New Arthrobotrys Nematode-Trapping Species (Orbiliaceae) from Terrestrial Soils and Freshwater Sediments in China

Arthrobotrys is the most complex genus of Orbiliaceae nematode-trapping fungi. Its members are widely distributed in various habitats worldwide due to their unique nematode-trapping survival strategies. During a survey of nematophagous fungi in Yunnan Province, China, twelve taxa were isolated from terrestrial soil and freshwater sediment habitats and were identified as six new species in Arthrobotrys based on evidence from morphological and multigene (ITS, TEF, and RPB2) phylogenetic analyses. These new species i.e., Arthrobotrys eryuanensis, A. jinpingensis, A. lanpingensis, A. luquanensis, A. shuifuensis, and A. zhaoyangensis are named in recognition of their places of origin. Morphological descriptions, illustrations, taxonomic notes, and a multilocus phylogenetic analysis are provided for all new taxa. In addition, a key to known species in Arthrobotrys is provided, and the inadequacies in the taxonomic study of nematode-trapping fungi are also discussed.


Introduction
Nematophagous fungi are a group of fungi that parasitize, capture, and poison nematodes and important balancing agents of the nematode population in nature [1-3]. They were divided into different groups according to their mode of action on nematodes: (1) nematode-trapping fungi capture nematodes with specialized hypha structure, (2) endoparasitic fungi infect nematodes with spores, (3) egg parasitic fungi invade nematode eggs and females with hypha tips, and (4) toxin-producing fungi produce toxins that paralyze and kill nematodes [3][4][5]. Among these, nematode-trapping fungi have been the focus of related studies due to their highly specialized, sophisticated, and diverse trapping structures. Since Corda described the first nematode-trapping species (Arthrobotrys superba Corda) [6], more than 120 species have been discovered in Zygomycota (Zoopagaceae), Basidiomycota (Nematoctonus), and Ascomycota (Orbiliomycetes) over the past 180 years [5,7,8]. Nematode-trapping fungi in Zygomycota (Zoopagaceae) are poorly understood due to their immature isolation and culture methods [8,9]. All nematode-trapping fungi in Basidiomycota catch nematodes with adhesive knobs or adhesive spores, and all of them belong to Nematoctonus [8,[10][11][12]. All nematode-trapping fungi in the Ascomycota belong to Orbiliaceae (the only family of Orbiliomycetes), accounting for more than 80% provide more materials for the biological control of parasitic nematodes, and add precious research objects for evolutionary studies of nematode-trapping fungi.

Sampling, Fungal Isolation and Morphological Observation
The strains included in this study were isolated from terrestrial soil and freshwater sediment collected in Yunnan Province, China. Terrestrial soil samples were collected from 0-10 cm depth using a 35 mm-diameter soil borer after removing fallen leaves from the soil surface [49][50][51]. Freshwater sediment samples were removed from the water with a Peterson bottom sampler (HL-CN, Wuhan Hengling Technology Company, Limited, Wuhan, China). The samples were placed into a zip-lock bag, and relevant site information were recorded. The samples were stored at 4 • C until processing.
Samples of 1-2 g of soil or sediment were spread on the surface of cornmeal agar (CMA) plates with sterile toothpicks. Approximately 5000 nematodes (Panagrellus redivivus Goodey, free-living nematodes) were added as bait to promote the germination of the nematode-trapping fungi [4,32,52,53]. The plates were incubated at 26 • C for three weeks and then observed under a stereomicroscope; the spores of nematode-trapping fungi were transferred to fresh CMA plates using a sterile needle. This step was repeated until a pure culture was obtained [4,5].
The pure cultures were transferred to fresh CMA plates with observation well (a square slot 2 × 2 cm created by removing agar in each plate) using a sterile needle and incubated at 26 • C until the mycelium spread beyond the well. Approximately 1000 living nematodes were placed in the well to induce the formation of the trapping device [4,5]. The types of trapping devices were checked using a stereomicroscope. All micromorphological features were photographed and measured with an Olympus BX53 microscope (Olympus Corporation, Tokyo, Japan).

DNA Extraction, PCR Amplification and Sequencing
Total genomic DNA was extracted from mycelium grown on potato dextrose agar (PDA) plates using a rapid fungal genomic DNA isolation kit (Sangon Biotech Company, Limited, Shanghai, China). The ITS, TEF, and RPB2 regions were amplified with the primer pairs ITS4-ITS5 [54], 526F-1567R [55], and 6F-7R [56], respectively. The PCR amplification was performed as follows: 4 min of pre-denaturation at 94 • C; followed by 35 cycles of 45 s denaturation at 94 • C; 1 min of annealing at 52 • C (ITS), 55 • C (TEF), or 54 • C (RPB2), and 1.5-2 min of extension at 72 • C; with a final extension of 10 min at 72 • C. The PCR products were purified with a DiaSpin PCR Product Purification Kit (Sangon Biotech Company, Limited, Shanghai, China). The purified PCR products of the ITS and RPB2 regions were sequenced in the forward and reverse directions using PCR primers, and the primer pair 247F-609R [57] was used to sequence the TEF genes (BioSune Biotech Company, Limited, Shanghai, China). SeqMan v. 7.0 (DNASTAR, Madison, WI, USA) [58] was used to check, edit, and assemble the sequences. The sequences generated in this study were deposited in the GenBank database at the National Center for Biotechnology Information (NCBI; https://www.ncbi.nlm.nih.gov/, accessed on 26 February 2022), and the accession numbers are listed in Table 1.

Phylogenetic Analysis
The sequences generated in this study were compared against the NCBI GenBank database using BLASTn (https://blast.ncbi.nlm.nih.gov/, accessed on 11 February 2022). The BLASTn search results and the morphological features of these six species indicated that they belong to the genus Arthrobotrys. This genus was searched in the Species Fungorum (http://www.speciesfungorum.org, accessed on 13 February 2022), and all relevant records were checked individually according to the relevant documents to ensure that all Arthrobotrys taxa were considered in this study [4,5,8,13,19]. All reliable ITS, TEF, and RPB2 sequences of Arthrobotrys taxa were downloaded from GenBank database (Table 1). Three genes were aligned using the online program MAFFT v.7 (http://mafft.cbrc.jp/alignment/server/, accessed on 15 February 2022) [67] and manually adjusted using BioEdit v7. 2.3 [68]; they were then linked with MEGA6.0 [69]. Vermispora fusarina YXJ13-5 and Dactylaria higginsii CBS 121934 were selected as outgroups. Phylogenetic trees were inferred via maximum likelihood (ML), maximum parsimony (MP), and Bayesian inference (BI) analyses.
The SYM+I+G, GTR+I+G, and GTR+I+G models were selected via jModelTest v2.1. 10 [70] as the best-fit optimal substitution models for ITS, TEF, and RPB2, respectively, for maximum likelihood (ML) and Bayesian inference (BI) analysis.
Maximum likelihood (ML) analysis was implemented using IQ-Tree v1. 6.5 [71]. The dataset was partitioned, and each gene was analysed with the corresponding model. The statistical bootstrap support values (BS) were computed using rapid bootstrapping with 1000 replicates [72]. PAUP 4. a168 on XSEDE [73] in the CIPRES Science Gateway v. 3.3 web resource was used to generate the maximum parsimony (MP) analysis. Trees were inferred using the heuristic search option with TBR branch swapping and 1000 random sequence additions. Max-trees were set up at 5000 and no-increase. Clade stability was assessed via a bootstrap analysis with 1000 replicates [72]. Tree length (TL), consistency index (CI), retention index (RI), rescaled consistency index (RC), and homoplasy index (HI) values were calculated for all trees generated under different optimality criteria. All of the above parameters were edited into the PAUP block in the NEX file.
Bayesian inference (BI) analysis was conducted with MrBayes v. 3.2. 6. [74]. The multiple sequence alignment file was converted into a MrBayes-compatible NEXUS file using FastaConvert [75]. The dataset was partitioned, and the optimal substitution models of each gene were equivalently replaced to conform to the setting of MrBayes. Six simultaneous Markov chains were run for 10,000,000 generations, and trees were sampled every 100 generations. The first 25% of the trees were discarded, and the remaining trees were used to calculate the posterior probabilities (PP) in the majority rule consensus tree. All of the above parameters were edited in the MrBayes block in the NEX file.
The tree was visualized with FigTree v1. 3.1 [76]. The backbone tree was edited and reorganized using Microsoft PowerPoint (2013) and Adobe Photoshop CS6 software (Adobe Systems, San Jose, CA, USA).

Phylogenetic Analysis
A total of 118 Arthrobotry related taxa were listed in the Species Fungorum (http:// www.speciesfungorum.org/ (accessed on 6 March 2022)), representing 59 valid Arthrobotrys species. Among them, 51 species had confirmed molecular data. Therefore, the combined ITS, TEF, and RPB2 alignment dataset contained 64 Arthrobotrys isolates representing 57 Arthrobotrys species (plus our 12 isolates and 6 new species) and other related species in Orbiliaceae (Dactylellina: 4 species and Drechslerella: 4 species). The final dataset comprised 1918 characters (551 for ITS, 547 for TEF, and 820 for RPB2), among which 872 bp were constant, 1004 bp were variable, and 748 bp were parsimony informative. The maximum likelihood analysis of a best-scoring tree was performed with a final ML optimization likelihood value of −6304. 618465. Within the MP analysis, a strict consensus MP tree was obtained from the three most equally parsimonious trees (TL = 3443, CI = 0.546, RI = 0.510, RC = 0.298, HI = 0.419). For the Bayesian analysis (BI), the consensus tree was calculated with the remaining 75% of trees, and the Bayesian posterior probabilities were evaluated with a final average standard deviation of the split frequency of 0.009254. Although the trees inferred by ML, MP, and BI showed slightly different topologies in some clusters, all trees showed that all six species clustered together with known Arthrobotrys species, with distinct divergence from other species. The best-scoring ML tree was selected for presentation ( Figure 1).

Discussion
In this phylogenetic analysis, 65 species of nematode-trapping fungi used in this study were clustered into two large clades according to their mechanisms of catching nematodes. Clade I contained species that catch nematodes with adhesive trapping devices (adhesive nets and knobs). Clade II contained species that catch nematodes with active traps (constricting rings). Within clade I, species were clustered into two clades according to their trap types: one clade contained all species that produce adhesive nets, and the other contained those species that produce adhesive knobs. The results were consistent with previous studies [8,15,57,80] and again emphasized the importance of different types of trapping devices in the division of genera among nematode-trapping fungi. At the genus level, the taxonomy of Orbiliaceae nematode-trapping fungi remains an open question, especially in Arthrobotrys, which contains the greatest number of species. Morphologically, 61 species of Arthrobotrys can be divided into different groups according to the morphologies of their conidiophores and conidia [19]; however, phylogenetic studies have not supported this division; many phylogenetic clades show low support values, and the phylogenetic position of some Arthrobotrys species are unclear. The reason for this dilemma is the lack of molecular data for many species, and the existing data cannot provide a stable phylogenetic placement. Therefore, to thoroughly analyse the taxonomy of nematode-trapping fungi, we should use more comprehensive molecular data in future studies.
The emergence of molecular phylogenetic methods has led to unprecedented breakthroughs in the study of fungal taxonomy. Phylogenetic studies based on only a few molecular barcodes cannot provide sufficient and reliable information for the definition of fungal species; therefore, morphological descriptions of each species are still extremely important [81,82]. However, a significant problem facing fungal taxonomy studies is that the description of species is too shallow [83]. This problem is particularly prominent in Orbiliaceae nematode-trapping fungi and is mainly reflected in two aspects. (1) The descriptions of some morphological characteristics are too indistinct. Among six described species in this study, only A. eryuanensis and A. shuifuensis could be easily distinguished from known species based on their distinct morphological characteristics. The remaining four species required more detailed characteristics (such as the size of conidia) to be identified from known species. When mycologists measure the size of conidia, they are accustomed to uniformly calculating the size data of conidia with different shapes and septate numbers, and the sizes of these conidia usually show significant differences. This causes the size range of conidia to be too extensive for effective comparisons of different species [4,19]. (2) There are too few morphological features that can be used for species identification; although the description of a species includes many features, such as its trap type, conidia, chlamydospores, and hyphae, only the trap type, conidia, and conidiophores can be used for species identification [4,19]. As an increasing number of new species are established, it is difficult to distinguish some similar species based on these three characteristics only.
In conclusion, we should screen all potential morphological features in future studies to identify more features with significance for species identification. On the other hand, we should establish a unified standard morphological feature description model to facilitate comparisons between different species.
After the first nematode-trapping fungus was established in 1839 [6], the history of studies on the diversity of nematode-trapping fungi can be divided into three periods. In the nursery period, from 1839 to 1929, due to the limitation of separation methods, only five species were discovered over 90 years. In the rapid development period, from 1931 to 2009, the separation method improved gradually with the contributions of Drechsler et al. [24,25], and nearly 90 species were described over 80 years. From 2010 to 2019, only three species were discovered over 10 years (http://www.speciesfungorum.org (accessed on 6 March 2022)). These data indicated that the excavation of nematode-trapping fungi seems to have reached a plateau, and over time, it is unlikely that many new species will be discovered. However, in recent years, we have investigated nematode-trapping fungi in Yunnan Province and collected 10 new species (four previously published and six reported in this study) [18], which indicates that there are still many nematode-trapping fungi in nature that have not been discovered. Previous studies on the diversity of nematode-trapping fungi have mainly focused on soil habitat, whereas there have been considerably fewer investigations of aquatic nematode-trapping fungi [48,84,85]. However, three of the six new species described in this paper are from freshwater sediment, suggesting that aquatic habitats may also be important sources of nematode-trapping fungi and should not be ignored in future studies.

Data Availability Statement:
The data that support the finding of this study are contained within the article.