Highly selective analysis of pyrrolizidine alkaloids in herbal extracts 

With SCIEX QTRAP^® 6500+ LC-MS/MS System and SelexION^®+ Differential Mobility Separation Technology

Jasmin Meltretter¹, Anja Bürkli², Axel Besa^1
1 SCIEX, Darmstadt, Germany, ² Coop Genossenschaft, Basel, Switzerland

Introduction

Pyrrolizidine alkaloids (PAs) are a large group of natural secondary metabolites with alkaloid structures that are produced by plants as a defense against herbivores. They have been detected in significant concentrations in plant-based food, such as herbal infusions, black or green tea, green leafy vegetables, and honey products. Within the last ten years, more and more attention has been drawn to the presence of PAs in food, since various studies have demonstrated their hepatotoxicity as well as their genotoxic and carcinogenic potential. For these reasons, the European Food Safety Authority concluded that PAs represent a serious health concern and ranked them among the substances in food requiring careful monitoring.¹ Still, no legal limits for maximum allowed concentrations of PAs in food are available yet. As a result, the generally accepted maxim these days is to keep PAs concentrations in food as low as reasonably achievable.²

From an analytical point of view, PAs represent an especially challenging group of compounds. While protocols for the sample preparation and analysis are available,³ the LC-MS/MS detection remains particularly demanding. PAs are a collection of several isomeric substances that cannot be distinguished by mass spectrometry alone since the molecular weight, as well as the fragment masses, are not specific for individual compounds. Even with modern ultra-high performance LC separation technologies, it is impossible to fully chromatographically resolve some of the enantiomeric compounds, such as indicine/lycopsamine/echinatine or intermedine-N-oxide/indicine-N-oxide.^5,6

In this paper, an innovative method to analyze 27 pyrrolizidine alkaloids in herbal extracts is presented using a SCIEX QTRAP 6500+ System equipped with the SelexION+ Differential Mobility Separation Device (Figure 1). The presented method covers all 17 PAs that the EFSA recommends monitoring in food,¹ other PAs that are suggested for routine analysis by the Federal Institute for Risk Assessment in Germany (BfR),⁴ and some additional isomeric species of PAs.

Key features of SelexION+ Technology for separation of pyrrolizidine alkaloids

• Differential mobility separation using SelexION+ Technology is able to distinguish between enantiomeric species that are not separated by chromatography
• SelexION+ Technology cand be coupled with MRM analysis on a QTRAP 6500+ System for a fast quantitative method for 27 pyrrolizidine alkaloids in herbal extracts

Methods

Sample preparation: Extracts from fennel, chamomile, and peppermint were prepared as described by the German Federal Institute for Risk Assessment.³ Briefly, pyrrolizidine alkaloids from 2 g of sample were extracted twice with 20 mL of 0.05M sulfuric acid in an ultrasonic bath. Following centrifugation, filtration and pH adjustment, aliquots of the extracts were purified and concentrated by solid-phase extraction using C18 cartridges. The eluate was dried under nitrogen flow, and the residue was reconstituted in 1 mL of 5% aqueous methanol. For spiking experiments, 180 µL of the blank extracts were spiked with 20 µL of a standard solution, yielding final concentrations of 0.125-25 ng/mL, corresponding to 0.25-50 ng PAs per gram sample.

Chromatography: Pyrrolizidine alkaloids were chromatographically separated on a Phenomenex Luna^® Omega C18 column (100 x 2.1 mm, 1.6 μm), using the SCIEX ExionLC™ AD System. Mobile phase was 5 mM ammonium formate and 0.1% formic acid in water (eluent A) and in methanol (eluent B), respectively. Oven temperature was set at 50 °C. Injection volume was 5 µL. Maximum pressure throughout the chromatographic run was <580 bar.

Mass spectrometry: The SCIEX QTRAP 6500+ System was equipped with a SelexION+ Differential Mobility Separation Device. The analysis was performed in positive polarity with electrospray ionization. MS parameters were optimized on a column by injecting a standard solution at different settings. Final parameters were as follows: source temperature, 600 °C; curtain gas, 20 psi (40 psi for experiments without SelexION⁺ Device); GS1, 40 psi; GS2, 60 psi; ion spray voltage, 4500V; CAD gas, 9 psi; entrance potential, 10 V. Declustering potentials, collision energies, and cell exit potentials were set depending on the compound. The separation voltage, SV, of the SelexION+ Device was set at 3800V. Isopropanol was continuously infused into the curtain gas flow at 249 µL/min. The ion mobility cell was heated to 150 °C. Resolution gas was set to medium. For each analyte, the specific optimized compensation voltage, COV, was applied. Data was acquired using the Scheduled MRM™ Algorithm with a target scan time of 0.6 seconds using an MRM detection window of 40 seconds.

LC separation of PAs

Phenomenex Luna Omega Polar C18 column with 10 cm length was employed for the analysis of pyrrolizidine alkaloids in herbal tea extracts and allowed for the chromatographic separation of most of the compounds except some of the isomeric species (Figure 2). The demonstrated use of SelexION+ Technology significantly increased the selectivity of this approach. Therefore, a chromatographic baseline separation of the isomers is not necessary and, thus, the total run time could be reduced to 12 minutes.

Separation using differential mobility spectrometry

Figure 3 shows the principle of the separation in the ion mobility device. The DMS cell is composed of two flat plates that are parallel and define a mobility region. While the ions are drawn by the transport gas flow towards the MS, the power supply outputs two sine waves of different frequencies on each of the two DMS electrodes, resulting in an oscillating high and low voltage field (separation voltage, SV). Due to the difference between high and low field ion mobility coefficients, ions will migrate toward the walls on a wavelike path and leave the flight path unless their trajectory is corrected by a counterbalancing voltage, a DC potential referred to as compensation voltage, COV. This compensation voltage is highly specific for each compound and enables the removal of matrix interferences and the separation of isomers.

Tuning of SelexION+ Technology parameters

Tuning of SelexION+ Technology parameters is easily done using infusion of standards. The compensation voltage can then be ramped while varying settings for other parameters such as separation voltage, resolution gas or the addition of a modifier solvent to the curtain gas flow. The aim is to get the best separation of the isomers while maintaining the maximum intensity possible.

Figure 4 shows the tuning results for a standard mixture of intermedine, indicine, lycopsamine, and echinatine, and their oxidized species. These analytes are especially demanding in LC-MS analysis, since they are difficult to separate by chromatography and, moreover, they have the same transitions in MS analysis.

For both the non-oxidized and oxidized compounds, a slight broadening of the peak was observed when the COV was ramped at increasing separation voltages (Figure 4A), indicating a beginning, very slight separation of the isomers. In the next step, the resolution gas was adjusted, which increases the residence time of the ions in the cell and thereby the resolution of the separation. With the resolution gas flow set to high, shoulders are detected on the signals for both analyte groups, but still the four isomers are far from being separated (Figure 4B).

To further improve the resolution, isopropanol was used as a modifier and continuously infused into the curtain gas flow. Since the modifier forms clusters with the analyte ions, the mobility of the latter in the cell can be significantly modified. Finding the right modifier strategy can lead to a better separation. Figure 4D shows the COV ramps with different SV settings, with the resolution gas set to medium and isopropanol as a modifier. With increased separation voltage, the enantiomers are more separated from each other.

In the final test, using isopropanol modifier and an SV of 3800 V based on the previous tests, the resolution gas was varied. The medium setting is the best compromise between sufficient separation of the non-oxidized species and transmission losses of some of the oxidized species (Figure 4D). 

Added selectivity with differential mobility separation

While indicine, lycopsamine, and echinatine, as well as indicine-N-oxide and intermedine-N-oxide, could not be chromatographically separated (Figure 5A), they could be easily resolved when SelexION+ Technology was used (Figure 5B). The compound-specific compensation voltage, determined during SelexION Device tuning, served as an efficient parameter for increased selectivity.

Next, the selectivity of using DMS with a Q1 Multiple Ion was compared to the industry standard MRM approach (Figure 6) using a peppermint extract. DMS provided an increase in the selectivity of the Q1 MI scan compared to a Q1 MI scan without DMS and similar selectivity to the classical MRM experiment. For this reason, it is considered sufficient to monitor one single MRM transition for each PA compound when working with the ion mobility device, which allows to increase dwell times and thus to improve signal-to-noise at lower concentrations.

Figure 7 demonstrates that the SelexION Device not only separates isomeric compounds, but removes matrix interferences. As shown in Figure 7A, a matrix compound which is present already in the peppermint blank extract elutes very closely to heliotrine-N-oxide, hampering the accurate and precise quantification at low concentrations. With SelexION+ Technology, however, the interference is removed providing a cleaner blank such that the signal of heliotrine-N-oxide can be easily integrated. As another example, the lasiocarpine-N-oxide trace shows a high background noise level, whereas the noise is completely eliminated with SelexION Technology (Figure 7B).

Sensitivity and linearity of the analysis of pyrrolizidine alkaloids in herbal extracts

With the final method using SelexION+ Technology, blank extracts from fennel, chamomile, and peppermint spiked with pyrrolizidine alkaloids to final concentrations from 0.125-25 ng/mL (corresponding to 0.25-50 ng/g raw material) were analyzed. Depending on the compound and matrix, the lowest detectable concentration was between 0.125 ng/mL and 1.25 ng/mL. Figure 8 shows the chromatograms at 2.5 ng/mL.

All compounds behaved linearly in the concentration range from 0.25-25 ng/mL with coefficients of correlation greater than 0.99 (Figure 9). Higher concentrations were not tested, but it can be expected that the concentration range could be extended due to the excellent dynamic range of the SCIEX QTRAP 6500+ LC-MS/MS system.

Quantification of pyrrolizidine alkaloids in real herb samples

Extracts from different peppermint, chamomile, fennel, savory, vervain and rooibos samples were analyzed using the developed targeted method. Table 2 summarizes the detected levels of the various PAs in the samples. The results underline the need for a selective and sensitive analytical method, since the different samples contained the enantiomeric species in varying concentrations. Only the selectivity of ion mobility and thus the definite structural identification of the PAs allow the precise toxicological evaluation of the samples.

Conclusions

SelexION+ Technology is a powerful technology to significantly increase the selectivity of an MRM experiment. It is especially useful 1) to remove matrix interferences which hamper the sensitive detection or precise and accurate integration of the analytes and 2) to distinguish between isomeric compounds which cannot be separated by chromatography or mass spectrometry alone. The presented method demonstrates the sensitive quantification of 27 pyrrolizidine alkaloids, which are relevant for the analysis of herbal and tea extracts. Clear identification of these different structures can help to evaluate more precisely the toxicological relevance of the presence of PAs in food.

References

1. Risks for human health related to the presence of pyrrolizidine alkaloids in honey, tea, herbal infusions and food supplements. EFSA CONTAM Panel (EFSA Panel on Contaminants in the Food Chain), EFSA Journal (2017) 15(7): 4908.
2. Aktualisierte Risikobewertung zu Gehalten an 1,2-ungesättigten Pyrrolizidinalkaloiden (PA) in Lebensmitteln. Stellungnahme Nr. 020/2018 des BfR vom 14. Juni 2018.
3. Bestimmung von PA in Pflanzenmaterial mittels SPE-LC-MS/MS, Bundesinstitut für Risikobewertung, BfR-PA-Tee-2.0/2014.
4. Vorläufige Empfehlungen des BfR zur Analytik von Pyrrolizidinalkaloiden (PA) in Kräutertee und Tee (Analysespektrum und Probenahme-verfahren), Mittteilung Nr. 002/2016, BfR, 5. January 2016.
5. Dzuman et al. (2020) Development of a new LC-MS method for accurate and sensitive determination of 33 pyrrolizidine and 21 tropane alkaloids in plant-based food matrices. Anal. Bioanal. Chem. 412, 7155.
6. An accurate and sensitive method for the quantification of 33 pyrrolizidine and 21 tropane alkaloids in plant-based food matrices. SCIEX technical note RUO-MKT-02-12768-A.
   
7. SelexION^® Technology: The solution to selectivity challenges in quantitative analysis, SCIEX technical note RUO-MKT-02-3251-A.

Acknowledgements

The authors from SCIEX would like to thank the Central Laboratory COOP in Basel for providing the samples analyzed in this study, especially Benjamin Hölzle for the sample preparation and Dr. Jürg Cabalzar from SCIEX for careful review of the paper.