The genus Arthrinium (Ascomycota, Sordariomycetes, Apiosporaceae) from marine habitats from Korea, with eight new species

Species of Arthrinium are well-known plant pathogens, endophytes, or saprobes found in various terrestrial habitats. Although several species have been isolated from marine environments and their remarkable biological activities have been reported, marine Arthrinium species remain poorly understood. In this study, the diversity of this group was evaluated based on material from Korea, using morphological characterization and molecular analyses with the internal transcribed spacer (ITS) region, β-tubulin (TUB), and translation elongation factor 1-alpha (TEF). A total of 41 Arthrinium strains were isolated from eight coastal sites which represented 14 species. Eight of these are described as new to science with detailed descriptions. Supplementary Information The online version contains supplementary material available at 10.1186/s43008-021-00065-z.


INTRODUCTION
The genus Arthrinium, which belongs to Apiosporaceae in Xylariales (class Sordariomycetes in Ascomycota), was first recognized and established more than 200 years ago, with A. caricicola as type species (Schmidt and Kunze 1817). To date, it comprises approximately 88 species worldwide (Index Fungorum: http://www. indexfungorum.org).
Arthrinium species have traditionally been classified based on morphological characteristics such as conidial shape, conidiophores, and the presence or absence of sterile cells and setae (Schmidt & Kunze 1817;Hughes 1953;Minter 1985). Among these characteristics, conidial shape appears to be diagnostic for distinguishing species (Singh et al. 2013). However, morphological variation is often observed depending on the growth substrate and incubation period (Crous & Groenewald 2013). As such, species identification based on morphological characteristics is problematic and impractical. To address this problem, DNA sequences of the internal transcription spacer (ITS), translation elongation factor 1-alpha (TEF), and β-tubulin gene (TUB) were employed to delimit and recognize closely related Arthrinium species and infer their phylogenetic relationships (Crous & Groenewald 2013).
Arthrinium species have been globally reported as endophytes, plant pathogens, and saprobes and are commonly isolated from various terrestrial environments, including air, plants, and soil (Kim et al. 2011;Crous & Groenewald 2013;Wang et al. 2018). More recently, isolation from various marine environments, including seawater, seaweed, and the inner tissues of marine sponges, has been reported (Miao et al. 2006;Tsukamoto et al. 2006;Suryanarayanan 2012;Flewelling et al. 2015;Hong et al. 2015;Wei et al. 2016;Elissawy et al. 2017;Li et al. 2017). Arthrinium species isolated from sponges, egg masses of sailfin sandfish, and seaweeds showed promising bioactive properties, including high enzymatic activity, antifungal activity, and antioxidant capacity (Elissawy et al. 2017;Li et al. 2017;Park et al. 2018). Some species (A. arundinis, A. phaeospermum, A. rasikravindrae, A. sacchari, and A. saccharicola) have been detected in both marine and terrestrial environments (Wang et al. 2018). Whether these species have specific adaptations to survive in seawater requires further investigation. A recent study showed that marine Arthrinium species developed strategies to adapt to marine environments, such as a symbiotic partnership with seaweed (Heo et al. 2018). In marine systems, dissolved organic matter in seawater can absorb ultraviolet radiation and produce reactive oxygen species (ROS), which cause oxidative stress on marine microorganisms (Mopper & Kieber 2000). Heo et al. (2018) detected relatively high antioxidant activity and radical-scavenging activity in marinederived Arthrinium species. The antifungal activity of seaweed-pathogenic fungi has also been studied (Hong et al. 2015;Heo et al. 2018). Arthrinium saccharicola (KUC21342) has the potential to inhibit the growth of Asteromyces cruciatus, a pathogenic fungus that attacks brown algae (Heo et al. 2018). The discovery of the promising bioactivities of marine Arthrinium species was one of the reasons motivating our subsequent investigation of the diversity of marine Arthrinium in Korea.
Six species of Arthrinium have previously been reported from marine environments in Korea: A. arundinis, A. marii, A. phaeospermum, A. rasikravindrae, A. sacchari, and A. saccharicola (Hong et al. 2015;Heo et al. 2018;Park et al. 2018). However, many marine species remain unidentified owing to the lack of resolution in ITS-based phylogenies and the paucity of morphological characteristics. The aim of this study was to investigate marine Arthrinium species from coastal environments in Korea and to identify them using morphological characteristics and multigene phylogenies (ITS, TEF, and TUB).

Sampling and isolation
The seaweed Sargassum fulvellum and unidentified seaweeds were collected from two locations, Taean-gun on the west coast of Korea and Jeju Island south of Korea. To isolate the fungi, the seaweeds were washed with distilled water and cut into small pieces (approximately 5 mm diam) using a sterile surgical blade. The pieces were treated with 70% ethanol for 60 s and washed in sterile distilled water for 10 s. Each piece was placed on 2% malt extract agar (MEA) supplemented with 0.01% streptomycin and 0.01% ampicillin to inhibit bacterial growth. The plates were incubated at 25°C for 7-15 d. Suspected Arthrinium colonies were transferred onto potato dextrose agar (PDA, Difco, Sparks, MD, USA) plates. The colonies were subsequently identified as belonging to Arthrinium based on ITS sequences (see below). A total of 14 Arthrinium strains were isolated in this study and an additional 27 Arthrinium strains were obtained from the Seoul National University Fungus Collection (SFC), Seoul, Korea. Each strain is stored in 20% glycerol at − 80°C in the Korea University Fungus Collection (KUC), Seoul, Korea. Type specimens were deposited in the Korean Collection for Type Culture, Daejeon, Korea (KCTC), with ex-type living cultures deposited in KUC.

DNA extraction, PCR amplification, and sequencing
Genomic DNA was extracted using an Accuprep Genomic DNA extraction kit (Bioneer, Korea) according to the manufacturer's protocol. PCR targeting the ITS, TUB, and TEF regions was carried out according to a previously described method (Hong et al. 2015). For the ITS region, the primers ITS1F and ITS4/LR3 were used (White et al. 1990;Gardes & Bruns 1993); for TUB, we employed Bt2a/T10 and Bt2b/T2 (Glass & Donaldson 1995;O'Donnell & Cigelnik 1997), and for TEF, we used EF1-728F and EF2 (O'Donnell et al. 1998;Carbone & Kohn 1999). All PCR products were checked on a 1% agarose gel and purified with the AccuPrep PCR/Gel DNA Purification Kit (Bioneer, Seoul, Korea). DNA sequencing was performed at Macrogen (Seoul, Korea) on an ABI3730 automated DNA Sequencer (Applied Biosystems, Foster City, CA) using the same set of primers for each locus. Additional DNA sequences of some strains were obtained from previous studies (Hong et al. 2015;Heo et al. 2018). All new sequences generated in this study were deposited in GenBank (Table 1).

Phylogenetic analysis
ITS sequences were assembled, proofread and edited using MEGA v. 7 (Kumar et al. 2016) and subsequently aligned with Arthrinium reference sequences from Gen-Bank using MAFFT 7.130 (Katoh and Standley 2013). To adjust the ambiguous alignment manually, maximum likelihood analysis was performed using all sequence where ambiguous regions excluded using G-block. Then, the original sequences were aligned based on the supported clades, and ambiguous regions were manually adjusted.
Maximum likelihood (ML) analyses were conducted using RAxML v. 7.03 (Stamatakis 2006) and a GTR + G model with 1000 bootstrap replicates. Bayesian tree inference (BI) was carried out using MrBayes version 3.2    (Ronquist et al. 2012), with the best model (HKY + I + G) selected for each marker based on the Bayesian information criteria using jModeltest v. 2.1.10 (Darriba et al. 2012). To achieve stationary equilibrium, 20 million trees were generated, and trees were sampled every 1000 generations. The first 25% of the trees was discarded as burn-in, and the remaining 75% was used for calculating posterior probabilities (PP) in the majority rule consensus tree. All analyses were performed on the CIPRES web portal (Miller et al. 2010). The sequences of the other two loci (TEF and TUB) were individually aligned with Arthrinium reference sequences from GenBank using the same approach described for the ITS. ML and BI analyses also followed the above criteria. The models for TEF and TUB were HKY + I + G and K80 + I + G, respectively. The ITS taxa for the multigene tree were different from those of the single ITS tree, so the model test for the ITS region was redone for the multigene analysis. As a result, the SYM + G model was applied to ITS region in the multigene tree. Finally, sequence concatenation was performed using the same methods and models assigned for each locus described above.

Morphological observation
Strains were grown on oatmeal agar (OA, Difco™), PDA, and MEA at 15, 20, and 25°C in darkness for 14 d. The culture characteristics, such as surface structure, presence of aerial mycelium and the colour of the mycelium, colour of colony or medium, and sporulation (Crous et al. 2009), were recorded. Colors and the corresponding codes were evaluated according to the Munsell color chart (Munsell Color, 2009). To determine fungal growth rates, the diameter of each colony was measured every 24 h, and each measurement was performed in triplicate. Microscopic characters were observed with an Olympus BX51 light microscope (Olympus, Tokyo, Japan). Samples were mounted in water to take pictures of conidiophores and conidia, and pictures were taken using a DP20 microscope camera (Olympus, Tokyo, Japan). At least 30 individuals were measured for each microscopic character. To illustrate the range of variation, 5% of the extreme measurements from each end of the range are given in parentheses.
Scanning electron microscope (SEM) was used to observe detailed morphological characters. Colonies sporulating abundantly on PDA, MEA, and OA were freezedried. Ion coating and observation were performed by Wooyoung Solution Inc. (Suwon, Korea), using an S-5200 scanning electron microscope (Hitachi, Tokyo, Japan). The SEM images were taken under 1500x to 8000x magnifications.

RESULTS
A total of 41 Arthrinium strains were identified, representing six known and eight new species. Of these strains, 26 were isolated from various seaweeds, 14 from the eggs of sailfin sandfish, and one from beach sand. The dominant species were three of the new species, A. agari (5 strains), A. arctoscopi (5 strains), and A. marinum (5 strains) (Table 1).
A total of 21 ITS (580-1150 bp), 24 TEF (420-970 bp), and 22 TUB (400-560 bp) sequences were newly generated for the 41 Arthrinium strains. The ITS phylogeny contained 124 terminals, including Nigrospora gorlenkoana as outgroup. The concatenated three-gene phylogeny contained 95 terminals, consisting of 749, 613, and 503 characters respectively, including gaps. Preliminary identification was based on the ITS region, and multigene analysis was used to test the identifications, determine the phylogenetic relationships among the taxa, and to resolve closely related species. Both the ML and Bayesian analyses showed the same tree topologies and the ML tree is represented (Figs. 1, 2). The 41 Arthrinium strains obtained in this study formed five clades (A, B, C, D, and E), both in the ITSbased and combined phylogeny analyses (Figs. 1, 2). In the ITS tree, many Arthrinium species were distinguished from one another. However, some were not clearly separated (clades B and D) and the relationships of the others (clades C and D) were not resolved. The above problem was solved in the individual trees of TEF and TUB (Figs. 1S, 2S), and the multigene tree based on the ITS, TUB, and TEF regions (Fig. 2). The multigene analysis supported the conclusion that six taxa corresponded to known species. Eight putatively novel species were classified into five clades (Fig. 2). The eight species were clearly separated from the previously sequenced taxa, each forming a clade with high support (over 99% of BS, 0.99 of PP) (Fig. 2). Arthrinium agari and A. koreanum. Were included in clade A, A. piptatheri and A. fermenti were in clade D, and A. pusillispermum and A. taeanense were in clade E. Comparison with morphoanatomical and other data of species that have so far not been sequenced supported our interpretation of these eight entities representing novel species. Description: Mycelium of smooth, hyaline, branched, septate, hyphae 2.0-3.5 μm diam. Conidiogenous cells aggregated in clusters on hyphae or solitary, at first hyaline, becoming pale green, cylindrical, sometimes ampulliform. Conidia brown, smooth to granular, globose to subglobose in surface view, (8.5-)9.0-10.5 × (7.0-)7.5-8.5 (− 9.0) μm ( x = 9.5 × 8.1 μm, n = 30); lenticular in side view, with equatorial slit, 5.5-7.0 μm wide ( x = 6.4 μm, n = 30), elongated cell observed.
Culture: PDA: colonies thick, concentrically spreading with aerial mycelium, margin irregular; mycelia creamy white; sporulation was not observed; pigment absent in medium; odour indistinct. MEA: colonies flat, concentrically spreading with aerial mycelium, margin irregular;   Notes: Arthrinium arctoscopi is closely related to A. obovatum (98.84% similarity in the ITS region, 96.10% in the TEF region, and 94.31% in the TUB region) and A. aquaticum (99.80% similarity in the ITS region). However, A. arctoscopi can be distinguished from A. obovatum by the conidial shape and growth rate; the conidia of A. arctoscopi are globose to subglobose, whereas those of A. obovatum are obovoid or occasionally elongated to ellipsoid in shape (Wang et al. 2018). In addition, the growth rate of A. arctoscopi (7-9 mm in 7 d at 25°C, PDA) is slower than that of A. obovatum The reference species were cited from the following marks: a (Crous and Groenewald  (covering a 90 mm Petri dish in 7 d at 25°C, PDA) (Wang et al. 2018). The conidial shape of A. arctoscopi is also slightly different from that of A. aquaticum (globose to subglobose conidia, 9-11 × 8-10 μm, x = 10 × 9 μm, n = 20). Two non-sequenced species, A. algicola and A. sinensis, are morphologically similar to A. arctoscopi. The longer length and narrower width of A. algicola conidia (10.5-15 × 6-8 μm) and lageniform conidiogenous cell of A. sinensis distinguish them from A. arctoscopi (Table 2). Arthrinium fermenti S.L. Kwon, S. Jang & J.J. Kim, sp. nov.
Notes: Arthrinium fermenti is closely related to A. pseudospegazzinii (98.96% similarity in the ITS region, 92.47% in the TEF region, and 95.00% in the TUB region) (Figs. 1,  2). It can be distinguished from the latter by conidial shape and colony colour. The conidia of A. fermenti are globose to elongate-ellipsoid, whereas A. pseudospegazzinii has uniformly globose conidia (Crous & Groenewald 2013). Moreover, while the colonies of A. pseudospegazzinii were light orange on PDA and dirty white with an olivaceous grey patch on OA and MEA (Crous & Groenewald 2013), A. fermenti colonies had a yellowish to reddish colour on OA and MEA and a strong yeast odour. Arthrinium globosum (non-sequenced species) has a conidia shape similar to that of A. fermentiglobose to subglobose. However, a lenticular shape in side view was not observed in A. globosum (Table 2).
Notes: Arthrinium koreanum is closely related to A. qinlingense (98.48% similarity in the ITS region, 94.92% in the TEF region, and 94.85% in the TUB region) (Figs. 1,  2). They can be distinguished by their conidial sizes; 7.5-11 × 5.5-10 μm in A. koreanum vs. 5-8 μm in diameter in A. qinlingense (Jiang et al. 2018). Arthrinium koreanum has a similar conidia shape to that of the two nonsequenced species, A. globosum and A. sphaerospermum. However, the conidia of the latter two species only have globose to subglobose shape, and lenticular shape is not observed in side view (Table 2).
Notes: Although Arthrinium marinum and A. rasikravindrae were not distinguished on ITS alone (100% similarity in the ITS region), these species formed two distinct clades based on the combined analysis of the ITS, TUB, and TEF regions (99.08% in the TEF region and 97.97% in the TUB region) (Figs. 1, 2). They can also be distinguished by their growth rates: A. marinum (16-17 mm in 5 d on PDA at 20°C) had a slower growth rate than A. rasikravindrae KUC21327 (34-39 mm in 5 d on PDA at 20°C).
Culture: PDA: colonies thick around centre, concentrically spreading with aerial mycelium, margin circular; mycelia white, pale yellow to grey; sporulation was not observed; greenish black (10GY 2.5/1) pigment diffused in medium; odour indistinct. MEA: colonies abundant, flat, concentrically spreading with sparse aerial mycelium, margin irregular; mycelia white to gray colored; sporulation was not observed; pigment absent in medium; odour indistinct. OA: colonies thick, concentrically spreading with aerial mycelium, margin irregular; mycelia white to pale brown and grey to dark grey; sporulation on hyphae around the centre after 2 weeks, spores black; greenish black (10Y 2.5/1) to very dark greenish grey (10Y 3/1) pigment diffused in medium; odour indistinct. Notes: Arthrinium pusillispermum is closely related to A. gutiae (99.44% similarity in the ITS region, 88.52% in the TEF region, and 98.98% in the TUB region) (Figs. 1,  2). Arthrinium pusillispermum is distinguished from A. gutiae by the shape of the conidiogenous cells and the substrate: A. pusillispermum has cylindrical conidiogenous cells and was isolated from seaweed, whereas A. gutiae has lageniform conidiogenous cells and was isolated from the gut of grasshoppers (Crous et al. 2015). Arthrinium pusillispermum can be distinguished from the 22 nonsequenced species by its small conidia size (Table 2).