 mL of the supernatant were transferred into a separatory funnel and liquid-liquid extracted with 30 mL of n-hexane to eliminate fats.Then, 5 mL of the ACN/ammonium carbonate fraction were evaporated to dryness at 40 °C using a rotary evaporator IKA-RV-10 (IKA, Staufen, Germany).The residues were redissolved in 5 mL of dichloromethane/ethyl acetate/methanol/25 % aqueous ammonia mixture (50:25:10:1 by volume) and sonicated in an ultrasonic bath.
Neutral alumina preheated overnight at 600 °C in a muffl e oven was used to clean up the extracts.Then, 1 mL of the extract prepared as above was directly transferred on the 300 mm×20 mm glass chromatographic column fi lled with 3 g of alumina.The column was subsequently eluted with 9 mL of dichloromethane/ethyl acetate/methanol/25 % aqueous ammonia mixture (50:25:10:1 by volume).The eluate was collected into a round-bo om fl ask and evaporated at 40 °C using a rotary evaporator.The residues were redissolved in 2 mL of ACN/ammonium carbonate solution mixture (50:50 by volume).The samples were fi ltered through 0.2-µm nylon syringe fi lters (Chromacol, Welwyn Garden City, UK) before injection into the HPLC.

Ergot alkaloid standards
First, 0.5 mL of ergot alkaloid standard and 1 mL of the corresponding epimer standard were transferred into a 10-mL measuring fl ask and dissolved in ACN.The solution was subsequently transferred into an amber glass vial.The following concentrations of individual ergot alkaloid standards were obtained (in µg/mL): ergometrine, ergocornine, ergocristine and ergosine each 5.0; ergokryptine and ergotamine each 5.1; and ergometrinine, ergocorninine, ergocristinine, ergokryptinine, ergosynine and ergotaminine each 2.6.All standards were stored frozen at a temperature below -20 °C to prevent alkaloid epimerisation.The standards could also be stored at room temperature when dissolved in chloroform, although ACN solutions are more convenient for the HPLC analysis (19).

Calibration curves
A separate calibration curve was plo ed for each determined compound.Then, 0.5 mL of standard mixture was transferred into a 10-mL measuring fl ask and fi lled with ACN.The concentration of individual ergot alkaloid in the calibration mixture was (in µg/mL): ergometrine, ergocornine, ergocristine and ergosine 0.25, ergotamine and ergokryptine 0.26, and other compounds 0.13.Portions of 1, 3, 5, 13, 25, 50, 75 and 125 µL of the mixture were subsequently added into autosampler vials.The solvent was evaporated under a gentle stream of nitrogen, and the residues were redissolved in 1 mL of the 1:1 ACN/ ammonium carbonate buff er.

LC-IT-MS/MS analysis
An LCQ Advantage Max (Thermo Fisher Scientifi c, Austin, TX, USA) mass spectrometer with an ion trap operated in the positive ionisation electrospray mode was used for the determinations.The fl ow of nebulising gas (nitrogen) was set at 25 arbitrary units (AU), the fl ow of make-up gas (also N 2 ) was 10 AU, capillary bias was 34 V, nebuliser bias 5 kV, capillary temperature 260 °C and the ion source current 80 µA.The optimised instrument operating parameters, monitored daughter ions and precursor ions for fragmenting the individual studied compounds are shown in Table 1.
Chromatographic separations of six basic ergot alkaloids and their epimers were performed on a 150/2 Nucleodur ® Sphinx RP 1.8 µm (Macherey-Nagel, Düren, Germany) chromatographic column.Bifunctional octadecyl/phenyl groups in the column bed provided high selectivity for both hydrophobic and hydrophilic aromatic compounds in the pH range of 1-10.The mobile phase was selected according to the literature data (2,3,5,7,14,20,21).Effi cient chromatographic separation was obtained when the mobile phase was composed of water, 200 mg/L of ammonium carbonate buff er (pH=9.0±0.3) and ACN.Two mobile phases were used in a gradient setup: aqueous ammonium carbonate 200 mg/L (phase A) and ACN (phase B).The used gradient was: 40 % phase A from 0 to 4 min, 45 % from 5 to 7 min, 50 % from 10 to 20 min, 80 % from 25 to 37 min and 40 % phase A from 40 to 60 min.The cleaned-up samples were dissolved in a 1:1 ACN/ aqueous ammonium carbonate buff er mixture and maintained in the autosampler at 10 °C to slow down the epimerisation processes.The chromatographic column was maintained at 50 °C.A typical chromatographic separation of the ergot alkaloid standard mixture is shown in Fig. 2.

Results of LC-IT-MS/MS analysis
The optimal values of ionisation parameters were established in a series of preliminary experiments.The mass spectrometer operating parameters were automatically optimised a er injecting a solution of the individual analytes into the ion source using a syringe pump.The optimised ionisation parameters and the monitored daughter ions for individual analytes are shown in Table 1.
The protonated molecular ions (M+H + ) are shown in Fig. 3.The MS signal intensity mainly depends on the source temperature, cone bias voltage, capillary bias voltage, nebulising gas fl ow rate, and make-up gas fl ow rate.The optimum source temperature was 260 °C.Losses of ions at a too high nebulising/make-up gas fl ow rate caused the spectrometer sensitivity to decrease.
The fragmentation and ionisation pa ern of ergot alkaloids in the electrospray source, which was operated in the positive ion (ESI) mode, was studied to establish the MS/MS operating conditions.This technique is most frequently applied to analyse ergot alkaloids because it provides the highest sensitivity at atmospheric pressure ionisation conditions (2).The negative ionisation mode was also applied, but it did not give satisfactory results.Daughter ions of individual ergot alkaloids and their epimers that were produced in the MS/MS mode are shown in Fig. 4. In all cases, except ergometrine (Em) and its derivative ergometrinine (Emn), fragmentation ions were produced by dissociating a water molecule and a aching a proton (M-H 2 O+H + ).The single type fragmentation ions were produced at any given excitement bias voltage.The fragmentation of ergometrine ions consisted in cleaving the CONHCH(CH 3 )CH 2 OH chains off the basic molecules, so the most intense peak was observed at m/z=223.2.
Fragmentation of ergometrinine ions consisted in cleaving the CH 3 CH 2 CH 2 OH chains off the epimer molecules, so the most intense peak was observed at m/z=265.2.
Müller et al. (14) reported a neutral alumina-based technique for this purpose.In their work, ergot alkaloid extracts dissolved in the 50:25:5:1 (by volume) dichloromethane/ethyl acetate/methanol/25 % aqueous ammonia mixture were transferred directly into a glass column fi lled up with heated neutral alumina.In this alkaline medium, the ergot alkaloids did not exhibit practically any retention on alumina compared to the retention of impurities.Thus, the initial fractions of the eluent contained highly purifi ed analytes.However, in this work, somewhat more polar 84:16 ACN/ammonium carbonate buff er extraction mixture was used.The residues that remained a er the extraction solvent evaporated were redissolved in the dichloromethane/ethyl acetate/methanol/25 % aqueous ammonia mixture.This change to a more selective solvent helped us effi ciently separate the analytes from the matrix impurities because the la er were more strongly adsorbed on neutral alumina.Using the modifi ed method, we were able to elute more effi ciently the com-Fig.3. Molecular ions of individual ergot alkaloids and their epimers produced in the MS mode.Abbreviations as in Fig. 2 pounds that were treated with the previously described extraction/purifi cation procedure, which exhibits low recovery rates, i.e. ergometrine, ergometrinine, ergosine and ergotamine.The elution pa ern of ergot alkaloids from a neutral alumina-based column (extracted with ACN/ammonium carbonate buff er mixture, extraction solvent evaporated, and residues redissolved in dichloromethane/ethyl acetate/methanol/25 % aqueous ammonia mixture) is shown in Fig. 5.
The eff ect of the alumina deactivation degree on the analyte recovery from fortifi ed rye samples was also assessed.Alumina was deactivated at 600 °C overnight and subsequently mixed with water at 1, 2, 3, 4 and 5 %.No signifi cant correlation between the recovery rate of the studied compounds and the deactivation level was noticed.However, at higher water contents (2, 3, 4 and 5 %), the background noise related to the sample matrix appeared to increase in the chromatograms, which aff ected the repeatability of the instrumental readback.The lack of eff ect of the deactivation level on the recovery rate was most probably related to the composition of the phase that was used for the column elution.The solvent mixture was composed of polar solvents, i.e. ethyl acetate, methanol and water, which eff ectively bound to the existing active sites on the alumina surface.The eff ect of the addition of water (deactivation level) to alumina on the analyte recovery rate is shown in Fig. 6.

Method validation
The obtained ergopeptine mass fractions were within the range of 3-325 µg/kg, and the ergopeptinine mass fractions were within the range of 1-163 µg/kg (Table 2).The coeffi cient determination (R 2 ) was higher than 0.98 for every analysed compound.The limit of quantifi cation (LOQ, the lowest point on the calibration curve) for indi-    2 vidual analytes was found between 1 and 3 µg/kg (signal to noise ratio S/N>170).An example of calibration curve for ergometrine is shown in Fig. 7.
To prepare the spiked samples, 25, 100 or 300 µL of the calibration stan dard mixture were added to fl our samples (5 g) that were earlier tested for the presence of ergot alkaloids.The samples were le intact in a dark place at room temperature to let the solvent evaporate.The obtained mass fractions at the three spiking levels were 25, 100 and 300 µg/kg for ergopeptines, and 12.5, 50 and 150 µg/kg for ergopeptinines.The mean recoveries ranged from 63.0 % (ergometrine) to 106.4 % (ergosine, ergocorninine), depending on the alkaloid and its spiking level.
The method precision (relative standard deviation, RSD) was below 18 % (Table 3).The method accuracy was assessed by analysing a certifi ed material (ergot alkaloids in rye fl our BRM 003020) according to the Eurachem/CIT-AC guidelines: absolute diff erences between the measured/certifi ed mass fractions were compared against the combined measurement uncertainties and the certifi ed value uncertainties (Table 4).
The combined uncertainty of the determined mass fractions was estimated based on the identifi ed uncertainty sources in the proposed analytical method.The uncertainty was expressed as an expanded value at k=2 and α=0.05.The expanded uncertainty value was used to assess the diff erences in the ergot mass fractions that were determined for the studied certifi ed reference material samples and reference concentration values.The diff erence did not exceed the combined measurement uncertainty in any case.In other words, no statistically significant diff erence between the obtained results and values that was declared in the material certifi cate was noted.

Levels of ergot alkaloids in rye-based food products and ergot sclerotia
The developed analytical procedure was used to estimate the levels of ergot alkaloids in rye-based food products supplied by local manufacturers.In total, 65 samples were tested (rye grain 18, fl our 34, bran 12, fl akes 1).The results are shown in Table 5.
Ergot alkaloids were detected in 83 % of the tested rye grain, 94 % of rye fl our and 100 % of rye bran and fl ake samples.The mean total mass fractions of the studied alkaloids in the tested rye grain and rye fl akes were found to be (14.8±9.8) and (127.8±9.9)µg/kg, respectively.A relatively high mass fraction of 1215.5 µg/kg was found in one rye fl our sample.A chromatogram of fl our that was naturally contaminated with ergot alkaloids is shown in Fig. 8. Ergotamine (48 samples), ergocornine (46 samples) and ergosine (45 samples) are the most commonly found alkaloids, whereas ergometrinine (3 samples) and ergometrine (20 samples) are the least commonly found ones.
These results are consistent with the literature data.Additionally, alkaloids were determined in ergot sclerotia that were isolated from rye grains.The total content was nearly 0.01 % (97.9 mg/kg).However, the alkaloid profi le was dominated by ergocristine at 44.7 mg/kg (45.6 % share in total), which is not among the most commonly found alkaloids in the tested food products.Ergocorninine at the mass fraction of 0.2 mg/kg (0.2 % share in total) was the least abundant alkaloid.The absolute levels and contribution of individual compounds to the total content of alkaloids in ergot sclerotia are shown in Fig. 9.The chromatogram of alkaloids in the ergot sclerotia sample is shown in Fig. 10.The proportions of various alkaloids in ergot sclerotia and contaminated food products depend on the fungal species, geographic location, and plant on which the fungi parasitise.However, the exact mechanisms that produce the diff erences remain unknown (2,28).

Conclusions
A liquid chromatography/ion trap mass spectrometry-based method has been developed to determine six ergot alkaloids and their isomers in rye-based food prod-ucts.The analyte recovery rates ranged from 63.0 to 104.6 %, depending on the spiking level and analyte type.Relative standard deviation was below 18 %.The method performance was verifi ed by analysing ergot alkaloids in rye certifi ed reference material (BRM 003020).
Ion traps are relatively rarely used to determine ergot alkaloids because they require the analytes to be carefully purifi ed to eliminate matrix ions that may easily degrade the performance of the mass spectrometer (the space charge eff ects within the trap); only a few similar methods have been reported in the literature.The originally developed cleaning method can effi ciently separate alkaloids from polar impurities of the sample matrix on a neutral alumina bed.
The developed analytical procedure was applied to determine the levels of ergot alkaloids in 65 samples of selected rye-based food products.Measurable levels of alkaloids were found in most of the analysed samples (particularly rye fl our samples).A single sample of fl our contained ergot alkaloids at a combined level of 1215.5 µg/kg.The aforementioned results, which were obtained from a relatively small number of samples, suggest that ergot alkaloids in rye-based food products (and fodder) should be monitored on a wider scale.

Fig. 4 .
Fig. 4. Fragmentation ions of individual ergot alkaloids and their epimers produced in the MS/MS mode.Abbreviations as in Fig. 2

Fig. 6 .
Fig.6.The infl uence of the alumina deactivation level on the ergot alkaloid recovery from the rye grain sample.Abbreviations as in Fig.2 Reinhard et al. (9) tested 14 rye fl our samples and found the total ergot alkaloid content in the range of 18-519 µg/ kg.Storm et al. (27) tested 34 samples of rye fl our that were produced from grain harvested in both ecological

Fig. 8 .
Fig. 8. Chromatogram of a fl our sample naturally contaminated with ergot alkaloids.Abbreviations as in Fig. 2

Fig. 9 .
Fig.9.Absolute levels and profi le of alkaloids in ergot sclerotia.Abbreviations as in Fig.2

Table 1 .
MS/MS precursor ions/fragmentation ions and optimal ionisation parameters for individual ergot alkaloids

Table 2 .
Parameters of the calibration curves determined for individual ergot alkaloids X=mass fraction (µg/kg), R 2 =correlation coeffi cient, LOQ=limit of quantifi cation, S/N=signal to noise ratio, LOD=limit of detection.Abbreviations as in Table1

Table 4 .
Method accuracy assessed in the analyses of ergot alkaloids in rye certifi ed reference material (CRM)

Table 3 .
Mean recoveries and relative standard deviations (RSD) for rye fl our samples fortifi ed at three spiking levels (25, 100 and 300 µL of calibration standard mixture)

Table 5 .
Levels of ergot alkaloids in rye-based food products