Three New Species of Apiospora (Amphisphaeriales, Apiosporaceae) on Indocalamus longiauritus, Adinandra glischroloma and Machilus nanmu from Hainan and Fujian, China

Apiospora is widely distributed throughout the world, and most of its hosts are Poaceae. In this study, Arthrinium-like strains were isolated from non-Poaceae in the Hainan and Fujian provinces of China. Based on the combined DNA sequence data of the internal transcriptional spacer (ITS), partial large subunit nuclear rDNA (LSU), translation extension factor 1-α gene (TEF1-α) and β-tubulin (TUB2), the collected Apiospora specimens were compared with known species, and three new species were identified. Based on morphological and molecular phylogenetic analyses, Apiospora adinandrae sp. nov., A. bawanglingensis sp. nov. and A. machili sp. nov. are described and illustrated.


Introduction
The genus Apiospora (Apiosporaceae, Amphisphaeriales, Sordariomycetes, Ascomycota) was identified and established by Saccardo (1875) with Apiospora montagnei as the type species [1].The Arthrium-like taxa under Apiosporaceae contains the genus with apiosporous hyaline ascospores and basauxic conidiogenesis [2,3].Thus, after Ellis (1965), the biological relationship between Apiospora and Arthrinium became widely accepted, and Apiospora was considered a synonym for Arthrinium until the study of Pintos and Alvarado, the former was usually considered the sexual morph and the latter was considered the asexual morph [3][4][5].During this time, the dual nomenclature was abandoned, and the old name Arthrinium was recommended for unitary nomenclature [6].
The genus Arthrinium was introduced by Kunze and Schmidt (1817) with Ar. caricicola as the type species [7].In order to clarify the exact status of the Arthrium-like taxa, most species were recollected, and some species were reassigned to different genera by Link (in Willdenow 1824) due to different conidial shapes, but were subsequently reclassified back to Arthrinium by von Höhnel 1925 and Cooke 1954, because of similar conidiophores [8][9][10].Pintos et al. (2019) published the genetic information of the first model species Ar. caricicola, followed by the phylogenetic analysis of Pintos and Alvarado (2021), which found that Arthrinium formed a separate clade from other sequences of Apiospora, and Apiospora and Arthrinium were phylogenetically distinguished [4,11].
The morphological similarities between the two genera (Apiospora and Arthrinium) make it difficult to determine the boundary between them by morphological methods alone.Morphologically, most of the conidia of Apiospora are nearly spherical in the front and lenticular at the side, and sometimes the conidia develop into pinholes, and the conidia of Arthrinium vary in shape (angular, spherical, curved, boat-shaped, fusiform, polygonal), with some species having thick black intervals [12][13][14][15].
Ecologically, Apiospora was widely distributed in various climatic zones around the world and was found in many species of hosts, whereas Arthrinium was rarely distributed in tropical and subtropical regions and had fewer hosts [13,14].Species of Apiospora have been found in Poaceae during past studies, while Arthrinium has been found in the family Cyperaceae and Juncaceae [4,[15][16][17][18][19][20].The ecological differences further support the genetic and morphological division of the two genera, and all the evidence is sufficient to support the separation of the two genera in taxonomic status.In addition, the host association and geographical distribution were discussed based on the existing research progress.
Currently, there are 120 records of Arthrinium and 157 records of Apiospora on Index Fungorum (http://www.indexfungorum.org/,accessed on 20 November 2023).In the present study, the authors isolated strains of Arthrinium-like taxa in China.To clarify the taxonomic status, morphological and phylogenetic studies were conducted.Three new species, Apiospora adinandrae sp.nov., A. bawanglingensis sp.nov.and A. machili sp.nov., were identified and classified in Apiospora by multilocus analysis of tandem internal transcribed spacer (ITS), 28S large subunit ribosomal RNA gene (LSU), translation elongation factor 1-alpha gene (TEF1α) and beta-tubulin (TUB) datasets.

Strains Isolation
The four strains were isolated from three host plants in two provinces of China (Table 1).These strains were obtained by tissue separation and the single spore isolation method [21].Surface sterilization was performed on 5 × 5 mm fragments at the junction of the leaf lesion edge and healthy tissue.The fragments were soaked in 75% alcohol for 90 s, washed with sterile water for 45 s, then immersed in 5% sodium hypochlorite solution for 60 s and finally washed with sterile water 3 times.The sterilized wet fragments were dried on sterilized filter paper and then cultured on potato dextrose agar (PDA: 220 g potato, 20 g agar, 18 g dextrose, 1000 mL sterile water and natural pH) 25 • C for 3 days.Subsequently, morphologically different colonies could be observed on PDA, and the tip of a hyphae with strong growth capacity growing on the edge of the colony was picked and transferred to another Petri dish containing PDA and culture was continued at 25 • C.

Morphological Studies
The colonies were observed morphologically on the 7th and 14th day of culture and photographed by digital camera (Canon Powershot G7X, Canon, Tokyo, Japan).A stereomicroscope (Olympus SZX10, OLYMPUS, Tokyo, Japan) and a microscope (Olympus BX53, OLYMPUS, Tokyo, Japan) were used to observe the microscopic morphological characteristics of the colonies.Both microscopes were equipped with high-definition color digital cameras to capture the conidia of fungal structures.
In order to facilitate further research, the four strains in this study were kept in 10% sterilized glycerol or sterile water at 4 • C. Voucher specimens were stowed in two institutions, the Herbarium of the Department of Plant Pathology (HSAUP) of Shandong Agricultural University, Taian, China, and the Herbarium Mycologicum Academiae Sinicae (HMAS) of the Institute of Microbiology, Chinese Academy of Sciences, Beijing, China.Ex-holotype living cultures were stored in the Shandong Agricultural University Culture Collection (SAUCC).All taxonomic information on the new taxa obtained in this paper has been submitted to MycoBank (http://www.mycobank.org,accessed on 20 November 2023).

DNA Extraction and Amplification
Fungal tissue was obtained on PDA and genomic DNA was extracted from the mycelium.DNA extraction was performed by a magnetic bead kit (OGPLF-400, Ge-neOnBio Corporation, Changchun, China) and the CTAB method [22,23].The polymerase chain reaction (PCR) procedure was performed using the primer pairs in Table 2, which contains the entire internal transcriptional spacer (ITS) of the intervening 5.8S rRNA gene, part of the large subunit nrRNA gene (LSU), part of the translation extension factor 1-α gene (TEF1α) and part of the beta-tubulin gene (TUB2).
The polymerase chain reaction was carried out using an Eppendorf Master Thermocycler (Hamburg, Germany).The amplification reaction was performed in a 20 µL reaction system consisting of 10 µL 2 × Hieff Canace ® Plus PCR Master Mix (With Dye) (Yeasen Biotechnology, Cat No. 10154ES03, Shanghai, China), the forward and reverse primer (TsingKe, Qingdao, China) each 0.7 µL per 10 µM, and 1.4 µL template genomic DNA, at last replenishing the total volume to 20 µL with distilled deionized water.The PCR products were separated by 1% agarose gel electrophoresis with GelRed added, and the bands were observed under ultraviolet light [24].Then, we used a Gel Extraction Kit (Cat: AE0101-C) (Shandong Sparkjade Biotechnology Co., Ltd., Ji'nan, China) for gel recovery.The bidirectional sequencing of DNA samples was completed by the Tsingke Company Limited (Qingdao, China), and the resulting sequences were processed by MEGA 7.0 to achieve consistency [25].The sequences data obtained and used in this article were all uploaded to GenBank (Table S1).

Phylogenetic Analyses
The new sequences obtained in present study were all compared in NCBI's Genbank nucleotide database (https://www.ncbi.nlm.nih.gov/,accessed on 11 November 2023) and all reference sequences for relevant species were downloaded.Multi-sequence analysis was performed using MAFFT 7 online services and the auto policy (http://mafft.cbrc.jp/alignment/server/, accessed on 20 November 2023) to compare the newly generated sequences with other related sequences.To identify isolates at the species level, a phylogenetic analysis of each marker was first performed and then combined (ITS-LSU-TEF1α-TUB2) (See Supplementary File S1).
Phylogenetic analysis of multi-labeled data was performed based on Bayesian inference (BI) and maximum likelihood (ML) algorithms.Both ML and BI were run on the CIPRES Science Gateway portal (https://www.phylo.org/,accessed on 20 November 2023) or offline software (The ML was operated in RaxML-HPC2 on XSEDE v8.2.12 and BI analysis was operated in MrBayes v3.2.7a with 64 threads on Linux) [30][31][32][33][34][35].For ML analyses, the default parameters were used and 1000 rapid bootstrap replicates were run with the GTR+G+I model of nucleotide evolution, and the BI analysis was performed using a fast bootstrap algorithm with an automatic stop option.The BI analyses included 2 million generations of sixty-four parallel threads with the stop rule options and 100 generations of sampling frequency.The burn-in score was set to 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, accessed on 20 November 2023), or ITOL: Interactive Tree Of Life (https://itol.embl.de/,accessed on 21 November 2023) [36], and the layout of the trees was produced in Adobe Illustrator CC 2019.

Phylogenetic Analyses
Phylogenetic analysis was performed on 99 isolates representing Apiospora species, of which 97 isolates were considered to be the ingroup and 2 strains of Arthrinium caricicola (CBS 145127) were used as the outgroup.The final alignment contained 2843 concatenated characters, viz.1-838 (ITS), 839-1674 (LSU), 1675-2293 (TEF1α), 2294-2843 (TUB2), 1619 were constant, 421 were variable and parsimony-uninformative and 803 were parsimonyinformative.The topology of the ML tree confirms the tree topology obtained from Bayesian inference; therefore, only the ML tree is presented (Figure 1).The alignment has 1456 distinct alignment patterns.The proportion of gaps and completely undetermined characters in this alignment: 33.53%.The estimated base frequencies were as follows: A = 0.234029, C = 0.250395, G = 0.257923, T = 0.257653; substitution rates AC = 1.257520,AG = 3.156093, AT = 1.159778,CG = 0.983356, CT = 4.511962 and GT = 1.000000; gamma distribution shape parameter α = 0.263195.Final ML optimization likelihood: −25299.135520.The GTR+I+G model was proposed for ITS, LSU, TEF1α and TUB2.BI analysis of these four tandem genes was performed over 3,400,000 generations in 68,002 trees.The first 17,000 trees representing the burn-in phase of the analysis are discarded, while the remaining trees are used to calculate the posterior probability in the majority rule consensus tree (Figure 1; first value: PP > 0.90 shown).The alignment embodied a total of 1469 unique site patterns (ITS: 475, LSU: 204, TEF1α: 440, TUB2: 350).

Discussion
In this study, three new species, viz., Apiospora adinandrae sp.nov., A. bawanglingensis sp.nov.and A. machili sp.nov., are introduced and described based on their morphological characters and phylogenetic status. A. bawanglingensis was collected from diseased leaves of Indocalamus longiauritus in Bawangling National Forest Park, Hainan Province, China. A. adinandrae and A. machili were collected from Wuyi Mountain, Fujian Province, China; the host of the former was Adinandra glischroloma and the latter was Machilus nanmu.
The asexual morphology of the species in this study is consistent with the basic characteristics of the genus.Morphologically, the biological relationship between Apiospora and Arthrinium has been controversial due to their similar morphological characteristics in having basauxic conidiogenesis [57].Although some morphological features of Arthrinium species are difficult to observe in Apiospora, such as the conidiophores having black thick septa, and it is still difficult to define the boundary between Apiospora and Arthrinium by the asexual form alone [4,58].It has been demonstrated that ascogenous Apiospora can reproduce the mycelial asexual state on artificial media such as PDA and MEA [57,60].Despite this cultural and molecular evidence of asexual to sexual transition, the sexual morphology of Apiospora is still rarely reported [61].
Ecologically, Apiospora is widely distributed in subtropical, tropical, temperate and even cold regions, including Africa, America, Asia, Australia, and Europe, according to reported data [13,15].As endophytes, plant pathogens and humus, Apiospora is ubiquitous in a wide range of terrestrial environments, such as soil, atmosphere and even marine substrates, but its main hosts are still plants, especially Poaceae [60].Of all the Apiospora species that have been reported, more than 60% of the hosts are Poaceae, of which about half are from bamboos [59].The new species A. bawanglingensis in this paper was also collected from Indocalamus longiauritus, which further enriches the diversity of bamboo fungi.Based on the existing statistical data of the USDA fungal database (https://nt.arsgrin.gov/fungaldatabases/,accessed on 10 November 2023.) and the collation of related literature published later on in the genus Apiospora, only 14 records of the genus Apiospora were isolated from woody plants (trees, shrubs, small shrubs), accounting for less than 10% of the total records, and about half of these hosts are Arecaceae.The strains A. adinandrae and A. machili in this study were isolated from Adinandra glischroloma and Machilus nanmu, respectively, which is the first time that Apiospora has been isolated from non-Poaceae hosts in China after Lu, B. (2000) [62].
Based on morphological and phylogenetic analysis, our study identified three new species in which only asexual morphology was found.Apiospora adinandrae sp.nov., A. bawanglingensis sp.nov.and A. machili sp.nov.are morphologically similar to their sister taxa, with significant differences in sequence.A. bawanglingensis was collected from Indocalamus longiauritus.Because most fungi on bamboo were not pathogens and there was no obvious plaque on the host, we believe for the time being that A. bawanglingensis is not a pathogen, and its pathogenicity still needs further study.A. adinandrae and A. machili were isolated from Adinandra glischroloma and Machilus nanmu, respectively.Based on the existing statistical data of the USDA fungal database, this is the first time that Apiospora has been found on Adinandra glischroloma and Machilus nanmu, enriching the host diversity of Apiospora.

Figure 1 .
Figure 1.A maximum likelihood tree showing the phylogenetic relationships of Apiospora inferred from the ITS, LSU, TEF1α and TUB2 sequences, and the roots on Arthrinium caricicola (CBS 145127).The Bayesian inference posterior probability (left, BIPP ≥ 0.90) and the maximum likelihood bootstrap value (right, MLBV ≥ 50%) are shown as BIPP/ML above the nodes.Strains marked with a star "*" and bolded represented are ex-types or ex-holotypes.Strains from the present study are in orange.The scale in the bottom middle indicates 0.1 substitutions per site.In order to make the layout of the evolutionary tree beautiful, some branches are shortened by two diagonal lines ("/") with the number of times.

Figure 1 .
Figure 1.A maximum likelihood tree showing the phylogenetic relationships of Apiospora inferred from the ITS, LSU, TEF1α and TUB2 sequences, and the roots on Arthrinium caricicola (CBS 145127).The Bayesian inference posterior probability (left, BIPP ≥ 0.90) and the maximum likelihood bootstrap value (right, MLBV ≥ 50%) are shown as BIPP/ML above the nodes.Strains marked with a star "*" and bolded represented are ex-types or ex-holotypes.Strains from the present study are in orange.The scale in the bottom middle indicates 0.1 substitutions per site.In order to make the layout of the evolutionary tree beautiful, some branches are shortened by two diagonal lines ("/") with the number of times.

Table 1 .
Strains associated with host and region in this study.

Table 2 .
Molecular markers and their PCR primers and programs used in this study.

Table 3 .
The asexual morphological characters of some Apiospora species.