Contribution of a novel gene to lysergic acid amide synthesis in Metarhizium brunneum

The fungus Metarhizium brunneum produces ergot alkaloids of the lysergic acid amide class, most abundantly lysergic acid α-hydroxyethylamide (LAH). Genes for making ergot alkaloids are clustered in the genomes of producers. Gene clusters of LAH-producing fungi contain an α/β hydrolase fold protein-encoding gene named easP whose presence correlates with LAH production but whose contribution to LAH synthesis in unknown. We tested whether EasP contributes to LAH accumulation through gene knockout studies. We knocked out easP in M. brunneum via a CRISPR/Cas9-based approach, and accumulation of LAH was reduced to less than half the amount observed in the wild type. Because LAH accumulation was reduced and not eliminated, we identified and mutated the only close homolog of easP in the M. brunneum genome, a gene we named estA. An easP/estA double mutant did not differ from the easP mutant in lysergic acid amide accumulation, indicating estA had no role in the pathway. We conclude EasP contributes to LAH accumulation but is not absolutely required. Either a gene encoding redundant function and lacking sequence identity with easP resides outside the ergot alkaloid synthesis gene cluster, or EasP plays an accessory role in the synthesis of LAH.


Introduction
Ergot alkaloids derived from lysergic acid have impacted humankind for millennia as toxins in contaminated grain crops and as pharmaceuticals to treat dementia, migraines, hyperprolactinemia, and other disorders [1][2][3]. Many pharmaceutical ergot alkaloids are derivatives of lysergic acid amides. Naturally occurring lysergic acid amides include ergonovine (called ergometrine in Europe) and lysergic acid α-hydroxyethylamide (LAH). Ergonovine is the primary lysergic acid amide in the rye ergot fungus Claviceps purpurea, which also produces abundant ergopeptines [2,4]; whereas, the plant root symbiont and insect pathogen Metarhizium brunneum produces ergonovine and LAH, with the concentration of LAH dwarfing that of ergonovine by approximately 200 to 1 [5,6]. In studies to date, all genes required for ergot alkaloid synthesis are clustered in the genomes of the producing fungi in ergot alkaloid synthesis (eas) clusters [7][8][9][10][11][12].
Both ergonovine and LAH are derived from the intermediate lysergyl-alanine [4,6] which is synthesized via a complex of two monomodular peptide synthetases, lysergyl peptide synthetase (Lps) 2 and Lps3. Lps2 recognizes lysergic acid, activates it by adenylation, binds it as a thioester, and condenses it with alanine that has been similarly recognized, activated, and thioesterified by Lps3. Lps3 contains a carboxy terminal reductase domain that can reduce thioesterified lysergyl-alanine to ergonovine [4]. The reductase domain uses hydride ions obtained from its cofactor NADPH to reduce the carbonyl carbon of the alanyl moiety of lysergyl-alanine to an aldehyde and then a primary alcohol, liberating ergonovine from enzyme bound lysergyl-alanine [4] (Fig. 1). In fungi that produce LAH as well as ergonovine (e.g., M. brunneum, the paspalum ergot fungus Claviceps paspali, and the morning glory symbiont Periglandula ipomoeae) there are two additional genes in the eas cluster: easO, encoding a Baeyer-Villiger monooxygenase (BVMO) required for synthesis of LAH [6], and easP, encoding an α/β hydrolase fold protein, a role for which in ergot alkaloid synthesis has not yet been demonstrated but whose presence correlates perfectly with the ability to produce LAH [7,8,10]. As a BVMO, EasO is hypothesized to insert an oxygen between the alpha carbon and carbonyl carbon of the alanyl portion of lysergyl-alanine [6]. Gene knockout and stable-isotope labeling data support this hypothesis. The hypothesized intermediate produced by EasO provides a substrate that could be converted to LAH through the activity of a carboxyl esterase (potentially an α/β hydrolase fold protein encoded by easP) or through the activity of the reductase domain of Lps3 ( Fig. 1) [6]. In the present study, we mutated easP through a CRISPR/Cas9-based approach to test its contribution to synthesis of lysergic acid amides.
The easP locus of M. brunneum was knocked out via a transient CRISPR/Cas9-mediated approach based on the protocol described by Davis et al. [14]. An sgRNA was synthesized from the template 5′-TTC TAA TAC GAC TCA CTA TAG TCT GCT CCA TGG AGG CTC CTGTT TTA GAG CTA GA-3′ (the 20-nt target sequence is underlined and an additional G was inserted immediately preceding the target sequence) with the EnGen sgRNA synthesis kit (New England Biolabs, Ipswich, MA, USA). The sgRNA was complexed with EnGen Spy Cas9 NLS (New England Biolabs) and co-transformed into protoplasts along with a phosphinothricin-resistance conferring fragment [14]. Transformants were screened for mutations at easP in PCRs primed with oligonucleotides PcrspF (5′-CAC ACT CTA CTC CCT CAC AAGG-3′) and PcrspR (5′-CCG CTC CAG GCA TCG TCA C-3′). Reaction conditions included an initial denaturation at 98 °C for 30 s, followed by 35 cycles of 98 °C for 15 s, 66 °C for 15 s, and 72 °C for 80 s. PCR products that differed in length from that of wild-type M. brunneum were Sanger sequenced at Eurofins Genomics (Louisville, KY, USA).
The estA gene was knocked out via a similar approach, except the template for the sgRNA was 5′-TTC TAA TAC GAC TCA CTA TAG TCC AAG ACG TAG CCG ACT TCGTT TTA GAG CTA GA-3′ (where the target sequence is underlined and preceded by an additional G), and the selectable marker was the hygromycin resistance-conferring plasmid pBC-hygro [15]. Transformants were screened by PCR with oligonucleotides estAF (5′-GAA CAC AAC TTC ACC ACA TCGC-3′) and estAR (5′-GAC GTC ACC GAG CTC TCT G-3′) in reactions with conditions as described above, except the annealing temperature was 64 °C and the extension time 180 s.
Ergot alkaloids and ergosterol were extracted from infected larvae of G. mellonella by bead beating in methanol [5,6]. To detect ergot alkaloids, 20 µL of extract was analyzed by high performance liquid chromatography (HPLC) with fluorescence detection (excitation 310 nm/emission 410 nm) [5,6,16]. Ergot alkaloid quantities in each sample were normalized to the quantity of the fungus-unique sterol ergosterol in that same extract as described by Steen et al. [6]. Data were compared by Brown-Forsythe tests to assess equality of variances. Data that passed a Brown-Forsythe test (P > 0.05) were analyzed by single-factor ANOVA. For data sets involving multiple comparisons, a Tukey-Kramer Honestly Significant Difference test was employed. In three comparisons (ergonovine and lysergyl-alanine in the easP knockout compared to wild type, and ergine in the easP/estA double knockout), data did not pass a Brown-Forsythe test (P < 0.05) and subsequently were compared nonparametrically, initially with a Wilcoxon rank sums test and then (for the ergine data) a Steel-Dwass nonparametric multiple comparison test. Statistical analyses were conducted in JMP Pro 14 (SAS, Cary, NC, USA).

Results
Introduction of Cas9 complexed with an sgRNA targeting easP, which was co-transformed with a selectable marker for phosphinothricin resistance, resulted in knockout of easP. As a haploid fungus, M. brunneum contains only a single allele of easP. Sequence analysis of the easP locus in two independent transformants indicated Cas9 cut three bp before the PAM site and a portion of the cotransformed phosphinothricin-resistance marker was ligated into the easP locus during repair (Additional file 1: Figure S1). Knockout of easP had a significant effect on accumulation of LAH relative to fungal biomass (estimated by measuring the fungus-unique sterol ergosterol) in M. brunneum-infected larvae of the model insect G. mellonella (Fig. 2). The easP knockout resulted in a significant decrease (approximately 70%) in accumulation of LAH (P < 0.0001). The concentrations of alternate lysergic acid amides, ergonovine and lysergyl-alanine were not affected in the easP knockout. Ergine, the simple amide of lysergic acid, accumulated to a higher concentration in the easP knockout than in the wild type (P = 0.003). Ergine can arise as a spontaneous hydrolysis product of LAH [17,18] (Fig. 1) and also from hydrolysis of other lysergic acid derivatives [19]. A similar biochemical phenotype was observed qualitatively in the second easP knockout. Whereas the primary purpose of this study was to assess the contribution of easP to LAH biosynthesis, the easP knockout also was investigated for changes in radial growth and sporulation in vitro and colonization of G. mellonella larvae (measured by ergosterol accumulation). No significant differences were detected in these traits (Additional file 2: Table S1).
Because the easP knockout mutation led to reduction as opposed to elimination of LAH, we investigated the possibility that a similar enzyme with redundant function was encoded in the M. brunneum genome. Since the eas cluster contained no additional genes with unidentified function, we queried the haploid genome of M. brunneum ARSEF 3297 for genes with the capacity to encode proteins similar to EasP by tblastn search. The M. brunneum genome encoded only one additional significant match (E value 2e −05 ; no other match had an E value less than 1). The homolog's translation product corresponds to the α/β hydrolase fold protein under GenBank accession XP_014542068. We named this gene estA, for esterase A, and knocked it out by a CRISPR/Cas9 approach in the easP knockout strain while selecting for a co-transformed hygromycin resistance marker (Additional file 3: Figure S2). Knockout of estA did not affect ergot alkaloid accumulation relative to the easP knockout (Fig. 3), indicating that EstA plays no role in the ergot alkaloid pathway. In this data set, LAH was reduced by a mean of 54% in the easP and easP/estA knockouts relative to wild type. Concentrations of ergonovine and lysergyl-alanine did not differ significantly among strains. The concentration of ergine was again higher in the easP knockout (and Fig. 2 Accumulation of LAH and other lysergic acid amides in wild-type M. brunneum versus an easP knockout strain. Extracts were prepared from larvae of the model insect Galleria mellonella seven days after infection with the indicated strains. Moles of ergot alkaloids were measured relative to a standard curve of ergonovine and normalized to mass of the fungus-unique sterol ergosterol as a surrogate for fungal biomass. Bars indicate means (n = 6 fungus-infected larvae), and error bars indicate standard error. Different letters within an alkaloid comparison indicate that fungal strain had a significant effect on accumulation of the indicated lysergic acid amide (P < 0.05) in the easP/estA double knockout) than in the wild type, with P values of 0.035 and 0.014, respectively, for the two easP mutant strains compared to the wild type in a Steel-Dwass multiple comparison test.

Discussion
Our data indicate that EasP contributes significantly to LAH accumulation but is not absolutely required. One hypothesis to account for this observation is that a similar enzyme encoded in the M. brunneum genome has a redundant function. Functional analysis of the only homolog of easP in the M. brunneum genome did not support this hypothesis. If another enzyme is performing a function redundant with that of EasP, the enzyme must be unrelated in terms of amino acid sequence and also encoded outside the eas cluster. To date all genes required for ergot alkaloid synthesis have been encoded in the eas cluster [2,3,7,8]. In a recent study relevant to this point, Jones et al. [10] reported that three species of Aspergillus synthesize LAH and evolved the steps for incorporation of lysergic acid into LAH independently of M. brunneum (and other fungi in the Clavicipitaceae) yet had similarly constituted eas clusters with no additional genes present that might be redundant with easP. The possibility that two lineages evolved similar pathways for incorporating lysergic acid into LAH yet kept one unrelated and redundant gene outside the eas cluster appears unlikely.
An alternate hypothesis to account for the remaining LAH in the easP knockout backgrounds is that EasP plays an accessory role in generating LAH and that LAH can be synthesized in the absence of EasP or any carboxyl esterase. Steen et al. [6] suggested the reductase domain of Lps3 might liberate LAH from a carboxyl ester/thioester diester intermediate arising from the activity of the Baeyer-Villiger monooxygenase EasO on Lps3-bound lysergyl-alanine (Fig. 1). Liberation of LAH via reduction of the carbonyl carbon of this intermediate could produce LAH without the requirement for esterase activity. The reductase domain of Lps3 reduces that same carbonyl carbon during the production of ergonovine in C. purpurea [4]. Although not required for LAH synthesis in this scenario, EasP might still contribute to LAH accumulation via one or more of several possible mechanisms. One mechanism would be for EasP to hydrolyze LAH from Lps3-bound carboxyl ester/thioester derivative of lysergyl-alanine (Fig. 1) prior to reduction by the Lps3 reductase domain. A second mechanism by which EasP could increase LAH accumulation via this same model would be to salvage LAH from an LAH formate ester that would result if the activity of the reductase domain of Lps3, acting on that same carbonyl carbon, resulted in bond breakage on the thioester side of the diester intermediate. The resulting formate ester might be hydrolyzed to LAH by EasP acting as a carboxyl esterase. No novel fluorescent peaks stood out in the easP knockout extracts, but increased ergine noted in easP mutants may have arisen at least in part from hydrolysis of lysergic acid amides other than LAH which, as a typical source of ergine, was present in lower concentration in the easP knockout backgrounds. Testing these hypothetical roles for EasP in conjunction with Lps3 will require mutation of the reductase domain of Lps3 while retaining the earlier functions of the enzyme (adenylation and thiolation of alanine). Such experiments are the objectives of future work in our laboratory.

Limitations
We showed EasP contributes significantly to LAH accumulation, but we have not demonstrated the mechanism by which it acts or identified the origin of the remaining LAH in the easP knockout. We lack the genetic resources to test whether the remaining LAH in an EasP mutant results from activity of the reductase domain of Lps3. Investigation of whether a gene outside the eas cluster is required for synthesis of the remaining LAH will require novel approaches. Our study focused on the contribution of easP to LAH biosynthesis; we did not investigate its potential contribution to other pathways.
Abbreviations ANOVA: Analysis of variance; ARSEF: Agricultural Research Service collection of entomopathogenic fungi; BVMO: Baeyer-Villiger monooxygenase; Cas9: Clustered regularly interspaced palindromic repeats-associated protein 9; CRISPR: Clustered regularly interspaced palindromic repeats; Eas: Ergot alkaloid Fig. 3 Effect of estA/easP double mutation relative to easP mutation on accumulation of LAH and other lysergic acid amides in M. brunneum. Alkaloids are expressed as moles of ergot alkaloids per mass of the fungus-unique sterol ergosterol as a surrogate for fungal biomass. Data represent means (n = 6 fungus-infected G. mellonella larvae), and error bars represent standard error. Extracts were prepared seven days after infection. Strains, within an alkaloid type, marked with different letters differed significantly (P < 0.05)