Amanita Section Phalloideae Species in the Mediterranean Basin: Destroying Angels Reviewed

Simple Summary Whitish lethal species of Amanita sect. Phalloideae (‘destroying angels’) are known to be among the most poisonous fungi worldwide due to their production of amatoxins. The taxonomy of species occurring in the Mediterranean region is here revised, clarifying the identity of several names. Amanita decipiens, A. porrinensis, and A. virosa var. levipes are here considered later heterotypic synonyms of A. verna, A. phalloides, and A. amerivirosa, respectively, while a new name, A. vidua, is proposed for a spring-occurring taxon. The amatoxins and phallotoxins present in Mediterranean destroying angels were characterized, and their epidemiology discussed on the basis of the case study of available data from Spain. Abstract In Europe, amatoxin-containing mushrooms are responsible for most of the deadly poisonings caused by macrofungi. The present work presents a multidisciplinary revision of the European species of Amanita sect. Phalloideae based on morphology, phylogeny, epidemiology, and biochemistry of amatoxins and phallotoxins. Five distinct species of this section have been identified in Europe to date: A. phalloides, A. virosa, A. verna, the recently introduced North American species A. amerivirosa, and A. vidua sp. nov., which is a new name proposed for the KOH-negative Mediterranean species previously described as A. verna or A. decipiens by various authors. Epitypes or neotypes are selected for species lacking suitable reference collections, namely A. verna and A. virosa. Three additional taxa, Amanita decipiens, A. porrinensis, and A. virosa var. levipes are here considered later heterotypic synonyms of A. verna, A. phalloides, and A. amerivirosa, respectively.


Introduction
Lethal species of the genus Amanita Pers. (Amanitaceae, Agaricales, Agaricomycetes, Basidiomycota, Fungi), including the white-coloured taxa in sect. Phalloideae, which are commonly called 'destroying angels', are among the most well-documented causes of intoxication by fungi, with reports dating back to ancient history. Despite the information campaigns periodically launched by public health services and mycological societies, intoxications caused by these species are still reported every year, resulting in the most severe cases of mushroom poisoning in Europe and North America [1][2][3]. The first lethal species to be formally named were A. phalloides (Vaill. ex Fr.) Link (basionym Agaricus phalloides Vaill. ex Fr.) and A. verna Bull. ex. Lam. (protonym Ag. bulbosus vernus Bull.), both found near Paris, France [4,5], and A. virosa Bertill. (basionym Ag. virosus Fr.) from

Phylogenetic Studies
Total DNA was extracted from herbarium specimens using a standard method based on cetyltrimethylammonium bromide (CTAB) [40], or with the REDExtract-N-Amp tm Plant PCR Kit (Sigma-Aldrich, St. Louis, MO, USA), following the manufacturer's instructions. Polymerase chain reactions (PCRs) [41] included 35 cycles with an annealing temperature of 54 • C. Primers ITS1F and ITS4 [42,43] were employed to amplify the internal transcribed spacer nuc-rDNA region (ITS), while LR0R, LR1, and LR7 [44][45][46] were used for the 28S nuc-rDNA large ribosomal subunit (LSU). PCR products were checked in 1% agarose gels, and positive reactions were sequenced with one or both PCR primers (depending on the quality of the resulting sequences). Chromatograms were edited by hand to correct reads at heteromorphic sites and other putative errors using MEGA5 [47] or Codon Code Aligner 4.1.1 (CodonCode Corp., Centerville, MA, USA).
The ITS and LSU sequences obtained were aligned with the closest matches from BLAST searches in the public databases. The sequences retrieved (Supplementary Table S1) came mainly from [48][49][50][51][52], among others. The resulting dataset was aligned in MEGA5 software using its ClustalW application followed by manual correction. The concatenated loci (677 nt ITS, 554 nt LSU) were subjected to Gblocks [53], which removed 221 ambiguously aligned positions (165 nt ITS, 56 nt LSU), to obtain a final alignment with 512 nt from ITS (301 nt variable), and 498 nt from LSU (139 nt variable). Each region was loaded in PAUP* 4.0b10 [54] and subjected to MrModeltest 2.3 [55] to infer the best-fitting evolutionary model. MrBayes 3.2.6 [56] was employed to conduct a Bayesian inference (BI) analysis using model GTR + G+I (nst = 6, rates = invgamma, statefreqpr = dirichlet) in each partition (ITS and LSU), two simultaneous runs, four chains, temperature set to 0.2, and sampling every 100th generation. Finally, a full search for the best-scoring maximum likelihood (ML) tree was performed in RAxML 8.2.10 [57] using the standard search algorithm with the GTRCAT approximation as recommended by the manual for data sets >50 taxa, data partitioned as for Bayesian analysis, and 2000 bootstrap replications. All analyses were run locally. The significance threshold was set above 0.95 posterior probability (PP) for BI and above 70% bootstrap proportion (BP) for ML.

Toxicological Studies
High performance liquid chromatography (HPLC) was employed to quantify the amatoxins present in selected samples from the species inferred by phylogenetic analyses (Supplementary Table S2). All extractions were repeated three times. Since the pileus of poisonous species of Amanita is known to contain large amounts of toxins [58,59], a portion of the cap (50 mg of dry material) was taken from each sample and finely ground. One ml of a mixture of MeOH/H 2 O/HCl (5:4:1, v/v/v) was added as previously described [25,[59][60][61][62], sonicated for 15 min, and then filtered and centrifuged. Three different fruitbodies of PAM19050401 (harvested in the same location and time, named a, b, and c) were analyzed, each one being extracted three times as the other samples.
Chromatographic separation and detection for quantitative analyses were performed on an Ultimate 3000 instrument that included a quaternary pump, a degasser, an automatic sampler, and a DAD detector (Thermo Fisher Scientific Inc., San Jose, CA, USA). The system was operated using the Chromeleon software, version 2.0. Chromatographic separation was achieved on an ODS Hypersyl C18 column (250 × 4.6 mm, 5 µm, Thermo Fisher Scientific Inc., San Jose, CA, USA), with a column temperature maintained at 35 • C. Analyses were performed at a flow rate of 1 mL/min, using solvent A (0.02 M ammonium acetate/acetonitrile 90:10 v/v) and solvent B (0.02 M ammonium acetate/acetonitrile 76:24 v/v) as previously described [60,61,63]. HPLC-grade acetonitrile (Carlo Erba Reagents, Val de Reuil, France) and water (PanReac Quimica SLU, Castellar del Valles, Spain) were employed. The gradient profile was set up as follows: 100% A for 4 min, then 57% B for 16 min, then 100% B for 10 min, and finally 100% A.
Standards of α-amanitin, β-amanitin, phallacidin, and phalloidin were purchased from Sigma-Aldrich (Saint-Louis, MO, USA) and stocked at −20 • C. The toxins were freshly prepared and mixed together to obtain calibration solutions at 1. 95, 3.91, 7.81, 15.63, 31.25, 62.5, 125, 250, and 500 µg/mL in MeOH/H 2 O (50:50). Individual calibration curves were produced for each toxin and shown to be linear over the range of interest (R2 > 0.9995, n = 3, Supplementary Figure S1). The UV/Vis spectra were recorded in the 200-400 nm range and chromatograms were acquired at 254 nm (unspecific wavelength), 295 nm (λ max for phallacidin and phalloidin), and 305 nm (λ max for α-amanitin and β-amanitin). Extracts obtained from the different species of Amanita studied in the present work were analyzed directly without any prior dilution. For each peak, a UV spectrum was obtained and compared with those of commercial standards to confirm their identification. Contents of amatoxins and phallotoxins were expressed in mg of toxin/g of dried cap.

Epidemiological Studies
The epidemiological analysis focused on phalloidian intoxications in Spain, targeting two main indicators: (1) cumulative incidence of poisoning according to time of year and (2) geographical and seasonal distribution of poisoning incidence. Primary data were gathered from the following sources: 1.
Ministry of Science and Innovation-National Center of Epidemiology (CNE)-Sistema de Vigilancia de la Red Nacional de Vigilancia Epidemiológica (RENAVE), SiViEs platform: for the period 1980-2020, outbreaks in which the agent was a fungal toxin and those associated with another agent in which mushrooms were consumed. In this database, only clusters (n ≥ 2 patients) are registered. To target putative phalloidian cases, the criterion "Incubation time" was selected for values > 6 h (when recorded) coupled with "hospitalization", assuming that phalloidian poisonings required hospitalizations. Cases involving non-phalloidian fungal species were discarded. 2.

Phylogeny
The Bayesian inference of phylogeny based on ITS and LSU data ( Figure 1) is consistent with previously published organization of Amanita sect. Phalloideae [31,[33][34][35][36][37]64], although the limited signal provided by the ribosomal regions alone does not resolve most supraspecific nodes supported by the analyses, including protein-coding genes. The Mediterranean 'destroying angels' include at least seven different species. Five of them cluster inside sect. Phalloideae: (1) Amanita phalloides (= A. phalloides var. alba, = A. porrinensis),  (6) is nested within sect. Strobiliformes (Amanita strobiliformis), and the last one (7) belongs in sect. Validae (A. phalloides var. larroquei = A. citrina). Amanita phalloides is significantly related to A. subjunquillea S. Imai, which includes a white variety, A. subjunquillea var. alba Zhu L. Yang. While both species did not receive reciprocal significant support in the present analysis, they are considered independent taxa until a more complete analysis is done to test their actual status. Amanita virosa and A. amerivirosa/A. virosa var. levipes are significantly related to A. subpallidorosea Hai J. Li, but their phylogenetic relation with A. ocreata Peck, found in the multigenic analysis performed by Codjia et al. [34] could not be recovered with the present analysis. All Mediterranean species in sect. Phalloideae are nested within a 'Northern Hemisphere lineage', significantly distinct from the remaining taxa, which in previous works [34][35][36][37], constitute three different monophyletic lineages. ITS rDNA sequences of A. strobiliformis are 92% similar to those of A. sabulicola S. Morini et al., which is by now the closest relative [65]. In turn, the ITS sequence of A. phalloides var. larroquei (specimen provided by its author F. Massart) is almost 98% similar to sequences of A. citrina Pers. (i.e., MH508311), and therefore probably belongs to this species. The taxonomy of these species is updated below according to these results.
Habitat and distribution: very frequent, isolated, or scattered basidiomes can be found in autumn, forming mycorrhizal associations with Quercus as well as with other broadleaved trees, such as Castanea, Corylus, Betula or Fagus, less commonly with conifers (Cedrus, Pinus). Found in temperate, Mediterranean, and boreal habitats. Notes: Amanita phalloides is a very common autumnal species described in detail in many publications and web sites, i.e., [38,67,68]. Freire & Castro [69] did not recognize A. phalloides in the odd-shaped specimen described by them (and validated by Castro [70]) as A. porrinensis; these two species cannot be separated from one another with the current sequencing results, and so they are considered conspecific. Therefore, we propose to downgrade the unusual Mediterranean taxon observed by Freire & Castro [69] and redescribed by Neville et al. [71] to a form of A. phalloides.

2.
Amanita Pileus campanulate, with a wide and obtuse central umbo, 5-7.5 cm in diam.; surface pure white, not changing, smooth, not striate nor fibrillose, lacking any remnants of the universal veil, difficult to separate from the context underneath; margin not striated, inrolled at first, then flat when mature. Lamellae very crowded free, with pure white lamellulae; edge white; spore print white. Stipe 10-20 × 2-3 cm, cylindrical, sinuose, enlarged towards the base, pure white, sericeous to slightly floccose; annulus white, hanging high on the stipe, membranous, thin, soon disappearing; universal veil forming a fragile sack-like membranous volva, adhering, detachable, without remnants on the pileus surface. Flesh white, not changing, odourless; taste slightly sweet. Reactions: pileipellis turning citrine yellow with KOH 10% according to Freire & Castro [69], although the same authors report a weak pale yellowish reaction in the herbarium notes. Neville et al. [71] report an intensely yellow reaction on lamellae, weaker in the pileus and stipe; positive to Wieland test.
Notes: a very rare albino form of A. phalloides with a characteristic obtuse wide umbo and squamose or sericeous stipe. Only a few collections are known, found in autumn, in Spain [69], as well as Italy and Switzerland [71][72][73]. A toxic species according to the results of the Wieland test performed by the original authors [69]. The present results based on ribosomal data suggest that A. porrinensis is conspecific with A. phalloides. Sequences from several protein-coding genes (tef1, rpb2) obtained from a recent collection from Italy (ALV30282) do not show differences with those obtained from A. phalloides available in public databases (data not shown). Amanita porrinensis is here downgraded to a form of A. phalloides to warn about this unusual, but probably lethal, morphological variant.

3.
Amanita Pileus 4-15 cm in diam., originally conical-campanulate, expanding to convex-flat with a persistently convex center, never depressed; surface pure white and remaining so or sometimes becoming light ochraceous-pinkish at the center with age, greasy to nearly dry, shiny-silky when dry, smooth; margin incurved then expanded and flexuose when mature, smooth, never striated. Lamellae crowded, 80-100, reaching the stipe, typically with one lamellula per lamella, smooth to nearly smooth, cream white, with slight pinkish tones with age; edge white, floccose. Stipe 9-20 (-25) × 0.8-2.5 cm, cylindrical throughout or progressively thinner towards the apex, bulbous-sphaerical at the base (which can be up to 3.5 cm thick), pure white, slightly floccose towards the apex, below the annulus subtly silky-zebrate, becoming glabrous with age; partial veil a membranous, persistent annulus hanging 2-3 cm down the apex, rarely laciniate and then hanging at the margin, not adhering to the edges; universal veil forming an ample membranous volva, rarely dissociated on the pileus, white, with pale pinkish-ochraceous patches on outer surface. Context white, slightly foxy-spotted on damaged surfaces, light lemon yellow when dry; smell distinctly iodine-like in the bulb, elsewhere not distinct, fungoid by alteration; taste mild, fungoid. Reaction to 5% KOH chrome yellow on all surfaces and context. Exsiccates turn entirely light lemon yellow after a few days. Spore print pure white.
Notes: Amanita amerivirosa is a very recently described taxon [37], so far only documented by DNA sequences from NE America where it has been recorded for a long time under the name A. virosa. We confirm here Neville & Poumarat's [38] hypothesis that Amanita virosa var. levipes described from the French Atlantic coast is the same fungus, likely not indigenous in Europe.

4.
Amanita Pileus 3-9 cm diam., globose at first then convex to flat, soon flattened to somewhat depressed at center; surface smooth, slightly greasy when moist, pure white at first, somewhat ochraceous towards disk with age; margin convex, only later expanding, never striate. Lamellae crowded, 70-80 reaching the stipe, 1 (-2) lamellulae per lamella, free to somewhat adnexed, pure white when young, often with a subtle pinkish tone when ageing; edge white, thin, smooth to somewhat serrulate. Stipe 6-15 × 0.8-1.5 cm, equal to slightly narrowed at apex, abruptly bulbous-sphaerical at base, not or only faintly striate at apex, smooth, silky to subtly floccose below annulus; partial veil a membranous, persistent but thin annulus usually hanging high on the stipe, white, slightly striate on the upper side, smooth; universal veil a membranous, rather thin volva, white to somewhat ochraceous on surface. Flesh white, unchanging; smell none, somewhat fungoid on adults; taste mild. Reaction chrome-yellow to 5% KOH on all surfaces and context, except inner layer of volva.
If the original concept of A. verna is considered a synonym of A. phalloides var. alba [76,77], the name A. decipiens would be available for the KOH+ specimens (see Discussion below). However, we think that this option would not contribute to stabilizing the nomenclature of this group, and the interpretation of A. verna as a KOH+ species is here preferred. Reaction chrome-yellow to 5% KOH on all surfaces and context, except inner layer of volva.  [38]. In the present study, no spore print was available, and spores were observed on exsiccata as deposits at the apex of the stipe or near the edge. Most collections (including the epitype) studied here displayed only a moderate variability of spore shape, which ranged from subglobose to ellipsoid ( Figure 6A,C). Specimens with narrowly ellipsoid to cylindrical spores such as those described as A. verna f. ellipticospora by Gilbert [78] or those described by Bertault [80] from specimens collected in Morocco, could not be revised here and a special attention should be given to such collections in the future. The description of A. gilbertii f. subverna Bertault & Parrot from Morocco, about which Bertault [81] wrote: "seules les spores cylindriques permettent de les séparer" (only the cylindrical spores make a difference [with A. verna]), suggests that this taxon, which Neville & Poumarat [38] validated by selecting a holotype found in the island of Porquerolles (south eastern France), could be related also to A. verna.
Although Amanita verna is univocally described in the literature as having a white to more or less ochre pileus, several collections from Italy with a greenish yellow pileus have been found by Raumi [82], described under the provisional name 'Amanita verna f. xanthoviridis' ad. int. The ITS rDNA, TEF1, and RPB2 sequences from one of these collections (ALV21204, kindly shared by M. Raumi, 1 pileus deposited at LIP n • 0002275) do not differ from those of typical samples of A. verna (Figure 1). The yellow tinge seems to be due to the presence in the subpellis of radially oriented hyphae 5-6.5 µm wide with thick, yellowish walls (up to 1 µm thick), which were not observed in white specimens; the spores are in the range of the type, [1/22] 9.2-10.4 × 7.0-8.2 µm, Q = 1.2-1.3, Q av = 1.2, mostly broadly elliptical, with a few macrospores likely issued from 2-spored basidia, up to 13.5 × 10.5 µm. This remarkable form might be confused with A. phalloides, from which it differs by the fruiting season and the strong yellow reaction of all parts of the basidiome to KOH (Figure 2). No complete specimen could be obtained for the purpose of this article, and thus a full description of this form and the validation of its provisional name could not be provided here.

5.
Amanita vidua Gasch, G. Moreno [86] Etymology: from Latin noun vidua: widow, an ironic reference to its high toxicity and to the fact that this species became 'widowed' after the name A. verna could not be employed for it anymore.
Habitat and distribution: infrequent vernal species that probably establishes ectotrophic mycorrhizae with Mediterranean species of Quercus (Q. ilex ssp. ballota, Q. pyrenaica, Q. suber, etc.), rarely also with other trees, i.e., Castanea, Fagus, Pinus. Known from acidic soils of the Mediterranean basin, from Spain to Lebanon, as well as Morocco [76], from April to early June.     Description of the neotype: Pileus 4-11 cm, cylindrical to conical-obtuse at first, later flexuose with the center persistently convex or sometimes umbonate, never flattened or depressed, greasy to almost viscid, shiny-silky when dry, pure white, disc slightly ochraceous when old; margin incurved for a long time, then irregularly expanded, smooth, never striated, white, usually with traces of a floccose partial veil hanging. Lamellae crowded, rather thick, free, collariate at maturity, pure white, with subtle pinkish tones when mature; edge thick, heavily floccose, usually with adherent floccose patches of the partial veil. Stipe 7-18 × 0.6-1.8 cm, cylindrical or narrowed towards the apex, equal, slender, often curved at maturity, inflated at the base with a sphaerical bulb up to 3 (-4) cm wide; surface entirely floccose, with bandlets or ascendant scales on the lower part with age, somewhat striate at the apex, pure white and remaining so with age; partial veil membranose-floccose, fragile, easily dissociated, usually partly hanging high on stipe, often absent in mature specimens, partly adherent to lamellae and to the pileus margin, white; universal veil forming a resistant membranous volva, white to somewhat ochraceous on surface. Context white, unchanging with age. Herbarium samples remaining white, or faintly pale yellow at the center of the pileus; fresh specimens lack any smell, but they can develop a fungoid, unpleasant odor (especially in the bulb) with age; taste mild, fungoid. Reaction to 5% KOH chrome yellow on all surfaces and context except the inner layer of the volva.
Notes: In Europe, Amanita virosa cannot be easily confused with other species, being the only Amanita sect. Phalloideae growing on acidic organic soils, with a creamy ring and conical pileus. However, although being a strictly Eurasian species (as suggested by the sequences available in GenBank and UNITE databases), its name has been misapplied to various white extra-European species. Its neotypification with a collection from Sweden will fix the name and avoid further misidentifications. Examined

Toxicology
Results from the HPLC analysis ( Figure 9) show that the contents of amatoxins and phallotoxins are strongly associated with the genetic identity of the samples. Amanita virosa contains α-amanitin, phallacidin, and phalloidin but lacks β-amanitin. This is partially in accordance with the literature [62,[87][88][89]. However, according to other authors, some samples of A. virosa from North America [90,91] or Japan [92] contain β-amanitin, but no sequencing was performed to check the actual identity of these samples, which could represent other taxa. Samples of A. amerivirosa analyzed in the present work do not show amanitins or phallacidin, resembling the profile exhibited by other samples of A. virosa from North America [90]. Samples of A. verna analyzed in the present work contain large amounts of αand β-amanitins (5.47 and 10.26 mg/g dry matter, respectively) and phallacidin (4.69 mg/g dry matter), and a lower amount of phalloidin (0.64 mg/g dry matter) in accordance with other works [25,89]. With regard to A. phalloides, different contents of α-amanitin were found in samples LIP:0402239 (from a Pinus pinaster forest with sandy soil) and SMM2018-10 (from a Quercus ilex stand with Arbutus unedo). Soil properties are thought to influence the toxicological profile of A. phalloides [58,93], and this could explain the differences found. However, Bon & Andary [94] did not find αor β-amanitins by thin layer chromatography (TLC) in the type specimen of Amanita dunensis (=A. phalloides). The method of extraction and/or quantification (HPLC vs. TLC) might account for these differences. In addition, what remains of A. dunensis type collection at LIP (Bon 4119) is a half-specimen in bad condition, probably already very mature when collected. This could be another factor that influences the toxicological results. More samples from the Atlantic coast should be analyzed to check whether actual differences exist or not. Finally, the analyses conducted on specimens of A. vidua PAM1905041 at various developmental stages revealed low inter-specimen variations. Large amounts of all amatoxins (α-amanitin: 4.73-7.18 mg/g dry matter, β-amanitins: 5.02-8.50 mg/g dry matter) and phallotoxins were detected (phallacidin: 6.53-7.78 mg/g dry matter), suggesting that A. vidua is highly toxic and must be considered as another deadly 'destroying angel', maybe the most poisonous of the entire group.
Biology 2022, 11, x FOR PEER REVIEW 24 of 33 of α-amanitin were found in samples LIP:0402239 (from a Pinus pinaster forest with sandy soil) and SMM2018-10 (from a Quercus ilex stand with Arbutus unedo). Soil properties are thought to influence the toxicological profile of A. phalloides [58,93], and this could explain the differences found. However, Bon & Andary [94] did not find α-or β-amanitins by thin layer chromatography (TLC) in the type specimen of Amanita dunensis (=A. phalloides). The method of extraction and/or quantification (HPLC vs. TLC) might account for these differences. In addition, what remains of A. dunensis type collection at LIP (Bon 4119) is a half-specimen in bad condition, probably already very mature when collected. This could be another factor that influences the toxicological results. More samples from the Atlantic coast should be analyzed to check whether actual differences exist or not. Finally, the analyses conducted on specimens of A. vidua PAM1905041 at various developmental stages revealed low inter-specimen variations. Large amounts of all amatoxins (α-amanitin: 4.73-7.18 mg/g dry matter, β-amanitins: 5.02-8.50 mg/g dry matter) and phallotoxins were detected (phallacidin: 6.53-7.78 mg/g dry matter), suggesting that A. vidua is highly toxic and must be considered as another deadly 'destroying angel', maybe the most poisonous of the entire group.

Epidemiology
The analysis of records obtained from Spanish databases was divided into two periods: (1) spring period (from 1 February to 30 July), and (2) fall period (from 1 August to 31 January). These periods are clearly separated by the cold or dry periods when Amanita species do not fruit, i.e., mid-winter (January), in which no case of phalloidian poisoning was registered in any database, and mid-summer (July) with only one case known, attributable to A. verna. Ninety-four fungal poisoning clusters (≥2 patients), representing 406 symptomatic cases, were retrieved from the RENAVE database (1980-2020, source 1). Five clusters in the spring period (Feb-Jul) and ten clusters in the fall period contained cases compatible with phalloidian poisonings. Spring clusters led to 10 hospitalizations without registered deaths and originated in the provinces of Aragon (2 clusters, 5 cases), Andalusia (2 clusters, 3 cases), and Castile and León (1 cluster, 2 cases). During the fall

Epidemiology
The analysis of records obtained from Spanish databases was divided into two periods: (1) spring period (from 1 February to 30 July), and (2) fall period (from 1 August to 31 January). These periods are clearly separated by the cold or dry periods when Amanita species do not fruit, i.e., mid-winter (January), in which no case of phalloidian poisoning was registered in any database, and mid-summer (July) with only one case known, attributable to A. verna. Ninety-four fungal poisoning clusters (≥2 patients), representing 406 symptomatic cases, were retrieved from the RENAVE database (1980-2020, source 1). Five clusters in the spring period (Feb-Jul) and ten clusters in the fall period contained cases compatible with phalloidian poisonings. Spring clusters led to 10 hospitalizations without registered deaths and originated in the provinces of Aragon (2 clusters, 5 cases), Andalusia (2 clusters, 3 cases), and Castile and León (1 cluster, 2 cases). During the fall (August-January), 10 clusters were retained as possibly phalloidian, representing 28 hospitalizations and 2 deaths. One fall case was attributed explicitly to Lepiota brunneoincarnata, while the others were either attributed to Amanita spp. or not identified in RENAVE; fall intoxications occurred in Andalusia (3 clusters, 9 cases), Castile-La Mancha (2 clusters, 4 cases), Catalonia (2 clusters, 4 cases), Aragon (1 cluster, 2 cases), Galicia (1 cluster, 2 cases), and Castile and León (1 cluster, 2 cases).
The CMBD (1997CMBD ( -2015 and RAE-CMBD (2016-2020) databases (source 2) contain 1817 cases of putative fungal intoxications, 219 of them presenting acute hepatitis, characteristic of the phalloidian syndrome. During the spring period (Feb-Jul), 16 cases were recorded, which originated in Andalusia (4) (17), Aragon (12), and Andalusia (9) regions; Murcia is the only community without any phalloidian cases in the fall. It is noticeable that all spring cases in Andalusia occurred in the western part of the region (Huelva, Seville) close to Extremadura, which has a spring incidence nearly 10 times higher than the national average. Global cumulative incidence of phalloidian poisonings in Spain for the 1997-2020 period was 4.8/1,000,000 hab. (0.34 in spring and 4.34 in fall), calculated from the population in the middle of the period (detailed for each community in Figure 10), and average cumulative incidence was 0.20/1,000,000 inhabit. per year.  (17), Aragon (12), and Andalusia (9) regions; Murcia is the only community without any phalloidian cases in the fall. It is noticeable that all spring cases in Andalusia occurred in the western part of the region (Huelva, Seville) close to Extremadura, which has a spring incidence nearly 10 times higher than the national average. Global cumulative incidence of phalloidian poisonings in Spain for the 1997-2020 period was 4.8/1,000,000 hab. (0.34 in spring and 4.34 in fall), calculated from the population in the middle of the period (detailed for each community in Figure 10), and average cumulative incidence was 0.20/1,000,000 inhabit. per year.

Discussion
Amanita verna was originally described by Bulliard ([5], pl. 108) under the trinomial name Agaricus bulbosus vernus. The original plate illustrates four complete, entirely white

Discussion
Amanita verna was originally described by Bulliard ([5], pl. 108) under the trinomial name Agaricus bulbosus vernus. The original plate illustrates four complete, entirely white basidiomes and a longitudinal section of the biggest one. In the extensive legend of the plate, the author describes a spring-fruiting Amanita with pure white, convex to depressed pileus, rounded bulb, and a well-formed membranous ring placed high on a smooth stipe. Bulliard indicates that the species is common in the woods around Paris, and cites a greenish form, which he later described in detail (pl. 506) under the name Agaricus bulbosus Bull. Shortly after this publication, the taxon was upgraded to species status by Lamarck [95] as Amanita verna (Bull.) Lam., with a similar description.
The history of the name Amanita verna, sanctioned by Fries [96] at species rank, has been reviewed by various authors, with opposing conclusions about the real identity of Bulliard's species, according to personal interpretations of the aforementioned elements (Table 1). Konrad & Maublanc [76], for instance, interpreted A. verna as the pigmentless form of A. phalloides, also named A. phalloides var. alba or A. andaryi (invalid name), based on the fact that Bulliard included in his Ag. bulbosus white and yellow specimens without other distinction. However, most authors accepted A. verna as a vernal species, much rarer than mentioned by Bulliard (at least in Northern France), which differs from A. virosa by its smooth stipe. In a posthumous publication (a catalogue of macrochemical reactions of many species without descriptions), Bataille [77] revealed a new feature observed in specimens identified by him as A. verna: the absence of reaction to 10% KOH on the pileus, in contrast to A. virosa, which turned bright chrome-yellow to this reagent. However, many authors after him reported a positive reaction on their own collections of A. verna [67,86,97], among others, and so did Trimbach [84], who accommodated the KOH+ samples in a separate variety, A. verna var. decipiens Trimb., expressing his own doubts about the taxonomic value of the KOH reaction. A couple of years later, Trimbach coined the name A. verna var. tarda for a variety found in Switzerland with less truncate lamellae, fruiting in autumn, and lacking any visible reaction to KOH.
Bertault [80,83]  The KOH-samples are here named A. vidua, since they lack any other existing name. The spectacular yellow reaction to KOH displayed by A. verna is present also in A. virosa and A. amerivirosa, and therefore this macrochemical feature could have been present in the common ancestor of the whole 'North Hemisphere lineage' but lost in A. phalloides (although somewhat present in f. porrinensis) and strongly weakened in A. vidua. This reaction appears under optical microscope to be mainly located at the abundant epiparietal mucoid deposits on most hyphae, and extends to spore content, which shows a yellowish lipidic content in 5% KOH, in all species with the exception of A. phalloides, and is only scarcely visible in A. vidua. The spores themselves have a distinctive shape, mainly subglobose in A. amerivirosa and A. virosa (Q av = 1-1.15), mainly broadly ellipsoid with a higher Q av (1.13-1.22) and usually narrower (<7.8 µm wide) in A. phalloides and A. vidua. Amanita verna has a remarkable variability in spore shape, observed among different collections as described above (see Figure 6A,C,D), but also among spores obtained from the same basidiome as illustrated by Malençon & Bertault [79].
The four toxins studied in the present work are apparently occurring across the whole sect. Phalloideae, suggesting that all species are probably toxic, but there are some remarkable differences among them. In the virosa clade, β-amanitin is not detectable or at a very low concentration, and A. amerivirosa has a very low content of α-amanitin, completely lacking phallacidin. In addition, chromatograms of A. amerivirosa display some peaks absent in A. virosa, with UV spectra corresponding to compounds lacking 6-hydroxytryptophan [62]. The identification of these compounds would be of great toxicological interest. The actual amount of toxins is highly dependent on the extraction conditions, with major impacts of solvent, temperature, and processing times during the maceration step. Freshness of the samples analyzed also influences the results of HLPC, but little information exists about the degradation of amatoxins and phallotoxins during storage. Stijve & Seeger [99] observed a decrease in amanitins (α-and β-) and phalloidin contents over time in specimens collected in the same location in different years and suggest that phallotoxins might be less stable during storage than amanitins, whose stability is known to depend on treatment and storage of the samples [100][101][102]. Results obtained in the present work support this hypothesis, since at least αand β-amanitins seem less abundant in older specimens (i.e., A. vidua from 1987 vs. A. verna from 2001). However, these observations were not the purpose of the present work, and therefore a more detailed study should be carried out to further clarify this issue.
Amanita amerivirosa has been reported from western France for more than 30 years, first as A. decipiens [103], then as 'American A. virosa', and finally named A. virosa var. levipes [38]. Carefully surveyed by local mycologists who could follow its fast expansion from the Nantes area to Bordeaux and Normandy [104,105], it has become especially abundant in the last 15 years, recently reaching the Paris area [106]. Fortunately, in spite of the abundance of this species in western France, the number of poisoning reports caused by A. amerivirosa in France is still limited [2,107], with only two cases confirmed in 2018 (identified as A. virosa var. levipes; Moreau, unpublished data), and they lacked associated symptomatology. This could be related to the low amount of toxins detected in the collections of A. amerivirosa studied by Beutler & Der Marderosian [90], Bonnet & Basson [91], and the present work. However, this needs to be confirmed by additional analyses, and so this species should not be excluded from the list of lethal 'destroying angels' for the time being. On the other hand, A. vidua and A. verna are evidently highly poisonous on the basis of the amatoxins detected in these species. Both species may grow in the same sites and season as A. ponderosa Malençon & R. Heim (the so-called "gurumelo"), a highly prized edible fungus endemic in southern Portugal and south-western Spain (Huelva, Badajoz, Cáceres; [74,108]). This could explain the abnormally high incidence of spring intoxications compared to fall intoxications in the Spanish provinces of Extremadura and Andalusia ( Figure 10). Amanita boudieri Barla (Amanita sect. Lepidella), which can cause another lethal syndrome (acute renal failure; [109]), has also been mistaken for A. ponderosa but differs morphologically by the absence of a membranous volva and ring [108].
The average incidence of phalloidian poisonings (mostly due to lethal Amanita species) in Spain during the period analyzed in the present work was 0.20/1,000,000 inhabit. per year. This value is comparable to botulism (0.20 per year in the last decade; [110]), another food-borne disease, and thus represents a comparable challenge for public health. More detailed and homogeneous data on similar poisoning cases in Europe, as well as more precise identification of the responsible species, would be necessary to evaluate and manage the risk, which seems to differ between regions. Especially, information campaigns targeting unexperienced mushroom collectors should warn them of the presence of the lethal 'destroying angels'. Since A. vidua can be found in the whole Mediterranean basin (from Spain to Levant, and probably also in North Africa) and shows a high toxicity, a special focus should be placed on this species in the entire area. However, spring intoxications are apparently not relevant in some areas, such as Apulia (southern Italy; [111]), suggesting that either A. vidua and A. verna are not present there, or else that in the absence of similar edible early-fruiting species or of local picking traditions attached to them, they are not accidentally consumed.