Phylogeny, Global Biogeography and Pleomorphism of Zanclospora

Zanclospora (Chaetosphaeriaceae) is a neglected, phialidic dematiaceous hyphomycete with striking phenotypic heterogeneity among its species. Little is known about its global biogeography due to its extreme scarcity and lack of records verified by molecular data. Phylogenetic analyses of six nuclear loci, supported by phenotypic data, revealed Zanclospora as highly polyphyletic, with species distributed among three distantly related lineages in Sordariomycetes. Zanclospora is a pleomorphic genus with multiple anamorphic stages, of which phaeostalagmus-like and stanjehughesia-like are newly discovered. The associated teleomorphs were previously classified in Chaetosphaeria. The generic concept is emended, and 17 species are accepted, 12 of which have been verified with DNA sequence data. Zanclospora thrives on decaying plant matter, but it also occurs in soil or as root endophytes. Its global diversity is inferred from metabarcoding data and published records based on field observations. Phylogenies of the environmental ITS1 and ITS2 sequences derived from soil, dead wood and root samples revealed seven and 15 phylotypes. The field records verified by DNA data indicate two main diversity centres in Australasia and Caribbean/Central America. In addition, environmental ITS data have shown that Southeast Asia represents a third hotspot of Zanclospora diversity. Our data confirm that Zanclospora is a rare genus.


Introduction
Zanclospora [1], typified with Z. novae-zelandiae, was established for dematiaceous hyphomycetes observed on plant litter or decaying wood and bark and characterised by setiform conidiophores, discrete phialides arranged in whorls and hyaline, unicellular, non-setulate conidia in slimy masses enveloping the conidiophores [2-10]. However, the morphological characters of conidiophores, phialides and conidia vary among species and contribute to the phenotypic heterogeneity of the genus. Conidiophores are simple or branched, occasionally accompanied by setae, branches are fertile, resembling the main stalk with secondary and tertiary branches often developed, or they are sterile and setiform, inserted into the main stalk. Conidiogenous cells are either tightly appressed to the conidiophore in multiple whorls forming a compact fertile zone or divergent in several loose whorls. Phialides possess indistinct or well-defined, flared to tubular collarettes. The conidial shape varies from falcate, obovoid to bacilliform. The teleomorph-anamorph connection has been established for only two species, both with a teleomorph attributed to Chaetosphaeria, namely the Z. brevispora anamorph of Ch. brevispora [11] and Zanclospora sp. anamorph of Ch. lateriphiala [12].
To date, ten species and two varieties were introduced in Zanclospora [1-10], but little is known about the systematic placement, relationships and global geographical distribution of these taxa. Moreover, the genus is under-represented in culture collections. Using molecular data, Fernández et al. [13] and Hernández-Restrepo et al. [9] confirmed the placement of Z. iberica and Ch. lateriphiala in the Chaetosphaeriaceae. Zanclospora is similar to Cryptophiale [14,15], Cryptophialoidea [16], and Kionochaeta [17] in having pigmented conidiophores with a setiform extension, lateral phialides usually arranged in fertile zones and hyaline conidia. Our observations indicate that conidial structures similar to two dematiaceous hyphomycete genera, Stanjehughesia [18] and Phaeostalagmus [19], can occur on a natural substrate and in culture.
Although there are only a few published records of Zanclospora, Z. novae-zelandiae seems to be an exception to the rule. In the protologue of Z. novae-zelandiae [1], conidia were reported with a relatively large range of lengths, 18-35 µm long. On the other hand, the specimens listed by these authors [1] shared the same morphological diagnostic characters such as colourless disk-like excrescences on the upper setiform part of conidiophores and branches, branched conidiophores and falcate conidia. However, the large range of conidial lengths provided an opportunity to expand the species concept further, and the published morphological profiles of Z. novae-zelandiae vary considerably. Several authors [8,[20][21][22][23] reported Z. novae-zelandiae from different geographical regions with a variable conidial length and introduced other features. The conidiophore wall lacked ornamentation, and some collections contained uniformly unbranched conidiophores. These differences may indicate cryptic species within Z. novae-zelandiae or high intraspecific variability. Unfortunately, DNA sequence data of Z. novae-zelandiae are not available to provide answers to these hypotheses.
Our knowledge about the biogeography of Zanclospora is fragmentary due to the lack of records verified by molecular data and its extreme scarcity. Published data of members of Zanclospora suggest a worldwide geographical distribution. Species were recorded from the tropics of Brazil, Brunei, Cuba, India, Ecuador, Ivory Coast, Kenya, Nigeria, Seychelles, Taiwan and Vietnam, but also from the temperate and subtropical climate zones of the Southern and Northern Hemispheres in Japan, New Zealand, South Africa, Spain and the USA [1][2][3][4][5][6][7][8][9][10][11][12][20][21][22][23][24][25][26][27][28]. Almost all published records were from decaying bark and wood, less often from fallen leaves, and were obtained by direct observation on natural substrates. Thus, it is unknown if these fungi also occur in related substrates such as soil or healthy plant tissues (as endophytes), where they remain overlooked due to their slow growth in culture and rarity.
To date, research of geographic distribution patterns of fungi has relied heavily on public nucleotide sequence databases such as NCBI GenBank [29] and UNITE [30], which enable blasting (BLASTn search) [31] against fungal barcodes generated by Sanger technology. Data mining of DNA barcodes is especially helpful for biogeography and diversity studies of abundant and globally distributed taxa e.g., [32]. However, most of the barcode sequences generated so far come from massively parallel sequencing technologies, whose data have been stored in various public repositories, not allowing for easy data mining in multiple studies. As a result, any biogeographic evaluation is laborious and limited to a small number of source datasets [33]. This gap has recently been filled by the creation of the GlobalFungi database of fungal ITS data [34,35] collected from terrestrial biomes of soil, dead or live plant material. Such a tool is particularly useful for studying members of the Chaetosphaeriaceae, which are usually less abundant and inhabit substrates covered by GlobalFungi.
This study aims to assess the systematic placement of Zanclospora and investigate intraspecific and interspecific variability of its members by using comparative morphology on natural substrates and in culture along with phylogenetic analyses. Other objectives include the description and experimental verification of teleomorph-anamorph connections and anamorphic phenotypes, and the determination of geographical distribution and ecology of species of Zanclospora.

Fungal Strains
During our study, we gathered several Zanclospora inhabiting decaying plant material in various localities from the south temperate climate zone of New Zealand, and north temperate climate zone of Europe in Portugal and Spain, and North America in the USA. Other specimens were obtained from the Fungarium of the Illinois Natural History Survey (ILLS, Champaign, IL, USA) and New Zealand Fungarium (PDD, Auckland, New Zealand). Holotypes and specimens collected in this study were deposited at CBS, ILLS and PDD (as dried voucher specimens or dried cultures).
Axenic cultures were derived from freshly collected material (see Section 2.2). Additional cultures were obtained from BCCM/MUCL Agro-food and Environmental Fungal Collection (MUCL, Université Catholique de Louvain, Louvain, Belgium), the International Collection of Microorganisms from Plants (ICMP, Auckland, New Zealand) and Westerdijk Fungal Biodiversity Institute (CBS, Utrecht, The Netherlands). Representative strains and ex-type strains isolated from our collections were deposited at CBS and ICMP.
Isolates, their sources and GenBank accession numbers of sequences generated in this study are listed in Table 1. Fungal novelties were registered in MycoBank.

Gene Markers, Sequence Alignments and Phylogenetic Analyses of Fungal Strains
Sequences of six gene markers: ITS1-5.8S-ITS2 (ITS) of the nuclear rRNA cistron, the small subunit 18S ribosomal DNA gene (18S) and the large subunit 28S ribosomal DNA gene (28S) (approximately 1800 base pairs at the 5 -end), domains 5-7 of the second largest subunit of RNA polymerase II (rpb2), the intermediate section of the coding region of the translation elongation factor 1-alpha (tef1-α) and coding and non-coding regions of betatubulin (tub2) marked by exons 2−6, were analysed to assess evolutionary relationships of Zanclospora and similar fungi. GenBank accession numbers for sequences retrieved from GenBank and published in other studies [9,39, are listed in Table S1.
Sequences were aligned manually in Bioedit v.7.1.8 [76] and introns were excluded from the alignments. The sequences were combined into four datasets that were partitioned into ITS, 18S, 28S, rpb2, tef1-α and coding and non-coding regions of tub2 subsets of nucleotide sites for which we assumed rate heterogeneity. Single-locus data sets were evaluated using PartitionFinder2 [77], implemented in the CIPRES Science Gateway v.3.3 [78,79], to find the best partitioning scheme for our datasets and to select best-fit models under corrected Akaike information criteria. Conflict-free data sets were concatenated, and four alignments (deposited in TreeBASE) were subjected to subsequent phylogenetic analyses.
Since there are no previous phylogenetic studies on Zanclospora and the standard use of particular nuclear loci vary among fungal groups, we conducted four phylogenetic analyses to assess relationships of the genus, based on the preliminary results of the BLASTn search. The phylogenetic analysis of Zanclospora and members of the Sordariomycetes were based on 18S, 28S and rpb2 markers. The relationships within the Chaetosphaeriaceae were assessed with the ITS and 28S sequences. The intraspecific relationships of Zanclospora were evaluated with ITS, 28S, tef1-α and tub2 genes, and the phylogenetic analysis of two Zanclospora strains with affinity to the Xylariales were assessed in the analysis of the combined ITS, 28S, tef1-α and rpb2 sequences.
Phylogenetic reconstructions were performed using Bayesian Inference (BI) and Maximum Likelihood (ML) analyses through the CIPRES Science Gateway v.3.3. ML analyses were conducted with RAXML-HPC v.8.2.12 [80] with a GTRCAT approximation. Nodal support was determined by non-parametric bootstrapping (BS) with 1000 replicates. BI analyses were performed in a likelihood framework as implemented in MrBayes v.3.2.6 [81]. Two Bayesian searches were performed using default parameters. The B-MCMCMC analyses lasted until the average standard deviation of split frequencies was below 0.01 with trees saved every 1000 generations. The first 25% of saved trees, representing the burn-in phase of the analysis, were discarded. The remaining trees were used for calculating posterior probabilities (PP) of recovered branches. The BI and ML phylogenetic trees were compared visually for a topological conflict among supported clades.
Histograms of intraspecific and interspecific distances of Zanclospora s. str. were created for each of the four markers (ITS, 28S, tef1-α, and tub2) used in the phylogenetic analyses in order to illustrate the amount of overlap for each gene. Matrix of pairwise distances was computed with MEGAX [82] using the Kimura two parameters (K2P) model, and the histogram was plotted in GraphPrism 7.03 software (Graphpad Software Inc., LaJolla, CA, USA) using a bin size of 0.001.

Phylogeny of Environmental Sequences and Biogeography
Initially, the interspecies genetic distance (p-dist) was calculated for ITS1 and ITS2 datasets of 12 Zanclospora species using MEGAX [82] to obtain sequence similarity thresholds for species delimitation in Zanclospora. The obtained value and the limit of full coverage were used for the search in the GlobalFungi v.0.9.6 (release version 1.0) database containing data from 20,000 samples originating from 207 studies [34]. For each taxon, data about occurrence across environmental samples and metadata related to the particular samples (location, substrate, biome, climatic data, pH) were obtained (Table S2). Taxa related to Zanclospora were used for comparison, e.g., Chaetosphaeria minuta, Cryptophiale, Cryptophialoidea and Kionochaeta (Table S3).
In order to study Zanclospora diversity hidden among environmental sequences, the full-length ITS1 and ITS2 sequences of 12 Zanclospora species were blasted against the GlobalFungi database. The sequences with a similarity of 89-100% and full-length coverage were downloaded. The Zanclospora genus boundaries were inferred from ML trees of ITS1 and ITS2 sequences computed in Phyml v.3.1 [83] using the GTR model and 500 bootstrap replicates. The same procedure, i.e., blasting, downloading of related sequences and phylogenetic analyses, was performed against sequences deposited in NCBI GenBank and UNITE database. Virtual taxa, consisting of environmental sequences only, were defined as arbitrary phylotypes in the ML phylogenetic trees. Metabarcoding data can contain pseudogenous copies, which may lead to an overestimation of diversity. Thus, GC-content and ITS2 secondary structure stability of obtained sequences were compared as recommended in [84].

Phylogenetic Analyses
In order to examine the evolutionary relationships of Zanclospora within the Sordariomycetidae, phylogenetic analysis was based on the combined 18S, 28S and rpb2 sequences of 108 representatives of the Sordariomycetes. Adelosphaeria catenata, Melanotrigonum ovale and Pleurotheciella erumpens (Pleurotheciales, Hypocreomycetidae) served as the outgroup. One hundred-twenty nucleotides (nt) at the 5 -end of 18S, 85 nt at the 5 -end and 483 nt at the 3 -end of 28S were excluded from the alignment because of the incompleteness in the majority of sequences. The full dataset consisted of 4139 characters including gaps (18S = 1634 characters, 28S = 1342, rpb2 = 1163) and 2213 unique character sites (RAxML). For the BI analysis, the GTR+I+G model was selected for all partitions. The BI and ML trees were not in conflict; the ML tree is shown in Figure 1. The subclass Sordariomycetidae included 30 well-supported clades (≥75% ML BS/≥1.0 PP) representing orders and families and one incertae sedis lineage. This subclass was resolved with four major subclades. The first subclade (96/1.0) included ten orders and families with mostly phialidic and tretic conidiogenesis, rarely holoblastic, namely Boliniales, Cephalothecales, Chaetosphaeriales, Coniochaetales, Cordanales, Helminthosphaeriaceae, Leptosphaerellaceae, Phyllachorales, Pseudodactylariales, Sordariales and Tracyllales. The second subclade included Vermiculariopsiellales (100/1.0) and an unsupported clade with Mirannulata samuelsii and Teracosphaeria petroica (both incertae sedis). Species with prevalent holoblastic conidiogenesis, if known, attributed to 13 orders and families, formed a strongly supported subclade (98/1.0), which was inferred as sister to the fourth subclade (99/1.0) containing Calosphaeriales, Diaporthales, Jobellisiales and Togniniales with phialidic conidiogenesis. The Xylariomycetidae were resolved as a strongly supported clade (100/1.0) encompassing five representatives of the Xylariales. Zanclospora novae-zelandiae clustered in the Chaetosphaeriales (100/1.0), while Z. stellata nested in the Vermiculariopsiellales and was transferred to a new genus Stephanophorella. Zanclospora urewerae was inferred as a member of the Xylariales (100/1.0) and accommodated in the new genus Brachiampulla. A non-type strain of Selenosporella curvispora CBS 102623, the generic type, clustered in the Helminthosphaeriaceae.
Relationships of Zanclospora with four Chaetosphaeria, so far known to produce only a phaeostalagmus-like anamorph or their anamorph is unknown [12,85,86], and other members of the Chaetosphaeriaceae were assessed in the phylogenetic analysis of a data set that included ITS and 28S sequences of 89 representative species of the family. Leptosporella arengae and L. bambusae (Leptosphaerellaceae), and Tracylla eucalypti and T. aristata (Tracyllaceae) served as an outgroup. Seventy-one nt at the 5 -end and 663 nt at the 3 -end of 28S were excluded from the alignment.
The alignment had 1722 characters including gaps (ITS = 605 characters, 28S = 1117) and 861 unique character sites (RAxML). For the BI analysis, the GTR+I+G model was selected for both partitions. No conflicts occurred between BI and ML trees; the ML tree is shown in Figure 2. The Chaetosphaeriaceae included 47 lineages representing genera or natural groups of species. Zanclospora was resolved as a strongly supported monophyletic clade (99/1.0). Four Chaetosphaeria, namely Ch. jonesii, Ch. phaeostalacta, Ch. sylvatica and Ch. tropicalis, clustered in the Zanclospora clade. Kionochaeta was shown as paraphyletic, a non-type strain of K. ramifera MUCL 39164 [87], the type species of the genus, clustered with two other Kionochaeta as a monophyletic lineage (100/1.0), while K. ivoriensis nested on a separate branch close to Cryptophiale and Cryptophialoidea. Phaeostalagmus (as P. cyclosporus CBS 663.70, the generic type) and Stanjehughesia (as S. hormiscioides CBS 102664), two hyphomycete genera whose similar phenotypes appear in the life cycle of Zanclospora, were resolved as separate lineages.
In order to evaluate relationships among 16 strains of Zanclospora and five strains of Chaetosphaeria, some of which form a phaeostalagmus-like anamorph in culture, we analysed a data set of the combined ITS, 28S, tef1-α and tub2 sequences. Three Dictyochaeta were used as an outgroup to root the tree. The alignment had 4770 characters including gaps (ITS = 487, 28S = 1842, tef1-α = 992, tub2 = 1449) and 729 unique character sites (RAxML). For the BI analysis, the GTR+G model was selected for ITS and tef1-α, GTR+I+G for 28S and tub2 coding region, and GTR+I for the tub2 non-coding partition. The ML tree is shown in Figure 3. Zanclospora was resolved with two subclades containing 12 species. Molecular data confirmed a close relationship among species with the typical Zanclospora conidiophores and those exhibiting phaeostalagmus-and stanjehughesia-like morphotypes.
The first subclade (-/1.0) comprised nine species including Z. novae-zelandiae and five new species, namely Z. aurea, Z. clavulata, Z. falcata, Z. ramifera and Z. xylophila, described below. The second subclade (100/1.0) contained three species formerly attributed to Chaetosphaeria. The differences between BI and ML trees were in the position of several species. In the BI tree, Z. aurea was shown on a separate branch, and Z. phaeostalacta and Z. xylophila were resolved as sister species.

Figure 2.
Combined phylogeny using ITS and 28S of members of the Chaetosphaeriaceae. Species names given in bold are taxonomic novelties; T, E, I and P indicate ex-type, ex-epitype, ex-isotype and ex-paratype strains. Thickened branches indicate branch support with ML BS = 100%, PP values = 1.0. Branch support of nodes ≥75% ML BS and ≥0.95 PP is indicated above or below branches. Abbreviation: p.p. after a genus name (pro parte).

Figure 2.
Combined phylogeny using ITS and 28S of members of the Chaetosphaeriaceae. Species names given in bold are taxonomic novelties; T, E, I and P indicate ex-type, ex-epitype, ex-isotype and ex-paratype strains. Thickened branches indicate branch support with ML BS = 100%, PP values = 1.0. Branch support of nodes ≥75% ML BS and ≥0.95 PP is indicated above or below branches. Abbreviation: p.p. after a genus name (pro parte).
Relationships of Z. urewerae were assessed in the phylogenetic analysis of the combined ITS, 28S, tef1-α and rpb2 sequences of 81 representatives of the Xylariales. Bactrodesmium abruptum and B. diversum (Savoryellaceae), Helicoascotaiwania lacustris and Pleurotheciella erumpens (Pleurotheciaceae) were used to root the tree. Eighty-five nt at the 5 -end and 925 nt at the 3 -end of 28S were excluded from the alignment. The alignment had 3964 characters including gaps (ITS = 764 characters, 28S = 857, tef1-α = 1148, rpb2 = 1195) and 2307 unique character sites (RAxML). For the BI analysis, the SYM+I+G model was selected for ITS, while the GTR+I+G model was selected for 28S, tef1-α and rpb2 partitions. There were no conflicts between ML and BI trees; the ML tree is shown in Figure 4. The Xylariales included 33 well-supported clades representing families and one incertae sedis lineage. Zanclospora urewerae was clustered as a sister to Xyladictyochaeta of the Xyladictyochaetaceae (99/1.0). Morphologically similar genera and species such as Selenodriella (Microdochiaceae) and Ceratocladium polysetosum (incertae sedis) formed separate lineages.

Barcode Analysis
Although we lacked enough representatives for species to be able to fully explore the barcoding gap of Zanclospora s. str., we used the current four-gene dataset to examine pairwise genetic distances, visualize them and evaluate the species-delimiting ability of each marker (Table S4). The barcoding gap separating intraspecific and interspecific variability of Zanclospora was present in all studied markers, with the biggest gap found in tef1-α, followed by tub2, ITS and 28S. In ITS, the minimal interspecific divergence occurred among species of the Z. novae-zelandiae species complex (0.64-1.3%), the next minimum and maximum distances between other species ranged between values 1.3-16.18%. In tef1-α, the minimum genetic distance occurred between the sibling species, Z. falcata and Z. novae-zelandiae (0.52%), but ranged from 1.05 to 5.24% in other species. The situation in tub2 gene was complicated by various lengths of the sequenced fragments. Nonetheless, the performance of the gene is comparable to tef1-α and delimits closely related species; the minimum and maximum interspecific distances ranged between values 1.45-15.14%. The genes tef1-α, tub2 and ITS are recommended as barcodes for Zanclospora.

Analysis of Zanclospora Diversity in Environmental Samples, Biogeography and Ecology
For the ITS1, the lowest interspecies distance was 0.012 (Z. clavulata vs Z. novaezelandiae). For the ITS2, the lowest distance ranged from 0 (Z. novae-zelandiae vs Z. xylophila) followed by Z. novae-zelandiae vs Z. iberica and Z. xylophila vs Z. iberica, both 0.012. The obtained values showed that for ITS1 and ITS2 there is no generally valid sequence similarity threshold for species delimitation in Zanclospora. However, 99-100% sequence similarity was applicable for most of the species and was used for the search in GlobalFungi. The only exceptions were Z. novae-zelandiae and Z. xylophila, where the criterion of full sequence similarity in ITS2 was used.
The BLAST search resulted in 559 unique ITS1 sequences (similarity 89-100% to Zanclospora queries, see Section 2.5). The dereplicated dataset had 33 sequences, 185 characters, from which 82 were variable and 16 singletons. The ML tree (Phyml) was rooted in a branch leading to the Dictyochaeta clade and is shown in Figure 5. The environmental sequences were clustered into seven phylotypes. Among them, one can be linked with Z. jonesii. For the ITS2, 108 unique sequences were found, which resulted in 79 sequences attributable to Zanclospora. The dereplicated dataset had 48 sequences, 166 characters, from which 87 were variable and 20 singletons. The ML tree (Phyml) was rooted in a branch leading to the Dictyochaeta clade and is shown in Figure 6. The environmental sequences clustered into 15 phylotypes. One was linked with Z. clavulata, while the other three contained sequences of the whole ITS region, and thus may be linked with phylotypes defined by the ITS1 marker. The same procedure was applied to data from the NCBI GenBank and UNITE databases and resulted in the single sequence (Ascomycete sp., DQ124120, unpublished), which was linked with phylotype ITS1-ENV5 ( Figure 5). The sequence similarity in the ITS2 region has little differentiation power in the Z. novae-zelandiae clade. Interestingly, Z. novae-zelandiae and Z. xylophila share identical ITS2, while they are distinct in ITS1 and other studied genes. The best hit of Z. novae-zelandiae/Z. xylophila was 99.32% and is considered as ITS2-ENV2 phylotype ( Figure 6, Table S3). Habitat of the environmental sequences is indicated by brown, grey and green boxes corresponding to dead wood, soil or roots.  Habitat of the environmental sequences is indicated by brown, grey and green boxes corresponding to dead wood, soil or roots. Asterisk (*) indicates phylotypes that can be linked with phylotypes defined by the ITS1 marker. Habitat of the environmental sequences is indicated by brown, grey and green boxes corresponding to dead wood, soil or roots. Asterisk (*) indicates phylotypes that can be linked with phylotypes defined by the ITS1 marker.
Concerning the diversity presented in environmental sequences, only Z. clavulata and Z. jonesii were traced in the GlobalFungi database at the defined similarity level. Another six and 11 phylotypes, respectively, roughly corresponding to the level of species, were identified in ITS1 and ITS2. Biogeography and ecology of newly identified phylotypes inferred from the GlobalFungi database (Tables S2, S3 and S5) and known species (Table S6)   . Geographical distribution and substrate affinity of Zanclospora species with known ITS sequence data. The map summarizes data from the GlobalFungi database (shown by circles) and field collections verified by sequencing (species [1][2][3][4][5][6][7][8][9][10][11][12] or based only on published data (species 13-20) (shown by squares). See Tables S2 and S6 for primary data. Each symbol (circle or square) represents a unique sample. The substrates are differentiated by colours. Note that environmental taxa defined by ITS1 sequences can overlap with those from the ITS2 dataset.

Taxonomy
Habitat and geographical distribution: Members of the genus are saprobes on decaying plant material with a worldwide distribution in Northern and Southern temperate, subtropical and tropical climate zones ( Figure 7). Although most of the field observations include specimens on decaying wood or fallen leaves, environmental ITS1 and ITS2 sequences attributable to Zanclospora originated also from soil and roots (Tables S2 and S5).
Moreover, environmental data suggested several new, likely undescribed species lineages from Southeast Asia, Australasia and South America.
Notes: Our Zanclospora strains derived from conidia and ascospore isolates exhibit an undescribed morphological variability in anamorphic characteristics that are associated with three anamorphic stages. Sterile or rarely fertile conidiophores resembling euseptate, cylindrical conidia of another hyphomycete Stanjehughesia [18] were often associated with ascomata and the typical Zanclospora conidiophores. On natural substrates and in culture, Zanclospora and stanjehughesia-like conidiophores occur irregularly and independently of each other. The Zanclospora conidiophores that arise on agar are usually less complex and significantly reduced in size becoming remarkably similar to Phaeostalagmus [19], or they are reduced to single conidiogenous cells. These reduced forms produce microconidia exclusively, compared to the complex Zanclospora conidiophores on natural substrates or sterile stems of U. dioica in vitro producing macroconidia. There is very little difference between the morphology of a reduced Zanclospora conidiophore and what can be called the phaeostalagmus-like morphotype. In the latter, the phialides are lateral, sessile, arranged in verticilli and also disposed terminally, sometimes on short branches, resembling P. cyclosporus, the type species of the genus. The phaeostalagmus-like morphotype occurs primarily in species whose axenic culture was derived from ascospores and the Zanclospora anamorph is unknown, i.e., Z. phaeostalacta Seventeen species and two varieties are accepted in Zanclospora, 12 of which have been verified using molecular data and are presented below. Their colonies on the four growth media are compared in Figures 8 and 9. New teleomorph-anamorph connections have been experimentally confirmed for Z. aurea, Z. falcata, Z. novae-zelandiae, Z. ramifera and Z. xylophila. Five other species, whose molecular data are unavailable, are provisionally accepted in Zanclospora based on morphological similarities, namely Z. austroamericana [3], Z. bicolorata [10], Z. bonfinensis [8], Z. brevispora [1] and Z. mystica [4]. Zanclospora indica [2], Z. stellata [6] and Z. urewerae [7] are excluded from the genus.
Zanclospora bonfinensis, Z. bicolorata and Z. mystica deviate from other Zanclospora in conidiophores that are darker and opaque in the upper setiform part. Moreover, the two former species possess tubular to narrowly wedge-shaped collarettes and bacilliform to suballantoid conidia that are unique in Zanclospora and better correspond to Z. stellata, transferred to the new genus Stephanophorella in this study. The conidiophores of other Zanclospora are paler towards the apex, the apex and the rest of the conidiophore have more or less the same colour, conidia are falcate to almost horseshoe-shaped, obovoid, clavate and collarettes are inconspicuous to short-flared.
Preparation of the identification key has proven challenging, mainly due to inconsistencies in the occurrence of teleomorphs, anamorphs and synanamorphs on natural material and in culture. Therefore, a synopsis table with diagnostic features of accepted species of Zanclospora is compiled to show the interspecific variability (Table 2).     Brazil unknown n/a n/a n/a n/a Z. bicolorata decaying leaf Ecuador unknown n/a n/a n/a n/a Z. bonfinensis decaying leaves Brazil unknown n/a n/a n/a n/a wood South Africa unknown n/a n/a n/a n/a Z. clavulata plant debris Portugal unknown n/a n/a n/a n/a Z. falcata decaying wood New Zealand present

3-septate fusiform
Z. iberica plant debris Spain unknown n/a n/a n/a n/a  Z. jonesii not observed n/a n/a n/a n/a n/a not observed n/a n/a n/a n/a n/a Z. ramifera not observed present, similar to main stalk smooth (30-)55-235 × 2-3.5, 3.5-6 at FZ n/a n/a Z. sylvatica not observed n/a n/a n/a n/a n/a Z. tropicalis not observed n/a n/a n/a n/a n/a Z. xylophila not observed n/a smooth n/a n/a n/a icana not observed n/a n/a n/a n/a [3] Z. bicolorata not observed n/a n/a n/a n/a [10] Z. bonfinensis not observed n/a n/a n/a n/a [8] Z. brevispora var. brevispora not observed n/a n/a n/a n/a [1] Z. brevispora var. transvaalensis not observed n/a n/a n/a n/a [5] Z. clavulata not observed Z. jonesii present, described as setae n/a n/a n/a n/a [86] Z. lateriphiala not observed not observed present not observed not observed [12] Z. mystica not observed n/a n/a n/a n/a [4] Z. novaezelandiae 364-675 × 4.  Figure 10).  Etymology: Aureus (L) golden, from aurum (gold), referring to the colour of goldenyellow colonies of the anamorph.
Culture characteristics: On CMD colonies 6-7 mm diam, circular, slightly convex, margin entire, velvety, finely furrowed, brown to beige-brown occasionally with a dark brown outer zone of submerged growth, reverse dark brown. On MLA colonies 5-9 mm diam, circular to irregular, pulvinate, margin entire, velvety, furrowed, brown, reverse dark brown. On OA colonies 7-8 mm diam, circular, convex, margin entire, velvety, light grey, olivaceous grey at the margin, light olivaceous pigment diffusing into agar, reverse dark olivaceous-grey. On PCA colonies 5-7 mm diam, circular, slightly convex, margin entire to finely lobate, velvety, furrowed, beige-brown with a dark brown outer zone of submerged growth, reverse dark brown. Sporulation was sparse on MLA and PCA, absent on CMD and OA.
Habitat and geographical distribution: Saprobe on decaying wood, known from New Zealand.
Notes: Zanclospora aurea is well-distinguishable from other members of the genus in the golden-yellow colonies formed on the natural substrate, presence of connective elements on conidiophores, and falcate, strongly curved to almost horseshoe-shaped conidia. In vitro, colonies were very slow growing; on MLA they appeared pulvinate of somewhat crustose consistency, while on other media colonies were effuse with a spreading edge. The stanjehughesia-like synanamorph was not observed on the natural substrate or any of the growth media.  Etymology: Clavulate (Latin) club-shaped, alternative form of clavate from clava (club), referring to the shape of microconidia.

Zanclospora clavulata
Culture characteristics: On CMD colonies 15-17 mm diam, circular, slightly raised, margin entire to finely fimbriate, lanose, floccose, grey-brown with an outer dark olivaceous brown zone of submerged growth, reverse dark brown. On MLA colonies 24-27 mm diam, circular, slightly convex, margin entire, lanose, beige-grey with an outer dark grey-brown zone of submerged growth, reverse dark brown. On OA colonies 17-18 mm diam, circular, flat to slightly convex, margin entire to finely fimbriate, sparsely lanose, floccose, grey with an outer olivaceous grey zone of submerged growth, reverse dark olivaceous grey. On PCA colonies 19-20 mm diam, circular, flat to slightly raised, margin finely fimbriate, lanose, light grey-brown with an outer olivaceous brown zone of submerged, reverse brown-grey. Sporulation was moderate on CMA with U. dioica stems, absent on CMD, MLA, OA and PCA.
Habitat and geographical distribution: Saprobe on decaying wood, known from Portugal. Environmental data indicate another occurrence in the soil in Tasmania in the mixed forest biome with Eucalyptus sp. as a dominant plant species (Figures 6 and 7) (Table S2).
Notes: Zanclospora clavulata is closely related to Z. falcata, Z. iberica and Z. novaezealandiae. Initially, it was listed under Z. iberica [9], but the present four-gene phylogeny revealed it is a separate species (Figure 3). Its wildtype is unknown, so comparison with other species is somewhat limited. The herbarium material, which is no longer available, contained only the stanjehughesia-like synanamorph, but Zanclospora and the stanjehughesia-like synanamorphs were formed on CMA with U. dioica stems. In vitro, conidiophores of Z. clavulata produced only microconidia that are narrower (0.7-1 µm) than in other species (1-2.5 µm). When compared to other Zanclospora that exist in culture, it is the fastest-growing species. Etymology: Falcate (Latin) from falx (sickle) meaning shaped like a sickle, and -ate (resembling), referring to the shape of conidia.

Zanclospora iberica
Material Habitat and geographical distribution: The species is a saprobe on decayed plant material and is so far known only from Spain ( [9], this study).
Notes: For additional description and illustrations refer to Hernández-Restrepo et al. [9]. Although conidiophores similar to Stanjehughesia were the only anamorphic structure observed on the natural substrate, Zanclospora and stanjehughesia-like ( Figure 14A,B) conidiophores were observed on Urtica stems in vitro. The Zanclospora conidiophores matched well the protologue but varied in size. Those which developed from mycelium on agar were smaller and less complex ( Figure 14J-L) than those formed directly on Urtica stems ( Figure 14C-I). The smaller conidiophores and single conidiogenous cells produced only microconidia.
Zanclospora iberica appears to be most similar to Z. novae-zelandiae. Because the wildtype of Z. iberica is unknown, both species can only be compared in culture. The sterile apical part of the conidiophore stalk and branches is smooth in both species. However, it is ornamented with disk-like excrescences in Z. novae-zelandiae on the natural substrate ( Figures 15J and 16E,F). In the absence of this diagnostic character, the species are practically indistinguishable when grown in culture. The size of their macroconidia overlaps significantly, i.e., (12.5-)15.5-25 × 2-3 µm (this study) and 12.5-22 × 2-3 µm fide Hernández-Restrepo et al. [9] in Z. iberica vs 16-24 × (1.5-)2-3 µm in Z. novae-zelandiae (ICMP 15781 ex-epitype strain). On the natural substrate, the macroconidia of Z. novae-zelandiae tend to be longer, 24-28.5 × (2-)2.5-3 µm in the epitype (PDD 80663), 23.5-39 × 2.5-3.5 µm in the holotype (PDD 20727). Thus, we expect that also macroconidia of Z. iberica may be longer in natural conditions. In comparison to Z. novae-zelandiae, the conidiophores of Z. iberica appeared to be less flexuous, and conidia were usually slightly inflated towards the basal end.  Habitat and geographical distribution: The species is a saprobe on decorticated wood, known so far from Asia in Thailand [86]. Environmental data confirm another occurrence in the soil in Guangdong province, southeastern China, in the forest biome with Schima superba and Michelia macclurei as the dominant plant species (Figures 5 and 7) (Table S2).
Notes: For description and illustrations, see Perera et al. [86]. The Zanclospora conidiophores were not observed, the stanjehughesia-like conidiophores, probably mistaken for setae, occurred on ascomata and the nature substrate around them (see Discussion). In the four-gene phylogenetic tree (Figure 3), Z. jonesii clustered as a sister to Z. tropicalis. Both species share cylindrical ascospores bent near the basal end, but differ in their size; the ascospores of Z. jonesii are shorter and narrower, 16.2-17.7 × 2.8-3.6 µm fide Perera et al. [86] than those of Z. tropicalis, 19-26 × 3.2-6.3 µm fide Fernández and Huhndorf [12]. Culture characteristics: On CMD colonies 11-14 mm diam, circular, raised, margin fimbriate, lanose, cobwebby at the margin, colony centre beige, dark brown at the margin, later with a distinct dark brown outer zone of submerged growth, reverse dark brown to nearly black. On MLA colonies 15-17 mm diam, circular, convex, margin fimbriate, lanose, floccose, aerial mycelium with numerous minute colourless exudates, beige, brown at the margin, reverse dark brown. On OA colonies 8-9 mm diam, circular, flat, margin fimbriate, cobwebby becoming mucoid, smooth towards the periphery, mainly comprising of submerged mycelium, olivaceous black, reverse black. On PCA colonies 10-11 mm diam, circular, flat, margin fimbriate to rhizoidal, lanose, floccose, beige-brown centrally, with a prominent dark brown to russet outer zone of the submerged growth, reverse dark brown. Habitat and geographical distribution: Saprobe on decorticated wood, known from North America in the USA (Illinois, Indiana, North Carolina, Wisconsin) [12].
Notes: For description and illustrations, see Fernández and Huhndorf [12]. According to these authors, repeated transfers of mycelium in vitro resulted in the loss of the typical Zanclospora conidiophores; they were reduced to single conidiogenous cells or whorls of cells on hyphae reminiscent of Phaeostalagmus and produced smaller, clavate conidia. It is in agreement with our observations in other Zanclospora, conidiophores formed on agar are generally less complex and produce only microconidia. Zanclospora lateriphiala is similar to Z. falcata but differs in the smooth setiform part of conidiophores and branches and smaller asci and ascospores. The stanjehughesia-like synanamorph has not been yet reported for Z. lateriphiala. For a full comparison with other species, see Table 2.
Culture characteristics: On CMD colonies 11-12 mm diam, circular to irregular, flat, margin fimbriate, velvety-lanose at the centre becoming cobwebby towards the margin, colony centre olivaceous brown, olivaceous beige at the margin, reverse dark brown. On MLA colonies 11-12 mm diam, circular, slightly convex, margin fimbriate, velvety, cobwebby at the margin, colony centre beige, dark brown at the margin, reverse dark brown. On OA colonies 6-7 mm diam, circular to irregular, flat, margin fimbriate to rhizoidal, cobwebby centrally, smooth towards the periphery, dark olivaceous brown, light brown pigment diffusing into the agar, reverse brown. On PCA colonies 5-7 mm diam, circular to irregular, flat, margin fimbriate, velvety, sometimes mucoid at the centre, beige-brown, with a dark brown outer zone of submerged growth, reverse dark brown. Sporulation was absent on CMD and OA, moderate on PCA and MLA, abundant on CMA with Urtica stems.
It is challenging to distinguish Z. novae-zelandiae from Z. iberica. In culture, when grown on Urtica stems, both species are similar in characters of conidia, phialides and smooth conidiophores. For comparison of both species and discussion, see notes to Z. iberica.
In addition to New Zealand, Z. novae-zelandiae was also reported from other geographical areas (Figure 7). However, all of these collections lacked the ornamentation of the conidiophore wall, or this character was not mentioned in the description [8,[20][21][22][23]. Besides, the size of conidia varied among these collections and corresponded to the longspored Z. novae-zelandiae s. str., the newly segregated short-spored Z. falcata, both from New Zealand, but also to Z. lateriphiala from North America. Examination of three collections identified as Z. novae-zelandiae by Schoknecht and Crane [21] from the USA with smooth conidiophores and conidia 15.5-23 × 2.3-3.3 µm revealed they represent Z. lateriphiala. The specimen of Z. novae-zelandiae recorded from Brazil by Almeida et al. [8] has conidia significantly shorter (10-16.5 × 1-2 µm) than Z. falcata, Z. lateriphiala and Z. novae-zelandiae, and likely represents another cryptic species in the Z. novae-zelandiae species complex. Interestingly, Mel'nik et al. [23] recorded a specimen of Z. novae-zelandiae from Vietnam with exclusively unbranched, smooth conidiophores and conidia 22-24(-26) × 2-2.4 µm. Although we are aware of inconsistencies in the published phenotypes of Z. novae-zelandiae, these records are listed above but need to be verified. We present the first molecular data of Z. novae-zelandiae; however, more concentrated sampling is required to assess its global geographical distribution. We should also consider the possibility that the species is endemic to New Zealand. Habitat and geographical distribution: Saprobe on decaying wood, known from New Zealand [85].
Culture characteristics: On CMD colonies 12-13 mm diam, circular, convex, margin entire to weakly fimbriate, lanose, somewhat floccose, beige at the centre, dark beige to brown towards the margin, reverse dark brown. On MLA colonies 15-16 mm diam, circular, slightly convex centrally, margin weakly fimbriate, lanose, cobwebby towards the periphery, zonate, beige at the centre, with dark brown to dark reddish-brown middle zone and paler brown outer zone, reverse brown. On OA colonies 6-8 mm diam, circular, flat, margin entire, smooth to cobwebby, dark olivaceous grey with an outer zone of a similar colour of submerged growth, reverse dark brown. On PCA colonies 5-6 mm diam, circular, flat, margin weakly fimbriate, cobwebby, brown with a dark brown outer zone of submerged growth, reverse dark brown. Sporulation was abundant on MLA, absent on CMD, OA and PCA.
Other material examined: NEW ZEALAND, West Coast, Grey district, Victoria Forest Park, Lake Christabel track, Palmer's Hut ca. 18  Habitat and geographical distribution: Saprobe on decaying wood of Nothofagus sp. and another unidentified host, known from New Zealand.
Notes: For description and illustrations, refer to Fernández and Huhndorf [12]. In the four-gene phylogeny, Z. sylvatica was inferred as a sister to Z. jonesii and Z. tropicalis. Zanclospora sylvatica forms only the phaeostalagmus-like synanamorph in vitro, the stanjehughesia-like and Zanclospora conidiophores were not observed. The species is characterised by symmetrical, 3-septate, fusiform ascospores in contrast to the asymmetrical ascospores of Z. jonesii and Z. tropicalis. Habitat and geographical distribution: Saprobe on decaying wood, known only from the Caribbean (Puerto Rico) and Central America (Costa Rica) [12].
Habitat and geographical distribution: Saprobe on decaying wood of Nothofagus sp. and other unidentified hosts, known from New Zealand.

Doubtful and Excluded Species
This section includes species retained in Zanclospora based on morphology but not verified by molecular DNA data, as well as species excluded or transferred to other genera on the basis of molecular evidence and/or morphological data. Habitat and geographical distribution: Saprobe on the bark of Eucalyptus propinqua, known from Brazil [3].
Notes: For description and illustrations, see Sutton and Hodges [3]. This species closely resembles other species in the genus with typical Zanclospora conidiophores and falcate conidia. It can be distinguished in having simple conidiophores with conidiogenous cells confined to two separate regions. Habitat and geographical distribution: Saprobe on decaying leaves of an unidentified plant, known from Ecuador [10].

Zanclospora bicolorata
Notes: For description and illustrations, see Villavicencio et al. [10]. Zanclospora bicolorata has conidiogenous cells disposed of more or less in the middle of the simple, setiform conidiophores with suballantoid conidia, similar to Z. bonfinensis. However, their conidiophores differ in colour and ornamentation from conidiophores of other members of the genus. The conidiophores are pale brown or brown at the base becoming dark reddish-brown to dark brown and almost opaque toward the apex. The conidiophores are smooth-walled in Z. bicolorata, while Z. bonfinensis has the conidiophores smooth at the base, becoming verrucose at the apex [8,10]. Habitat and geographical distribution: Saprobe on decaying leaves of unidentified dicotyledonous plant, known from Brazil [8].
Notes: For description and illustrations, see Almeida et al. [8]. For comparison, see comment under Z. bicolorata and Table 2. Notes: For description and illustrations, see Hughes and Kendrick [1]. Two varieties were described under this species, Z. brevispora var. brevispora [1] and var. transvaalensis [5]. They are distinguished mainly by the number of conidiogenous cells and conidial characters. Zanclospora brevispora var. brevispora has more conidiogenous cells and curved somewhat smaller conidia [1,5]. This species fits well in the concept of Zanclospora, however further studies are needed to resolve, whether it is one or two species. . This species deviates from the generic concept of Zanclospora in the morphology of the conidiogenous cells and the way they are inserted on the conidiophore. The phialides of Z. indica are broadly lageniform, extend into a narrow neck and elongate percurrently to form a secondary phialide. The collarette is well-defined. In addition, some of the secondary phialides are illustrated in a lateral position on the primary phialide suggesting a sympodial elongation. The phialides diverge from the conidiophore and are arranged in several whorls below the transverse septa. Other Zanclospora differs from Z. indica in having phialides with an indistinct collarette; they are tightly appressed to the conidiophore and arranged in compact fertile zones. Based on comparative morphology, Z. indica is not accepted in the genus. Habitat and geographical distribution: Saprobe on leaf litter, known only in Ivory Coast [4].
Notes: Although the species was described with conidiogenous cells with a single phialidic opening [7], the examination of the paratype, other herbarium material and living cultures revealed that the conidiogenous cells are polyphialidic, indeterminate, the upper part is sympodially elongating and contains numerous openings within minute collarettes. The holotype PDD 76621 of Z. urewerae did not contain any herbarium material, only a dried culture. A personal note on the holotype of Z. urewerae by J.A. Cooper, dated 2 July 2010, was posted on the website of the PDD herbarium and reads as: "A subsequent examination of conidiogenous cells under SEM suggests this is Selenosporella".
Based on a detailed comparison of the revised material of Z. urewerae and the description and illustration of Selenosporella verticillata [94], we consider both species identical. Therefore, a new genus Brachiampulla is proposed for S. verticillata in the Xyladictyochaetaceae, and Z. urewerae is reduced to synonymy under the former species. Brachiampulla verticillata resembles S. acicularis [95] and S. aristata [96] in the morphology of conidiogenous cells with minute phialidic openings formed after sympodial elongation. On the other hand, highly similar Selenosporella species characterised by ampulliform, polyblastic, sympodially elongating conidiogenous cells arranged in whorls and unicellular, hyaline conidia but with holoblastic conidiogenesis include S. curvispora [97], the type species of Selenosporella, and also S. nandiensis [98] and S. setosa [99]. Etymology: Stephanos (Greek) crown, -phora (Greek) bearing, from pherein (to bear), referring to a group of short branches resembling a crown at the conidiophore apex, -ella, diminutive, used as a name-forming suffix.
Habitat and geographical distribution: The species is a saprobe on plant debris, known so far from Africa in Nigeria [6]. Notes: For additional description and illustrations, see Calduch et al. [6]. Stephanophorella stellata resembles Zanclospora in setiform conidiophores, lateral, determinate phialides arranged along the midsection in a fertile zone and hyaline conidia, but differs in phialides with a well-defined collarette and a crown of setiform branches at the conidiophore apex.

Morphology, Interspecific Variability and Life History of Zanclospora
Our phylogenetic analysis of the combined 18S, 28S and rpb2 sequences has provided strong support for the recognition of Zanclospora as a polyphyletic genus, with species distributed among three distantly related evolutionary lineages in the Sordariomycetes (Figure 1). The core of the genus, including Z. novae-zelandiae, clustered in the Chaetosphaeriaceae (Chaetosphaeriales), Z. stellata is positioned in the Vermiculariopsiellaceae (Vermiculariopsiellales), and Z. urewerae is nested in the Xyladictyochaetaceae (Xylariales).
Within the Chaetosphaeriaceae, Zanclospora is resolved as a well-supported monophyletic clade (Figure 2). It is a holomorphic genus encompassing 17 species and two varieties. Cultivation studies, morphological comparisons on natural substrates and phylogenetic analysis of four markers (ITS, 28S, tef1-α and tub2) of 21 strains representing 12 species revealed unknown pleomorphism in Zanclospora. Our barcoding gap analysis showed that the barcodes widely used in Ascomycota (i.e., ITS, tef1-α and tub2) are applicable for species delimitation in Zanclospora.
Three different conidiophore morphotypes hitherto considered unrelated, occur in the life cycle of several members of the genus. They include the anamorph with the typical Zanclospora conidiophores and phaeostalagmus-and stanjehughesia-like synanamorphs. Based on novel molecular and phenotypic data, the generic concept of Zanclospora is emended to include teleomorphic and anamorphic characters. Ten Zanclospora species have known teleomorph-anamorph connections, which were either experimentally established ( [12], this study) or estimated, based on the juxtaposition of both morphs [11].
The diagnostic characters of Zanclospora include erect, pigmented, setiform conidiophores that bear one to several whorls of pale brown, sessile monophialides arising just below the septa. The phialides are appressed to the conidiophore; they form a compact fertile zone and produce hyaline, aseptate conidia without setulae. Macroconidia form only on natural conditions and vary in shape from falcate, horseshoe-shaped, obovoid to bacilliform, whereas microconidia form only in culture and are clavate to oblong-clavate, ellipsoidal to fusiform. However, the Zanclospora anamorphs growing on agar or natural substrate/Urtica stems in culture, providing semi-natural conditions, differ in size, overall complexity and appearance of conidiophores and also in conidial morphology. The synanamorph similar to Phaeostalagmus [19] occasionally occurs when the species is grown in culture. It shares with Zanclospora the arrangement of lateral phialides in whorls on the conidiophore, or they can be disposed of in a verticillate fashion on short branches. To a certain extent, it may represent a simplified Zanclospora under in vitro conditions. A similar analogy can be found between Zanclospora and the stanjehughesia-like synanamorph. The latter forms dark brown, multiseptate conidiophores, which are wider and several times longer than those of Zanclospora, but remain sterile, rarely with one or two lateral phialides ( Figure 14A). Although conidiophores of this synanamorph are strikingly reminiscent of conidia of Stanjehughesia (conidiophores are absent, reduced to conidiogenous cells), a dematiaceous hyphomycete segregated from Sporidesmium [18], they represent a different structure with a different function. The stanjehughesia-like synanamorph has been frequently observed on natural substrates as well as in culture. Phaeostalagmus and Stanjehughesia form separate lineages in the Chaetosphaeriaceae tree (Figure 2).
De Hoog [100] addressed plasticity and variation in conidiogenesis of yeast-like fungi and distinguished synanamorphs into two basic categories, i.e., pleoanamorphy dependent on environmental conditions and spontaneous pleoanamorphy that exists under identical environmental conditions. The phaeostalagmus-like synanamorph was observed only when grown in culture. We assume that it may represent the group of synanamorphs that are influenced by environmental conditions. The identification of Zanclospora became challenging because of the high degree of variability in the anamorphic morphology and their irregular presence on the natural substrate and in culture. This is especially true in cases when the Zanclospora conidiophores are absent on material from nature and form only in culture, and the stanjehughesialike synanamorph is the only anamorphic phenotype present, i.e., Z. clavulata, Z. iberica, Z. ramifera, Z. xylophila. In addition, some strains derived from ascospores lack both Zanclospora and stanjehughesia-like conidiophores, and only the phaeostalagmus-like synanamorph is formed in culture, namely Z. phaeostalacta, Z. sylvatica and Z. tropicalis. Although the anamorph of Z. jonesii is unknown, brown, sinuous, cylindrical 'setae' arising from the base of ascomata and in their vicinity were described and illustrated by Perera et al.
Although the stanjehughesia-like conidiophores only mimic conidia of Sporidesmium and its segregates, we can compare both groups in terms of the formation of phialides on these morphologically similar structures or in their life cycle. Phialidic synanamorphs are rare in Sporidesmium and similar taxa. Stanjehughesia hormiscioides (teleomorph Umbrinosphaeria, Chaetosphaeriaceae) form a chloridium-like synanamorph when grown in culture [101]. Kirk [102] described a phialidic synanamorph accompanying Sporidesmium clarkii; discrete phialides producing filiform, bent microconidia are born directly on conidia or separate conidiophores, solitarily or on compact branches.
In the phylogenetic tree inferred from the four combined loci (Figure 3), Z. novaezelandiae and three other morphologically similar species, i.e., Z. clavulata, Z. falcata and Z. iberica, were resolved as closely related but separate species lineages. Except for Z. clavulata, which forms only microconidia, they are remarkably similar in macroconidia, conidiogenous cell and conidiophore morphology. Ornamentation of the setiform, apical part of the conidiophores of Z. novae-zelandiae [1], originally unique to this species, is newly described in Z. falcata. Moreover, the excrescences on the conidiophore surface do not develop in culture. Molecular data suggest that Z. novae-zelandiae is a species complex. The four-gene phylogeny and morphological comparison of specimens tentatively identified as Z. novae-zealandiae with the holotype of this species facilitated their correct identification and allowed interpretation of the type material. As a result, a new species, Z. falcata, was designated and separated from Z. novae-zelandiae by primarily anamorphic features and Z. novae-zelandiae was epitypified. Given the known broad geographical distribution and the described variability in conidial size and conidiophore characters of Z. novae-zelandiae, this molecular study is the first to suggest that it is an unexplored complex that may contain other cryptic species.

Global Biogeography of Zanclospora
Zanclospora was identified as a low-diverse genus comprising 12 verified species and a relatively low number of phylotypes inferred from the environmental DNA (Figures 3, 5 and 6). We could not link most of the phylotypes between ITS1 (6) and ITS2 (14) datasets, except for three, which contain the whole ITS, but we assume that due to their geography and ecology ( Figure 7) they overlap or some represent known but so far not sequenced species. In our study, the minimal number of taxa from environmental DNA data, roughly corresponding to the level of species, was 14 ( Figure 6). Data from NCBI GenBank and UNITE databases contributed to this diversity study by only one sequence (DQ124120) belonging to the phylotype ITS1-ENV5 ( Figure 5). These results confirm GlobalFungi, the most comprehensive atlas of global fungal distribution, as a powerful tool for diversity studies.
Short reads attributable to Zanclospora were extremely rare. Theoretically, it could be due to the PCR amplification bias, resulting in an underestimation of studied fungi. In general, Zanclospora can be amplified and sequenced under standard conditions with commonly used ITS primers. On the contrary, data mining of sequence-only members of the Chaetosphaeriaceae resulted in tens of thousands of hits so PCR bias appears unlikely as an explanation for the lack of Zanclospora sequences in these datasets (Tables S2 and S3).
We, therefore, expect that our findings reflect reality rather than being due to limitations of PCR and massively parallel sequencing.
We can further confirm that Zanclospora inhabits a spectrum of substrates including decaying bark, wood and fallen leaves, but also living roots and more often soil, which was not yet known. Our data mining confirms fully the pattern of distribution known from classical studies and expands its distribution and ecology. Known centres of geographical distribution (New Zealand, Central America and Caribbean, South America) were confirmed from metabarcoding data and also have shown that Southeast Asia represents another hotspot of Zanclospora diversity. In conclusion, based on accepted species, most of which are known from the type collection only, and data from the GlobalFungi database, we have demonstrated that Zanclospora is a very rare genus of worldwide distribution, living in humid natural forests (mostly temperate rainforest and tropical rainforest zones) in soil and on decaying plant matter. Our study is the first application of the GlobalFungi database for diversity, biogeography and ecology survey of a group of fungi.
The present phylogenies of metabarcoding based on ITS1 or ITS2 for the majority of sequences demonstrate the importance of environmental DNA sequences in phylogenybased taxonomic studies. There has been much discussion in the mycological community on using metagenomic DNA from environmental samples as holotypes and in general for taxon naming [103][104][105][106][107][108][109][110]. In addition, the first studies on naming DNA-based taxa have already been published, e.g., De Beer et al. [111], Lücking and Moncada [112], Kalsoom Khan et al. [113]. The ITS1 and ITS2 phylotypes attributed to Zanclospora represent either new lineages or already described species that have not yet been sequenced. They expand the known geographical distribution and ecology and help to estimate the number of existing species; the number of recovered phylotypes almost doubles the number of known Zanclospora species. Using metabarcoding data, we gathered most information on the distribution, ecology and phylogeny of Zanclospora except morphology. However, Zanclospora includes three anamorphic phenotypes, which represent most of the morphological variability of the genus. We, therefore, prefer to define the environmental sequences of the 'dark' Zanclospora taxa as phylotypes. We hope that future sampling and increased efforts to cultivate these fascinating fungi will reveal new connections between cultivable fungi and their counterparts identified from environmental DNA.

Zanclospora and Its Allies
Zanclospora can be compared to Cryptophiale [15], Cryptophialoidea [16] and Kionochaeta [17], all members of the Chaetosphaeriaceae, which form separate lineages (Figure 2). They share pigmented, mononematous, setiform conidiophores with phialides arranged in fertile zones and hyaline conidia. The main distinguishing characters are the arrangement of phialides on the conidiophore, branching pattern of the conidiogenous apparatus and presence or absence of a shield-shaped plate. In Cryptophiale, Cryptophialoidea and Zanclospora the phialides are discrete, sessile, while in Kionochaeta they are integrated, disposed at the apex of conidiophores or branches. Zanclospora is the only of these four genera with phialides arranged vertically along the main axis of the conidiophore in whorls and around the entire conidiophore perimeter. The phialides have a poorly defined collarette and are not obscured by a shield or any sterile structure. Cryptophiale, on the other hand, encompasses fungi with the fertile region composed of sessile phialides with indistinct collarettes that arise at a right angle to the conidiophore axis, arranged in one or two palisade rows, and covered and partly enclosed by a shield-shaped plate of sterile, fused cells. The conidiophores are apically sterile, branched or unbranched, or with lateral branches. Cryptophiale species with phialides with a well-defined collarette arranged on only one side of the conidiophore and lacking the protecting shield were segregated into Cryptophialoidea (Cr.) by Kuthubutheen and Nawawi [16]. Molecular data, which are available only for non-type strains of Cr. fasciculata [114] and two Cryptophiale [75], suggest a close relationship of both genera. The fundamental character of Kionochaeta is the branching pattern of the conidiogenous apparatus. It comprises compact or loosely arranged branches bearing conidiogenous cells, irregularly branched or in a penicillate fashion. The conidiophores are unbranched or with lateral branches inserted in or above the fertile region.
In the ITS-28S phylogeny of the Chaetosphaeriaceae (Figure 2), the closest relatives to Zanclospora were Chaetosphaeria minuta [12] and a clade containing Cryptophiale, Cryptophialoidea, Kionochaeta ivorensis and Conicomyces. In the arrangement and morphology of phialides, the anamorph of Ch. minuta resembles Cryptophialoidea [16]. The majority of species of Cryptophialoidea are monophialidic, e.g., Cr. ramosa, Cr. secunda, the type species, or Cr. uncispora [16,114,115], mono-or rarely polyphialidic in Cr. fasciculata [24,114], or polyphialidic in Cr. manifesta [87]. Moreover, some Cryptophialoidea have phialides arranged in discrete unilateral bundles (Cr. fasciculata, Cr. manifesta, Cr. ramosa), while in other species the phialides are evenly distributed along the conidiophore. The conidia vary in shape and septation; they are 0-1-septate, fusiform, falcate or apically hooked. The lack of molecular data does not allow assessing the taxonomic value of these features in Cryptophialoidea. The separate position of Cr. fasciculata and Ch. minuta with the cryptophialoidea-like anamorph suggests further variability of this small genus, which currently includes five species [116]. The ex-type strain of K. ivorensis [17,117] is shown as a sister to Cryptophiale and Cryptophialoidea, while K. ramifera [87], the type species of the genus, and two other Kionochaeta form a separate, monophyletic lineage ( Figure 2). The grouping of K. ivorensis suggests even wider phenotypic plasticity of the Cryptophiale clade than has been described.
Conicomyces [118], on the other hand, is morphologically well distinguishable from Zanclospora. The genus was introduced for synnematous, pigmented hyphomycetes with integrated phialides and septate, setulate conidia.
In the phylogenetic analysis of the combined 18S, 28S and rpb2 sequences (Figure 1), the ex-type strain of Z. stellata clustered in the Vermiculariopsiellales on a single branch basal to a clade comprising Tubulicolla [42] and Vermiculariopsiella [9]. Therefore, Z. stellata is excluded from Zanclospora into a new genus, Stephanophorella. Members of the Vermiculariopsiellales are saprobic, dematiaceous hyphomycetes that form effuse colonies or sporodochia. Stephanophorella resembles Zanclospora in setiform conidiophores and the arrangement of sessile, lateral phialides, but differs mainly in well-defined collarettes and the dark, opaque, setiform part of the conidiophore with branches inserted in a stellate fashion at the apex.

Selenosporella and Morphologically Similar Taxa
Due to the newly discovered morphological characters of conidiogenous cells, Z. urewerae was found to be conspecific with Selenosporella verticillata [94] and excluded from Zanclospora. A new genus Brachiampulla is proposed for Z. urewerae and its systematic placement is resolved with the four combined loci in the Xyladictyochaetaceae (Xylariales). Brachiampulla, based on B. verticillata, includes saprobes on fallen leaves, which are morphologically reminiscent of Selenosporella.
Although Sutton and Hodges [94] described the conidiogenous cells of B. verticillata as polyphialidic, they added that: "conidiogenesis could well be holoblastic rather than enteroblastic".
We agree with Sutton and Hodges [94] to interpret the structures surrounding the conidiogenous loci of B. verticillata as minute collarettes on polyphialidic conidiogenous cells. A similar mode of conidiogenesis was described in Xyladictyochaeta, a sister genus of Brachiampulla in the Xyladictyochaetaceae [9]. It is likely that Selenosporella and selenosporella-like fungi accommodated in several phylogenetically different groups vary in the mode of conidiogenesis. In addition, Selenosporella and Brachiampulla share similar morphology of the conidiogenous cells that are ampulliform to lageniform, undetermined with sympodially extending apex, arranged in whorls along the conidiophore and conidia that are hyaline, unicellular, lunate or falcate.
Within the Xylariales, Brachiampulla is comparable to Selenodriella [133,134] in pigmented macronematous conidiophores, ampulliform, lateral, polyblastic conidiogenous cells arranged in whorls or terminally, and unicellular, hyaline conidia, but Selenodriella differs in holoblastic-denticulate conidiogenesis. Ceratocladium [135], represented by C. polysetosum [136] in our phylogeny, shares with Brachiampulla polyblastic, discrete, lateral, ampulliform conidiogenous cells and unicellular, hyaline conidia, but differs in the presence of setae and conidiogenous cells growing on climbing fertile hyphae. The species was resolved as a separate lineage without affinity to any known families. Its relationship to morphologically similar Circinotrichum was investigated by Hernández-Restrepo et al. [9]. A similar arrangement of sympodial conidiogenous cells found in Brachiampulla occurs, for example, in Umbellidion [137]. In Umbellidion (genera incertae sedis), the conidiogenous cells are cylindrical to lageniform and form only the apical whorl, occasionally the pigmented conidiophore proliferates and a second whorl is developed.

Conclusions
Zanclospora is pleomorphic genus rarely encountered on decaying bark, wood or leaf litter. Several species added to the genus have broadened its boundaries, but the generic concept has never been evaluated with molecular DNA data. Our knowledge of Zanclospora biogeography is minimal; the field records are unverified with molecular data, and only one or a handful of collections have been recorded for each species. Using six genetic markers, Zanclospora was shown to be polyphyletic, with three distantly related lineages in the Sordariomycetes. Zanclospora s. str. was resolved as a strongly supported monophyletic clade in the Chaetosphaeriaceae. Based on the results of phylogenetic analyses and phenotypic data, two new segregate genera, Brachiampulla and Stephanophorella, were proposed for Z. urewerae and Z. stellata, respectively. They represent two distantly related lineages in the Xylariales and Vermiculariopsiellales.
Zanclospora produces teleomorphs previously classified in the genus Chaetosphaeria. However, more frequently Zanclospora produces anamorphs characterised by erect, pigmented, setiform, often branched conidiophores, sessile monophialides with indistinct collarettes arranged in whorls and tightly appressed to the conidiophores or branches, and hyaline, aseptate macroconidia (on the natural substrate; falcate to horseshoe-shaped to obovoid) and microconidia (in culture; usually clavate to ellipsoidal) without setulae. Seventeen species are accepted, 12 of which have been verified with DNA sequence data. We have discovered variability in anamorphic characteristics associated with three anamorphic stages, of which phaeostalagmus-and stanjehughesia-like are newly described. Phylogenetic analyses of environmental ITS1 and ITS2 sequences retrieved from the Glob-alFungi database provided insight into the global biogeography of Zanclospora. Seven and 15 phylotypes have been identified in samples derived from soil, dead wood and roots. The field records verified by DNA data indicated two main diversity cores in Australasia and Caribbean/Central America. Environmental ITS data suggested Southeast Asia as a third hotspot of Zanclospora diversity. Interestingly, environmental sequences of these fungi were completely missing in Europe and North America, which are the best-sampled continents in the GlobalFungi database.
Our study demonstrated the importance of in vitro studies to assess anamorphic plasticity and systematics of Zanclospora and the use of environmental sequences to expand our knowledge on biogeography and unknown interspecific diversity. It has also confirmed that different phenotypes distinguished within Zanclospora are phylogenetically distinct. We hope that future sampling and increased efforts to cultivate these fascinating fungi will reveal new connections between cultivable fungi and their counterparts identified from environmental DNA. Although we addressed issues related to evaluating taxonomic diagnostic criteria at the generic level, we were unable to obtain all species to assess their phylogenetic relationships. They are retained in Zanclospora based on morphology.
Supplementary Materials: The following are available online at https://www.mdpi.com/article/10 .3390/microorganisms9040706/s1, Table S1: Taxa of the Sordariomycetes, their collection numbers and accession numbers for sequences retrieved from GenBank, Table S2: A list of environmental samples attributable to Zanclospora with references to all studies in the GlobalFungi database, Table S3: Taxa related to Zanclospora and comparison of their BLAST search in the GlobalFungi database, Table  S4: Estimates of evolutionary divergence between ITS rDNA, 28S rDNA, tub2 and tef1-α sequences, Table S5: The biogeography, substrate and habitat affinity of Zanclospora and outgroup taxa inferred from the GlobalFungi database, Table S6: Published records of Zanclospora with host, substrate, country of the collection and references.