Diatrypella macrospora, a new host and geographical record from Forlì-Cesena, Italy

Diatrypaceae consists of 22 genera and approximately 1035 species (Wijayawardene et al. 2020, Hyde et al. 2020b, Konta et al. 2020; Dissanayake et al. 2021). The taxa are generally characterized by perithecial ascomata immersed in a stroma, asci with long pedicels, allantoid ascospores and coelomycetous or hyphomycetous asexual morph (Shang et al. 2017, Hyde et al. 2020b). The asexual morphs of Diatrypaceae are morphologically similar and thus cannot be used for taxonomic differentiation (Acero et al. 2004). Genera in Diatrypaceae are delineated mainly based on stromatal morphology e.g., pustulate, valsoid, eutypoid, well-developed or poorly-developed (Vasilyeva & Stephenson 2005, Senwanna et al. 2017). Early treatments of Diatrypaceae were based on morphology, while recent studies have included phylogenetic data (Mehrabi et al. 2016, Dayarathne et al. 2020, Konta et al. 2020).

Diatrypella was introduced as a segregate of Diatrype with polysporous asci (Croxall 1950). Characteristics of Diatrypella include conical or truncate, discoid or cushion-like stromata delimited by a black zone on the host tissues, perithecial ascomata, umbilicate or sulcate ostioles, and numerous ovoid to allantoid ascospores. The asexual morph is described as libertella-like (Senwanna et al. 2017, Shang et al. 2017, Hyde et al. 2020a). Diatrypella includes approximately 115 species (Wijayawardene et al. 2020). In the last decade, several new species have been introduced using evidence from morphology coupled with ITS and β-tubulin sequence analyses. These include D. iranensis, D. macrospora, D. tectonae and D. yunnanensis (Mehrabi et al. 2015, 2016, Shang et al. 2017, Hyde et al. 2020a). Molecular studies, have however, revealed that Diatrypella is polyphyletic (Senwanna et al. 2017). Acero et al. (2004) suggested that polysporous asci is a characteristic that evolved independently multiple times during evolution of Diatrypaceae (Acero et al. 2004, Senwanna et al. 2017). This suggests that current features used in the taxonomy of Diatrypaceae may not reflect the evolutionary history (Mehrabi et al. 2016), hence highlighting the importance of using molecular data for delimitation of genera within the family.

Diatrypella taxa have broad geographic distribution as saprobes (e.g. D. heveae, D. quercina, D. yunnanensis), endophytes (e.g. D. favacea, D. frostii) or occasionally as suspected pathogens (e.g. D. japonica, D. vulgaris) mainly on woody angiosperms (Pitt et al. 2013, de Almeida et al. 2016, Senwanna et al. 2017, Rudolph et al. 2018, Rashmi et al. 2019, Hyde et al. 2020a, Li et al. 2020). Some species have a broad host range, others such as D. vitis (on grapevines) have been reported only on one host genus (Acero et al. 2004, Farr & Rossman 2020).

In the present study we report a new host and geographical record of Diatrypella macrospora on Quercus cerris from Italy based on combined ITS and β-tubulin sequence analysis. Our comprehensive analysis also highlights the need for extensive revisions within Diatrypaceae.

Sample collection, morphological studies and isolation

A dead land branch of Quercus cerris was collected in the Province of Forlì-Cesena, Italy in July 2019. Samples were brought to the laboratory and kept in paper envelopes. Macroscopic characters (enlarged host surface and ascomata) were examined using a Motic SMZ 168 series stereomicroscope. Microscopic features were observed using a Nikon DS-Ri2 digital camera fitted to a Nikon Eclipse 80i compound microscope. Thin cross sections of stromata were prepared manually and mounted in water on glass microscopic slides. Photomicrographs were prepared with Adobe Photoshop CC v. 20.0.5 (Adobe Systems, USA) and all character measurements were made with Tarosoft Image FrameWork v. 0.9.0.7.

Single ascospore isolation was carried out as described in Senanayake et al. (2020). After germination, ascospores were transferred aseptically to malt extract agar (MEA) medium and incubated at room temperature. A pure culture was obtained and its characteristics were also observed. Dried specimens and living cultures were deposited in the Mae Fah Luang University herbarium, Thailand (MFLU) and the Mae Fah Luang University Culture Collection (MFLUCC) respectively.

DNA extraction, PCR amplification and phylogenetic analyses

The methods used in DNA extraction, PCR amplification, sequence analyses and genetic analyses are as outlined in Dissanayake et al. (2020) with the following or modifications. The internal transcribed spacer (ITS) and β-tubulin loci were amplified by polymerase chain reaction. The primers used for amplification were: ITS5 and ITS4 (White et al. 1990) for ITS and T1 and Bt2b (Glass & Donaldson 1995, O 'Donnell & Cigelnik 1997) for β-tubulin. The PCR conditions for both ITS and β-tubulin were set as follows: initial denaturation of 94℃ for 3 mins, followed by 35 cycles of denaturation at 94℃ for 30s, annealing at 55℃ for 50s, elongation at 72℃ for 1 min and final extension at 72℃ for 10 mins. DNA sequencing was performed at BGI Shenzhen, China. The closest related strains and sequences spanning the diversity of Diatrypaceae were compiled following recent publications (Dayarathne et al. 2020, Konta et al. 2020) (Table 1). For the Maximum likelihood (ML) analysis, the optimal ML tree was obtained using 1000 separate runs and GTR+GAMMA was used as the model for nucleotide substitution. For Bayesian inference analyses (BI), two parallel runs, each consisting of four simultaneous Markov chains were executed for 4, 000, 000 generations, sampling one tree every 1000^th generation. 25% of the trees were discarded as the burn-in phase in the analysis. The remaining trees were used to calculate posterior probabilities in the majority rule consensus tree. Convergence was determined when the average standard deviation of split frequencies reached 0.01.

Phylogenetic analyses

Phylogenetic analyses of a combined ITS and β-tubulin sequence dataset comprised 131 ingroup taxa and two outgroup taxa, namely Kretzschmaria deusta and Xylaria hypoxylon (Xylariaceae). The combined matrix contained 1240 nucleotide sites (ITS: 1-515; β-tubulin: 516-1240). The ML analysis yielded a best-scoring tree with a final ML optimization likelihood value of -18323.132499. The matrix had 887 distinct alignment patterns, with 38.83% of gaps and completely undetermined characters. The ML and BI analyses yielded trees with similar topologies. The clades recovered in the phylogenetic tree are similar in topologies to previous studies (Acero et al. 2004, Dayarathne et al. 2020, Konta et al. 2020). Diatrypella species clustered in five distinct clades as A1, A2, B, C1 and D1 (Fig. 1). Our newly sequenced strain grouped separately from the type species Diatrypella verruciformis in clade D1. It clustered with the type of D. macrospora (IRAN 2344C) and with Diatrype sp. (H2/4b) with statistical support, MLBS 84%, BYPP 1.00 (Fig. 1).

Figure 1.  Maximum likelihood phylogenetic tree generated from combined ITS and β-tubulin sequence data of 131 Diatrypaceae taxa and 1240 sites. For each node maximum likelihood bootstrap support values (MLBS) are given first, followed by Bayesian posterior probabilities (BYPP). ML bootstrap support values ≥ 75% and BYPP ≥ 0.95 are indicated at the nodes. Lower values are indicated by a dash (-). Nodes that were not recovered are indicated by asterisks (*). The new isolate is in red bold font. Ex-type sequences are in black bold font. The tree is rooted to Kretzschmaria deusta (CBS 826.72) and Xylaria hypoxylon (CBS 122.620).

Taxonomy

Diatrypella macrospora Mehrabi, Hemmati, Vasilyeva & Trouillas, Phytotaxa 252(1): 47 (2016)

      Fig. 2

Figure 2.  Diatrypella macrospora (MFLU 19-2401, new record). a Stromata erupting from bark. b Longitudinal section through stroma. c Section through a perithecium. d Peridium. e Paraphyses. f-i Asci. j-l Ascospores. m Germinated ascospore. n, o Top and bottom view of culture. Scale bars: a = 2 mm, b = 500 μm, c = 100 μm, d = 50 μm, e-j = 20 μm, k-m = 5 μm.

Index Fungorum number: IF813001; Facesoffungi number: FoF01891

Saprobic on dead land branch of Quercus cerris. Sexual morph: Stromata 1.2-2.8 mm (x = 1.8 mm, n = 10) wide, discoid to irregular, scattered or aggregated, immersed to erumpent arising from cracks in the bark, black, separated from host tissue by a black zone, perithecia arranged in groups of 3- 8, entostroma well-developed, white to yellow to brown. Perithecia 260-490 µm (x = 375 µm, n = 10) in diameter, globoid to flask-shaped, sometimes deformed by compression. Perithecial neck about 110- 350 μm (x = 240 µm, n = 10), ostiolar canals sulcate, compressed, converge together at the apex, black, ostiole brown, opening separately through host bark, periphysate. Peridium composed of light brown to brown, somewhat flattened cells of textura angularis, becoming hyaline towards the inner region. Paraphyses elongate, filiform, aseptate, unbranched. Asci spore-bearing part 72-120 µm × 8-14.5 µm (x = 94 × 11 µm, n = 20), basal part, filiform, 24-53 µm (x = 40 µm, n = 20), polysporous, unitunicate, elongate, cylindrical to clavate, obtuse apex, with a J-apical ring. Ascospores 7.5-12 μm × 2-3.1 µm (x = 10 × 2.5 µm, n = 50), allantoid, subhyaline, yellowish in mass, aseptate, usually bi-guttulate, thin, smooth-walled. Asexual morph: For morphological description see Mehrabi et al. (2016).

Culture characteristics - Colonies on MEA reaching 75 mm diam. after 2 weeks at room temperature. Colonies circular, slightly dense, flat, with fimbriate margin, white to light brown to black from above, similar colour from below.

Material examined - Italy, Province of Forlì-Cesena, Forlì, Farazzano, on a dead land branch of Quercus cerris (Fagaceae), 22 July 2019, E. Camporesi, IT 4432 (MFLU 19-2401, new record, dried culture), living culture MFLUCC 21-0010.

GenBank accession numbers - ITS: MW647094, β-tubulin: MW677962

Notes - Our new collection of Diatrypella macrospora resembles the holotype (IRAN 2344C) as it has subhyaline and allantoid ascospores, elongate asci, overlap in the number of perithecia as well as sizes of stromata, perithecia and asci (Table 2) (Mehrabi et al. 2016). However, our collection differs from D. macrospora (IRAN 2344C) in having smaller ascospores which are considerably more curved and shorter perithecial necks. Based on curvature and spore size, D. macrospora from our collection shows more resemblance with D. quercina (Croxall 1950). In the combined ITS and β-tubulin phylogeny, our new isolate MFLUCC 21-0010 clusters with D. macrospora (IRAN 2344C) and with Diatrype sp. (H2/4b), with high statistical support (MLBS 84%, BYPP 1.00) (Fig. 1, Clade D, D1). There are 2/510 (0.39%) and 9/349 (2.28%) base pair differences between our isolate and D. macrospora (IRAN 2344C) for ITS and β-tubulin sequences respectively. For the ITS sequence of our isolate and Diatrype sp. (H2/4b) there are 2/510 (0.39%) base pair differences. The two strains of Diatrypella macrospora formed a clade with D. iranensis, D. sp. (C6, ENQ55 and MNQ75B), D. quercina, Diatrype bullata, D. disciformis, D. sp. (H2/5c and H3/2b), and D. spilomea, with 78% MLBS and 0.99 BYPP statistical support (Fig. 1, Clade D). Previously, Diatrypella macrospora had been reported only in Iran on Quercus brantii. Our new isolate represents a new host and geographical record of D. macrospora on Quercus cerris (Fagaceae) in Italy.

In this study, we identify our collection as Diatrypella macrospora because it is morphologically similar to the holotype, has a short phylogenetic distance, and little ITS base pair differences with the type strain. It also occurs on the same host genus as the holotype. Previously, D. macrospora had only been reported on Quercus brantii in Iran (Mehrabi et al. 2016). Our collection occurs on Quercus cerris, commonly known as Turkey oak. The tree is widespread in the Italian Peninsula (Taffetani et al. 2012) and it is similar to other oak trees but its wood is more prone to cracking and splitting (Vidrinskas & Deveikis 2016). Nonetheless, there are some differences between our strain and the type specimen. The observed differences in morphology may be attributed to differences in host and geographical location. Alternatively, the morphological distinctions along with differences in the nucleotide sequence of β-tubulin also indicate the possibility of a new species rather than a new strain of D. macrospora. However, there is only a single collection of D. macrospora (as well as D. iranensis) and it is therefore difficult to determine species boundaries given the lack of data. Ideally, several differences in more than one locus are required to introduce a new species. With the data currently available, we take a conservative approach and list our isolate as a new record of D. macrospora.

Our analysis also provides insights into Diatrypella. The phylogenetic tree (Fig. 1) clearly shows that the genus is not monophyletic, confirming previous studies (Acero et al. 2004). Diatrypella sequences formed five distinct clades in the tree (Clades A1, A2, B, C1, D1). The newly identified strain along with D. macrospora (IRAN 2344C) formed a clade with D. iranensis and D. quercina, both of which are closely associated with oak trees (Croxall 1950, Farr & Rossman 2020). Further collections are required to determine host-specificity of Diatrypella species in Clade D1. In the phylogenetic tree, the above three species group together in the same clade, distant from the type species, D. verruciformis (Fig. 1), a consistent finding across studies (Mehrabi et al. 2015, 2016, 2019, Konta et al. 2020). Diatrypella verruciformis clusters as sister to Diatrype sequences separately from other Diatrypella. Diatrypella iranensis, D. macrospora and D. quercina could be considered as part of Diatrype, since they form a strongly supported clade with the type D. disciformis. This placement is congruent with previous studies (Acero et al. 2004, Mehrabi et al. 2019, Konta et al. 2020). In these investigations, the three species are treated as part of Diatrype. However, the D. disciformis sequence data used in these studies are not from ex-type strains, thus they might be misidentified. Senanayake et al. (2015) proposed a reference specimen for D. disciformis (MFLU 15-0722, MFLUCC 15-0538), for which LSU and ITS sequences are available. However, these data were not included in the present analysis. In fact, only a few taxa in our dataset have available LSU sequences. Moreover, the ITS sequence for D. disciformis (MFLUCC 15-0538) appears to be problematic. When the strain was subjected to BLAST search, the closest match were members of Nectriaceae, a different family. Therefore, the taxonomic placement of the three Diatrypella species cannot be conclusively confirmed based on molecular data. Similarly, D. favacea, D. pulvinata, D. yunnanensis also form a distinct clade (Clade C1). Clade C1 and D1 could actually represent two distinct and novel genera. However, it is difficult to separate them with certainty from Diatrypella, since there is no sequence from the ex-type of D. verruciformis.

Transfer of D. iranensis, D. macrospora and D. quercina to Diatrype would need revision of the generic concepts or a new genus should be erected. Cesati & De Notaris (1863) separated Diatrypella and Diatrype solely based on the number of spores per ascus (Croxall 1950, Acero et al. 2004). In early taxonomic studies, it was recognized that, although convenient, this classification system is likely to be artificial and not a good reflection of evolutionary history of the two genera (Glawe & Rogers 1984, Rappaz 1987). In fact, phylogenetic analyses show that number of spores per ascus is highly variable throughout the evolution of Diatrypaceae and multispored asci appeared several times independently (Dayarathne et al. 2020, Acero et al. 2004). This explains the mixed clades (Clade A, C, D) comprising both Diatrypella and Diatrype species recovered in this study and previous studies (Acero et al. 2004, de Almeida et al. 2016, Dayarathne et al. 2020, Konta et al. 2020). Apart from the number of spores per ascus other unique characteristics should be used to delineate between genera (Carmarán et al. 2006). However, from the data available (Table 2) there is no unique common characteristic separating Diatrypella species in Clade C1 from those in Clade A.

As mentioned by Acero et al. (2004), several early taxonomists have highlighted the morphological distinction between D. quercina and other Diatrypella species (Acero et al. 2004). Specifically, although the species has polysporous asci it also has a well-developed ectostroma, which is unusual of Diatrypella, but common in Diatrype (Wehmeyer 1926). Diatrypella macrospora (MFLUCC 21-0010), which groups in the same clade as D. quercina, also displays a higher degree of ectostromatic development. No details were given for the ectostroma of D. iranensis and D. macrospora (IRAN 2344C). Thus, even though ectostroma could be used as a taxonomic character this cannot be conclusively determined due to lack of information. This could be because the ectostroma is normally not used as taxonomic characteristic for herbarium as it is present only on immature specimens (Rappaz 1987). Also, some taxonomists argue that microscopic characteristics are more appropriate for delineating diatrypaceous taxa as compared to the more variable macroscopic characters (de Almeida et al. 2016).

Another difference observed by Croxall (1950) in D. quercina is its strongly curved ascospores. However, curvature in ascospore appears only as a difference between species rather than between genera (Table 2). Molecular analysis confirmed the distinction of D. quercina from other Diatrypella species (Acero et al. 2004). The morphological characteristics related to these molecular differences are not clear. A summary of recent morphological descriptions of species in Clade D is also provided by Thiyagaraja et al. (2019). The data presented is coherent with that reviewed in our study except for the colour of the entostroma which seems to have been mixed with the colour of stromata. Still, the additional details do not provide insights into a more reliable classification. From the morphological descriptions available, D. iranensis, D. macrospora and D. quercina show no clear common feature that separate them from other Diatrypella species or explain their phylogenetic position among Diatrype species (Table 2). Collectively, these results suggest that the currently used morphological features have a high degree of overlap and their taxonomic value might need to be reconsidered.

Missing data, inaccessibility of type specimens and inadequate original descriptions make resolving the taxonomic problems mentioned above challenging. The genus concept of Diatrypella and Diatrype as well as the classification of Diatrypaceae as a whole should be reviewed. Sequencing of the reference specimen for Diatrype should be checked and an epitype should be provided for Diatrypella (de Almeida et al. 2016). More collections of diatrypaceous taxa and a combination of reliable molecular and morphological information would greatly aid resolving the taxonomy of Diatrypaceae.