Occurrence and chemotaxonomical analysis of amatoxins in Lepiota spp. (Agaricales) Author Sarawi, Sepas * & Mycology Working Group, Institute of Ecology, Evolution, and Diversity, Goethe University Frankfurt, Max-von-Laue-Str. 13, 60438, Frankfurt am Main, Germany Author Shi, Yan-Ni Author Lotz-Winter, Hermine Author Reschke, Kai Author Bode, Helge B. Author Piepenbring, Meike text Phytochemistry 2022 113069 2022-03-31 195 1 8 http://dx.doi.org/10.1016/j.phytochem.2021.113069 journal article 10.1016/j.phytochem.2021.113069 1873-3700 8235405 2.2. Distribution and identification of amatoxins in Lepiota spp HPLC profiles obtained from extractions of the benchmark A . phalloides with methanol or with MeOH:H 2 O:0.01 M HCl (5:4:1) showed significant differences concerning the extraction efficiency ( Fig. 3 ). The solvent mixture was more efficient in extracting amatoxins and phallotoxins than methanol alone, and mostly avoids extracting unrelated components with low polarity (peaks with retention times between 11 and 14 min). The amatoxins and phallotoxins had similar maximal UV absorption wavelengths of around 250 nm , 295 nm , and 305 nm ( Sgambelluri et al., 2014 ). By comparing the UV absorption values, HRESIMS signals in the positive mode, and molecular formula with published data, β- amanitin ( 1 ), α- amanitin ( 2 ), γ- amanitin ( 3 ), Ɛ- amanitin ( 4 ), phallisin ( 5 ), amanin ( 6 ), phalloidin ( 7 ), phallisacin ( 8 ), phallacidin ( 9 ), and phalloin ( 10 ) were identified ( Table 2 ) ( Jansson et al., 2012 ; Sgambelluri et al., 2014 ). Peak 1 had an accurate mass of m/z 920.3387 [M+H] + , corresponding to the molecular formula C 39 H 53 N 9 O 15 S of 1 . In its MS/MS spectrum, it contained the major and diagnostic fragment ions at m/z 902.3285, 884.3166, 750.2324, 374.1558, 259.1287, 228.1336, 171.1126, 146.0588, and 143.1171 [M+H] + ( Fig. S1 ; Table S1 ). The fragment ion at m/z 902.3285 was formed by the loss of H 2 O from [β- amanitin +H] + and is the precursor of m/z 884.3166 [M+H] + which results by the loss of a further molecule of H 2 O. Fragment ions at m/z 750.2334, 374.1158, 171.1126, and 228.1336 [M+H] + were generated by the cleavages of amino bonds. The fragment ion at m/z 143.1171 [M+H] + was produced by the neutral loss of CO from the ion at m/z 171.1126 [M+H] + . Structures of fragment ions at m/z 259.1287 [M+H] + and m/z 146.0588 [M+H] + are proposed to be as shown in Fig. S2 . Peak 2 showed mass data similar to 1 , but by the signal at m/z 919.3530 [M+H] + corresponds to the molecular formula C 39 H 54 N 10 O 14 S of 2 . In its MS/MS spectrum, it showed fragment ions at m/z 901.3439, 883.3364, 749.2533, and 373.1729 [M+H] + as well as fragment ions at m/z 259.1289, 228.1346, 171.1130, 146.0596, and 143.1172 [M+H] + ( Table S1 ) similar to those of 1 ( Fig. S2 ). The differences of ions at m/z 901.3439, 883.3364, 749.2533, and 373.1729 [M+H] + in comparison to those at m/z 902.3285, 884.3166, 750.2324, and 374.1158 [M+H] + in 1 can be attributed to different R 1 residues, which are –OH for 1 and –NH 2 for 2 . Table 1 Structures of amatoxins and phallotoxins. Numbers after compound names refer to peak numbers in the UHPLC spectra (Fig. 3).

Compounds	Structure	R1	R2	R3	R4	R 5	R 6
β- amanitin ( 1 ) α- amanitin ( 2 ) γ- amanitin ( 3 ) Ɛ- amanitin ( 4 ) amanin ( 6 ) amaninamide ( 12 ) amanullin amanullinic acid proamanullin	OH NH2 NH2 OH OH NH2 NH2 OH NH2	OH OH OH OH OH OH OH OH H	OH OH OH OH OH OH H H H	OH OH H H OH OH H H H	OH OH OH OH H H OH OH OH
phallisin ( 5 ) phalloidin ( 7 )	CH3 CH3	CH3 CH3	OH OH	OH H	OH OH	OH OH
phallisacin ( 8 ) phallacidin ( 9 ) phalloin ( 10 ) prophalloin	CH(CH3)2 CH(CH3)2 CH3 CH3	COOH COOH CH3 CH3	OH OH OH H	OH H H H	OH OH OH OH	OH OH H H
phallacin	CH(CH3)2	COOH	OH	H	OH	H

Fig. 1. Fruiting bodies of Lepiota spp . ( Agaricales ) used for the present analysis of amatoxins. A L. elaiophylla (SeSa 23), scale bar: 2 cm. B L. brunneoincarnata (SeSa 4), scale bar: 3 cm. C L. felina (SeSa 7), scale bar: 1.5 cm. D L. aspera (SeSa 14), scale bar: 3 cm. E L. magnispora (S 21), scale bar: 1.8 cm. F L. cristata (SeSa 1), scale bar: 3.8 cm. G L. oreadiformis , scale bar: 4 cm. H L. fuscovinacea (SeSa 8), scale bar: 3.5 cm. Based on the MS/MS spectra of 1 and 2 , the fragmentation mechanism was tentatively proposed as shown in Fig. S2 and verified by 3 , 4 , and 6 . The fragment structures of amatoxins ( 3, 4, and 6 ) were predicted, and their corresponding ion fragments were found in their MS/ MS spectra ( Fig. S1 and S 2 ). In the fragment pathway, there are nine types of ion fragments [A] + –[I] + . Type [A] + and [B] + were gradually formed by the loss of H 2 O. Type [C] + , [D] + , [F] + , [G] + , and [H] + were formed by cleavages of amino bonds. Type [I] + was generated by the loss of CO from type [D] + . The difference of ion fragments [A] + –[C] + , [E] + , [G] + , and [H] + of amatoxins ( 1–4, and 6 ) are due to different residues R 1 , R 4 , and R 5 . The signal of [E] + is the strongest of the nine ion types and can be used for diagnosing the residue R 4 . When R 4 is –OH, the mass of [E] + is about m/z 259.1287 [M+H] + . When R 4 is –H, its mass is about m/z 243.1339 [M+H] + . The signal of ion type [H] + can be used for identifying R 5 . When R 5 is –OH, its mass is about m/z 146.0594 [M+H] + while when it is –H, its mass is about m/z 130.0642 [M+H] + . When R 4 is similar in two molecules, signals of [G] + can be used for diagnosing R 1 . In those five amatoxins, 4 and 6 possessed the same molecular formula. Even without pure standard compounds, they can be easily distinguished by comparing MS/MS spectra. When fragment ions at m/z 886.3358 [A] + , 868.3274 [B] + , 734.2515 [C] + , 171.1125 [D] + , 228.1338 [F] + , and 143.1169 [I] + (data from Ɛ- amanitin MS/MS spectra) are similar, the different fragment ions at m/z 243.1338 [E] + , 357.1774 [G] + , and 146.0592 [H] + in 4 and fragment ions at m/z 259.1286 [E] + , 374.1551 [G] + , and 130.0643 [H] + in 6 can be attributed to different residues at R 4 and R 5 . All the fragment ions and their corresponding elemental composition, which were calculated by DataAnalysis software, are summarized in Table S 1 . By comparing the retention times, UV absorptions, HRESIMS, calculated molecular formulas, and MS/MS spectra with the data obtained from the benchmark species, 2 and 3 were detected for all samples of L. subincarnata . 2 and 12 were found in all samples of L. brunneoincarnata . 12 was not detected in the benchmark sample which contained 3 with the same molecular weight and molecular formula. However, the retention time and MS/MS spectrum of 12 differs from those of 3 . Based on the proposed fragmentation pathway, nine types of fragment ions of 12 were obtained and all the corresponding ions data were found in its MS/MS spectrum ( Fig. S1 and S 2 ). The detected fragment ions at m /z 259.1289 and 130.0642 [M+H] + in 12 were different from those at m /z 243.1339 and 146.0594 [M+H] + in 3 due to differences concerning R 4 and R 5 . Besides the MS/MS spectra, the UV absorptions were also different for these two toxins. 12 possessed higher UV absorption values than 3 at 295 and at 305 nm , and 3 showed higher values at 305 nm ( Fig. 3 , Sgambelluri et al., 2014 ). 1 and 2 were found in Lepiota elaiophylla Vellinga & Huijser. Furthermore , there are two amatoxin-like compounds, D1 ( 13 ) detected in L. boudieri and D2 ( 14 ) detected in L. elaiophylla . D1 ( 13 ) was found in all L. boudieri specimens (n = 5) at m/z 920.3704 [M+H] + (calculated for C 39 H 56 N 10 O 14 S, Δppm = 1.2 ppm) which is similar to 1 , but its mass did not correspond to the masses of any known amatoxin or phallotoxin ( Table 2 ). Its calculated molecular formula probably is C 39 H 55 N 10 O 14 S and fragment ions at m/z 902.3605, 884.3166, 171.1130, and 130.0643 [M+H] + from MS/MS spectra ( Fig. S1 ) are consistent with those produced from known amatoxins. D2 ( 14 ) from one specimen of L. elaiophylla at m/z 917.3442 [M+H] + (calculated for C 39 H 53 N 10 O 14 S, Δppm = +1.6 ppm) is considered as a further amatoxin derivatives. Fragment ions at m/z 899.3362, 881.3216, 228.1339, 171.1121, 146.0573, and 143.1154 [M+H] + from MS/MS spectra ( Fig. S1 ) are consistent with those produced from known amatoxins. Therefore, these compounds are probably undescribed amatoxins. The specimen SeSa 13 identified as L. cf. boudieri , in addition to peaks corresponding to D1 ( 13 ) showed an additional compound ( 15 ) at m/z 901.4233 [M+H] + . The fragmentation pattern of this compound, however, did not correspond to an amatoxin ( Fig. S1 ). Fig. 2. Phylogenetic tree based on sequences of the ITS region of nuclear rDNA. Posterior probabilities obtained from Bayesian analysis and maximum likelihood bootstrap values are shown on the upper side of each branch and transfer bootstrap expectations are shown on the lower side. GenBank accession numbers and information on the geographic origin, if available, are shown. Sequences obtained in the context of the present study are highlighted in bold. The occurrence of amatoxins and amatoxin derivatives (D1, D2) detected by mass spectroscopy (MS) are indicated on species level by colored bars. : α- amanitin,: β- amanitin,: γ- amanitin,: amaninamide,: D1,: D2,: unknown, non-amatoxin compound,: species analyzed by MS with no amatoxins being detected. For specimen specific results see Fig. S3. No amatoxins or phallotoxins were detected in any specimen of L. aspera (Pers.) Qu´el., L. castanea Qu´el., L. clypeolaria (Bull.) P. Kumm. , L. cristata ( Bolton ) P. Kumm., L. erminea (Fr.) P. Kumm. , L. felina (Pers.) P. Karst. , L. fuscovinacea F.H. Møller & J.E. Lange , L. lilacea Bres., L. magnispora Murrill , L. oreadiformis Velen. , L. pseudolilacea Huijsman, L. sp. (SeSa 4), and L. subalba Kühner ex P.D. Orton , based on the applied methods. For specimens of L. boudieri , L. cf. boudieri (SeSa 13), L. brunneoincarnata , L. elaiophylla , and L. subincarnata no phallotoxins were detected. Fig. 3. Results of UHPLC analyses of fruiting bodies of Lepiota spp . Lepiota boudieri (S 4), L. cf. boudieri (SeSa 13), L. brunneoincarnata (SeSa 4), L. elaiophylla (S 23), and L. subincarnata (SeSa 28) extracted in (B) MeOH:H 2 O:0.01 M HCl (5:4:1). Results of UHPLC analyses of fruiting bodies of the benchmark species Amanita phalloides (S105), (A) extracted in methanol, (B) extracted in MeOH:H 2 O:0.01 M HCl (5:4:1). Results obtained by absorption of UV light of diverse wavelengths are indicated by different colours. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.) 2.3. Phylogeny and chemotaxonomy of Lepiota spp . based on amatoxins The results obtained by molecular phylogenetic and chemical analyses are combined in Fig. 2 . The phylogram shows that species of Lepiota containing known amatoxins in their fruiting bodies detected in the context of the present study belong to a single monophyletic group within the section Ovisporae . Further putative amatoxin derivatives were detected in L. boudieri (D1, 13 ) which belongs to the section Stenosporae and in L. elaiophylla (D2, 14 ) which belongs to the section Ovisporae . The specimen of L. cf. boudieri SeSa 13 contained an additional compound ( 15 ) with an amatoxin-like molecular weight. As evident by its fragmentation pattern, however, it does not correspond to an amatoxin. This compound was not detected in specimens of L. boudieri . The specimen SeSa 13 forms a distinct lineage close to other specimens identified as L. boudieri in the phylogenetic tree. Table 2 Masses and sum formulas of amatoxins and phallotoxins observed in extracts of the benchmark fungus Amanita phalloides ( 1–11 ), Lepiota brunneoincarnata ( 2, 12 ), L. subincarnata ( 2, 3 ), L. boudieri ( 13 ), L. elaiophylla ( 1, 2, 14 ), and L. cf. boudieri ( 13 , 15 ).

Compound	Retention Time (min)	Detected Mass [M H]+ +	Calculated Mass [M H]+ +	Error (Δppm)	Sum formula [M H]+ +
β- amanitin ( 1 )	3.98	920.3387	920.3455	+7.3	C39H54N9O15S
α- amanitin ( 2 )	3.99	919.3530	919.3614	+9.2	C39H55N10O14S
γ- amanitin ( 3 )	4.56	903.3585	903.3665	+8.9	C39H55N10O13S
Ɛ- amanitin ( 4 )	4.56	904.3461	904.3505	+4.9	C39H54N9O14S
phallisin ( 5 )	4.67	805.3149	805.3185	+4.5	C35H49N8O12S
amanin ( 6 )	4.77	904.3426	904.3505	+8.8	C39H54N9O14S
phalloidin ( 7 )	5.05	789.3197	789.3236	+4.9	C35H49N8O11S
phallisacin ( 8 )	5.15	863.3170	863.3240	+8.1	C37H51N8O14S
phallacidin ( 9 )	5.34	847.3223	847.3291	+8.0	C37H51N8O13S
phalloin ( 10 )	5.72	773.3248	773.3287	+5.0	C35H49N8O10S
unknown ( 11 )	5.72	889.3322	889.3331	+1.0	C38H53N10O11S2
amaninamide ( 12 )	4.70	903.3589	903.3665	+8.9	C39H55N10O13S
unknown, D1 ( 13 )	4.77	920.3704	920.3693	1.2	C39H56N10O14S
unknown, D2 ( 14 )	4.12	917.3442	917.3458	+1.6	C39H53N10O14S
unknown ( 15 )	5.68	901.4236	901.4223	1.4	C40H65N6O15S

3. Discussion The results obtained concerning the occurrence of toxins in species of Lepiota show that amatoxins can be used as taxonomic characteristics to differentiate some species of Lepiota which were synonymized for a long time in the literature, namely L. elaiophylla from L. xanthophylla P.D. Orton and infraspecific groups in the L. boudieri species complex. In contrast to the current state of knowledge, we show that the number of species of Lepiota containing amatoxins is probably lower than generally assumed. However, many species of Lepiota still need to be investigated to complete our knowledge about the toxicity and diversity of species of Lepiota worldwide. Amatoxins were detected in L. brunneoincarnata , L. elaiophylla , and L. subincarnata . The presence of α- amanitin ( 2 ) in L. brunneoincarnata and L. subincarnata , the presence of amaninamide ( 12 ) in L. brunneoincarnata , as well as the absence of β- amanitin ( 1 ) and the presence of γ- amanitin ( 3 ) in L. subincarnata are in accordance with HPLC-MS results from Sgambelluri et al. (2014) and Yilmaz et al. (2015) as well as TLC results from Haines et al. (1986) . In contrast to Sgambelluri et al. (2014) and Yilmaz et al. (2015) we were not able to detect 1 nor amanin ( 6 ) in L. brunneoincarnata . This difference might be due to different climate, environmental conditions, and/or development stages of the fungal fruiting bodies ( Yilmaz et al., 2015 ). L. aspera , L. castanea , L. clypeolaria , L. cristata , L. erminea , L. felina , L. fuscovinacea , L. lilacea , L. magnispora , L. oreadiformis , L. pseudolilacea , L. sp. (SeSa 5), and L. subalba did not contain any amatoxins or phallotoxins. Furthermore, L. boudieri , L. cf. boudieri , L. brunneoincarnata , L. elaiophylla , and L. subincarnata did not contain any phallotoxins. Differences between our results and TLC results of previous studies might be the consequence of different species concepts and infraspecific variation in the production of natural compounds that may be influenced by ecological factors. Differences in sensitivity and resolution of the applied methods, i.e., TLC and HPLC-MS, should also be considered. To the best of our knowledge, amatoxins are reported here for Lepiota elaiophylla for the first time. Besl et al. (1984) , however, reported 1 and 2 for L. xanthophylla based on TLC studies. L. elaiophylla and L. xanthophylla are morphologically rather similar. For a long period of time, it was unclear whether L. elaiophylla and L. xanthophylla should be treated as synonyms or whether they represent different species ( Vellinga and Huijser, 1997 ). Vellinga (2001) differentiated L. elaiophylla from L. xanthophylla by cellular structures on their pileus and the geographical distribution of the two species. According to these species concepts, L. xanthophylla has a pileus covering composed of hyphae with small clavate terminal cells and grows in forests in temperate Europe. L. elaiophylla has a pileus covering consisting of hyphae with long cylindrical terminal cells and grows in tropical regions or in flowerpots of tropical plants. The separation of these two species is confirmed by ITS sequences, showing that L. elaiophylla and L. xanthophylla are not closely related to each other ( Vellinga, 2003 ). Therefore, we assume in accordance with Vellinga and Huijser (1997) that the Lepiota sample investigated by Besl et al. (1984) might correspond to L. elaiophylla and not to L. xanthophylla . This assumption is supported by drawings of L. “ xanthophylla ” ( Besl et al., 1984 ) showing a pileus covering similar to the one described for L. elaiophylla by Vellinga (2001) . By combining the results on the occurrence of known amatoxins with the phylogenetic analysis of Lepiota spp ., a monophyletic clade is recognized including all species of Lepiota for which known amatoxins have been detected, namely L. brunneoincarnata , L. elaiophylla , L. spiculata , L. subincarnata , and L. venenata . This statement is based on data obtained by the present study as well as results published by Angelini et al. (2020) for L. spiculata and Long et al. (2020) for L. venenata . According to molecular sequence data, this clade also includes L. farinolens Bon & G. Riousset for which amatoxins could not be confirmed because no specimen was available for the analysis. Further investigation is necessary to verify the presence of amatoxins in L. farinolens to support the assumption of a monophyletic distribution of amatoxins in this clade. Outside this “amatoxin clade”, L. boudieri showed a compound (D1, 13 ) corresponding by its mass and MS/MS fragmentation data to amatoxins. TLC results (G´erault and Girre, 1975 ) of L. boudieri (cited as L. fulvella Rea , a synonym) support this observation. The amatoxin derivative D2 ( 13 ) was detected from all L. boudieri specimens at m/z 920.3704 [M+H] + (calcd for C 39 H 56 N 10 O 14 S, Δppm = 1.2 ppm). It is very similar to 1 and does not correspond to the masses of any known amatoxin or phallotoxin ( Table 2 ). The calculated results show that its molecular formula probably is C 39 H 56 N 10 O 14 S, and fragment ions at m/z 902.3605, 884.3166, 171.1130, and 130.0643 [M+H] + from MS/MS are consistent with those produced from known amatoxins ( Fig. S1 ). This compound probably is an undescribed amatoxin. In addition to D1 ( 13 ), the specimen SeSa 13 called L. cf. boudieri presented an additional compound ( 15 ) with a mass of m/z 901.4236 [M+H] + which was not detected in other specimens of L. boudieri . However, the fragmentation pattern of this compound did not correspond to those of amatoxins. The specimen SeSa 13 also differs from other specimens of this species by its ITS sequence and accordingly occupies an isolated position in the phylogenetic tree. Apparently, this specimen represents a taxon closely related to L. boudieri but not identical to it. Based on morphological characteristics, in the past several species and infraspecific taxa were distinguished in L. boudieri that are presently included as synonyms in L. boudieri ( Bon, 1993 ) . The observed difference in natural products may reflect the fact that L. boudieri is a complex of species that may be differentiated by morphology and chemotaxonomic characteristics in future. According to previous studies, one gene ( AMA 1 ) may encode for 2 and 3 in species of Amanita and Galerina ( Hallen et al., 2007 ; Luo et al., 2012 ). This may explain the joint occurrence of both, 2 and 3 , in L. subincarnata . AMA 1 may also be present in L. brunneoincarnata , which contains 2 and 12 that has a molecular weight and structure similar to 3 and that may result from a different posttranslational modification. Another gene is proposed to be responsible for the synthesis of 1 and ε- amanitin ( 4 ) ( Hallen et al., 2007 ). In L. elaiophylla , 1 and 2 were detected, so the presence of both genes in L. elaiophylla is assumed. Further studies are necessary to determine the gene encoding for the undescribed amatoxin derivative (D2, 14 ) and to resolve conflicting data regarding the presence of amatoxins, e.g., in L. elaiophylla , L. xanthophylla and the L. boudieri species complex. Table 3 Specimens of Lepiota spp . used for UHPLC-QTOF-MS/MS analyses.

species	specimens
L. aspera	SeSa 6, SeSa 14, SeSa 38
L. boudieri	SeSa 17, SeSa 46, S 4, HLW 3313
L. cf. boudieri	SeSa 13
L. brunneoincarnata	SeSa 4, SeSa 54, S 8
L. castanea	S 7, KaiR 1179
L. clypeolaria	SeSa 31, SeSa 37, SeSa 40, SeSa 41, SeSa 42, SeSa 52. SeSa
53
L. cristata	SeSa 1, SeSa 3, SeSa 24, SeSa 26, SeSa 27, SeSa 35, SeSa 55,
HLW 3075
L. elaiophylla	S 23
L. erminea	S 18, S 38
L. felina	S 22
L. fuscovinacea	SeSa 8, S 17, S 36
L. lilacea	SeSa 23, SeSa 25
L. magnispora	S 21, S 37
L. oreadiformis	SeSa 21, SeSa 29, HLW 3337
L. pseudolilacea	WPR 201009-005
L. sp.	SeSa 5
L. subalba	SeSa 11
L. subincarnata	SeSa 28, SeSa 36, S 6

The presence of known amatoxins in species Lepiota forming a monophyletic clade around Lepiota brunneoincarnata as already presented by Vellinga (2001) is confirmed and chemotaxonomically supported, while the presence of known amatoxins in other species of Lepiota , as sporadically reported in literature, is not confirmed. In order to decide whether conflicting data on amatoxin occurrences are due to infraspecific variability, ecological factors affecting gene expression, or limitations of methods, further analyses are necessary. The structure and the toxicity of the two putative amatoxin derivatives detected in this study still need to be elucidated. Characteristic ions detected as results of fragmentation pathways can be used as diagnostic ions to quickly identify known toxins or to discover unknown toxins. All these pieces of information are important in order to know the toxic potential of fruiting bodies of Lepiota spp . being ingested by human beings.