Cryptodiscus muriformis and Schizoxylon gilenstamii, two new species of Stictidaceae (Ascomycota)

Cryptodicus muriformis and Schizoxylon gilenstamii (Stictidaceae, Ostropales, Lecanoromycetes, Ascomycota) are described as new to science from collections made in Europe (Sweden, Austria, Switzerland) and North America (Arizona), based on phylogenetic analyses of three loci. Both species show unique morphological characters that distinguish them from closely related species; Cryptodiscus muriformis, growing on decorticated or sometimes corticated branches of Picea abies and Pseudotsuga menziesii, that resembles C. foveolaris and C. tabularum in the yellowish-orange ascomatal disc, but is clearly distinguished by the muriform ascospores. Schizoxylon gilenstamii grows on dead stems of Globularia vulgaris and has a characteristic unique red-purple pigment in the epithecium and upper part of the excipulum that changes to bluish-green with KOH. We also report for the first time the anamorph of Stictis brunnescens.


Introduction
The family Stictidaceae (Ostropales, Lecanoromycetes, Ascomycota) contains predominantly saprotrophic discomycetes, but also lichens, parasites, and endophytes (e.g., Gilenstam 1969;Fernández-Brime et al. 2011;Baloch et al. 2013;Aptroot et al. 2014;Jahn et al. 2017). Sherwood (1977aSherwood ( , 1977b) produced a detailed and still very useful worldwide taxonomic and nomenclatural revision of the family with main focus on taxa from the USA. With the advent of DNA technologies, molecular phylogenetic studies have been crucial to revise established generic limits and species concepts in the family, although these studies are still in their infancy. Wedin et al. (2004Wedin et al. ( , 2005Wedin et al. ( , 2006 investigated taxa living on Populus and Salix in Northern Scandinavia. They found that several species of Stictis and Schizoxylon can live either as lichens or as saprotrophs depending on the substrate they inhabit, a phenomenon they called Boptional lichenization^. These results led to a widened concept of Stictis and Schizoxylon, which until then only included saprotrophic species. Wedin et al. (2006) pointed out that these two genera are potentially heterogeneous, but as molecular data was available only for a small number of species they could not pursue this further at that time. Finally, Cryptodiscus was studied in detail in Scandinavia by Baloch et al. (2009), resulting in a revision of the generic concept and the description of a new species.
Our knowledge of the Stictidaceae is obviously still very limited. The family contains small and generally droughttolerant fungi, which are easily overlooked and rarely collected. For this reason, in recent years we have continued exploring plant debris and, as a result, we found specimens of stictidaceous fungi that did not fit any described taxa. This paper introduces two new species collected in Sweden, Austria, Switzerland and USA, and provides detailed descriptions and a three-locus phylogeny that places them in the here adopted genera within the context of the Stictidaceae.
Anamorphs have been reported from several Stictidaceae species. Pycnidial anamorphs have been found in Acarosporina microspora, Conotrema urceolatum, Cryptodiscus gloeocapsa, Cyanodermella oleoligni, and Stictis radiata (e.g., Gilenstam 1969;Sherwood 1977a;Johnston 1983;Baloch et al. 2009;Nieuwenhuijzen et al. 2016). Johnston (1983) established the connection between anamorphs and teleomorphs of several Stictis species using cultures. In all cases, the anamorphs described were pycnidia. As a result from recent field work in Sweden, we found several conidiomata, which are presumably the anamorph of Stictis brunnescens. We have also used the three-locus phylogeny to test the connection between the anamorph and teleomorph.

Morphological and anatomical studies
Specimens were examined morphologically and anatomically and preserved in the herbaria GJO, S, UPS (herbarium codes according to Thiers [2017, continuously updated]) and H. O. Baral's private herbarium (H.B.). Hand-cut sections mounted in water were used to measure margin, hymenium, asci, paraphyses, and ascospores. Freezing microtome sections, 20-30 μm thick, mounted in water or stained with lactic blue or Congo red, were produced for a more detailed study of the anatomy. Lugol's iodine solution, before and after treatment with 10% potassium hydroxide (KOH), was used to detect amyloid structures. Thereby, Btype RR^means that the ascus wall reacts red at both low and high concentration before applying KOH. For the anatomical features measurements, the extreme values are shown after rejecting 10% of the highest and the lowest values; the highest and lowest absolute values are given in parentheses.

DNA extraction, PCR amplification, and sequencing
Total DNA was extracted from a selected number of samples with the DNeasy Plant Mini Kit (Qiagen, Germany) following the manufacturer's instructions or with the E.Z.N.A. Forensic DNA Kit following an improved protocol (Stielow et al. 2013).
The selected markers for this study were the internal transcribed spacer complete repeat (ITS), the first part (D1-D2 region, ca. 600 bp) of the large subunit (nLSU) of the nuclear ribosomal DNA, and the small subunit of the mitochondrial ribosomal DNA (mtSSU). Primer combinations for amplification of the three loci were: ITS1F (Gardes and Bruns 1993) and ITS2, ITS3, and ITS4 (White et al. 1990) for ≈ 0.6 kb ITS, LR0R and LR3 (Vilgalys and Hester 1990) for ≈ 0.6 kb nLSU, and mrSSU1 and mrSSU3R (Zoller et al. 1999) for the ≈ 0.8 kb mtSSU. Symmetric PCR amplifications were performed using IllustraTM Hot Start PCR beads, according to the manufacturer's instructions, with the following settings for the three loci: initial denaturation 94°C for 5 min, followed by five cycles (94°C for 45 s, 54°C for 30 s, and 72°C for 60 s), 30 cycles (94°C for 45 s, 52°C for 30 s, and 72°C for 60 s), and a final extension 72°C for 8 min. After examination by gel electrophoresis, amplification products were purified using ExoSAP-IT (USB Corp., U.S.A.). Sequencing of both strands was performed with the Big Dye Terminator technology kit v3.1 (ABI PRISM, U.S.A.) using the PCR primers.

Sequence alignment and phylogenetic analyses
Sequence fragments were edited and assembled with Sequencher 4.9 (Gene Codes Corp., USA) and consensus sequences subjected to BLAST searches to rule out amplification of contaminants. The newly produced sequences were included in individual alignments for each locus, together with sequences retrieved from GenBank representing morphologically similar species from the family Stictidaceae. The taxon Karstenia rhopaloides was selected as an outgroup as it is closely related to Stictidaceae, but Baloch et al. (2009) showed that it does not belong to the family. Sequences were aligned with the online version of MAFFT v7 (http://mafft. cbrc.jp/alignment/server/) using the algorithms L-INS-I for ITS and nLSU, and E-INS-I for mtSSU. The ambiguously aligned regions were identified using the G-blocks v0.91b (http://molevol.cmima.csic.es/castresana/Gblocks_server. html) and removed before analysis. Alignments were submitted to TreeBase (http://www.treebase.org; ID number 21812).
Phylogenetic analyses were performed for each locus individually with RAxML v 8.2.0 (Stamatakis 2014), using the rapid bootstrap analysis and search for best-scoring maximum likelihood algorithm (−f a), the GTRGAMMA model and 1000 bootstrap replicates (ML-BS). Then, the individual phylogenies were compared to assess congruence, considering there to be a significant conflict when two alternative relationships corresponding to different loci were both supported with bootstrap values ≥70% (Mason-Gamer and Kellog 1996). As no significant conflicts were detected, the three alignments were concatenated into a single dataset with five partitions (ITS1, 5.8S, ITS2, nLSU and mtSSU). Phylogenetic relationships and confidence were inferred on the combined dataset using maximum likelihood (ML) and Bayesian inference (BI). For the ML analysis, the same settings used for the individual gene analyses were used in RAxML, with the GTRGAMMA model applied to each partition. The Bayesian inference of the phylogeny was carried out by a Markov chain Monte Carlo with Metropolis coupling (MCMCMC) as implemented in MrBayes v3.2.6 (Ronquist et al. 2012). The substitution models estimated using the Bayesian Information Criterion (BIC) as currently implemented in jModeltest v.0.1.1 (Posada 2008) were: GTR + I + G for ITS1 and mtSSU, K80 + I for 5.8S and SYM + G for ITS2 and nLSU. Two parallel runs with four independent chains each were conducted for 20 million generations, with trees sampled at intervals of 500 generations and a random starting tree. A burn-in sample of the first 25% of the trees was discarded for each run, and the remaining trees were used to estimate branch lengths and posterior probabilities (BI-PP). All analyses were run in the CIPRES Science Gateway.

Phylogenetic analyses of combined sequence data
The combined dataset comprised 43 specimens, for which 41 ITS, 37 nLSU and 43 mtSSU sequences were available. The 21 newly generated sequences were submitted to GenBank and the voucher information and the accession numbers are available in Table 1. The data matrix contained 1381 aligned characters after the exclusion of 901 sites corresponding to ambiguously aligned regions, introns, and both ends of the alignments. Of the total number of analysed characters, 913 were constant and 468 were variable, of which 387 were parsimony-informative characters.
Both ML and BI analyses of the combined dataset resulted in very similar topologies; as the ML tree that resulted was slightly more resolved, we show here the best-scoring RAxML tree with support values indicated for both analyses (Fig. 1). Internodes were considered significantly supported when ML-BS ≥ 70% and/or BI-PP ≥ 0.95.
The monophyly of the putative new species Cryptodiscus muriformis and Schizoxylon gilenstamii received maximum support in both analyses (ML-BS: 100%, BI-PP: 1). Cryptodiscus muriformis appeared nested inside the Cryptodiscus clade with a highly supported (ML-BS: 99%, BI-PP: 1) sister relationship with C. tabularum. Schizoxylon gilenstamii appeared nested inside Schizoxylon and sister to the only specimen of S. berkeleyanum present in our phylogeny. The three samples belonging to a green conidiophore structure clustered unequivocally with the specimens of Stictis brunnescens (ML-BS: 100%, BI-PP: 1).
Etymology: the species name refers to the characteristic muriform ascospores.
Substrate: on decorticated or rarely corticated, dead, 8-20 mm thick branches of Picea abies and once on Pseudotsuga menziesii, on wood, rarely bark; found from 0.2 to 1.8 m above the ground, still attached to living or fallen trunks. Associated species: Mellitiosporium propolidoides, Orbilia alpigena, Propolis rhodoleuca, P. hillmanniana and different lichens, in North America Claussenomyces
MycoBank: MB 823298. Diagnosis: black erumpent apothecia 0.4-0.8 mm in diam., in section medullary excipulum green, which continues below the subhymenium, with a red-purple pigment in the upper part of the excipulum and the epithecium that changes to KOH+ bluish-green.
Etymology: the species is named after the Swedish lichenologist and mycologist Gunnar Gilenstam. Gunnar has had a long interest in the Stictidaceae and has contributed enormously to our understanding of the biology and diversity of these fungi in northern Scandinavia.
Known distribution: Sweden, all collections are from the island of Öland.

Discussion
We introduce the new species Cryptodiscus muriformis and Schizoxylon gilenstamii in the family Stictidaceae, based on the presence of differences in morphological characters and phylogenetic inference.
Cryptodiscus muriformis is nested within the highly supported clade representing Cryptodiscus sensu Baloch et al. (2009) (Fig. 1). Its phylogenetic placement is supported by the presence of characteristic traits of Cryptodiscus (Sherwood 1977a;Baloch et al. 2009): apothecia round to ellipsoid in outline, persistently immersed in the substrate, deeply concave disc; margin without crystals and periphysoids and without differentiation into layers; asci cylindrical to clavate, 8-spored, with amyloid ascus wall (I+ red, KOH/I+ blue) and distinct apical dome, paraphyses filiform, simple; ascospores hyaline. So far, all described species of Cryptodiscus produce narrowly cylindric-ellipsoid to ellipsoid or ellipsoid-fusoid ascospores with 1-7 transversal septa. However, C. muriformis has broadly ellipsoid, muriform ascospores. Most Cryptodiscus are saprotrophs except for three lichenicolous (Pino-Bodas et al. 2017) and three lichenized species that form a thin thallus (Baloch et al. 2009). Green patches corresponding to algae were observed on the wood substrate and around the fruiting bodies of C. muriformis (Fig. 2 a-b); similar algal patches have been observed in other Cryptodiscus species (see Fig. 3 c, Baloch et al. 2009). However, there is currently nothing that suggests that these algae are associated with the fungus as they appear scattered throughout the wood rather than forming the algal clumps observed in other loosely lichenized Stictidaceae (Wedin et al. 2004).
Ascus amyloidity was found to be variable in C. muriformis. The apical dome includes a large, distinctly amyloid ring (Fig. 2h) in the samples from Sweden, Steiermark (Bärentalalm), and Osttirol, but when reexamining the latter collection, the asci were inamyloid even regarding the entire ascus wall (Fig. 2i). The amyloid ring was also absent in the samples from Bern and Arizona. In the apothecia with inamyloid asci from Osttirol a faintly amyloid perihymenial medullary excipulum was observed. However, not every collection was thoroughly studied for amyloidity. Contrary to the European samples, the apothecia in the sample from Arizona grew on bark of a corticated branch of Pseudotsuga, but morphologically it well concurred with those from Europe.
In the resulting phylogeny, C. muriformis appears closely related to C. foveolaris and C. tabularum. All three species share a distinctly yellowish-orange disc, and C. tabularum, the sister taxon of C. muriformis, is also similar in having the hymenium Lugol + red-brown. However, C. foveolaris has one-septate ascospores and grows on wood of deciduous trees, and C. tabularum has 3-septate ascospores and it has been found so far only on decorticated wood of Pinus sylvestris.
Schizoxylon gilenstamii clearly forms a monophyletic group with the other Schizoxylon species included in the phylogeny (Fig. 1). This is supported by the combination of morphological characters that Sherwood (1977a) stated as diagnostic for the genus: cartilaginous apothecia, with a margin lacking periphysoids, long cylindrical asci with a thickened apex opened by a broad pore, and long filiform multiseptate ascospores disarticulating into part-spores. All Schizoxylon species are saprotrophs, except for S. albescens that can live either as a saprotroph or as loosely lichenized with algal clumps associated with its fruiting bodies (Wedin et al. 2006;Muggia et al. 2011). No algae were observed associated with S. gilenstamii.
The new Schizoxylon species is, to our knowledge, the first in the genus having red-purple pigments in the epithecium and the upper part of the excipulum changing K+ bluish-green. Based on Sherwood's monographs (1977a, b), which are the most complete studies of Schizoxylon up to date, other species having 8-spored asci less than 300 μm long and ascospores disarticulating in unicellular part-spores are S. ligustri, S. nigrellum (recorded from Norway), and S. sepincola. Schizoxylon ligustri has larger fruiting bodies that are not completely erumpent, but remain partly immersed in the substrate when mature, and are never globose but have flat grey disc and plane or reflexed pruinose white margin. Schizoxylon nigrellum and S. sepincola both have black erumpent apothecia of the same size as S. gilenstamii (0.5-0.8 mm in diam.). However, apart from lacking red-purple pigments in the epithecium and excipulum, in both species the margin in cross-section is brown (with colourless crystals in S. nigrellum), not brown-olivaceous as in S. gilenstamii, and have shorter asci and ascospores. All these species have been found on various types of woody substrate, but never on herbaceous remnants. Schizoxylon berkeleyanum is similar to S. gilenstamii because it also grows on herbaceous plant debris and may sometimes have non-septate partspores, but the apothecia in S. berkeleyanum are larger and have a thick margin covered with white to yellow pruina that reacts with KOH+ yellow, and the paraphysis tips are olivaceous brown and not red-purple as is characteristic for S. gilenstamii.
Schizoxylon gilenstamii is an ephemeral species that grows on last year's dead flower pedicles of Globularia vulgaris. Globularia is, in Sweden, considered a relict from a warmer period, and occurs here only in the alvar limestone pavement vegetation on the Baltic islands of Öland and Gotland. Schizoxylon gilenstamii should be searched for on Gotland and in the other parts of its host's distribution area (Spain and southern France).