Four new species of Trichoderma in the Harzianum clade from northern China

Abstract The Harzianum clade of Trichoderma comprises many species, which are associated with a wide variety of substrates. In this study, four new species of Trichoderma, namely T. lentinulae, T. vermifimicola, T. xixiacum, and T. zelobreve, were encountered from a fruiting body and compost of Lentinula, soil, and vermicompost. Their colony and mycelial morphology, including features of asexual states, were described. For each species, their DNA sequences were obtained from three loci, the internal transcribed spacer (ITS) regions of the ribosomal DNA, the gene encoding the second largest nuclear RNA polymerase subunit (RPB2), the translation elongation factor 1-α encoding gene (TEF1-α). The analysis combining sequences of the three gene regions distinguished four new species in the Harzianum clade of Trichoderma. Among them, T. lentinulae and T. xixiacum clustered with T. lixii, from which these new species differ in having shorter phialides and smaller conidia. Additionally, T. lentinulae differs from T. xixiacum in forming phialides with inequilateral to a strongly-curved apex, cultural characteristics, and slow growth on PDA. Trichoderma vermifimicola is closely related to T. simmonsii, but it differs from the latter by producing phialides in verticillate whorls and smaller conidia. Trichoderma zelobreve is the sister species of T. breve but is distinguished from T. breve by producing shorter and narrower phialides, smaller conidia, and by forming concentric zones on agar plates. This study updates our knowledge of species diversity of Trichoderma.


Introduction
The genus Trichoderma Pers., introduced by Persoon (1794), is cosmopolitan, including saprotrophs and mycoparasites in a diversity of ecosystems, such as agricultural fields, prairies, forests, salt marshes, and fungal fruiting body (Gazis and Chaverri 2010;Chaverri et al. 2015;Qiao et al. 2018). Species of this genus have been widely used in the biocontrol of plant pathogens Degenkolb et al. 2015; Bunbury-Blanchette and Walke 2019) and production of enzymes and bioactive compounds (Sun et al. 2016). Nevertheless, some of them are associated with green mold diseases in the commercial production of mushrooms (Innocenti et al. 2019;Sun et al. 2019a). Morphologically, the asexual-morphs are similar in producing branched treelike conidiophores with cylindrical to nearly subglobose phialides and ellipsoidal to globose conidia, but their variation is insufficient to differentiate the Trichoderma species Qin and Zhuang 2017;Qiao et al. 2018). Multilocus molecular phylogeny, based on combined sequence data of the internal transcribed spacer (ITS) regions, RNA polymerase II subunit (RPB2), and the translation elongation factor 1-α gene (TEF1-α), enables rapid and accurate identification of the Trichoderma species Atanasova et al. 2013;Chaverri et al. 2015). Currently, the combination of multi-gene phylogenetic analysis and phenotypic characteristics is extensively applied in species delimitation of Trichoderma (du Plessis et al. 2018;Qiao et al. 2018;Innocenti et al. 2019).
Trichoderma harzianum Rifai is one of the most well-known Trichoderma species, due to its antifungal properties and effective bio-control ability, used to suppress soil-borne plant pathogens Degenkolb et al. 2015; Bunbury-Blanchette and Walker 2019). As a cosmopolitan and ubiquitous fungus, it has been isolated from diverse substrates, such as soil, plant tissue, and mushrooms Jaklitsch and Voglmayr 2015;Innocenti et al. 2019;Sun et al. 2019b). Since Chaverri et al. (2015) provided a systematic revision of species in the Harzianum clade, numerous new species have been described (Jaklitsch and Voglmayr 2015;Qin and Zhuang 2016a;Sun et al. 2016;Chen and Zhuang 2017b;Qiao et al. 2018). Currently, more than 60 species are placed in the Harzianum clade (Jaklitsch and Voglmayr 2015;Zhuang 2016a, b, 2017;Chen and Zhuang 2017b;Qiao et al. 2018;Phookamsak et al. 2019;).
It is estimated that 136 new species of Trichoderma have been recognised since 2015 (www.indexfungorum.org 2020), with 84 among these reported from China (Sun et al. 2012;Zhuang 2016a, b, 2017;Chen and Zhuang 2017a, b;Qiao et al. 2018), which evidenced that China has a high species diversity of Trichoderma (Zhu and Zhuang 2015;Jiang et al. 2016). In our survey of Trichoderma, eighteen isolates were obtained from soil, mushroom substrates, and vermicompost from northern China. Four new species belonging to the Harzianum clade were identified based on morphological features and DNA sequence data at three loci: the genes encoding RNA polymerase II subunit (RBP2) and translation elongation factor 1-α gene (TEF1-α), and the internal transcribed spacer (ITS) regions of the nuclear ribosomal RNA gene.

Sampling sites and strains isolation
Since Trichoderma is easily isolated from soil, mushroom substrates, and earthworm substrates, the soil, mushroom substrates, and earthworm were therefore collected from Yinchuan, Ningxia Hui Autonomous Region, and Chaoyang district, Beijing, China. All the samples were stored at 4 °C before fungal isolation. Trichoderma strains were isolated by gradient dilution and the spread plate method or directly from the mushroom substrates. Three dilutions (10 -1 , 10 -2 , and 10 -3 ) were prepared with 1 g soil and sterile water, and 100 µl of each dilution was spread on a 9 cm diameter Petri dish of PDA agar with100 mg/L chloramphenicol added. The plates were then incubated at 25 °C. Each of the individual colonies was transferred to a new PDA dish after 1-3 days and incubated at 25 °C. Dried cultures from the single spore or specimens of new species were deposited in the Herbarium Mycologicum Academiae Sinicae (HMAS) and the ex-type strains were preserved in the China General Microbiological Culture Collection Center (CGMCC)

Morphological analysis
For morphological studies, we used three different media: cornmeal dextrose agar (CMD, Difco, BD Science, USA), PDA (Difco, BD Science, USA), and synthetic low nutrient agar (SNA, Difco, BD Science, USA) . Each strain was first cultured on an SNA plate for 3 days and a small agar piece of 0.5 cm diameter with mycelium was then transferred, respectively, to new CMD, PDA, and SNA plates. Strains were incubated in 9 cm diam with three replicates. Petri dishes at 25 °C with a 12 h natural light and 12 h darkness interval. Colony diameter at 25 °C was measured three days after inoculation, and the time when mycelium entirely covered the surface of the agar plate was also recorded. Micromorphological characters were examined from the cultures of one-week-old colonies on SNA . A Nikon Ellipse 80i light microscope, equipped with differential interference contrast (DIC) optics, was used to capture digital images.

DNA extraction, PCR and sequencing
Genomic DNA of each strain was extracted from fresh mycelium growing on PDA after 5 days of growth following the rapid "thermolysis" method described in Zhang et al. (2010). For the amplification of ITS, RPB2, and TEF1-α gene fragments, ITS4 and ITS5 for ITS (White et al. 1990), EF1-728F (Carbone and Kohn 1999) and TEF1LLErev (Jaklitsch et al. 2005) for TEF1, and RPB2-5F and RPB2-7R for rpb2 (Liu et al. 1999) were used. Each PCR reaction consisted of 12.5 µl T5 Super PCR Mix (containing Taq polymerase, dNTP, and Mg 2+ , Beijing TsingKe Biotech Co. Ltd., Beijing), 1.0 µl of forward primer (10 µM), 1.0 µl of reverse primer (10 µM), 0.5 µl DMSO, 3 µl DNA template and 7 µl double sterilized water. PCR reactions were in Eppendorf Mastercycler, following the protocols described by Sun et al. (2016). PCR products were purified with the PCR product purification kit (TIANGEN Biotech, Beijing, China), and sequencing was carried out in both directions on an ABI 3730 XL DNA sequencer (Applied Biosystems, Foster City, California) with primers used during PCR amplification.

Phylogenetic analyses
Preliminary BLAST searches with ITS, RPB2, and TEF1-α gene sequences of the new isolates against NCBI, TrichOKey (Druzhinina and Kopchinski 2006), and TrichoBlast  databases identified species closely related to our isolates. Based on this information, sequences of ITS, RBP2, and TEF1-α of 133 strains, representing 59 species were downloaded from GenBank, following recent publications (Qin and Zhuang 2017;Qiao et al. 2018;Innocenti et al. 2019). Among them, 139 strains are belonging to the Harzianum clade, and Trichoderma ceramicum, T. parestonicum, and T. estonicum were chosen to represent the outgroup.
Tree alignment files were generated by using MAFFT version 7.03 with the Q-INS-I strategy (Katoh and Standley 2013). Conserved blocks were selected from the initial alignments with Gblocks 0.91 b (Castresana 2000). The appropriate nucleotide substitution model for each gene was determined by using MrModeltest v2.4 (Nylander 2004). HKY + I + G was estimated as the best-fit model for RPB2, and GTR + I + G was estimated as the best-fit model for TEF1-α and ITS under the output strategy of AIC. The partition homogeneity test (p = 0.01) indicated that the individual partitions were not significantly incongruent (Cunningham 1997), thus the aligned sequences of ITS, RPB2, and TEF1-α were combined for analyses. The multi-locus phylogenetic analyses included 1065 characters for RBP2, 587 characters for TEF1-α, and 555 characters for ITS. All characters were weighted equally and gaps were treated as missing characters.
Maximum Likelihood (ML) analyses were performed by RAxML (Stamatakis 2006), using the GTR-GAMMA-I model. The maximum likelihood bootstrap proportions (MLBP) were using 1000 replicates. Bayesian Inference (BI) analyses were conducted with MrBayes v3.2.6 (Ronquist et al. 2012). Metropolis-coupled Markov Chain Monte Carlo (MCMC) searches were calculated for 10,000,000 generations, sampling every 100 th generation with the best best-fit model for each gene. Two independent analyses with six chains each (one cold and five heated) were carried out until the average standard deviation of the split frequencies dropped below 0.01. The initial 25% of the generations of MCMC sampling were discarded as burn-in. The refinement of the phylogenetic tree was used for estimating Bayesian inference posterior probability (PP) values. The Tree was viewed in FigTree v1.4 (Rambaut 2012), values of Maximum likelihood bootstrap proportions (MLBP) greater than 50% and Bayesian inference posterior probabilities (BIPP), greater than 95% at the nodes, are shown along branches. The final alignments and the trees obtained have been deposited in TreeBASE (TreeBASE accession number: 25400).
On PDA after 72 h, colony radius 55-58 mm at 25 °C, covering the plate at 30 °C, 8-9 mm at 35 °C. Colony white to yellow-white; mycelium dense and radial. Aerial conspicuous. No diffusing pigment, not distinct odor (Fig. 5A). Conidial production noted after 3 days, starting around the inoculum, effuse in the aerial hyphae, first white, turning green after 4 d. Chlamydospores unobserved.

Discussion
A combination of phylogenetic, morphological, ecological, and biogeographical data has robustly resolved the taxonomy of Trichoderma (Jaklitsch and Voglmayr 2015;Qin and Zhuang 2016a;Sun et al. 2016;Chen and Zhuang 2017b;Qiao et al. 2018). In this study, phylogenetic analysis based on a single gene of ITS could not distinguish species of Trichoderma in the Harzianum clade from each other (Suppl. material 1: Fig. S1), which confirmed that the ITS region is not suitable for species delimitation of Trichoderma (Jaklitsch et al. 2012;Qin et al. 2018). Sequences of RPB2 and TEF1-α were powerful due to their suitable interspecific variations (Jaklitsch and Voglmayr 2015), and these have extensively been used in solving the taxonomy of Trichoderma (Jaklitsch and Voglmayr 2015;Qin and Zhuang 2016a;Chen and Zhuang 2017a, b;Qiao et al. 2018). Despite the phylogenetic analyses based on the single gene of RPB2 and TEF1-α generally revealed the phylogenetic relationship within the Harzianum clade (Suppl. material 1: Fig. S2, Suppl. material 3: Fig S3), but the relationships among T. lentinulae, T. xixiacum, T. vermifimicola, T. zelobreve, and their closed taxa were not well distinct. Consideration of the universality and reliability of barcodes for species in the Trichoderma genus (Qiao et al. 2018), combined ITS, RPB2, and TEF1-α dataset was used for phylogenetic analysis in this study, revealing phylogenetic relationship among species in Harzianum clades, and suggesting that T. lentinulae, T. xixiacum, T. vermifimicola, and T. zelobreve are distinguishable from each other and species within and outside of Harzianum clade as well. Trichoderma lentinulae was phylogenetically close to T. xixiacum and T. lixii but represents a taxon (Fig. 1). Morphologically, it differed from T. xixiacum in producing less frequently lageniform phialides with inequilateral to a strongly-curved apex.
Currently, the Harzianum clade contains more than 60 species which were isolated from soil, plant tissues, and other fungi (Jaklitsch and Voglmayr 2015;Qin and Zhuang 2016a;Chen and Zhuang 2017b;Qiao et al. 2018;Sun et al. 2019a, b). Several studies have confirmed that species in this clade are important because of their mycoparasitism Chen and Zhuang 2017a;Sun et al. 2019). When numerous biological control agents were explored deriving from species in the Harzianum clade , several taxa, such as T. atrobrunneum T. pleuroti, and T. pleuroticola were recognized as causing agents of "Green mold" disease of cultivated mushroom (Innocenti et al. 2019;Sun et al. 2019a, b). In this study, T. lentinulae was isolated from a fruiting body and the cultivated substrates of L. edodes, causing the decay of the host as well. How T. lentinulae affect the cultivation of Lentinula edodes is worthy of further studies. Since T. lentinulae was isolated from mushroom, T. lentinulae and T. vermifimicola were isolated from the mushroom spawn and substrates for earthworm cultivation, T. xixiacum and T. zelobreve were isolated from soil, confirming that species in the Harzianum clade have flexible nutrition modes (Chaverri and Samuels 2013;Zhang et al. 2018). The new species introduced here are not only potential candidates for biological agent exploration, but also improve our understanding of the diversity of Trichoderma, especially of the Harzianum clade in China.