Loose Ends in the Cortinarius Phylogeny: Five New Myxotelamonoid Species Indicate a High Diversity of These Ectomycorrhizal Fungi with South American Nothofagaceae

This paper is a contribution to the current knowledge of taxonomy, ecology and distribution of South American Cortinarius (Pers.) Gray. Cortinarius is among the most widely distributed and species-rich basidiomycete genera occurring with South American Nothofagaceae and species are found in many distinct habitats, including shrublands and forests. Due to their ectomycorrhizal role, Cortinarius species are critical for nutrient cycling in forests, especially at higher latitudes. Some species have also been reported as edible fungi with high nutritional quality. Our aim is to unravel the taxonomy of selected Cortinarius belonging to phlegmacioid and myxotelamonioid species based on morphological and molecular data. After widely sampling Cortinarius specimens in Patagonian Nothofagaceae forests and comparing them to reference collections (including holotypes), we propose five new species of Cortinarius in this work. Phylogenetic analyses of concatenated rDNA ITS-LSU and RPB1 sequences failed to place these new species into known Cortinarius sections or lineages. These findings highlight our knowledge gaps regarding the fungal diversity of South American Nothofagaceae forests. Due to the high diversity of endemic Patagonian taxa, it is clear that the South American Cortinarius diversity needs to be discovered and described in order to understand the evolutionary history of Cortinarius on a global scale.


DNA Extraction, PCR Amplification and Sequencing
To establish phylogenetic relationships, ITS-rDNA sequences were produced as previously described [44] using the primers ITS1 and ITS4 [48]. The rDNA LSU region was amplified with the primer combination LR0R and LR05 [48]. PCR amplifications of RPB1 domains A-C were made with the primer combination RPB1-A and RPB1-C [49]. Sequences were assembled and edited with Sequencher 4.1 (Gene Codes, Ann Arbor, Mich., USA). As a first step, a Blast search was conducted in UNITE (https://unite.ut.ee 1 April 2017). Sequences of closely related Cortinarius species were then downloaded from GenBank (http://ncbi.nlm.nih.gov 1 April 2017) and UNITE. ITS sequences from several types of specimens were also included in the study (Table 1). A total of 28 ITS sequences from the five new species were produced for this study. In addition, eleven LSU sequences and five RPB1 sequences were generated. The newly generated sequences were submitted to GenBank under the accession numbers MN707570-90, MT925622-25; MT925951-53, MW546828-32.

Data Analysis
Data analysis was carried out in a two-step process: A first analysis was based on ITS rDNA sequences only. The best reference sequence database is available for this DNA barcoding region, including sequences from several Patagonian holotypes. As a second step, we aimed at placing our terminal clades into known Cortinarius sections or lineages based on concatenated ITS-LSU and ITS-LSU-RPB1 sequences. This was done separately due to the fact that different reference sequences were available for the LSU and RPB1 markers.
A total of 64 rDNA ITS sequences were aligned and manually adjusted in MEGA X [50]. Reference sequences were selected and downloaded for closely relates species based on morphology or based on sequence similarity resulting from BLAST analyses. The evolutionary history was inferred by using the Maximum Likelihood method based on the Hasegawa-Kishino-Yano model + G, parameter = 0.2332. All positions with less than 95% site coverage were eliminated. There was a total of 560 positions in the final dataset.
To evaluate the robustness of the branches in the phylogenetic trees, parsimonybased bootstrap analyses were applied. The bootstrap analyses were conducted using 1000 replications, an SPR search method, and search level 5. The alignment is composed of 653 nucleotides (including gaps). Bayesian Posterior Probabilities were calculated with Mr Bayes 3.2. [51]. Bayesian analysis was carried out with two independent four-chain runs, sampling over 2 million generations.
In addition, two combined phylogenetic analyses were carried out: the first was based on rDNA data only (ITS and LSU), and the second on a combined dataset containing combined sequences spanning RPB1 regions, the ITS regions, and about 600 bases of the 5 -terminal large subunit (LSU) domain (D1/D2). The alignment of the combined ITS and LSU data contained 1328 positions, and 52 taxa. ML analysis was carried out based on the best model (GTR + G parameter = 0.2131) and 1164 positions were analyzed. The tree with highest log likelihood was −10619.28. The alignment of the combined RPB1, ITS, and LSU sequences contained 99 sequences and 2773 positions after the exclusion of ambiguous regions. ML analysis was carried out based on the best model (Tamura 3-parameter + G, parameter = 0.2462), and 1697 positions were analyzed. Two separate Mr Bayes runs were run under the general time-reversible model with gamma-distributed rate variation. Runs included four incrementally heated chains that were run for 10 million generations each, sampling every 100th generation and with the first 2.5 million generations discarded as burn-in. For further evaluation of branch robustness, parsimony-based bootstrap analyses were applied as described above, with 1000 replications.
Statistical analyses were performed with the width, length, and volume of spores. The width, length, and volume of spores between the species did not meet the assumptions of normal distribution and equal variances using the Shapiro-Wilk and Levene tests [52]. Therefore, differences in the width, length, and volume of spores between species were analyzed using non-parametric Kruskal-Wallis ANOVAs performed at the 0.05 significance level, using the statistical package InfoStat for Windows version 2017 [53]. Test for normal distribution and QQ-Plots were performed with R package (R Core Team 2019).

Molecular Data
The ITS-based phylogeny with the best ML log likelihood -2174.22 allowed for the best comparison to available reference sequences, including sequences generated from type specimens. All five species described here form well-supported clades (Bayesian posterior probability(BPP) > 0.99, Bootstrap Score (BS) > 95%) in the ITS phylogeny ( Figure 1). The sister-group relationships are well-resolved in C. gracilentus only, where closely related reference sequences of C. avellaneus are available. Cortinarius egonii is sister to C. rhodophyllus Moser & Horak (BPP = 0.95), but only one reference sequence is available for C. rhodophyllus, but it was not obtained from type material. The ITS sequence generated from the type of section Myxotelamonia, C. cinereobrunneus IB19630258, could not be aligned with the sequences of the five new species, showing that these new species do not belong to section Myxotelamonia. The most closely related sequences from Moser and Horak s South American holotypes were C. mitis (Subgenus Myxacium, Section Ochroleuci), which is related to the clade with the new species C. neuquensis (BPP = 0.83); C. micaceus (subgenus Sericeocybe strips Nothoanomalus), C. cinereus (Section Telamonia) and C. nitellinus (Section Formiores) are weakly related to the new species C. voluptatis (BPP = 0.77); and C. avellaneus (Section Myxotelamonia), C. semiamictus (Subgenus Paramyxacium, stirps Myxacioides) and C. macilentus (Section Myxotelamonia) are related to the new species C. gracilentus (BPP = 1.00).
The concatenated analysis of the LSU-ITS rDNA and RPB1 regions (best ML tree with log likelihood −17366.88) indicates a possible common origin for C. pseudoxiphidipus, C. voluptatis and C. egonii (BPP 0.929). Our data also suggest that C. neuquensis could be related to a clade containing C. lustratus Fr., C. cretax Soop and C. pinophilus Soop (BPP 0.871) (Supplementary Figure S1).

Taxonomic Data
All the species included in the morphological study differ significantly in the dimensions of their spores (Kruskal Wallis, H = 1957.22, p < 0.0001), confirming statistically different clouds of data ( Figure 2). Statistical analysis of basidiospore measurements confirmed that the spores of C. neuquensis are significantly smaller (in both width and length) than spores of other Cortinarius species from this study. On the other hand, C. voluptatis' spores are significantly bigger (in both width and length) from any other species in this study (Kruskal Wallis, H = 1957.22, p < 0.0001, Figure 3). All species can clearly be separated from each other based on their spore size and shape (as a function of Q = length/with). The species epithet refers to Dr. Egon Horak, globally recognized expert on the genus Cortinarius. His work has inspired many mycologists around the world to discover the fascinating world of Cortinarius taxonomy, including the authors of this paper.

Taxonomic Data
All the species included in the morphological study differ significantly in the dimensions of their spores (Kruskal Wallis, H = 1957.22, p < 0.0001), confirming statistically different clouds of data ( Figure 2). Statistical analysis of basidiospore measurements confirmed that the spores of C. neuquensis are significantly smaller (in both width and length) than spores of other Cortinarius species from this study. On the other hand, C. voluptatis' spores are significantly bigger (in both width and length) from any other species in this study (Kruskal Wallis, H = 1957.22, p < 0.0001, Figure 3). All species can clearly be separated from each other based on their spore size and shape (as a function of Q = length/with).  MycoBank MB 836828

Etymology
The species epithet refers to Dr. Egon Horak, globally recognized expert on the genus Cortinarius. His work has inspired many mycologists around the world to discover the fascinating world of Cortinarius taxonomy, including the authors of this paper.  MycoBank MB 836828

Etymology
The species epithet refers to Dr. Egon Horak, globally recognized expert on the genus Cortinarius. His work has inspired many mycologists around the world to discover the fascinating world of Cortinarius taxonomy, including the authors of this paper.     MycoBank MB 836579

Etymology
The species epithet refers to the slender habitus of the basidiomata.

Diagnosis
Cortinarius gracilentus ( Figures 4B and 5D) is characterized by a glutinous, hygrophanous, cinnamon brown pileus, stipe cylindrical, dry, fibrous, white, with the remains of a caramel veil, and cocoa brown lamellae in young specimens. Basidiospores are elliptic, verrucose, 11.1 ± 0.7 × 7.1 ± 0.5 µm. Basidia stain with cotton blue and usually grow alone. Macrocharacters PILEUS 2.7-4.4 cm in diam., convex in young specimens, hemispherical to convex and plano-convex with age, pileus margin slightly bent, in young specimens slightly involute. Pileus surface glutinous, slightly hygrophanous, smooth. Pileus color varies between cream (9D2) to cork (13B7) at the margin with clearly darker colors towards the pileus center ranging between artificial brown (8L6) and Tuscany brown (7L11-7L12), later mixed with an orange tone.
Context ochraceous when fresh, with paler colors towards the margin of the stipe. Smell fungal to sweetish in gills. Taste mild. Usually growing in groups but not cespitose.
Clamp connections present in all tissues.

Ecology and Distribution
Forest type-Nothofagus dombeyi, N. pumilio and N. antarctica; observed in May. The monthly average temperature for May in the area is 8 • C (max/min 13/5 • C), with a total of 95 mm precipitation (weather station El Bolsón Aero, data from 2017). Soil pH = 5.8.
Other Notes: Cortinarius egonii has a mean of 0 bp within species variation in the ITS region, except for collection HCFC C80, that differs from the ITS sequence of the holotype by 0.7% (five substitutions and indels). C. egonii is the only representative of the UNITE SH1142013.08FU and differs by 4% (22 substitutions and indels) from the most closely related reference sequence C. rhodophyllus (GenBank Acc. No. KJ421051, Chile). The difference to all other species ranges between 4.5 and 9.0% (29 and 50 substitutions and indels) (MW = 33, SD = 6).
Clamp connections present in all tissues.

Ecology and Distribution
Forest type-Nothofagus dombeyi and N. antarctica; observed in April. The monthly average temperature in the area is 8.9 • C (max/min 13.7/2 • C), with a total of 82 mm precipitation in April (weather station Lago Cholila, data from 2017). Soil pH = 5.8.
Other Notes: Cortinarius gracilentus has no within species ITS differences and differs by 0.1% (one substitution or indel) from the included subclade (MES-1597, MES-1801). The most closely related species is C. avellaneus Moser, which differs by 4.5% in the ITS region (23 substitutions or indels). The difference to all other species ranges from 4.5 to 9% (23-49 substitutions or indels MW 31 + −8 bp).
Based on ITS sequences only, where the majority of reference sequences are available, C. gracilentus is closely related to C. avellaneus. Morphological characters confirm this relationship: C. avellaneus has darker avellaneous to umber brow dry pilei, yellow-rusty brown lamellae and further differs by ellipsoid to amygdaliform, strongly verrucose spores with an inconspicuous plage. The second species belonging to this stirps Avellaneus, C. fulvoconicus Moser, differs by the vividly red-brown pileus colours, the dry pileus, and narrower spores. Moser and Horak (1977)

Diagnosis
Cortinarius neuquensis (E and A) has medium-sized basidiomata (3.5-5 cm pileus diam.) and is characterized by a light ochre brown to reddish brown, glutinous, not hygrophanous pileus, and a white, cylindrical, fragile, and dry stipe. The lamellae are pale argillaceous, with an entire margin and sinuate. Basidiospores are amygdaliform to elliptical and finely verrucose, (6. Macrocharacters PILEUS 3.6-5.0 cm in diam., low convex, occasionally conical in young specimens, pileus margin somewhat wavy with age. Pileus surface glutinous, not hygrophanous, smooth. Pileus color varies between light ochre (10C1), golden light brown (10H4) to reddish brown, becoming with age honey ochre (12J6). The margin of the pileus is straight, and in young specimens slightly bent.
Clamp connections present in all tissues.

Ecology and Distribution
Forest type-mostly in forest of Nothofagus antarctica, but also in forests of N. dombeyi and in mixed forests of Lophozonia alpina-L. obliqua; observed during May. The monthly average in May temperature in the area is 9 • C (max/min 14/6 • C), with a total of 88 mm precipitation (weather station Lago Ñorquinco, data from 2017). Soil pH = 6.1. This species has also been detected in Chile (KY462703, KY462509) Other The difference from all other species ranges 5-8% (29-50 substitutions and indels, MW = 32, SD = 6). Cortinarius verniciorum and C. ducamarus Soop from New Zealand belong to the same section Verniciori Soop. These two species differ by orange brown basidiomes with strongly viscid pilei. Cortinarius viscovenetus Horak is morphologically similar. However, the spores are significantly larger (10-11 × 6.8-7.2 µm). The ITS sequence similarity of these two species is <97% (unpublished).
CONTEXT (flesh) firm, corky with pale colors. Smells slightly fruity. Tastes mild. Usually growing in scattered small groups. Macrochemical reactions; 20% KOH on exsiccate material slightly yellowish. Yellowish reaction on exsiccate lamella with 3% KOH. No fluorescence was detected at 350 nm nor at 254 nm (in exsiccate material).

Ecology and Distribution
Forest type-Nothofagus dombeyi; observed in April. The monthly average temperature in the area is 8.9 • C (max/min 13.7/2 • C), with a total of 82 mm precipitation in April (weather station Lago Cholila, data from 2017). Soil pH = 5.8.
Other Notes: Cortinarius pseudoxiphidipus has a mean within species ITS difference of 0 bp and differs by >5% (28-48 substitution and indels MW 29 ± 4 bp) from the other species included in the phylogeny, and the difference to C. xiphidipus, which it resembles morphologically, is >6%. Cortinarius pseudoxiphidipus has smaller basidiomes, lamellae with copper colors, and larger basidiospores than C. xiphidipus. Cortinarius xiphidipus has argillaceous lamellae, and elliptical, finely warty spores of 6-8 × 4-4.8 µm. From Latin, the one who provides pleasure or joy. The species epithet refers to the red wine colors of the pileus in young specimens.
Usually growing in scattered small groups. Macrochemical reactions; 20% KOH on exsiccate. Pileus slightly yellowish. No fluorescence was detected at 350 nm nor at 254 nm (in exsiccate material).
Clamp connections present in all tissues.

Ecology and Distribution
Forest type-Lophozonia alpina and L. obliqua but occasionally with N. antarctica; observed in May. The monthly average temperature in the area is 7 • C (max/min 12/2 • C), with a total of 136 mm precipitation in May (weather station Chapelco Aero, data from 2017). Soil pH = 5.8.
Other Notes: Cortinarius voluptatis has a mean within species ITS difference of 0.1% (one substitution or indel) and differs by 2.3% (13 substitutions or indels) from the most closely related species C. cinereus Moser. The difference to all other species ranges between 3 and 9% (17 and 49 substitutions or indels MW = 27, SD = 7).
Based on morphology, C. voluptatis could belong to Section Myxotelamonia Subsect. Lilacifolii [5]. The type of the section, C. roseopurpurascens Moser & Horak, somewhat resembles C. voluptatis, but differs by the stipe being lilaceous, it has significantly smaller spores (10-12.5 × 6.5-7.5 µm) and it lacks a veil. Moser & Horak [5] hypothesized that this section could represent an endemic South American complex of species without morphologically similar species in the Northern Hemisphere. Old specimens of Cortinarius voluptatis could be confused with C. pseudoxiphidipus because of the size and fruiting habit, but C. voluptatis can be differentiated by the darker color of the veil and the larger spores. Cortinarius juglandaceus Soop resembles C. voluptatis but is not closely related. C. juglandaceus is viscid whereas C. voluptatis has a glutinous pileus surface and has larger spores [54]. Cortinarius juglandaceus occurs in Nothofagus forests in New Zealand.

Discussion
Five new species of Cortinarius are proposed here based on morphological and molecular data. Cortinarius is an ectomycorrhizal genus [55] and it is possible that associations with specific host tree species help to explain high regionalism and habitat relationship. Several Cortinarius species described from Patagonia are thought to associate only with a specific Nothofagaceae host tree species, including, C. magellanicus Speg., C.  [1]. Due to undersampling and the difficult taxonomy of Cortinarius in Patagonia, it is possible that specific host associations could be more frequent than previously assumed. The newly described C. pseudoxiphidipus could represent one of these species with strong host preferences or host specificity since it was found only associated with N. dombeyi. Similarly, C. egonii and C. gracilentus showed specificity at the genus level and were only found with Nothofagus species. Within the genus Cortinarius there are also several ectomycorrhizal species with more generalist plant host associations, such as C. albocanus They are all reportedly associated with Nothofagaceae species present in Patagonia [1]. We observed the same generalist pattern for C. voluptatis and C. neuquensis, which are associated with both Lophozonia and Nothofagus spp.
The phylogenetic placement of the five new species proposed in this work into subsections, sections, or even subgenera was not possible due to limited phylogenetic resolution and the lack of closely related reference taxa. The combined phylogenetic analysis of ITS, LSU and RPB1 resulted in a well-supported clade consisting of C. pseudoxiphidipus, C. voluptatis and C. egonii. They all have a similar habitus. However, we are reluctant to define new, possibly endemic South American sections of Cortinarius based only on three species. The lack of reference data shows that the South American Cortinarius diversity is still widely underexplored and it also suggests the presence of endemic Cortinarius lineages in this area. Up until now there is very little overlap between Cortinarius species described from Australia and New-Zealand [38] with the species reported from South American Nothofagaceae forests. Given the high diversity of Southern Hemisphere Cortinarius diversity and the lack of available reference sequences, it is not surprising that the phylogeny of Cortinarius species remains largely unresolved. Based on available data it seems likely that that endemic South American Cortinarius lineages exist. The recently proposed Cortinarius section Austramericani [16] (p. 1130) could be an example for such an endemic lineage of Patagonian Cortinarius species associated with Nothofagus species. More intense investigations on the diversity of this genus in South America are urgently needed, as this will not only clarify the Patagonian Cortinarius diversity, but will also provide fascinating insights into the evolution of ectomycorrhizal associations on a global scale.
"Southern Gondwana" connections are often explained by the presence of specific host plants in the Southern Hemisphere that are absent in other regions and vice versa [56][57][58]. Due to their association with Nothofagaceae forests, Southern Gondwanan connections could also be assumed for several lineages of Cortinarius, as already proposed for other genera of ectomycorrhizal fungi [59][60][61][62]. For example, the /Pseudotriumphans clade of Cortinarius is shared between South America and Oceania (Australia-Tasmania and New Zealand), and therefore very likely to represent a Southern Hemisphere lineage of Cortinarius with wide Nothofagaceae host range [24] (p. 1049) and [23] (p. 1467). However, at the moment the available data only allow for speculation. A better knowledge of fungal diversity is needed to understand the evolutionary history of ectomycorrhizal fungi in Patagonian forests.
The ITS region is frequently used for fungal species identification [37,[63][64][65][66]. However, the ITS region has only minimal variation across Cortinarius and therefore probably underestimates the true diversity of this genus in natural ecosystems by up to 20% [67]. Barcoding is a powerful tool for ecological, environmental, or taxonomic research [68]. However, fungi occurring in the Southern Hemisphere are still largely under-represented in public databases [61], and data concerning their diversity and distribution are far from being complete, even when Cortinarius spp. are often quite dominant in studies of ectomycorrhizal fungi communities on roots and in soil [1,13,61,68,69]. This is especially true and important for fungal groups with immense species richness like Cortinarius, a widespread and important ectomycorrhizal genus from the South American Nothofagaceae forests with high ecological, forest restoration interests, and important non-timber forest products.

Conclusions
The diversity of Cortinarius from the Northern Hemisphere and from New-Zealand and Australia have been extensively studied (e.g., [24,38,68,[70][71][72]). However, aside from the pioneering work from Moser and Horak [5,26] there are relatively few studies that have focused on South American Cortinarius diversity [25]. Recent studies describing several new species or even new sections highlight this knowledge gap [8,12,15,16,67]. Based on our current knowledge, the Cortinarius species associated with Nothofagaceae species rarely or never occur outside the distribution range of their host trees [26,[71][72][73][74][75], making it highly likely that, after losing the connection via Antarctica, endemic Cortinarius lineages evolved in South America and Australasia. Cortinarius harbors a high diversity in these habitats, including many taxa that are waiting to be discovered [15,16]. In the future, systematic approaches will be important to fully sample Cortinarius from South America. These approaches should include DNA barcording of South American Cortinarius herbarium and systematic sampling of Nothofagaceae forests by South American mycologists.
Supplementary Materials: The following are available online at https://www.mdpi.com/article/10 .3390/life11050420/s1, Figure S1: Phylogenetic relationship (Bayesian consensus tree) of four new species of Cortinarius from South American Nothofagaceae forests, based on combined rDNA ITS-LSU and RPB1 sequences. Bayesian Posterior Probabilities are provided beside nodes. C. gracilentus was not included in the analysis. Sequences generated from the new species presented are highlighted in red.