Mycologia, 106(6), 2014, pp. 1090–1105. DOI: 10.3852/13-346 2014 by The Mycological Society of America, Lawrence, KS 66044-8897 # Novel endophytic lineages of Tolypocladium provide new insights into the ecology and evolution of Cordyceps-like fungi Romina Gazis1 sapwood of trees as habitat for fungal endophytes challenges our understanding of horizontally transmitted endophytes because the bark of trees not only imposes additional filters to the entry of the fungal inocula but also to their necessary exit for dispersal. Most tropical trees are long-lived (William et al. 2004); therefore, in order to disperse endophytic fungi either must migrate (from sapwood) into the foliage and sporulate on senescent leaves or persist until host death and use their wood-decaying abilities or that of other wood-decomposers to be released into the environment (Oses et al. 2008, Parfitt et al. 2010, Promputtha et al. 2010). Not all species that are isolated as endophytes constitute part of the core endophytic community of their respective host because some of these species might be incidental colonizers that remain physiologically quiescent within their host (Stone et al. 2004). Since sapwood of tropical trees harbors a high diversity of fungal species (Evans et al. 2003, Gazis and Chaverri 2010), a corresponding diversity of strategies for infection, transmission and functions also is expected. Many fungal species with ecological roles such as saprobes, plant pathogens, entomopathogens and mycoparasites have been found to spend a portion of their life cycle as ‘‘endophytes’’, that is as inconspicuous or symptomless fungal infections of living plants (Vega et al. 2008, Promputtha et al. 2010, Alvarez-Loayza et al. 2011). For instance, connections between the saprophytic and endophytic phases of fungal species have been made through molecular studies (Parafitt et al. 2010, Chaverri and Gazis 2011). Common genera of plant pathogens such as Botryosphaeria and Colletotrichum have been isolated from healthy plant tissue (Slippers and Wingfield 2007, Rojas et al. 2010). Several species of Trichoderma with proven mycoparasitic activity found ex planta also have been isolated as endophytes (Samuels et al. 2006, Bailey et al. 2008). In addition, species of entomopathogenic fungi (e.g. Beauveria, Metarhizium, Purpureocillium) have been isolated as endophytes (Evans et al. 2003, Vega et al. 2008). The relationships between endophytes and their hosts varies from species to species, and even within the same fungal species the relationship may vary between mutualism to parasitism depending on environmental factors (Schulz and Boyle 2005, Kogel et al. 2006, Alvarez-Loayza et al. 2011). Mutualism between endophytes and their hosts is difficult to Clark University, Biology Department, 950 Main Street, Worcester, Massachusetts 01610 Demetra Skaltsas University of Maryland, Department of Plant Science and Landscape Architecture, 2112 Plant Sciences Building, College Park, Maryland 20742 Priscila Chaverri University of Maryland, Department of Plant Science and Landscape Architecture, 2112 Plant Sciences Building, College Park, Maryland 20742, and Universidad de Costa Rica, Escuela de Biologı́a, Apdo. 11501-2060, San Pedro, San José, Costa Rica Abstract: The objective of this study was to identify a group of unknown endophytic fungal isolates from the living sapwood of wild and planted Hevea (rubber tree) populations. Three novel lineages of Tolypocladium are described based on molecular and morphological data. Findings from this study open a window for novel hypotheses regarding the ecology and role of endophytes within plant communities as well as trait evolution and potential forces driving diversification of Cordyceps-like fungi. This study stresses the importance of integrating asexual and sexual fungal states for a more complete understanding of the natural history of this diverse group. In addition, it highlights the study of fungi in the sapwood of tropical trees as habitat for the discovery of novel fungal lineages and substrate associations. Key words: biocontrol, entomopathogens, Hevea, inter-kingdom host jump, sapwood, species concept, species delimitation INTRODUCTION Undiscovered fungal diversity may be hidden in the tropics, especially within trees growing in undisturbed forests (Arnold and Herre 2003, Rodriguez et al. 2009, Blackwell 2011). Leaves from many plant species throughout the world have been examined for endophytic fungi (Arnold and Lutzoni 2007, Hyde and Soytong 2008). Fewer studies have looked for endophytes inhabiting stem or trunk tissues of plants (Evans et al. 2003, Rubini et al. 2005, Oses et al. 2008, Gazis and Chaverri 2010, Gazis et al. 2012). The Submitted 16 Oct 2013; accepted for publication 6 Mar 2014. 1 Corresponding author. E-mail: rgazis@clarku.edu 1090 GAZIS ET AL.: NOVEL LINEAGES OF ENDOPHYTIC TOLYPOCLADIUM assess, but it has been shown in some studies. For instance, Arnold et al. (2003) demonstrated that by inoculating a diverse array of naturally occurring endophytic fungi into cacao trees (Theobroma cacao), the damage of a common tropical tree pathogen could be reduced. In addition, it has been suggested that fungal endophytes play a defensive role in tropical plants by influencing leaf-cutting ant foraging preferences (Coblentz and Van Bael 2013). This defensive mutualism has been studied primarily in vertically transmitted endophytes (directly from parent to offspring), such as the protection against herbivores induced by Neotyphodium species to their grass hosts (Gundel et al. 2010, Eaton et al. 2011, Faeth and Saari 2012). Nonetheless, horizontally transmitted endophytes (among individuals) also are being explored and tested for their potential applications as biological control agents (Bailey et al. 2008, Rocha et al. 2011). During a survey of endophytes associated with Hevea brasiliensis and H. guianensis (Euphorbiaceae, Malpighiales) in its wild habitat and plantations, many isolates were obtained that constitute novel lineages of Tolypocladium (Ascomycota, Pezizomycotina, Sordariomycetes, Hypocreales, Ophiocordycipitacae). These novel lineages were circumscribed based on the congruence of multiple molecular markers and morphological data. The relationships of these new lineages to other species of Tolypocladium is reported here. We include species previously classified as Chaunopycnis, generally found as endophytes or in the soil, and Elaphocordyceps, known as entomopathogens and mycoparasites. Chaunopycnis and Elaphocordyceps are synonyms of Tolypocladium (Quandt et al. 2014), thus in this study we use Tolypocladium. Results found here broaden the spectrum of substrata used by this group of fungi and cast new light on our understanding of ecological functions of tropical endophytes. MATERIALS AND METHODS Isolations.—As part of a project to characterize the fungal endophyte diversity inhabiting rubber trees (Hevea spp.), sapwood and leaf endophytes were isolated from H. brasiliensis and H. guianensis populations distributed inside and outside their native range (Gazis 2012). The present study focuses on 46 isolates collected from tissue of 35 individual Hevea trees in these localities: i. Amazon Conservatory of Tropical Studies (ACTS) Biological Station, Loreto, Peru, 12u30943.200S, 70u3934.090W, seven trees; ii. Los Amigos Biological Station, Madre de Dios, Peru, 3u14952.30S, 72u54953.80W, three trees; iii. Madre Selva Biological Station, Loreto, Peru, 3u37914.900S, 72u14948.330W, 14 trees; iv. Rurópolis, Pará, Brazil, 4u06949.70S, 55u00924.90W, two trees; v. Novo Horizonte, Aveiro, Brazil, 4u02955.40S, 1091 55u21911.70W, one tree and (vi) Tabasco rubber plantation, Huamanguillo, Mexico, 17u58932.530N, 93u23913.650W, eight trees. Most of the isolates (45 out of 46) were found in sapwood with only one isolate from leaf tissue. Therefore, only the methods for sapwood-endophyte isolation are described here. Leaf endophytes were isolated following the protocols described in Gazis and Chaverri (2010). Three chips of ca. 3 3 6 cm dead bark were cut from each tree at shoulder height (1.5 m) and from three different parts of its circumference with a knife sterilized in the field by immersion in 95% ethanol and flaming. After exposing the sapwood (one section at a time), three pieces of ca. 5 3 5 mm tissue were excised from each exposed area with the aid of a sterilized scalpel and quickly transferred to Petri dishes containing CMD (BBLTM cornmeal agar + 2% dextrose) and 1% neomycin-penicillin-streptomycin (Sigma-Aldrich, St Louis, Missouri). Petri dishes were kept at ca. 4–8 C until they were processed in the laboratory (Department of Plant Science and Landscape Architecture, University of Maryland, College Park). Once in the laboratory, cultures were incubated up to 2 mo and emerging colonies were subcultured in DifcoTM potato dextrose agar (PDA) to obtain pure isolates. Morphological studies.—Isolates were grown on PDA, MEA (DifcoTM malt extract agar) and SNA (Spezieller Nährstoffarmer Agar; Nirenberg 1976) for up to 3 wk at 25 C with alternating 12 h/12 h fluorescent light/darkness. Microscopic observations of the cultures were made with a Leica DFC2500 microscope. Measurements of conidia and phialides (length and width) were made with the Leica Application Suite 3.5.0. Continuous measurements (at 31000) were based on at least 100 measured units and are reported as the mean 6 one standard deviation. Images were captured with a Leica DFC420 digital camera. Statistical analysis was conducted with SPSS 13.0 for Windows. A dried culture of the type specimen corresponding to each novel species was deposited at the U.S. National Fungus Collections (BPI), and additional representatives and ex-type cultures were deposited in the Centraalbureau voor Schimmelcultures (CBS), Utrecht, the Netherlands. DNA extraction, PCR and sequencing.—Pure cultures of the isolates were grown in DifcoTM potato dextrose broth (PDB) at 25 C for 1 wk. Genomic DNA was extracted from the mycelial mat with Power PlantTM DNA isolation kit (MO BIO Laboratories Inc., Solana Beach, California) according to the manufacturer’s instructions with these modifications: Mycelial tissue was stored at 280 C several days before extraction, and instead of the vortex a FastPrepH-24 (MP Biomedicals, Solon, Ohio) machine was used to improve tissue lysis. Seven loci were sequenced: the internal transcribed spacers (ITS 5 ITS1 + 5.8S + ITS2), the nuclear small subunit (nucSSU), nuclear large subunit (nucLSU), subunit 1 of RNA polymerase II (RPB1), partial b-tubulin gene, elongation factor 1a (EF-1a) and mitochondrial ATP synthase subunit 6 (mtATP6) (SUPPLEMENTARY TABLE I). We also attempted to sequence subunit 2 of RNA polymerase II (RPB2) but obtained a nonspecific region with primers RPB2-5F2 (forward) and fRPB2-7cR (reverse) (Castlebury et 1092 MYCOLOGIA al. 2004). All PCR reactions were assembled as follows: 12.5 mL GoTaqH Green Master Mix (Promega Corp., Madison, Wisconsin), 1.25 mL 10 mM reverse primer, 1.25 mL 10 mM forward primer, 1 mL dimethyl sulfoxide (DMSO, Sigma-Aldrich, St Louis, Missouri), a maximum of 25 ng/mL of genomic DNA and double-distilled water to complete the total volume (25 mL). PCR products were cleaned with ExoSAP-ITH (USB Corp., Cleveland, Ohio) and sequenced at MCLAB laboratories (www.mclab.com). SequencherTM 4.9 (Gene Codes Corp., Ann Arbor, Michigan) was used to assess the quality of sequence chromatograms. Sequences were aligned with MAFFT 6 under the EINS-i strategy (Katoh et al. 2009), and the alignments were viewed and refined with Mesquite 2.75 (Maddison and Maddison 2011). Ambiguously aligned regions were excluded from the alignment with Gblocks 0.91b (Talavera and Castresana 2007) with reduced stringency settings by allowing gaps within final blocks and less strict flanking positions. Newly produced sequences were deposited in GenBank (SUPPLEMENTARY TABLE II), and the alignments and corresponding phylogenetic trees were submitted to TreeBASE (S15407). Multilocus datasets.—Because the BLAST algorithm from NCBI placed the unknown strains close to species in the Cordyceps-like group (Clavicipitaceae s.l.), we selected the dataset from Sung et al. (2007a) as a phylogenetic framework to infer their taxonomic placement. The taxon matrix was reconstructed with sequences downloaded from GenBank. (Refer to Sung et al. 2007a for taxa and accession numbers.) Bootstrap analysis with RAxML as implemented in raxmlGUI 1.3 (Stamatakis 2006, Silvestro and Michalak 2012) was conducted separately on each locus to detect topological incongruence before concatenation of loci (1000 replicates, GTR+GAMMA model). Conflict was assumed to be significant if a group of taxa was supported at $ 70% as monophyletic with one locus but supported as nonmonophyletic by another locus (reciprocal 70% ML bootstrap support criterion; Reeb et al. 2004). The mtATP6 locus has been reported to show topological incongruence with other fungal markers in this study (Sung et al. 2007b); we found no conflict in the targeted clade (Ophiocordycipitaceae). Two datasets comprising representatives of the different families within the Hypocreales were assembled: a seven-locus (nucLSU+nucSSU+RPB1+RPB2+b-tubulin+EF1a+mtATP6) dataset consisting of 202 taxa, including ambiguously aligned regions and a seven-locus (nucLSU+ nucSSU+RPB1+RPB2+b-tubulin+EF-1a+mtATP6) dataset consisting of 202 taxa, excluding ambiguously aligned regions. In addition, a dataset containing only Tolypocladium strains and representatives from its assumed synonyms (Chaunopycnis and Elaphocordyceps) was assembled with eight loci (nucLSU+nucSSU+RPB1+RPB2+ITS+b-tubulin+ EF-1a+mtATP6) and comprising 60 taxa. Ambiguously aligned regions were excluded from the matrix. Taxa placed in Chaunopycnis and Tolypocladium are not well represented in GenBank; therefore only the ITS, nucLSU and in some cases the nucSSU regions were included for those species. Even though the RPB2 region was missing for the unknown endophytic strains (see DNA extraction, PCR and sequencing), this region was included in the analyses because it was available for the majority of the other taxa. Outgroup for the 202-taxa datasets was composed by two isolates of Glomerella cingulata and for the 60-taxa dataset by two isolates of Ophiocordyceps gracilis. Phylogenetic analyses and species delimitation.—Phylogenetic relationships and node robustness were estimated for the three datasets described above, using maximum likelihood (ML) as implemented in raxmlGUI 1.3 (Silvestro and Michalak 2012) and Bayesian Inference (BI) as implemented in MrBayes 3.1.2 (Ronquist and Huelsenbeck 2003). Analyses were run with Clark University’s computing cluster. In the ML and the BI analyses the three datasets (202-taxon with and without ambiguously aligned regions and the 60taxon dataset) were partitioned by locus. The best tree and bootstrap support were estimated with 1000 replicates with the general time-reversible evolutionary model with a gamma distribution to account for rate heterogeneity among sites (GTRGAMMA). Bayesian analysis was conducted with Markov chain Monte Carlo (MCMC) under a GTR+I+G model. Four default chains were sampled every 1000 generations and run for a total of 30 000 000 generations. Convergence of log likelihood scores (2Ln) was assessed with TRACER 1.4 (Rambaut and Drummond 2007), and stationarity was assumed when a stable equilibrium value was reached (Ronquist and Huelsenbeck 2003). A burn-in sample of 20 000 trees was discarded for each run. The remaining 20 000 trees (10 000 from each run) were used to estimate posterior probabilities (PP) with the MAJORITY RULE CONSENSUS TREE command in Mesquite. Individual nodes were considered well supported when ML bootstrap values (BS) were equal to or greater than 70% and when PP values were equal to or greater than 0.95. To delimit species within the collection of Hevea endophytic strains, we used a combination of the genealogical concordance phylogenetic species recognition (GCPSR) concept (Taylor et al. 2000) and recombination analyses (i.e. phylogenetic networks). Phylogenetic networks can detect potential evidence of recombination or incomplete lineage sorting among strains. Networks were constructed based on the multilocus alignment containing only Hevea strains and using SplitsTree 4.12.6 (Huson and Bryant 2006). We used the neighbor-net procedure and applied the GTR distance correction. Bootstrap values were obtained with 1000 replicates. Terminal clades that were highly supported in the multilocus analysis (eight loci, 60 taxa) and present in the recombination analyses were considered distinct evolutionary lineages and therefore described here as new species. In addition we evaluated whether morphological characters were correlated to the putative phylogenetic species. Ancestral state reconstruction (ASR).—To study the evolution of host affiliation within the Tolypocladium (5 Chaunopycnis 5 Elaphocordyceps) clade, we reconstructed the ancestral host affiliation state of the most common recent ancestor (MCRA) of each lineage. Host or substrate affiliations were assigned as character states. We codified taxa under one of two states: associated with insect 5 0 (entomopathogen) or associated with other fungi (myco- GAZIS ET AL.: NOVEL LINEAGES OF ENDOPHYTIC TOLYPOCLADIUM parasite) 5 1; taxa isolated as endophytes were assigned an unknown state (?) to denote our uncertainty in determining their true ecological role within their host (Chaverri and Samuels 2013). Data on host identity was retrieved from Bissett (1983), Samson and Soares (1984), Bills et al. (2002) and Sung et al. (2008). Ancestral states were reconstructed under Bayesian, maximum likelihood and parsimony models. In all analyses we used 1000 rooted trees drawn randomly from the post burn-in 20 000-tree pool derived from the MrBayes analysis of the eight-gene, 60-taxa dataset. Before analyses, endophytic strains that were not representative of independent lineages were removed from the trees. Maximum likelihood and parsimony ancestral reconstruction were inferred with the modules implemented in Mesquite 2.75 (Maddison and Maddison 2011). In Mesquite the option TRACE CHARACTER OVER TREES under ANALYSIS was selected to reconstruct ancestral character states assuming an Mk1 and an AsymmMk class models (with rates estimated by Mesquite from the data) and unordered characters. Parsimony reconstructions were optimized with MOST PARSIMONIOUS RECONSTRUCTIONS (MPR) option. The program BayesTraits 1.0 (Pagel and Meade 2007, www.evolution. rdg.ac.uk) was used to reconstruct the ancestral states of key nodes in the Tolypocladium phylogeny. The BayesMultistate method was implemented with the MCMC option. Various priors and parameter bounds then were explored. Two sets of priors were tested, a uniform 0, 100 (default settings) and a gamma hyperprior 0, 1 (HYPERPRIORALL command). A rate deviation of 130 was found to give acceptance rates with a mean of 30–40%. Acceptance rates 20–40% are considered optimal according to Pagel and Meade (BayesTraits 1 manual) and hence the above model was used in all further MCMC analyses. The MCMC analyses were carried out with 5 000 000 generations, sampling every 100th generation, with a burn-in of 50 000. Alternative ancestral states for the key nodes were compared with the FOSSIL command to fix internal nodes to either state 0 or 1 and the BayesFactor (BF) tests (BF 5 2[log {harmonic mean (best model)} – log harmonic mean {alternative model}]) was used to compare the harmonic means of the alternative states. Support for any particular state was regarded as positive when BF 5 was . 2 (Pagel et al. 2004). The key nodes for the ASR analysis using BayesTraits were: i. clade containing all the Tolypocladium strains and ii. clade containing only T.album, T. pustulatum and the novel endophytic lineages. Mining of ecological data with ITS sequences deposited in GenBank.—To include other studies that have isolated strains belonging to the Tolypocladium clade from living plant tissues (as endophytes) or from other not yet documented environments, we retrieved sequences that have not been sufficiently identified but are believed to be closely related to these genera for an ITS phylogenetic analysis. For this task we used the USER-SPECIFIED GENUS tool implemented in the Emerencia webserver (Ryberg et al. 2009). To maximize the exploration of potential ecologies, we also included sequences that blasted to our query (using ITS sequences from the Hevea endophytes) with more than 97% similarity and 90% coverage. Two sequences representing the ITS barcode of the type species of Chaunopycnis 1093 (C. alba, AF389191 5 Tolypocladium album) and Tolypocladium (T. inflatum, AY245643) were added to the dataset (Seifert et al. 2011). In summary, the final ITS matrix was composed by representatives of Tolypocladium (as Elaphocordyceps in the Sung et al. 2007a multilocus analysis), our unknown Hevea endophytic strains, ITS-barcode sequences for ‘‘Chaunopycnis’’ and Tolypocladium and sequences retrieved directly from GenBank or through Emerencia algorithm that showed similarity with ‘‘Chaunopycnis’’, ‘‘Elaphocordyceps’’ or Tolypocladium sequences. This dataset had a total of 130 ITS sequences (SUPPLEMENTARY TABLE III). Following Wang et al. (2011), we aligned the ITS dataset with SATé (simultaneous alignment and tree estimation, Liu et al. 2012). SATé alignment was performed by letting the iteration run 24 h under custom settings, using MAFFT for alignment, RAxML for phylogenetic inference and a centroid decomposition strategy with a maximum subproblem of 20% of the sequences. The SATé alignment of the complete ITS (ITS1-5.8S -ITS2) then was edited in Mesquite to remove the residues of 18S and 28S rDNA at the fringes. The aligned and edited ITS dataset was used to build an ITS-based phylogeny in raxmlGUI 1.3 (Silvestro and Michalak 2012). The phylogeny based on the eight-locus dataset (nucLSU+nucSSU+RPB1+RPB2+ITS+b-tubulin+EF1a+mtATP6) consisting of 60 taxa was used as a constraint tree. To mine ecological data we revised the metadata and the publications associated with the GenBank accessions used in the analysis. RESULTS Phylogenetic placement and species delimitation of the endophytic fungi of Hevea.—Both multilocus datasets (202 taxa) placed the endophytic strains of Hevea as a monophyletic group within Tolypocladium (5 Chaunopycnis 5 Elaphocordyceps), with high support values (BS: 100, PP: 1; FIG. 1A, B; SUPPLEMENTARY FIG. 1). The evolutionary relationships among lineages within Tolypocladium could not be elucidated with confidence due to lack of support (FIG. 1B). Nevertheless, our analyses placed ‘‘Elaphocordyceps’’ capitata, ‘‘E.’’ fracta, ‘‘E.’’ japonica and ‘‘E.’’ longisegmentis as basal lineages to the rest of the species of Tolypocladium (BS: 78; PP: 1). Based on the GCPSR concept and on the phylogenetic network analysis, the endophytic clade was segregated into four lineages (clades I–IV in FIG. 1B and SUPPLEMENTARY FIG. 2). Among the four clades, three appeared to be novel species of Tolypocladium while one had ‘‘Chaunopycnis’’ alba nested within it. Additional phylogenetic structure within each of these clades was detected, but we did not find sufficient molecular and morphological evidence to further segregate them into distinctive lineages. Ancestral state reconstruction of the novel endophytic lineages.—The ancestral host association for the genus Tolypocladium was reconstructed unequivocally 1094 MYCOLOGIA FIG. 1. A. Large phylogeny including taxa used in Sung et al. 2007 (seven loci, 202 taxa dataset) with large clades (at the family level) collapsed. B. Eight loci phylogeny (60 taxa dataset) only with Tolypocladium members, including their synonyms Elaphocordyceps and Chaunopycnis. Bootstrap support and posterior probability are indicated as BS/PP. as fungal in the parsimony analysis (FIG. 2), suggesting that isolates on insects and as endophytes were derived from those occurring on fungi (5 mycoparasites). Within this clade it appears that several interkingdom host jumps occurred (i.e. from fungus to insect). The most recent common ancestor (MRCA) of the clade containing the Hevea endophytic lineages was derived from an insect-associated lineage (FIG. 2). The lack of support found for the relationships among the lineages within the Tolypocladium clade prevented us from confidently tracing the evolutionary history of the character states (host affiliation) under more complex models (i.e. likelihood and Bayesian). Reconstruction under the mentioned models were equivocal for the key nodes (Tolypocladium node: P [insect] 5 0.5; P [fungi] 5 0.5; T. album, T. pustulatum, T. endophyticum, T. tropicale, T. amazonense node: P [insect] 5 0.5; P [fungi] 5 0.5). However, BayesTraits analyses gave consistently higher likelihoods for the ancestral state at the root of the Tolypocladium clade as being associated with a fungal host rather than with an insect host, but the BayesFactor test was not significant (BF 5 0.44). For the second key node (connecting T. album, T. pustulatum, T. endophyticum, T. tropicale and T. amazonense) BayesTraits analyses produced consistently higher likelihoods for the ancestral state being associated with an insect host than with a fungal host, but again the BayesFactor test was not significant (BF 5 0.06). Novel ecological associations based on mining GenBank sequences and associated metadata.—The top hits from the BLAST query of the NCBI database were species of Tolypocladium, including species under its syno- GAZIS ET AL.: NOVEL LINEAGES OF ENDOPHYTIC TOLYPOCLADIUM 1095 FIG. 2. Results from the ancestral state reconstruction analysis based on parsimony. The phylogeny in which the hostassociation states were mapped was produced by Bayesian analysis of a concatenated dataset of eight loci (nucLSU+ nucSSU+RPB1+RPB2+ITS+b-tubulin+EF-1a+mtATP6). nyms Chaunopycnis and Elaphocordyceps. With the 130 ITS sequence dataset we found that strains belonging to Tolypocladium have broader ecological niches than previously reported (FIG. 3). Strains of Tolypocladium album have been isolated from ant nests, marine sponges (endozoic) and plant roots (endophytic). Because of the low support for the internal nodes, typical of the ITS region (Wang et al. 2011), only a few strains could be placed with confidence as part of the novel lineages described here or as part of the known species included in the analysis. On the other hand, most of the Tolypocladium strains included in this this analysis formed a clade (BS: 72%) with the ITS barcode sequence for Tolypocladium inflatum (5 Elaphocordyceps subsessilis). TAXONOMY Tolypocladium endophyticum Gazis, Skaltsas, & P. Chaverri., sp. nov. MycoBank MB807965 Etymology: In reference to the habit of the strains. Colonies in PDA and MEA forming abundant mycelium, white, floccose, superficial. Conidial spor- ulation promoted by growing in low nutrient media (SNA). Conidiomata absent or forming discrete pustules on nutrient-poor media (SNA) and on the margins of nutrient-rich media (PDA). Conidiophores with trichodermoid branching pattern. Conidiogenous cells phialidic, discrete, lageniform, hyaline, smooth-walled, 4.1 6 0.9 mm long 3 1.6 6 0.2 mm broad; intercalary phialides often produced. Conidia abundantly produced and aggregated in slimy heads, globose, hyaline, aseptate, smooth, 1.3 6 0.2 mm diam. Chlamydospores present, hyaline, intercalary. Habitat: Isolated as endophytes from living sapwood of Hevea brasiliensis and H. guianensis trees in their natural habitat (Amazon Basin) and in plantations in and outside their native range. Known distribution: Brazil, Mexico and Peru. Holotype: MEXICO. TABASCO: Huamanguillo, 17u58932.530N, 93u23913.659W, , 10 m, endophytic in living sapwood of cultivated Hevea brasiliensis, Apr 2010, coll. R. Gazis, (HOLOTYPE BPI892890; cultures EX-HOLOTYPE MX575 5 CBS 136896). Additional cultures examined: MEXICO. TABASCO: Huamanguillo, 17u58932.530N, 93u23913.650W, , 10 m, 1096 MYCOLOGIA FIG. 3. A. ITS mining results (ITS RAxML phylogeny) and (B) inset showing the same phylogeny as in A but displaying branch lengths. A total of 130 sequences were used to construct the phylogeny, including curated and noncurated sequences. Clade support was obtained with 1000 bootstraps. GAZIS ET AL.: NOVEL LINEAGES OF ENDOPHYTIC TOLYPOCLADIUM endophytic on sapwood of cultivated Hevea brasiliensis, coll. R. Gazis, Apr 2010 (MX66 5 CBS 136900, MX329, MX335 5 CBS 136898, MX451 5 CBS 136901, MX485 5 CBS 136891, MX486); PERU. MADRE DE DIOS: Manu, 12u30943.200S, 70u3934.090W, Los Amigos biological station, , 280 m, endophytic on living sapwood of Hevea guianensis, Jun 2008, coll. R. Gazis (LA196, LA264 5 CBS 136903); LORETO: Napo, 3u37914.900S, 72u14948.330W, Madre Selva biological station, , 100 m, endophytic on living sapwood of wild Hevea brasiliensis, Jun 2010, coll. R. Gazis (MS237). Tolypocladium tropicale Gazis, Skaltsas, & P. Chaverri. sp. nov. MycoBank MB807964 Etymology: In reference to the climatic region where the strains were isolated. Similar to T. endophyticum except for the size of conidiogenous cells (phialides) and conidia. Conidiogenous cells 4.6 6 1.2 mm long 3 1.5 6 0.3 mm wide. Conidia 1.5 6 0.1 mm diam. Habitat: Isolated as endophytes from sapwood and leaf tissue of Hevea brasiliensis trees in their natural habitat (Amazon Basin) and plantations outside their native range (Mexico). Known distribution: Mexico and Peru. Holotype: PERU. LORETO: Napo, Amazon Conservatory of Tropical Studies (ACTS) biological station 3u14952.30S, 72u54953.80W, , 100 m, endophytic on sapwood of wild Hevea brasiliensis, Jun 2009, coll. R. Gazis (HOLOTYPE BPI892888; cultures EX-HOLOTYPE IQ214 5 CBS 136897). Additional cultures examined: MEXICO. TABASCO: Huamanguillo, 17u58932.530N, 93u23913.650W, , 10 m, endophytic on sapwood of cultivated Hevea brasiliensis, Apr 2010, coll. R. Gazis (MX337 5 CBS 137292, MX338 5 CBS 136894); PERU. LORETO: Napo, Amazon Conservatory of Tropical Studies (ACTS) biological station 3u14952.30S, 72u54953.80W, , 100 m, endophytic on living sapwood of wild Hevea brasiliensis, Jun 2009, coll. R. Gazis (IQ35 5 CBS 136892). Tolypocladium amazonense Gazis, Skaltsas, & P. Chaverri. sp. nov. MycoBank MB807966 Etymology: In reference to area where the strains were isolated. As in T. endophyticum, except for the size of conidiogenous cells (phialides) and conidia. Conidiogenous cells 4.6 6 1.2 mm long 3 1.5 6 0.3 mm wide. Conidia 1.4 6 0.2 mm diam. Habitat: Isolated as endophytes from sapwood of Hevea brasiliensis and H. guianensis trees in their natural habitat (Amazon Basin). 1097 Known distribution: Peru. Holotype: PERU. LORETO: Napo, 3u37914.900S, 72u14948.330W, Madre Selva biological station, , 100 m, endophytic on living sapwood of wild Hevea brasiliensis, Jun 2010, coll. R. Gazis (HOLOTYPE BPI892889; cultures EX-HOLOTYPE MS308 5 CBS 136895). Additional cultures examined: PERU. MADRE DE DIOS: Manu, 12u30943.200S, 70u3934.090W, Los Amigos biological station, , 280 m, endophytic on living sapwood of Hevea guianensis, Jun 2008, coll. R. Gazis (LA100 5 CBS 137291, LA108 5 CBS 136899); LORETO: Napo, 3u37914.900S, 72u14948.330W, Madre Selva biological station, , 100 m, endophytic on sapwood of wild Hevea brasiliensis, Jun 2010, coll. R. Gazis (MS507 5 CBS 136893). Tolypocladium album (W. Gams) Quandt, Kepler & Spatafora, IMA Fungus, (2014). Basionym: Chaunopycnis alba W. Gams, Persoonia 11:75 (1980). Description as in T. endophyticum except for the size of conidiogenous cells (phialides) and conidia. Conidiogenous cells 8.2 6 1.7 mm long 3 3.5 6 1.3 mm broad. Conidia 2.3 6 0.5 mm diam. Cultures examined: PERU. LORETO: Napo, 3u37914.900S, 72u14948.330W, Madre Selva biological station, , 100 m, endophytic on sapwood of wild Hevea brasiliensis, Jun 2010, coll. R. Gazis (MS108, MS271, MS460, MS490, MS506 5 CBS 136902). Descriptions and illustrations: Gams, Persoonia 11:75 (1980); Möller and Gams, Mycotaxon 48:441– 450 (1993); Seifert et al. (p 561, plate 79C) Utrecht, the Netherlands: CBS-KNAW Fungal Biodiversity Centre (2011). Tolypocladium pustulatum (Bills, Polishook & J.F. White) Quandt, Kepler & Spatafora, IMA Fungus (2014). Basionym: Chaunopycnis pustulata Bills, Polishook & J.F. White, Mycol. Progr. 1:8 (2002). Descriptions and illustrations: Möller and Gams, Myco taxon 48:441–450 (1993); Bills et al. Mycol Prog 1:8 (2002). DISCUSSION This study uncovered three novel species of Tolypocladium and one previously described (T. album) as endophytes of Hevea spp. We detected phylogenetic structure within each of these clades but did not find sufficient molecular and morphological evidence to further segregate them into distinctive lineages. The latter can be a sign that these novel clades are species complexes, but more sampling would be needed to further explore if their genetic divergence is corre- 1098 MYCOLOGIA lated with geographic or host associations. Our analyses also support the monophyletic relationship among the asexual genera Chaunopycnis and Tolypocladium and members of the sexual genus Elaphocordyceps (Quandt et al. 2014). Species in Chaunopycnis and Tolypocladium were not included in the comprehensive revision of the Clavicipitaceae sensu lato by Sung and colleagues (2007a) but had been placed within this clade on the basis of teleomorphanamorph connection and molecular studies (Hodge et al. 1996, Bills et al. 2002, Stensrud et al. 2005). Chaunopycnis currently includes three species, namely the type species, C. alba, and C. ovalispora and C. pustulata. Chaunopycnis alba first was described from soil and litter (Gams 1980), and it is now thought to be a cosmopolitan species with high genetic diversity (Möller et al. 1996). The second species, C. ovalispora, was described as an endosymbiont of the Antarctica lichen Caloplaca regali (Möller and Gams 1993). The third species, C. pustulata, was described from soil, twigs and as an endophyte of Quercus segregatus (Bills et al. 2002). No sexual stage has been found for any of the Chaunopycnis species. Only Tolypocladium album (5 Chaunopycnis alba) was found as endophyte of Hevea spp. Quandt and colleagues recently have revised the taxonomy of the Ophiocordycipitaceae and made the nomenclatural revisions needed under the 1F:1N (Quandt et al. 2014). Index fungorum lists 11 species under Tolypocladium, of which only three are represented by sequence data and were included in the phylogenetic analysis. Nevertheless the three species described here were considered novel because their morphology did not correspond to any of the previously described Tolypocladium species. Morphological characterization.— In this study we recognize three lineages of Tolypocladium as new species based on molecular phylogenetic and morphological evidence. Three anamorph morphologies have been associated with this clade (i.e. Chaunopycnis, Tolypocladium and Verticillium). Verticillium-like anamorphs are distinguished from Tolypocladiumand Chaunopycnis-like anamorphs by having a broad conidiophore main axis that bears verticilliate branches and phialides as in T. microsporum (refer to Bissett 1983 for comparison of Tolypocladium against other clavicipitaceous anamorphs). Verticillium sensu stricto belongs in the Plectosphaerellaceae (not Hypocreales; order incertae sedis). Tolypocladium-like anamorphs are said to be distinguished from Chaunopycnis-like anamorphs by having flask-shaped phialides with a swollen base and narrow, often hooked neck, in addition to the lack of sporodochium-like conidiomata (Seifert et al. 2011). Howev- er, species of Tolypocladium might have more than one branching morphology. For instance, Möller and Gams (1993) described Tolypocladium ovalisporum (as C. ovalispora) as sometimes having phialides with a swollen base and a thinner neck that often arises in lateral position. Bills and colleagues (2002) reported the formation of sporodochia in T. pustulatum (as C. pustulata). Based on morphology, the three new lineages described in this study appear to be more closely related to the species described as Chaunopycnis (i.e. C. alba [type species], C. pustulata and C. ovalispora) but can be separated by few morphological differences (FIG. 4A–G, TABLE I). Only one Tolypocladium species (T. inflatum) have been connected to a sexual state Elaphocordyceps subsessilis. Evolutionary history of the nutritional mode in the Tolypocladium clade.— Ophiocordycipitaceae includes species with an animal-based nutritional mode, a trait that is considered to be the ancestral in the family (Spatafora et al. 2007, Sung et al. 2008). A shift to a fungal host in the Tolypocladium MRCA, in addition to further reversals to insect hosts within that clade, has been reported for this genus (Spatafora et al. 2007, Sung et al. 2008). Our study supports the aforementioned scenario (FIG. 2). Based on our phylogenetic analyses, the novel endophytic lineages appeared late in the diversification of the Tolypocladium clade. Unfortunately, even with the inclusion of eight loci, we were unable to infer with confidence the majority of the relationships among species within this clade. This lack of support also was reported by Sung et al. (2007a) and could be a sign a rapid diversification within this genus and also a sign of incomplete lineage sorting (ILS). Nevertheless, our analyses placed ‘‘Elaphocordyceps’’ capitata, ‘‘E.’’ fracta, ‘‘E.’’ japonica and ‘‘E.’’ longisegmentis as basal lineages to the rest of the species of Tolypocladium (BS: 78; PP: 1). These species have been found in their sexual state parasitizing species of the false truffle Elaphomyces. Ancestral state reconstruction analysis (ASR), under the parsimony model, reconstructed the MRCA of the endophytic Tolypocladium clade as having an insect host (5 entomopathogen), implying that there was a shift from mycoparasitism to entomoparasitism. Several interkingdom host shifts and reversals to ancestral states have been reported for diverse Cordyceps-like lineages (Nikoh and Fukatsu 2000, Spatafora et al. 2007, Sung et al. 2008). For instance, the ancestral state for the grass endophyte Epichloë and the mycoparasitic Tolypocladium have been reconstructed as pathogens of arthropods (Spatafora et al. 2007, Sung et al. 2008). This interkingdom host jump (from insect to plant host and from insect to GAZIS ET AL.: NOVEL LINEAGES OF ENDOPHYTIC TOLYPOCLADIUM 1099 FIG. 4A–C. Graphs comparing conidia diameter and phialides length and width. D. Micrograph showing intercalary phialides (T. endophyticum, strain 5 MX335 5 CBS 136898). E. conidiophore bearing phialides and conidia (T. endophyticum, strain 5 LA264 5 CBS 136903). F. Conidia (T. album, strain 5 MS271). G. Chlamydospore (T. amazonense, strain 5 LA100 5 CBS 137291). Abbreviations: TAL 5 T. album, TAM 5 T. amazonense, TEN 5 T. endophyticum, TTR 5 T. tropicale. fungal host) has been explained by the ‘‘host habitat hypothesis’’ (Nikoh and Fukatsu 2000). Under this hypothesis, interkingdom host jumps are facilitated by the proximity of habitats occupied by the hosts. For instance, Elaphomyces, the main host of sexual Tolypocladium (i.e. Elaphocordyceps) are hypogeous, truffle-like fungi that grow in mycorrhizal symbiosis with plant roots. This habitat is shared by coleopteran larvae and cicada nymphs, which also are hosts of Elaphocordyceps. The proximity of the mycorrhizal Elaphomyces and the cicada nymphs might have facilitated the host jump of the parasite, which later diversified as insect-associated species. Host shifts from animal to plant hosts have been reported for the lineages of the endophytic species Epichloë, the plant pathogen Claviceps and the plant parasite Ascopolyporus (Bischoff et al. 2005, Spatafora et al. 2007). Under a similar scenario and using the Nikoh’s ‘‘host habitat hypothesis’’ as framework, we propose that the lineages of endophytic Tolypocladium could have arisen by coming in contact with plant hosts through a mycorrhizal association (e.g. Tolypocladium parasitizing a mycorrhizal species) or through plantassociated insects (e.g. Tolypocladium parasitizing borer or piercing insects). Species of Elaphomyces are apparently globally distributed (Reynolds 2011). However, there are no reports of this species being associated with Hevea trees, and only recently was this genus reported from South America (Castellano et al. 2012). Most of the Elaphomyces species have been described for northern temperate forests and are known to form ectomycorrhizal (ECM) associations with plants in the Pinaceae, Fagaceae and Betulaceae. The only Elaphomyces species described from tropical lowland forest was found in association with Dipterocarpaceae in Singapore (Corner and Hawker 1953). To date, only two species of Elaphomyces have been described from South America, both associated with monodominant Dicymbe forest (Castellano et al. 2012) distributed along the Guyana Shield (ter Steege et al. 2006). Lowland Amazonian plant communities are dominated by vesicular arbuscular mycorhizae (Janos 1980). Based on the apparent absence of Elaphomyces in lowland Amazonia and on the ASR conducted in this study, we hypothesize that the MRCA of these new endophytic strains gained the ability to become endophytic (plant parasitic) from an insect parasitic ancestor. Several plant pathogenic fungi are known to be transmitted by borer or piercing-sucking insects (Mitchell 2004, Jacobi et al. 2013) and nongrass endophytes also have been reported to be dispersed by phytophagous insects (Devarajan and Suryanarayanan 2006). In addition entomopathogenic fungi have been reported affecting bark beetles (Brownbridge et al. 2010). Under any of these scenarios, the MCRA of the novel Tolypocladium lineages could have been inoculated into plant populations and become endophytic through evolutionary time. present spherical, range 2.13(2.10–2.16 mm diam living sapwood Peru present spherical, range 1.95(1.92–1.99 mm diam living sapwood Mexico, Peru, Brazil b a Geographic distribution presentb globose to subglobose, 1.5–2.5 mm diam soil, plant litter, living sapwoodb cosmopolitan Chlamydospores Conidial shape and dimensions Substrata white to pale yellow Colony pigments Data from Möller and Gams (1993) and from Bills et al. (2002). As described in this study. Antarctica absent variable, globose to ellipsoidal, range 4.5–9 3 2.5–3.5 mm lichens white Another plausible scenario involves the ecology of the anamorphs. Species of Tolypocladium frequently are isolated from diverse soil types, including tropical forest soils (Möller et al. 1996, Bills et al. 2002). The MRCA of the novel endophytic strains might have entered their host through the vascular system. Such a phenomenon has been observed under laboratory conditions with species of Trichoderma (Hypocreales) where host plants become systemically colonized from inoculum added to the soil (Bailey et al. 2008). This also has been described for soilborne fungal pathogenic species (Sukno et al. 2008, Zhang et al. 2013). present spherical, range 2.07(2.03–2.10) mm diam living sapwood and leaves Mexico, Peru open or absent, pustular to sporodochial white open or absent, pustular to sporodochial white to pale pink closed, membranous Conidiomata open or absent, pustular to sporodochial white, pink, to vinaceous gray Not reported ellipsoidal to obovate, range 2–3 3 1.5–2.5 mm dead and living plants, soil Mexico, Spain, USA closed, membranous open or absent, pustular to sporodochial white T. amazonense T. album T. pustulatum T. ovalisporum T. tropicale T. endophyticum MYCOLOGIA Character TABLE I. Distinguishing characteristics of previously described Tolypocladium (5 Chaunopycnis)a and the three novel lineages 1100 Mining the ITS database for ecological data.—From our survey of the nutritional modes/host associations of members of the Tolypocladium clade obtained through mining ITS sequences in GenBank (FIG. 3), several species were reported as inhabitants of diverse substrata such as different types of soil (i.e. from soil inside bat caves to tundra and Antarctic soils, leaf litter and ant nests) (Connell et al. 2006, Rodrigues et al. 2011, Lorch et al. 2013, Slemmons et al. 2013). However, these species also have been reported as endosymbionts of living plants, roots and marine sponges (Caballero-George et al. 2010, Rodrigues et al. 2011, Passarini et al. 2013, Wu et al. 2013). None of the Chaunopycnis-like Tolypocladium strains included in the ITS-mining analysis were isolated from insects. The only exception was GenBank FJ824614 (‘‘C. pustulata’’), which corresponded to an isolate collected from the surface of a bark beetle (Ips typographus) hibernating under the bark of a living Norway spruce (Picea abies). Isolates collected from ant nests in Texas by Rodrigues et al. (2011) clustered with the Hevea endophytes belonging to Clade IV identified as Tolypocladium album (FIG. 3). The nests harboring the isolates belong to the fungus-gardening ant Trachymyrmex septentrionalis. If T. album is a common endophyte of the plant foraged by this ant, the fungus could be transported into the nests through the leaf material. Studies have reported fungal species known as endophyte-inhabiting ant nests (Fisher et al. 1996, Rodrigues et al. 2008). The role that these species play inside the fungus gardens remains unknown, but leaf-cutting ants (Atta columbica) not only select plants that have a lower incidence of fungal endophytes but they also reduce the amount of endophytic fungi in leaves before planting them in their gardens (van Bael et al. 2009, Coblentz and Van Bael 2013). The latter suggests that some of these endophytes may represent a threat to the ant gardens (Rodrigues et al. 2005, 2008). On the other hand, it is possible that ants use the endophytes, especially those that produce secondary metabolites with antibiotic and GAZIS ET AL.: NOVEL LINEAGES OF ENDOPHYTIC TOLYPOCLADIUM antifungal properties, to protect their gardens against pathogens (i.e. against the ‘‘garden weed’’ Escovopsis) and for that reason these beneficial endophytic species are not eliminated from the colony. As in other studies (Ryberg et al. 2008, Bonito et al. 2010, Wang et al. 2011) mining the GenBank ITS database and associated metadata revealed many other interesting findings. The strain AY354236 (5 ‘‘Cordyceps’’ sp. olrim125) isolated as a xylem endophyte of Betula pendula (Lygis et al. 2004) clustered with Elaphocordyceps (5 Tolypocladium) ophioglossoides (BS: 74%). The latter is a parasite of Elaphomyces, an ectomycorrhizal associate with species of Betula. An interesting finding was that the GenBank AF284132 obtained as root endophyte from an environmental sample of ericoid mycorrhizal roots (Allen et al. 2003) appeared nested with high support (BS: 100%) within Ophiocordycipitaceae. This sequence blasted to Purpureocillium lilacinum (99% identity, 100% coverage), which has been reported as entomopathogen, pathogen of nematodes, soil inhabitant, mycoparasite and endophyte (Gupta et al. 1993, Luangsa-ard et al. 2011, Johny et al. 2012). This finding provides some clues about how fungal species with different ecological roles also can occur as endophytes. Ecological role of the novel endophytic Tolypocladium lineages.—The ecology of the Tolypocladium clade has been studied mainly based on species with known sexual stages. Tolypocladium species are mostly parasites of the truffle-like fungal genus Elaphomyces and to a lesser extent of soil-inhabiting cicada nymphs (Cicadidae, Hemiptera) and wood-inhabiting beetle larvae (Coleoptera) (Sung et al. 2007a). Members of this clade and of the Ophiocordycipitaceae are known for their symbiosis with hosts, which in most cases are host-specific. For instance, Ophiocordyceps unilateralis only parasitizes formicinae ants (Formicinae, Hymenoptera) and to date Elaphomyces species are the only fungal host reported for Tolypocladium (Spatafora et al. 2007). Asexual and sexual states of the same species appear in different ecological roles often associated with different substrates. As an example the sexual state of Tolypocladium inflatum (5 Elaphocordyceps subsessilis) is found fruiting on beetle larvae, whereas its asexual state is often isolated from soil (Hodge et al. 1996). Nevertheless, both states of one fungal species should be considered as a whole, with its ecological role, life cycle and evolutionary history evaluated from this perspective. Connecting sexual and asexual states increases our understanding of the natural history of this diverse and potentially useful fungal group. Moreover, the inclusion of asexual and sexual states in the multilocus-based ancestral character reconstruction analyses contrib- 1101 utes to our understanding of substrate association and host jumps within the clade. We were unable to determine whether the novel endophytic species of Tolypocladium have a strictly endophytic life history, i.e. obtaining their nutrition solely from the plant host, if they are saprophytic, (escaping the host at plant senescence), if they are insect or fungal parasites and/or if they produce antiherbivory chemicals that protect the plant host. These novel strains may reside inside tropical trees waiting for their true host (i.e. an insect) or they could be persisting as endophytes until they are able to sporulate, possibly after host death. As in many other hypocrealean fungi, these Tolypocladium lineages could vary their ecological role depending on the resource availability (Ownley et al. 2008). For example, these species could be facultative entomopathogens when living outside their plant host. This is the case of Metarhizium anisopliae and M. robertsii, both entomopathogenic fungi that also can live as active mutualistic endophytes (Sasan and Bidochka 2012). Other genera of entomopathogenic fungi that have been isolated as endophytes include Beauveria, Cladosporium, Clonostachys, Lecanicillium, Paecilomyces and Purpureocillium (Vega 2008). In many cases the entomopathogenic fungal endophytes are thought to deter insects by the production of secondary metabolites (Vega et al. 2008, Ownley et al. 2008). An alternative scenario could be that the Tolypocladium species reported here are incidental Hevea colonizers and do not form part of their core endophytic community (Stone et al. 2004). Based on the geographic distribution of the novel lineages (Brazil, Peru, Mexico), their presence in wild and managed settings, their close phylogenetic placement with saprotrophic and endophytic Tolypocladium species (T. album, T. pustulatum) and from the results obtained from the ITS-mining exercise, we suggest that these lineages have evolved to become true endophytes and have primarily a plant-based nutritional mode. Thomas et al. (2008) reported strains closely related to T. album and T. pustulatum as endophytes of sapwood of Theobroma cacao distributed in the wild while Giordano et al. (2009) isolated T. pustulatum from sapwood of Pinus sylvestris. Unfortunately ITS sequences were not available for any of these studies; therefore these strains could not be included in our analyses. Based on all these findings, it seems likely that these novel clades are not specific to Hevea and are able to colonize other tropical tree species. The fact that only one strain was isolated from leaf tissue does not imply that they are unable to frequently colonize foliage tissue. Their presence could have been overlooked due to the inherent biases introduced by cultured- 1102 MYCOLOGIA based collection methods. Leaves of tropical trees harbor a diverse fungal community, but fast growing fungal species are the most commonly isolated (Hyde and Soytong 2008). Species of Colletotrichum and Pestalotiopsis have been found to dominate Hevea leaf tissue, especially in agricultural settings (Gazis et al. 2011). Therefore it is likely that these novel Tolypocladium strains frequently inhabit both sapwood and leaf tissue, escaping their host to fruit after leaf senescence. Another plausible scenario, supported by our ASR analysis, is that these novel endophytes are facultative entomopathogens. The MCRA of the Tolypocladium clade was reconstructed as ‘‘insectassociated’’ (5 entomopathogenic); therefore the properties needed (i.e. genetic background) to become entomopathogenic might still be present in the extant species of this group and consequently in the three novel Tolypocladium species. This scenario is supported by the study by Carlsen (2002) in which the entomopathogenic species T. cylindrosporum was isolated from a surfaced-sterilized root of Carex capillaris, indicating that this species also might become endophytic. ACKNOWLEDGMENTS We thank Akiko Hirooka (UMD), Peter O’Halloran (UMD) and Ikenna Okafor (UMD) for their help with laboratory sample processing. We thank several people for their invaluable help in collecting trips: Janette Barrios, Maribel Espinoza and Dr Enrique Arevalo (ICT) in Peru, Dr Olinto Liparini Pereira and Dr Gisele Barata in Brazil and Dr Maribel Domingues-Domingues in Mexico. We also thank Catalina Salgado (UMD), László Nagy (Clark) and Amy Rossman (Agricultural Research Service, United States Department of Agriculture [ARS-USDA]) for their very helpful comments on this article. This project was financially supported by NSF grants DEB-925672 and DEB1019972 to P. Chaverri and grants from the Amazon Conservation Association and the Latin American Studies Center (UMD) to R. Gazis. R. Gazis worked in this article while a postdoctoral fellow in the Open Tree of Life project, supported by the NSF (grant DEB-12008809). LITERATURE CITED Allen TR, Millar T, Berch SM, Berbee ML. 2003. Culturing and direct DNA extraction find different fungi from the same ericoid mycorrhizal roots. New Phytol 160: 255–272, doi:10.1046/j.1469-8137.2003.00885.x Alvarez-Loayza P, White JF Jr, Torres MS, Balslev H, Kristiansen T, Svenning J-C, Gil N. 2011. Light converts endosymbiotic fungus to pathogen, influencing seedling survival and niche-space filling of a common tropical tree, Iriartea deltoidea. PLoS ONE 6:e16386, doi:10.1371/journal.pone.0016386 Arnold AE, Herre EA. 2003. Canopy cover and leaf age affect colonization by tropical fungal endophytes: ecological pattern and process in Theobroma cacao (Malvaceae). Mycologia 95:388–398, doi:10.2307/3761880 ———, Lutzoni F. 2007. Diversity and host range of foliar fungal endophytes: Are tropical trees biodiversity hot spots? Ecology 88:541–549, doi:10.1890/05-1459 ———, Mejı́a LC, Kyllo D, Rojas EI, Maynard Z, Robbins N, Herre EA. 2003. Fungal endophytes limit pathogen damage in a tropical tree. Proc Natl Acad Sci USA 100: 5649–15654, doi:10.1073/pnas.2533483100 Bailey BA, Bae H, Strem MD, Crozier J, Thomas SE, Samuels GJ, Vinyard BT, Holmes KA. 2008. Antibiosis, mycoparasitism and colonization success for endophytic Trichoderma isolates with biological control potential in Theobroma cacao. Biol Control 46:24–35, doi:10.1016/ j.biocontrol.2008.01.003 Bills GF, Polishook JD, Goetz MA, Sullivan RF, White JF Jr. 2002. Chaunopycnis pustulata sp. nov., a new clavicipitalean anamorph producing metabolites that modulate potassium ion channels. Mycol Prog 1: 3– 17, doi:10.1007/s11557-006-0001-3 Bischoff JF, Chaverri P, White JF. 2005. Clarification of the host substrate of Ascopolyporus and description of Ascopolyporus philodendrus sp. nov. Mycologia 97:710– 717, doi:10.3852/mycologia.97.3.710 Bissett J. 1983. Notes on Tolypocladium and related genera. Can J Bot 61:1311–1329, doi:10.1139/b83-139 Blackwell M. 2011. The Fungi: 1, 2, 3 … 5.1 million species? Am J Bot 98:426–438, doi:10.3732/ajb.1000298 Bonito GM, Gryganskyi AP, Trappe JM, Vilgalys R. 2010. A global meta-analysis of Tuber ITS rDNA sequences: species diversity, host associations and long-distance dispersal. Mol Ecol 19:4994–5008, doi:10.1111/j.1365-294X. 2010.04855.x Brownbridge M, Reay SD, Cummings NJ. 2010. Association of entomopathogenic fungi with exotic bark beetles in New Zealand pine plantations. Mycopathologia 169:75– 80, doi:10.1007/s11046-009-9229-1 Caballero-George C, Bolaños J, Ochoa E, Carballo JL, Cruz JA, Arnold AE. 2010. Protocol to isolate spongeassociated fungi from tropical waters and an examination of their cardioprotective potential. Curr Trends Biotechnol Pharm 4:881–889. Carlsen TA. 2002. Molecular diversity of root endophytes in an alpine Bistorta vivipara-Kobresia myosuroides tundra plant community [master’s thesis]. Norway: Univ. Oslo Press. 53 p. Castellano MA, Henkel TW, Miller SL, Smith ME, Aime MC. 2012. New Elaphomyces species (Elaphomycetaceae, Eurotiales, Ascomycota) from Guyana. Mycologia 104: 1244–1249, doi:10.3852/12-061 Castlebury LA, Rossman AY, Sung GH, Hyten AS, Spatafora JW. 2004. Multigene phylogeny reveals new lineage for Stachybotrys chartarum, the indoor air fungus. Mycol Res 108:864–872, doi:10.1017/S0953756204000607 Chaverri P, Gazis RO. 2011. Linking ex planta fungi with their endophytic stages: Perisporiopsis, a common leaf litter and soil fungus, is a frequent endophyte of Hevea GAZIS ET AL.: NOVEL LINEAGES OF ENDOPHYTIC TOLYPOCLADIUM spp. and other plants. Fungal Ecol 4: 94– 102, doi:10.1016/j.funeco.2010.09.003 ———, Samuels GJ. 2013. Evolution of habitat preference and nutrition mode in a cosmopolitan fungal genus with evidence of interkingdom host jumps and major shifts in ecology. Evolution 67:2823–2837. Coblentz KE, van Bael SA. 2013. Field colonies of leafcutting ants select plant materials containing low abundances of endophytic fungi. Ecosphere 4:66, doi:10.1890/ES13-00012.1 Connell L, Redman R, Craig S, Rodriguez R. 2006. Distribution and abundance of fungi in the soils of Taylor Valley, Antarctica. Soil Biol Biochem 38:3083– 3094, doi:10.1016/j.soilbio.2006.02.016 Corner EJH, Hawker LE. 1953. Hypogeous fungi from Malaya. J Br Mycol Soc 36:125–137, doi:10.1016/S00071536(53)80057-4 Devarajan PT, Suryanarayanan TS. 2006. Evidence for the role of phytophagous insects in dispersal of non-grass fungal endophytes. Fungal Divers 23:111–119. Eaton CJ, Cox MP, Scott B. 2011. What triggers grass endophytes to switch from mutualism to pathogenism? Plant Sci 180:190–195, doi:10.1016/j.plantsci.2010.10.002 Evans HC, Holmes KA, Thomas SE. 2003. Endophytes and mycoparasites associated with an indigenous forest tree, Theobroma gileri, in Ecuador and a preliminary assessment of their potential as biocontrol agents of cocoa diseases. Mycol Prog 2:149–160, doi:10.1007/ s11557-006-0053-4 Faeth SH, Saari S. 2012. Fungal grass endophytes and arthropod communities: lessons from plant defense theory and multitrophic interactions. Fungal Ecol 5: 364–371, doi:10.1016/j.funeco.2011.09.003 Fisher PJ, Stradling DJ, Sutton BC, Petrini LE. 1996. Microfungi in the fungus gardens of the leafcutting ant Atta cephalotes: a preliminary study. Mycol Res 100: 541–546, doi:10.1016/S0953-7562(96)80006-2 Gams W. 1980. Chaunopycnis alba, gen. et sp. nov., a soil fungus intermediate between Moniliales and Sphaeropsidales. Persoonia 11:75–79. Gazis R. 2012. Evaluating the endophytic fungal community in planted and wild rubber trees (Hevea brasiliensis) [doctoral dissertation]. College Park: Univ. Maryland Press. 264 p. ———, Chaverri P. 2010. Diversity of fungal endophytes in leaves and stems of rubber trees (Hevea brasiliensis) in Tambopata, Peru. Fun Ecol 3:240–254, doi:10.1016/ j.funeco.2009.12.001 ———, Miadlikowska J, Lutzoni F, Arnold AE, Chaverri P. 2012. Culture-based study of endophytes associated with rubber trees in Peru reveals a new class of Pezizomycotina: Xylonomycetes. Mol Phylogenet Evol 65:294–304, doi:10.1016/j.ympev.2012.06.019 ———, Rehner S, Chaverri P. 2011. Species delimitation in fungal endophyte diversity studies and its implications in ecological and biogeographic inferences. Mol Ecol 20:3001–3013, doi:10.1111/j.1365-294X.2011.05110.x Giordano L, Gonthier P, Varese GC, Miserere L, Nicolotti G. 2009. Mycobiota inhabiting sapwood of healthy and 1103 declining Scots pine (Pinus sylvestris L.) trees in the Alps. Fungal Divers 38:69–83. Gundel PE, Omacini M, Sadras VO, Ghersa CM. 2010. The interplay between the effectiveness of the grass-endophyte mutualism and the genetic variability of the host plant. Evol Appl 3:538–546, doi:10.1111/j.1752-4571. 2010.00152.x Gupta SC, Leathers TD, Wicklow DT. 1993. Hydrolytic enzymes secreted by Paecilomyces lilacinus cultured on sclerotia of Aspergillus flavus. Appl Microbiol Biot 39: 99–103, doi:10.1007/BF00166856 Hodge KT, Krasnoff SB, Humber RA. 1996. Tolypocladium inflatum is the anamorph of Cordyceps subsessilis. Mycologia 88:715–719, doi:10.2307/3760965 Huson DH, Bryant D. 2006. Application of phylogenetic networks in evolutionary studies. Mol Biol Evol 23:254– 267, doi:10.1093/molbev/msj030 Hyde KD, Soytong K. 2008. The fungal endophyte dilemma. Fungal Divers 33:163–173. Jacobi WR, Koski RD, Negron JF. 2013. Dutch elm disease pathogen transmission by the banded elm bark beetle Scolytus schevyrewi. For Pathol 43: 232– 237, doi:10.1111/efp.12023 Janos DP. 1980. Mycorrhizae influence tropical succession. Biotropica 12:56–64, doi:10.2307/2388157 Johny S, Kyei-Poku G, Gauthier D, van Frankenhuyzen K. 2012. Isolation and characterization of Isaria farinosa and Purpureocillium lilacinum associated with emerald ash borer, Agrilus planipennis in Canada. Biocontrol Sci Techn 22:723–732, doi:10.1080/09583157.2012.677808 Katoh K, Asimenos G, Toh H. 2009. Multiple alignments of DNA sequences with MAFFT. Methods Mol Biol 537: 39–64, doi:10.1007/978-1-59745-251-9_3 Kogel KH, Franken P, Hückelhoven R. 2006. Endophyte or parasite—What decides? Curr Opin Plant Biol 9:358– 363, doi:10.1016/j.pbi.2006.05.001 Liu K, Warnow TJ, Holder MT, Nelesen SM, Yu J, Stamatakis AP, Linder CR. 2012. SATe-II: very fast and accurate simultaneous estimation of multiple sequence alignments and phylogenetic trees. Syst Biol 61:90–106, doi:10.1093/sysbio/syr095 Lorch JM, Lindner DL, Gargas A, Muller LK, Minnis AM, Blehert DS. 2013. A culture-based survey of fungi in soil from bat hibernacula in the eastern United States and its implications for detection of Geomyces destructans, the causal agent of bat white-nose syndrome. Mycologia 105:237–252, doi:10.3852/12-207 Luangsa-ard J, Houbraken J, van Doorn T, Hong SB, Borman AM, Hywel-Jones NL, Samson RA. 2011. Purpureocillium, a new genus for the medically important Paecilomyces lilacinus. FEMS Microbiol Lett 321: 141–149, doi:10.1111/j.1574-6968.2011.02322.x Lygis V, Vasiliauskas R, Stenlid J. 2004. Planting Betula pendula on pine sites infested by Heterobasidion annosum: disease transfer, silvicultural evaluation and community of wood-inhabiting fungi. Can J Forest Res 34:120–130, doi:10.1139/x03-202 Maddison WP, Maddison DR. 2011. Mesquite 2.75: a modular system for evolutionary analysis. Available from http://mesquiteproject.org 1104 MYCOLOGIA Mitchell PL. 2004. Heteroptera as vectors of plant pathogens. Neotrop Entomol 33:519–545, doi:10.1590/ S1519-566X2004000500001 Möller C, Gams W. 1993. Two new hyphomycetes isolated from Antarctic lichens. Mycotaxon 48:441–450. ———, Weber G, Dreyfuss MM. 1996. Intraspecific diversity in the fungal species Chaunopycnis alba: Implications for microbial screening programs. J Ind Microbiol Biot 17:359–372, doi:10.1007/BF01574767 Nikoh N, Fukatsu T. 2000. Interkingdom host jumping underground: phylogenetic analysis of entomoparasitic fungi of the genus Cordyceps. Mol Biol Evol 17:629–638, doi:10.1093/oxfordjournals.molbev.a026341 Nirenberg HI. 1976. Untersuchungen über die morphologische und biologische Differenzierung in der Fusarium-Section Liseola. Berlin-Dahlem: Mitt Biol Bundesanst für Land- und Forstwirt, 169. p 1–117. Oses R, Valenzuela S, Freer J, Sanfuentes E, Rodriguez J. 2008. Fungal endophytes in xylem of healthy Chilean trees and their possible role in early wood decay. Fungal Divers 33:77–86. Ownley BH, Griffin MR, Klingeman WE, Gwinn KD, Moulton JK, Pereira RM. 2008. Beauveria bassiana: endophytic colonization and plant disease control. J Invertebr Pathol 98: 267– 270, doi:10.1016/j.jip. 2008.01.010 Pagel M, Meade A. 2007. BayesTraits 1.0 computer package. Available from http://www. evolution.reading.ac.uk/ BayesTraits.html ———, ———, Barker D. 2004. Bayesian estimation of ancestral character states on phylogenies. Syst Biol 53: 673–684, doi:10.1080/10635150490522232 Parfitt D, Hunt J, Dockrell D, Rogers HJ, Boddy L. 2010. Do all trees carry the seeds of their own destruction? PCR reveals numerous wood decay fungi latently present in sapwood of a wide range of angiosperm trees. Fungal Ecol 3:338–346, doi:10.1016/j.funeco.2010.02.001 Passarini MR, Santos C, Lima N, Berlinck RG, Sette LD. 2013. Filamentous fungi from the Atlantic marine sponge Dragmacidon reticulatum. Arch Microbiol 195: 99–111, doi:10.1007/s00203-012-0854-6 Promputtha I, Hyde KD, McKenzie EHC, Peberdy JF, Lumyong S. 2010. Can leaf degrading enzymes provide evidence that endophytic fungi becoming saprobes? Fungal Divers 41:89–99, doi:10.1007/s13225-010-0024-6 Quandt AC, Kepler RM, Gams W, Araújo JPM, Ban S, Evans HC, Hughes D, Humber R, Hywel-Jones N, Li Z, Luangsa-ard JJ, Rehner SA, Sanjuan T, Sato H, Shrestha B, Sung G-H, Yao Y-J, Zare R, Spatafora JW. 2014. Phylogenetic-based nomenclatural proposals for Ophiocordycipitaceae (Hypocreales) with new combinations in Tolypocladium. IMA Fungus (in press). Rambaut A, Drummond AJ. 2007. Tracer 1.4. Available from http://beast.bio.ed.acuk/Tracer Reeb V, Lutzoni F, Roux C. 2004. Contribution of RPB2 to multilocus phylogenetic studies of the euascomycetes (Pezizomycotina, Fungi) with special emphasis on the lichen-forming Acarosporaceae and evolution of polyspory. Mol Phylogenet Evol 32: 1036– 1060, doi:10.1016/j.ympev.2004.04.012 Reynolds HT. 2011. Systematics, phylogeography and ecology of Elaphomycetaceae [doctoral dissertation]. Durham, North Carolina: Duke Univ. Press. 182 p. Rocha AC, Garcia D, Uetanabaro AP, Carneiro RT, Araújo IS, Mattos CR, Góes-Neto A. 2011. Foliar endophytic fungi from Hevea brasiliensis and their antagonism on Microcyclus ulei. Fungal Divers 47:75–84, doi:10.1007/ s13225-010-0044-2 Rodrigues A, Bacci M Jr, Mueller UG, Ortiz A, Pagnocca FC. 2008. Microfungal ‘‘weeds’’ in the leafcutter ant symbiosis. Microb Ecol 56:604–614, doi:10.1007/ s00248-008-9380-0 ———, Mueller UG, Ishak HD, Bacci M Jr, Pagnocca FC. 2011. Ecology of microfungal communities in gardens of fungus-growing ants (Hymenoptera: Formicidae): a year-long survey of three species of attine ants in central Texas. FEMS Microbiol Ecol 78:244–255, doi:10.1111/j.1574-6941.2011.01152.x ———, Pagnocca FC, Bueno OC, Pfenning LH, Bacci M Jr. 2005. Assessment of microfungi in fungus gardens free of the leaf-cutting ant Atta sexdens rubropilosa (Hymenoptera: Formicidae). Sociobiology 46:329–334. Rodriguez RJ, White JF, Arnold AE, Redman RS. 2009. Fungal endophytes: diversity and functional roles. New Phytol 182:314–330, doi:10.1111/j.1469-8137. 2009.02773.x Rojas EI, Rehner SA, Samuels GJ, van Bael S, Herre EA, Cannon PF, Chen R, Pang J, Wang RW, Zhang Y, Peng YQ, Sha T. 2010. Colletotrichum gloeosporioides s.l. associated with Theobroma cacao and other plants in Panamá: multilocus phylogenies distinguish host-associated pathogens from asymptomatic endophytes. Mycologia 102:1318–1338, doi:10.3852/09-244 Ronquist F, Huelsenbeck JP. 2003. MrBayes 3: Bayesian phylogenetic inference under mixed models. Bioinformatics 19:1572–1574, doi:10.1093/bioinformatics/ btg180 Rubini MR, Silva-Ribeiro RT, Pomella AWV, Maki CS, Araújo WL, Dos Santos DR, Azevedo JL. 2005. Diversity of endophytic fungal community of cacao (Theobroma cacao L.) and biological control of Crinipellis perniciosa, causal agent of witches’ broom disease. Int J Biol Sci 1: 24–33, doi:10.7150/ijbs.1.24 Ryberg M, Kristiansson E, Sjökvist E, Nilsson RH. 2009. An outlook on the fungal internal transcribed spacer sequences in GenBank and the introduction of a webbased tool for the exploration of fungal diversity. New Phytol 181:471–477, doi:10.1111/j.1469-8137.2008. 02667.x ———, Nilsson RH, Kristiansson E, Töpel M, Jacobsson S, Larsson E. 2008. Mining metadata from unidentified ITS sequences in GenBank: a case study in Inocybe (Basidiomycota). BMC Evol Biol 8:50, doi:10.1186/ 1471-2148-8-50 Samson RA, Soares GG. 1984. Entomopathogenic species of the hyphomycete genus Tolypocladium. J Invertebr Pathol 43:133–139, doi:10.1016/0022-2011(84)90130-7 Samuels GJ, Suarez C, Solis K, Holmes KA, Thomas SE, Ismaiel A, Evans HC. 2006. Trichoderma theobromicola and T. paucisporum: two new species isolated from GAZIS ET AL.: NOVEL LINEAGES OF ENDOPHYTIC TOLYPOCLADIUM cacao in South America. Mycol Res 110:381–392, doi:10.1016/j.mycres.2006.01.009 Sasan RK, Bidochka MJ. 2012. The insect-pathogenic fungus Metarhizium robertsii (Clavicipitaceae) is also an endophyte that stimulates plant root development. Am J Bot 99:101–107, doi:10.3732/ajb.1100136 Schulz B, Boyle C. 2005. The endophytic continuum. Mycol Res 109:661–686, doi:10.1017/S095375620500273X Seifert K, Morgan-Jones G, Gams W, Kendrick B. 2011. The genera of Hyphomycetes. Utrecht, the Netherlands: CBS-KNAW Fungal Biodiversity Centre. 997 p. Silvestro D, Michalak I. 2012. raxmlGUI: a graphical frontend for RAxML. Org Divers Evol 12:335–337, doi:10.1007/s13127-011-0056-0 Slemmons C, Johnson G, Connell LB. 2013. Application of an automated ribosomal intergenic spacer analysis database for identification of cultured Antarctic fungi. Antarct Sci 25:44–50, doi:10.1017/S0954102012000879 Slippers B, Wingfield MJ. 2007. Botryosphaeriaceae as endophytes and latent pathogens of woody plants: diversity, ecology and impact. Fungal Biol Rev 21:90– 106, doi:10.1016/j.fbr.2007.06.002 Spatafora JW, Sung GH, Sung JM, Hywel-Jones NL, White JF. 2007. Phylogenetic evidence for an animal pathogen origin of ergot and the grass endophytes. Mol Ecol 16:1701–1711, doi:10.1111/j.1365-294X.2007.03225.x Stamatakis A. 2006. RAxML-VI-HPC: maximum likelihoodbased phylogenetic analyses with thousands of taxa and mixed models. Bioinformatics 22: 2688– 2690, doi:10.1093/bioinformatics/btl446 Stensrud Ø, Hywel-Jones NL, Schumacher T. 2005. Toward a phylogenetic classification of Cordyceps: ITS nrDNA sequence data confirm divergent lineages and paraphyly. Mycol Res 109:41–56, doi:10.1017/S095375620400139X Stone JK, Polishook JD, White JRJ. 2004. Endophytic fungi. In: Biodiversity of fungi: inventory and monitoring methods. Elsevier. p 241–270. Sukno SA, Garcı́a VM, Shaw BD, Thon MR. 2008. Root infection and systemic colonization of maize by Colletotrichum graminicola. Appl Environ Microb 74: 823–832, doi:10.1128/AEM.01165-07 Sung GH, Hywel-Jones NL, Sung JM, Luangsa-ard JJ, Shrestha B, Spatafora JW. 2007a. Phylogenetic classification of Cordyceps and the clavicipitaceous fungi. Stud Mycol 57:5–59, doi:10.3114/sim.2007.57.01 ———, Poinar GO, Spatafora JW. 2008. The oldest fossil evidence of animal parasitism by fungi supports a Cretaceous diversification of fungal–arthropod symbioses. Mol Phylogenet Evol 49:495–502, doi:10.1016/ j.ympev.2008.08.028 ———, Sung JM, Hywel-Jones NL, Spatafora JW. 2007b. A multigene phylogeny of Clavicipitaceae (Ascomycota, 1105 Fungi): identification of localized incongruence using a combinational bootstrap approach. Mol Phylogenet Evol 44:1204–1223, doi:10.1016/j.ympev.2007.03.011 Talavera G, Castresana J. 2007. Improvement of phylogenies after removing divergent and ambiguously aligned blocks from protein sequence alignments. Syst Biol 56:564–577, doi:10.1080/10635150701472164 Taylor JW, Jacobson DJ, Kroken S, Kasuga T, Geiser DM, Hibbett DS, Fisher MC. 2000. Phylogenetic species recognition and species concepts in fungi. Fungal Genet Biol 31:21–32, doi:10.1006/fgbi.2000.1228 ter Steege H, Pitman N, Phillips OL, Chave J, Sabatier D, Duque A, Molino J-F, Prevost M-F, Spichiger R, Castellanos H, von Hildebrand P, Vasquez R. 2006. Continental-scale patterns of canopy tree composition and function across Amazonia. Nature 443:444–447, doi:10.1038/nature05134 Thomas SE, Crozier J, Aime MC, Evans HC, Holmes KA. 2008. Molecular characterization of fungal endophytic morphospecies associated with the indigenous forest tree, Theobroma gileri, in Ecuador. Mycol Res 112:852– 860, doi:10.1016/j.mycres.2008.01.008 van Bael SA, Fernández-Marı́n H, Valencia MC, Rojas EI, Wcislo WT, Herre EA. 2009. Two fungal symbioses collide: endophytic fungi are not welcome in leafcutting ant gardens. Proc R Soc B London 276:2419– 2426, doi:10.1098/rspb.2009.0196 Vega FE. 2008. Insect pathology and fungal endophytes. J Invertebr Pathol 98: 277– 279, doi:10.1016/j.jip. 2008.01.008 ———, Posada F, Aime MC, Pava-Ripoll M, Infante F, Rehner SA. 2008. Entomopathogenic fungal endophytes. Biol Control 46:72–82, doi:10.1016/j.biocontrol. 2008.01.008 Wang Z, Nilsson RH, Lopez-Giraldez F, Zhuang WY, Dai YC, Johnston PR, Townsend JP. 2011. Tasting soil fungal diversity with earth tongues: phylogenetic test of SATé alignments for environmental ITS data. PloS ONE 6: e19039, doi:10.1371/journal.pone.0019039 William L, Nascimento HEM, Laurance SG, Condit R, D’Angelo S, Andrade A. 2004. Inferred longevity of Amazonian rainforest trees based on a long-term demographic study. Forest Ecol Manag 190:131–143, doi:10.1016/j.foreco.2003.09.011 Wu L, Han T, Li W, Jia M, Xue L, Rahman K, Qin L. 2013. Geographic and tissue influences on endophytic fungal communities of Taxus chinensis var. mairei in China. Curr Microbiol 66:40–48, doi:10.1007/s00284-012-0235-z Zhang WW, Jiang TF, Cui X, Qi FJ, Jian GL. 2013. Colonization in cotton plants by a green fluorescent protein-labeled strain of Verticillium dahliae. Eur J Plant Pathol 135:867–876, doi:10.1007/s10658-012-0131-1