The genus Purpureocillium was established in 2011 to accommodate Paecilomyces lilacinus, a medically and agriculturally important fungus, based on comparative analyses of ribosomal RNA gene fragments and translation elongation factor sequences. The genus belongs to the family Ophiocordycipitaceae within the order Hypocreales, a lineage famous for its entomopathogenic members, including the so-called zombie-ant fungi of the genus Ophiocordyceps. Purpureocillium species inhabit soil, insects, nematodes, and even marine sediments worldwide, and two species in particular, P. lilacinum and P. lavendulum, are prized as broad-spectrum biological control agents against plant-parasitic nematodes, whiteflies, thrips, aphids, spider mites, and leaf-cutter ants.

Yet species boundaries within the genus have remained incompletely resolved, and several new species have been described since 2024 as researchers begin to sample more intensively. Against this backdrop, a team led by Xiaoyan Liu, Hao Wang, Lijun Luo, Baomin Fan, and Hong Yu set out to reconstruct the phylogeny of Purpureocillium using a multilocus approach, combining five nuclear loci with mitochondrial gene sequences and comparative morphology. Their new species, isolated from an infected ant collected in September 2022 at an altitude of 1,380 meters, proved to be the centerpiece of the effort.

The researchers used a rigorous molecular toolkit. Nuclear phylogenies were built from concatenated sequences of the internal transcribed spacer region, the small and large ribosomal RNA subunits (18S and 28S), the translation elongation factor 1-alpha gene (TEF1-alpha), and the RNA polymerase II largest subunit gene (RPB1). Sequences were aligned with Clustal X and MEGA6, tested for incongruent phylogenetic signals with PAUP*, and analyzed under both maximum likelihood and Bayesian inference frameworks, with substitution models selected by ModelFinder under the Bayesian information criterion. Bootstrap support of 5,000 replicates in the maximum likelihood analysis and posterior probabilities in the Bayesian analysis provided convergent statistical validation of the resulting trees.

The nuclear analyses consistently and strongly placed the Yunnan isolate within the Purpureocillium clade as a distinct lineage. Pairwise comparisons revealed substantial divergence: 65 nucleotide differences in the ITS region and 44 differences in TEF1-alpha between the new species and P. lilacinum, and 69 and 44 to 45 differences, respectively, versus P. lavendulum. Divergence was lowest against P. atypicola, with only 14 differences in TEF1-alpha and 20 in RPB1, yet the combined multilocus dataset still recovered the new isolate as evolutionarily separate. The authors caution that the phylogenetic positions of recently described species such as P. roseum and P. sodanum remain provisional, because sequence data for them are limited to the ITS and 28S regions and show incongruence with other loci.

Morphology reinforced the molecular story. Under the microscope, P. xinfangense produces conidiophores bearing two or three whorls of phialides, the spore-forming cells, which are predominantly short and cylindrical with a pointed, mucronate apex. Its conidia are long and cylindrical, arranged in chains, measuring roughly 2.62 to 4.67 micrometers in length but only 0.65 to 1.59 micrometers wide. This is a striking contrast with its relatives: P. lilacinum produces subglobose to ellipsoidal conidia, P. takamizusanense produces ellipsoidal spores of intermediate length, and P. lavendulum produces subglobose to lemon-shaped conidia with an apiculate base. Colony growth on malt extract agar was also distinctive, reaching only 24 to 26 millimeters after fourteen days, with white fluffy mycelium and a pink-purple sporulation zone.

The mitochondrial genomes added an even more dramatic dimension. The team sequenced and compared the complete mitogenomes of four species: P. lilacinum at 23,498 base pairs, P. lavendulum at 23,475 base pairs, P. takamizusanense at 26,396 base pairs, and P. xinfangense at a towering 121,341 base pairs. All four genomes are closed circles and share the same core gene content: fifteen canonical protein-coding genes, including the ATP synthase subunits atp6, atp8, and atp9, cytochrome c oxidase subunits cox1 through cox3, seven NADH dehydrogenase subunits, the cytochrome b gene cob, and the ribosomal protein gene rps3, along with 25 transfer RNA genes and two ribosomal RNA genes. Among these genes, nad5 is the longest at 2,007 base pairs and atp8 the shortest at 147 base pairs, while tRNA genes range from 71 to 87 base pairs.

Why is the new species’ mitochondrial genome so enormous? The authors cannot yet say definitively, but they offer testable hypotheses grounded in prior work. A 112,465-base-pair mitogenome for P. atypicola is known to harbor numerous introns, and comparative analyses have shown that intron gain and loss are the primary drivers of mitochondrial genome size divergence within Purpureocillium, with the core gene order largely conserved. Similar cases exist across the fungal kingdom: intron invasion expanded the mitogenomes of Rhynchosporium species, and the 135-kilobase mitochondrial genome of the cultivated mushroom Agaricus bisporus is packed with group I and group II introns and plasmid-related sequences. Kingdom-wide surveys likewise link fungal mitogenome size variation to accessory elements such as introns, homing endonuclease genes, and unidentified open reading frames. Whether intron proliferation, repeat accumulation, or extra open reading frames account for the expansion in P. xinfangense awaits detailed repeat annotation and intron mapping.

Nucleotide composition offered a further signature. GC skew was positive in all four genomes, but AT skew differed sharply: it was slightly negative in P. lilacinum, P. lavendulum, and P. takamizusanense, yet positive in P. xinfangense. Such strand-specific compositional asymmetry is thought to reflect asymmetric mutational pressures during mitochondrial replication and transcription, so the compositional fingerprint of the new species differs from its congeners in a measurable way. Mitochondrial phylogenetic analysis of fourteen concatenated protein-coding genes across 52 hypocrealean taxa placed P. xinfangense in a phylogenetically isolated position, sister to all other Purpureocillium species, with high topological stability and strong support for the monophyly of all six sampled families within Hypocreales.

The discovery carries ecological intrigue as well. The type specimen came from a naturally infected ant cadaver associated with an Ophiocordyceps unilateralis species complex, neither of which could be identified to species level. This suggests a possible hyperparasitic association, in which one fungus colonizes another, but the authors emphasize that the exact relationship remains uncertain and will require direct observation, infection assays, and genomic analyses to confirm. Although the new species is currently known from a single collection, the combination of distinctive morphology, a rare ant-associated substrate, and concordant support from both nuclear and mitochondrial datasets justified its formal description. With five new Purpureocillium species described since 2024, including taxa associated with spiders in Brazil and karst landscapes in China, the genus is clearly far more diverse than its two well-studied biocontrol representatives suggest, and the authors call for broader taxonomic sampling to resolve species boundaries and unlock the biocontrol potential of these enigmatic purple-spored fungi.

Subject of Research: Phylogenetics and mitochondrial genomics of the fungal genus Purpureocillium and description of a new species from southwestern China

Article Title: Phylogenetic reconstruction of the genus Purpureocillium based on multilocus nuclear and mitochondrial DNA sequences, with the description of Purpureocillium xinfangense sp. nov. (Ophiocordycipitaceae, Hypocreales) from Southwestern China

Article References: Liu, X., Wang, H., Luo, L., Liu, Z., Yu, Y., Fan, B., & Yu, H. (2026). Phylogenetic reconstruction of the genus Purpureocillium based on multilocus nuclear and mitochondrial DNA sequences, with the description of Purpureocillium xinfangense sp. nov. (Ophiocordycipitaceae, Hypocreales) from Southwestern China. Frontiers in Fungal Biology, 7, Article 1972355. https://doi.org/10.3389/ffunb.2026.1972355

Image Credits: AI Generated

DOI: 10.3389/ffunb.2026.1972355

Keywords: Purpureocillium, Purpureocillium xinfangense, Ophiocordycipitaceae, Hypocreales, mitochondrial genome, phylogenetics, fungal taxonomy, entomopathogenic fungi, Yunnan China, multilocus phylogeny, biocontrol fungi, species discovery

The story of Tolypocladium begins in 1971, when the Dutch mycologist Walter Gams established the genus to house three soil fungi distinguished by a suite of distinctive microscopic features: sparingly branched spore-bearing stalks called conidiophores, swollen spore-producing cells known as phialides with conspicuously bent necks, and single-celled hyaline spores that gather together in slimy droplets at their tips. For decades, identifying members of this genus was largely a matter of matching these morphological traits and noting the host organism the fungus had killed or colonized. But as molecular phylogenetics swept through mycology in the 1990s and 2000s, cracks began to appear. Hodge and colleagues first connected the asexual Tolypocladium inflatum to its sexual stage, the fungus now known as Cordyceps subsessilis, revealing that this apparently simple mould was in fact part of a much larger lineage of entomopathogenic fungi.

The subsequent history of the genus is a case study in how the One Fungus-One Name principle, formally adopted by the International Code of Nomenclature for algae, fungi, and plants, has reshaped fungal classification. In 2007, Sung and Spatafora introduced the genus Elaphocordyceps for sexual species that parasitize Elaphomyces, the false truffles, along with certain arthropods, linking them to Tolypocladium anamorphs through multigene phylogeny. Seven years later, Quandt and colleagues synonymized Elaphocordyceps and another asexual genus, Chaunopycnis, under the older name Tolypocladium, consolidating the group within the family Ophiocordycipitaceae in the order Hypocreales. Recent surveys by Yu and colleagues in 2021 and Dong and colleagues in 2022 expanded the genus further, describing several new species and eventually bringing the count to 48 species worldwide, only nine of which had been reported from China.

Despite this growing census, one persistent problem loomed: molecular evidence kept hinting that Tolypocladium, as traditionally circumscribed, was polyphyletic, meaning its members did not all share a single recent common ancestor. The most conspicuous anomaly was Tolypocladium extinguens, a pathogen of the New Zealand glowworm Arachnocampa luminosa first described in 1984. Preliminary analyses by Dong and colleagues had already suggested it sat outside the core Tolypocladium lineage, but with limited sampling at the time, no formal move was made. This left mycologists with an uncomfortable question: how much of what they called Tolypocladium actually belonged together?

To answer it, the research team assembled one of the most taxonomically rich datasets ever built for this group, combining 202 fungal taxa and 3,424 DNA characters drawn from five loci: the internal transcribed spacer region and the large subunit ribosomal RNA gene, which are standard barcoding regions, plus three protein-coding genes, rpb1 and rpb2, which encode the two largest subunits of RNA polymerase II, and tef-1α, which encodes translation elongation factor 1-alpha. The specimens came from fieldwork on Fanjing Mountain in Jiangkou County, Tongren City, a UNESCO World Heritage site in southwestern China where dead insects infected with fungi were collected from the litter layer, transported in sterile bags, and cultured on potato dextrose agar under carefully controlled conditions of 25 degrees Celsius with alternating twelve-hour light and dark cycles.

The phylogenetic results were unambiguous. Using both Bayesian inference, run for ten million generations in MrBayes, and maximum likelihood analysis in IQ-TREE with 10,000 ultrafast bootstrap replicates, the team produced largely congruent trees showing that Tolypocladium species are scattered across three different families of Hypocreales: Ophiocordycipitaceae, Cordycipitaceae, and Clavicipitaceae. Two species in particular betrayed the genus. Tolypocladium extinguens, the glowworm pathogen, turned out to belong squarely within Cordycipitaceae, clustering near genera such as Dimorphspora, Flavocillium, Gamszarea and Parengyodontium. Meanwhile Tolypocladium lignicola, described in 1983 from rotting wood in Ontario, Canada, where it parasitizes bdelloid rotifers, landed firmly inside the genus Metapochonia in Clavicipitaceae.

The consequences were formal and deliberate. Because Tolypocladium extinguens could not remain in the genus that bears its name, the authors erected a new genus, Tolypocladiopsis, a name that acknowledges the morphological resemblance while signaling a fundamentally different evolutionary identity within Cordycipitaceae. The species is now Tolypocladiopsis extinguens. Similarly, Tolypocladium lignicola becomes Metapochonia lignicola, a new combination that mirrors the earlier transfer of another rotifer-associated species, Metapochonia parasitica, out of Tolypocladium. The team notes that the Bayesian support for Tolypocladiopsis was moderate, a limitation they attribute candidly to the fact that the ex-type strain of the species has only ITS and LSU sequences available, lacking the more informative protein-coding loci and represented by a single strain. Nevertheless, the lineage’s consistent recovery across analyses, its stable placement in Cordycipitaceae, and its unique paecilomyces-like reproductive structures and ellipsoidal to kidney-shaped spores justify its recognition.

The Fanjing Mountain expedition also delivered a bonanza of new species. The six newly collected strains formed three distinct, well-supported clades within Tolypocladium sensu stricto, each confirmed as a separate species using the Genealogical Concordance Phylogenetic Species Recognition approach, including a pairwise homoplasy index test that returned a value of 0.05942, just above the 0.05 threshold indicating the species are genuinely distinct rather than recombining populations. Tolypocladium guizhouense, with phialides averaging 17.6 by 1.9 micrometers, was found on a Cordyceps host. Tolypocladium tongrenense, bearing the shortest phialides at 11.9 by 1.1 micrometers and producing two distinct conidial types, infected a dead ant. Tolypocladium zongqii, named in honor of Professor Zongqi Liang for his contributions to arthropod-pathogenic fungus research, was isolated from a lepidopteran pupa and carries phialides measuring 13.1 by 1.6 micrometers with subglobose to ellipsoidal spores. Each species is anchored by holotype specimens deposited at the Institute of Fungus Resources at Guizhou University.

Beyond the formal nomenclature, the study delivers a broader lesson about the pitfalls of morphology-driven classification in fungi. The simple, reduced conidiogenous structures that characterize tolypocladium-like asexual forms have demonstrably evolved on multiple occasions across the Hypocreales, creating a case of convergent evolution so convincing that it fooled taxonomists for fifty years. The authors also report unexpected placements for other lineages, including Metarhiziopsis microspora, which forms a sister clade to Tolypocladium sensu stricto within Ophiocordycipitaceae despite its historically assumed position in Clavicipitaceae, and several other genera whose relationships remain stubbornly unresolved even with expanded sampling.

The discovery of three new Tolypocladium species from a single locality in Guizhou Province underscores how much fungal diversity remains undocumented in the biodiverse forests of southwestern China, where Fanjing Mountain may harbor further undescribed lineages. The research team has also provided a comprehensive taxonomic key covering all recognized Tolypocladium species, a practical tool for future surveys. As genomic-scale datasets and additional molecular markers become available, mycologists expect further realignments within Ophiocordycipitaceae, whose internal generic boundaries the authors describe as more complex than previously appreciated. For now, the revision stands as a vivid demonstration that in the fungal world, appearance can be profoundly deceptive, and that only by marrying classical microscopy with the molecular record can science begin to draw the true map of fungal life.

Subject of Research: Phylogenetic reassessment and taxonomic revision of the fungus genus Tolypocladium in the order Hypocreales

Article Title: Phylogeny and reassessment of the genus Tolypocladium s. lato.: Proposing a new genus, three new species and two combinations

Article References: Chen, W.-H., Xiong, C.-X., Shu, H.-L., Li, D., Zeng, Q., Liang, J.-D., Wijayawardene, N. N., Han, Y.-F., & Zhao, J.-H. (2026). Phylogeny and reassessment of the genus Tolypocladium s. lato.: Proposing a new genus, three new species and two combinations. MycoKeys, 141, 337-379. https://doi.org/10.3897/mycokeys.141.204222

Image Credits: AI Generated

DOI: 10.3897/mycokeys.141.204222

Keywords: Tolypocladium, mycology, phylogenetics, Ophiocordycipitaceae, Cordycipitaceae, Clavicipitaceae, new species, taxonomic revision, entomopathogenic fungi, Fanjing Mountain, Guizhou Province, Hypocreales