Analysis of the Mushroom Nephrotoxin Orellanine and Its Glucosides

Article pubs.acs.org/jnp

Analysis of the Mushroom Nephrotoxin Orellanine and Its Glucosides Anders Herrmann,*,† Heidi Hedman,‡ Johan Rosén,† Daniel Jansson,§ Börje Haraldsson,‡ and Karl-Erik Hellenas̈ † †

National Food Agency, Box 622, SE-751 26 Uppsala, Sweden Institute of Medicine, University of Gothenburg, SE-413 45 Gothenburg, Sweden § Swedish Defence Research Agency, FOI CBRN Defence and Security, SE-901 82 Umeå, Sweden ‡

ABSTRACT: Orellanine is a nephrotoxin found in various Cortinaceae mushroom species. Unintentional consumption after these species were confused with edible mushrooms such as Cantharellus tubaeformis has caused several casualties. In this work, a quantitative HPLC-ESI-MS/MS method for total orellanine in Cortinarius rubellus, spiked blood plasma, and a mushroom stew prepared from C. tubaeformis with the addition of a single specimen of C. rubellus is presented. The existence of mono- and diglucosylated orellanine in C. rubellus was also proven, although quantitative analysis could not be obtained for the glucosides due to rapid hydrolyzation to orellanine in the extract. Extraction with 3 M HCl or water mainly yielded orellanine, while MeOH or acidified MeOH mainly extracted mono- and diglucosylated orellanine. The highest recovery of total orellanine was obtained with 3 M HCl, which was subsequently used for quantitative analysis. A C18 HPLC column and low pH in the eluents retained all these toxins. Orellanine could be detected at a 4.9 ng/mL level in all extracts, which is well below the threshold for acute toxic effects. Additionally, the fragmentation pattern of orellanine upon electrospray MS/MS was probed. The method described is useful for two important applications. First, it allows quantitative analysis of processed food products that may be contaminated by orellanine from Cortinaceae mushrooms. Second, orellanine is currently being evaluated as a potential cure of metastatic renal cancer, and this work provides a method for monitoring orellanine at low concentrations within the therapeutic interval in blood serum.

O

identity of these mushrooms with edible mushrooms such as Cantharellus tubaeformis and Cantharellus cibarius.1−7 The latency period from intake until symptoms of acute renal failure occur can vary from 2 to 17 days. Common symptoms such as vomiting, nausea, and oliguria occur as signs of uremia.8,9 The mode of action involves both induction of free radicals10−12 and eradication of the intracellular defense mechanisms.11 The diagnosis of orellanine poisoning can be made by chemical and mycological testing or by toxicological analysis of a renal biopsy specimen. There is currently no cure for orellanine poisoning, and the final outcome mainly depends on the amount of toxin ingested and an individual's sensitivity. The toxin is widespread in the Cortinarius genus and has been identified in 34 species.13 The first paper on orellanine poisoning was published in 1957, reporting 102 cases in which 12 people died.14 It took another five years before the active compound was isolated and its structure was determined.15 There has been debate concerning the structure of orellanine,16 but after the first synthesis17 and crystal structure analysis18 there is now consensus.

rellanine (Figure 1A) is a highly nephrotoxic bipyridine N-dioxide found in various mushrooms in the Cortinariaceae family, including fool’s webcap (Cortinarius orellanus) and deadly webcap (Cortinarius rubellus, formerly named C. speciosissimus). These two species are regarded as two of the world’s most poisonous mushrooms and have claimed several lives in Europe and North America after confusion of the

Figure 1. Structure of (A) orellanine (3,3′,4,4′-tetrahydroxy-2,2′bipyridine-N,N′-dioxide) and (B) orellanine-4,4′-diglucopyranoside. © XXXX American Chemical Society and American Society of Pharmacognosy

Received: February 21, 2012

A

dx.doi.org/10.1021/np300135k | J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products

Article

Table 1. QTOF-MS/MS Analysis of Diglucosylated Orellanine and Orellanine Together with Proposed Ion Formulas and Molecular Speciesa m/z measured d

577.1516 415.0986e 253.0457d 236.0429f 219.0402f 191.0450f 163.0500f

proposed ion molecular formulab

corresponding theoretical m/z

proposed molecular species

theoretical molecular weightc

C22H29N2O16 C16H19N2O11 C10H9N2O6 C10H8N2O5 C10H7N2O4 C9H7N2O3 C8H7N2O2

577.1512 415.0983 253.0455 236.0428 219.0400 191.0451 163.0502

diglucosylated orellanine monoglucosylated orellanine orellanine loss of OH loss of 2OH loss of 2OH and CO loss of 2OH and 2CO

576.1437 414.0908 252.0379

a

All molecular weights are monoisotopic. bSee Experimental Section for details. cThe theoretical molecular weights were calculated by assuming that the proposed ion molecular formulas are [M + H]+ ions. dPrecursor ion. eProduct ion from diglucosylated orellanine. fProduct ion from orellanine.

A reliable, sensitive, and simplified method for extraction and quantification of orellanine is needed for two main purposes. First, it is needed for the analysis of mushroom material and complex foods by food regulatory agencies within their preventive food safety monitoring, as well as for clinical or forensic investigations of suspected poisonings. It could also be useful to quantitate the amount of orellanine in body fluids or tissue after poisoning, although patients will normally not develop any symptoms until after a few days, at which time the toxin is found only in the kidneys. Second, orellanine is currently being tested as a potential treatment for metastatic renal cancer based on its highly selective toxicity to renal cells.19 Within a few years, orellanine may be used in clinical trials to treat patients with metastatic renal cancer. In these patients, it will be necessary to quantify the amount of orellanine to keep the concentrations within therapeutic range. Several methods for isolation and analysis of orellanine have been published in which the complicated chromatographic behavior together with poor solubility and stability in solution have been frequently discussed.10,20−22 These difficulties are associated with its acid−base properties: the toxin has two ring systems, each of which contains three hydroxyl groups, with pK values of 1.5, 5.8, and 11.0.12 Another complicating factor is that orellanine occurs mainly as 4,4′-diglucopyranoside, at least in C. rubellus and C. orellanus (Figure 1B).22 Isolation of orellanine from Cortinarius species has been mainly performed using methanolic solvents and different workup procedures such as removal of fatty mushroom material with organic solvents10 and solid-phase extraction.22 The toxin has been analyzed using TLC,13 electrophoresis,13 and HPLC with photodiode array,21 electrochemical,20 and UV10 detection. In the latter two methods, reversed-phase ion-pair HPLC with phosphate-containing eluents20 and amide and C18 columns with phosphoric acid as the eluent10 were employed. Orellanine has been detected in single MS mode using electron impact (EIMS)20,23 and electrospray ionization (ESIMS).21 HPLCESIMS/MS has also been used but only for qualitative analysis of the diglucoside of orellanine.22 In the present work, optimal extraction and chromatographic procedures and the electrospray MS/MS fragmentation pattern of orellanine and its naturally occurring glucosides have been investigated. The method was used to analyze orellanine in C. rubellus, spiked blood plasma, and a stew prepared from C. tubaeformis spiked with one specimen of C. rubellus.

dominant ion. In electrospray MS/MS (0 to 45 eV), product ions with m/z 163, 191, 219, and 236 were most abundant. These product ions were also analyzed by high-resolution QTOF-MS/MS. The product ions 236 and 219 corresponded to the loss of one and two OH groups, respectively, and 191 and 163 to the loss of two OH groups and one and two CO groups, respectively (Table 1 and Figure 2A). The correspond-

RESULTS AND DISCUSSION MS Method Development. Using a triple quadrupole instrument, a standard of orellanine was analyzed in full-scan mode (m/z 200−700) where [M + H]+ (m/z 253) was the

Figure 2. QTOF-MS/MS of (A) orellanine and (B) diglucosylated orellanine and proposed fragmentation pattern. Orellanine was analyzed by direct infusion using a syringe pump (10 μg/mL), while the diglucoside was analyzed by HPLC-MS/MS (C. rubellus extract, elution time 1.1 min).

■

B



Article

ing [M•]+ ions have previously been identified in EIMS analysis of orellanine,23 and it was suggested that these product ions were formed after N−O ring closure and elimination of hydroxyl radicals followed by ring rearrangement through phenolic CO eliminations. The high level of similarity in the MS/MS spectra generated using electrospray and EIMS/MS indicates a common fragmentation pattern, although these two fragmentation techniques will not necessarily produce an identical fragmentation pattern. The transitions 253 > 191 (quantifier), 253 > 219, and 253 > 163 (qualifiers) were selected for the method used in the remaining triple quadrupole HPLC-MS/MS experiments. HPLC Method. Optimum chromatographic performance was obtained with an Agilent Eclipse Plus C18 column using ammonium formate (pH 2.5) and MeOH with formic acid (0.2%) as eluents (retention factor 5.3) (Figure 3). Under these

1.2. Besides these columns, ZIC-HILIC, Hypercarb, and polymeric Shodex columns were tested, but they were all unable to retain or produce a chromatographic peak of orellanine despite testing a range of buffer salts, pH-values, and solvents. Figure 3 shows chromatograms of orellanine injected into the evaluated columns. Extraction and Identification of Orellanine and Its Glucosides. C. rubellus was subjected to extraction with water, 3 M HCl, acidified MeOH, and MeOH in four separate experiments. After 3 h of extraction, the extracts were analyzed by quadrupole full-scan MS (m/z 200−700). In both the MeOH and acidified MeOH extracts, one small peak corresponding to orellanine eluted at 3.1 min, and another two groups of unresolved peaks at 1.1−1.8 and 2.1−2.5 min, respectively (Figure 4A). The latter two had m/z values of 415 and 577, which correspond to mono- and diglucosylated orellanine. The molecular formulas of these ions were determined by their accurate mass and isotope distributions using TOF-MS and corresponded to these glucosides (Table 1). Furthermore, QTOF-MS/MS of these two toxins showed that diglucosylated orellanine readily fragmented into both monoglucosylated orellanine and orellanine, while monoglucosylated orellanine fragmented into orellanine (Figure 2B). Additionally, all major product ions of orellanine were also formed from both mono- and diglucosylated orellanine, which further confirmed their identities (Figure 2A and B). The MS/ MS spectra of these peaks were identical over the whole cluster of signals for the respective glucoside. Whether these unresolved peaks correspond to α- and β-anomers, glycosides with different isomeric sugar identities (e.g., galactose or glucose), or other types of isomers was not determined in this work. Previous NMR studies suggested that the diglucoside isolated from C. rubellus and C. orellanus (MeOH) was a βglucoside, but no chromatogram from the preparative HPLC procedure used in that work was presented, and it was not mentioned whether the glucoside eluted as one or multiple peaks.22 However, in earlier published HPLC-UV (a C18 column with ion-pair agents in the eluent) chromatograms of methanolic C. orellanus extracts, two groups of peaks eluted before orellanine with a similar unresolved character to the peaks of the glucosides observed in the present study.10 The identities of the compounds responding to these peaks were not determined and were merely referred to as “other constituents of the mushroom extracts”. Extraction using acidified MeOH (3 h) yielded monoglucosylated (63% of the total toxin content) and diglucosylated (25%) orellanine and a low amount of orellanine (12%) (Figure 4B). MeOH (3 h) yielded even less orellanine (2%) and about half the amount of total toxin as acidified MeOH. In both these extractions, the total toxin content did not increase even after another 23 h of extraction. However, the distribution between the three toxins had changed over time in the MeOH−3 M HCl (10:1) extract but not in the MeOH extraction (data not shown). Monitoring the toxin content in the supernatant of the centrifuged mushroom extracts (MeOH−3 M HCl (10:1)) showed that the glucosides hydrolyze over time, and after 72 h the percentage of orellanine was 80%, the percentage of the monoglucoside was 19%, and the percentage of the diglucoside was only 1% (Figure 4C). In the MeOH extract, virtually no hydrolyzation was observed. Another important observation in the MeOH−3 M HCl (10:1) extraction is that the summarized peak areas for all three forms of toxins did not change over the hydrolyzation process, which

Figure 3. MRM traces (253 > 191) of orellanine in 3 M HCl (1 μg/ mL) with retention times using six columns.

conditions, orellanine eluted a single peak and with low relative standard deviation (RSD) of peak areas for four subsequent injections (1.2%). A higher pH (4.0) in eluent A resulted in poor retention and peak shape, while a lower pH (2.0) decreased the MS signal without improvement in retention. The Hypersil C18 column successfully retained orellanine but with a lower retention factor (3.6) than the Agilent Eclipse column. The BEH-HILIC column with ammonium acetate (pH 3) and acetonitrile as eluents yielded a retention factor as low as C



Article

Figure 4. Extraction of orellanine and its glucosides from dried C. rubellus mushroom. (A) Summarized extracted ion chromatograms (from full-scan MS) of diglucosylated (m/z 577), monoglucosylated (m/z 415), and glucose-free (m/z 253) orellanine after 3 h of extraction. (B) Content and distribution of orellanine and its glucosides in extracts after extraction using different extraction solvents (3 h). (C) Monitoring of toxin distribution in the centrifuged mushroom extract (MeOH−3 M HCl (10:1)) over 72 h. (D) Monitoring of the toxin distribution during 26 h of extraction with water.

M HCl extracts of spiked blood plasma, C. rubellus, and a mushroom stew spiked with one specimen of C. rubellus, it was shown that 60 min was sufficient time for extraction (Figure 5). Extraction of orellanine with 3 M HCl followed by analysis with no further workup meant that the relatively highly abundant coextracted components could coelute with orellanine, generate ion suppression effects, and hence lower the sensitivity. The slopes of the standard curves of orellanine in 3 M HCl and the 3 M HCl extracts were used to estimate the degree of these effects. Compared to orellanine dissolved in 3 M HCl, the detector response was 60% when the toxin was dissolved in plasma extracts. Co-extracted compounds from C. tubaeformis and the mushroom stew affected the MS response only slightly ( 163, 253 > 191, and 253

Figure 5. Orellanine concentration over time during a 3 M HCl extraction. A small portion of the extract ( 219. The mono- and diglucosides of orellanine were detected in positive ESI full-scan mode (m/z 200−700) by their [M + H]+ of 415 and 577, respectively. A range of columns were tested. When using the Agilent Eclipse Plus C18, Hypersil C18, Hypercarb, and Shodex columns, 4 mM ammonium formate (pH 2.0, 2.5, 3.0, 3.5, or 4.0) and MeOH with formic acid (0.2%) were used as eluents. For the BEH-HILIC and ZIC-HILIC columns, the eluents consisted of 1 mM ammonium acetate (pH 2.0, 2.5, 3.0, 3.5, or 4.0) and acetonitrile. In the final method, the Agilent Eclipse C18 column was used, where eluent A consisted of 4 mM ammonium formate (pH 2.5) and eluent B consisted of MeOH with formic acid (0.2%). Orellanine and its glucosides were eluted using a linear gradient: 0 min: 3% B, 3 min: 20% B, 5−7 min: 95% B. The flow was 0.2 mL/min, the column temperature was set to 35 °C, and the injection volume was 5 μL. HPLC-QTOF-MS. The HPLC-QTOF-MS experiments were performed on a Dionex UltiMate 3000 LC coupled to a Bruker Daltonics MaXis UHR-TOF (mass resolution 20 000). The drying gas temperature was set to 190 °C, and the capillary voltage was set to 4.5 kV. MS data (m/z 100−800) and MS/MS data (m/z 80−600) were acquired in ESI positive mode with an acquisition rate of 4 Hz. The collision energies were alternated between 0 and 45 eV. Mass spectra were internally calibrated with a sodium formate solution with imidazole (25 mL water−25 mL 2-propanol−0.5 mL 1 M NaOH−50 μL formic acid−10 mg imidazole). Methyl stearate (1 mg/mL in 2propanol, m/z 299.294457) was used for lock-mass calibration. All relevant features were extracted in the Compass DataAnalysis 4.0 processing software. The molecular formulas were generated by SmartFormulaManually using the following settings: tolerance: 2 mDa, max H/C: 3, electron configuration: odd or even. The formulas with the highest score (based on mass accuracy and isotopic distribution) were selected. The errors in mass accuracy were below 2.2 ppm, and the errors in isotopic pattern (milli-sigma) below 30. The experiments were performed in nonfocus mode with a resolution of 200 000. The isolation width for MS/MS was 10. Orellanine was analyzed by direct infusion of orellanine (10 μg/mL in 3 M HCl) using a KD Scientific syringe pump with a Hamilton 500 μL syringe at 180 μL/h. The glucosides of orellanine (C. rubellus extract) were analyzed through HPLC-MS experiments, where eluent A consisted of 5 mM ammonium formate−MeOH (90:10) and formic acid (0.02%) and eluent B consisted of 5 mM ammonium formate in MeOH with formic acid (0.02%). Mono- and diglucosylated orellanine were eluted using a linear gradient: 0 min: 1% B, 3 min: 39% B, 14 min: 99.9% B, 16 min: 99.9% B. The flow was 0.2 mL/min, the column temperature was set to 30 °C, and the injection volume was 2 μL. Preparation of Mushroom Stew. Blank mushroom stew was prepared with 10 g of dried C. tubaeformis, 7.5 g of butter, 15 g of flour, and 100 mL of milk (3% fat). Another mushroom stew was prepared in the same manner but with the addition of 1 g of dried C. rubellus (one dried specimen). The mushrooms were soaked in 10 mL of water for 15 min before use. The prepared stews were weighed to 143 and 150 g, respectively, and homogenized before extraction. Extraction Procedure for Total Orellanine. In four different experiments, 80 mg of C. rubellus was subjected to extraction with 8 mL of water, 3 M HCl, MeOH, and MeOH−3 M HCl (10:1) for 3 h in a 15 mL test tube and then centrifuged (10 min, 3600g). As 3 M HCl extracted total orellanine most efficiently, this solvent was used for the application experiments as described below. Rat blood plasma (20 μL) spiked with orellanine was mixed with 180 μL of 3 M HCl in a 1.5 mL Eppendorfer tube, incubated for 1 h on a shaking table at room temperature and in the dark followed by centrifugation (5 min, 8000g). In two separate experiments, 1 and 10 g of the each mushroom stew were transferred to 50 mL test tubes, extracted with 4 and 40 mL of 3 M HCl for 60 min, centrifuged (10 min, 3600g), and filtered using centrifuge tube filters (10 min, 16000g). The extraction was performed in triplicate. For determination of the total orellanine content in C. rubellus and spiked C. tubaeformis, 80 mg of mushroom material was subjected to extraction

using 8 mL of 3 M HCl for 3 h in a 15 mL test tube and then centrifuged (10 min, 3600g). Determination of Extraction Recoveries, Ion Suppression Effects, and Limits of Detection. In three separate experiments, blood plasma was spiked with orellanine at 1, 5, and 20 μg/mL, including a blank sample, and extracted as described above. After the extraction, the blank sample was spiked with orellanine to 5000 ng/ mL, from which a 7-fold 1:4 dilution series down to 1.2 ng/mL was constructed and used as a standard curve. For C. tubaeformis, the same extraction procedure and standard curve were employed with 80 mg of dried mushroom with spiking levels of 0.02, 0.2, and 2 mg/g mushroom sample. As a solvent blank, 3 M HCl was spiked with orellanine to 5 μg/mL, and the same standard curve was constructed as for blood plasma and C. tubaeformis. The recoveries were determined from the standard curves, and the ion suppression effects estimated by comparing their slope to that of the solvent blank. The limit of detection was defined at the concentration where the signal-tonoise value exceeded 3.0. All extractions were performed in triplicate. Determination of Orellanine Content in C. rubellus and Spiked Mushroom Stew. The orellanine content in C. rubellus was determined by a calibration curve from spiked C. tubaeformis extracts (4.9−5000 ng/mL, 1:4 dilution). The C. rubellus extract was diluted 80 times before the HPLC-MS/MS analysis. The orellanine content in the spiked mushroom stew was determined in the same manner using a matrix-matched calibration curve from spiked C. tubaeformis stew extract. The spiked stew extract was diluted 100 times before HPLCMS/MS analysis. The extractions were performed in triplicate.

■

AUTHOR INFORMATION

Corresponding Author

* Tel: +4618171479. Fax: +4618105848. E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■ ■

ACKNOWLEDGMENTS Financial support from the Swedish Contingencies Agency is gratefully acknowledged. REFERENCES

(1) Bouget, J.; Bousser, J.; Pats, B.; Ramee, M. P.; Chevet, D.; Rifle, G.; Giudicelli, C. P.; Thomas, R. Intensive Care Med. 1990, 16, 506−10. (2) Calvino, J.; Romero, R.; Pintos, E.; Novoa, D.; Guimil, D.; Cordal, T.; Mardaras, J.; Arcocha, V.; Lens, X. M.; Sanchez-Guisande, D. Am. J. Nephrol. 1998, 18, 565−9. (3) Holmdahl, J.; Mulec, H.; Ahlmen, J. Hum. Toxicol. 1984, 3, 309− 13. (4) Horn, S.; Horina, J. H.; Krejs, G. J.; Holzer, H.; Ratschek, M. Am. J. Kidney Dis. 1997, 30, 282−6. (5) Schumacher, T.; Hoiland, K. Arch. Toxicol. 1983, 53, 87−106. (6) Short, A. I.; Watling, R.; MacDonald, M. K.; Robson, J. S. Lancet 1980, 2, 942−4. (7) Wörnle, M.; Angstwurm, M. W.; Sitter, T. Am. J. Kidney Dis. 2004, 43. (8) Danel, V. C.; Savic, P. F.; Garon, D. Toxicon 2001, 39, 1053−60. (9) Prast, H.; Werner, E. R.; Pfaller, W.; Moser, M. Arch. Toxicol. 1988, 62, 81−8. (10) Cantin, D.; Richard, J. M.; Alary, J. J. Chromatogr. 1989, 478, 231−7. (11) Nilsson, U. A.; Nyström, J.; Buvall, L.; Ebefors, K.; BjörnsonGranqvist, A.; Holmdahl, J.; Haraldsson, B. Free Radical Biol. Med. 2008, 44, 1562−9. (12) Richard, J. M.; Cantin-Esnault, D.; Jeunet, A. Free Radical Biol. Med. 1995, 19, 417−29. (13) Oubrahim, H.; Richard, J. M.; Cantin-Esnault, D.; SeigleMurandi, F.; Trecourt, F. J. Chromatogr. A. 1997, 758, 145−57. (14) Grzymala, S. Z. Pilzkunde 1957, 23, 139−142. (15) Grzymala, S. Bull. Soc. Mycol. Fr. 1962, 78, 394. F



Article

(16) Kürnsteiner, H.; Moser, M. Mycopathologia 1981, 74, 65−72. (17) Dehmlow, E. V.; Schulz, H.-J. Tetrahedron Lett. 1985, 26, 4903− 4906. (18) Antkowiak, W. Z.; Gessner, W. P. Tetrahedron Lett. 1979, 20, 1931. (19) Hedman, H.; Buvall, L.; Nilsson, U.; Johansson, M.; Nyström, J.; Haraldsson, B. NC-001 induces apoptosis and necrosis in clear cell renal cell carcinoma in vitro and in a xenograft model in vivo. In LB118, AACR annual meeting in Chicago, Chicago, 2012. Available at http://www.abstractsonline.com/Plan/ViewAbstract.aspx?sKey= c8ad5867-cb39-4be6-a181-346c764cb755&cKey=d2eb8eb5-e8944e50-a0f7-f3f1fe0a4b3e&mKey=%7B2D8C569E-B72C-4E7D-AB3B070BEC7EB280%7D. (20) Holmdahl, J.; Ahlmen, J.; Bergek, S.; Lundberg, S.; Persson, S. A. Toxicon 1987, 25, 195−9. (21) Koller, G. E.; Høiland, K.; Janak, K.; Størmer, F. C. Mycologia 2002, 94, 752−6. (22) Spiteller, P.; Spiteller, M.; Steglich, W. Angew. Chem., Int. Ed. 2003, 42, 2864−7. (23) Antkowiak, Z.; Antkowiak, R.; Wyrzykiewicz, E.; Czerwinski, G. Heterocycles 1994, 39, 477−484. (24) Prast, H.; Pfaller, W. Arch. Toxicol. 1988, 62, 89−96. (25) Richard, J. M.; Louis, J.; Cantin, D. Arch. Toxicol. 1988, 62, 242−5. (26) Delpech, N.; Rapior, S.; Cozette, A. P.; Ortiz, J. P.; Donnadieu, P.; Andary, C.; Huchard, G. Presse Med. 1990, 19, 122−4.

G


Analysis of the Mushroom Nephrotoxin Orellanine and Its Glucosides

Recommend Documents