Skip to main content
Browse Subject Areas

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Isolation of strains and their genome sequencing to analyze the mating system of Ophiocordyceps robertsii

  • Melvin Xu,

    Roles Investigation, Methodology, Writing – original draft, Writing – review & editing

    Affiliation School of BioSciences, The University of Melbourne, Parkville, Victoria, Australia

  • Nathan A. Ashley,

    Roles Conceptualization, Funding acquisition, Resources, Writing – review & editing

    Affiliation GhostMothLabs, 20 Lynch Drive, Echuca, Victoria, Australia

  • Niloofar Vaghefi,

    Roles Data curation, Formal analysis, Investigation, Methodology, Writing – original draft, Writing – review & editing

    Affiliations Centre for Crop Health, University of Southern Queensland, Darling Heights, Queensland, Australia, School of Agriculture and Food, The University of Melbourne, Parkville, Victoria, Australia

  • Ian Wilkinson,

    Roles Conceptualization, Funding acquisition, Resources, Writing – review & editing

    Affiliation GhostMothLabs, 20 Lynch Drive, Echuca, Victoria, Australia

  • Alexander Idnurm

    Roles Conceptualization, Formal analysis, Funding acquisition, Investigation, Supervision, Writing – original draft, Writing – review & editing

    Affiliation School of BioSciences, The University of Melbourne, Parkville, Victoria, Australia


The fungal genus Ophiocordyceps contains a number of insect pathogens. One of the best known of these is Ophiocordyceps sinensis, which is used in Chinese medicine and its overharvesting threatens sustainability; hence, alternative species are being sought. Ophiocordyceps robertsii, found in Australia and New Zealand, has been proposed to be a close relative to O. sinensis, but little is known about this species despite being also of historical significance. Here, O. robertsii strains were isolated into culture and high coverage draft genome sequences obtained and analyzed. This species has a large genome expansion, as also occurred in O. sinensis. The mating type locus was characterized, indicating a heterothallic arrangement whereby each strain has an idiomorphic region of two (MAT1-2-1, MAT1-2-2) or three (MAT1-1-1, MAT1-1-2, MAT1-1-3) genes flanked by the conserved APN2 and SLA2 genes. These resources provide a new opportunity for understanding the evolution of the expanded genome in the homothallic species O. sinensis, as well as capabilities to explore the pharmaceutical potential in a species endemic to Australia and New Zealand.


Ophiocordyceps robertsii is an entomopathogenic fungus of historical note and of potential pharmaceutical value. O. robertsii is pathogenic to ghost moth larvae in the Hepialidae family, is endemic in Australia and New Zealand, and was the first fungus described from New Zealand by Europeans [1]. When assigned its Linnaean name, the fungus was already well known to the Māori culture, who used āwheto as a source of ink for their ritual tattoos [2].

O. robertsii was proposed as the closest relative of Ophiocordyceps sinensis [3], which is known as dōng chóng xià cǎo (‘summer grass, winter worm’) and is China’s, if not the world’s, most valued medicinal fungus. On parity with the cost of gold, this is the most valuable fungal fruiting body by weight, well over that of white truffles [4]. With its limited geographic distribution on the Qinghai-Tibetan Plateau, O. sinensis harvesting has forced extensive pressures into this sensitive region and on this species. Producing this fungus in cultivation may protect its natural habitat [5, 6]. Alternatively, closely related species may also provide a new resource for use in medicine and/or as origins for new pharmaceutical chemicals.

The biology, population structure, diversity and relationships with other species of O. robertsii are unknown. In particular, the life cycle of O. robertsii is uncharacterized. If to be used as a source of beneficial chemicals, e.g. as reported anticancer agents [7], that would ideally be isolated from in vitro cultures, it is essential to determine the mating system especially as some fungal secondary metabolites may only be produced at certain stages of the life cycle. When comparing O. robertsii with O. sinensis, a number of questions emerge. The genome of O. sinensis shows an expansion in size relative to other members of the Ophiocordycipitaceae family and the species has a homothallic life style [8, 9]: how did this arise? Which species, as currently known, is the closest relative to O. sinensis, and could that species be informative for a better understanding of the evolution of species within the Ophiocordyceps genus? With these questions in mind, the primary aim of this research was to define the mating system of O. robertsii, by developing genomic resources that could also contribute further to understanding the evolution of entomopathogenic fungi.

Materials and methods

Fungal material

A fruiting body of O. robertsii and associated parasitized larva was collected in 2018 near Queenstown, Tasmania (under Tasmanian Department of Primary Industries, Parks, Water and Environment permit FL 18158). A second fruiting body was donated by a member of the public, having been collected in an undescribed location in southeastern Victoria (received under Victorian Department of Environment, Land, Water and Planning permit 10008557). The fruiting bodies were pressed onto potato dextrose agar (PDA) medium supplemented with chloramphenicol and cefotaxime (30 μg/ml and 100 μg/ml, respectively) to reduce bacterial growth, to release ascospores. Colonies derived from single ascospores were excised with a scalpel and transferred to new PDA plates.

Insect identification

Genomic DNA was extracted from the posterior end of the larva from Tasmania, and a fragment of the mitochondrial COX1 gene was amplified with primers modified from Lco1490 (5´-GGTCAACAAATCATAAAGATATTG-3´) and Hco2198 (5´-TAAACTTCAGGGTGACCAAAAAAT-3´) [10]. The amplicon was cloned with the pCR2.1-TOPO TA cloning kit (Invitrogen), the reaction transformed into chemically-competent Escherichia coli cells, and colonies selected on Luria-Bertani medium containing kanamycin (50 μg/ml). Plasmids were extracted from multiple E. coli cultures, and then the inserts sequenced using the universal M13F and M13R primers with Sanger chemistry at the Australian Genome Research Facility (AGRF).

Fungal DNA extraction and genome sequencing

Mycelium was cultured in potato dextrose broth, and genomic DNA extracted from freeze-dried mycelium. DNA from strain UoM1 was sequenced using Illumina paired reads of 125 nucleotides on a HiSeq 2500 machine by the AGRF. DNA from strain UoM4 was sequenced using Illumina paired reads of 150 nucleotides using a NovaSeq 6000 machine at the Murdoch Children’s Research Institute.

Genome assemblies

A preliminary genome assembly of strain UoM1 was created from Illumina sequencing reads in Geneious version 8.1.9, using Velvet [11]. A k-mer of 71 was selected based on Velvet optimization aimed at obtaining longest contigs. Gene predictions were made using Augustus version 3.3.3 with the parameters based on Neurospora crassa or Aspergillus nidulans [12].

A second round of genome assembly was undertaken, as follows. Quality filtering and adapter removal from raw Illumina reads was conducted using BBduk (k = 23, trimq = 15, minlen = 45) from the BBmap suite v.36.86 [13]. Two approaches were adopted to produce draft genome assemblies of strains UoM1 and UoM4. First, Unicycler v.0.4.8 [14] was used, which is a SPAdes-optimiser assembler, with automatically selected k-mer values. A second de novo assembly was generated for both strains using SOAPdenovo2 v.2.40 [15] (SOAPdenovo-127mer) testing the k-mer length parameter between 71 and 91, in 2 bases increments. Assembled genomes were screened and filtered for potential bacterial contamination using Kraken v.2.1.1 [16]. The completeness of the genome assemblies was evaluated through identification of Benchmarking Universal Single-Copy Orthologs vis BUSCO v.1.2 [17]. Genome statistics were compared using Quast v.2.0.5 [18]. Based on the results from BUSCO analysis, genomes produced by Unicycler were selected for further analyses. OcculterCut v.1.1 [19] was used to scan the generated genomes to determine their percent GC content and distribution. The analysis was also conducted on the genome of O. sinensis strain IOZ07 (GenBank accession GCA_012934285.1 [20]) to compare the GC content distribution of the O. robertsii genomes produced here.

To be able to conduct evidence-based annotation of the genomes, RNA sequencing data from O. sinensis was obtained from the NCBI SRA database (PRJNA687052, PRJNA673413 and PRJNA673413) and aligned against the genome of strain UoM4 using HISAT2 V.2.2.1 [21], and those reads that aligned to the genome were further extracted to produce a transcriptome assembly using Trinity v.2.10.0 [22]. Repeat libraries for both genomes were generated using RepeatModeler v.2.0.1 [23] to allow for repeat masking prior to annotation. Genome annotations were conducted using Maker v.2.31.9 [24]. The resulting annotations from the first round were further used to produce hidden Markov model (HMM) profiles for the genomes, which was further refined with a second round of SNAP training and used for the final annotation [24].

Phylogenetic analyses

The internal transcribed spacer (ITS) regions were amplified from O. robertsii genomic DNA using primers ITS1 (5´-TCCGTAGGTGAACCTGCGG-3´) and ITS4 (5´-TCCTCCGCTTATTGATATGC-3´) [25], and sequenced using Sanger chemistry at the AGRF. Sequences from Ophiocordyceps species identified from the literature or from BLAST comparisons with O. robertsii sequences were downloaded from GenBank (accessions are provided in S1 Table); note that only relatively close relatives were selected to enhance subsequent phylogenetic resolution. O. agriotidis was selected as an outgroup. O. robertsii strain UoM1 sequences, other than for the ITS regions, were obtained from the next generation sequencing reads. The sequences were aligned in clustalW, imported into MEGA X [26], and alignments adjusted by eye. Phylogenetic trees were created using maximum likelihood with support for relationships tested from 1000 bootstrap replicates.

For Bayesian analysis, the best nucleotide substitution model for individual loci were determined using PAUP v.4.0 [27] and MrModeltest v. 2.3 [28], to be (GTR+I+G) for TEF, LSU and RBP2; (GTR+I) for SSU; and (GTR+I) for RBP1. Two MCMC chains were run for each individual locus and the concatenated alignment using MrBayes v. 3.2.4 [29], saving one tree per 100 generations. The run ended automatically when the standard deviation of split frequencies reached below 0.01. The 50% majority rule consensus trees were generated after 25% burn-in of the saved trees.

Targeted sequence assembly and analysis of the mating type (MAT) locus

Defining the mating type (MAT) locus sequences used a combination of PCR and Sanger chemistry sequencing and next generation read assembly. First, inverse PCR was used to establish the presence of a second mating type in strains of O. robertsii after the identification of the putative MAT locus in strain UoM1. Genomic DNA of strain UoM4 was digested with restriction enzymes (EcoRI, KpnI and NcoI) then circularized using T4 DNA ligase, and this template used for PCR amplification with primers MAI0553 (5´-GCCAGAGGACTCTGCAGG-3´) and MAI0562 (5´-CCAGTGGTGGTTGTGCAGAC-3´). This established a region of DNA in strain UoM4 that was not present in the genome assembly of strain UoM1, and became a starting point for focused PCRs. Second, the genome of UoM4 was sequenced, and a targeted assembly generated based on partial sequences already obtained from PCR and sequencing and by mapping to strain UoM1. Third, the genes adjacent to the MAT locus and within each idiomorph were predicted using BLAST comparisons to GenBank.

Once the MAT locus idiomorph sequences were resolved, to test the proportion of mating types in the 11 strains derived from the sample from Tasmania in 2018 and the four strains derived from the sample from Victoria in 2019, a region of the MAT1-1-1 gene was amplified with primers MAI0663 (5´-CTTCTGAACTTCCTCACG-3´) and MAI0664 (5´-GCAAAGTGAAAGTCGTCC-3´) and from the MAT1-2-1 gene with primers MAI0665 (5´-AATCCCAATCCTCAGTGG-3´) and MAI0666 (5´-CGCAGTGTTGGAATCACC-3´). Amplicons were resolved on 1% agarose gels stained with ethidium bromide.

Results and discussion

Isolation of Ophiocordyceps robertsii strains into culture

Ascospores from an O. robertsii fruiting body collected from Tasmania in 2018 were plated onto PDA medium, and 11 strains isolated. The strains grew slowly in culture, reaching a colony diameter of about 2 cm after a month (Fig 1A). Cultures also produced fruiting bodies, but in a stochastic and irregular manner (Fig 1B). A number of attempts and growth conditions were used to optimize the induction of this fruiting process. The most reliable conditions observed, to date, are up to eight months incubation in a starting volume of 25 ml potato dextrose broth in a 20 cm high petri dish, in darkness at 14°C. All 11 strains were able to produce these structures, but they did not progress further towards the generation of sexual perithecia. O. xuefengensis, a heterothallic close relative of O. robertsii, also makes asexual stromata; the formation of these structures also have a preference for darkness [30].

Fig 1. Properties of O. robertsii in culture.

A. Radial vegetative growth on PDA after 32 days. Bar = 1 cm. B. Production of asexual stromata after ~ 3 months. Bar = 1 cm.

To ensure the growth of these isolates was representative of the species, a second fruiting body collected in Victoria in 2019 was used to isolate four additional strains. These four strains also grew slowly in culture, and this trait in strains from both sources is consistent with recent observations of the growth rate of O. robertsii strains isolated from New Zealand [31].

The sample from Tasmania included the parasitized insect larva, and its identification was sought by amplification, cloning and sequencing a diagnostic fragment of the COX1 gene (sequence submitted to GenBank as accession OP376821). Comparison with the GenBank nr database revealed >99% identity to species in the ghost moth genus Oxycanus. The species is unknown, due to ambiguities in species identification from which DNA samples were sequenced [32, 33].

Draft genome assemblies of O. robertsii indicate a larger size compared to other Ophiocordycipitaceae species

The genomic DNA samples of two O. robertsii strains (UoM1 and UoM4) were sequencing using short-read Illumina technology (deposited as GenBank PRJNA630494). Strain UoM1 was sequenced first, and a draft genome obtained after a Velvet k-mer optimization process designed to optimize maximum contig length. Using a k-mer option of 71 and a minimum cut off contig length of 1 kb, 12,910 contigs were assembled for a total genome size of 97.8 Mb. Analysis of this assembly using Augustus predicted 10,818 or 11,280 genes (N. crassa and A. nidulans parameters, respectively).

Genome assemblies constructed using Unicycler were found to have no gaps (represented by N’s in the assemblies; S2 Table) and resulted in more complete genomes compared to those generated with SOAPdenovo2, based on sordariomycetes_odb10 BUSCO results (97% completeness for UoM1 and 97.2% completeness for UoM4) (S3 Table), and therefore were selected for subsequent genome annotation and OcculterCut analyses. The differences in genome predictions likely reflect the differences in the underlying assembly parameters. Unicycler was originally developed for bacterial genome assembly [14], hence, may outperform SOAPdenovo in haploid genome assembly. Unicycler assembly for strain UoM1 consisted of 10,216 contigs and a total size of 95.8 Mb, while the assembly for strain UoM4 included 5,720 contigs with a total genome size of 102.7 Mb. These assemblies are in GenBank as BioProject PRJNA630494 (JAPEBV000000000 for UoM1 and JAPEBW000000000 for UoM4) and the annotations are in Figshare ( Analyses of the assemblies using evidenced-based annotation in Maker predicted 8,904 and 8,934 genes for UoM1 and UoM4, respectively.

The distribution of genes within the O. robertsii genome assemblies is uneven, with many of the largest contigs having no genes predicted to be present on them, and a pattern that is consistent with previous observations of large genome expansions in the Ascomycota not correlating with increased gene content [34]. This instead correlates with expanded large regions with high AT content, with contigs being unevenly distributed in AT vs. GC content, and correspondingly fewer genes predicted in the AT-rich regions (Fig 2). OcculterCut results showed that the GC content distribution in both genomes was bimodal, similar to that of O. sinensis strain IOZ07 (Fig 2A–2C). However, the proportion of AT-rich regions in O. sinensis strain IOZ07 (78%) was higher compared to that of UoM1 (70%) and UoM4 (72%). OcculterCut analysis also found gene density in the AT-rich regions of UoM1 and UoM4 was 0.148 and 0.081 genes per Mb, respectively, while gene density in the GC-balanced regions of these strains was 312 and 313 genes per Mb (example shown in Fig 2D). This ‘bipartite’ structure is reminiscent of the genomes of a number of fungi often associated with plants, such as Leptosphaeria maculans, a plant pathogen in which this genome structure was first observed [35]. The structure is attributed to repetitive regions, usually transposable elements, being targeted by repeat induced point (RIP) mutation, converting cytosine to thymine and effectively raising the AT content. RIP mutation is predicted to occur in many fungi, and has been experimentally demonstrated in the classes Sordariomycetes, Dothideomycetes and Eurotiomycetes [3638]. RIP occurs in Trichoderma reesei [39], which is in the same order (Hypocreales) as O. robertsii.

Fig 2. The GC content distribution of O. robertsii strains in bimodal.

A. Strain UoM1 and B. strain UoM4 show a similar structure to that of C. O. sinensis strain IOZ07. Vertical blue lines show the GC cut-off points selected by OcculterCut [19] to classify genome segments into distinct AT-rich and GC-balanced regions. D. An example of an O. robertsii contig that illustrates its bipartite genome, comprising AT-rich and gene poor regions vs. GC-rich and gene dense regions. AT and GC content is measured using 100 bp sliding windows, with two genes marked with the arrows. The homologous region from O. sinensis strain IOZ07 is shown for comparison; note that the AT-rich regions extend for more than 30 kb and 80 kb on either side of the two genes. BLASTx of the surrounding region returns hits in GenBank to retrotransposable elements; in the two species they are inactive as feature numerous stop codons.

The genome sequence of O. robertsii appears to be considerably larger than other species in the Ophiocordycipitaceae family, which typically range between 21.9 to 49.3 Mb [40, 41], but is similar to that of the far larger O. sinensis. There have now been at least five genome sequencing attempts on O. sinensis towards complete coverage and assembly, with the best current estimate of a genome sized at 110–120 Mb [9, 20, 4244]. As a consequence of the well-established challenge to sequence O. sinensis, attention was not placed on trying to optimize further the genome assembly parameters for the Illumina short reads of O. robertsii, as a better understanding of the genome structure, genome expansion, gene content, and gene arrangements relative to O. sinensis will require generation of a high quality version of the genome by incorporating long read sequencing technologies coupled to RNA sequencing to aid in gene annotation. However, some points are of relevance in comparing the genomes of O. robertsii and O. sinensis. One is that more than two thirds of the O. sinensis genome is comprised of repetitive elements, and the species is proposed to have an active RIP mutation system, and, unlike in other fungal species, is hypothesized to have targeted the DNA repeats for the ribosomal RNA molecules for RIP, which has led to difficulties using the ITS regions for phylogenetic studies [42, 45, 46].

It has been hypothesized that the expansion in genome size in O. sinensis occurred during the rising in altitude of the Tibetan Plateau, and that the genome expansion contributed to its adaptation to this environment [9, 42]. However, O. robertsii has a similar genome structure, and one set of the strains were isolated from Queenstown, which is 129 m above sea level. Furthermore, there are just under 500 reports of O. robertsii in the Atlas of Living Australia (, a citizen-scientist database, all from southeastern Australia and New Zealand locations that are well below the altitude of the Tibetan Plateau.

Phylogenetic analyses place O. robertsii as one of the close relatives of the medicinal fungus O. sinensis

Sequences of six DNA regions were obtained from GenBank for the closest relatives of O. robertsii, such as O. sinensis, O. xuefengensis [47], O. macroacicularis [48], O. lanpingensis [49], O. karstii [50] and others, and then compared with those same regions from O. robertsii by alignment and the construction of phylogenetic trees.

The six trees show limited concordance between them for the relationships between the species and often low bootstrap support or Bayesian posterior probabilities (S1 Fig), which has been observed previously [see for example the phylogeny of Ophiocordyceps by [51]]. Concatenation of five of the regions (i.e. excluding the ITS region) yielded a tree with increased support at many nodes (Fig 3). Different sets of genes, or potentially whole genome comparisons, are required if using DNA sequencing information to provide stronger relationship indicators between Ophiocordyceps species.

Fig 3. Phylogenetic tree of the relationships between Ophiocordyceps species closely related to O. robertsii and O. sinensis generated using maximum likelihood.

H is the abbreviation for Hirsutella. O. agriotidis was used as the outgroup. Concatenation of DNA sequences of regions of SSU, LSU, TEF1α, RBP1 and RBP2. Numbers adjacent to nodes indicate bootstrap % from 1000 replicates if above 70% / Bayesian posterior probability if above 0.95. The General Time Reversal +G model was used.

Several features emerge from the phylogenetic analyses. The first of these is that O. robertsii is not the closest relative to O. sinensis, at least based on the phylogenetic trees generated. The strains isolated in this work have identical ITS sequences (the GenBank accession for the strains derived from Victoria is deposited as OP326271) and are very similar to O. robertsii collected in 1994 from Australia, while YHORZT007 from China [49] is misidentified as O. robertsii and instead may represent a novel species. The ITS sequences in GenBank from O. robertsii strains from New Zealand share less than 97% identity with Australian strains, which is potentially an indication that specimens classified as O. robertsii may be two distinct species in the two countries. Another Ophiocordyceps species that is entomopathogenic and endemic to New Zealand is Cordyceps hauturu [52]. While similar to O. robertsii in physical appearance, the one ITS sequence available from C. hauturu is only 91% similar to that of O. robertsii and therefore is a distinct species. One final point is that O. stylophora and O. appendiculata are likely the same species.

Analysis of the predicted mating type (MAT) locus of O. robertsii supports a heterothallic reproductive strategy

The genome sequencing assembly of strain UoM1 was searched, using BLAST, for homologs of the mating type genes or two flanking genes that are adjacent to the MAT locus in ascomycetes. One region was identified, and initial PCR analysis suggested that this was found in only a subset of the O. robertsii strains. To obtain additional MAT sequences, inverse PCR was used to identify the idiomorphic region in strain UoM4. Due to challenges and limited efficiency in this method, strain UoM4 was then also sequenced and assembled, with attention paid to assembling the MAT locus by a combination of reiterative targeted assemblies of short reads coupled to PCRs and sequencing of those amplicons. The DNA sequence was completed except for one region that could be amplified by PCR, but could not be fully resolved by sequencing off those amplicons, in one of the strains (this gap is marked in Fig 4A). The sequences of the two MAT loci have been deposited to GenBank as accessions MT436030 and MT436031.

Fig 4. DNA sequence analysis suggests that O. robertsii has a heterothallic life cycle.

A. Diagram of the MAT locus from two strains of O. robertsii. The genes in white indicate those found adjacent to MAT loci in other fungi. DNA sequence similarity is low in the idiomorph region (grey box), in which one strain encodes three predicted genes and other encodes two predicted genes. The gap indicates a region of low sequence coverage. B. Agarose gels resolving PCR amplification products using primers specific to either MAT1-1-1 or MAT1-2-1 from 11 isolates of O. robertsii. M indicates the DNA ladder; H2O indicates no DNA was added to the PCRs. Original gel images are provided as S2 Fig.

Comparison of the two O. robertsii strains indicates that each contains an idiomorphic DNA region that has minimal similarity between each other (Fig 4A), while encoding proteins commonly done so by genes within the MAT loci of other fungi (S3 Fig) [53, 54]. The MAT1-1 idiomorphic region of 4.5 kb contains three predicted genes similar to those found in MAT loci of other fungi. The MAT1-2 idiomorphic region of 3.9 kb contains two predicted genes. Comparison of the predicted proteins within the MAT locus by BLAST against GenBank returned highest matches to O. sinensis and O. xuefengensis, which is consistent with the inter-species relationships observed using other genes (Fig 3).

MAT1-1-1 encodes the ‘MATα’ high mobility group (HMG)-box protein that is a fundamental determinant of mating type (e.g. as α in Saccharomyces cerevisiae). MAT1-1-2 is commonly found in MAT loci in the Sordariomycetes: deletion of the homolog in Cordyceps militaris results in a strain able to form fruiting bodies but with sterile perithecia [55]. MAT1-1-3 encodes another HMG-domain protein (MATA-type).

The MAT1-2-1 gene encodes the HMG-domain protein (MATA-type) found in ascomycete MAT loci. The MAT1-2-2 gene encodes a protein of unknown function that is not commonly associated with MAT loci in ascomycete species. The sequence available for the MAT1-2 idiomorph of O. xuefengensis is likely truncated [56]; nonetheless it includes sequence of the likely the start of the MAT1-2-2 homolog (S3 Fig). Furthermore, BLAST of genome sequences available for other Ophiocordyceps species and analysis of the flanking genes indicates that MAT1-2-2 is also found in the predicted MAT locus of O. sinensis (at least in strain CO18), O. australis and O. camponoti-rufipedis (S3 Fig).

The predicted amino acid sequences of the O. robertsii and O. sinensis homologs of the MAT-encoded proteins were aligned with each other and related Ophiocordycipitaceae species (S3 Fig). These sequences show a high level of identify, further supporting the prediction that these are responsible for mating in O. robertsii.

Fragments of the MAT regions were amplified from the 11 strains isolated from the fruiting body collected in Tasmania in 2018. Each contained only one of the regions (Fig 4B). Of the four strains isolated from the fruiting body from Victoria in 2019, two were MAT1-1 and two were MAT1-2. Hence, the DNA sequences, similarities in predicted proteins and idiomorph distribution are consistent with O. robertsii having a heterothallic lifestyle.

During our research, the nature of the MAT locus of O. xuefengensis was explored indicating that this species is also heterothallic [56, 57]. The sexual reproduction system in O. xuefengensis has been characterized, and sheds light on the fruiting structures formed by single strains of O. robertsii. That is, O. xuefengensis and presumably O. robertsii produce an asexual stroma that is subsequently fertilized by the opposite mating type. With slow growth in culture, completion of the O. robertsii life cycle in vitro is yet to be achieved. The formation of the asexual stromata was difficult to achieve reliably in culture, and it may be other growth conditions, or infection of insects, would be needed to complete the sexual cycle. The resources developed here, including genome assemblies and annotations that are available from GenBank or Figshare and strains, one of each mating type that have been deposited in the Queensland Herbarium culture collection (BRIP 75182 a and BRIP 75183 a), will contribute to this endeavor.


Ophiocordyces robertsii is endemic to Australia and New Zealand, and a notable species in New Zealand’s culture. While this research focused on the cultivation of the species, for the prospective benefits such cultures could bring, it also has ramifications about the evolution and biology of O. sinensis, which is one of the world’s most expensive fungi. O. sinensis has been difficult to cultivate ex situ due to its niche geographical range, and only a handful of companies have commercial production [5, 6]. This research established that O. robertsii could serve for comparative analysis with O. sinensis and other Ophiocordyceps species. The converse is that the medicinal benefits associated with O. sinensis are underexplored in O. robertsii. Last, it is currently unclear if O. robertsii in Australia and New Zealand represents one or multiple species.

Supporting information

S1 Fig. Phylogenetic trees of the relationships between Ophiocordyceps species closely related to O. robertsii and O. sinensis generated using maximum likelihood for six DNA regions.

H is the abbreviation for Hirsutella and in the ITS tree C is the abbreviation for Cordyceps. O. agriotidis was used as the outgroup. Numbers adjacent to nodes indicate bootstrap % from 1000 replicates if above 70% / Bayesian posterior probability if above 0.95. Models used were Kimura-2 +G (SSU), Tamura-Nei +G (RBP2, TEF1), Hasegawa-Kishino-Yano +G (ITS, LSU) and Tamura-3 (RBP1).


S2 Fig. Original agarose gel images used in Fig 4B.

A. MAT1-1-1. B. MAT1-2-1. Both gels are loaded in a similar manner as follows. Nothing was loaded in wells 1, 16 and 17. A DNA ladder (sizes provided on the right) was loaded in wells 2 and 15. Amplification of progeny DNA was loaded into wells 3–14, with the PCR water control loaded in well 14. * indicates an additional amplification product. Images captured 27 April, 2020 using a Bio-Rad® gel documentation system.


S3 Fig. Alignment of predicted amino acid sequences from O. roberstii with those encoded in candidate MAT loci of Ophiocordycipitaceae species.

The two strains of O. robertsii encode five predicted proteins. Best matches in GenBank or of relevance for genome arrangement are illustrated. H = Hirsutella, T = Tolypocladium and camp.-rufi. = camponoti-rufipedis. The … symbol indicates missing sequence information.


S1 Table. GenBank accessions of the DNA sequences used to create phylogenetic trees.

Grey shading indicates no sequence data are available.


S2 Table. Genome statistics for de novo assemblies of Ophiocordyceps robertsii strains UoM1 and UoM4 produced in this study estimated using Quast v.2.0.5 and genome completeness for each of the assemblies estimated using BUSCO v.1.2.


S3 Table. Genome completeness for de novo assemblies of Ophiocordyceps robertsii strains UoM1 and UoM4 produced in this study estimated using BUSCO v.1.2 based on sordariomycetes_odb10 database.



We thank Christine Wilson for finding samples of O. robertsii.


  1. 1. Hooker WJ. Icones Plantarum. London: Longman, Rees, Orme, Brown, Green & Longman; 1837.
  2. 2. Best E. The Uhi-Maori, or native tattooing instruments. J Polynesian Soc. 1904; 13:166–72.
  3. 3. Sung G-H, Hywel-Jones NL, Sung J-M, Luangsa-ard JJ, Shrestha B, Spatafora JW. Phylogenetic classification of Cordyceps and the clavicipitaceous fungi. Studies Mycology. 2007; 57:5–59.
  4. 4. Zhang Y, Li E, Wang C, Li Y, Liu X. Ophiocordyceps sinensis, the flagship fungus of China: terminology, life strategy and ecology. Mycology. 2012; 3:2–10.
  5. 5. Qin Q-l, Zhou G-l, Zhang H, Meng Q, Zhnag J-h, Wang H-t, et al. Obstacles and approaches in artificial cultivation of Chinese cordyceps. Mycology. 2018; 9:7–9. pmid:30123655
  6. 6. Li X, Liu Q, Li W, Li Q, Qian Z, Liu X, et al. A breakthrough in the artificial cultivation of Chinese cordyceps on a large-scale and its impact on science, the economy, and industry. Critical Rev Biotechnol. 2019; 39:181–91. pmid:30394122
  7. 7. Beattie KD, Ulrich R, Grice ID, Uddin SJ, Blake TB, Wood KA, et al. Ethanolic and aqueous extracts derived from Australian fungi inhibit cancer cell growth in vitro. Mycologia. 2011; 103:458–65. pmid:21262985
  8. 8. Bushley KE, Li Y, Wang W-J, Wang X-L, Jiao L, Spatafora JW, et al. Isolation of the MAT1-1 mating type idiomorph and evidence for selfing in the Chinese medicinal fungus Ophiocordyceps sinensis. Fungal Biol. 2013; 117:599–610.
  9. 9. Xia E-H, Yang D-R, Jiang J-J, Zhang Q-J, Liu Y, Liu Y-L, et al. The caterpillar fungus, Ophiocordyceps sinensis, genome provides insights into highland adaptation of fungal pathogenicity. Sci Rep. 2017; 7:1806.
  10. 10. Folmer O, Black M, Hoeh W, Lutz R, Vrijenhoek R. DNA primers for amplification of mitochondrial cytochrome c oxidase subunit I from diverse metazoan invertebrates. Mol Mar Biol Biotechnol. 1994; 3:294–9.
  11. 11. Kearse M, Moir R, Wilson A, Stones-Havas S, Cheung M, Sturrock S, et al. Geneious Basic: an integrated and extendable desktop software platform for the organization and analysis of sequence data. Bioinformatics. 2012; 28:1647–9. pmid:22543367
  12. 12. Stanke M, Diekhans M, Baertsch R, Haussler D. Using native and syntenically mapped cDNA alignments to improve de novo gene finding. Bioinformatics. 2008; 24:637–44. pmid:18218656
  13. 13. Bushnell B. BBMap: a fast, accurate, splice-aware aligner (No. LBNL-7065E). Berkeley, CA, USA: Lawrence Berkeley National Lab (LBNL); 2014.
  14. 14. Wick RR, Judd LM, Gorrie CL, Holt KE. Unicycler: resolving bacterial genome assemblies from short and long sequencing reads. PLoS Comput Biol. 2017; 13:e1005595. pmid:28594827
  15. 15. Li R, Li Y, Kristiansen K, Wang J. SOAP: short oligonucleotide alignment program. Bioinformatics. 2008; 24:713–4. pmid:18227114
  16. 16. Wood DE, Salzberg SL. Kraken: ultrafast metagenomic sequence classification using exact alignments. Genome Biol. 2014; 15:R46. pmid:24580807
  17. 17. Simão FA, Waterhouse RM, Ioannidis P, Kriventseva EV, Zdobnov EM. BUSCO: assessing genome assembly and annotation completeness with single-copy orthologs. Bioinformatics. 2015; 31:3210–2. pmid:26059717
  18. 18. Gurevich A, Saveliev V, Vyahhi N, Tesler G. QUAST: quality assessment tool for genome assemblies. Bioinformatics. 2013; 29:1072–5. pmid:23422339
  19. 19. Testa AC, Oliver RP, Hane JK. OcculterCut: a comprehensive survey of AT-rich regions in fungal genomes. Genome Biol Evol. 2016; 8:2044–64. pmid:27289099
  20. 20. Shu R, Zhang J, Meng Q, Zhang H, Zhou G, Li M, et al. A new high-quality draft genome assembly of the Chinese Cordyceps Ophiocordyceps sinensis. Genome Biol Evol. 2020; 12:1074–9.
  21. 21. Kim D, Paggi JM, Park C, Bennett C, Salzberg SL. Graph-based genome alignment and genotyping with HISAT2 and HISAT-genotype. Nat Biotechnol. 2019; 37:907–15. pmid:31375807
  22. 22. Grabherr MG, Haas BJ, Yassour M, Levin JZ, Thompson DA, Amit I, et al. Full-length transcriptome assembly from RNA-Seq data without a reference genome. Nat Biotechnol. 2011; 29:644–52. pmid:21572440
  23. 23. Smit AFA, Hubley, R. RepeatModeler Open-1.0. 2008–2015 2008.
  24. 24. Cantarel BL, Korf I, Robb SMC, Parra G, Ross E, Moore B, et al. MAKER: an easy-to-use annotation pipeline designed for emerging model organism genomes. Genome Res. 2008; 18:188–96. pmid:18025269
  25. 25. White TJ, Bruns T, Lee S, Taylor J. Amplification and direct sequencing of fungal ribosomal RNA genes for phylogenetics. PCR Protocols: a guide to methods and applications: Academic Press, Inc.; 1990. p. 315–22.
  26. 26. Kumar S, Stecher G, Tamura K. MEGA7: Molecular Evolutionary Genetics Analysis version 7.0 for bigger datasets. Mol Biol Evol. 2016; 33:1870–4. pmid:27004904
  27. 27. Swofford DL. PAUP Phylogenetic Analysis Using Parsimony (and other methods). Version 4. Sunderland, MA, USA: Sinauer Associates; 2003.
  28. 28. Nylander JAA. MrModeltest v2. Program distributed by the author. Sweden: Evolutionary Biology Centre, Uppsala University; 2004.
  29. 29. Ronquist F, Teslenko M, van der Mark P, Ayres DL, Darling A, Höhna S, et al. MrBayes 3.2: efficient Bayesian phylogenetic inference and model choice across a large model space. Syst Biol. 2012; 61:539–42. pmid:22357727
  30. 30. Zhong C, Zhou R, Jin J, Liu H, Xie J, Zhen L, et al. Cloning and analysis of the photoreceptor genes Oxwc-1 and Oxwc-2 from ethnopharmacological fungus Ophiocordyceps xuefengensis. Mycoscience. 2021; 62:189–97.
  31. 31. Casonato SG, Hill MG. Effects of amendments to agar media on the in vitro growth of the entomopathogenic fungus Ophiocordyceps robertsii (Hook.) G.H. Sung, J.M. Sung, Hywel-Jones & Spatafora (Hypocreales: ophiocordycipitaceae). NZ J Bot. 2022; 60:107–12.
  32. 32. Beaver EP, Moore MD, Velasco-Castrillón A, Stevens MI. Three new ghost moths of the genus Oxycanus Walker, 1856 from Australia (Lepidoptera: Hepialidae). Zootaxa. 2020; 4732:351–74.
  33. 33. Hebert PDN, deWaard JR, Zakharov EV, Prosser SW, Sones JE, McKeown JTA, et al. A DNA ’barcode blitz’: rapid digitization and sequencing of a natural history collection. PLoS One. 2013; 8:e68535. pmid:23874660
  34. 34. Kelkar YD, Ochman H. Causes and consequences of genome expansion in fungi. Genome Biol Evol. 2012; 4:13–23. pmid:22117086
  35. 35. Rouxel T, Grandaubert J, Hane JK, Hoede C, van de Wouw AP, Couloux A, et al. Effector diversification within compartments of the Leptosphaeria maculans genome affected by repeat-induced point mutations. Nat Commun. 2011; 2:202.
  36. 36. Cambareri EB, Jensen BC, Schabtach E, Selker EU. Repeat-induced G-C to A-T mutations in Neurospora. Science. 1989; 244:1571–5.
  37. 37. Urquhart AS, Mondo SJ, Mäkelä MR, Hane JK, Wiebenga A, He G, et al. Genomic and genetic insights into a cosmopolitan fungus, Paecilomyces variotii (Eurotiales). Front Microbiol. 2018; 9:3058.
  38. 38. Van de Wouw AP, Elliott CE, Popa KM, Idnurm A. Analysis of repeat induced point (RIP) mutations in Leptosphaeria maculans indicates variability in the RIP process between fungal species. Genetics. 2019; 211:89–104.
  39. 39. Li W-C, Huang C-H, Chen C-L, Chuang Y-C, Tung S-Y, Wang T-F. Trichoderma reesei complete genome sequence, repeat-induced point mutation, and partitioning of CAZyme gene clusters. Biotechnol Biofuels. 2017; 10:170.
  40. 40. Kobmoo N, Wichadakul D, Arnamnart N, Rodríguez De La Vega RC, Luangsa-ard JJ, Giraud T. A genome scan of diversifying selection in Ophiocordyceps zombie-ant fungi suggests a role for enterotoxins in co-evolution and host specificity. Mol Ecol. 2018; 27:3582–98.
  41. 41. de Bekker C, Ohm RA, Evans HC, Brachmann A, Hughes DP. Ant-infecting Ophiocordyceps genomes reveal a high diversity of potential behavioral manipulation genes and a possible major role for enterotoxins. Sci Rep. 2017; 7:12508.
  42. 42. Liu J, Guo L, Li Z, Zhou Z, Li Z, Li Q, et al. Genomic analyses reveal evolutionary and geologic context for the plateau fungus Ophiocordyceps sinensis. Chinese Medicine. 2020; 15:107.
  43. 43. Jin LQ, Xu ZW, Zhang B, Yi M, Weng C-Y, Lin S, et al. Genome sequencing and analysis of fungus Hirsutella sinensis isolated from Ophiocordyceps sinensis. AMB Express. 2020; 10:105.
  44. 44. Hu X, Zhang YJ, Xiao GH, Zheng P, Xia YL, Zhang XY, et al. Genome survey uncovers the secrets of sex and lifestyle in caterpillar fungus. Chinese Science Bulletin. 2013; 58:2846–54.
  45. 45. Li Y, Jiang L, Wang K, Wu H-J, Yang R-H, Yan Y-J, et al. RIP mutated ITS genes in populations of Ophiocordyceps sinensis and their implications for molecular systematics. IMA Fungus. 2020; 11:18.
  46. 46. Li Y, Jiao L, Yao Y-J. Non-concerted ITS evolution in fungi, as revealed from the important medicinal fungus Ophiocordyceps sinensis. Mol Phylogenet Evol. 2013; 68:373–9.
  47. 47. Wen T-C, Zhu R-C, Kang J-C, Huang M-H, Tan D-B, Ariyawansha H, et al. Ophiocordyceps xuefengensis sp. nov. from larvae of Phassus nodus (Hepialidae) in Hunan Province, southern China Phytotaxa. 2013; 123:41–50.
  48. 48. Ban S, Sakane T, Nakagiri A. Three new species of Ophiocordyceps and overview of anamorph types in the genus and the family Ophiocordyceptaceae. Mycol Progress. 2015; 14:1017.
  49. 49. Chen Z-H, Dai Y-D, Yu H, Yang K, Yang Z-L, Yuan F, et al. Systematic analyses of Ophiocordyceps lanpingensis sp. nov., a new species of Ophiocordyceps in China. Microbiol Res. 2013; 168:525–32.
  50. 50. Li GJ, Hyde KD, Zhao RL, Hongsanan S, Abdel-Aziz FA, Abdel-Wahab MA, et al. Fungal diversity notes 253–366: taxonomic and phylogenetic contributions to fungal taxa. Fungal Divers. 2016; 78:1–237.
  51. 51. Long F-Y, Qin L-W, Xiao Y-P, Hyde KD, Wang S-X, Wen T-C. Multigene phylogeny and morphology reveal a new species, Ophiocordyceps vespulae, from Jilin Province, China. Phytotaxa. 2021; 478:33–48.
  52. 52. Dingley JM. The Hypocreales of New Zealand. V. The genera Cordyceps and Torrubiella. Trans Roy Soc New Zealand. 1953; 81:329–43.
  53. 53. Pöggeler S. Mating-type genes for classical strain improvements of ascomycetes. Appl Microbiol Biotechnol. 2001; 56:589–601. pmid:11601605
  54. 54. Martin T, Lu S-W, van Tilbeurgh H, Ripoll DR, Dixelius C, Turgeon BG, et al. Tracing the origin of the fungal α1 domain places its ancestor in the HMG-box superfamily: implication for fungal mating-type evolution. PLoS One. 2010; 5:e15199.
  55. 55. Lu Y, Xia Y, Luo F, Dong C, Wang C. Functional convergence and divergence of mating-type genes fulfilling in Cordyceps militaris. Fungal Genet Biol. 2016; 88:35–43.
  56. 56. Zou J, Zeng T-T, He Z-M, Zhang P, Chen Z-H. Cloning and analysis of Ophiocordyceps xuefengensis mating type (MAT) loci. FEMS Microbiol Lett. 2019; 366:fnz070.
  57. 57. Zou J, Wu L, He Z-M, Zhang P, Chen Z-H. Biological characteristics and cultivation of Ophiocordyceps xuefengensis. Mycosystema. 2017; 36:1104–10.