﻿New species and new combinations in the genus Paraisaria (Hypocreales, Ophiocordycipitaceae) from the U.S.A., supported by polyphasic analysis

﻿Abstract Molecular phylogenetic and chemical analyses, and morphological characterization of collections of North American Paraisaria specimens support the description of two new species and two new combinations for known species. P.cascadensissp. nov. is a pathogen of Cyphoderris (Orthoptera) from the Pacific Northwest USA and P.pseudoheteropodasp. nov. is a pathogen of cicadae (Hemiptera) from the Southeast USA. New combinations are made for Ophiocordycepsinsignis and O.monticola based on morphological, ecological, and chemical study. A new cyclopeptide family proved indispensable in providing chemotaxonomic markers for resolving species in degraded herbarium specimens for which DNA sequencing is intractable. This approach enabled the critical linkage of a 142-year-old type specimen to a phylogenetic clade. The diversity of Paraisaria in North America and the utility of chemotaxonomy for the genus are discussed.


Introduction
Paraisaria is an asexual morph-typified genus of entomopathogenic fungi, originally described by Samson and Brady in 1983, characterized by synnemata with verticillately-branched conidiophores and flask-shaped sympodially proliferating phialides (Samson and Brady 1983).These asexual morphs were derived from larvae (Delacroix 1893) and from cultured isolates of the sexual morphs of species in the genus Cordyceps (Samson and Brady 1983;Li et al. 2004), which were later transferred to Ophiocordyceps (Sung et al. 2007).Paraisaria was later proposed for suppression, along with four other genera then in use, in favor of recognizing a broad concept of Ophiocordyceps (Quandt et al. 2014).This limited the number of new combinations required to accommodate 1F1N rules following the abolition of the dual system of nomenclature in which sexual states and asexual states of fungi were classified separately.In molecular analyses, Paraisaria has been recovered as a distinct monophyletic clade, being referred to as the "gracilis subclade" within the "ravenelii subclade" of Ophiocordyceps by Sanjuan et al. (2015).Paraisaria was ultimately resurrected in 2019, segregated from Ophiocordyceps, and amended to include sexual morphology (Mongkolsamrit et al. 2019).Paraisaria species possess distinctive sexual morphs characterized by a globose fertile terminal portion of the stroma with immersed perithecia.Thus, Paraisaria constitutes a distinct, and robustly monophyletic clade deserving a unique genus classification, though its segregation from Ophiocordyceps rendered Ophiocordyceps into several paraphyletic clades.Ultimately, a comprehensive analysis of Ophiocordyceps sensu Sung et al. (2007), is needed to establish robust generic concepts and restore global monophyly.A major sticking point for this action is the uncertain placement of the type of Ophiocordyceps, O. blattae, among the paraphyletic subclades of Ophiocordyceps.
In North America, Paraisaria species are unique among most Cordyceps sensu lato in that they form fruiting bodies in the spring, whereas most other insect pathogens fruit in the summer, fall, or winter months, which is evident in herbarium records on MycoPortal (MycoPortal 2023) and observations on the community science platform iNaturalist (https://www.inaturalist.org/projects/north-american-cordyceps-sensu-lato).Most Paraisaria species, and thus far, all known Paraisaria species occurring in North America, form fruiting bodies on subterranean insect hosts.Some of the insect hosts of Paraisaria species are sought as food and their contamination by Paraisaria species could pose a human health concern.Doan et al. (2017) reported a series of poisonings and one fatality in Southern Vietnam, among people who had consumed cicadae infected with a fungus identified as Paraisaria heteropoda (=Cordyceps heteropoda, Ophiocordyceps heteropoda), between 2008 and 2015.The toxicity was attributed to the presence of mycotoxins in the otherwise edible cicadae, and the toxic agent was putatively identified as ibotenic acid.The potential role of entomopathogenic fungi in causing food-borne mycotoxin poisonings underscores the need to describe the biological and chemical diversity present in this group of fungi.
In addition to their impact on human and animal health, fungal natural products can be highly useful phenotypic characters for taxonomic purposes.Chemical fingerprints can be used to identify chemical families that constitute a generic chemotype for a taxonomic group, and also unique suites of compounds within a chemical family can be used to resolve species.For example, Cedeño-Sanchez et al. (2023) profiled chemical extracts from stromata to characterize and distinguish species and genera in the family Hypoxylaceae.
Only two studies (Krasnoff et al. 2005;Umeyama et al. 2011) have reported a total of five natural products from Paraisaria species, both of which investigated fungi identified as the cicada pathogen, Paraisaria heteropoda.A third study reports leucinostatin analogs from an organism reported as Ophiocordyceps heteropoda (=Parasiara heteropoda) (Kil et al. 2020), but which is evidently a Purpureocillium species based on ITS phylogeny and chemotaxonomy.Doan et al. (2017) also report the amino acid, ibotenic acid from this species, but no analytical chemistry data are presented to confirm this.There are currently no published genome sequences available to mine the specialized metabolic potential of Paraisaria species, although the sequenced genomes of other Ophiocordycipitaceae species display a familial trend of high biosynthetic capacity for specialized metabolites.The first chemical study of a member of this genus resulted in the discovery of the new 8-residue antimicrobial peptaibiotics, cicadapeptins I and II, which possessed a unique two consecutive 4-hydroxyproline residues at the N-terminus (Krasnoff et al. 2005).The known antifungal and immunosuppressant sphingosine analog, myriocin was also isolated in this study.Heteropodamides A and B are N-methylated cyclic heptapeptides reported as cytotoxins from P. heteropoda (Umeyama et al. 2011).Their absolute structures are yet to be determined.The further discovery of Paraisaria species and their natural products presents fertile grounds for investigation.
In the course of ongoing investigations for the discovery of biologically active natural products from Paraisaria species (Tehan 2022), it became critical to perform a taxonomic analysis of North American Paraisaria to better understand the biological diversity present in this group.In this study, we examined 29 recent collections of Paraisaria to investigate the diversity of North American Paraisaria.We also analyzed the type collections of Ophiocordyceps insignis and O. monticola, both of which were anticipated to belong in Paraisaria based on morphological description, ecology, and phenology.One phylogenetically informative DNA sequence was afforded from the 87-year old O. monticola specimen.The 142-year old O. insignis type did not permit successful DNA sequencing, however, chemical analysis of the newly characterized paraisariamide family of compounds by LC-HRMS provided robust support for the combination of both species into Paraisaria, as well as the correct identification of a species of importance to human health, as P. insignis.This study provides a novel framework for the use of minimally destructive chemical analysis in taxonomic assessment of type specimens where DNA sequencing is not possible.The combined analysis of molecular data, morphology, ecology, phenology, and chemical data support the circumscription of two new species and two new combinations, and provides an initial overview of the diversity of American Paraisaria species.

Specimens and isolates
Twenty nine new collections of Paraisaria specimens and their insect hosts were examined.Macroscopic characters were examined from fresh stromata, and microscopic characters were examined from fresh and dried stromata, including ascospores discharged from fresh stromata when possible and sections of dried specimens.Colors are in general terms of the senior author.Specimens are deposited in the Oregon State University Herbarium mycological collection.Culture isolates of fungi were made from tissue dissected from the context of stromata, placed on PDA with 50 µg/ml ampicillin and 100 µg/ml streptomycin, or from ascospores germinated on PDA.Agar plugs were taken from outgrowth of stromatic tissue and subcultured onto PDA and CMA at 20 °C.Cultures are deposited at the USDA ARS Collection of Entomopathogenic Fungal Cultures (ARSEF).

DNA extraction and sequencing
DNA was extracted from the ascogenous portion of dried stromata, ground with mortar and pestle in CTAB buffer (1.4 M NaCl, 100 mM Tris-HCl pH 8.0, 20 mM EDTA pH 8.0, 2% CTAB w/v) and processed following the method of Kepler et al. (2012).Samples were extracted with 25:24:1 phenol:chloroform:isoamyl alcohol, (affymetrix), and DNA was precipitated with 3 M sodium acetate (pH 5.2) and 95% ethanol.PCR amplification was performed on the Internal Transcribed Spacer (ITS), amplified using ITS4 and ITS5 primers (White et al. 1990).Alternatively, ITS1F (Gardes and Bruns 1993) was used as a forward primer for samples where ITS4 did not work.For samples in which amplification of the ITS region did not succeed, individual amplification of the ITS1 and ITS2 loci was attempted using primer sets ITS5 and ITS2 (White et al. 1990) for the ITS1 locus, and ITS3 (White et al. 1990) and ITS4 for the ITS2 locus.Nuclear small subunit (nucSSU) was amplified using nucSSU131 and NS24 (Kauff and Lutzoni 2002), nuclear large subunit (nucLSU) using LROR (Rehner and Samuels 1994) and LR7 (Vilgalys and Hester 1990), subunit 1 of RNA polymerase II (RPB1) using RPB1-A f and RPB1-6R1asc (Hofstetter et al. 2007).Alternatively, CRPB-1 (Castlebury et al. 2004) was used as a forward primer for samples where RPB1-A f did not work.Elongation factor 1α (EF-1α) was amplified using 983F and2218R (Castlebury et al. 2004).PCR was performed with an iCycler (Bio-Rad, USA), with a total of 20 μl reaction mixture containing 1× PCR Buffer (Promega), 1× TBTpar prepared as in Samarakoon et al. (2013), 2.5 mM MgCl 2 , 0.5 µM each forward and reverse primers, 200 µM of each of the four dNTPs, and 0.5 U Taq polymerase.For ITS, SSU, LSU, and TEF, the PCR thermal cycle consisted of an initial 1 min denaturation at 95 °C; 34 cycles of 30 s at 94 °C, 1 min at 52 °C, 1.5 min at 72 °C, and a termination with an elongation 7 min at 72 °C.For RPB1 and RPB2, the PCR thermal cycle consisted of an initial 1.5 min denaturation at 95 °C; 39 cycles of 30 s at 94 °C, 1 min at 47 °C, 2 min at 72 °C, and a termination with an elongation 4 min at 72 °C.Sequencing was performed by the Sanger method at the Center for Quantitative Life Sciences at Oregon State University.The sequences obtained in this study were deposited to GenBank (Table 1).

Data analysis
Sequences derived from the SSU, LSU, TEF, RBP1, RPB2, and ITS were aligned with MUSCLE 5.1 (Edgar 2004).Ambiguous and phylogenetically uninformative regions were manually removed and the trimmed alignments were concatenated for analysis using Geneious Prime® 2023.0.4.A Maximum Likelihood Tree was made using the GTR+I+A algorithm and 1000 bootstrap replicates.

Chemical extraction and LCMS analysis
Excisions (0.4-6.7 mg) were made from the endosclerotia of nineteen dried Paraisaria collections, individually placed in MeOH (1 ml, HPLC-grade), sonicated for 5 min, and extracted for 1 hr at 35 °C, then 24 h at ambient temperature.
The twenty separate extracts were filtered through syringe filters (0.2 µm PTFE) and dried in vacuo before dissolution in MeOH (0.1 mg/ml, LC-MS-grade) for analysis by LC-MS, injecting 3 µl on a Phenomenex Kinetex column (2.6 µm C18 100 Å, 50 × 2.1 mm), with H 2 O + 0.1% Formic Acid (A) MeCN + 0.1% Formic Acid (B) as mobile phase solvents at 0.4 ml/min.The LC method was as thus: 0.5 mins at 20% B, a linear gradient from 20-90% B over 14 mins, 4 min at 90% B, a linear gradient from 90-100% B over 0.5 mins, 4.5 mins at 100% B, followed by a linear return to 20% B over 3 mins, and re-equilibration at 20% B for 5 mins, before the next injection.High resolution (Agilent 6545 QToF) mass data were acquired for 26 mins from m/z 100-3200, with MS/MS spectra obtained using data-dependent ion selection for up to five precursor ions per duty cycle, excluding precursor ions with m/z less than 210, and fragmenting with collision energies of 20, 40, and 60 eV.LCMS data files were converted to mzML format and deposited on the public repository MassIVE (MSV000092591).Extracted ion chromatograms were produced for m/z 690-875, corresponding to the mass range for the paraisariamide peptide family (Tehan 2022).

Molecular networking
Unprocessed LC-MS files were converted to mzML format and uploaded to the GNPS online molecular networking platform (version 30) (Wang et al. 2016) using the default network settings but with minimum peak intensity set to 3000.

Principal component analysis
LC-MS data were processed in MZmine v2.53 (Pluskal et al. 2010).Feature detection was performed with noise level set to 1×10 4 .Chromatograms were built using a minimum group size of 5, group intensity threshold set to 1×10 2 , minimum highest intensity set to 2×10 4 , and m/z tolerance was set to m/z 0.001 or 10 ppm.Chromatogram deconvolution was performed with minimum peak height set to 1×10 4 , peak duration was set to 0.1-10 mins, and the baseline was set to 5×10 2 .Isotope peaks were grouped with mass tolerance m/z 0.001 or 15 ppm, RT tolerance was set to 1, with the most intense ion taken as the representative, and max charge was set to 2. Peaks were aligned with mass tolerance m/z 0.001 or 12 ppm, RT tolerance set to 0.8 mins, with m/z weighted 75% and RT weighted 25%.Feature list rows were filtered for features falling within the range m/z 690-875, and RT 5-14 mins, with a minimum of 2 peaks per row, and a minimum of 2 peaks in an isotopic pattern.Gap filling was performed with an intensity tolerance of 10%, mass tolerance m/z 0.001 or 15 ppm, and RT tolerance 0.6 mins.The resulting feature list was subjected to Principal Component Analysis (PCA).
Cordyceps kyushuensis and C. militaris (Cordycipitaceae) were designated as outgroup taxa.All genera, with the exception of Ophiocordyceps, are supported as monophyletic clades.A clade comprising several species morphologically similar to the well-known cicada pathogen, P. heteropoda, referred to here as the "P.heteropoda complex", is resolved into five well-defined species as well as additional samples revealing cryptic diversity.Two new species within the P. heteropoda complex, Paraisaria cascadensis and Paraisaria pseudoheteropoda, are supported as monophyletic clades, and are described below.Ophiocordyceps insignis samples produced a monophyletic clade within the P. heteropoda complex supporting its combination into Paraisaria, and is redescribed based on a fresh collection, which is designated here as an epitype.The type collection of Ophiocordyceps monticola also occurred within the genus Paraisaria, grouping closely with P. yodhathaii and P. alba.It was the only North American Paraisaria species analyzed in this study which did not fall within the P. heteropoda complex.

LC-MS analysis
Molecular Network Analysis of nineteen Paraisaria endosclerotium extracts revealed a prominent subnetwork identified as the paraisariamide family of cyclopeptides, with constituent molecular ion masses ([M+H] + ) ranging from m/z 694.49-860.56 (Fig. 2A.).All endosclerotium extract samples were observed to possess a subset of paraisariamide congeners with partial overlap between species.Production of paraisariamide cyclopeptides in host/endosclerotium is thus supported as a conserved chemotype for Paraisaria.Paraisariamides can thus potentially be used as a generic diagnostic character.Chromatograms generated from the extracted ion range m/z 690-875, corresponding to the mass range for the peptide family of paraisariamides, were unique to and consistent within each species (Fig. 2B).From the processed mass data, a feature list was produced comprising 59 LC-MS ion features (Suppl.material 1).A PCA plot generated from this feature list afforded three major clusters (Fig. 2C).Samples derived from P. insignis and P. pseudoheteropoda were resolved in distinct clusters.Samples derived from P. cascadensis together with samples from its sister clade, "Paraisaria sp.1", grouped apart from other samples.

Taxonomy
Culture characteristics.Colonies on PDA 61 days at 20 °C, 28 mm, white to yellow, reverse reddish brown to orange.Mycelium septate, smooth-walled hyaline.No conidial state was observed.
Description.Stromata capitate, unbranched, growing singly to gregarious, in groups of up to four stromata on a single host.Stromata 20-52.5 mm long.Ascogenous portion brown, globose to oblong, 8-22 mm long × 7-16 mm wide, papillate with ostioles of perithecia.Stipe golden yellow to reddish orange, sometimes furfuraceous toward upper half, 14-25 × 4-9 mm long, attached to hypogeous host by thick mats of fibrous, tangled, yellow to reddish orange rhizomorphs, extending 25-45 mm.Mycelial growth occurring between, and sometimes over, larval segments, forming a thin membrane.Perithecia embedded, obclavate, brown, (520-)640-800(840) × (160-)185-250(-270) µm.Asci hyaline, cylindrical, up to 380 µ long × (3.8-)4.0-5.9(-7.5)µm, possessing abruptly thickened apex.Ascospores hyaline, filiform, smooth, disarticulating into 64 part-spores.Part-spores, cylindrical, 6.3-9.0(-10.5)× 2.5-3.5 µm.Notes.Recent collections of this species were initially determined to not match any described species and were given the provisional name Paraisaria tortuosa, which was used in a doctoral dissertation (Tehan 2022), and in confer-ence presentations.The conspecificity with Ophiocordyceps insignis (=Cordyceps insignis) was considered but it was difficult to reconcile Cooke's description of the stroma as "livid purple".However, that species was described from a dried specimen and the true colors of the fresh specimen were evidently not observed by the authority.Petch (1935) cast doubt on the accurate description of the color of C. insignis and though the original host is not able to be precisely identified, Petch's analysis here is helpful, suggesting based on morphology that the host is one that pupates in wood, which accords with the host of recent collections identified as Prionus imbricornis.Ultimately, chemical comparison of fresh collections to the holotype was definitive in the identification of the fresh collections, and strongly supports the combination into Paraisaria.Lloyd (1920) reported C. gryllotalpae from a mole cricket collected in Louisiana, USA, but that specimen was immature, and bore only cylindrical immature stromata with no ascogenous tissue.Owing to the absence of microanatomical character data available for C. gryllotalpae, and the lack of genetic data available for either species, future studies could compare P. monticola to C. gryllotalpae by chemical means, focusing on paraisariamide content of the fungal endosclerotium.P. monticola is only known from the type collection.

Additional Paraisaria specimens examined
Two additional collections were examined which were phylogenetically closest to P. cascadensis but occurring on undetermined insect hosts, outside of the known geographic distribution of Cyphoderris monstrosa, the host of P. cascadensis.Together they form a clade which is sister to P. cascadensis.We do not consider these collections to be conspecific to P. cascadensis, but their formal description was not within the scope of the present study owing to lack of adequate sampling and host data.We anticipate that they represent two distinct new species, the description of which requires further sampling.U.S.A., California: Mendocino County, Ukiah, at approximately 39.1568, -123.2328,elevation: 352 m, 5 April 2019, on undetermined insect host buried in soil, collected by Warren Cardimona (OSC-M-052011) U.S.A., Iowa: Johnson County, Solon, at approximately 41.7572, -91.5457, elevation: 238 m, 30 June 2022, on undetermined insect host buried in soil, collected by Ross Salinas (OSC-M-052026).

Discussion
In this study, two new Paraisaria species are described and two known species are combined into Paraisaria.The entomopathogenic fungal genus Paraisaria thus currently comprises 18 formally described species which occur on six continents, as deduced from a combination of herbarium records (MycoPortal 2023) and citizen science observations (iNaturalist 2023).The extent of Paraisaria diversity both in North America and worldwide is not comprehensively reflected in this study, which warrants future studies of this group.The results of our phylogenetic and chemical analyses support the presence of additional cryptic diversity yet to be elucidated.For such a geographically widespread genus, there has been a relative paucity of sampling and analyses of Paraisaria specimens globally.Continued study of this group promises to reveal additional new Paraisaria species, each with the potential for new specialized metabolite discovery.In this study, Paraisaria populations in North America prove to be enriched in species falling within the Paraisaria heteropoda complex.Species in this clade are characterized by fruiting bodies with yellow, brown, and reddish hues and prodigious orange to brown rhizomorphs attaching to hypogeous insect hosts.Aboveground portions of the fruiting bodies in some respects resemble the truffle parasite, Tolypocladium capitatum, with which they have been compared (Cooke 1883), and with which they are frequently confused.Numerous host shifts have accompanied speciation in the P. heteropoda complex with species occurring on insect hosts in orders Hemiptera, Diptera, Coleoptera, and Orthoptera.Host identification is critical for field identification of North American Paraisaria species.P. insignis and P. pseudoheteropoda overlap extensively in fruiting body morphology and geographic distribution but are easily distinguished by their respective distinct hosts.P. insignis occurs strictly on coleopteran hosts and P. pseudoheteropoda is the only known Paraisaria species to occur on cicadas in North America.P. cascadensis and P. monticola both occur on orthopteran hosts, but the geographic distribution of P. cascadensis appears to be restricted to montane regions of the Pacific Northwest, which accords with the distribution of its host, Cyphoderris monstrosa.P. monticola is only known from the type specimen collected in Vonore, TN.Re-collection efforts for this species would be valuable and could focus on records of its host Neocurtilla hexadactyla, in the vicinity of the type locality.Notably, N. hexadactyla is widely distributed, and may support a wide distribution of P. monticola.
The life cycles of Paraisaria species, including mode of infection of their insect hosts, their possible occurrence in soil, as endophytes, saprophytic, and nematophagous nutritional modes, are not well characterized.Owing to the observation that Paraisaria species produce fruiting bodies in spring months in North America, we hypothesize that they colonize their insect hosts in the prior season and overwinter as endosclerotia which are observed to possess high concentrations of cyclopeptide specialized metabolites.The molecular structures, biological activities, and chemical ecology of Paraisaria specialized metabolites are the focus of ongoing studies (Tehan 2022).
The targeted LC-MS analysis of specialized metabolites from fungi that are only partially represented in phylogenetic analyses represents a robust application of chemotaxonomy to resolve species.Fungi that produce cyclopeptides may be especially good candidates for chemotaxonomic profiling as many cyclopeptides are particularly resistant to degradation by oxidation, heating, or proteolytic cleavage (Haque and Grayson 2020).Chemotaxonomic profiling of stable metabolites also provides a framework for the analysis of fungal groups lacking genetic data for type specimens, whereby type specimens that afford only chemical data can be linked to samples for which both chemical and genetic data are available, if both types of data resolve species groups.The lack of genetic data for type material is especially challenging when type specimens are very old and possess degraded, highly-fragmented DNA, and for which no suitable neotype has been designated.Micromorphological characters lack robustly distinct differences between Paraisaria species for use in reliable species di-agnoses.It was thus critical to compare chemical profiles of recent collections of P. insignis to the holotype to rigorously establish their conspecificity.Conservation of the general paraisariamide chemotype also supports paraisariamides as chemotaxonomic markers for genus Paraisaria, as these compounds were detected in the endosclerotia of all Paraisaria specimens analyzed.These markers are substantially more durable than DNA over long periods of time as is evident from the definitive detection of these compounds in the 142-year-old holotype of P. insignis.Notably, the shape of chromatograms was visually identical between old and new specimens, indicating that even the relative abundance of paraisariamide congeners within a sample is preserved.LC-MS/MS profiling surveys should be conducted across Paraisaria species and related groups of fungi to assess the extent of the paraisariamide molecular family and confirm the utility of these metabolites as chemotaxonomic markers.
Other specialized metabolite families may offer promise as critical chemotaxonomic markers, depending on the relative stability of their biosynthetic genes over time, and whether or not they are reliably expressed.For example, genomic analyses show that the cyclosporin genotype is highly conserved within the insect pathogen, Tolypocladium inflatum (Ophiocordycipitaceae), whereas peptaibiotics have evolved rapidly (Olarte et al. 2019) though neither cyclosporins nor peptaibiotics are detected by LCMS in every Tolypocladium strain exhibiting those genotypes (Blount 2018;Tehan et al. 2022).
Ophiocordyceps blattae, the type species of the large genus Ophiocordyceps, presents another system for potential chemotyping to compare with the various paraphyletic clades of Ophiocordyceps.Grounding of genus Ophiocordyceps in a type species to strictly define a core Ophiocordyceps clade and circumscribe other clades, has remained a longstanding problem owing to the rarity of the type species, and age of its holotype specimen.Increasingly routine chemical profiling by high resolution LC-MS and metabolomics analysis applied to the characterization of fungi in taxonomic studies adds an additional layer of phenotypic assessment that could be indispensable for taxon circumscriptions.Increasing efforts to profile and characterize specialized metabolites in fungi will not only provide useful data for taxonomists but is critical for understanding fungal ecology and may also guide pharmaceutical drug discovery efforts.These pursuits are highly complementary, as demonstrated here and in ongoing research.The isolation, structure elucidation, organic synthesis, biosynthesis, biological characterization, and chemical ecology of the paraisariamides are the focus of ongoing research.

Figure 1 .
Figure 1.Maximum likelihood tree based on the combined dataset of SSU, LSU, TEF, RPB1, RPB2, and ITS sequences displaying the relationship of Paraisaria species within family Ophiocordycipitaceae.

Figure 2 .
Figure 2. Chemical comparison of paraisariamide content in the endosclerotia of Paraisaria species collected in the USA A molecular network of the paraisariamide molecular family of cyclic peptides detected in methanol extracts of endosclerotia of Paraisaria specimens.Nodes are displayed as pie charts conveying the relative abundance of paraisariamide mass ion features in each Paraisaria species (Orange = P. cascadensis, Purple = P. pseudoheteropoda, Green = P. insignis, Yellow = "Paraisaria sp.1", Red = P. monticola) B extracted ion chromatograms of m/z 690-875 for methanol extracts of endosclerotia of Paraisaria specimens C principal component analysis of mass features m/z 690-875 from methanol extracts of endosclerotia of Paraisaria specimens, color-coded by phylogenetic clade.

Figure 3 .
Figure 3. Paraisaria cascadensis A OSC-M-052017 B fertile head C cross section of fertile head showing arrangement of perithecia D perithecia E ascus F Ascus apex G-I ascospores J part-spores K, L colonies on PDA 61 d (K obverse, L reverse).

Figure 4 .
Figure 4. Paraisaria pseudoheteropoda A OSC-M-052022 B fertile head C, D cross section of fertile head showing arrangement of perithecia E perithecia F ascus G ascus apex H, I ascospores J ascospore tip K part-spores L, K colonies on PDA 61 d (L obverse, M reverse).

Figure 5 .
Figure 5. Paraisaria insignis A OSC-M-052013 Epitype B fertile head C, D cross section of fertile head showing arrangement of perithecia E rhizomorphs F perithecia G ascus H, I asci apices J-L ascospores M part-spores N, O colony on PDA 70 d (N obverse, O reverse).

Figure 6 .
Figure 6.Paraisaria monticola A holotype BPI 634610 B fertile head C ascus D ascus apex E portion of ascospore F part spores.

Table 1 .
Sequences used in phylogenetic tree construction.