A Baeyer-Villiger Monooxygenase Gene Involved in the Synthesis of Lysergic Acid Amides Affects the Interaction of the Fungus Metarhizium brunneum with Insects

ABSTRACT

Several fungi, including the plant root symbiont and insect pathogen Metarhizium brunneum, produce lysergic acid amides via a branch of the ergot alkaloid pathway. Lysergic acid amides include important pharmaceuticals and pharmaceutical lead compounds and have potential ecological significance, making knowledge of their biosynthesis relevant. Many steps in the biosynthesis of lysergic acid amides have been determined, but terminal steps in the synthesis of lysergic acid α-hydroxyethylamide (LAH)—by far the most abundant lysergic acid amide in M. brunneum—are unknown. Ergot alkaloid synthesis (eas) genes are clustered in the genomes of fungi that produce these compounds, and the eas clusters of LAH producers contain two uncharacterized genes (easO and easP) not found in fungi that do not produce LAH. Knockout of easO via a CRISPR-Cas9 approach eliminated LAH and resulted in accumulation of the alternate lysergic acid amides lysergyl-alanine and ergonovine. Despite the elimination of LAH, the total concentration of lysergic acid derivatives was not affected significantly by the mutation. Complementation with a wild-type allele of easO restored the ability to synthesize LAH. Substrate feeding studies indicated that neither lysergyl-alanine nor ergonovine were substrates for the product of easO (EasO). EasO had structural similarity to Baeyer-Villiger monooxygenases (BVMOs), and labeling studies with deuterated alanine supported a role for a BVMO in LAH biosynthesis. The easO knockout had reduced virulence to larvae of the insect Galleria mellonella, indicating that LAH contributes to virulence of M. brunneum on insects and that LAH has biological activities different from ergonovine and lysergyl-alanine.

IMPORTANCE Fungi in the genus Metarhizium are important plant root symbionts and insect pathogens. They are formulated commercially to protect plants from insect pests. Several Metarhizium species, including M. brunneum, were recently shown to produce ergot alkaloids, a class of specialized metabolites studied extensively in other fungi because of their importance in agriculture and medicine. A biological role for ergot alkaloids in Metarhizium species had not been demonstrated previously. Moreover, the types of ergot alkaloids produced by Metarhizium species are lysergic acid amides, which have served directly or indirectly as important pharmaceutical compounds. The terminal steps in the synthesis of the most abundant lysergic acid amide in Metarhizium species and several other fungi (LAH) have not been determined. The results of this study demonstrate the role of a previously unstudied gene in LAH synthesis and indicate that LAH contributes to virulence of M. brunneum on insects.

INTRODUCTION

Ergot alkaloids are important specialized metabolites that have affected humankind in several capacities. They accumulate in plants infected by some fungi of the family Clavicipitaceae and adversely affect the health of animals and humans that consume them (1–3). Natural and semisynthetic derivatives of lysergic acid have been used as pharmaceuticals to treat dementia, migraines, hyperprolactinemia, and several other disorders (4–7). Ergot alkaloids are derived from 4-prenylated tryptophan through the concerted activities of several enzymes, all of which are encoded by contiguously clustered genes in the genomes of the producing fungi (1, 8, 9). Different branches of the ergot alkaloid pathway lead to different families of end products, including clavine-type ergot alkaloids and lysergic acid-derived ergot alkaloids. In some members of the Clavicipitaceae, lysergic acid is incorporated into peptides to make a family of ergot alkaloids known as ergopeptines; in others, it is incorporated into simpler lysergic acid amides such as ergonovine (called ergometrine in Europe), lysergic acid α-hydroxyethylamide (LAH), and ergine (Fig. 1). Both ergonovine and LAH are derived from an intermediate consisting of lysergic acid-alanine peptide (lysergyl-alanine) (10, 11). Ergine arises as a spontaneous hydrolysis product from LAH (12, 13) and other amide intermediates (14) in the ergot alkaloid pathway; no evidence of a role for a catalyst has been published.

FIG 1.

Steps from lysergic acid to lysergic acid amides. Genes controlling or hypothesized to control (indicated by question marks) certain steps are indicated. Dashed lines indicate hypothetical hydrolysis to ergine. R, connection to Lps3 via pantetheine; LAH, lysergic acid α-hydroxyethylamide.

The biosynthesis of ergonovine from lysergic acid and alanine has been clearly elucidated and requires the multiple activities of two peptide synthetases (Lps2 and Lps3, encoded by lpsB and lpsC, respectively) (10, 15). Lps2 recognizes lysergic acid, activates it by adenylation, and then binds it as a thioester to Lps2-bound pantetheine. Alanine is similarly activated and thioesterified by Lps3, and a peptide bond is formed between lysergic acid and alanine, with the resulting lysergyl-alanine covalently tethered to Lps3 via the thioester bond of alanine to Lps3-bound pantetheine (Fig. 1). The alanine-derived carbonyl group of lysergyl-alanine is then reduced to an alcohol by the reductase domain of Lps3, releasing ergonovine.

Whereas the synthesis of ergonovine has been clearly demonstrated (10), the synthesis of the alternate lysergic acid amide, LAH, is less well understood. Claviceps purpurea makes ergonovine as its sole lysergic acid amide. Other fungi make predominantly LAH with lesser quantities of ergonovine. These LAH producers include Claviceps paspali (an ergot fungus found on Paspalum species) (12, 13), Periglandula species that grow symbiotically with plants of the morning glory family (Convolvulaceae) (16, 17), and Metarhizium brunneum and some other Metarhizium species which grow primarily as plant root symbionts and also infect insects and produce ergot alkaloids in insects (18, 19).

Genes involved in the synthesis of ergot alkaloids are clustered in the genomes of ergot alkaloid-producing fungi (3, 9, 20, 21). The ergot alkaloid synthesis (eas) clusters of fungi that produce LAH contain two genes, easO and easP, that are conspicuously absent from the gene clusters of ergot alkaloid-producing fungi that do not make LAH (8, 18, 20). Based on DNA sequence, easO is predicted to encode a flavin-containing oxidoreductase, and easP is predicted to encode an alpha/beta fold hydrolase. We hypothesize that easO and easP control important steps, downstream of lysergyl-alanine formation, in the synthesis of LAH. In this study, we knocked out easO and observed the effects on the ergot alkaloid profile of the resulting mutant and the interactions of the mutant with insects. Our results provide novel information on the role of the product of easO (EasO) in LAH biosynthesis and on the significance of LAH to the success of M. brunneum as an entomopathogen.

RESULTS

Knockout of easO alters the ergot alkaloid profile of M. brunneum.

To engineer an easO knockout mutant, we cotransformed M. brunneum ARSEF 9354 with Cas9 complexed with an single guide RNA (sgRNA) specific for easO along with a phosphinothricin resistance (bar) construct. A preliminary high-performance liquid chromatography (HPLC) screen of resulting colonies indicated that two of eight colonies had an altered profile of lysergic acid amides with diminished LAH and increased ergonovine. We selected one of the transformants and purified it to nuclear homogeneity by culturing from single spores. We amplified the easO locus, and Sanger sequencing demonstrated disruption of the easO coding sequences and insertion of a fragment of the bar gene used as a selectable marker (see Fig. S1 in the supplemental material). This easO knockout mutant was then injected into larvae of the model insect Galleria mellonella (Lepidoptera) to increase the quantity of ergot alkaloids (18), and the phenotype of the mutant lacked LAH but featured abundant ergonovine and lysergyl-alanine (Fig. 2). Adding a wild-type copy of the easO locus back to the mutant (Fig. S2) restored its ability to produce LAH (Fig. 2). Liquid chromatography-mass spectrometry (LC-MS) confirmed that the compounds that accumulated in the easO knockout mutant were ergonovine (Fig. S3) and lysergyl-alanine (Fig. S4).

FIG 2.

HPLC analyses of extracts from G. mellonella larvae infected with M. brunneum strains. Extracts were prepared 8 days and 12 days postinoculation. Black and red lines (for wild-type and complemented strains, respectively) are superimposed over much of each chromatogram. Fluorescence detection was observed with excitation at 310 nm and emission at 410 nm. LA, lysergic acid, LAH, lysergic acid α-hydroxyethylamide.

Cultivation of M. brunneum strains in G. mellonella larvae demonstrated quantitative and time-dependent differences in ergot alkaloid profiles of the parent strain, easO knockout strain, and complemented strain. Since alkaloid values were obtained from infected larvae where fungal biomass cannot be measured directly, alkaloid data were normalized to the concentration of ergosterol as a marker of fungal biomass (Table 1). LAH was not detectable in the easO knockout, and the quantities of lysergyl-alanine and ergonovine were significantly increased in the easO knockout compared to the wild-type and complemented strains (P < 0.05). Considering the loss of LAH and increase of lysergyl-alanine and ergonovine, the total molar quantity of ergot alkaloids did not differ significantly among strains at 8 days postinfection (P = 0.57) or 12 days postinfection (P = 0.78). Comparison of the day 8 and day 12 data indicated a significant difference in the profile of the mutant over time. Lysergyl-alanine was a major component of the ergot alkaloid profile of the easO knockout at day 8 but then decreased significantly by day 12, while lysergic acid increased, presumably from catabolism of lysergyl-alanine. Ergonovine concentrations in easO knockout-infected larvae were high at both day 8 and day 12; the apparent increase from day 8 to day 12 was not statistically significant (P = 0.43).

TABLE 1.

Ergot alkaloid accumulation (μmol/g^a ergosterol) in G. mellonella larvae colonized by M. brunneum strains^e

Strainb	Ergot alkaloid accumulation (μmol/g ergosterol)e
	Lysergic acidd	Lysergyl-alanine	Ergonovined	LAHd	Ergine	Total
8 days postinoculationc
M. brunneum WT	ND A	11 ± 3 A	ND A	884 ± 221 A	19 ± 5 A	914 ± 228 A
easO KO	6 ± 2 B	761 ± 203 B	322 ± 41 B	ND B	7 ± 2 A	1096 ± 238 A
easO comp	ND A	11 ± 4 A	ND A	740 ± 161 A	18 ± 5 A	769 ± 168 A
12 days postinoculationc
M. brunneum WT	2 ± 0.6 A	ND A	ND A	776 ± 233 A	23 ± 8 A	801 ± 242 A
easO KO	302 ± 105 B	3 ± 1 B	415 ± 103 B	ND B	8 ± 1 A	728 ± 187 A
easO comp	1 ± 0.2 A	ND A	ND A	590 ± 110 A	18 ± 4 A	609 ± 114 A

^aCalculated from sums of stereoisomers and based on fluorescence relative to ergonovine.

^bWT, M. brunneum ARSEF 9354; KO, knockout; comp, complemented.

^cComparisons of means apply only within the same sampling date.

^dND, not detected; limit of detection = 0.05 μmol/g.

^eValues followed by different capital letters differ significantly (P < 0.05) in a multiple-comparison test; the nonparametric Steel-Dwass test was used for data sets containing “ND” values, and the Tukey-Kramer test was used for the remaining data sets.

Activity of EasO.

Based on the accumulation of ergonovine and lysergyl-alanine in the easO knockout and on previous work on the pathway to lysergic acid amides (10, 22), we can envision three potential substrates for EasO—ergonovine, free lysergyl-alanine, or lysergyl-alanine bound to lysergyl peptide synthetase 3 (Lps3) as a thioester which occurs as an intermediate in the biosynthesis of ergonovine. To address this question, we mutated the gene lpsC (encoding Lps3) by a CRISPR-Cas9 approach (Fig. S5) to obtain a strain incapable of producing ergonovine and LAH but retaining easO and easP, which are hypothesized to act downstream of Lps3 (Fig. 1). The lpsC mutant did not produce LAH or ergonovine but instead accumulated lysergic acid (Fig. S6). We tested the substrate-product relationship between ergonovine and LAH and free lysergyl-alanine and LAH by feeding these potential substrates the lpsC knockout cultured in vitro on sucrose-yeast extract medium. These experiments did not result in production of LAH or any detectable modification of the fed substrates, indicating that neither ergonovine nor free lysergyl-alanine is a substrate for the product of easO. The lpsC knockout continued to produce lysergic acid, indicating that the ergot alkaloid pathway was active in these cultures. We do not presently have a means for directly testing Lps3-bound lysergyl-alanine as a substrate for EasO.

The sequence of the deduced product of easO has identity with Baeyer-Villiger monooxygenases (BVMOs). The top 20 matches for EasO in a BLASTp search of the Protein Data Bank (E values ranging from 7 × 10⁻⁶⁶ to 1 × 10⁻³³) were Baeyer-Villiger monooxygenases from bacteria and fungi, with the top match being the model BVMO cyclohexanone monooxygenase of the bacterium Rhodococcus sp. strain HI-31 (Fig. S7). BVMOs typically insert a molecule of oxygen between a carbonyl carbon and an adjacent carbon, without a leaving group, to create an ester (23). Activity of this type on Lps3-bound lysergyl-alanine would create the hypothetical ester illustrated in Fig. 3C without removing the hydrogen attached to carbon 2, the alpha carbon of the alanyl residue. In this scheme, the oxygen introduced onto carbon 2 would, upon hydrolysis or reduction, become the hydroxyl group of LAH (Fig. 3). Introduction of an oxygen onto carbon 2 by other mechanisms would eliminate the hydrogen attached to carbon 2 of the alanyl residue. To gather data distinguishing between these two possibilities, we fed l-alanine labeled with four molecules of deuterium—one on carbon 2 and three on the methyl carbon (carbon 3)—to M. brunneum ARSEF 9354. These experiments were conducted in infected G. mellonella larvae to maximize the yield of LAH. Overall, labeling of LAH with deuterated alanine was low (1.0 to 2.9% in individual fungus-infected larvae), presumably due to rapid metabolism of alanine to pyruvate in the insect hemolymph and/or in the fungus (24); nonetheless, the abundance of the ion at m/z 316 (which would result from a BVMO step in the synthesis of LAH and also from low-abundance natural isotopes in LAH) relative to unlabeled LAH (m/z 312) and relative to the ion measured at m/z 315 (which would result from loss of the deuterium at carbon 2 during LAH synthesis and also from low-abundance natural isotopes in LAH) was significantly greater in samples treated with deuterated alanine than in untreated samples (P = 0.0002) (Fig. 3), indicating retention of the deuterium on carbon 2 of the alanyl residue.

FIG 3.

Evidence for Bayer-Villiger monooxygenase activity in LAH biosynthesis. (A) Ratios of m/z intensities of ions obtained from extracts of M. brunneum ARSEF 9354 grown in unsupplemented G. mellonella larvae or larvae supplemented with l-alanine-2,3,3,3-D₄ (D4-ala). Error bars show the standard error of the mean; n = 7 larvae per treatment. M, molecular ion, which for LAH is m/z 312. (B) l-alanine-2,3,3,3-D₄-derived portion of Lps3-bound lysergyl-alanine. (C and D) Hypothetical oxygenated products of carbon 2 of the alanyl residue. (E) The side chain attached to the lysergyl group in LAH. Hydrolysis or reduction of the ester bond in C would yield LAH (E) while retaining all four deuteriums. Conversion of intermediate D to LAH requires a more complicated mechanism and would result in loss of one deuterium. Abbreviations: D, deuterium; R1, attachment to Lps3 via pantetheine; R2, attachment to carbonyl group of lysergic acid. The asterisk in panel C marks the carbon reduced by the reductase domain of Lps3.

Knockout of easO affects interaction of M. brunneum with a model insect.

Analysis of survival curves of G. mellonella larvae inoculated with 100 conidia of the wild-type parent strain, the easO knockout, and the complemented strain indicated that the wild-type fungus killed larvae at a significantly higher rate than the easO knockout (P < 0.0001) (Fig. 4). Complementation of the easO knockout restored the virulence of the mutant to levels comparable to that of the wild type (P = 0.69). The average time to 50% lethality (LT₅₀) was significantly greater in the easO knockout (164 h) than in the wild type or easO complemented strain (both 138 h) (P < 0.0001). Larvae infected with the easO mutant accumulated significantly fewer conidia compared to those inoculated with the wild type and complemented strain at 12 days postinoculation (Fig. 5). The observed decrease in sporulation may correspond to the documented delay in LT₅₀, because the number of conidia derived from easO knockout-killed larvae at days 13 and 14 was 25% and 75%, respectively, of the conidia obtained from the wild type at day 12.

FIG 4.

Kaplan-Meier survival curves for G. mellonella larvae inoculated with different strains of M. brunneum. Different letters next to curves indicate statistically significant differences in survival rates in log rank tests (Bonferroni corrected alpha = 0.008). Data are composited from three trials with 15 larvae per treatment per trial.

FIG 5.

Effect of easO mutation on conidiation. (A) Number of conidia per larva (n = 5) 12 days postinoculation with M. brunneum ARSEF 9354 (wt), M. brunneum easO knockout (ko), and the easO complemented strain (comp). Error bars represent the standard error. Treatments with the same letter do not differ significantly in a Tukey-Kramer honestly significant difference test (P < 0.05). (B) Photograph of larvae 12 days postinoculation.

DISCUSSION

Our data demonstrate that easO controls a required step in the synthesis of LAH downstream from the synthesis of lysergyl-alanine. Several lines of evidence summarized below indicate the product of easO acts as a Baeyer-Villiger monooxygenase to divert lysergyl-alanine from being used for ergonovine biosynthesis and reroute the lysergic acid amide pathway to predominantly accumulate LAH. Our data also indicate that LAH contributes to the virulence of M. brunneum in an insect model.

Analyses of the easO knockout mutant resulted in elimination of LAH and accumulation of the alternate lysergic acid amides lysergyl-alanine and ergonovine. In vitro substrate feeding studies did not support ergonovine or free lysergyl-alanine as substrates for EasO. Previous studies applying similar methodologies to feed substrates to cultures of other ergot alkaloid-producing fungi yielded successful conversion of substrates to products (25–28). Older labeling studies also did not support free lysergyl-alanine as a precursor to LAH (22). These data provide indirect evidence of Lps3-bound lysergyl-alanine as a substrate for EasO, but we do not have a mechanism for directly testing that hypothesis.

In the fully functional lysergic acid amide pathway, we hypothesize that once lysergyl-alanine is formed and bound to Lps3 as a thioester, it has three potential fates: (i) it can be reduced by the reductase domain of LpsC to yield ergonovine (10); (ii) it can be converted to LAH in a series of reactions initiated by EasO; and (iii) it can be hydrolyzed and released as free lysergyl-alanine (Fig. 1). In wild-type M. brunneum, LAH is far more abundant than ergonovine. In data presented by Leadmon et al. (18), LAH was over 200-fold more abundant than ergonovine in M. brunneum. In the present study, ergonovine was not detectable in larvae infected by wild-type M. brunneum, whereas LAH was the most abundant ergot alkaloid and accounted for approximately 97% of the total pool of lysergic acid derivatives measured (Table 1). Accordingly, EasO must be more abundant and/or more active than the reductase domain of Lps3. In our easO knockout mutant, Lps3-bound lysergyl-alanine has only two potential fates, reduction to ergonovine or liberation as free lysergyl-alanine. We observed both of these lysergic acid amides in high concentrations in G. mellonella larvae infected with our easO mutant (Table 1). Infected larvae analyzed at 8 days postinoculation had high concentrations of both lysergyl-alanine and ergonovine. The concentration of lysergyl-alanine declined severely by 12 days (with lysergic acid increasing, presumably as a catabolite), whereas the concentration of ergonovine relative to fungal biomass was maintained. The liberation of free lysergyl-alanine from Lps3 would require a thioesterase—presumably a type II thioesterase because the sequence of Lps3 contains no evidence of an intrinsic type I thioesterase domain. Perhaps expression or activity for that thioesterase is greater in younger than in older cultures, though we have no data directly supporting this explanation. The only potential candidate for a type II thioesterase in the eas cluster would be the alpha/beta fold hydrolase encoded by easP. Of course, candidate thioesterase genes also may be encoded outside the eas cluster. Fungi have housekeeping thioesterases to clean up stalled nonribosomal peptide synthetases (29). Moreover, in the scheme described above where EasO oxidizes Lps3-bound lysergyl-alanine to initiate synthesis of LAH, the reaction would leave a formate residue bound as a thioester on Lps3 that would need to be removed, presumably by the action of a thioesterase or the reductase domain of Lps3.

In silico analyses of the structure of the deduced product of easO indicated that it encodes a Baeyer-Villiger monooxygenase. BLASTp searches of the nonredundant database at NCBI returned many matches to hypothetical fungal monooxygenases without biochemically tested function; however, in a BLASTp search of the Protein Data Bank, the top 20 matches were Baeyer-Villiger monooxygenases from bacteria and fungi. Our labeling study with deuterated alanine indicated that LAH derived from labeled alanine retained the deuterium on carbon 2, consistent with the oxygen of LAH being added via a Baeyer-Villiger mechanism (23) (Fig. 3); that mechanism would be to insert an oxygen between the carbon 2 (the alpha carbon) and the carbonyl carbon of the alanine portion of lysergyl-alanine to create an ester that would serve as the immediate precursor to LAH (Fig. 3). That ester intermediate could theoretically yield LAH by either or both of two alternate mechanisms: (i) the carboxyl ester could be hydrolyzed by an esterase to yield LAH or (ii) LAH might be liberated as a result of the activity of the reductase domain of Lps3 on the carbonyl carbon of the ester (marked with an asterisk in Fig. 3). The reductase domain of Lps3 reduces that carbonyl carbon with a hydride ion derived from NADPH during the production of ergonovine from enzyme-bound lysergyl-alanine (10). Reduction of that very same carbon in the product of EasO would result in bond breakage on either the carboxyl ester side or the thioester side of the carbonyl carbon. Breakage on the carboxyl ester side would yield LAH directly; breakage on the thioester side would liberate a formyl ester of LAH, which might then be acted on by an esterase to yield LAH. One candidate for that esterase is the alpha/beta fold hydrolase encoded by easP, which also was mentioned above as a candidate for a type II thioesterase. The ability of the product of easP to fill both roles appears unlikely, which raises the possibility that other candidate genes reside outside the eas cluster. Functional analysis of easP and other potential esterases and thioesterases is a planned future line of study in our laboratory.

Survival studies indicated that LAH contributes to the virulence of M. brunneum on larvae of G. mellonella. This lepidopteran is a parasite of beehives; however, in that context it would not be a target for deliberate biocontrol by M. brunneum. It was used in our study as a model host. Although reduced in its virulence, the easO mutant was still pathogenic, supporting the roles of other virulence factors, presumably including those previously shown to be important in Metarhizium pathogenesis of insects (e.g., 30–32). We used a low concentration of conidia (100 conidia per larva) in our assays, resulting in a relatively high LT₅₀ value of almost 6 days (138 h) for the wild type. Elimination of easO resulted in an approximately 1-day (26-h) lag in the LT₅₀. The reduced accumulation of conidia also indicated diminished colonization of larvae by the easO knockout. HPLC data indicated that the total molar amount of ergot alkaloids in the mutant and the wild type were similar, so the differences in the interaction of the strains with their insect host must have been due to the changes in species of ergot alkaloids represented. Production of LAH instead of ergonovine requires additional enzymes and catalytic steps; the data from insect studies may represent a benefit for maintaining these extra steps as opposed to simply allowing the reductase domain of Lps3 to produce ergonovine from Lps3-bound lysergyl-alanine.

Our data show the necessity for easO in LAH biosynthesis and provide indirect evidence of its role. Identification of the gene encoding the esterase hypothesized to act on the product of EasO would support the proposed activity of EasO as a Baeyer-Villiger monooxygenase. The abundant lysergyl-alanine observed in the easO knockout is another observation that requires further investigation. Our data from larvae colonized for 8 days indicated the existence of a thioesterase that acted to liberate enzyme-bound lysergyl-alanine as frequently as the reductase domain of Lps3 acted to reduce thioesterified lysergyl-alanine to ergonovine. The data from our virulence assays in G. mellonella larvae indicate that the small difference in structure between LAH and ergonovine affects virulence of the fungus and suggest a rationale for the fungus maintaining additional enzyme activities to terminate its lysergic acid amide pathway with LAH as opposed to ergonovine.

MATERIALS AND METHODS

Genetic manipulation of fungi.

Metarhiziumbrunneum ARSEF 9354 was routinely cultured on sucrose yeast extract agar (SYE; per liter: 20 g sucrose, 10 g yeast extract, 1 g MgSO₄ · 7H₂O, 15 g agar). easO was transformed and mutated using a CRISPR-Cas9 strategy essentially as described by Davis et al. (19). To target easO for mutation, an oligonucleotide of sequence P-TTCTAATACGACTCACTATAGAGTAGTCAAGCACTTTAACTGTTTTAGAGCTAGA-OH (with the 20-nucleotide [nt] target sequence underlined) was synthesized at Eurofins Genomics (Louisville, KY) and converted to an sgRNA with the EnGen sgRNA synthesis kit (New England, BioLabs, Ipswich, MA). The sgRNA was designed to disrupt the coding sequences of easO. The coding sequences spanned coordinates 266649 to 268655 on strain ARSEF 3297 (GenBank accession number AZNG01000019.1), and the target sequence spanned coordinates 267148 to 267147, or nucleotides 500 to 519 within the 2,007-nt coding sequence including introns. Cas9-complexed sgRNA was cotransformed into protoplasts of M. brunneum ARSEF 9354 along with a phosphinothricin resistance-conferring fragment (bar; 19) as a selectable marker. Transformed colonies were screened preliminarily for changes in ergot alkaloid profile by analyzing approximately 400-μl samples of fungus-colonized SYE by HPLC with fluorescence detection as described below. Colonies with altered ergot alkaloid profiles were purified to nuclear homogeneity by culturing from single spores and screened for mutation at easO by PCR from primers P-GCCAAACCCTTCGTGGTCG-OH and P-GCAATCAATCGTCGTCCCATCAC-OH. PCR was performed with Phusion green master mix (Thermo Scientific, Waltham, MA) and the following program: initial denaturation at 98°C for 150 s, followed by 35 cycles of 98°C for 15 s, 66°C for 15 s, and 72°C for 240 s. PCR products were purified with the Zymo clean and concentrate kit (Zymogen, Irvine, CA) and Sanger sequenced at Eurofins Genomics.

The easO mutant was complemented with a fragment of DNA PCR amplified from wild-type genomic DNA with primers P-CTGACTTGAGCCACTTGCTTGC-OH and P-CGTGACATCGTGATACCCGTGTG-OH. PCR conditions were as described above except that the annealing temperature was 68°C and the extension time was 90 s. The product was 3,260 bp and contained the entire easO coding sequence along with 992 bp of 5′-untranslated sequences and 261 bp of 3′-untranslated sequences. The presence of an intact copy of the easO locus was detected in the complemented strain by PCR from the primers and PCR conditions listed immediately above.

Lps3 encoded by lpsC was mutated with two sgRNAs synthesized (as described above) from oligonucleotides P-TTCTAATACGACTCACTATAGGAGGAAGTTGAGGAACAAGGTTTTAGAGCTAGA-OH and P-TTCTAATACGACTCACTATAGAATCTCGCAGTTGCTAGCTGTTTTAGAGCTAGA-OH (with the 20-nt target sequence underlined) to delete the active site. The lpsC coding sequence (in reverse complement and containing one intron) spanned nucleotides 242413 to 237342 of strain ARSEF 3297 (GenBank accession number AZNG01000019.1), and the target sequences spanned coordinates 238025 to 238044 (nucleotides 4493 to 4512 within the 5,072-nt coding sequence including introns) and coordinates 237902 to 237921 (nucleotides 4370 to 4389 in the 5,072-nt coding sequence including introns). Transformed colonies were screened by PCR from primers P-CACTGGCTGAAGAGCTACGCG-OH and P-GCAGATGTTGTCACAGGCTCC-OH in a program as described above except that the annealing temperature was 67°C and the extension time was 60 s. PCR products were cleaned and Sanger sequenced as described above for the easO mutant.

Infection of Galleria mellonella larvae to measure survival and increase alkaloids.

Spores of M. brunneum strains were inoculated into larvae of G. mellonella essentially as described previously (18, 33) except that larvae were inoculated at a concentration of 5 spores/μl (20 μl volume/larva). The larvae were monitored every several hours until all inoculated larvae died (approximately 8 days). Three separate trials were conducted with 15 larvae per strain per trial. At 8 days and 12 days after inoculation, subsets of larvae were extracted by bead beating at 6 m/s for 30 s with 1 ml of methanol and 10 3-mm glass beads. An aliquot of the methanol extract was analyzed for ergot alkaloids by HPLC with fluorescence detection as described below, and another aliquot was analyzed for ergosterol concentration as a surrogate for fungal biomass. In order to quantify fungal sporulation, larvae were vortexed in 1 ml methanol, an aliquot was diluted in 0.1% Tween 20, and conidia were counted on a hemocytometer.

Analyses of ergot alkaloids by HPLC and LC-MS.

Methanol extracts from cultures or larvae were analyzed for ergot alkaloids by HPLC with fluorescence detection as described previously (18, 19, 34). The solid phase was a C₁₈ column (150 × 4.6 mm inside diameter [i.d.], 5-μm particle size, Prodigy C₁₈; Phenomenex, Torrance, CA), and the mobile phase was a binary, multilinear gradient of 5% acetonitrile in 50 mM aqueous ammonium acetate to 75% acetonitrile in 50 mM ammonium acetate. Ergot alkaloids were detected by fluorescence with excitation at 310 nm and emission at 410 nm to detect lysergic acid and its derivatives. Concentrations of ergot alkaloids were calculated by comparing peak areas of lysergic acid derivatives to an external standard curve prepared from ergonovine (Sigma, St. Louis, MO), which has the identical fluorophore to other lysergic derivatives. For lysergic acid derivatives that frequently occur in stereoisomers, areas of both isomers were combined. LC-MS analyses were performed with electrospray ionization in positive mode as described previously (35), but the mobile phase was a simple linear gradient from 5% acetonitrile in 0.1% formic acid to 75% acetonitrile in 0.1% formic acid over 20 min.

Quantification of fungal ergosterol in infected G. mellonella larvae.

Ergosterol, which was extracted from M. brunneum-infected larvae along with ergot alkaloids by bead-beating in methanol (as described above), was analyzed in an isocratic system by HPLC. The solid phase was a Phenomenex Prodigy C₁₈ column (150 × 4.6 mm i.d., 5-μm particle size) maintained at 30°C. The mobile phase was 97.5% methanol in water. Ergosterol was detected by absorbance at 280 nm and quantified relative to an external standard curve of ergosterol (Sigma). Under these conditions, ergosterol was eluted at approximately 12 min and samples were loaded every 20 min.

Precursor feeding assays.

Cultures containing 30,000 spores of M. brunneum lpsC knockout were grown in 200 μl of SYE and supplemented with ergonovine to 30 μM or lysergyl-alanine to 7.5 μM and incubated at 30°C for 8 days. Cultures were extracted with the addition of 200 μl of methanol and beating with five 3-mm glass beads at 6 m/s for 30 s. Extracts were clarified by centrifugation and analyzed by HPLC as described above.

Labeling studies with l-alanine-2,3,3,3-D₄ (deuterated alanine) were conducted by injecting larvae of G. mellonella with 20 μl of phosphate-buffered saline (PBS) containing 0.01% Tween 20 (33) containing a total of 75,000 conidia of M. brunneum ARSEF 9354. A high concentration of conidia was injected to try to promote a more rapid infection and accumulation of ergot alkaloids while deuterated alanine was present. At 1 and 3 days after inoculation, half the larvae were injected with 50 μl of 50 mg/ml deuterated alanine in PBS (D₄-ala treatment), and the other half received PBS without deuterated alanine (untreated larvae). Ergot alkaloids were extracted on day 7 after fungal inoculation, and extracts were analyzed by LC-MS as described above.

Statistical analyses.

For studies with deuterated alanine, relative intensities of ions of m/z 312, 315, and 316 were recorded from the scan occurring at the time of peak intensity for each of the two major stereoisomers of LAH (m/z 312) as well as from two scans immediately before peak intensity and two scans immediately after each of the two major LAH peaks. Ratios of scan intensities for m/z 316/312 and m/z 315/312 were calculated for each these 10 scans per sample, and an overall mean ratio for each individual sample was calculated. Means from seven larvae each for the deuterated alanine treatment and seven untreated controls were compared in an analysis of variance (ANOVA). Data were log transformed prior to ANOVA, so that variances between treatment and control groups approximated equality.

Data for more routine comparisons of ergot alkaloid accumulation or accumulations of conidia were checked for unequal variances with a Brown-Forsythe test. Data for which the Brown-Forsythe test yielded a P value of >0.05 were analyzed with ANOVA. When ANOVA indicated a difference among treatments (P < 0.05), means were separated using the Tukey-Kramer honestly significant difference test. When data did not pass the Brown-Forsythe test (P < 0.05; as was the case for samples for which certain ergot alkaloids were not detectable), means were compared nonparametrically with the Steel-Dwass test.

Larval survival after inoculation with different strains of M. brunneum was plotted on Kaplan-Meier survival curves, and differences between rates were compared by log rank tests with Bonferroni-corrected alpha set at 0.008 to compensate for multiple comparisons. All statistical analyses were performed in JMP (SAS, Cary NC).

Data availability.

The coding sequences (with introns) for easO of M. brunneum correspond to coordinates 266649 to 268655 of strain ARSEF 3297 (NCBI GenBank accession AZNG01000019.1); the sequence data originated from the work of Wang et al. (32). HPLC and insect survival data are available via Dryad at https://doi.org/10.5061/dryad.0rxwdbs0t.