Validation of two Amanita species from eastern North America: A.rhacopus sp. nov. and A.variicolor sp. nov.

Abstract Members of the mushroom genus Amanita usually can easily be identified to the genus in the field, however, species circumscription and identification are often problematic. Several names have been misapplied and cryptic species exist. Here, we formally describe and validate two new species of Amanitasect.Vaginatae from eastern North America that were recognised under the umbrella European names A.ceciliae by past authors: Amanitarhacopussp. nov. and Amanitavariicolorsp. nov.


Introduction
The genus Amanita is one of the best-known genera within the Agaricales. The genus contains edible and poisonous mushrooms, is distributed worldwide and is entirely or mostly ectomycorrhizal depending on which side one stands in its currently debated circumscription (Redhead et al. 2016, Tulloss et al. 2016, Hawksworth 2016. Many Amanita species are common, quite robust and readily recognisable in the field. However, cryptic species exist and several European names have been misapplied to taxa from other continents. In eastern North America, more than 75 species are documented and well characterised although several taxa are still not validly published (Lamoureux 2006, Tulloss 2017. Following an extensive survey conducted by one of us in the province of Québec, Canada, between 1985 and2005, it was concluded that the region hosts at least 52 taxa of Amanita (Lamoureux 2006), which is about twice the 27 species reported by Pomerleau in 1980 for the same region. Amongst these taxa, 11 appear new to science, not corresponding to any known American or European species. Although these species were provisionally named (Lamoureux 2006), the names are not valid since the descriptions of the species lacked a diagnosis and a reference to a holotype.
Two of the new species proposed by Lamoureux (2006) corresponded to what Pomerleau (1980) originally called A. ceciliae (Berk. & Broome) Bas or its later synonym A. inaurata Secr. ex Gillet. A. ceciliae is a European taxon first described in 1854 and this name was later used worldwide to label species with similar appearance. These two new species were described as medium size Amanita with cap striated at the margin, a cylindrical stipe without annulus and a friable universal veil, greyish with age with orange tint or not, often leaving some remnants on the cap and at the stipe base. A. rhacopus Y. Lamoureux nom. prov. is brown to dark greyish-brown whereas A. variicolor Y. Lamoureux nom. prov. has variable intensity of pileus colour but differs mainly by its orange tint especially at the stipe base. Both are in conifer forest mixed with Betula. The concept of these two Amanita species became well accepted in eastern North America, although A. ceciliae ss. auct. amer. is still used by many for both species to avoid the use of provisorum nomen (see Tulloss 2017). Recently, DNA sequences of several eastern North America and Québec collections of A. rhacopus and A. variicolor were released in Genbank (R. Tulloss, unpublished work). The present work aims at validating these two names by designating holotypes and providing diagnoses, detailed descriptions as well as ITS-rDNA barcode sequences (Schoch et al. 2012).

Specimens examined
Type specimens are deposited in the Cercle des mycologues de Montréal Fungarium, (CMMF). Additional specimens are in the Royal Ontario Museum Fungarium (TRTC), in CMMF or in the private collections of H. Lambert (HL), R. Lebeuf (HRL) or R. Labbé (RLA).

Morphological examination
Collections examined in this study were photographed in the field and macromorphological features were derived from both field notes and pictures. Microscopic studies were performed in saline solution on fresh material or in 3% ammoniac on exsiccata. Melzer's reagent was used for amyloidity, Cotton Blue for cyanophily and Congo Red for tissue staining. Dimensions of microscopic elements are given as: [a/b/c] (min) D1-D9 (max) where a, b and c represent the number of elements measured, the number of specimens and collections from which the elements were studied, respectively; min and max, the extreme values of the distribution; D1 and D9, the first and ninth decile. Q denotes the length/width ratio of a basidiospore in side view, Qm refers to the arithmetical mean. All microscopic elements were obtained using a Leitz Ortholux II or an Olympus CH-2 microscope equipped with digital camera and were measured from pictures using Piximetre software v.5.6 (Alain Henriot, France).

DNA sequencing and analyses
ITS-rDNA barcode sequences were obtained following Dentinger et al. (2010). DNA was extracted from dried herbarium specimens (Québec collections) or from fresh tissue blotted on FTA cards (Ontario collections). BLAST searches (Altschul et al. 1990) were conducted in Genbank and in UNITE (Kõljalg et al. 2005, Abarenkov et al. 2010 to compare the new sequences with those available in these databases.

DNA sequence analyses
The sequences we obtained from seven A. rhacopus collections were identical to each other, whereas those obtained from three specimens of A. variicolor differed at two nucleotide positions. The ITS sequences of the holotypes were subjected to BLAST searches against the GenBank and UNITE databases. The holotype sequence of A. rhacopus CMMF002171 retrieved 20 sequences with 100% identity deposited as The ITS sequence of the holotype of A. variicolor CMMM003787 has highest similarities with two sequences deposited in Genbank as Amanita sp. 'variicolor', one by Tulloss et al. (KP711844, specimen CMMF003463 from Québec;unpublished) and one by Bérubé et al. (KJ638268, specimen HL0846, also from Québec; unpublished), showing, respectively, zero to four mismatches out of 509 aligned nucleotides. Other highly similar sequences were from collections labelled Amanita sp., one from Arizona (MG518639, T.A. Clements, unpublished, which differs at six sites) and one from North Carolina (AY456335, Edwards et al. 2004, which differs at nine sites).  (1854) Diagnosis. Amanita rhacopus differs from other species of Amanita section Vaginatae by its brown to dark grey-brown pileus, stipe white at times covered with grey chevrons, universal veil grey leaving small to large flakes on pileus and annulus-like remnants at the stipe base, found in stands of conifers (Abies, Picea, Pinus, Tsuga) mixed with Betula.
Description. Pileus 40-80 mm wide, ovoid to rounded conic slightly umbonate to applanate, smooth, brown to greyish-brown, with time darker in the centre and over inner ends of marginal striations, often with grey velar flakes, margin striated. Lamellae free, crowded, greyish near the pileus margin or completely greyish with age, lamellulae numerous, truncated, of very diverse lengths, unevenly distributed, edges finely powdered, white to whitish. Stipe 70-120 × 7-13 mm, cylindrical (not bulbous), flocculose and white first, then smooth to appressed fibrillose, whitish to greyish, at times with chevron-forming greyish fibrils, without annulus, with grey annulus-like remnants at the base. Universal veil friable, grey, leaving flakes on the pileus and annuluslike remnants at the stipe base. Partial veil absent. Context whitish, unchanging when cut or bruised, odour and taste not distinctive.
Ecology and distribution. Solitary or scattered in stands of conifers (Abies, Picea, Pinus) mixed with Betula, on mesic to sub-mesic soil, never seen in plantations, from July to September in Québec and, according to sequences in Genbank, in all eastern North America down to Tennessee and Texas.
Etymology. The epithet rhacopus refers to the Greek ῤάκος, meaning piece of cloth and πούς, meaning foot.  Diagnosis. Amanita variicolor differs from other species of Amanita sec. Vaginatae by its versicolour (straw-yellow, orange-brown to blackish brown) pileus, stipe white then covered with brown olive to orange chevron-forming fibrils, stipe base dark orange to rusty, universal veil grey to orange-grey leaving small to large flakes on the pileus and one or two strips at the stipe base, found mainly with Abies and Betula.
Holotype. CANADA, Québec: Rawdon, in a mixed forest of Abies balsamea and Betula papyrifera, on moist soil close to a bog, 19 August 2003, CMMF003787, ITS Genbank accession number MG734656.
Etymology. The epithet variicolor refers to the very variable colour of the pileus.

Description.
Pileus 40-100 mm wide, ovoid to rounded conic at first, then plane with an umbo, smooth, olive yellow, straw yellow, bronze, olive brownish to brownblack and then tinged with olive or orange yellow towards the margin, at times darker in the centre and over inner ends of marginal striations, often with small to large grey or orange grey velar flakes, margin striated. Lamellae free, subcrowded, whitish, greyish to salmon tints. Stipe100-200 × 8-17 mm, cylindrical (not bulbous or barely), flocculose and white at first, typically covered all over by chevron-forming rusty-orange fibrils on a whitish background when mature, without annulus, base always rusty orange with one or two greyish-orange velar strips. Universal veil friable, grey to orange grey, often leaving small to large flakes on the pileus and one or two strips at the stipe base. Partial veil absent. Context white, unchanging when cut or bruised, odour not distinctive, taste not recorded.
Ecology and distribution. Solitary, sometimes scattered to gregarious, in stands of conifers (Abies, Picea, Pinus, Tsuga) mixed with Betula, on sub-hydric to mesic soil, never seen in plantations, from July to September in Québec and Ontario. Possibly also present further south (see Discussion).

Discussion
Amanita rhacopus and A. variicolor belong to subgenus Amanita section Vaginatae (Fr.) Quél. due to the absence of a basal bulb and a partial veil, inamyloid basidiospores and absence of clamp connections at the base of basidia (Bas 1969, Yang 1997. Tulloss (2017) lists 296 names in this section, of which 97 correspond to accepted species, 132 are cryptonyms and 57 are provisory names. Many ITS sequences from members of sect. Vaginatae have been deposited in public databases, however, most are unidentified, misidentified or refer to provisory names. In addition, ITS variation within the section is very high and comprises numerous insertion/deletion events that preclude unambiguous sequence alignment in many positions. For these reasons, we have refrained to place A. rhacopus and A. variicolor in an ITS phylogeny.
In general, A. variicolor is larger than A. rhacopus. The pileus of A. rhacopus is greyish-brown becoming darker with age, never with an orange tint as found on A. variicolor. The lamellae of A. rhacopus become greyish with time, whereas the lamellae of A. variicolor become greyish-salmon. The universal veil texture of A. rhacopus is stronger than in A. variicolor, leaving a pseudo-annulus at the stipe base, which is more distinct on young basidiomes. In contrast, the universal veil of A. variicolor leaves light orange strips on the stipe base. The basal extremity of the stipe is always rusty orange in A. variicolor, a characteristic which is never seen in A. rhacopus. Both species developed chevron-like motifs on the stipe surface and are growing under conifers and Betula mixed forest. A. variicolor appears to be less frequent than A. rhacopus. They were never observed in pure plantations, only in native forest. A. variicolor is growing on rich mossy soil, A. rhacopus on dry soil often in spine conifer litter.
Although nucleotide sequence variation alone cannot be used for the circumscription of fungal species, empirical observations as well as a comprehensive study by Nilsson et al. (2008) indicated that ITS intraspecific variation in the Basidiomycetes is typically in the 0-3% range. Here, we found that ITS sequences amongst 27 collections of A. rhacopus, from Québec to Texas, were identical and differ from at least 4% from other Amanita sequences that have been deposited in public databases so far. This fact, along with its morphological uniqueness, strongly supports the recognition of A. rhacopus as a new and widespread species in eastern and southern North America.
While A. rhacopus shows remarkable ITS sequence homogeneity from Québec to Texas, A. variicolor sequences vary by 0.78% amongst collections from Québec and Ontario and differ by 1.76% from two sequences deposited in Genbank as Amanita sp., one from North Carolina (AY456335) and one from Arizona (MG518639). The former sequence comes from a fruiting body collected in a loblolly pine forest, whereas the latter was retrieved from a soil sample under ponderosa pine, douglas fir and gambel oak. Whether or not these two sequences belong to A. variicolor needs further investigation. If this is the case, then the distribution range of this species would also encompass areas in the eastern and southern U.S.A.