Four new East Asian species of Aleurodiscus with echinulate basidiospores

Abstract Four new species of Aleurodiscus sensu lato with echinulate basidiospores are described from East Asia: A.alpinus, A.pinicola, A.senticosus, and A.sichuanensis. Aleurodiscusalpinus is from northwest Yunnan of China where it occurs on Rhododendron in montane habitats. Aleurodiscuspinicola occurs on Pinus in montane settings in Taiwan and northwest Yunnan. Aleurodiscussenticosus is from subtropical Taiwan, where it occurs on angiosperms. Aleurodiscussichuanensis is reported from southwest China on angiosperms in montane environments. Phylogenetic relationships of these four new species were inferred from analyses of a combined dataset consisting of three genetic markers, viz. 28S, nuc rDNA ITS1-5.8S-ITS2 (ITS), and a portion of the translation elongation factor 1-alpha gene, TEF1.

During a two-decade long, ongoing survey of corticioid fungi from mainland China and Taiwan, we have found four new species of Aleurodiscus with echinulate basidiospores based on morphological characters. In addition, phylogenetic analyses of a nuclear rDNA 28S D1-D2 domains (28S) dataset and analyses of a second dataset consisting of three genetic markers -nuc rDNA 28S D1-D2 domains (28S), nuc rDNA ITS1-5.8S-ITS2 (ITS), and translation elongation factor 1-alpha (TEF1) -are performed to complement our morphological observations and place the newly described species in a molecular phylogenetic framework.

Morphological and cultural studies
Macroscopic and microscopic studies were based on dried specimens. Color names from Rayner (1970) are capitalized. Thin free-hand sections of basidiocarps were prepared for microscopic study. For observations and measurements of microscopic characters, sections were mounted in 5% KOH to ensure rehydration. A blue-black color change with Melzer's reagent (IKI) indicates an amyloid reaction. Cotton blue (CB) was used as mounting medium to determine cyanophily. Sulphoaldehyde (SA) was used to detect a sulphuric reaction of gloeocystidia; a bluish black color change with SA indicates a positive reaction. The following abbreviations are used for basidiospore measurements: L = mean spore length with standard deviation, W = mean spore width with standard deviation, Q = variation in L/W ratio, and n = number of spores measured from each specimen. Apiculi and ornamentation were excluded in spore measurements. Living mycelia were isolated from the woody substratum beneath the basidiocarps, and were cultured on 1.5% malt extract agar (MEA). Fungal specimens and living cultures used in this study are deposited in the herbaria of the National Mu-seum of Natural Science of ROC (TNM; Taichung City, Taiwan) and Beijing Forestry University (BJFC; Beijing, China).
DNA extraction, polymerase chain reaction (PCR), and sequencing Dried specimens or the mycelial colonies cultured on MEA were used for DNA extraction, carried out with a Plant Genomic DNA Extraction Miniprep System (Viogene-Biotek Corp., New Taipei City, Taiwan). Liquid N and Tissue Lyser II (Qiagen, Hilden, Germany) were used to disrupt and homogenize the fungal tissues before DNA extraction process. The primer pairs ITS1/ITS4 or ITS1F/LR22 were used for the ITS region (White et al. 1990, Gardes andBruns 1993), and LR0R/LR3 and LR0R/LR5 were used for the 28S region (Vilgalys and Hester 1990). Efdf/1953R and 983F/2218R were used to amplify a portion of the TEF1 gene (Rehner & Buckley 2005;Matheny et al. 2007). PCR products were purified and directly sequenced by MB Mission Biotech Company (Taipei City, Taiwan). We examined the technical quality of the newly obtained sequences by comparison to entries in GenBank. Sequences were assembled using BioEdit v7.2.5 (Hall 1999). Newly obtained sequences (Supplementary Table 1) were submitted to either GenBank through the National Center for Biotechnology Information (NCBI) or DNA Data Bank of Japan (DDBJ) (Mashima et al. 2016, Benson et al. 2018.

Alignment and phylogenetic analyses
The newly generated sequences were added to the DNA sequence dataset employed by Dai and He (2016), so far the most inclusive alignments for analyzing Aleurodiscus s.l. based on three genetic markers. To achieve a comprehensive analysis, we also added some related taxa of the genera Boidinia Stalpers & Hjortstam, Conferticium Hallenb., Gloeocystidiellum Donk and Megalocystidium Jülich to the ingroup. We tried to include the type species of the genera as far as possible (Table 1). The phylogenetic tree of the 28S+ITS+ TEF1 dataset was inferred through Maximum likelihood (ML) and Bayesian inference (BI) methods using RAxML v. 8.2.4 (Stamatakis 2014) and MrBayes v. 3.2.6 (Ronquist et al. 2012), respectively (Ronquist et al. 2012, Stamatakis 2014. The alignments were inferred in MAFFT v. 7 using the FFT-N-i strategy for 28S and TEF1, and Q-INS-i strategy for ITS. For the BI analysis, the best-fit model for each alignment partition was estimated by jModelTest 2 (Darriba et al. 2012) using the Akaike information criterion (AIC). For ML bootstrapping, the extended majority-rule consensus tree criterion was specified under a GTRGAMMA model with 1000 replicates. In the BI analysis, four MCMC chains were run simultaneously from a random starting tree for ten million generations. Trees were sampled every 1000 generations resulting in 10000 trees in the posterior distribution; the first 25% trees were discarded as the burn-in. Posterior probabilities (PP) were calculated based on the post-burn-in trees. ML bootstrap values (BS) and BI posterior probability (PP) values ≥ 50% and ≥ 0.7 are indicated at the nodes of the ML tree. The final sequence alignments and the phylogenetic trees are available at TreeBASE (S23581; www.treebase.org).

Discussion
A number of phylogenetic studies of Aleurodiscus s.l. have been conducted in the past twenty years (Wu et al. 2001;Larsson and Larsson 2003;Miller et al. 2006;Larsson 2007;Dai and He 2016;Dai et al. 2017). Miller et al. (2006) and Larsson (2007) tried to establish a family level classification for Aleurodiscus s.l., as well as related taxa of the Russulales. However, a fully resolved and robust phylogeny of Aleurodiscus s.l. and related taxa was not achievable with ribosomal genes alone. Dai and He (2016) and our study have addressed this by including TEF1 for phylogenetic analyses. From our phylogenetic analyses of three DNA genetic markers (Fig. 1) we can conclude the following about evolutionary relationships in the Stereaceae: (i) Aleurodiscus s.l. is highly polyphyletic; (ii) Acanthophysellum is polyphyletic; (iii) Gloeocystidiellum is polyphyletic; (iv) Megalocystidium is polyphyletic; and (v) Conferticium is paraphyletic.
Aleurodiscus alpinus is reminiscent of Aleurodiscus s.s. (A. amorphus (Pers.) J. Schröt. and A. grantii Lloyd) due to the discoid basidiocarp and echinulate basidiospores, as well as the absence of acanthophyses. However, the gloeocystidia of Aleurodiscus s.s. are paraphysis-like, narrow and moniliform, while those of A. alpinus are much wider and not moniliform. In addition, A. alpinus has unbranched or branched hyphidia, which are lacking in Aleurodiscus s.s. Aleurodiscus alpinus formed a clade with A. sichuanensis (Fig. 1), however, the latter has simple-septate hyphae and acanthophyses. Aleurodiscus alpinus and A. cupulatus share most morphological features, except the latter has much wider basidiospores. Aleurodiscus alpinus grows on Rhododendron sp. in Yunnan of China, while A. cupulatus occurs on Pseudotsuga menziesii in Idaho of USA. No DNA sequence of the latter has been obtained to examine their relationship.
Aleurodiscus pinicola presents protuberances in the basidia and this is reminiscent of Acanthobasidium. However, this feature is not limited to Acanthobasidium spp. For example, basidia of Aleurodiscus mirabilis (Berk. & M.A. Curtis) Höhn. and A. wakefieldiae Boidin & Beller occasionally possess protuberances, but they and A. pinicola do not belong to Acanthobasidium (Fig. 1).
Aleurodiscus senticosus is macroscopically distinct in having more or less cracked hymenophore from the fusion of smaller basidiocarp patches; microscopically, its basidia bear a large, spiny, bladder-like structure that is unique among Aleurodiscus s.l. The present phylogenetic analyses (Fig. 1) indicated that A. senticosus formed a clade with Xylobolus and Acanthofungus, but without strong support. However, these two genera differ from A. senticosus by causing a white-pocket rot in wood and by bearing smooth basidiospores.
In conclusion, the status of each segregate genus of Aleurodiscus s.l. should be further examined by multi-gene analysis of more species to evaluate which ones can be recognized and which cannot. Although the four new species we introduce cannot be accommodated in any segregate genus of Aleurodiscus s.l. according to the present combined morphological and phylogenetic studies, they are still placed under the broad sense of Aleurodiscus at the present time.