Identification of sinicuichi alkaloids in human serum after intoxication caused by oral intake of a Heimia salicifolia extract

Abstract

A 26-year-old male came to hospital around midnight complaining about muscle pain of the extremities as well as the tongue and slightly raised temperature. He reported the intake of an unknown amount of sinicuichi tea he had fermented over 24 h by adding yeast and sugar. The patient was treated with Vomex A1 (dimenhydrinate) and released from hospital the following afternoon. A blood sample taken shortly after submission and a small amount of the used plant material were available for analysis. Herbal drugs are widely used as stimulants as a legal alternative to illegal psychoactive drugs or in traditional context. Among many others like Sassafras officinalis, Salvia divinorum [1] or Ephedra [2], Heimia salicifolia (‘‘sinicuichi’’), a species of the lythraceae family, is available via several online shops. Brewed up or fermented and consumed, the so-called sinicuichi tea may cause exhilarating feelings and an alteration of awareness accompanied by bradycardia, relaxation of the muscles and a pleasant faintness. Therefore Sinicuichi brew and heimia leaves are widely used for medication by the natives of Central and South America. After liquid extraction with acetone five different alkaloids were detected in the plant material by LC–MS/MS operated in the Q1 scan mode applying a TurboIonSpray source. Subsequently, Product Ion Spectra were recorded and after confirming the molecular formula by determining the accurate masses, possible structures of H. salicifolia alkaloids were assigned. The information of the Product Ion Spectra was then used to set up a sensitive multiple reaction monitoring (MRM) method. Applying the MRM method to the patient’s serum sample after alkaline liquid–liquid extraction all of the five heimia alkaloids detected in the plant material were also detected qualitatively in the serum extract, confirming the ingestion. # 2008 Elsevier Ireland Ltd. All rights reserved. Keywords: Sinicuichi; Herbal drugs; LC–MS/MS

1. Introduction

Sinicuichi tea is made of the leaves of Heimia salicifolia, a species of the lythraceae family basically found in Central and South America. Dried plant material is sold as sinicuichi or ‘‘Sun Opener’’ by several internet shops. Brewed up or fermented with water, yeast and sugar the so-called sinicuichi tea is believed to cause exhilarating feelings and an alteration of awareness accompanied by bradycardia, relaxation of the muscles and a pleasant faintness. It is also said to be an extreme mood enhancer and provides visual stimulation, hence the name ‘‘Sun Opener’’. Dating back to the Aztecs, H. salicifolia was used during shamanic rituals and till this day Sinicuichi brew and heimia leaves are widely used for medication by the natives of Central and South America. The species of the lythraceae family are known to contain a number of biphenyl quinolizidine lactone alkaloids: vertine, sinicuichine, anelisine, heimidine, lythrine, nesodine, lythridine, lyfoline, dehydrocodine, demethyllasubine I, demethyllasubine II, demethoxyabresoline and abresoline [3]. Among them vertine, lyfoline, lythrine and nesodine have been investigated best. Vertine (cryogenine), the alkaloid showing the highest concentration in the plant material, is considered to be the primary source of the effects of traditional H. salicifolia brew like anticholinergic, anti-inflammatory, sedative, tranquilizing and spasmolytic activity. Further detailed informations on H. salicifolia are given by Malone and Rother [3]. After the intake of an unknown amount of sinicuichi tea, fermented over 24 h by adding yeast and sugar, a 26-year-old male came to hospital around midnight complaining about muscle pain of the extremities and the tongue as well as slightly raised temperature. In the evening and during the night nausea, headache, singular vomitus and ague occurred. C-reactive protein and leukocyte levels were at normal range while creatinine kinase was slightly increased. He was treated with dimenhydrinate (Vomex A1) and released from hospital the following afternoon after complete recovery. A serum sample taken shortly after admission and a small amount of the used plant material were available for analysis.

2. Materials and methods

2.1. Materials

All chemicals were of analytical or HPLC grade: formic acid (98–100%), water, acetone, 1-chlorobutane, acetonitrile (HPLC grade), (Merck, Darmstadt, Germany), ammonium formate (Sigma–Aldrich, Steinheim, Germany). For the purpose of comparison, two Sinicuichi specimens were purchased from two different online shops, selling sinicuichi as a ‘legal herbal drug’. Serum samples were stored at 20 8C prior to preparation and analysis. The plant material was stored at ambient temperature in a dark and dry place.

2.2. Sample preparation

The received plant material was comminuted in a motar and 1 g of the powder was extracted with 10 mL acetone at ambient temperature for 24 h. For analysis, 10 mL of this solution were diluted with 990 mL eluent. The patient’s serum was extracted using alkaline chlorobutane liquid–liquid extraction [4]. 0.5 mL borate buffer (pH 9) and 1.5 mL 1-chlorobutane were added to 1 mL of the patient’s serum. After 3 min of mixing and 4 min of centrifugation, the organic phase was transferred to a HPLC-Vial and dried under a gentle flow of nitrogen at 40 8C. For analysis the residue was dissolved in 100 mL eluent.

2.3. Instrumentation

The LC–MS/MS system consisted of an API 365 triple-quadrupole mass spectrometer fitted with a TurboIonSpray interface (Applied Biosystems/Sciex, Darmstadt, Germany) and a Shimadzu HPLC system (two pumps LC10AD Shimadzu, Duisburg, Germany). Separation was performed using a polarendcapped phenylpropyl reversed phase column (Synergy Polar-RP 50 mm  2 mm, 4 mm) with an equivalent guard column (4 mm  2 mm) (Phenomenex, Aschaffenburg, Germany). All analyses were performed using solvent A (0.1% formic acid (v/v) with 1 mmol/L ammonium formate) and solvent B (acetonitrile: 0.1% formic acid 95:5 (v/v) with 1 mmol/L ammonium formate). Twenty microliters of the prepared sample were injected. The initial Q1 screening (Q1 scan) was performed using the following 28 min gradient: 0–1 min: 5% B; 1–5 min: 5–30% B linear; 5–15 min: 30–70% B linear; 15–19 min: 70–95% B linear; 19–22 min: 95% B; 22–24 min: 95–5% B linear; 24–28 min: 5% B. The total flow rate was set to 0.25 mL/min. For the selection of the protonated molecular ions as precursor ions, a single-quadrupole mass spectrum was acquired in scan mode using three different declustering potentials (20, 50, 80 V) [5]. Product Ion Spectra (PIS) of the investigated substances were measured using a short 10 min gradient (0–0.5 min: 5% B; 0.5–7.5 min: 5–95% B linear; 7.5–8 min: 95% B; 8–9 min: 95–5% B linear; 9–10 min: 5% B). Three MS/MS spectra with collision energies of 20, 35 and 50 eV were recorded for each alkaloid [6]. For multi-reaction monitoring (MRM) the Q1 scan gradient was shortened to 15 min retaining the flow rate of 0.25 mL/min (0–0.5 min: 5% B; 0.5–10 min: 5–95% B linear; 10–12 min: 95% B; 12–12.5 min: 95–5% B linear; 12.5– 15 min: 5% B for re-equilibration). The MRM method was set up using the transitions of the respective [M+H]+ to m/z 84 and the non-fragmented precursor ions ([M+H]+ ) (Table 1). Accurate masses were determined using a quadrupole time-of-flight mass spectrometer (micrOTOFQ, Bruker Daltonik, Bremen, Germany) with positive electrospray ionisation (inlet 4500 V, sprayer on ground). The quadrupole was operated in ‘‘RF-only modus’’ to obtain good transfer of ions in the required mass range, i.e. the TOF was used to acquire complete spectra. Standard source parameters for LC coupling were applied, the nebulizer pressure was set to 1 bar, the drying gas set to 200 8C at a flow rate of 6 L/min. Separation of the analytes was performed by a Dionex 3000 capillary HPLC-System (Dionex Corporation, Sunnyvale, USA). To maintain comparable retention times, LCconditions were used as described for the final MRM method.

3. Results and discussion Five m/z values, matching the molecular ions of the 14 common heimia alkaloids were detected in the plant extract by our standard Q1 screening (Fig. 1). According to the literature [3], the detected m/z values can be associated to the following substances: m/z 278: demethyllasubine I and demethyllasubine II; m/z 422: lyfoline, dehydrocodine and anelisine; m/z 424: desmethoxyabresoline and 10-epi-desmethoxyabresoline; m/z 436: sinicuichine, vertine, nesodine and lythrine; m/z 454: lythridine, heimidine and abresoline, where lyfoline and vertine should be the most abundant alkaloids in the plant extracts. Product Ion Spectra of the particular precursors showed poor fragmentation for all alkaloids. The only high abundant fragment (m/z 84) appearing in every spectrum, was identified as the 2,3,4,5-tetrahydropyridinium cation (Fig. 2). To ease further description we grouped the 14 alkaloids according to their molecular masses. The Product Ion Spectrum of each precursor and the possible corresponding structures were shown in Fig. 3. Using LC–MS/MS a clear identification of the alkaloids was not possible due to the poor fragmentation behaviour. Because pure reference standards were not available the exact structures could not be determined by NMR either. However, due to the molecular weight information in addition to a transition typical for the structures of these alkaloids it can strongly be assumed, that transitions of the specified precursors belong to the alkaloids listed above. An MRM of the patient’s plant material was recorded and compared to the MRM of sinicuichi which had been purchased from two different online shops. The three plant samples were treated as specified above. The five groups of alkaloids were detected in the patient’s plant specimen as well as in one of the purchased sinicuichi specimen. However, no alkaloids could be found in the second sinicuichi specimen purchased in the online shop. The patient’s serum, as well as a spiked serum sample (heimia extract, diluted 1:100 (v/v) with blank serum) were analyzed qualitatively with this LC–MS/MS procedure. Fig. 4 shows the total MRM ion chromatogram (A) of the patient’s serum extract and the extracted ion chromatograms (XICs) (B– F) of the respective non-fragmented [M+H]+ and the transitions of [M+H]+ to m/z 84 of the five types of heimia alkaloids shown above. For comparison the MRM of the spiked serum sample is shown in Fig. 5. All of the five heimia alkaloids were detected in the patient’s serum, providing evidence for the intake of plant extract (Figs. 4 and 5). Varying intensity levels as detected in the patient’s serum and the spiked serum may have occurred due to differences in the pharmacokinetics of the particular alkaloids in the body. The analysis of a blank serum sample showed no interference within the relevant retention time windows. As no reference material was available, suppression effects could not be determined by post column infusion. Therefore, six serum samples were spiked with 20 mL of plant extract after liquid–liquid extraction and the areas of the particular transitions were compared to the results of a matrix free sample with the same concentration. Only minor matrix effects could be observed as a slight ion suppression. The assumed molecular formulas of the precursors were confirmed by measuring accurate masses and verifying the correct isotopic patterns (SigmaFit) using a high resolution QTOF MS with a deviation better than 10 ppm using external calibration (Table 2).

4. Conclusions By applying the presented MRM method to the patient’s serum sample after alkaline liquid–liquid extraction all of the five types of heimia alkaloids detected in the plant material were also detected qualitatively in the serum extract confirming the ingestion, although the symptoms of the patient (muscle pains, fever) were not consistent with the expected clinical findings, as H. salicifolia should rather have sedative and spasmolytic effects in therapeutic doses. The three investigated sinicuichi specimens (obtained from three different sources) showed highly varying alkaloid contents, potentially leading to intoxications of even experienced herbal drug users. Due to the poor fragmentation behaviour and the variety of isomers, complete structural identification of the alkaloids by LC–MS/MS was not possible. However, high-resolution MS (TOF) confirmed the molecular formula of the expected alkaloids. The reported procedure can be used as a general approach for qualitative analysis in cases of possible intoxications with herbal drug alkaloids.