﻿Molecular phylogeny and morphology reveal two new entomopathogenic species of Ophiocordyceps (Ophiocordycipitaceae, Hypocreales) parasitic on termites from China

﻿Abstract Two new termite-pathogenic species, Ophiocordycepsglobiperitheciata and O.longistipes, are described from Yunnan Province, China. Six-locus (ITS, nrSSU, nrLSU, tef-1α, rpb1 and rpb2) phylogenetic analyses in combination with morphological observations were employed to characterize these two species. Phylogenetically, O.globiperitheciata is most closely related to Hirsutellacryptosclerotium and O.communis, whereas O.longistipes shares a sister relationship with O.fusiformis. However, O.globiperitheciata differs from H.cryptosclerotium by parasitizing Blattodea and producing clavate, unbifurcated stromata. Ophiocordycepsglobiperitheciata is distinguished from O.communis by multiple stromata, shorter asci and ascospores. Ophiocordycepslongistipes differs from O.fusiformis in producing larger stromata, perithecia, asci and ascospores, as well as smaller citriform or oval conidia. Morphological descriptions of the two new species and a dichotomous key to the 19 termite-pathogenic Ophiocordyceps species are presented.


Introduction
Invertebrate-associated fungi are intriguing and diverse, widely distributed around the world (Araújo et al. 2018;Luangsa-ard et al. 2018;Haelewaters and Kasson 2020;Wilson et al. 2021;Santamaria et al. 2023).There are two typical relationships between fungi and invertebrates.One is mutualism.Mutualism is reciprocally positive interactions between pairs of species (Bronstein 2009).For example, Termitomyces Heim (Lyophyllaceae, Agaricomycetes) can decompose plants to provide food for termites; in return, termites shelter Termitomyces from external threats (Da Costa et al. 2019).The other is parasitism.
Parasitism is the interaction between two species where one party (the parasite) benefits, while the other party (the host) suffers harm (Roper et al. 2019).As exemplified by species of Cordyceps Fr. sensu lato (s.l.), fungi parasitize invertebrates and eventually kill them.Invertebrate-pathogenic fungi are considered as the most well-known parasitic fungi (Araújo et al. 2018;Araújo et al. 2021;Wilson et al. 2021).They are ubiquitous inhabitants of forests worldwide, especially in tropical and subtropical regions.Invertebrate-pathogenic fungi are highly virulent and are known to have significant effects on host populations (Evans 1974).Cordyceps s. l. represents the most abundant and diverse group among invertebrate-pathogenic fungi (Araújo et al. 2021).Representatives of this group can colonize hosts in more than 10 invertebrate orders (Sanjuan et al. 2015;Araújo and Hughes 2016).They spread primarily through their hosts, evolving extensively in their morphologies and parasitic strategies.(Araújo and Hughes 2016).According to the current status of Cordyceps s. l. taxonomy, it belongs to four families: Clavicipitaceae, Cordycipitaceae, Ophiocordycipitaceae and Polycephalomycetaceae (Sung et al. 2007a;Xiao et al. 2023).Among them, the genus Ophiocordyceps Petch (Ophiocordycipitaceae) has received significant attention for its unique interactions with hosts and medical values (Zou et al. 2017;Araújo et al. 2018;Luangsa-ard et al. 2018;Khonsanit et al. 2019;Wang et al. 2021a;Zou et al. 2022;Tang et al. 2023a).
The majority of species in Ophiocordyceps exhibit clavate, entirely, or partially darkly pigmented stromata or synnemata, especially those species with a hirsutella-like anamorph, while some species possess brightly colored stromata with hymenostilbe-like anamorph.The stromata are mostly wiry, tough, leathery, and flexible.Perithecia are superficial to pseudo-immersed to fully immersed, and are vertically or obliquely inserted in the stromata.Asci are usually cylindrical with thickened apex and contain eight ascospores.Ascospores are typically cylindrical or clavate, multiseptate, either disarticulating into secondary spores or remaining whole after discharge (Sung et al. 2007a;Quandt et al. 2014;Luangsa-ard et al. 2018).Species in Ophiocordyceps mainly attack insects of Coleoptera, Diptera, Hemiptera, Hymenoptera, Lepidoptera, Odonata, and Orthoptera.Generally, they can attack all stages (larva, pupa, nymph, and adult) of the insects, with the majority of targets being larvae of Coleoptera and Lepidoptera living in wood or buried in soil (Sung et al. 2007a;Shrestha et al. 2016).Among species of Ophiocordyceps, only 17 species attack termites (Tasanathai et al. 2019;Araújo et al. 2021;Tang et al. 2022;Tasanathai et al. 2022;Xu et al. 2022).
Termites (Termitidae, Blattodea) are typically eusocial soil-dwelling insects, widely distributed around the world, especially in tropical and subtropical regions (Pearce 1997).Most termites are considered pests, causing significant impacts on forest ecosystems, and agricultural and forestry crops, with subterranean termites being particularly destructive (Rust and Su 2012;Scharf 2015).Some species of termite-pathogenic Ophiocordyceps have been regarded as potential biological agents to control termite populations (Rath 2000).
During surveys of invertebrate-pathogenic fungi in Yunnan Province, China, several specimens attacking termites were collected.Morphological and molecular evidence indicates that they belong to two different taxa distinct from previously described species.This study aims to introduce these two new species and discuss their evolutionary placement among related species.

Collection and isolation
Stromata emerging above fallen leaves were found in subtropical evergreen broad-leaved forests of Ruili City and Jinghong City, Yunnan Province, China.Specimens were documented and photographed in the field using a Canon 90D digital camera, and then each was placed in a sterilized 50 mL plastic centrifugal tube.All samples were stored in a cooler with ice packs until they were taken to the laboratory.Pure cultures were obtained on potato dextrose agar (PDA) with the composition of 200 g/L potato, 20 g/L dextrose, and 20 g/L agar, following the method previously presented (Wang et al. 2020).Subsequently, pure cultures were transferred to PDA slants and stored at the Kunming Institute of Botany Culture Collection (KUNCC), Chinese Academy of Sciences.Dried specimens were deposited in the Cryptogamic Herbarium of the Kunming Institute of Botany, Chinese Academy of Sciences (KUN-HKAS).

Morphological observations
The newly collected specimens were macroscopically examined with the Canon 750D camera and Olympus SZ60 stereo microscope.The characteristics of stromata (size, texture, shape, and color) were recorded.For the observation of teleomorph, perithecia were removed from the stromata and mounted on a glass slide with either 3% potassium hydroxide (KOH) (w/v) or 0.04% lactophenol cotton blue stain solution (w/v).Subsequently, the sizes and shapes of the perithecia, asci, and ascospores were measured under Olympus BX53 microscope.For each species, at least two specimens are measured, and each characteristic is measured at least 15 times repeatedly.The characteristics of pure cultures (size, texture, and color) were photographed using a Canon 750D camera after six weeks of culturing in an incubator at 25 °C.For the morphological description of anamorph, microscope slide cultures were prepared using the previous described method (Wang et al. 2020).Conidiogenous structures and conidia were measured and photographed using an Olympus BX53 microscope.

Sequencing alignments and phylogenetic analyses
We generated sequences for six loci from five specimens (Table 1).These were complemented with sequences of 125 related samples downloaded from NCBI GenBank based on BLAST searches and recent publications on Ophiocordycipitaceae (Tang et al. 2022;Xu et al. 2022).Tolypocladium inflatum Gams OSC 71235 and T. ophioglossoides (J.F.Gmel.)Quandt et al.CBS 100239 were selected as the outgroup.The sequence datasets were aligned using MAFFT v. 7, and alignments were manually corrected in MEGA v. 7.0 (Katoh and Standley 2013;Kumar et al. 2016).Ambiguously aligned sites were manually eliminated and gaps were regarded as missing data.ModelFinder (Kalyaanamoorthy et al.

Phylogenetic analyses
The    ,nrLSU,rpb1,rpb2,and ITS showing the relationship of two new species on termites from China with other Ophiocordyceps species.Values at the nodes before and after the backslash are BI posterior probabilities (BI-PP greater than 0.60) and ML bootstrap proportions (ML-BP greater than 70%), respectively.New species described in this paper are shown in bold red.
Habitat and ecology.Parasitic on termites buried in soil of the subtropical evergreen broad-leaved forests, emerging from fallen leaves on the forest floor.
Notes.Ophiocordyceps longistipes is characterized by solitary stromata, superficial and pyramidal to oval perithecia, filiform asci, and filiform ascospores, hirsutella-like anamorph with monophialidic or rarely polyphialidic, flask-shaped conidiogenous cells, and citriform or oval conidia embedded in a mucous sheath.Phylogenetically, all specimens of O. longistipes are clustered in the H. thompsonii subclade of Hirsutella lineages and form a monophyletic clade, which is placed sister to O. fusiformis with maximum support (Fig. 1).However, O. longistipes exhibits significant morphological differences from O. fusiformis in its both teleomorph and anamorph.For the teleomorph, O. longistipes produce longer stromata of 17-24 cm (up to 6 cm  3).
Phylogenetically, almost all Ophiocordyceps species parasitic on termites are placed in the H. thompsonii subclade, except for O. brunneirubra.Termite-pathogenic species exhibit significant morphological variation overall.Among these species, the length of stromata ranges from extremely short to very long, the existence pattern of perithecia from superficial to pseudo-immersed to immersed, and the size of perithecia ranges from about 100 to 600 µm (Tasanathai et al. 2019;Araújo et al. 2021;Tasanathai et al. 2022;Xu et al. 2022).However, some of these species exhibit minimal interspecific morphological variation, making it challenging to distinguish them only through morphological studies.Therefore, the use of molecular systematics is necessary to accurately identify these species.For example, O. asiatica and O. puluongensis, as well as O. khokpasiensis and O. termiticola, share similar morphological char-acteristics.Ophiocordyceps asiatica and O. puluongensis produce subglobose superficial perithecia, similar asci, ascospores, conidiogenous cells, and conidia (Tasanathai et al. 2019;Xu et al. 2022).Ophiocordyceps khokpasiensis and O. termiticola possess similar colored and shaped stromata, pseudo-immersed perithecia, and similar asci, ascospores, and conidiogenous cells (Tasanathai et al. 2019).Although these species are morphologically indistinguishable, phylogenetic analyses support them as separate taxa.
It's worth noting that the hosts of these termite-pathogenic Ophiocordyceps species are usually buried underground, typically 5 to 15 cm below the ground, which may be relevant to the subterranean living habits of the host termites (Martelossi et al. 2023).However, this can pose a challenge for species identification, as hosts are often lost due to separation from fragile stromata during excavation (Tasanathai et al. 2022).
Termites are notorious pests known for damaging wood, cultivated plants, buildings, pastures, forests, and even non-cellulosic materials like cables, causing annual economic losses amounting to tens of billions of dollars.Subterranean termites are responsible for about 80% of the total damage (Rust and Su 2012;Scharf 2015;Oi 2022).Therefore, the control of termites has become the focus of attention in various industries.Previously, many chlorinated hydrocarbon insecticides were used for termite control, but they were banned due to their irreversible environmental impact and negative effects on crop production.Consequently, environmentally friendly and sustainable control measures for termites are urgently needed.Entomopathogenic fungi may represent a potent solution (Afzal et al. 2019;Tasanathai et al. 2019;Oi 2022;Moon et al. 2023).These fungi, with strong infectivity, can continuously spread spores in the field to control pests and are considered environmentally non-polluting, so they have significant advantages in pest control (Shimazu et al. 1995;Meyling and Eilenberg 2007).Most members of H. thompsonii subclade have been found to obligately parasitize termites, they may have a regulatory effect on natural termite populations.Particularly, O. bispora, for which field investigations have revealed a high infection rate against termites, and laboratory experiments have also shown that it can effectively kill termites (Blackwell and Gilbertson 1984;Suh et al. 1998).Although laboratory experiments have not been conducted with O. longistipes, field observations have found that termites infected by this fungus often appear in groups.This may indicate that it has strong lethality against termites and possesses the potential to become a biological control agent for termites.

Figure 1 .
Figure1.Phylogenetic tree based on the combined dataset of nrSSU, nrLSU, tef-1α, rpb1, rpb2, and ITS showing the relationship of two new species on termites from China with other Ophiocordyceps species.Values at the nodes before and after the backslash are BI posterior probabilities (BI-PP greater than 0.60) and ML bootstrap proportions (ML-BP greater than 70%), respectively.New species described in this paper are shown in bold red.

Table 1 .
Voucher information and GenBank accession numbers for the sequences included in this study.

Voucher information GenBank accession no. Reference ITS nrSSU nrLSU tef-1α rpb1 rpb2
(Hoang et al. 2018)016)nalyses were performed on the concatenated data set.The BI analysis was conducted using the MrBayes v. 3.2(Ronquist et al. 2012).Four simultaneous Markov chains were run for 2,000,000 generations with a sub-sampling frequency every 100 generations.A burn-in of the first 25% of the total run was discarded.ML analysis was conducted using IQ-TREE v. 2.1.3(Nguyenetal.2015)underpartitionedmodels(Chernomoret al. 2016)with 1000 ultrafast bootstrap(Hoang et al. 2018).Trees were visualized with its Maximum-Likelihood bootstrap proportions (ML-BS) and Bayesian posterior probability (BI-PP) in FigTree v. 1.4.4 and edited with Adobe Illustrator CS6.0.

Table 2 .
Results of the best-ftting model for maximum likelihood (ML) and Bayesian inference (BI) for six loci partitions.

Table 3 .
Morphological comparison between Ophiocordyceps species parasitic on termites.