Morphology, Multilocus Phylogeny, and Toxin Analysis Reveal Amanita albolimbata, the First Lethal Amanita Species From Benin, West Africa

Many species of Amanita sect. Phalloideae (Fr.) Quél. cause death of people after consumption around the world. Amanita albolimbata, a new species of A. sect. Phalloideae from Benin, is described here. The taxon represents the first lethal species of A. sect. Phalloideae known from Benin. Morphology and molecular phylogenetic analyses based on five genes (ITS, nrLSU, rpb2, tef1-α, and β-tubulin) revealed that A. albolimbata is a distinct species. The species is characterized by its smooth, white pileus sometimes covered by a patchy volval remnant, a bulbous stipe with a white limbate volva, broadly ellipsoid to ellipsoid, amyloid basidiospores, and abundant inflated cells in the volva. Screening for the most notorious toxins by liquid chromatography–high-resolution mass spectrometry revealed the presence of α-amanitin, β-amanitin, and phallacidin in A. albolimbata.

In this study, a new member of A. sect. Phalloideae from tropical West Africa is described. Its macromorphological and micromorphological characteristics, as well as its phylogenetic relationships with other Amanita species are discussed. In addition, the screening of the species for the known toxins occurring in Amanita is reported.

Collections and Preservation
Specimens were opportunistically collected in Benin, West Africa (Figure 1), during the rainy season from June to September (2018-2019), especially in the forest dominated by Fabaceae/Leguminosae (Isoberlinia Craib and Stapf ex Holland, Anthonotha P. Beauv., Berlinia Sol. ex Hook. f.), Phyllanthaceae (Uapaca Baill.), and Dipterocarpaceae (Monotes A. DC.). Specimens were photographed in situ using a digital cameratype Canon EOS 60D. Macromorphological characteristics were recorded on fresh materials, according to Tulloss and Yang (2011). Color codes recorded from fresh materials follow Kornerup and Wanscher (1981). The fresh basidiomata were dried using an electric dryer Stockli Dorrex at 45 • C for 1 day and stored thereafter as exsiccates with their label in sealable plastic bag-type minigrip. The dried specimens along with the holotype of the newly described species are deposited in the Mycological Herbarium of the University of Parakou (UNIPAR). Duplicates of dried specimens and the isotype of the new species are conserved at the Herbarium of Cryptogams of Kunming Institute of Botany, Chinese Academy of Sciences (KUN-HKAS). Small pieces of fresh basidiomata were also stored in CTAB lysis buffer (2% cetyl trimethylammonium bromide, 100 mM Tris-HCl, 20 mM EDTA, 1.4 M NaCl) and dried with silica gel for molecular investigations. Nomenclature aspects, as well as authorities for scientific names, have been double checked against Index Fungorum 1 and in .

Micromorphological Investigations
Microscopic structures were studied from dried materials mounted in 5% KOH and stained with Congo red to depict all tissues. The Melzer's reagent was used to test the amyloidity of basidiospores. All measurements and line drawings were performed at 1,000× magnification, and a minimum of 20-30 basidiospores from each basidioma were measured in side view. Micromorphological investigations were performed by mean of a microscope-type Nikon Eclipse 50i. The abbreviation (n/m/p) is used to describe basidiospores where n is the number of basidiospores from m basidiomata of p collections. The basidiospores dimensions are provided using the notation (a)b -c(d) with the range b-c containing a minimum of 95% of the measured values, a and d in the brackets showing the two extreme values. Q is used for the ratio length/width of a spore in side view; Qm is the average Q of all basidiospores ± sample standard deviation. The measurements for the basidiospores were analyzed with Piximetre v5.10 (Henriot and Cheype, 2020). The descriptive terms are in accordance with Bas (1969), Yang (2005), Yang (2015), Cai et al. (2016), and Cui et al. (2018).

DNA Extraction, Amplification, and Sequencing
The total genomic DNA was obtained from materials preserved in CTAB or dried with silica gel following the modified CTAB procedure (Doyle and Doyle, 1987). Polymerase chain reaction (PCR) amplification and sequencing were performed in accordance with those described in Cai et al. (2014) and Cui et al. (2018). The following primer pairs were used for PCR amplification and sequencing: ITS1F and ITS4 to amplify ITS region (White et al., 1990;Gardes and Bruns, 1993); LROR and LR5 (Vilgalys and Hester, 1990) for nrLSU region; EF1-983F and EF1-1567R (Rehner and Buckley, 2005) for the translation elongation factor 1-α (tef1α) region; ARPB2-6F and ARPB2-7R for RNA polymerase II second largest subunit (rpb2) region; and Am-b-tubulin F and Am-b-tubulin R (Cai et al., 2014) for beta-tubulin (βtubulin) region.

Sequence Alignment and Phylogenetic Analyses
Thirty-four sequences (seven for ITS, eight for nrLSU, seven for rpb2, six for tef1-α, and six for β-tubulin) were newly generated for this study and deposited in GenBank (Table 1) 2 . Additional sequences were retrieved from previously published articles and GenBank ( Table 1). The sequences were aligned using MAFFT v7.310 (Katoh and Standley, 2013) and edited manually when necessary using BioEdit v7.0.9 (Hall, 1999). The poorly aligned portions and divergent regions were eliminated using Gblocks v0.91b (Castresana, 2000;Talavera and Castresana, 2007). A concatenated dataset (including ITS, nrLSU, rpb2, tef1-α, β-tubulin) comprising 312 sequences was constructed using Phyutility v2.2 (Smith and Dunn, 2008) and used for phylogenetic analyses. Before using the concatenated dataset for phylogenetic analyses, the Incongruence Length Difference test in PAUP v4.0a168 (Swofford, 2002) was performed to detect any conflicts between the gene regions. As no incongruence (P = 0.363000) was detected, the maximum likelihood (ML) and Bayesian inference (BI) were used on the concatenated alignment for phylogenetic tree inference. The ML analysis was performed using RAxML v7.9.1 (Stamatakis, 2006) under the GTR + GAMMA + I nucleotide substitution model and performing non-parametric bootstrapping with 1,000 replicates. The BI was performed in MrBayes v3. 2 (Ronquist et al., 2012). The best substitution model was determined using the Akaike Information Criterion implemented in jModeltest v.2 on the CIPRES Science Gateway v3.1 (Miller et al., 2010). The BI was conducted with the following parameters: two runs, each with four simultaneous Markov chains, and trees were summarized every 1,000 generations. The analyses were completed after 20,000,000 generations when the average standard deviation of split frequencies was 0.002200 for the five-gene analysis, and the first 25% generations were discarded as burn-in. The phylograms from ML and BI analyses were visualized with

Species name
Collection or collector no. Country of origin GenBank accession no.
Frontiers in Microbiology | www.frontiersin.org The new sequences generated in this study are highlighted in bold font.References to sequences retrieved from GenBank: Cai et al. (2014Cai et al. ( , 2016, Thongbai

Analysis of Toxins by Liquid Chromatography-High-Resolution Mass Spectrometry
Dried basidiomata of the target taxon have been used for toxin analyses using the method of Lüli et al. (2019). Toxins were extracted from basidiomata, using methanol-water-0.01 M hydrochloric acid (5:4:1, vol/vol) as the extraction buffer. Dried material (0.05 g) was crushed into fine powder in a mortar and pestle with liquid nitrogen. Then, 1.5 mL aforementioned buffer was added, and the suspension transferred into 1.5-mL centrifuge tubes. The tubes were kept at room temperature for 30 min, followed by centrifugation (12,000 rpm) for 3 min. Finally, the supernatant was transferred into new centrifuge tubes for mass spectrometry analysis. The presence of cyclic peptides, especially α-amanitin, β-amanitin, phalloidin, and phallacidin (standards provided by Sigma Chemical Co, United States), was evaluated through the liquid chromatography-high-resolution mass spectrometry (LC-HRMS) using 1290 Infinity II HPLC systems coupled with 6540 UHD precision mass Q-TOF instruments under the conditions listed in Table 2.

Species Delimitation
Most of the fatal mushroom poisonings are caused by lethal amanitas belonging to A. sect. Phalloideae (Bresinsky and Besl, 1990;Unluoglu and Tayfur, 2003;Cai et al., 2014Cai et al., , 2016Li et al., 2015;Cui et al., 2018). In tropical Africa, very few lethal amanitas have been reported (Walleyn and Verbeken, 1998;Fraiture et al., 2019;. Only six species are known from tropical Africa including A. alliodora, A. murinacea, and A. thejoleuca described from Madagascar and A. bweyeyensis, A. harkoneniana, and A. strophiolata described from DR Congo (central Africa). Amanita albolimbata represents a new lethal Amanita from tropical Africa and can be recognized by its white basidiomata with a convex or applanate pileus without umbo, a limbate volva with inner part composed of abundant inflated cells, and broadly ellipsoid to ellipsoid basidiospores.
The multigene phylogenetic analyses revealed that A. albolimbata is an independent lineage in A. sect. Phalloideae. Surprisingly, the species is genetically distant from other African species such as A. alliodora, A. bweyeyensis, and A. harkoneniana that form a clade. Among lethal amanitas from tropical Africa, A. albolimbata is similar to A. strophiolata because of the white basidiomata, but unfortunately, we do not have any material of the latter species to test its phylogenetic relationship to other species. However, A. strophiolata presents some distinct morphological characteristics that clearly separate it from A. albolimbata. Amanita strophiolata was described by Beeli (1927Beeli ( , 1935 from DR Congo and is distinguished by its umbonate pileus, often yellowish at center, the absence of volval remnants on pileus, ellipsoid to elongate basidiospores (10-11 × 6-7 µm), a distinctive annulus in the form of a funnel, and larger basidiomata.
In the multigene phylogenetic tree, the African and Australian taxa are basal. This suggests that the lethal amanitas originated from the palaeotropical areas. Cai et al. (2014) also suggested a possible palaeotropical origin of lethal amanitas and highlighted the need for more molecular-phylogenetic studies on collections from the tropics and the Southern Hemisphere.

Toxicity in Amanita
For centuries, wild mushrooms have been consumed massively and popular in the human diet because of their matchless taste, protein content, and medicinal properties (de Román et al., 2006;Cheung, 2010). However, the high interest on wild mushroom collections and consumption could increase the risk of poisoning by lethal mushrooms. During picking, confusions could easily be made between edible and poisonous mushrooms because of their morphological similarities. Many mushroom poisoning cases have been reported worldwide and have mainly been caused by members of A. sect. Phalloideae (Zhang et al., 2010;Cai et al., 2014Cai et al., , 2016Yang, 2015;Li et al., 2015Li et al., , 2020Thongbai et al., 2017). Consequently, much attention has been devoted to the species producing toxins within A. sect. Phalloideae Li et al., 2014;Garcia et al., 2015;Cai et al., 2016).
Until now, no lethal amanitas had been reported from West Africa. However, lethal amanitas have been documented from Central Africa and Madagascar (Fraiture et al., 2019). Amanita albolimbata represents the first lethal species of A. sect. Phalloideae known from West Africa. The most notorious toxins, α-amanitin, β-amanitin, and phallacidin, have also been detected in the species.
Numerous amanitoid taxa are harvested and consumed by local people in tropical Africa (Codjia and Yorou, 2014;Yorou et al., 2014;Boni and Yorou, 2015;De Kesel et al., 2017;Fadeyi et al., 2017;Milenge et al., 2018;Soro et al., 2019). Because of the whitish color of the basidiomata, A. albolimbata can be confused with A. subviscosa Beeli. Amanita subviscosa is commonly harvested and used as food by local people in Benin Boni and Yorou, 2015;Fadeyi et al., 2017Fadeyi et al., , 2019. Still, A. subviscosa displays contrasting morphological characteristics with A. albolimbata by a slightly squamulose and viscous pileus, slightly striated margin, slightly bulbous, and slightly furfuraceous and hollow stipe, with a distinctive membranous volva (Beeli, 1935). The lack of A. albolimbata in various ethnomycological investigations Boni and Yorou, 2015;Fadeyi et al., 2017;Soro et al., 2019) attests that either local people are aware about its toxicity, or some fatal but unrecorded cases did occur within rural communities. However, it is important to educate local people on the best ways to discriminate morphologically close taxa in order to avoid the consumption of lethal Amanita species for an effective prevention of future poisoning incidents.

DATA AVAILABILITY STATEMENT
The datasets generated for this study can be found in the online repositories. The names of the repository/repositories and accession number(s) can be found in the article/ Supplementary Material.

AUTHOR CONTRIBUTIONS
ZLY, JEIC, and NSY conceived and designed the research. JEIC collected the species, performed the molecular phylogenetic analyses and the taxonomic studies, and wrote the first draft of the manuscript. JEIC and QC generated the DNA sequences. JEIC and SWZ carried out the cyclic peptide toxins analyses. QC, HL, MR, NSY, and ZLY critically revised and approved the final manuscript. All authors contributed to the article and approved the submitted version.

FUNDING
This study was supported by the International Partnership Program of Chinese Academy of Sciences (No. 151853KYSB201 70026), Yunnan Ten- Thousand-Talents Plan -Yunling Scholar Project, and the FORMAS Grant (No. 226-20141109).

ACKNOWLEDGMENTS
We are grateful to Gang Wu and Bang Feng from Kunming Institute of Botany, CAS, for providing additional materials and images for this study. We thank Yang-Yang Cui, Pan Meng Wang, Si-Peng Jian, Xin Xu, and Kui Wu (Kunming Institute of Botany, CAS) for their kind assistance.