Neodendryphiella, a novel genus of the Dictyosporiaceae (Pleosporales)

Abstract In a survey of soil and herbivore dung microfungi in Mexico and Spain, several dendryphiella-like species were found. Phylogenetic analyses based on ITS and LSU sequences showed that these fungi belonged to the family Dictyosporiaceae (Pleosporales) and represent an undescribed monophyletic lineage distant from Dendryphiella. Therefore, the genus Neodendryphiella is proposed to accommodate three new species, N.mali, N.michoacanensis and N.tarraconensis. The novel genus shares morphological features with Dendryphiella such as differentiated conidiophores and polytretic integrated conidiogenous cells, that produce acropetal branched chains of conidia. Neodendryphiella differs in the absence of nodulose conidiophores bearing conidiogenous cells with pores surrounded by a thickened and darkened wall, typical features in the conidiogenous apparatus of Dendryphiella. In addition, the phylogenetic and morphological analysis of several reference strains of different Dendryphiella species, available for comparison, support the proposal of D.variabilissp. nov., which mainly differs from the other species of the genus by having conidia up to 7 septa and highlight that D.vinosa and D.infuscans are obscure species that require further taxonomic review.


Introduction
In an ongoing survey of asexual microfungi from soil and herbivore dung, several interesting specimens morphologically consistent with Dendryphiella were found from samples collected in Mexico and Spain. Dendryphiella is a dematiaceous hyphomycete proposed by Bubák and Ranojevič (Ranojevič 1914) and typified with D. interseminata, which is currently considered a synonym of D. vinosa (Reisinger 1968). Dendryphiella vinosa is a saprobic fungus commonly found on plant debris, especially on the decaying herbaceous stems of several plants (Ellis 1971, Mercado Sierra et al. 1997. The genus is characterised by pigmented conidiophores, with terminal or intercalary polytretic conidiogenous cells, with dark scarring on the nodose swellings, producing acropleurogenous, solitary or catenate conidia, which are commonly multi-septate and cylindrical with rounded ends (Ellis 1971). Although Index Fungorum and MycoBank list 17 taxa in Dendryphiella, a recent review of literature reported only 12 species are accepted, including the newly proposed D. fasciculata (Liu et al. 2017). Dendryphiella pitsanulokensis is the latter species added to the genus (Hyde et al. 2018). Previous phylogenetic studies, conducted mainly from sequence data of the 18S nrDNA (SSU), 28S nrDNA (LSU) and the internal transcribed spacer (ITS) nrDNA regions, showed that the marine species D. arenariae and D. salina were phylogenetically distant from the type D. vinosa and related to the Pleosporaceae (Gareth Jones et al. 2008, Suetrong et al. 2009). Both species were therefore moved to the genus Paradendryphiella (Woudenberg et al. 2013) and, more recently, D. vinosa was included in the family Dictyosporiaceae (Tanaka et al. 2015, Boonmee et al. 2016). However, DNA sequence data for Dendryphiella species is very limited to create a robust taxonomy for the genus. Only LSU and/or ITS sequences of D. eucalyptorum, D. fasciculata, D. paravinosa, D. pitsanulokensis and D. vinosa are available (Crous et al. 2014, Liu et al. 2017, Hyde et al. 2018). In addition, with the exception of the first four mentioned, there is no ex-type culture of other species of this genus and only reference strains of D. vinosa and D. infuscans are available in public collections for comparison.
Despite the similarity of our soil isolates to Dendryphiella, a preliminary study revealed that they showed a low sequence relationship with members of this genus. On the other hand, they were closely related to the strain CBS 139.95 of Diplococcium (Di.) asperum, which was proven to be related to the Dictyosporiaceae (Shenoy et al. 2010, Boonmee et al. 2016. It is well known that the genus Diplococcium is highly polyphyletic, with species distributed across different classes of the Ascomycota, with its type species, Di. spicatum, being related to the Helotiales in Leotiomycetes (Shenoy et al. 2010, Hernández-Restrepo et al. 2017.
The aim of the present study was to resolve the taxonomy of these dendryphiellalike fungi which, based on analysis of the ITS and LSU loci, might represent a new genus in Dictyosporiaceae.

Sampling and fungal strains studied
Soil and dung samples collected in different geographical regions (Mexico and Spain) were studied using the wood baiting technique, moist chambers and dilution plating method according to Calduch et al. (2004). Using the first two techniques, we found three interesting dendryphiella-like fungi, which were isolated on Potato Dextrose Agar (PDA; Pronadisa, Madrid Spain) and incubated at room temperature in the dark. Additionally, six strains from the Westerdijk Fungal Biodiversity Institute, Utrecht, The Netherlands (CBS), which corresponded to D. vinosa (CBS 117.14, CBS 118716, CBS 121797 and CBS 584.96), D. infuscans (CBS 381.81) and Di. asperum (CBS 139.95) were included in the study for morphological and sequence comparison (Table 1).

DNA extraction, sequencing and phylogenetic analysis
The isolates were cultured on PDA for 7 days at 25 °C in darkness. The DNA was extracted through the modified protocol of Werner et al. (1998). The primer pairs ITS5/ITS4 and NL1/NL4b were used to amplify ITS regions, including the 5.8S gene and the D1/D2 domain of the LSU of the nrDNA, respectively, following Cano et al. (2004). PCR products were purified and stored at -20 °C until sequencing. The same pairs of primers were used to obtain the sequences at Macrogen Europe (Macrogen Inc. Amsterdam, The Netherlands). Finally, the sequences were assembled and edited using SeqMan v. 7.0.0 (DNAStar Lasergene, Madison, WI, USA) to obtain the consensus sequences.
The sequences generated in the present study were compared with those of the National Center for Biotechnology Information (NCBI) using the Basic Local Alignment Search Tool (BLAST). Alignments for each locus were made with the MEGA (Molecular Evolutionary Genetics Analysis) software v. 6.0. (Tamura et al. 2013), using the ClustalW algorithm (Thompson et al. 1994) and refined with MUSCLE (Edgar 2004) or manually, if necessary, on the same platform. The alignment included our sequences complemented with available sequences of NCBI and NITE Biological Resource Center (NBRC) of species that conformed the different genera of the family Dictyosporiaceae (Table 1). This determined the phylogenetic position of the dendryphiella-like isolates in this group of fungi. Phylogenetic reconstructions with ITS and LSU sequences were made using Maximum Likelihood (ML) and Bayesian Inference (BI) approaches under the MEGA software v. 6.0. (Tamura et al. 2013) andMrBayes v. 3.2.6 (Ronquist et al. 2012), respectively.
For the ML phylogenetic analysis of the LSU region, the best nucleotide substitution model determined by the same programme was the Kimura 2-parameter with Gamma distribution and, for the ITS region, it was the General Time Reversible model with Gamma distribution. The combined analysis of these two phylogenetic markers was tested through Incongruence Length Difference (ILD) implemented in the Winclada programme (Farris et al. 1994). For the combined analysis of LSU and ITS sequences, the best nucleotide substitution model was the General Time Reversible with Gamma distribution and Invariant sites (G+I). ML bootstrap values (BML) ≥70% were considered significant. For the BI phylogenetic analysis, the best nucleotide substitution model was determined using jModelTest (Posada 2008). For the LSU region, we used the Kimura 2-parameter with Gamma distribution (K80+G) and, for the ITS symmetrical model, we used Gamma distribution (SYM+G). The parameter settings used were two simultaneous runs of 5M generations, four Markov chains, sampled every 1000 generations. The 50% majority-rule consensus tree and posterior probability values (PP) were calculated after discarding the first 25% of the samples. A PP value of ≥0.95 was considered significant.
The DNA sequences and alignments generated in this study were deposited in GenBank (Table 1) and in TreeBASE (http://treebase.org), respectively.

Phenotypic study
The microscopic characterisation of the fungi studied was carried out according to Marin-Felix et al. (2017), using autoclaved pine twig arranged on the surface of water agar (PNA) after 7 days at 25 °C in darkness. Measurements and descriptions of the structures were taken from the specimens mounted in Shear's solution. Photomicrographs were obtained using a Zeiss Axio-Imager M1 light microscope (Zeiss, Oberkochen, Germany) with a DeltaPix Infinity × digital camera.
Nomenclatural novelties and descriptions were deposited in MycoBank (Crous et al. 2004). Ex-type cultures and holotypes, which consisted of dried cultures, were deposited at the CBS. Additionally, living cultures of the new species were also preserved in the Faculty of Medicine in Reus (FMR, Spain).

Results
The BLAST query revealed that LSU sequences of our dendryphiella-like isolates (FMR 16098, FMR 16234 and FMR 17003) showed a high percentage of identity (99%) with that of the isolate CBS 139.95 of Di. asperum and all of them were related to the Dictyosporiaceae. However, they showed a sequence identity of between 96-97% with LSU sequences of Dictyosporium species and other members of this family, including several species of Dendryphiella deposited in the GenBank. The ITS sequences did not match significantly any of those deposited in the NCBI database.
We carried out individual and combined analyses with the LSU and ITS loci to assess relationships with members of the Dictyosporiaceae, including reference strains of D. vinosa and D. infuscans sequenced in the present study. Single phylogenies of LSU and ITS loci encompassed 31 and 30 sequences, respectively, representing 12 genera and including Paradendryphiella arenaria and P. salina (Pleosporaceae) as outgroup (Figs. S1 and S2 in the supplementary material). LSU analysis comprised 630 bp from which 111 bp were variable and 84 bp phylogenetically informative. The ITS comprised 496 bp, 266 bp being variable and 206 bp being phylogenetically informative. The topology of trees for single loci were very similar and the ILD test showed that the LSU and ITS datasets loci were congruent (P = 0.16) and could be combined. The final combined analysis encompassed 30 sequences and comprised 1126 bp (ITS 496 bp, LSU 630 bp). The ML tree showed that FMR 16098, FMR 16234, FMR 17003 and CBS 139.95 clustered together in a well-supported undescribed monophyletic lineage representing a new genus in the family (Fig. 1). The LSU and ITS sequence comparison of the four isolates revealed them as different taxa. The low identity values together with the morphological differences found amongst them allow us to propose three new species in this new genus, which are described below.
Regarding the five Dendryphiella strains included in this study, only three (CBS 118716, CBS 121797 and CBS 854.96) nested in the well-supported clade of Dendryphiella and none of them matched sequences representative of the type species of the genus D. vinosa (DQ 307316.1, EU848590.1 and NBRC-03266901) and used previously by other authors to establish the relationship of D. vinosa with the Dictyosporiaceae (Gareth Jones et al. 2008, Crous et al. 2014, Tanaka et al. 2015, Boonmee et al. 2016, Liu et al. 2017. The strains CBS 118716 and CBS 121797 matched the ex-type strain of D. paravinosa (CBS 141286); while CBS 584.96 nested in a terminal subclade with D. fasciculata and D. paravinosa, but it was placed in a single branch representative of a distinct taxa (Fig. 1). Its genetic difference and the production of conidia with up to 7 septa, a distinct morphological feature with respect to the accepted species of Dendryphiella (Liu et al. 2017, Hyde et al. 2018, justify the proposal of a new species in this genus. The other two isolates that had been received as Dendryphiella did not belong to this genus. The oldest reference strain of D. vinosa (CBS 117.14) corresponded to Drechslera biseptata and the strain previously identified as D. infuscans (CBS 381.81) matched Torula herbarum. The molecular identification of all the isolates included in this study is provided in Table 1. Description. Mycelium superficial and immersed, composed of septate, branched, smooth to verruculose, hyaline to pale brown hyphae of 1-3 μm wide. Conidiophores semi-macronematous to macronematous, mononematous, erect or slightly flexuous, branched or unbranched, up to 11-septate, cylindrical, up to 385 μm long, 3-4 μm wide, brown, usually darker toward the base, smooth to verrucose. Conidiogenous cells terminal and intercalary, mostly cylindrical, 8-38 × 3-4(-5) μm, with 1-4 pores. Ramoconidia 0-1-septate, with up to 3 terminal and lateral pores, pale brown, smooth to verruculose, mostly cylindrical, (11-)15-17(-21) × 3-4 μm. Conidia catenate, with up to 10 conidia in the terminal unbranched part, (0-)1-septate, usually not constricted at the septum, pale brown, verruculose to verrucose, ellipsoidal, doliiform or subcylindrical with more or less rounded ends, 4-15 × 3-5 μm.
Neodendryphilla mali is morphologically very similar to N. michoacanensis since both have conidia and ramoconidia 0-1-septate; however, N. michoacanensis has shorter conidiophores (up to 280 μm long) and terminal conidial branches with fewer conidia (up to 4 per branch), which measure 5-16(-18) × 3-6 μm. In addition, 2-septate conidia can also be present in N. michoacanensis and this species tends to grow faster than N. mali on PDA (34 mm vs 22 mm diam. after 14 d, respectively) and PCA (42 mm vs 23 mm diam. after 14 d, respectively). Neodendryphiella mali also resembles D. infuscans, but the latter exhibits longer conidiophores, up to 500 μm and smooth to minutely verruculose conidia with up to 2 septa (Ellis 1971). However, the protologue of D. infuscans (as Cladosporium infuscans; Thümen 1879), which was based on a specimen collected in Aiken (USA), describes conidia 0-1-septate, smooth-walled and up to 10 μm long. No living culture of the type specimen was preserved for further comparison.
As mentioned before, the strain CBS 139.95 was identified as Di. asperum and found by other authors to be related with dictyosporium-like fungi (Shenoy et al. 2010, Tanaka et al. 2015. However, the protologue of Di. asperum was characterised by single or fasciculate conidiophores, which were up to 250 μm long, bearing terminal or subterminal, short and unbranched chains of conidia with only 1 septum (Pirozynski 1972), morphological features that do not fit with those observed in the abovementioned strain. We therefore concluded that it was a misidentified strain and clearly represents a different species. At any rate, it is of note that the taxonomy of Di. asperum is controversial because of the different interpretation of the morphological features of Pirozynski's specimen (DAOM 133941c isotype). Holubová-Jechová (1982) described conidiogenous cells showing inconspicuous denticles or conidiogenous scars instead of the typical pores in conidiogenous cells of Diplococcium and suggested excluding this species from the genus. On the other hand, Goh and Hyde (1998) re-examined the isotype of Di. asperum and observed the typical pores of tretic conidiogenesis, considering it an acceptable species for Diplococcium. However, since only herbarium material is preserved for comparison (Pirozynski 1972), its phylogeny remains uncertain.
Distribution. México. Notes. Neodendryphiella michoacanensis morphologically resembles N. mali, in its conidiogenous apparatus with 0-1-septate ramoconidia, but the latter differs by having longer conidiophores (up to 385 μm), terminal conidial chains with up to 10 conidia and its conidia are 0-1-septate and smaller (4-15 × 3-5 μm). Neodendryphiella michoacanensis also resembles D. uniseptata in their conidial morphology, but ramoconidia of the latter species are often aseptate and can be up to 30 μm long (Matsushima 1971). Dendryphiella uniseptata is only known from the type material, which was collected in Honiara (Japan) and no ex-type culture was preserved. This species was considered a synonym of D. infuscans by Matsushima (1975) but not accepted by Liu et al. (2017).

Discussion
The present study proposes the genus Neodendryphiella based on the analysis of the ITS and LSU sequences, which represented an undescribed monophyletic lineage related but phylogenetically distant from the morphologically similar genus Dendryphiella.
Both genera belong to the Dictyosporiaceae (Dothideomycetes) and share similar conidiophore morphology with polytretic conidiogenous cells forming usually septate conidia arranged in acropetal branched chains. Dendryphiella can be differentiated by the presence of nodulose conidiophores and conidiogeneous cells with pores surrounded by a thickened and darkened wall, which are absent in Neodrendryphiella. Other genera of the Dothideomycetes, although accommodated in different orders or families with a similar conidiogenous apparatus are Dendryphion (Toluraceae, Pleosporales) (Crous et al. 2014, Crous et al. 2015, Dendryphiopsis (Kirschsteiniotheliaceae, Kirschsteiniotheliales) (Su et al. 2016, Hernández-Restrepo et al. 2017 and Paradendryphiella (Pleosporaceae, Pleosporales) (Woudenberg et al. 2013). However, the genus Diplococcium in Leotiomycetes also shows similar asexual propagules (Shenoy et al. 2010, Hernández-Restrepo et al. 2017, which complicates the classification of these fungi based exclusively on morphological features. Our phylogenetic study not only allowed us to distinguish very similar isolates in three distinct species, N. mali, N. michoacanensis and N. tarraconensis, but also helped us to correctly identify some strains that had previously been attributed to Dendryphiella (Table 1). In addition, it is of note that, considering the species accepted in Dendryphiella (Liu et al. 2017, Hyde et al. 2018, this genus seems to be morphologically heterogeneous and probably polyphyletic. It includes species with apparently polyblastic denticulate conidiogenous cells, such as D. eucalypti (Matsushima, 1983) or D. uniseptata (Matsushima, 1971), rather than polytretic conidiogenous cells typical of Dendryphiella (Rao and Narania 1974, Crous et al. 2014 or species that produce solitary conidia, such as D. cruzalmensis (Batista, 1946) or D. lycopersicifolia (Batista & Peres, 1961). In this scenario, therefore, Dendryphiella requires a further taxonomic re-evaluation. However, taking into account that only herbarium material is available for the type D. vinosa (preserved in the Kew herbarium, as Helminthosporium vinosum) there is a need to re-collect this species from the type locality (Cuba) for epitypification and giving nomenclature stability to the genus.