ABSTRACT
Genomic sequence data indicate that certain fungi in the genus Metarhizium have the capacity to produce lysergic acid-derived ergot alkaloids, but accumulation of ergot alkaloids in these fungi has not been demonstrated previously. We assayed several Metarhizium species grown under different conditions for accumulation of ergot alkaloids. Isolates of M. brunneum and M. anisopliae accumulated the lysergic acid amides lysergic acid α-hydroxyethyl amide, ergine, and ergonovine on sucrose-yeast extract agar but not on two other tested media. Isolates of six other Metarhizium species did not accumulate ergot alkaloids on sucrose-yeast extract agar. Conidia of M. brunneum lacked detectable ergot alkaloids, and mycelia of this fungus secreted over 80% of their ergot alkaloid yield into the culture medium. Isolates of M. brunneum, M. flavoviride, M. robertsii, M. acridum, and M. anisopliae produced high concentrations of ergot alkaloids in infected larvae of the model insect Galleria mellonella, but larvae infected with M. pingshaense, M. album, M. majus, and M. guizhouense lacked detectable ergot alkaloids. Alkaloid concentrations were significantly higher when insects were alive (as opposed to killed by freezing or gas) at the time of inoculation with M. brunneum. Roots of corn and beans were inoculated with M. brunneum or M. flavoviride and global metabolomic analyses indicated that the inoculated roots were colonized, though no ergot alkaloids were detected. The data demonstrate that several Metarhizium species produce ergot alkaloids of the lysergic acid amide class and that production of ergot alkaloids is tightly regulated and associated with insect colonization.
IMPORTANCE Our discovery of ergot alkaloids in fungi of the genus Metarhizium has agricultural and pharmaceutical implications. Ergot alkaloids produced by other fungi in the family Clavicipitaceae accumulate in forage grasses or grain crops; in this context they are considered toxins, though their presence also may deter or kill insect pests. Our data report ergot alkaloids in Metarhizium species and indicate a close association of ergot alkaloid accumulation with insect colonization. The lack of accumulation of alkaloids in spores of the fungi and in plants colonized by the fungi affirms the safety of using Metarhizium species as biocontrol agents. Ergot alkaloids produced by other fungi have been exploited to produce powerful pharmaceuticals. The class of ergot alkaloids discovered in Metarhizium species (lysergic acid amides) and their secretion into the growth medium make Metarhizium species a potential platform for future studies on ergot alkaloid synthesis and modification.
INTRODUCTION
Ergot alkaloids are a diverse family of fungus-specialized metabolites, including simple clavine ergot alkaloids and more complex lysergic acid derivatives. Lysergic acid-derived ergot alkaloids, in particular, have been very important in agriculture and in medicine. Amides and peptides containing lysergic acid have contaminated grain and forage crops resulting in human and animal toxicoses (1–4). Additionally, derivatives of lysergic acid are used to synthesize a range of pharmaceuticals used to treat dementia, migraines, Parkinson’s disease, and hyperprolactinemia (5–8). Lysergic acid-derived ergot alkaloids are produced by representatives of several genera in the fungal family Clavicipitaceae, including Claviceps spp., Epichloë spp., Balansia spp., and Periglandula spp. (4, 9). Each of these fungi is closely associated with plants in symbioses that range from mutualism to parasitism (10, 11). The significance of ergot alkaloids to these plant-fungal symbioses is still being studied, but data indicate roles in deterring mammalian herbivory (12–14) and deterring or killing insect herbivores (15–17).
The ecologically important fungal genus Metarhizium is a member of the Clavicipitaceae. Metarhizium species are noteworthy because of their ability to colonize plant roots in a beneficial way (18–21) and their ability to parasitize insects. In fact, formulations of several Metarhizium species are used as biocontrol agents to control insect pests (22). Examples include biocontrol of spittle bug on sugar cane in Brazil by M. anisopliae (23), biocontrol of locusts by formulations of M. acridum (e.g., Green Muscle and Green Guard) in Africa and Australia (24, 25), as well as controlled trials of malaria-vectoring mosquitoes in Burkina Faso by genetically modified M. pingshaense (26). Significant previous research on Metarhizium species indicate that hydrolytic enzymes (19, 27–29) and a class of toxins called destruxins (30) are both important in pathogenesis of insect hosts.
Although Metarhizium species have not been demonstrated previously to contain ergot alkaloids, genomes of M. robertsii ARSEF 23 (an isolate originally classified as M. anisopliae) and M. acridum CQMa 102 have been observed to contain a cluster of genes homologous to those required for ergot alkaloid biosynthesis in other fungi (31) (Fig. 1). The genes involved in ergot alkaloid synthesis are found adjacent to each other in clusters (4, 9, 32, 33), and the inclusion or exclusion of certain genes provides insight into the biosynthetic capacity of the fungus. The gene clusters of M. robertsii and M. acridum resemble those of fungi that produce lysergic acid amides in having a core set of eight genes required to produce lysergic acid (shown in black in Fig. 1), in addition to genes that have been demonstrated (i.e., lpsB and lpsC) (34, 35) or hypothesized (i.e., easO and easP) (4, 9, 32) to encode enzymes involved in derivatizing lysergic acid into amide forms (Fig. 1 and 2). Genomes of six additional Metarhizium species have been sequenced and are publicly available (36–38), though their capacity for ergot alkaloid synthesis has not been investigated.
Ergot alkaloid synthesis (eas) gene clusters from several fungi. Arrows indicate genes controlling steps in the ergot alkaloid pathway. The eight genes drawn in black and common to all clusters are required to produce lysergic acid. Genes drawn in color (refer to key) are required or hypothesized to derivatize lysergic acid into amides or peptides. The major products of each cluster and their relative abundance are indicated in parentheses. The inset key (upper right) shows the association of particular genes with specific ergot alkaloids. With the exceptions of M. majus ARSEF 297 (37) and M. rileyi ARSEF 4871 (38) (described in the text), all sequenced isolates of Metarhizium species contain an ergot alkaloid synthesis cluster with genes in the same order and orientation as pictured. The clusters have been redrawn from data presented by Gao et al. (31) and Schardl et al. (32, 33).
Pathway to synthesis of lysergic acid amides. Lysergic acid results from the activities of enzymes encoded by eight genes (dmaW, easF, easE, easC, easD, easA, easG, and cloA) (4, 9). Genes that control the modification of lysergic acid or its derivatives are indicated at relevant steps. Dashed gray arrows indicate alternate, hypothetical routes to LAH. Asterisks in the LAH structure indicate chiral carbons that yield stereoisomers as discussed in the text. Ergine arises from the spontaneous hydrolysis of LAH (42, 44, 45). Trp, l-tryptophan; DMAPP, dimethylallylpyrophosphate.
Considering the agricultural and pharmaceutical importance of ergot alkaloids and the presence of ergot alkaloid synthesis genes in the genomes of Metarhizium species associated with plants and marketed as biocontrol agents, our objective in this study was to determine the ergot alkaloid biosynthesis capacity of several Metarhizium species. Given the multiple niches occupied by these fungi, we analyzed them for ergot alkaloids in saprotrophic culture, in insects, and in plants by high performance liquid chromatography (HPLC) and liquid chromatography-mass spectrometry (LC-MS) approaches.
RESULTS
Ergot alkaloid synthesis gene clusters in Metarhizium genomes.Since the genomes of M. robertsii ARSEF 23 and M. acridum CQMa 102 had been noted to contain ergot alkaloid synthesis (eas) gene clusters (31), we analyzed the published genomes of M. album ARSEF 1941, M. anisopliae ARSEF 549, M. brunneum ARSEF 3297, M. guizhouense ARSEF 977 (37), and M. anisopliae Ma69 (36) for ergot alkaloid synthesis genes. Each of these genomes contained an eas gene cluster homologous to, and syntenic with, the clusters previously identified in M. acridum and M. robertsii (31) (Fig. 1). We also investigated the published genomes of M. majus ARSEF 297 (37) and M. rileyi ARSEF 4871 (38). The genome of M. majus contained homologs of easO and easP, as well as a partial copy of lpsC, but lacked homologs of several other eas genes, including the genes for the first nine steps of the pathway. The genome of M. rileyi lacked homologs of all tested eas genes.
Metarhizium spp. produce ergot alkaloids in culture in an isolate-specific manner.Three of 10 tested isolates of M. brunneum accumulated the ergot alkaloids lysergic acid α-hydroxyethylamide (LAH) and its degradation product ergine, as well as lesser quantities of ergonovine, when grown on sucrose-yeast extract agar (SYE) (Fig. 3A; Table 1). These alkaloids were identified by their fluorescence properties and retention times relative to an authentic standard (for ergonovine) or standards prepared from Periglandula sp.-infected Ipomoea leptophylla (LAH and ergine) (39). LC-MS fragmentation studies supported the identity of the most abundant ergot alkaloid in M. brunneum cultures as LAH (Fig. S1 in the supplemental material), and accurate mass LC-MS identified each of the four stereoisomeric forms of LAH with the identical molecular formula (Fig. S2). The accumulation of ergot alkaloids in M. brunneum depended on the culture medium on which the fungus was grown; isolates of M. brunneum produced ergot alkaloids on SYE but not on corn meal agar or malt extract agar. Several additional Metarhizium species, including M. acridum (two isolates), M. album (three isolates), M. anisopliae (four isolates), M. flavoviride (one isolate), M. guizhouense (two isolates), M. majus (two isolates), M. pingshaense (one isolate), and M. robertsii (two isolates), were tested for accumulation of ergot alkaloids on SYE. Among these additional isolates, only M. anisopliae ARSEF 7426 accumulated detectable ergot alkaloids on SYE when analyzed by HPLC with fluorescence detection (Table 1).
High performance liquid chromatography analyses of Metarhizium species from cultures (A) and in the insect Galleria mellonella (B). Periglandula sp.-infected Ipomoea leptophylla served as a reference for the indicated lysergic acid amides. Detection was by fluorescence with excitation at 310 nm and emission at 410 nm.
Accumulation of ergot alkaloids from isolates of Metarhizium species colonizing sucrose-yeast extract agar (SYE culture) or larvae of Galleria mellonella (insect)
Secretion of ergot alkaloids from M. brunneum.When grown on SYE broth, M. brunneum ARSEF 9354 secreted the majority of its ergot alkaloids into the culture medium. Culture fluids and fungal mycelia and conidia were harvested, separated, and measured after 7 days. Ergot alkaloids were quantified from solid and liquid fractions and expressed on a total culture basis. The proportion of alkaloids secreted into the medium was 0.78, 0.85, and 0.84 for ergine, LAH, and ergonovine, respectively (Fig. 4). The proportions secreted are more than seven times greater than the proportion of ergot alkaloids secreted from another ergot alkaloid-producing fungus, Neosartorya fumigata (syn. Aspergillus fumigatus) (40).
Mean accumulation of ergot alkaloids in hyphae versus growth medium of cultures of M. brunneum. Error bars represent standard error of the mean (n = 5).
Ergot alkaloids in Metarhizium brunneum are not associated with conidia.Since conidia of the fungus N. fumigata have been documented to contain high concentrations of ergot alkaloids (41), and the conidia of Metarhizium species are produced and dispersed at very high concentrations (e.g., 5 × 1010 conidia/g in commercial product, applied at a rate of 4.6 × 1010 conidia/m2) (US-EPA registration number 70127-10; https://www3.epa.gov/pesticides/chem_search/ppls/070127-00010-20130116.pdf, accessed January 29, 2020) for the purpose of insect biocontrol, it is important to understand whether conidia contain ergot alkaloids that may present a threat to applicator or producer health. To test whether ergot alkaloids accumulated in M. brunneum conidia, conidia from cultures on SYE were collected independently of mycelia and growth medium by inverting the culture dish and collecting conidia on the inside of the dish’s lid and suspending them in methanol. No ergot alkaloids were detected in three separate samples of M. brunneum conidia weighing 10 mg, 14 mg, or 15 mg. The fluorescence HPLC method employed in these assays has a limit of detection of 5 fmol per mg of tissue. The corresponding cultures (containing mycelia of the fungus embedded in colonized SYE medium) were checked and found to contain LAH at a mean concentration of 0.8 μM (± 0.1 μM SE), which corresponds to approximately 300 pmol per SYE culture volume required to yield 10 mg of conidia.
Multiple Metarhizium species produce ergot alkaloids in inoculated insects.Metarhizium brunneum ARSEF 9354 was studied extensively for its ability to produce ergot alkaloids in the model insect Galleria mellonella, and isolates of several other species were also screened in G. mellonella for ergot alkaloids. In alkaloid extracts obtained from M. brunneum ARSEF 9354-infected larvae, we observed stereoisomers of LAH eluting at 37 min and 38 min to be highly abundant in fresh extracts relative to the 41- and 43-min isomers (Fig. 3B). The isomers of LAH eluting at 37 min and 38 min remained in approximately equal proportions relative to each other over time (Fig. S3). To quantify these observations and make inferences on the nature of the isomers, we measured proportions of all isomers over time in methanol extracts. The 37-min and 38-min isomers comprised 95% of the total yield of LAH in the first measurement of each sample after extraction (recorded from 5 to 14.5 h after extraction) and declined over 12 days at room temperature to 69% of total the LAH measured (Fig. S3). There was no measurable effect of time on the ratio of 37-min to 38-min isomers (P = 0.32) (Fig. S3), whereas the ratio of 41-min and 43-min isomers to the 37-min and 38-min isomers increased over time (P < 0.0001). These data indicate that the 37-min and 38-min isomers are the forms found naturally in the fungi and that over time in methanol some proportion of them isomerizes to the 41- and 43-min isomers. This pattern is typical of lysergic acid derivatives in protic solvents and is due to keto-enol tautomerization at the carbon atom (customarily called carbon 8) where the amide side branch attaches to the ergoline ring system (marked with an asterisk in Fig. 2) (42, 43).
We observed ergine, the simple amide of lysergic acid, and its stereoisomer erginine in some but not all alkaloid preparations of Metarhizium species. Ergine is not a biosynthetic endpoint in itself but rather a hydrolysis product of LAH and some additional ergot alkaloids (42, 44, 45). We measured proportions of ergine/erginine relative to LAH (sum total of all four stereoisomers) over time in M. brunneum ARSEF 9354 methanolic extracts (Fig. S3). Consistent with previous observations on its origin, ergine constituted 1% of the total of ergine plus LAH at the time of the initial measurement (which was recorded from 5 to 14.5 h after extraction) and that proportion increased over time to 17% after 12 days of the extracts incubating at room temperature.
When spore suspensions of 26 isolates representing nine species of Metarhizium were inoculated into larvae of G. mellonella, ergot alkaloid accumulation varied by fungal species (Table 1). Inoculated insects died 2 to 3 days after inoculation and were analyzed 7 to 10 days after inoculation by HPLC. No control insects (inoculated with a spore-free solution) contained ergot alkaloids. Among isolates from the former M. anisopliae species complex (i.e., the clade containing M. anisopliae, M. brunneum, M. robertsii, and M. pingshaense) (46), 14 of 16 tested isolates accumulated ergot alkaloids. For the five species (10 isolates total) tested outside that clade, ergot alkaloids were detected in only M. flavoviride and M. acridum. No isolates of M. album (including the sequenced isolate ARSEF 1941, which was tested here) or M. guizhouense accumulated ergot alkaloids in insects, despite the reference genome for each of these species (37) containing the requisite cluster of genes. Both tested isolates of M. majus lacked ergot alkaloids, consistent with the genome of M. majus ARSEF 297 as described by Hu et al. (37) (not tested for alkaloids here) containing only a partial ergot alkaloid synthesis cluster. The profile of ergot alkaloids produced by M. acridum, M. flavoviride, and M. robertsii (species that had not produced ergot alkaloids in culture) was like that of M. anisopliae and M. brunneum in containing LAH, ergonovine, and ergine.
The mean molarity of LAH in M. brunneum ARSEF 9354-inoculated larvae was 154 μM (± 58 μM standard error of the mean [SE]), approximately 200-fold greater than the concentration that accumulated in SYE cultures of that same isolate of M. brunneum. A time course of alkaloid accumulation in G. mellonella larvae infected with M. brunneum ARSEF 9354 showed that ergot alkaloids were first detected at day three postinoculation and peaked at day 11 before declining in concentration (Fig. 5). Inoculation of larvae that were killed by freezing at −20°C and thawing or by prolonged exposure to N2 gas prior to injection with M. brunneum ARSEF 9354 conidia resulted in a significant reduction (P < 0.05) in the amount of ergot alkaloids produced per mass of fungus as estimated by fungus-associated ergosterol content in infected larvae (Fig. 6).
Mean accumulation of ergot alkaloids over time in larvae (n = 4) infected with M. brunneum ARSEF 9354. Insects died 2 to 3 days after inoculation. Relative proportion of ergine to LAH at day 5 indicates extra hydrolysis on samples from that date. Data are expressed as micrograms per larva as opposed to micrograms/mass of larva, because larval mass changed significantly over time as dead larvae dehydrated.
Mean accumulation of ergot alkaloids (sum of LAH, ergine, and ergonovine) in larvae alive at the time of inoculation compared to larvae killed prior to inoculation by freezing or N2 gas. Treatments marked with different letters were significantly different in a Tukey’s test (P < 0.05). Error bars represent standard error of the mean (n = 5).
Several brood V and brood VIII periodical cicadas (Magicicada spp.) naturally infected with a Metarhizium species were analyzed and found to contain a profile of ergot alkaloids typical of that observed in infected G. mellonella (Fig. S4), though with a higher proportion of ergine to LAH, perhaps due to the unknown and uncontrolled duration and conditions between infection and analysis. This observation indicates that adult insects representing a different order from G. mellonella (Hemiptera as opposed to Lepidoptera) are also suitable hosts for ergot alkaloid accumulation.
Lack of ergot alkaloid accumulation in corn and bean seedlings inoculated with Metarhizium species.Metarhizium brunneum is a common rhizosphere inhabitant (21, 47–50) and has been demonstrated to colonize corn roots and increase plant growth under experimental conditions (51). Metarhizium flavoviride also has been associated with roots (48, 49). Neither M. brunneum nor M. flavoviride accumulated ergot alkaloids when inoculated onto young seedlings of bean (Phaseolus vulgaris) or corn (Zea mays) grown on filter paper. Tissues samples found to be negative for ergot alkaloids when analyzed by HPLC with fluorescence detection included roots, hypocotyls, cotyledons, and first leaf tissues. Fungal mycelia were visible on the surface of the exposed root tissues of these filter paper-grown seedlings, indicating successful colonization (Fig. S5). To provide additional evidence of fungal colonization, a global metabolomics analysis was performed on three biological replicates each of M. brunneum ARSEF 9354-inoculated corn roots, M. flavoviride BC1163-inoculated roots, and non-inoculated corn roots. The data showed that corn roots treated with M. brunneum or M. flavoviride had altered accumulation of 66 metabolites compared to control plants that had not been treated with the fungi (Table 2; Table S1). Noteworthy among the compounds that could be identified on the basis of mass spectral fragmentation was the plant hormone γ-aminobutyric acid, which accumulated 20 log-fold higher in M. brunneum ARSEF 9354-infected roots and 18 log-fold higher in M. flavoviride-infected roots compared to the untreated roots (P = 0.0002). These observations indicate that the seedlings were in fact colonized by the fungi, however the plant tissue did not contain ergot alkaloids.
Numbers of compoundsd that varied between Metarhizium species-colonized corn roots compared to uninoculated corn roots by a minimum two log-fold differencee
DISCUSSION
Our data show that several Metarhizium species produce ergot alkaloids in a condition-dependent manner, an observation that has agricultural and pharmaceutical implications. Only isolates of M. brunneum and M. anisopliae accumulated alkaloids in culture, even though three other species were later shown to accumulate ergot alkaloids in insects. These observations indicate that regulation of expression of the ergot alkaloid pathway genes differs in these fungi. The lack of accumulation of ergot alkaloids in plants by M. brunneum ARSEF 9354, which produced ergot alkaloids in culture and abundant ergot alkaloids in insects, demonstrates further substrate-specific regulation of accumulation of these alkaloids. The increased accumulation when insect larvae were alive at the time of inoculation indicates a further level of regulation and suggests a role for ergot alkaloids in insect colonization.
All ergot alkaloid-positive species of Metarhizium produced lysergic acid amides: most abundantly LAH (along with its hydrolysis product ergine) and lesser quantities of ergonovine. LAH occurs in four stereoisomers as documented in previous work with Claviceps pasapali (42). The four stereoisomers result from all combinations arising from two chiral carbons: one at the attachment point of the amide side chain to the lysergic acid ring (i.e., lysergic versus isolysergic forms typical of all lysergic acid derivatives) and the other from l-Ala-derived versus d-Ala-derived hydroxylamide side chains. The relative proportions of stereoisomers derived from the attachment of the amide side chain to lysergic acid change over time due to keto-enol tautomerization in protic solvents (42, 43), whereas the stereoisomers derived from the conformation of alanine have no mechanism to change over time. For these reasons, we hypothesize that the 37-min and 38-min isomers represent the l-Ala-derived and d-Ala-derived isomers in the natural lysergic acid conformation and that they form when LAH is synthesized and subsequently have no mechanism to isomerize between d-Ala-derived and l-Ala-derived forms after synthesis. Some proportion of those initial forms of LAH (i.e., the 37-min and 38-min isomers) subsequently isomerize from natural lysergic acid conformation to isolysergic acid conformation (i.e., the 42-min and 43-min isomers) after extraction in protic solvents as previously documented for all lysergic acid derivatives (42, 43). Our proposed assignment of stereoisomers to peak pairs observed in our chromatograms agrees with the relative abundances observed by Flieger et al. (42) in their studies of C. paspali. Small quantities of ergonovine also were detected in most extracts of Metarhizium species. Ergonovine is another amide of lysergic acid and is derived by reduction of the carbonyl group of lysergyl-alanine to a primary alcohol by the enzyme LpsC (35). In our preparations from Metarhizium species, ergonovine was typically present at less than 2% of the concentration of LAH.
LAH has been detected previously in C. paspali, a pathogen of Paspalum species (Poaceae), and in several Periglandula species that grow as seed-transmitted and unculturable symbionts of morning glories (Convolvulaceae) (4, 39). The discovery of LAH in Metarhizium species, which can be genetically manipulated, presents an opportunity to functionally analyze genes required for LAH synthesis. The correlation of genes easO and easP with LAH producers (Fig. 1) indicates a potential role for these genes in later stages of LAH biosynthesis. Metarhizium brunneum may provide a more tractable system for functional analysis of these genes than the non-culturable Periglandula spp. or the plant pathogen C. paspali. Interestingly, M. brunneum ARSEF 9354 secreted the vast majority of its ergot alkaloids into the culture medium. In this way it differs from other ergot alkaloid-producing fungi, such as N. fumigata and Claviceps purpurea, which retain much of their ergot alkaloids in their mycelia, complicating extraction and purification. The ability of Metarhizium species to secrete ergot alkaloids would be an advantage for pharmaceutical production and development.
The lack of ergot alkaloids in conidia (typically used as inoculum for biocontrol) and in Metarhizium sp.-infected plants indicates that the ergot alkaloid biosynthesis capacity of these fungi should not present complications for the use of the fungi as biocontrol agents. Moreover, our metabolomic data showed differences in roots inoculated with M. brunneum and M. flavoviride indicating colonization of plant roots, but the association did not induce accumulation of ergot alkaloids. Colonization of corn was supported by metabolomics data, including the demonstration of increased quantities of the plant hormone γ-aminobutyric acid. These data show that some key metabolites such as γ-aminobutyric acid are upregulated, whereas others are downregulated in roots of corn that were treated with M. brunneum or M. flavoviride. Other metabolites varied between corn roots that were treated with different species of Metarhizium, indicating that corn responds differently to different species of Metarhizium. The lack of ergot alkaloids in infected plant tissue indicates the presence of Metarhizium species on the roots or in the rhizosphere of plants is not of concern from a safety perspective.
Metarhizium species are commonly used as mycoinsecticides to control pest insects. Data from inoculations of dead versus living G. mellonella larvae indicate that the living insect provides an environment that induces ergot alkaloid synthesis. Whether or not the ergot alkaloids play a direct role in killing of insects has not yet been established. Future genetic studies, such as analyses of planned gene knockout mutants lacking ergot alkaloids, should provide material to facilitate tests of whether ergot alkaloids play a role in insect pathogenesis.
MATERIALS AND METHODS
Fungi and culture conditions.Isolates of Metarhizium species studied and their substrates of origin are listed in Table 1. Isolates with ARSEF numbers were obtained from the United States Department of Agriculture-ARS Collection of Entomopathogenic Fungal Cultures (Ithaca, NY). Metarhizium guizhouense CBS 258.90 was obtained from the Westerdijk Fungal Biodiversity Institute (Utrecht, The Netherlands). Metarhizium flavoviride BC1163 and M. anisopliae BC1203 were isolated from the millipede Brachycybe lecontii collected in Arkansas and Oklahoma, respectively (52). Millipede-associated Metarhizium species were confirmed by sequencing the ribosomal internal transcribed spacer (ITS) (deposited as GenBank accessions MN963924 and MT040791) and a portion of the 28S rRNA gene (GenBank accession MN963928) genes. During our work with the isolate initially supplied as M. anisopliae ARSEF 9354, we noticed several discrepancies between the published M. anisopliae sequences and fragments of the ARSEF 9354 isolate we sequenced, and these observations spurred further research into the identity of this isolate. We amplified and sequenced the translation elongation factor 1α (EF1α) (GenBank accession MH712765) and ITS (GenBank accession MH711991) alleles from ARSEF 9354 and compared them to the corresponding alleles in the genomes of previously sequenced Metarhizium species. The best matches were to the genome of M. brunneum ARSEF 3297 (99.7% EF1α, 100% ITS), with our sequences providing a lower match to M. anisopliae ARSEF 549 (97.9% EF1α, 99% ITS). Query coverage was 100% in each of these comparisons. Based on these findings we concluded that ARSEF 9354 is an isolate of M. brunneum.
Two isolates of M. brunneum (ARSEF 9354 and ARSEF 3826) and one isolate of M. flavoviride (BC1163) were grown on three different culture media: sucrose-yeast extract agar (SYE), corn meal agar (CMA), and malt extract agar (MEA). Once SYE was determined to be the best of these media for promoting ergot alkaloid accumulation, isolates of additional Metarhizium species were tested exclusively on this medium. SYE contained (per liter) 20 g sucrose, 10 g yeast extract, 1 g of magnesium sulfate-heptahydrate, 2 ml of trace element solution (53), and 15 g agar. CMA contained (per liter) 2 g corn meal and 15 g agar. MEA contained (per liter) 6 g malt extract, 1.8 g maltose, 6 g glucose, 1.2 g yeast extract, and 15 g agar. All cultures were grown at room temperature (22 to 24°C) with ambient lighting for at least 7 days before testing for ergot alkaloids.
Ergot alkaloid extraction and analysis.Samples of approximately 400 μl (and consisting of colonized agar, mycelia, and conidia) were cut from colonized culture medium with the broad end of a 1,000-μl pipet tip 7 to 10 days after inoculation. Samples were suspended in 400 μl of 100% methanol in a 1.5-ml microcentrifuge tube, vortexed, and rotated end over end on a Glas-Col (Terre Haute, IN) minirotator at 40 rpm for 20 min. Methanol extracts were analyzed by HPLC with fluorescence detection based on the methods described in Panaccione et al. (43). The HPLC apparatus consisted of a Waters model 600 pump controller with an in-line degasser, a Waters model 717plus autosampler, and a Rainin (Woburn, MA) Fl2 fluorescence detector set at excitation and emission wavelengths of 310 nm and 410 nm (to detect lysergic acid derivatives), respectively, and a Linear 304 fluorescence detector with excitation at 272 nm and emission monitored at 372 nm (to detect simpler clavine-type ergot alkaloids and dihydroergot alkaloids). Approximate concentrations of LAH, ergonovine, and ergine were estimated by comparing peak area to an external standard curve prepared from ergotamine, which contains the same lysergic acid fluorophore, and adjusting for relative masses of analytes and standard.
The identities of the peak corresponding to ergonovine was determined by fragmentation compared to the ergonovine analytical standard obtained from Sigma (St. Louis, MO) in LC-MS assays. The identities of peaks corresponding to ergine and lysergic acid α-hydroxyethylamide (LAH) were determined by fragmenting analytes from M. brunneum compared to those obtained from Periglandula species-infected Ipomoea leptophylla, which had previously been demonstrated to contain these lysergic acid amides (39). LC-MS was performed in positive ion mode as described by Ryan et al. (54), and high-resolution data were obtained by injecting methanol extracts into an Accela 1290 UHPLC coupled to a Thermo Fisher Scientific Q Exactive mass spectrometer operated in positive polarity mode with a data-dependent acquisition scheme.
To assess secretion of ergot alkaloids from M. brunneum ARSEF 9354, cultures were grown from 50,000 conidia in 5 ml of SYE broth (as described above, however without the addition of agar) at room temperature in petri dishes that were 3.5 cm in diameter for 7 days. The entirety of an individual culture (hyphae, conidia, and medium) was filtered through a preweighed 0.2-μm nylon filter under vacuum to separate the fungal mycelia and spores from the culture medium. The harvested solid material was dried, weighed, and a measured sample was analyzed for ergot alkaloids (as described above) to determine the total amount of ergot alkaloids in the solid phase of that culture. The volume of the liquid phase was measured, a subsample was analyzed for ergot alkaloids as described above, and the total amount of ergot alkaloids in the liquid phase was calculated.
To assess whether ergot alkaloids were contained in or on conidia of M. brunneum, conidia were collected from the inner surface of the lids of inverted petri dishes with methanol. After incubation at room temperature for 1 h, the mass of conidia extracted was determined by collecting conidia on a preweighed 0.2-μm spin filter and weighing the filter with collected conidia after drying. The clarified methanol extracts were analyzed by fluorescence HPLC as described above.
Ergosterol was measured with a method adapted from Zill et al. (55). Methanol extracts were analyzed by reversed-phase separation with isocratic elution in 99% methanol on an Agilent 1100 HPLC instrument (Agilent Technologies, Santa Clara, CA). Ergosterol was monitored with the multiwavelength detector set at 282 nm and identified by comparison to an authentic standard (Sigma-Aldrich).
Statistical comparisons of ergot alkaloid data were made with JMP (SAS, Cary, NC). Data were checked for unequal variances with a Brown-Forsythe test, and means were compared with ANOVA and a post hoc Tukey’s test.
Insect inoculation assays.Larvae of Galleria mellonella, commonly referred to as wax worms, were purchased from New York Worms (Glen Cove, NY), and individuals were selected for infection based on appearance (adequate size and lack of black coloration, which may indicate prior infection) and mobility. Spore suspensions contained 40,000 conidia/μl in a modified phosphate-buffered saline (PBS) solution (pH 7.4) containing 9 mM Na2PO4, 1.6 mM KH2PO4, 123 mM NaCl, 0.01% wt/vol Tween 20, and 10 μg/ml rifampin. Each larva was injected with 20 μl of spore solution via an insulin syringe with a 29.5-gauge needle according to methods described by Panaccione and Arnold (56). Larvae in control groups were inoculated in the same manner with 20 μl of modified PBS solution lacking conidia. Inoculated larvae were incubated at room temperature. Once the larvae died, they were incubated routinely for 7 days at room temperature under ambient lighting. In a time course study, larval samples (n = 4) were extracted daily for up to 13 days. Their weights were then recorded and 1 ml of methanol and ten 3-mm glass beads were added. The larvae were homogenized in a Fastprep 120 bead beater (Bio101; Carlsbad, CA), rotated for 20 min on the minirotator, clarified by centrifugation, and analyzed via HPLC with fluorescence detection (as described above).
Plant inoculation assays.Corn (Zea mays) and bean (Phaseolus vulgaris) seeds were sterilized in an aqueous mixture of 20% bleach and 0.01% Tween 20 for 20 min rotating at 40 rpm on the Glas-Col (Terre Haute, IN) minirotator. Seeds were rinsed with sterilized distilled water repeatedly and placed in petri dishes containing filter paper saturated with Hoagland’s solution (57). Seeds were incubated under light that ranged from 50 to 60 μmol/m2/sec at room temperature. Seeds were illuminated for 16 h/day, and Hoagland’s solution was replenished as needed to maintain moisture. Roots were inoculated when the emerged radicle reached a length greater than that of the seed, and uninoculated roots were left as a control treatment.
Metarhizium flavoviride BC1163 and M. brunneum ARSEF 9354 were cultured for 7 days on malt extract agar prior to inoculation. Both cultures were then used to create suspensions of 200,000 conidia per ml in 0.01% Tween 20. The concentration of spores in each sample was calculated with the aid of a hemocytometer. Spore suspensions were administered directly to the roots of each plant via a micropipette.
Plant samples were grown on filter paper moistened with Hoagland’s solution and under light and temperature conditions as described above for 14 days, then were divided into roots, hypocotyls, cotyledons, and leaves. Tissues were dried and dry weight was recorded. Weighed subsamples of respective tissues were placed into 2-ml tubes along with 10 glass beads and 1 ml of methanol. Plant samples (50 mg dry weight per ml of methanol) were homogenized by bead beating as described above for insect samples. The samples were then clarified by centrifugation, and 20 μl of sample were injected into the fluorescence HPLC system described above to assay for ergot alkaloids.
Metabolomic analyses of control (untreated) corn roots, corn roots inoculated with M. brunneum ARSEF 9354, and corn roots inoculated with M. flavoviride BC1163 (n = 3 biological replicates for each) were performed in two stages. The first stage involved discovery of parent ions (analyzed in both positive ion mode and negative ion mode) that varied in abundance among treatments. These data were collected on an Agilent 6530 QToF at Protea Biosciences (Morgantown, WV). Metabolite abundance was compared in one-way ANOVAs and considered to be different if abundances varied by a log fold change factor of two (P < 0.05). In the second stage, identities of compounds yielding parent ions that varied in abundance among treatments in stage 1 were obtained in data-dependent assays in which tandem mass spectrometry (MS-MS) scans were compared in Compound Discover software (Thermo Fisher Scientific, Waltham, MA) to libraries of MS-MS data in KEGGS, BioCyc, and Metlin metabolite databases (mass tolerance 10 ppm). The data-dependent analyses were performed on a Thermo Q Exactive MS at the Shared Research Facilities at WVU.
Data availability.Millipede-associated Metarhizium species were confirmed by sequencing the ribosomal internal transcribed spacer (ITS) (deposited as GenBank accessions MN963924 and MT040791) and a portion of the 28S rRNA gene (GenBank accession MN963928) genes. During our work with the isolate initially supplied as M. anisopliae ARSEF 9354, we noticed several discrepancies between the published M. anisopliae sequences and fragments of the ARSEF 9354 isolate we sequenced, and these observations spurred further research into the identity of this isolate. We amplified and sequenced the translation elongation factor 1α (EF1α) (GenBank accession MH712765) and ITS (GenBank accession MH711991) alleles from ARSEF 9354.
ACKNOWLEDGMENTS
This research was funded by NIH grant 2R15-GM114774-2, with additional support from the USDA Hatch project NC1183. We thank Protea Biosciences and the WVU Shared Research Facilities for access to instrumentation and Matt Berger and John Plischke for infected cicadas.
Mention of trade names or commercial products in this publication is solely for the purpose of providing specific information and does not imply recommendation or endorsement by the USDA. The USDA is an equal opportunity provider and employer.
FOOTNOTES
- Received 12 February 2020.
- Accepted 3 May 2020.
- Accepted manuscript posted online 8 May 2020.
Supplemental material is available online only.
- Copyright © 2020 American Society for Microbiology.
