Morphological and phylogenetic analyses reveal a new genus and two new species of Tubakiaceae from China

Abstract Species of Tubakiaceae have often been reported as plant pathogens or endophytes, commonly isolated from a wide range of plant hosts. The isolated fungi were studied through a complete examination, based on multilocus phylogenies from combined datasets of ITS/LSU/rpb2 and ITS/tef1/tub2, in conjunction with morphological characteristics. Five strains isolated from Lithocarpusfohaiensis and Quercuspalustris in China represented a new genus of Tubakiaceae, Obovoideisporodochium and three species, viz. Obovoideisporodochiumlithocarpi sp. nov., Tubakialushanensis sp. nov. and T.dryinoides.


Introduction
Diaporthales represents an important order in Sordariomycetes containing taxa that are mainly isolated as endophytes, saprobes or plant pathogens on various hosts (Fan et al. 2018). Tubakiaceae is a family in Diaporthales, which has been studied in recent years by Braun et al. (2018) by incorporating morphological and molecular data with appropriate genes to resolve species limitations in the family. Tubakiaceae currently comprises eight genera including Apiognomonioides U. Braun et al., Involutiscutellula U. Braun & C. Nakash., Oblongisporothyrium U. Braun & C. Nakash., Paratubakia U. Braun & C. Nakash., Racheliella Crous & U. Braun, Saprothyrium U. Braun et al., Sphaerosporithyrium U. Braun et al. and Tubakia B. Sutton (Braun et al. 2018).
Tubakia, the type genus of Tubakiaceae, was introduced by Sutton (1973). Species of Tubakia are endophytes in leaves and twigs of many tree species, but can also cause conspicuous leaf symptoms as plant pathogens (Harrington et al. 2012;Harrington & McNew 2016Braun et al. 2018). The genus is characterised by unique pycnothyria, consisting of pigmented, radiating, seta-like cells (scutellum) on top of a columella, with small phialides on the underside of the scutellum producing ellipsoid, hyaline to brown conidia that are forced out from under the pycnothyrium for rain dispersal (Harrington & McNew 2018). Some species produce a second type of much smaller conidia (microconidia), either in "normal" pycnothyria or in separate, mostly smaller pycnothyria (Braun et al. 2018). Saccardo (1913) introduced the genus Actinopelte for A. japonica, a scutellate fungus found in Japan on Castanea crenata (= C. pubinervis). Saccardo (1913) confused the large conidia of this species with asci, which was clarified and corrected by Theissen (1913) who provided a detailed discussion, description and illustration (Theissen 1913) of A. japonica. Von Höhnel (1925 revisited Actinopelte, added a new species, A. americana and introduced the new combination A. dryina, based on Leptothyrium dryinum. Yokoyama & Tubaki (1971) discussed the history of this genus in detail, published results of comprehensive examinations of Japanese collections in vivo and in vitro and described A. castanopsidis, A. rubra and A. subglobosa, based on Japanese collections. Since Saccardo's Actinopelte turned out to be illegitimate (later homonym of Actinopelte Stitzenb. 1861), Sutton (1973) introduced the replacement name Tubakia and reallocated all species recognised and treated in Yokoyama & Tubaki (1971) to this genus. Twenty-one additional Tubakia species have subsequently been described including fifteen new Tubakia species and six combinations in Tubakia species (Yun & Rossman 2011;Harrington et al. 2012;Braun et al. 2014;Harrington & McNew 2018;Senanayake et al. 2017;Braun et al. 2018;Yun & Kim 2020).
During field trips to collect plant pathogens causing leaf spots symptoms in China, several specimens associated with typical diaporthalean symptoms were collected from various tree hosts, i.e. Betula dahurica (Betulaceae), Juglans regia (Juglandaceae), Prunus davidiana (Rosaceae), Lithocarpus fohaiensis, Quercus mongolica and Q. palustris (Fagaceae). Based on morphological analyses as well as phylogenetic data, this study presents a new genus of Tubakiaceae, Obovoideisporodochium and three species, viz. Obovoideisporodochium lithocarpi sp. nov., Tubakia lushanensis sp. nov. and T. dryinoides from diseased leaves of L. fohaiensis or Q. palustris.

Isolation and morphological studies
The samples were collected from the Shandong and Yunnan Provinces, China. The strains were isolated from diseased leaves of Lithocarpus fohaiensis and Quercus palustris using tissue isolation methods. Tissue fragments (5 mm × 5 mm) were taken from the margin of leaf lesions and surface-sterilised by consecutively immersing in 75% ethanol solution for 1 min, 5% sodium hypochlorite solution for 30 s and then rinsing in sterile distilled water for 1 min. The pieces were dried with sterilised paper towels and placed on potato dextrose agar (PDA). All the PDA plates were incubated in a biochemical incubator at 25°C for 2-4 days. The colonies from the periphery were picked out and inoculated on to new PDA plates. Colony photos after 7 days and 15 days were taken by a digital camera (Canon Powershot G7X). Micromorphological characters were observed using an Olympus SZX10 stereomicroscope and Olympus BX53 microscope, all fitted with Olympus DP80 high definition colour digital cameras to photo-document fungal structures. All fungal strains were stored in 10% sterilised glycerine at 4°C for further studies. The holotype specimens are deposited in the Herbarium of Plant Pathology, Shandong Agricultural University (HSAUP). Ex-type cultures are deposited in the Shandong Agricultural University Culture Collection (SAUCC). Taxonomic information of the new taxa was submitted to MycoBank (http://www.mycobank.org).
The PCR was performed using an Eppendorf Master Thermocycler (Hamburg, Germany). Amplification reactions were performed in a 25 μl reaction volume, which contained 12.5 μl Green Taq Mix (Vazyme, Nanjing, China), 1 μl of each forward and reverse primer (10 μM stock) (Biosune, Shanghai, China) and 1 μl template genomic DNA in amplifier, adjusted with distilled deionised water to a total volume of 25 μl. The PCR parameters were as follows: 94°C for 5 min, followed by 35 cycles of denaturation at 94°C for 30 s, annealing at a suitable temperature for 50 s, extension at 72°C for 1 min and a final elongation step at 72°C for 10 min. The annealing temperatures for the genes were 55°C for ITS, 52°C for LSU, 53°C for tub2, 48°C for tef1 and 56°C for rpb2. The PCR products were separated with the 1% agarose gel, with added GelRed and UV light used to visualise the fragments. Sequencing was done bidirectionally, conducted by the Biosune Company Limited (Shanghai, China). Consensus sequences were obtained using MEGA v. 7.0 (Kumar et al. 2016). All sequences generated in this study were deposited in GenBank (Table 1).

Phylogeny
The generated consensus sequences for each gene were subjected to megablast searches to identify closely-related sequences in the NCBI's GenBank nucleotide database (Zhang et al. 2000). For the ITS-LSU-rpb2 and ITS-tef1-tub2 analyses, subsets of sequences from the alignments of Braun et al. (2018) were used as backbones. Newlygenerated sequences in this study were aligned with additional related sequences downloaded from GenBank (Table 1) using MAFFT 7 online service with the Auto strategy (Katoh et al. 2019, http://mafft.cbrc.jp/alignment/server/). To establish the identity of the isolates at species level, phylogenetic analyses were conducted, first individually for each locus and then as combined analyses (ITS-LSU-rpb2 and ITS-tef1-tub2).
Phylogenetic analyses were based on Maximum Likelihood (ML) and Bayesian Inference (BI) for the multilocus analyses. For BI, the best evolutionary model for each partition was determined using MrModelTest v. 2.3 (Nylander 2004) and incorporated into the analyses. ML and BI were run on the CIPRES Science Gateway portal (https://www.phylo.org/) (Miller et al. 2012) using RAxML-HPC2 on XSEDE v. 8.2.12 (Stamatakis 2014) and MrBayes on XSEDE v. 3.2.7a (Huelsenbeck & Ronquist 2001;Ronquist & Huelsenbeck 2003;Ronquist et al. 2012), respectively. For the ML analyses, the default parameters were used and BI was carried out using the rapid bootstrapping algorithm with the automatic halt option. Bayesian analyses included four parallel runs of 5,000,000 generations, with the stop rule option and a sampling frequency of 50 generations. The burn-in fraction was set to 0.25 and posterior probabilities (PP) were determined from the remaining trees. All resulting trees were plotted using FigTree v. 1.4.4 (http://tree.bio.ed.ac.uk/software/figtree) and the layout of the trees was done in Adobe Illustrator CC 2019.

ITS/LSU/rpb2 phylogeny
The alignment contained 37 isolates representing Tubakia and allied taxa and a strain of Greeneria uvicola (FI12007) was used as outgroup. The final alignment contained a total of 2459 characters used for the phylogenetic analyses, including alignment gaps, viz. ITS: 1-676, LSU: 677-1545, rpb2: 1546-2459. Of these characters, 1858 were constant, 115 were variable and parsimony-uninformative and 486 were parsimony- Table 1. Species and GenBank accession numbers of DNA sequences used in this study. New sequences in bold. informative. MrModelTest recommended that the Bayesian analysis should use Dirichlet base frequencies for the ITS, LSU and rpb2. The GTR+I+G model was proposed for ITS, LSU and rpb2. The MCMC analysis of the three concatenated genes, run for 700,000 generations, resulted in 14,001 trees. The initial 3500 trees, representative of the analysis burn-in phase, were discarded, while the remaining trees were used to calculate posterior probabilities in the majority rule consensus trees ( Fig. 1; first value: PP > 0.74 shown). The alignment contained a total of 744 unique site patterns (ITS: 266, LSU: 128, rpb2: 350). The topology of the ML tree confirmed the tree topology obtained from the Bayes analyses and, therefore, only the ML tree is presented (Fig.  1). Bayesian posterior probability (> 0.74) and ML bootstrap support values (> 74%) are shown as first and second position above nodes, respectively. The 37 strains were assigned to 25 species clades, based on the three-gene phylogeny (Fig. 1).

ITS/tef1/tub2 phylogeny
The alignment contained 37 isolates representing Tubakia and allied taxa and a strain of Greeneria uvicola (FI12007) was used as outgroup. The final alignment contained a total of 1939 characters used for the phylogenetic analyses, including alignment gaps, viz. ITS: 1-676, tef1: 677-1358, tub2: 1359-1939. Of these characters, 1077 were constant, 136 were variable and parsimony-uninformative and 726 were parsimony-informative. MrModelTest recommended that the Bayesian analysis should use Dirichlet base frequencies for the ITS, tef1 and tub2 data partitions. The GTR+I+G model was proposed for ITS and HKY+I+G for tef1 and tub2. The MCMC analysis of the three concatenated genes, run for 170,000 generations resulted in 3401 trees. The initial 850 trees, representative of the analysis burn-in phase, were discarded, while the remaining trees were used to calculate posterior probabilities in the majority rule consensus trees ( Fig. 2; first value: PP > 0.74 shown). The alignment contained a total of 997 unique site patterns (ITS: 266, tef1: 416, tub2: 315). The topology of the ML tree confirmed the tree topology obtained from the Bayes analyses and, therefore, only the ML tree is presented (Fig. 2). Bayesian posterior probability (> 0.74) and ML bootstrap support values (> 74%) are shown as first and second position above nodes, respectively. The 37 strains were assigned to 25 species clades, based on the three-gene phylogeny (Fig. 2).
Based on phylogenetic data (Figs. 1 and 2) and morphological analyses, the present study revealed a new genus of Tubakiaceae, Obovoideisporodochium and three species, viz. Obovoideisporodochium lithocarpi sp. nov., Tubakia lushanensis sp. nov. and T. dryinoides.  The BI and ML bootstrap support values above 0.74 and 74% are shown at the first and second position, respectively. The tree is rooted to Greeneria uvicola (culture FI12007) and ex-type cultures are indicated in bold face. Strains from the current study are in red. Some branches were shortened for layout purposes -these are indicated by two diagonal lines with the number of times a branch was shortened indicated next to the lines.

Taxonomy
Etymology. Composed of "obovoideisporo-" (obovoid spores) and "-dochium" (referring to the conidioma, i.e. sporodochium). The BI and ML bootstrap support values above 0.74 and 74% are shown at the first and second position, respectively. The tree is rooted to Greeneria uvicola (culture FI12007) and ex-type cultures are indicated in bold face. Strains from the current study are in red. Some branches were shortened for layout purposes -these are indicated by two diagonal lines with the number of times a branch was shortened indicated next to the lines.

Obovoideisporodochium lithocarpi
Culture characteristics. Cultures incubated on MEA at 25°C in darkness, attaining 52.0-58.0 mm diam. after 14 d (growth rate 3.5-4.0 mm diam./d), greywhite to creamy white with irregular margin, spread like petals from the inside and outside, reverse dark to light brown, distributed in an irregular circle. Conidial formation not observed.
Culture characteristics. Cultures incubated on MEA at 25°C in darkness, attaining 52.0-56.0 mm diam. after 14 d (growth rate 3.7-4.0 mm diam./d), creamy white to pale brown with regular margin, grey near the centre and hyphae clusters, reverse brown to dark brown rings, heterogeneous colour, with creamy-white edge. Conidial formation not observed.

Discussion
In the study of the phylogenetic affinity and position of Tubakia in the Ascomycota hierarchical system, Senanayake et al. (2017) placed this genus in the newly-introduced family Melanconiellaceae. However, the recently-published phylogenetic analyses, including sequence data of the type species of Tubakia, confirmed that Tubakia warranted a family of its own, Tubakiaceae (Braun et al. 2018) (Braun et al. 2018). The family comprises genera and species with sporodochia, crustose to pustulate pycnidioid stromatic conidiomata and superficial scutellate pycnothyria, monophialidic, colourless, conidiogenous cells, often with collarettes and conidia formed singly, mostly globose to broad ellipsoid-obovoid, aseptate, hyaline to pigmented, often with basal frill or truncate peg-like hilum.
The present study found two new species, one of which represents a novel genus in Tubakiaceae. In order to support the validity of the new species, we followed the guidelines of Braun et al. (2018). Based on ITS/LSU/rpb2 and ITS/tef1/tub2 molecular data, phylogenetic analyses revealed that two of the obtained isolates (SAUCC 0745 and SAUCC 0748) cluster in a separate lineage, fully supported at genus-level and related to the genera Racheliella, Oblongisporothyrium and Paratubakia. The new genus is named Obovoideisporodochium gen. nov. (type species: Obovoideisporodochium lithocarpi sp. nov.). The phylogenetic analyses also revealed that three isolates (SAUCC 1921, SAUCC 1923and SAUCC 1924 pertain to the genus Tubakia. Owing to different nucleotides in the concatenated alignment and morphology, two isolates (SAUCC 1921 andSAUCC 1923) of Tubakia were identified as a new species, namely T. lushanensis sp. nov, whereas the third isolate (SAUCC 1924) was identified as T. dryinoides.
The centre of genetic diversity of Tubakia appears to be in East Asia, where Quercus and other genera of Fagaceae are the most common hosts (Harrington & McNew 2018). Our study supports this phenomenon well. Tubakia lushanensis (SAUCC 1921 andSAUCC 1923) and T. dryinoides (SAUCC 1924) were isolated from Quercus palustris (Fagaceae), thereby increasing the genetic diversity of Tubakia in East Asia.