Morpho-molecular characterization of Discosia ravennica sp. nov. and a new host record for Sporocadus rosigena

Abstract Collections of fungal samples from two dead leaf specimens from Italy were subjected to morphological examination and phylogenetic analyses. Two coelomycetous taxa belonging to two different genera in Xylariomycetidae, Sordariomycetes, namely Discosia and Sporocadus, were identified. The Discosia taxon is revealed as a new species and is herein introduced as Discosia ravennicasp. nov. while the Sporocadus taxon is identified as Sporocadus rosigena. Multi-locus phylogeny based on DNA sequence data of the large subunit (LSU) and internal transcribed spacer (ITS) of nuclear ribosomal genes, β-tubulin (β-tub) and RNA polymerase II second largest subunit (rpb2) showed that D. ravennica is related to D. neofraxinea but it forms an independent lineage that supports its new species status. The new taxon also differs from other Discosia species by its unilocular to bilocular, superficial and applanate conidiomata with basal stroma composed of cells of textura angularis, elongate-ampulliform conidiogenous cells and conidia smaller in size. Sporocadus rosigena is here reported as a new host record from Quercus ilex from Italy. Descriptions, illustrations and molecular data for both species are provided in this paper.

After Libert (1837) established Discosia, it was re-studied by Subramanian and Reddy (1974) who designated D. strobilina Lib. ex Sacc. as lectotype for the genus (Nag Raj 1993;Tanaka et al. 2011). Later, when Sphaeria artocreas Tode was transferred to the genus and combined under D. artocreas (Tode) Fr., the latter was chosen as lectotype of the genus (Fries 1849;Vanev 1991). Morgan-Jones (1964) investigated both D. artocreas [same material examined by Fries (1849)] and D. strobilina and reported them as two different species. Subramanian and Reddy (1974) did not examine the type of D. artocreas, but the features of D. strobilina they observed did not match the same reported by Morgan-Jones (1964). The status of D. artocreas as type species of Discosia, therefore, has not been confirmed (Sutton 1980). Nevertheless, it is currently accepted as the type species of the genus (Crous et al. 2013; Index Fungorum, http:// www.indexfungorum.org/Names/Names.asp). Recently, an epitype for D. artocreas was designated .
Delineation of Discosia taxa was earlier, primarily focused on morphological characteristics such as septation of the conidia, varying proportional lengths of the conidial cells and the conidium size (Subramanian and Reddy 1974;Sutton 1980;Vanev 1991Vanev , 1992Vanev , 1996Nag Raj 1993). However, these similar morphological characters have been found to be overlapping for most Discosia species (Sutton 1977(Sutton , 1980Nag Raj 1993;Jeewon et al. 2002;Barber et al. 2011;Tanaka et al. 2011). Species of Discosia were earlier also divided into four sections based on the size, septation and pigmentation of the conidia (Subramanian and Reddy 1974). Later, six sections for the species were proposed based on the same conidial morphology (Vanev 1991). Acquisition of DNA sequence data for Discosia species followed by phylogenetic analyses have, however, shown that the concept of subdivision based on morphology alone has been inaccurate and that proper delineation of species must rely on both morphology and molecular phylogeny (Tanaka et al. 2011).
Sporocadus is a recently resurrected genus, characterized by integrated or discrete conidiogenous cells and generally 3-septate, ellipsoid, cylindrical or obovoid conidia which lack appendages . The genus was originally introduced to accommodate four species, including S. herbarum Corda, S. georginae Corda, S. lichenicola Corda and S. maculans Corda (Corda 1839). No type species for the genus was designated when these species were introduced. However, S. lichenicola was chosen as the lectotype by Hughes (1958). Although Wijayawardene et al. (2016) followed the synonymy of Sporocadus under Seimatosporium by Sutton (1975), Brockman (1976 and Nag Raj (1993) did not accept this. Recently, multi-loci phylogenetic analyses showed that Sporocadus and Seimatosporium are two separate genera, with the former genus usually accommodating taxa without appendages and epitypified by S. lichenicola .
Documenting fungal species, whether they are novel species or new records, is an important contribution to diversity, taxonomy and plant pathology. It is also imperative that these fungal taxa are studied as a number of them are recognized to be potential emerging plant pathogens and they can impact on disease management strategies (Dugan et al. 2009;Giraud et al. 2010;Ghelardini et al. 2016;Rodeva et al. 2016;Jayasiri et al. 2019;Jayawardena et al. 2020). The aim of this paper is to introduce a new Discosia species collected from Italy based on morphology supported by phylogenetic analyses of combined LSU, ITS, β-tub and rpb2 sequence data. In addition, we report a new host record for a sporocadus-like taxon, identified as Sporocadus rosigena, isolated from Quercus ilex (Fagaceae) in Italy.

Sample collection and isolation
Samples of plant materials bearing discosia-like and sporocadus-like fungi were collected from dead land leaves of Pyrus sp. and Quercus ilex in the provinces of Ravenna, Oriolo dei Fichi-Faenza and Forlì-Cesena, Fiumana di Predappio, Italy, respectively. They were brought to the laboratory in paper bags and labelled initially as IT 3632 and IT 3569. The specimens were then examined using a dissecting microscope (Motic SMZ-168).
Single-spore isolation was carried out as described in Senanayake et al. (2020). Conidia of the sporocadus-like taxon successfully germinated and were transferred aseptically to malt extract agar (MEA) plates. The cultures were incubated at 18 °C for 2-3 weeks with frequent observations to assess the colony color and other characters.

Morphological studies
Free-hand sections of conidiomata of the Discosia taxon were prepared to examine their morphological characters. The following structures were observed and measured: height, diameter, and shape of conidiomata, conidiomatal wall cell structure, shape and dimen-sions of conidiophores and conidiogenous cells, length and width of conidia. Morphology of the representatives of the Sporocadus species was obtained from the culture and the morphological characters examined included conidiomata, conidiophores, conidiogenous cells and conidia. All the fungal characters were examined with a fluorescence microscope (Nikon Eclipse E600) and digital images were captured with a Nikon DS-U2 and Cannon 750D camera. All measurements were made using the Tarosoft (R) Image Frame Work software v.0.9.0.7. Images used for photo plates were processed with Adobe Photoshop CS6 v. 12.0 (Adobe Systems, USA).

Material deposition
The holotype of the newly described taxon herein was deposited in the Mae Fah Luang University Herbarium (MFLU), Chiang Rai, Thailand while the isotype at the Cryptogamic Herbarium, Kunming Institute of Botany Academia Sinica (HKAS), Chinese Academy of Sciences, Kunming, China. Herbarium specimen for S. rosigena was also deposited in MFLU while its living culture in Mae Fah Luang University Culture Collection (MFLUCC). Facesoffungi and MycoBank numbers are provided as described in Jayasiri et al. (2015) and MycoBank (http://www.MycoBank.org) respectively. Species concepts are discussed following Jeewon and Hyde (2016).

DNA extraction, PCR amplification and sequencing
Fresh mycelium from the culture of S. rosigena (MFLUCC 18-0387) scraped from the margin of colonies on MEA plates (incubated at room temperature for 4 weeks), and conidiomata of the new taxon (MFLU 18-0131) from natural substrate were used for DNA extraction. Around 20 conidiomata of the new taxon (MFLU 18-0131) were carefully picked from the sterilized material using a fine sterile needle, observed through a stereomicroscope and collected in a 1.5 ml micro-centrifuge tube for subsequent DNA extraction. Genomic DNA was extracted using Forensic DNA Kit (D3591-01, OME-GA bio-tek), following the manufacturer's instructions. The loci LSU, ITS, β-tub and rpb2 were amplified using primers LR0R/LR5 (Vilgalys and Hester 1990;Rehner and Samuels 1994), ITS5/ITS4 (White et al. 1990;Ward and Adams 1998), BT-2a/BT-2b (Glass and Donaldson 1995) and fRPB2-5F/fRPB2-7cR (Liu et al. 1999;Sung et al. 2007) respectively. Polymerase Chain Reactions (PCR) were conducted in an Applied Biosystems C1000 Touch TM Thermal Cycler with the following PCR conditions for LSU, ITS, β-tub and rpb2 regions: initial denaturation at 95 °C for 3 min followed by 34 cycles of denaturation at 95 °C for 30 s and 30 s of annealing and elongation at 72 °C for 1 min, and a final extension at 72 °C for 10 min. The annealing temperatures were 52 °C for LSU and 58 °C for ITS, β-tub and rpb2. The PCR reaction mixture, 25 µL in final volume, was composed of 0.3 µL of TaKaRa Ex-Taq DNA polymerase (TaKaRa, China), 2.5 µL of 10x Ex-Taq buffer (TaKaRa, China), 3.0 µL (2.5 µM) of dNTPs (TaKaRa, China), 1 µL of genomic DNA, 1 µL (0.4 µM) of each primer, and 16.2 µL of double-distilled H 2 O. Sequencing of PCR products was carried out with the same primers as mentioned above at the Beijing Biomed Gene Technology Co., Ltd, and Sangon Biotech, Shanghai China. The newly generated sequences were deposited in GenBank (Table 1).

Phylogenetic analyses
Newly generated sequences from LSU, ITS, β-tub and rpb2 during this study (Table 1) were analyzed with other sequences obtained from GenBank along with recently published relevant phylogenies (Wanasinghe et al. 2018;Liu et al. 2019). Sequences for each locus (LSU, ITS, β-tub and rpb2) were aligned using MAFFT V.7.036 (http:// mafft.cbrc.jp/alignment/server/; Katoh et al. 2019), with L-INS-i Iterative refinement methods and manually improved when necessary in BioEdit v. 7.0 (Hall 2004). Phylogenetic analyses of the aligned data were based on maximum likelihood (ML) and Bayesian inference (BI) analyses with details as outlined by Tang et al. (2007Tang et al. ( , 2009. RAxML-HPC2 on XSEDE (v. 8.2.8) (Stamatakis et al. 2008;Stamatakis 2014) in the CIPRES Science Gateway platform (Miller et al. 2010) was used to generate the ML trees. Optimal ML tree search was conducted with 1000 separate runs, using the default algorithm of the program from a random starting tree for each run. The ultimate tree was selected among suboptimal trees from each run by comparing likelihood scores under the GTRGAMMA substitution model.
Bayesian analysis was executed in MrBayes v. 3. 1. 2 (Huelsenbeck and Ronquist 2001) through Markov Chain Monte Carlo (MCMC) sampling to calculate the posterior probabilities (PP) (Rannala and Yang 1996;Zhaxybayeva and Gogarten 2002). Partitioning of data was initially done by locus and then the parameters of the nucleotide substitution models for every partition were selected independently using MrModeltest v. 2.3 (Nylander 2004). Six Markov chains were run in parallel for 5M generations with trees being sampled every 1000 th generation. The distribution of log-likelihood scores was examined to determine the stationary phase for each search and to decide whether additional runs were required to reach convergence, using the program Tracer 1.5 (Rambaut and Drummond 2007). Convergence was declared when the average standard deviation of split frequencies at the end of the total MCMC generations was at 0.01. First 20% of generated trees was discarded as burn-in and the remaining 80% was used to calculate PP of the majority rule consensus tree

Phylogenetic analyses
The combined gene dataset (LSU, ITS, β-tub and rpb2) used to generate ML tree in Fig. 1 comprised 51 taxa including the newly generated sequences. Pestalotiopsis hollandica (CBS  Discosia taxa were divided into two separate clades (A and B). Clade A, consisting of 3 strains of Discosia, grouped with and was sister to Sporocadus with strong statistical support (100% ML, 1.00 PP). Clade B, comprising 21 strains of Discosia, was basal to both Sporocadus and clade A with strong statistical support (100% ML, 1.00 PP). Our strain MFLU 18-0131 was positioned in clade A, basal to both strains of D. neofraxinea (MFLU 15-0375 and MFLUCC 12-0670 = NTIT469), forming an independent lineage with good statistical support (96% ML/ 1.00 PP).
All the Sporocadus species formed a monophyletic clade with strong statistical support (100% ML, 1.00 PP). The strain MFLUCC 18-0387 from this study clustered with the other existing S. rosigena strains with a bootstrap support of 91% ML and 0.98 PP (Fig. 1).

Discussion
Discosia ravennica sp. nov. forms an independent lineage, basal to the two strains of D. neofraxinea (96% ML/ 1.00 PP) (Fig. 1). It is different from D. neofraxinea in its unilocular to bilocular, applanate conidiomata along with elongate-ampulliform conidiogenous cells and conidia smaller in size (Table 2). With regard to DNA sequence data comparison, D. ravennica differs from both strains of D. neofraxinea  in having 14 out of 531 (2.6 %) and 8 out of 512 (1.6%) different base pairs (bp) in the ITS alignments respectively. Moreover, 13 bp out of 229 (5.7%) and 82 bp out of 832 (9.9%) differences in the ß-tub and rpb2 alignments respectively can be observed between D. ravennica and D. neofraxinea (MFLUCC 12-0670 = NTIT469). Sequence data of ß-tub and rpb2 are not available for the strain of D. neofraxinea  in GenBank and hence could not be compared. Similarly, no molecular data for D. fraxinea are accessible in GenBank, following which the new species, D. ravennica, has been delineated based on morphology ( Table 2). The 5.7% and 9.9% differences in nucleotides in ß-tub and rpb2 respectively may acceptably support the establishment of a new species . Following this assumption along with the above-mentioned morphological differences and high statistical support, D. ravennica is herein established as a new species.
A peculiar finding from our DNA sequence analyses is the placement of D. neofraxinea and D. ravennica. Both of them constitute a strongly supported independent clade (clade A) basal to species of Sporocadus. One might argue that given their distinct phylogenetic nature, a new genus accommodating these two species might be a possibility. However, in this particular scenario, we would rather take a more conservative and lumping taxonomic approach and maintain the latter two species in Discosia. The reasons we would advocate are that there is a lot of morphological resemblance between members of clades A and B. For instance, when we compare D. neofraxinea and D. ravennica (clade A) with the type species, D. artocreas (clade B), they all have stromatic conidiomata, conidiophores which arise from the upper cell layer of the basal stroma, and hyaline to sub-hyaline, usually 3-septate conidia bearing two appendages (Nag Raj 1993;Senanayake et al. 2015;Liu et al. 2019). The main difference is that D. neofraxinea and D. ravennica have the third cell of their conidia from the base longer than the second cell while D. artocreas has the second cell of its conidia from the base longer than the third cell (Nag Raj 1993) or both median cells of almost equal length ). However, this distinctive characteristic is not sufficient enough for the establishment of a new genus. It might be that the genus is paraphyletic, but until more species are recovered and analyzed to provide further taxonomic insights, we refrain from making any taxonomic amendments. It might also be possible that there is a need to establish species complexes given the wide intraspecies variation as we have seen in other genera such as Phyllosticta ).
The second recovered species from this study, Sporocadus rosigena, clusters with other S. rosigena strains in a well-supported clade (91% ML / 0.98 PP) in our 4-gene phylogeny (Fig. 1). The latter shows similar topology to the 5-gene phylogeny reported by Liu et al. (2019). Sporocadus rosigena has earlier been reported as saprobic or endophytic on species of Rosa, Rubus, Pyrus (Rosaceae), Rhododendron (Ericaceae) and Vitis (Vitaceae) (Wanasinghe et al. 2018;Liu et al. 2019). In this study, the species was found from Quercus ilex (Fagaceae) and is therefore introduced as a new host record. Different fungi have equally been reported from Quercus ilex in Italy; for instance, the genera Alternaria (Lunghini et al. 2013), Beltrania (Pirozynski 1963), Endothia (Spaulding 1961), Monochaetia (Nag Raj 1993), Neognomoniopsis , Pestalotia (Nag Raj 1993), Xylaria and Zygosporium (Lunghini et al. 2013), indicating a broad diversity of fungi on the same host. All Sporocadus species in their asexual stage possess 3-septate, obovoid, fusoid to cylindrical conidia, which do not have any appendage. The only exceptions are S. trimorphus and S. rosarum, which are known to produce conidia both with and without appendages .
Fungal diversity and classification are always ever-changing and require an ongoing assessment (Hyde and Soytong 2008;Jeewon et al. 2017). This becomes especially essential in cases where taxa are described from genera which usually accommodate pathogens. Discosia, for instance, is known to comprise the plant pathogen D. yakushimensis which causes leaf spots on plants such as Symplocos prunifolia (Tanaka et al. 2011). Identifying novel species in a genus may also potentially imply the discovery of emerging pathogens which can cause damage to crops of economic importance (Jayawardena et al. 2019a(Jayawardena et al. , 2019b. Evolutionary relationships and ecological roles of fungi have been reported to be intricately linked to the emergence of new species (Zhang et al. 2008;Hyde et al. 2020). However, such phenomena also extend to the recognition of existing species from new hosts, as is the case for S. rosigena in the present study. Documenting records from new hosts has become useful repertoires for mycologists who aim to understand evolution of fungi, host jumping, expanding host diversity and adaptations to different environmental conditions ). These are equally important for proper quarantine measures, whereby potential pathogens or species known to have a wide host diversity are to be closely monitored with a view to avoid unintentional disturbance to a specific environment (Cai et al. 2011).