The lolines are a class of bioprotective alkaloids that are produced by Epichloë species, fungal endophytes of grasses. These alkaloids are saturated 1-aminopyrrolizidines with a C2 to C7 ether bridge, and are structurally differentiated by the various modifications of the 1-amino group: -NH2 (norloline), -NHCH3 (loline), -N(CH3)2 (N-methylloline), -N(CH3)Ac (N-acetylloline), -NHAc (N-acetylnorloline), and -N(CH3)CHO (N-formylloline). Other than the LolP cytochrome P450, which is required for conversion of N-methylloline to N-formylloline, the enzymatic steps for loline diversification have not yet been established. Through isotopic labeling, we determined that N-acetylnorloline is the first fully cyclized loline alkaloid, implying that deacetylation, methylation, and acetylation steps are all involved in loline alkaloid diversification. Two genes of the loline alkaloid biosynthesis (LOL) gene cluster, lolN and lolM, were predicted to encode an N-acetamidase (deacetylase) and a methyltransferase, respectively. A knockout strain lacking both lolN and lolM stopped the biosynthesis at N-acetylnorloline, and complementation with the two wild-type genes restored production of N-formylloline and N-acetylloline. These results indicated that lolN and lolM are required in the steps from N-acetylnorloline to other lolines. The function of LolM as an N-methyltransferase was confirmed by its heterologous expression in yeast resulting in conversion of norloline to loline, and of loline to N-methylloline. One of the more abundant lolines, N-acetylloline, was observed in some but not all plants with symbiotic Epichloë siegelii, and when provided with exogenous loline, asymbiotic meadow fescue (Lolium pratense) plants produced N-acetylloline, suggesting that a plant acetyltransferase catalyzes N-acetylloline formation. We conclude that although most loline alkaloid biosynthesis reactions are catalyzed by fungal enzymes, both fungal and plant enzymes are responsible for the chemical diversification steps in symbio.
Citation: Pan J, Bhardwaj M, Nagabhyru P, Grossman RB, Schardl CL (2014) Enzymes from Fungal and Plant Origin Required for Chemical Diversification of Insecticidal Loline Alkaloids in Grass-Epichloë Symbiota. PLoS ONE 9(12): e115590. https://doi.org/10.1371/journal.pone.0115590
Editor: Richard A. Wilson, University of Nebraska-Lincoln, United States of America
Received: September 14, 2014; Accepted: November 29, 2014; Published: December 22, 2014
Copyright: © 2014 Pan et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: The authors confirm that all data underlying the findings are fully available without restriction. All relevant data are within the paper and its Supporting Information files.
Funding: This research was supported by Kentucky's National Science Foundation Experimental Program to Stimulate Competitive Research grant 0814194 (http://www.kynsfepscor.org/), and National Institute of General Medical Sciences grant R01GM086888, and United States Department of Agriculture grant 2012-6701319384. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Many cool-season grasses (Poaceae, subfamily Poöideae) are infected by endophytic fungi of genus Epichloë. The endophyte systemically colonizes aboveground tissues of the grass and grows along the longitudinal axis between plant cells. Many Epichloë species rely completely on the host for dissemination through the seeds, whereas some are additionally or solely transmitted horizontally to new plants via sexual spores. Although the horizontally transmitted species are somewhat pathogenic — causing choke disease in forming their sexual reproduction structures (stromata) — all Epichloë species spend most or all of their life cycle in plant intercellular spaces and cause no obvious harm to their hosts. This close interaction of the two organisms tends to foster mutualistic interaction in which the endophyte helps to defend its host grass against invertebrate or vertebrate herbivory through production of various kinds of alkaloids. Four classes of alkaloids are produced by these endophytic fungi — ergot alkaloids, lolitrems, peramine, and loline alkaloids — all of which confer herbivory resistance . The loline alkaloids specifically deter invertebrate herbivores , . Lolines found in grass-Epichloë symbiota (Fig. 1) differ in modifications of the 1-amino group: -NH2 (norloline), -NHCH3 (loline), -N(CH3)2 (N-methylloline = NML), -N(CH3)Ac (N-acetylloline = NAL), -NHAc (N-acetylnorloline = NANL), and -N(CH3)CHO (N-formylloline = NFL).
Substitutions on the nitrogen at C1 differentiate the lolines.
The asexual seed-transmitted Epichloë species are often found to be inter- or intra-species hybrids –, which (among other benefits) can pyramid multiple alkaloid biosynthesis gene clusters and promote diversification of alkaloid profiles , . A likely benefit of diversifying defensive alkaloids is to reduce the potential for herbivores to develop resistance, provided that different alkaloids have different modes of specificity . Another potential advantage is that different alkaloids may affect different herbivores and hence broaden the spectrum of host protection. The ability of grass-Epichloë symbiota to produce a variety of protective alkaloids greatly enhances the plant-defensive arsenal . Even among loline alkaloids, differently modified forms have been reported to exert different effects on insects . For example, at high concentrations (800 and 1600 µg/g), NFL reduces the growth and development of Argentine stem weevil (Listronotus bonariensis) larvae, whereas NANL causes high mortality of Argentine stem weevil larvae but with little effect on the growth or development of the larvae . These interesting observations raise the question of how the diverse loline alkaloids are synthesized and how profiles of different loline alkaloids have evolved.
Several genomes of loline alkaloid-producing Epichloë species have been sequenced, and a LOL cluster that encodes enzymes involved in loline alkaloid biosynthesis has been found in all of them. The endophytes that produce NFL together with the other loline alkaloids have a total of 11 LOL genes positioned in the cluster in the order lolF, lolC, lolD, lolO, lolA, lolU, lolP, lolT, lolE, lolN and lolM . Others that produce NANL or the bicyclic exo-1-acetamidopyrrolizidine (AcAP) as the end product have subsets of the 11 aforementioned genes, either lacking genes or having inactivating mutations in late pathway genes . Functions of enzymes encoded by LOL genes have been suggested based on bioinformatic analysis . LolC has been proposed to catalyze an unusual γ-substitution reaction that condenses proline and homoserine, and it has been established to be involved in the pathway by an RNA-interference (RNAi) experiment . A predicted cytochrome P450 enzyme, LolP, is required for oxygenation of NML to produce NFL and is not involved in earlier biosynthetic steps . Recently, it has been established that ether-bridge formation requires the 2-oxoglutarate-dependent non-heme iron oxygenase, LolO . The other loline alkaloid diversification steps have not yet been identified. Here we establish the basis for the diversity of loline alkaloids, demonstrating that LolN and LolM are required for biosynthetic steps from NANL to NML, and also showing that the host plant converts loline to NAL.
Materials and Methods
Biological materials and general experimental procedures
Fungal strains (Table 1) were isolated and cultured on potato dextrose agar (PDA) at 22°C as previously described . Infection of the grass with or without the endophyte was checked by tissue-print immunoblot , then grow-out of the fungus from surface-sterilized grass tissue, and confirmation of the endophyte strain using polymerase chain reaction (PCR) with specific primers. General experimental procedures and reagents were as previously reported . Primers used in this study are listed in S1 Table.
Loline alkaloids were extracted with chloroform, using quinoline as internal standard, and analyzed by GC-MS as previously reported . Extraction from plant and minimal-medium cultures followed previously reported methods .
Preparation of tetradeuterated AcAP
This procedure was conducted as previously reported for preparation of AcAP , except that we used (±)-3,3-dideutero-1-oximinopyrrolizidine  as the starting material to label the pyrrolizidine ring on C3, and we used [2H6]-acetic anhydride to add a trideuteroacetyl group to the exo-1-aminopyrrolizidine. The resulting products were 13.5% d3-, 85% d4-, and only trace amounts of d5-exo-1-acetamidopyrrolizidine. Substitution of deuterium with protium from H2O must have occurred during the Raney Ni reduction, rendering exo-1-trideuteroacetamido-3-deuteropyrrolizidine ([2H4]AcAP) as the major product. (In fact, washing the Raney Ni with D2O before adding the oxime yielded products with up to seven D atoms incorporated.) Fortunately, this less highly deuterated material remained suitable for the experiments for which it was intended.
Application of [2H4]AcAP to Epichloë uncinata e167 culture
Epichloë uncinata e167 was grown in minimal medium to induce loline-alkaloid production as described previously . The labeled compound was applied at the onset of loline alkaloid production (9th day post inoculation) at a final concentration of 3.75 mM. L-[U-2H7]Pro (4 mM) or Pro (4 mM) was applied following the same procedure as control. To each of the treatments DMSO (1.5%) was also added to increase permeability of the cells. The cultures were shaken at 55 rpm, 22°C, for another 35 days and then checked for loline alkaloids.
Construction of gene-replacement and complementation plasmids
In Epichloë festucae E2368, lolN and lolM are adjacent and divergently transcribed. A plasmid was constructed to replace lolM and most of lolN in E. festucae E2368 with a modified hph marker gene  via homologous recombination. Using E2368 genomic DNA as the template, a 2.5-kb fragment upstream of the target lolN-lolM segment was amplified by PCR with primers upFLKf (XbaI) and upFLKr (XbaI), digested with XbaI, and ligated with pKAES173  digested with XbaI. The ligated product was introduced by transformation into Escherichia coli XL1-blue competent cells, and transformants were screened for hph by colony-PCR with primers hphf and hphr. Positive colonies were extracted for plasmid DNA, which was then tested by restriction-endonuclease digestions and Sanger sequencing to identify the construct that had the correct orientation. The proper construct was then digested with XmaI and XhoI. A 1.9-kb fragment downstream of the lolN-lolM segment was amplified by PCR with primers lolAANMTdf (XmaI) and lolAANMTdr (XhoI), and then digested with XmaI and XhoI. The digested plasmid and PCR product were gel purified and ligated to generate plasmid pKAES323, which has flanking sequences of the target lolN-lolM region on either side of the hph expression construct.
To complement the lolN and lolM double-knockout strain, lolN and lolM of E2368, together with their common, divergent promoter, were amplified by PCR with primers upFLKf and lolNMkops using Phusion Hot Start High-Fidelity DNA Polymerase (Thermo Scientific, Ratastie, Vantaa, Finland) with HF buffer (with 1.5 mM MgCl2) provided from the manufacturer. The temperature conditions were 98°C for 3 min, 35 cycles of 98°C for 10 s, 62°C for 10 s, and 72 for °C 3 min, then a final 5 min incubation at 72°C. The PCR product was purified and digested with XbaI. Plasmid pKAES215 (S1 Protocol) was digested with XbaI and EcoRV. The two fragments were gel purified and ligated to generate pKAES341, which confers hygromycin resistance and expresses lolN and lolM driven by their native, divergent promoter. All plasmid constructs were confirmed by Sanger sequencing.
Fungal transformation and screening for knockouts
Protoplast preparation and transformation procedures were performed as previously described  with minor changes . After electroporation, lolN and lolM knockout transformants of E2368 were selected on regeneration medium with 450 µg/ml hygromycin B. The same selection was also used in single spore isolation and fungal isolate maintenance on PDA. After two single-spore isolations, the transformants were extracted for DNA with DNeasy 96 Plant Kit (Qiagen, Valencia, CA, USA) and screened for the lolN gene by PCR with primers aamAup1 and aamAd1. The transformants that were negative in the lolN screening were then positively screened with primers upFLKf and lolNMkops, which produced a 4.0 kb product from the knockout transformant and a 5.7 kb product from the wild-type strain. As controls, an ectopic transformant with pKAES323 and an empty-vector transformant with plasmid pKAES173 were also generated and confirmed by PCR.
Transformation of the knockout strain with pKAES341 for complementation of lolN and lolM was performed by co-transforming linearized pKAES341 and pII99, which confers geneticin resistance , at a molar ratio of 5∶1. Selection of transformants and single-spore isolations were carried out on PDA plates with 800 µg/ml G418. The transformants were single-spore isolated twice and extracted for DNA to screen for lolN by PCR with primers aamAup1 and aamAd1, and for lolM with primers aamAup1 and lolNMkops.
Inoculation of the strains to endophyte-free meadow fescue was carried out as described by Latch and Christensen . Loline-alkaloid profiles of the plants were checked approximately 6 months after inoculation.
Genomic DNA was digested with HindIII and 1 µg of each DNA was electrophoresed in 0.7% agarose gel (Phenix Research, Candler, NC, USA) in 0.5 x TBE buffer [44.5 mM tris(hydroxymethyl)aminomethane (Tris) base, 44.5 mM boric acid, 1 mM ethylenediaminetetraacetic acid (EDTA), pH 8.2] for 20 hr in a 4°C cold room. The DNA was then transferred to Hybond H+-nylon membrane (GE Healthcare, Piscataway, NJ, USA) with a GENIE electroblotter (Idea Scientific, Minneapolis, MN, USA). Immobilized DNA on the membrane was denatured in 0.4 M NaOH for 10 min and neutralized in 2x SSC for 10 min. The DNA was cross-linked to the membrane with UV light using a Spectrolinker (Spectronics, Westbury, NY, USA). Probes for lolN and lolM were prepared by PCR with primers Nexpf and Nexpr, and Mexpf and Mexpr, respectively, using E2368 DNA as template. A 974-bp PCR product for hph probe was prepared by PCR with primers hphf and hphr using pKAES173 as template. The PCR products were purified and labeled with α-[32P]dCTP (GE Healthcare) using the Prime-a-Gene Labeling System (Promega, Madison, WI, USA). Hybridization was performed as in Starnes et al. . Hybridizations of probes for lolN and lolM were performed sequentially on the same membrane, whereby the probe from each previous hybridization was removed by incubating the membrane twice with 0.4 M NaOH for 30 min at 45°C, followed by washing with 0.1x SSC 0.1% SDS for 15 min at 45°C. Subsequent hybridization to an hph probe was used to verify the quality and approximately equal quantities of the DNA in each lane.
Yeast expression of lolM and western blot analysis
The full-length lolM was amplified from cDNA prepared from mRNA of E. uncinata e167 that was induced to produce lolines by ca. 20 day culture in minimal medium . Primer pair MexpYNfBglII and MexpYNrBglII was used to amplify lolM with BglII incorporated in the 5′ ends so that lolM would be expressed with an N-terminal FLAG tag . The PCR products were purified and digested with BglII, and ligated with BglII-digested pESC-LEU (Agilent Technologies, Santa Clara, CA, USA). Resulting plasmids with the correct orientation were chosen through endonuclease digestion and further confirmed by Sanger sequencing. The resulting plasmid pKAES350 has lolM under control of the promoter GAL10, with an N-terminal FLAG tag. Yeast (Saccharomyces cerevisiae) strain YPH499 (MATa ura3-52 lys2-801amber ade2-101ochre trp1-Δ63 his3-Δ200 leu2-Δ1)  (Agilent Technologies) was used to express lolM. Yeast competent cell preparation and transformation were performed according to the manufacture's protocol (Agilent Technologies), and transformants were selected on synthetic dextrose minimal medium without leucine (SD/-Leu) agar plate. The growing colonies were then streaked twice on the same medium before feeding experiments.
Expression of the LolM protein was confirmed by western blot analysis as follows. Yeast crude protein was extracted following the methods described in Rajendran et al. ; then the proteins were fractioned by electrophoresis in a SDS-PAGE gel (10%). The primary antibody was α-FLAG (Sigma-Aldrich, St. Louis, MO), and immunoblots were developed using an ECL detection kit (Roche Applied Science, Basel, Switzerland).
Application of norloline or loline to yeast crude protein extract
Yeast transformed with pKAES350 or empty vector pESC-LEU were grown in SD/-Leu broth at 30°C overnight, washed twice with synthetic galactose minimal medium without leucine (SG/-Leu), and resuspended in SG/-Leu medium to OD600 of 0.5. The cultures were grown at 30°C, 250 rpm for 2 days before sampling. The yeast cells were centrifuged at 1000×g for 5 min, and the cell pellet was weighed. Approximately 300 mg of yeast pellet was used for each crude protein extraction following the procedure described in  with modifications. A Fastprep machine (20 sec at 4 m/s, repeated four times) was used to break the cells in a Tris-HCl buffer (50 mM Tris-HCl, pH 7.5, 10% glycerol, 10 mM dithiothreitol), with proteinase (cOmplete ULTRA Tablets, Mini, EDTA-free, Roche) added right before extraction. Either loline or norloline (4 mM), and S-adenosyl methionine (AdoMet) (1 mM), were added to 90 µl of the crude extract for the enzyme assay. Crude protein extract with loline or norloline but without adding AdoMet was also assayed as a check that the reaction requires AdoMet. The assay mixture was incubated at 30°C overnight, and the reaction was stopped by adding 100 µl of 1M NaOH. Then the loline alkaloids were extracted with chloroform and analyzed by GC-MS.
Loline used in this study was prepared from endophyte infected tall fescue seeds according to the previously reported procedure (Petroski et al., 1989). Norloline was prepared from Adenocarpus decorticans seeds (100 g). The ground seeds were basified by adding 1 M NaOH to pH 11, then extracted with chloroform (5 x, 10 ml/g seed). Then the alkaloid mixture was concentrated by rotary evaporation, during which the loline alkaloids were hydrolyzed to norloline due to concentration of the base. Then the mixture was applied on silica gel column (chloroform: methanol: NH4OH = 49.5: 49.5: 1 as the running solvent) to purify approximately 10 mg of norloline.
Genome sequencing and phylogenetic analysis
Fungal DNA was isolated and sequenced by pyrosequencing on a Roche/454 GS FLX+ platform, and assembled as previously described , . Genes were annotated as previously described  from sequenced genomes of Epichloë species and Atkinsonella hypoxylon, available at www.endophyte.uky.edu and the National Center for Biotechnology Information. The lolN sequences of Epichloë sp. e4686 and e4779 were obtained by PCR with primers Nexpf and lolNampr and sequencing of the PCR products. The sequences were deposited in GenBank with accession numbers KJ653809 for e4686 lolN and KJ653810 for e4779 lolN. Alignment of DNA sequences and amino acid sequences was performed on www.phylogeny.fr , and edited with MacVector v. 12.7.5.
Results and Discussion
Retention of the deuterium atoms from [2H4]AcAP in loline alkaloid product
A previous study showed that LolO is required in formation of the ether bridge of loline alkaloids , but it did not determine whether AcAP is the direct precursor of NANL. An alternative hypothesis is that exo-1-aminopyrrolizidine (1-AP) is the real precursor to be oxidized to form norloline, whereas AcAP and NANL are both shunt products generated by acetylation of 1-AP and norloline, respectively. To test these alternative possibilities, we prepared [2H4]AcAP with one deuterium atom at C3 of the pyrrolizidine ring system and three deuterium atoms on the acetyl group, and we applied it to a culture of E. uncinata e167 under loline alkaloid-production conditions . If the acetylation of 1-AP occurred before ether-bridge formation, we expected NANL to retain four deuterium atoms, whereas if AcAP were first deacetylated to 1-AP before ether-bridge formation, we expected it to retain only one deuterium. In the event, we observed a compound that was determined by MS to be tetradeuterated NANL at the front edge of the NANL peak, with parent ion m/z = 187 (m/z = 183+4), and major fragment ions with m/z = 83 (m/z = 82+1) and m/z = 157 (m/z = 153+4) (Fig. 2). This result clearly demonstrated the linear nature of the biosynthetic pathway and established AcAP and NANL as true intermediates. Hence, the likely pathway for the diversification of loline alkaloids from NANL would involve deacetylation, methylation, and acetylation.
Deletion of the lolN and lolM genes resulted in accumulation of NANL and eliminated NFL production
The lolN and lolM genes are found in all NFL-producing Epichloë species, and in the genome assembly of Epichloë festucae E2368 they are linked with the known LOL cluster . Bioinformatic analyses of these genes suggested that LolN is an acetamidase and LolM is a methyltransferase. The functions fit well into our hypothesis that NANL must be deacetylated to produce norloline, which is then methylated to form loline and NML, the precursor of NFL . Hence, it appeared likely that LolN and LolM are involved in NFL biosynthesis from NANL. In E. festucae E2368, lolN and lolM share a common, divergent promoter, so in order to investigate functions of lolN and lolM, we constructed a plasmid to replace both genes with marker gene hph through homologous recombination (Fig. 3A).
(A) Schematic representation of lolN-lolM replacement by the hph marker gene via homologous recombination. Shown are maps of the wild-type lolN and lolM in Epichloë festucae E2368 (WT), targeting vector (pKAES323), and the locus after homologous recombination (KO). Black bars represent DNA sequence, and filled arrows represent genes. Bent blue lines on the bars represent HindIII digestion sites. Colored arrowheads represent primers used to generate pKAES323 and to screen the transformants. (B) Southern-blot analysis of E. festucae strains. Wild-type E2368 and transformants were probed with a lolN fragment or lolM gene amplified from E2368 (old probe was stripped off the membrane before new hybridization). Lanes contained HindIII-digested genomic DNA from E2368 (WT), lolN-lolM knockout transformant (KO), ectopic transformant of E2368 with pKAES323 (Ect), and E2368 transformed with the empty vector pKAES173 (WT+vec).
One lolN and lolM double-knockout transformant was obtained from 150 transformants. This knockout strain was confirmed by Southern-blot analysis to lack full-length lolN and lolM (Fig. 3B). The knockout strain, an ectopic transformant, and an empty-vector transformant were inoculated back into endophyte-free meadow fescue seedlings to check their loline-alkaloid profiles (E. festucae produces lolines only in symbio). Plants with the knockout mutant contained NANL as the only loline alkaloid detected, whereas plants with the control and wild-type strains contained NFL, NAL, and NANL (Fig. 4). To confirm that the change of loline alkaloid profile was due to loss of a functional copy of lolN and lolM, a complementation plasmid that had lolN and lolM under their native promoter was introduced into the knockout mutant. Three independent complementation strains were obtained and inoculated into meadow fescue to check the loline alkaloid profile. All three transformants restored production of NFL and NAL in planta (Fig. 4). These results indicate that lolN and lolM are required in the pathway from NANL to the biosynthetic end product, NFL.
(A) The lolN-lolM knockout (KO), (B) an empty-vector control transformant (WT+vec), and (C and D) complementation strains (KO+lolN+lolM). The numbers after complementation strains represent different meadow fescue plants inoculated with independent transformants. (E) Proposed roles of LolN and LolM (this work), and reported role of LolP , in the biosynthetic pathway from N-acetylnorloline (NANL) to the final product, N-formylloline (NFL).
Natural NANL accumulators have mutations or deletions of lolN
If lolN and lolM are indeed involved in the production of NFL from NANL as we hypothesized, it is conceivable that endophytes that produce NANL as the end product may have mutations in or deletions of lolN. In order to test this hypothesis, LOL clusters of several loline-alkaloid producers were compared. Endophytes that showed one of the three chemotypes, NANL, AcAP, or NFL as the end product shared most of the LOL genes except for lolO, lolP, lolN or lolM (Table 2). Strains with mutations in, or complete lack of, lolO accumulated AcAP, consistent with the previously reported function of lolO . Interestingly, these strains also lacked functional copies of lolN, lolM, and lolP. Strains that had a functional lolO, but lacked or had a mutated lolN, all accumulated NANL. Conversely, those strains from plants that accumulated NANL but not NFL, NML, or NAL had lost or obviously defective lolN, lolM and lolP genes, with the sole exception of Epichloë coenophiala strain e4309 (discussed next).
In contrast to other NANL accumulators, strain e4309 had a LOL cluster with the complete set of 11 genes and no obvious inactivating mutation in any of them. Therefore, we aligned and compared lolN gene sequences and predicted amino acid sequences of e4309 to those of known NFL producers. LolN of e4309 had two nonsynonymous changes, G495D and E551K, compared to LolN of strains that produced NFL (Fig. 5). Moreover, when strain e4309 was compared to two other E. coenophiala strains that produce NFL (e19 and e4163), amino acid sequences of all LOL genes were identical except for the aforementioned G495D and E551K changes in lolN of e4309. Reverse transcription-PCR (RT-PCR) indicated that lolN and lolM were expressed by e4309 in symbio. Thus, one or both of the two alterations in LolN observed in e4309 might have rendered the protein dysfunctional for deacetylation of NANL.
Red-framed sequences are three different E. coenophiala isolates. Ecoe = Epichloë coenophiala, Eaot = Epichloë aotearoae, Efes = Epichloë festucae, Echis = Epichloë chisosa, Esig = Epichloë siegelii, Eunci = Epichloë uncinata.
Results from experiments with the lolN and lolM double knockout, and comparison of LOL clusters of various species and strains, consistently showed that NANL accumulation was associated with lolN dysfunction, supporting its predicted function as an acetamidase that would deacetylate NANL to form norloline in the biosynthetic pathway.
LolM expressed in yeast catalyzed methylation of norloline and loline
To test if the lolM gene product functions as a methyltransferase to methylate norloline in consecutive steps to form loline and NML, we expressed lolM in yeast. Crude protein extract from LolM-expressing yeast was prepared and incubated with AdoMet plus either loline or norloline. Loline and a smaller amount of NML were observed from application of norloline to the crude protein extract when incubated with AdoMet (Fig. 6). Extract of yeast transformed with the empty vector failed to catalyze methylation of norloline. Similarly, the protein extract from LolM-expressing yeast catalyzed the conversion of almost half of the added loline to NML, whereas no increase in NML was observed after incubating loline and AdoMet with protein extract from yeast transformed with the empty vector (Fig. 6). These results demonstrated that LolM was responsible for methylation of norloline to loline, and loline to NML.
(A) Chromatogram of loline alkaloids from incubation of norloline and AdoMet with protein extract of yeast transformed with empty vector. (B) Chromatogram of loline alkaloids from incubation of norloline and AdoMet with crude protein extract from yeast that expresses LolM. (C) Chromatogram of loline alkaloids from incubation of loline and AdoMet with protein extract of yeast transformed with empty vector. (D) Chromatogram of loline alkaloids from incubation of loline and AdoMet with crude protein extract from yeast that expresses LolM. (E) Proposed scheme of loline and N-methylloline (NML) formation from norloline. AdoHcy = S-adenosyl homocysteine.
Plant acetyltransferase responsible for NAL production
The acetylated form of loline, N-acetylloline (NAL), is also a common loline alkaloid observed in many infected grasses together with NANL and NFL. It has been reported that symbiota comprising the endophyte species FaTG-3 and its natural host, tall fescue (Lolium arundinaceum), contain NFL and NAL, but when FaTG-3 is inoculated to perennial ryegrass (Lolium perenne), the resulting symbiota contain NFL, but lack NAL . Furthermore, we observed that tall fescue and meadow fescue symbiotic with Epichloë siegelii accumulated NFL, NANL, and NAL, whereas some perennial ryegrass plants symbiotic with the same fungus contained NFL and NANL, but lacked NAL, indicating that plant genotype affects production of NAL. In addition, no NAL was detected in loline-alkaloid producing cultures of E. uncinata e167, consistent with a previously reported observation , whereas meadow fescue symbiotic with e167 possessed NAL and other loline alkaloids. A culture of e167 sometimes gave a small peak of similar m/z as NAL, which migrated in GC near, but not coincident with, NAL, when compared to extracts from meadow fescue-E. uncinata e167 symbiota (S1 Fig.). We judged this peak as probably representing N-propionylnorloline (also known as decorticasine) based on comparison with extracts from Adenocarpus decorticans, in which decorticasine is the major pyrrolizidine alkaloid , .
We hypothesized that a plant enzyme is required to acetylate loline to form NAL. To test this hypothesis, loline was applied to asymbiotic meadow fescue and perennial ryegrass, and the alkaloid profile of the symbiota was analyzed. Compound NAL was observed in extracts from loline-fed meadow fescue, but not from loline-fed perennial ryegrass (Fig. 7), suggesting that a grass acetyltransferase enzyme catalyzes the conversion of loline to NAL. This result was surprising, considering that most of the known loline biosynthetic pathway is catalyzed by enzymes encoded in the fungal LOL cluster , , . This interesting finding raises questions for future investigations, such as the following: Does the conversion of loline to NAL occur within or outside of the plant cells? What is the evolutionary origin of this plant acetyltransferase? Is it specific for NAL production, or does it catalyze other acetylation reactions as well? Is there genotypic variation for NAL production within grass species? Are there other metabolic pathways with both fungal and plant components?
Shown are loline alkaloids extracted from loline applications to (A and B) endophyte-free (E-) meadow fescue (MF), and (C and D) E- perennial ryegrass (PRG), (E) mass spectrum of N-acetylloline (NAL) from application of loline to E- meadow fescue, and (F) proposed scheme of NAL formation from loline. Quinoline was added as internal standard (istd). Unlabeled peaks are non-loline alkaloid compounds. Numbers after MF or PRG indicate independent trials.
An acetyltransferase is also needed to produce AcAP from 1-AP, but this enzyme does not seem to be of plant origin, because AcAP is produced in E. uncinata cultures outside of the plant system. It will be intriguing to learn how the two partners of the symbiosis communicate with each other and regulate the possible trafficking of loline alkaloids between them.
Evolution and implications of loline alkaloid diversity
Alkaloid profiling showed that Epichloë species produced different loline alkaloids depending on the gene content of their LOL clusters. Fig. 8 summarizes current knowledge of the biosynthetic pathway, and indicates the genes that, by their presence or absence, determine loline alkaloid diversity. When there was a full complement of the 11 LOL genes, NFL, NAL, and NANL were produced in symbio, but when there was mutation in or deletion of lolO or lolN, the accumulated end-product was AcAP or NANL, respectively (Table 1). Remnants of lolN, lolO, lolM, or lolP were often found in strains with NANL or AcAP as end-products, indicating that evolution of these strains involved loss of genes from the ancestral 11-gene LOL cluster. In strains for which NANL was the end product, lolN genes showed nonsynonymous point mutations (e4309), a partial deletion (E2772), or complete absence (E57 and e4815). Among strains with AcAP end-product, lolO had been inactivated by apparently independent mutations . The multiple inactivations and losses of late pathway genes that alter loline alkaloid profiles suggest that loline alkaloid production may be under diversifying selection.
Labeled arrows are for steps that contribute to diversity of the lolines. Presence or absence of functional copies of lolO, lolN, lolM, or lolP, or a plant acetyltransferase activity, determine which alkaloids accumulate in the symbiotic plant as the pathway end-products.
We also found that the plant can directly contribute to diversity of these fungal metabolites. An acetyltransferase activity associated with the host plant is responsible for conversion of loline to NAL, adding a layer of chemotypic complexity to this close fungus-plant symbiotic interaction. Although it is well established that symbiotic organisms may be interdependent in production of indispensible metabolites , we are unaware of any other report of one partner modifying specialized (secondary) metabolites produced by the other in a mutualistic symbiosis.
Loline alkaloids show broad-spectrum anti-insect activity, and different lolines may have different biological activities. In vitro tests of NFL, NAL, NML, and semisynthetic loline derivatives with long carbon-chain acylations on the 1-amine have shown that many are effective against both fall armyworm (Spodoptera frugiperda Smith) larvae and European corn borer (Ostrinia nubilalis Hübner) larvae, but the effects seem to differ depending on the modifications. N-Formylloline reduces the weight gain of fall armyworms by deterring feeding, and does not significantly affect corn borers. In contrast, NAL reduces the weight gain of corn borer larvae without changing larval feeding behavior, indicating that its effect is due to metabolic toxicity . N-Formylloline, NAL, and NML are almost as potent as nicotine in insecticidal activity against green bugs (Schizaphis graminum Rondani). Furthermore, NML shows a similar regression curve to nicotine sulfate, indicating a similar pharmacological mechanism . It has also been shown that lolines (mainly NFL and NAL) reduce oviposition of adult Argentine stem weevils . Another study found that NFL reduces the growth, development, and survival of Argentine stem weevil larvae at high concentrations (800 and 1600 µg/g dry weight), whereas NANL causes high mortality of Argentine stem weevil larvae, but with little effect on the growth or development of the larvae . These findings clearly show that the chemically diverse loline alkaloids exert complex and multimode effects on grass herbivores. Moreover, loline alkaloids also affect insect parasitoids , , thereby contributing to multitrophic endophyte effects. Thus, the diversification of loline alkaloids produced by Epichloë species probably reflects selection imposed by the complex and variable biotic environments of their plant hosts.
Chromatograms of N-propionylnorloline (decorticasine) and N-acetylloline (NAL) in loline-alkaloid-producing systems. (A) Comparison of loline alkaloids from Adenocarpus decorticans seed, Epichloë uncinata e167 in inducing culture, and meadow fescue symbiotic with e167. Only decorticasine and NAL peaks are labeled. (B) Structures of decorticasine and NAL.Chromatograms of N-propionylnorloline (decorticasine) and N-acetylloline (NAL) in loline-alkaloid-producing systems. (A) Comparison of loline alkaloids from Adenocarpus decorticans seed, Epichloë uncinata e167 in inducing culture, and meadow fescue symbiotic with e167. Only decorticasine and NAL peaks are labeled. (B) Structures of decorticasine and NAL.
Oligonucleotides used in this study.
We are grateful to Carolyn A. Young at The Samuel Roberts Noble Foundation and Jennifer A. Rudgers at the University of New Mexico for providing plants; University of Kentucky AGTC members Jennifer S. Webb for genome sequencing, Jolanta Jaromczyk for genome assembly, and Cagney Coomer for Sanger sequencing. We are also grateful to John May at the University of Kentucky ERTL facility for GC-MS analysis, W. John Layton for assistance with NMR analysis, Neil Moore and Daniel R. Harris for genome annotations, Hexiang Luan for help with western-blot analysis, and J. Douglas Brown for maintaining the plants in the greenhouse. Capable technical assistance was also provided by laboratory manager Walter Hollin and undergraduate Trevor Kellen. This is publication number 14-12-090 of the Kentucky Agricultural Experiment Station, published with permission of the director.
Conceived and designed the experiments: JP RBG CLS. Performed the experiments: JP MB PN. Analyzed the data: JP MB PN. Contributed reagents/materials/analysis tools: JP MB PN. Wrote the paper: JP RBG CLS.
- 1. Schardl CL, Florea S, Pan J, Nagabhyru P, Bec S, et al. (2013) The epichloae: alkaloid diversity and roles in symbiosis with grasses. Curr Opin Plant Biol 16:480–488.
- 2. Schardl CL, Grossman RB, Nagabhyru P, Faulkner JR, Mallik UP (2007) Loline alkaloids: currencies of mutualism. Phytochemistry 68:980–996.
- 3. Moon CD, Craven KD, Leuchtmann A, Clement SL, Schardl CL (2004) Prevalence of interspecific hybrids amongst asexual fungal endophytes of grasses. Mol Ecol 13:1455–1467.
- 4. Ghimire SR, Rudgers JA, Charlton ND, Young C, Craven KD (2011) Prevalence of an intraspecific Neotyphodium hybrid in natural populations of stout wood reed (Cinna arundinacea L.) from eastern North America. Mycologia 103:75–84.
- 5. Kang Y, Ji Y, Zhu K, Wang H, Miao H, et al. (2011) A new Epichloë species with interspecific hybrid origins from Poa pratensis ssp. pratensis in Liyang, China. Mycologia 103:1341–1350.
- 6. Schardl CL, Young CA, Hesse U, Amyotte SG, Andreeva K, et al. (2013) Plant-symbiotic fungi as chemical engineers: multi-genome analysis of the clavicipitaceae reveals dynamics of alkaloid loci. PLoS Genet 9:e1003323.
- 7. Schardl CL, Young CA, Pan J, Florea S, Takach JE, et al. (2013) Currencies of mutualisms: sources of alkaloid genes in vertically transmitted epichloae. Toxins (Basel) 5:1064–1088.
- 8. Georghiou GP (1972) The evolution of resistance to pesticides. Annu Rev Ecol Evol Syst 3:133–168.
- 9. Schardl CL, Chen F (2010) Plant defences against herbivore attack. Encyclopedia of Life Sciences: John Wiley & Sons, Ltd. doi: doi:https://doi.org/10.1002/9780470015902.a0001324.pub2.
- 10. Riedell WE, Kieckhefer RE, Petroski RJ, Powell RG (1991) Naturally-occurring and synthetic loline alkaloid derivatives – insect feeding-behavior modification and toxicity. J Entomol Sci 26:122–129.
- 11. Popay AJ, Tapper BA, Podmore C (2009) Endophyte-infected meadow fescue and loline alkaloids affect Argentine stem weevil larvae. N Z Plant Prot 62:19–27.
- 12. Schardl CL, Grossman RB, Nagabhyru P, Faulkner JR, Mallik UP (2007) Loline alkaloids: currencies of mutualism. Phytochemistry 68:980–996.
- 13. Spiering MJ, Moon CD, Wilkinson HH, Schardl CL (2005) Gene clusters for insecticidal loline alkaloids in the grass-endophytic fungus Neotyphodium uncinatum. Genetics 169:1403–1414.
- 14. Spiering MJ, Faulkner JR, Zhang DX, Machado C, Grossman RB, et al. (2008) Role of the LolP cytochrome P450 monooxygenase in loline alkaloid biosynthesis. Fungal Genet Biol 45:1307–1314.
- 15. Pan J, Bhardwaj M, Faulkner JR, Nagabhyru P, Charlton ND, et al. (2014) Ether bridge formation in loline alkaloid biosynthesis. Phytochemistry 98:60–68.
- 16. Blankenship JD, Houseknecht JB, Pal S, Bush LP, Grossman RB, et al. (2005) Biosynthetic precursors of fungal pyrrolizidines, the loline alkaloids. ChemBioChem 6:1016–1022.
- 17. An ZQ, Siegel MR, Hollin W, Tsai HF, Schmidt D, et al. (1993) Relationships among non-Acremonium sp. fungal endophytes in five grass species. Appl Environ Microbiol 59:1540–1548.
- 18. Faulkner JR, Hussaini SR, Blankenship JD, Pal S, Branan BM, et al. (2006) On the sequence of bond formation in loline alkaloid biosynthesis. ChemBioChem 7:1078–1088.
- 19. Blankenship JD, Spiering MJ, Wilkinson HH, Fannin FF, Bush LP, et al. (2001) Production of loline alkaloids by the grass endophyte, Neotyphodium uncinatum, in defined media. Phytochemistry 58:395–401.
- 20. Tsai HF, Siegel MR, Schardl CL (1992) Transformation of Acremonium coenophialum, a protective fungal symbiont of the grass Festuca arundinacea. Curr Genet 22:399–406.
- 21. Panaccione DG, Johnson RD, Wang J, Young CA, Damrongkool P, et al. (2001) Elimination of ergovaline from a grass-Neotyphodium endophyte symbiosis by genetic modification of the endophyte. Proc Natl Acad Sci U S A 98:12820–12825.
- 22. Inoue I, Namiki F, Tsuge T (2002) Plant colonization by the vascular wilt fungus Fusarium oxysporum requires FOW1, a gene encoding a mitochondrial protein. Plant Cell 14:1869–1883.
- 23. Latch GCM, Christensen MJ (1985) Artificial infections of grasses with endophytes. Ann Appl Biol 107:17–24.
- 24. Starnes JH, Thornbury DW, Novikova OS, Rehmeyer CJ, Farman ML (2012) Telomere-targeted retrotransposons in the rice blast fungus Magnaporthe oryzae: agents of telomere instability. Genetics 191:389–406.
- 25. Hopp TP, Prickett KS, Price VL, Libby RT, March CJ, et al. (1988) A short polypeptide marker sequence useful for recombinant protein identification and purification. Nat Biotechnol 6:1204–1210.
- 26. Sikorski RS, Hieter P (1989) A system of shuttle vectors and yeast host strains designed for efficient manipulation of DNA in Saccharomyces cerevisiae. Genetics 122:19–27.
- 27. Rajendran KS, Nagy PD (2004) Interaction between the replicase proteins of Tomato Bushy Stunt virus in vitro and in vivo. Virology 326:250–261.
- 28. Pogany J, Stork J, Li Z, Nagy PD (2008) In vitro assembly of the Tomato bushy stunt virus replicase requires the host Heat shock protein 70. Proc Natl Acad Sci U S A 105:19956–19961.
- 29. Dereeper A, Guignon V, Blanc G, Audic S, Buffet S, et al. (2008) Phylogeny. fr: robust phylogenetic analysis for the non-specialist. Nucleic Acids Res 36:W465–W469.
- 30. Ball OJ-P, Tapper BA (1999) The production of loline alkaloids in artificial and natural grass/endophyte associations. Proceedings of the NZ Plant Protection Conference 52:264–269.
- 31. Ribas I, Alonso De Lama JM (1953) Alkaloids of the papilionoideas. XII. Alkaloids of Adenocarpus hispanicus D.C. from the Sierra del Guadarrama (Madrid). Farmacognosia 13:367–374.
- 32. Veen G, Greinwald R, Canto P, Witte L, Czygan FC (1992) Alkaloids of Adenocarpus hispanicus (Lam.) DC varieties. Z Naturforsch C 47:341–345.
- 33. Spiering MJ, Wilkinson HH, Blankenship JD, Schardl CL (2002) Expressed sequence tags and genes associated with loline alkaloid expression by the fungal endophyte Neotyphodium uncinatum. Fungal Genet Biol 36:242–254.
- 34. Zientz E, Dandekar T, Gross R (2004) Metabolic interdependence of obligate intracellular bacteria and their insect hosts. Microbiol Mol Biol Rev 68:745–770.
- 35. Jensen JG, Popay AJ, Tapper BA (2009) Argentine stem weevil adults are affected by meadow fescue endophyte and its loline alkaloids. N Z Plant Prot 62:12–18.
- 36. Bultman TL, Rodstrom JL, Radabaugh KR, VanDop JD, Librizzi JM, et al. (2009) Influence of genetic variation in the fungal endophyte of a grass on an herbivore and its parasitoid. Entomol Exp Appl 130:173–180.
- 37. Bultman TL, Borowicz KL, Schneble RM, Coudron TA, Bush LP (1997) Effect of a fungal endophyte on the growth and survival of two Euplectrus parasitoids. Oikos 78:170–176.