Cool-season grasses of the Poaceae sub-family Pooideae, including the important temperate forage grasses, tall fescue, and perennial ryegrass, are often infected with clavicipitaceous fungal endophytes that include sexual and asexual Epichloë species (Felitti et al. 2006; Posselt et al. 2006; Card et al. 2014). The associations between asexual Epichloë species and their host grasses are generally considered to be mutualistic, due to provision of benefits to both partners (Clay 1988). The plant provides certain benefits to the endophyte such as shelter, nutrition, reproduction, and distribution (Scott and Schardl 1993; West 1994; Schardl et al. 2004). The resident endophyte relies entirely on the host plant for dissemination through either seed or growth of vegetative structures (Philipson and Christey 1986; Schardl et al. 2004). Conversely, as part of the mutualistic association, the presence of endophyte provides several advantages to the host plant (Khan et al. 2010). Benefits for the host related to abiotic stress tolerance are obtained through enhanced growth, increased seedling vigour, persistence, particularly under water stress and nutrient deficiency (Ravel et al. 1997; Malinowski and Belesky 1999). The host plant is able to tolerate a broad range of environmental conditions, due to growth advantages provided by the endophyte, which contributes significantly to ecological success (Easton 1999). The endophyte also confers biotic stress tolerance to the host grass, through production of several classes of biologically active alkaloids that are toxic to both arthropod and mammalian herbivores. Production of secondary metabolites also enhances the resistance of ryegrass to nematodes and some fungal pathogens (Kimmons et al. 1990; Gwinn and Gavin 1992; Schardl and Phillips 1997; Johnson et al. 2003). Of the alkaloids produced by Epichloë endophytes, the ergot alkaloids and indole diterpenes are active against vertebrate and invertebrate herbivores (Leuchtmann et al. 2000; Schardl et al. 2013a). Epoxy-janthitrems are tremorgenic mycotoxins, but are much less toxic to grazing vertebrates than lolitrem B, and may also exhibit anti-invertebrate activity (Philippe 2016). In contrast, peramine and lolines are highly active specifically against insect pests but have not been correlated with any toxic symptoms in grazing mammals (Schardl et al. 2013b).

A considerable amount of recent research has been performed on the selection of Epichloë endophytes that produce minimal detrimental effects on livestock, while retaining the ability to produce alkaloids that provide adequate control of insects and nematodes (Philippe 2016). For example, novel fungal endophyte strains that lack or produce very low concentrations of lolitrem B and ergovaline such as AR1 and NEA2 (both of which produce peramine as the major alkaloid) have been used to infect perennial ryegrass and have been subjected to commercialisation (Rasmussen et al. 2009). Introduction of peramine biosynthesis into non-peramine-producing strains as well as increased levels in peramine-producing strains would be anticipated to further increase resistance to invertebrate pests and provide corresponding benefits to agriculture.

Peramine was originally identified in extracts of perennial ryegrass which had formed associations with Epichloë festucae var. lolii, and provides potent protection against herbivorous insect pests such as the Argentinian stem weevil (Listronotus bonariensis), an economically important exotic pest of perennial ryegrass in New Zealand (Bush et al. 1997; Schardl et al. 2012, 2013b). Peramine deters feeding of both adults and larvae at concentrations of 0.1 µg/g and 10 µg/g, respectively (Rowan et al. 1990). Furthermore, deterrence of feeding by peramine has the ability to significantly enhance the competitiveness of endophyte-positive (E+) grasses (Schardl et al. 2004). Other possible beneficial properties of peramine production include allelopathy, and inhibition of pathogens or seed predators (Faeth et al. 2002).

A single, multifunctional fungal gene, designated perA, is required for peramine biosynthesis (Tanaka et al. 2005). Although perA sequences are broadly distributed across both sexual and asexual Epichloë species, not all strains share the ability to synthesise peramine in planta (Scott et al. 2007). Presence of an intact copy of perA is directly correlated with the capacity to synthesise peramine (Schardl et al. 2013b). DNA sequence analysis has identified a large-scale deletion in the 3′-terminal region of the gene in some endophyte strains which are unable to produce the metabolite (Scott et al. 2007). The presence of miniature inverted repeat transposable elements (MITEs) has been associated with such deletion events (Fleetwood et al. 2011). Of four distinct taxa that colonise perennial ryegrass, peramine-producing strains that belong to E. festucae var. lolii (LpTG-1; Lolium perenne taxonomic group) and LpTG-2 possess intact perA genes, while non-peramine-producing strains (belonging to LpTG-3 and LpTG-4) are characterised by partial deletions and premature stop codons (Hettiarachchige et al. 2015). The complete perA gene encodes two adenylation (A1 and A2) domains, two thiolation (T1 and T2) domains, a methylation (M) domain, a condensation (C) domain, and a reductase (R) domain, which may be sufficient to produce peramine (Schardl et al. 2013b; Berry et al. 2015). Several variants of perA with variability of the number of functional domains have been identified in different Epichloë species. Expression of a perA variant which lacks the reductase domain may produce a metabolite related to peramine (Schardl et al. 2013b; Berry et al. 2015).

The development of endophyte associations with minimal or no detrimental effects, while retaining beneficial characteristics, is important for successful pastoral agriculture. To achieve this outcome, various different methodologies, from conventional selective breeding to modern biotechnological approaches, have been used in contemporary grass breeding programs (Gundel et al. 2013). Genetic modification of endophytes with useful genes may be used to confer additional valuable attributes on the host, following inoculation with the suitably altered endophyte (Mei and Flinn 2010). The genes required for the biosynthesis of alkaloids provide obvious targets for such modification. For example, ergovaline production has been eliminated from an Epichloë–grass association through ablation of a peptide synthetase gene, named lps, which is essential for ergopeptine biosynthesis (Panaccione et al. 2001).

The initial aim of the present study was to study the feasibility of artificially combining biosynthesis of two alkaloids by a single endophyte strain through introduction of a functional perA gene copy into an LpTG-3 endophyte strain by A. tumefaciens-mediated transformation. Subsequently, the feasibility of enhanced peramine production through introduction of an additional perA gene copy/copies into LpTG-1 and LpTG-2 endophyte strains was assessed. Finally, RNA-Seq-based expression analysis of perA genes from selected ryegrass endophytes belonging to distinct taxonomic groups was performed, to determine the functional status of the perA promoter, and the potential for restoration of truncated perA genes by genome editing technology.

Materials and methods

Strains and culture conditions

The properties of the selected endophyte strains, in terms of taxon (Hettiarachchige et al. 2015), alkaloid biosynthesis profile, observed growth characteristics, number of perA gene copies, and peramine production phenotype are summarised in Table 1. Endophyte cultures were grown either on potato dextrose agar (PDA; Fluka, St. Louis, Missouri, USA), or in PDB at 22 °C in the dark for a period of 7–10 days, depending on growth rate. A. tumefaciens strain AGL-1, which was used for transformation, and Escherichia coli strain DH5α (Thermo Fisher Scientific, Waltham, MA, USA) which was used during construction and maintenance of plasmid constructs, were grown at 28 °C and 37 °C, respectively, on either LB (Luria–Bertani) agar (5 g yeast extract, 10 g tryptone, 10 g NaCl in 1L of ddH2O) plates or in LB broth. PDA and LB media were supplemented with appropriate antibiotics as necessary.

Table 1 Properties of Epichloë endophytes used for transgenic modification

Plasmid construction

Gene cassettes were supplied by GeneArt (Regensburg, Germany) as follows: the constitutive glyceraldehyde-3-phosphate dehydrogenase promoter (gpdP) and tryptophan biosynthesis C gene terminator (trpCT) of Aspergillus nidulans, for cloning of the first reading frame A [RFA-A] cassette; the hygromycin B resistance gene (hph) under the control of the A. nidulans trpC promoter (trpCP) and terminator; the perA gene sequence of SE, flanked by attB1 and attB2 recombination sites.

The Gateway™-enabled destination vector (pEND0002) was constructed through modifications of the T-DNA region of pPZP200 (Hajdukiewicz et al. 1994).

A 2.1 kb XbaI/KpnI fragment containing the hph expression cassette (from pMA2) was force-cloned across the corresponding sites of pPZP200, so generating pEND0001. The RFA-A cassette (1.7 kb; Invitrogen, Carlsbad, California, USA) was cloned between gpdP and trpCT at the SmaI site of the GeneArt vector, pMA1. Single colonies were screened by restriction endonuclease analysis for the presence of the RFA-A cassette in the desired (sense) orientation. The 2.9 kb gpdP-(RFA-A)-trpCT cassette was then released by digestion with SbfI and HindIII and force-cloned across the corresponding sites of pEND0001 to generate pEND0002.

The 8562 bp perA gene fragment containing attB1 and attB2 sites was excised from the 11ABFUUC peramine pMA vector using the restriction endonucleases AscI and PacI and was cloned into the pDONR 221 (Invitrogen, Carlsbad, California, USA) vector using BP clonase following the manufacturer’s recommended method (Invitrogen, Carlsbad, California, USA) to generate the entry clone, pEND0024. This was combined with the destination vector pEND0002 using LR clonase following the manufacturer’s recommended method (Invitrogen, Carlsbad, California, USA) to generate a final expression vector, pEND0025 (Supplementary table S1). The structures of all expression clones were verified by Sanger sequencing.

Sensitivity of non-transgenic endophyte mycelia to hygromycin B

The sensitivity of non-transformed endophyte mycelia to treatment with hygromycin B was determined by plating 400 µL of endophyte culture onto PDA plates containing increasing concentrations of antibiotic (50, 100, 150, 200, 250, and 300 µg/mL), and incubation of plates at 22 °C for 4 weeks.

A. tumefaciens-mediated transformation

A. tumefaciens cells (AGL1) were transformed with the binary vectors by electroporation. Transformants were isolated on LB agar plates supplemented with 20 mg/L rifampicin and appropriate antibiotics for selection of the A. tumefaciens strain and binary vector. A. tumefaciens-mediated transformation was performed as previously described (Tanaka et al. 2007), with minor modifications.

Mitotic stability of transformed endophytes

Putative endophyte transformants (appearing 2–4 weeks after transformation) were transferred to fresh selection medium and mitotic stability was determined by sub-culturing five successive times every 2–3 weeks, alternating between media containing and lacking 200 µg/mL hygromycin B. The stability of the putative transformants was also confirmed by PCR detection of the hph gene.

Molecular analysis of transformed endophytes

Genomic DNA was extracted from freeze-dried fungal mycelia of putative transformants using the QIAGEN DNeasy plant mini kit (QIAGEN, Hilden, Germany). PCR analysis was used to confirm the presence of T-DNA using the following primer pairs: gpdP-perA-F (5′- AGTCCCTGGTAGGCAGCTTT -3′), gpdP-perA-R (5′- CAGAGGCCAAGATTGTTTGG-3′), hph-F (GATGTTGGCGACCTCGTATT) and hph-R (GAATTCAGCGAGAGCCTGAC). PCR cycling conditions consisted of an initial denaturation cycle of 10 min at 95 °C, followed by 30 cycles with 30 s denaturation (95 °C), 30 s annealing (59 °C) and 1 min polymerisation (72 °C).

Metabolic profile analysis of endophyte cultures using liquid chromatography/mass spectroscopy (LC/MS)

Sample preparation

A randomly selected sub-set of PCR-positive transformants were tested for peramine production in in vitro culture. Five putative transformants, along with the corresponding non-transgenic endophyte strains, were sub-cultured on PDA plates containing 15 mM urea. Four replicates from each endophyte strain were prepared and grown for 8 weeks. Plates were stored at − 80 °C until required. Samples were freeze-dried for 48 h and ground to a fine powder using a genogrinder (SPEX SamplePrep 2010 Geno/GrinderR, Metuchen, USA). The resulting powder was extracted by adding two volumes of 80% (v/v) methanol. Samples were vortexed briefly, sonicated for 10 min, vortexed again, and centrifuged at 10,000 rpm for 10 min. Three extractions were performed for each sample and all supernatants were combined. The solvent was evaporated using a SpeedVac Concentrator (Savant SPD2010, Thermo Fisher Scientific, Waltham, MA, USA). The residue was dissolved in 2 mL of 80% (v/v) methanol and collected in a vial for LC/MS analysis.

LC/QTOF mass spectrometry

The analytical system used was the Agilent UHPLC 1290 Infinity binary LC system (Agilent Technologies, Santa Clara, California, USA) and Agilent Technologies 6538 UHD Accurate-Mass QTOF Mass Spectrometer with Electrospray Ionisation (ESI) source. Analysis of samples with full-scan acquisition mode in positive ionisation mode was performed. A methanol-only negative control (blank) was run as every sixth sample (five samples followed by a blank) of analysis. The data acquisition software was Agilent MassHunter Qualitative Analysis software (version B.06.00). Chromatographic separation was performed using a Hypersil GOLD™ column (150 × 2.1, 1.9 µm) (Thermo Fisher Scientific, Waltham, MA, USA). The mobile phase was composed of 0.1% (v/v) aqueous formic acid (solvent A) and 0.1% (v/v) formic acid in acetonitrile (solvent B). The gradient elution program is detailed in Supplementary Table S2. Peramine standard [Peramine nitrate (BDG Synthesis, Wellington, New Zealand)] was used to construct concentration curves at 1, 20, 40, 80, 200, 400, 800, 1000, 2000, 4000, and 10,000 ng/mL in matrix and peramine concentrations of samples were calculated in parts per million (ppm). Elution time for peramine standard was 3.6 ± 0.05 min.

QTOF conditions included a gas temperature 8 L/min at 300 °C; nebulizer pressure 30 psig, capillary voltage 3.5 kV. Signals in the m/z (mass-to-charge ratio) 100–1700 range were recorded. MS/MS conditions: collision energy of 20 eV was used to fragment the parent ions in the m/z 100–1700 range. Acquisition rate was1 spectrum/s.

Data analysis

Data processing was performed with the Agilent MassHunter Qualitative Analysis software (version B.06.00 [6.0.633.10]). Identification of peramine was based on retention time and accurate mass. Identity of peramine was further confirmed from fragmentation data. Confidence of identification was expressed by an overall identification score, the weights of these parameters being: mass score = 100, isotope abundance score = 60, isotope spacing score = 70, MS mass 2.0 mDa + 5.6 ppm, MS isotope abundance 7.5%, and MS/MS mass 5.0 mDa + 7.5 ppm.

Average peramine concentration produced by each transgenic endophyte was compared to that of non-transgenic endophyte and statistical analysis was performed using the Mann–Whitney–Wilcoxon Test (Mann and Whitney 1947).

Seedling inoculation of transgenic endophytes

Endophyte-devoid (E) perennial ryegrass cultivar Alto seeds were obtained from New Zealand Agriseeds (NZA), Christchurch, New Zealand. These seeds have been generated from an endophyte-free parental population which has never been inoculated with endophyte. Prior to use for inoculations, DNA was extracted from a representative sample of this seed batch using the DNeasy Plant Mini Kit (QIAGEN, Hilden, Germany) and screened using three endophyte-specific SSR markers (NLESTA1QA09, NLESTA1CC05, and NLESTA1NG03) (Kaur et al. 2015; van Zijll de Jong et al. 2008) to confirm the absence of endophyte. A total of 50–100 seedlings were inoculated with each transgenic endophyte strain, together with non-transgenic counterparts. The number of inoculations performed was based on the documented success rates for the respective non-transgenic endophyte strains (K. Guthridge, unpublished).

E seeds were surface-sterilised (5% [w/v] NaOCl and Tween 20) for 3 h with shaking at room temperature. Following washing, seeds were plated onto water agar (26 g/L bacteriological agar) and incubated in dark at 20–22 °C for 6–8 days. Fresh endophyte mycelium was inserted into a parallel incision made across the shoot meristem. Inoculated seedlings were incubated in a growth room (23 °C, 80 µMm−2sec−1, 8–16 h photoperiod) for 7 days and then transferred to the glasshouse (16 h photoperiod, 60% relative humidity, natural lighting and an average controlled temperature of 25 °C) and planted in 42-cell plant trays filled with potting mix (30 L of seeding raising mix [Van Schaik’s Biogro, Dandenong South, Victoria, Australia] mixed with 0.9 L of fine vermiculite, 0.6 L of fine perlite, 25 g of micro nutricote, 20 g of water-holding granules, and 2.5 g of trace elements) and grown for 16 weeks under glasshouse conditions. Non-inoculated E Alto seeds were also germinated under the same conditions for use as a control.

Screening for endophyte presence

Two-to-three tiller samples (c. 0.5 cm from the base) of each endophyte-inoculated and endophyte-free plant were harvested, following 16 weeks growth in soil. DNA extraction and SSR genotyping were performed as previously described using three endophyte-specific SSR markers (NLESTA1QA09, NLESTA1CC05, and NLESTA1NG03) (van Zijll de Jong et al. 2008; Kaur et al. 2015). Four plants from each host–endophyte combination (Alto-SE, Alto-NEA11, and Alto-NEA12) and E Alto were then transplanted into 8 cm × 8 cm pots and maintained under glasshouse conditions as described before. When most of the plants had grown to a density of > 30 tillers (one transgenic SE-inoculated plant failed to survive and one non-transgenic SE-inoculated plant exhibited poor growth with a few tillers at this stage and was removed from further analysis), individual plants were split into seven clonal replicates and transplanted into pots of the same size, followed by maintenance under the same glasshouse conditions. Total number of transgenic and non-transgenic samples tested for peramine quantification were 21 for SE (seven clonal replicates of three different host plant genotypes) and 28 for NEA11 and NEA12 (seven clonal replicates of four different host plant genotypes).

Metabolic profile analysis of Alto–endophyte associations using liquid chromatography/mass spectroscopy (LC/MS)

The pseudostem component of plants was harvested and snap-frozen in liquid nitrogen. Samples were then freeze-dried and extracted twice from 20 mg of each clonal replicate using 80% (v/v) methanol, and the final residue was dissolved in 200 µL of 80% (v/v) methanol and collected in a vial for the LC/MS analysis.

The extracts were analysed using a 100 mm × 2.1 mm Thermo Hypersil Gold 1.9 µm HPLC column fitted to a Thermo Fisher Scientific Vanquish liquid chromatograph (Thermo Fisher Scientific, Waltham, MA, USA). Metabolites were eluted from the column using a gradient mobile phase, (A) 0.1% formic acid in water, (Thermo Fisher Scientific, Waltham, MA, USA) and (B) 0.1% formic acid in acetonitrile (Thermo Fisher Scientific, Waltham, MA, USA) at 0.3 ml/min. The initial conditions (98% A) preceded initiation of a linear gradient to 0% A over 11 min, and this was maintained for 4 min before returning to the initial gradient conditions. The compounds were detected with a Thermo Fisher QExactive Plus mass spectrometer (Thermo Fisher Scientific, Waltham, MA, USA), operating in the ESI mode with an HESI probe for positive data acquisition. Standards (peramine [Peramine nitrate {BDG Synthesis, Wellington, New Zealand}], ergotamine [Ergotamine D-tartrate {Sigma–Aldrich, St. Louis, MO, USA}], and lolitrem B (Reddy et al. 2015)) were used to construct concentration curves at 1, 10, 25, 50, 100, 250, 500, 1000, and 2000 ng/mL in matrix.

For liquid chromatography–mass spectrometry analysis, 3 µL of sample extract was injected onto the LC/MS system. Each extract was analysed using a QExactive Mass Spectrometer (Thermo Fisher Scientific, Waltham, MA, USA) in FT positive mode over a mass range of 80–1200 amu with resolution set at 35,000. Typical mass accuracy for the alkaloids was 3 ppm or better.

The presence of janthitrem I was confirmed by accurate mass and by MS2 analysis of the ion at m/z 646.377 with the same ion at the same retention time in the Alto-AR37 controls (AR37 being used as the positive control for janthitrem I production, as no standards are currently commercially available for this class of alkaloid). Elution times for peramine, ergovaline, lolitrem B, and janthitrem I were 3.57 ± 0.02 min, 5.37 ± 0.01 min 11.17 ± 0.02 min and 11.19 ± 0.02 min, respectively.

Concentrations of peramine were calculated from the standard curve and average peramine concentration produced by each transgenic endophyte infected plant and the overall average peramine concentration produced by all transgenic endophyte infected plants were compared to the overall average peramine concentration produced by non-transgenic endophyte infected plants.

Analysis of perA gene expression

E perennial ryegrass cultivar Alto seeds, and those containing the endophyte strains SE (Standard Endophyte), NEA11, and NEA12 were obtained from New Zealand Agriseeds (NZA), Christchurch, New Zealand. Seeds were surface-sterilised in 5% (w/v) NaOCl with gentle shaking for 3 h. Seeds were then washed thoroughly with sterile water and germinated on wet sterile filter papers in the dark for 2 days and in the light for a further 8 days (23 °C, 80 µMm–2sec−1, 8–16 h photoperiod). RNA was extracted from 10-day-old seedlings using a CTAB-based method as previously described (Chang et al. 1993) with minor modifications. Four biological replicates were prepared for each endophyte strain. Libraries were prepared using the Agilent SureSelect strand-specific RNA library preparation kit (Agilent Technologies, Santa Clara, California, USA), and sequencing was performed on an Illumina Hiseq 3000 platform ( (Illumina Inc., San Diego, California, USA).

Raw sequence reads were quality trimmed using the Gydle nuclear program ( (Gydle Inc., Quebec, Canada), such that all reads ≥ 50 bases long were retained using a base quality cut-off value of 20. All high-quality reads for a given endophyte strain were mapped separately to the SE perA gene sequence using Bowtie (version 1.0.0) (Langmead et al. 2009) and visualised with Tablet (version (Milne et al. 2013).


A. tumefaciens-mediated transformation of endophyte mycelia

The sensitivities of non-transgenic SE, NEA11, and NEA12 to the selective antibiotic, hygromycin B, were assessed prior to inception of the transformation experiments, to determine whether the hph gene could be used as the selectable marker. Colonies of each strain grew on hygromycin B concentrations up to 150 µg/mL, and at higher concentrations, negligible growth was observed. A higher threshold concentration of hygromycin B (200 µg/mL) was consequently chosen for the selection of putative transformants.

The expression vector (pEND-perA) generated by replacing the Gateway RFA-A cassette in vector pEND0002 by the perA gene of SE was transformed into A. tumefaciens strain AGL1. PCR analysis was performed to confirm the presence of the transformed plasmid (data not shown). The A. tumefaciens strain was co-cultivated with finely cut endophyte mycelium from the experimental strains SE, NEA11, and NEA12. Hygromycin B-resistant colonies started to develop 2–4 weeks after transformation, suggesting that all three endophyte strains (and hence taxa) were receptive to A. tumefaciens-mediated T-DNA transfer. The number of resistant transformants generated for SE, NEA11, and NEA12 was 39, 43, and 78 per 1 g of mycelia, respectively. The negative controls that were used, being the corresponding non-transgenic endophyte or A. tumefaciens strain, did not give rise to resistant colonies, providing confidence that genuine transformation events were selected.

Stable transformation events were confirmed through the ability of these putative transformants to grow after sub-culturing onto fresh selection plates containing higher hygromycin concentrations (300 µg/mL). Successful expression of hph, a commonly used selectable marker for fungal transformation, suggests that the A. nidulans trpC promoter is effective for the expression of transgenes in the target endophyte strains.

PCR analysis and mitotic stability assessment

A total of 36 hygromycin B-resistant transformants (12 from each endophyte strain) that were selected on antibiotic selection media were subjected to PCR to identify the presence of introduced perA and hph genes. Primers for the perA gene were designed that flank both the gpd promoter and the introduced perA gene, to permit discrimination from the endogenous genes that are present in all three strains. The expected sizes of PCR products for perA (201 bp) and hph (414 bp) were observed for all transformants, confirming the presence of stable transformation events. No amplification products were produced by non-transgenic endophyte strains.

PCR-positive transformants maintained the ability to grow on hygromycin B after alternate sub-culturing on to PDA, with and without hygromycin B, for a minimum of five sub-cultures, so confirming mitotic stability. The stability of the putative transformants was also confirmed by PCR detection of the hph gene.

Metabolic profiling

The ability of transformants to produce peramine in in vitro culture was assessed following molecular characterisation of transformants. For this purpose, a total of five randomly selected PCR-positive transformants from each endophyte strain were tested. The presence of peramine was identified based on retention time and accurate mass, while further confirmation was achieved using multiple mass spectrometry (MS/MS). All five transformants containing the perA gene for each endophyte strain that was analysed demonstrated the presence of a specific peak with a retention time of 3.8 ± 0.15 min, and were well within 6.5 ppm of the m/z 248.15063 [M + H]+, so confirming the ability of transformants to produce peramine in culture. This result demonstrates that the presence of a functional copy of the perA gene is capable of restoring production to an otherwise non-producing strain, NEA12, and further confirms the sole responsibility of the perA gene for peramine biosynthesis. Non-transgenic SE and NEA11 also produced the same specific peak with the expected retention time, while no specific peak was observed for non-transgenic NEA12, as expected. The successful expression of the perA gene suggests that the A. nidulans gpd promoter is effective for the expression of transgenes in the target endophyte strains, which represent three distinct taxa.

The identity of peramine as a metabolite was further confirmed based on mass spectrometric fragmentation pattern. The fragment ion spectra of m/z (mass-to-charge ratio) 248.15063 [M + H] + for 13 out of 15 of the transformants included all of the distinct fragment ions m/z 149, m/z 175, m/z 206, and m/z 231 (Fig. 1). The remaining two transformants generated only m/z 175 and m/z 248 of those ions, probably due to low abundance. Non-transgenic SE and NEA11 generated ions only with m/z 175 (data not shown).

Fig. 1
figure 1

Mass fragment spectrum derived from the parent ion of peramine. a Transgenic  SE_1, b Transgenic NEA11_1, and c Transgenic NEA12_1. The distinct fragment ions m/z 149, m/z 175, m/z 206, and m/z 231 are indicated by arrows

Level of peramine production

Variable levels of peramine production were observed for transgenic SE, NEA11, and NEA12 endophyte strains containing perA genes in in vitro culture. Differences between expression levels of transformants of the same strain could be due to variability of the number of perA gene copies that have been integrated, as well as the influence of the genomic site of integration in endophyte genomes. Very low levels of peramine were detected for non-transgenic SE and NEA11 (Fig. 2).

A significant difference (p < 0.05) in peramine production was observed for all transgenic SE and NEA11 when compared to the corresponding non-transformants. This result is consistent with the observation that small amounts of peramine are produced by endophytes in culture, compared to higher levels detected in E+ grass tissue. The activity of the constitutive gpd promoter, rather than the native perA promoter, was presumably responsible for the elevated level of peramine biosynthesis observed in culture for transgenic SE and NEA11 when compared to their non-transgenic counterparts. A higher level of peramine production was observed for most NEA12 transformants when compared to SE and NEA11 transformants.

Fig. 2
figure 2

Variation in peramine production by non-transgenic and transgenic SE, NEA11, and NEA12 endophyte strains in in vitro culture. Each bar shows the average peramine concentration (ppm) produced by four replicates of each endophyte strain and error bars denote standard errors of the mean. *Indicates a significant statistical difference (p < 0.05) between transgenic and corresponding non-transgenic endophyte samples. Non-transgenic NEA12 is not shown as it did not produce any peramine in culture

Seedling inoculation of transgenic endophytes

Inoculation of ryegrass seedlings with PCR-verified transgenic endophyte strains that produce peramine in culture, together with the corresponding non-transgenic strains, was performed to determine the capacity to form effective associations and produce peramine in planta. Putative associations were screened 16 weeks after transfer to soil with Epichloë-specific SSR marker assays, as explained above. All three combinations of non-transgenic and transgenic endophyte strains established associations with their hosts, suggesting that transgene presence did not exert inhibitory effects on inoculation ability. The infection frequencies obtained are summarised in Table 2.

Table 2 Infection frequencies obtained for different endophyte strains

E+ plants (both non-transgenic and transgenic) were transferred to pots from cell trays and maintained under glasshouse conditions, followed by re-analysis 12 weeks later using the same SSR marker assay, to assess vegetative stability of the association. PCR-based screening for the presence of transgenic sequences was also confirmed. For each strain, four transgenic and four corresponding non-transgenic endophyte infected plants were used for the detection of peramine, ergovaline, lolitrem B, and janthitrem I. Analysis of LC/MS data revealed that all strains produced the expected alkaloid profiles.

Interestingly, transgenic NEA12-infected plants were able to produce both peramine and janthitrem I, while non-transgenic NEA12-infected plants produced only janthitrem I. Variable levels of peramine production were observed for non-transgenic and transgenic SE and NEA11 as well as transgenic NEA12-infected plants: such variation may be due to genotypic differences between host plants. The levels of peramine production detected for transgenic NEA12 infected plants were low compared to those of SE and NEA11 (Fig. 3). This observation differs from that made for peramine production in culture (Fig. 2) where most NEA12 transformants produced high level of peramine and suggests that in planta peramine production in SE and NEA11 is primarily due to the activity of the endogenous gene copies under the up-regulated native perA promoter. Conversely, the low level of peramine produced in planta by transgenic NEA12-infected plants is due to control by the constitutive fungal promoter.

Fig. 3
figure 3

Variation in peramine production by non-transgenic and transgenic SE, NEA11, and NEA12 in planta. Error bars denote standard errors of the mean. *Indicates a significant statistical difference (p < 0.05) in peramine production between transgenic and corresponding non-transgenic endophyte infected plants. a First three bars from the top show average peramine concentration produced by each transgenic SE-infected plant (n = 7). The fourth and fifth bars show the overall average peramine concentration produced by transgenic and non-transgenic SE-infected plants (n = 21), respectively. b, c First four bars from the top show average peramine concentration produced by each transgenic NEA11- and NEA12-infected plant (n = 7), respectively. The fifth and sixth bars show the overall average peramine concentration produced by transgenic and non-transgenic NEA11- and NEA12-infected plants, respectively (n = 28)

Plants with expected alkaloid profile were split into seven clonal replicates, grown for another 12 weeks and re-analysed, producing alkaloid profiles consistent with the previous results. Only one transgenic NEA11-infected plant produced a significantly high (P < 0.05) peramine, while no significant difference (P < 0.05) in peramine production was detected for the other three transgenic NEA11-infected plants when compared to the corresponding non-transgenic NEA11. Furthermore, no significant difference was observed between overall average levels of peramine produced by non-transgenic and transgenic NEA11. Of three transgenic SE-infected plants that were analysed, two produced a significantly higher level of peramine, and one produced significantly less peramine when compared to the corresponding non-transgenic counterparts. As a consequence, the overall average peramine produced by transgenic SE in planta was significantly higher than that of non-transgenic SE (Fig. 3). However, this effect may be mostly due to the genotypic effect of the host plant on peramine production.

Expression of the perA gene in different Alto–endophyte associations

The level of peramine produced in planta under the control of the A. nidulans gpd promoter by transgenic NEA12-infected plants was very low compared to the other peramine-producing endophytes used in this study, despite the widespread use of a fungal constitutive promoter. Consequently, an expression analysis of the perA gene was performed to test the possibility of introducing peramine production into a non-peramine-producing strain using the native promoter and novel genome editing techniques.

High-quality reads from each Alto–endophyte association were mapped to the SE-derived perA gene sequence, numbering 1294, 3791, and 523 for Alto-SE, Alto-NEA11, and Alto-NEA12 associations, respectively. Read distribution along the gene length demonstrated full coverage for SE and NEA11 and partial coverage for NEA12 (Fig. 4). Transcription of the perA gene up to coordinate 5475 bp was observed for NEA12, which was beyond the premature stop codon identified at coordinate 2923 bp from gene sequence analysis. Expression of the truncated perA gene of NEA12 confirms functionality of the native promoter, supporting the potential for using genome editing techniques to restore full gene function.

Fig. 4
figure 4

Tablet output showing Illumina HiSeq-derived reads mapped to the perA gene of SE: a full-length perA gene (Berry, Takach, et al. 2015), b Alto-SE, c Alto-NEA11, and d Alto-NEA12 associations


A. tumefaciens-mediated transformation of the perA gene

The SE-derived perA gene sequence was used to construct an expression vector under the control of the A. nidulans gpd promoter, which is known to function across a broad range of Ascomycete species (Mikkelsen et al. 2003). The expression vector was constructed using Gateway™ cloning, which provides a simple and highly efficient system for gene cloning and rapid target gene transfer between vectors (Zhu et al. 2009). Modifications of the conventional vectors by introduction of the Gateway cassette have been used extensively for functional studies in filamentous fungi (Oliveira et al. 2008; Shafran et al. 2008; Zhu et al. 2009).

A. tumefaciens-mediated transformation is effective for the transformation of filamentous fungi due to versatility of choice in starting material (such as protoplasts, hyphae, spores, or blocks of mycelial tissue), high transformation efficiency, and a less labour-intensive nature than protoplast-requiring transformation protocols (de Groot et al. 1998; Mullins et al. 2001; Amey et al. 2002; Fitzgerald et al. 2003; Meyer et al. 2003; dos Reis et al. 2004). Growth of endophyte mycelia was completely inhibited by hygromycin B at 200 µg/mL, similar concentrations (200–300 µg/mL) having been successfully used for A. tumefaciens-mediated transformation of both E. festucae var. lolii and other ascomycete species (Mikkelsen et al. 2001; Panaccione et al. 2001; Knight et al. 2009).

Metabolic profile analysis of transformants

PCR-positive transformants were analysed for the production of peramine in in vitro culture. Restoration of peramine production by NEA12 transformants confirms that the peramine-devoid phenotype of NEA12 is due to the effects of the deletion and premature stop codon within the perA gene sequence. It also confirms the sole responsibility of the perA gene for peramine biosynthesis, consistent with the outcomes of other studies (Tanaka et al. 2005).

Endophytes utilise nitrogen both for growth and for the biosynthesis of alkaloids, which are nitrogen-rich compounds (Faeth and Fagan 2002). Alkaloid levels have been reported to be potentially enhanced by increased nitrogen availability in E+ grasses, especially under moderate water stress (Faeth et al. 2002). In contrast, reductions of both alkaloid concentrations and endophyte prevalence have been found in some studies when plants were grown at high nitrogen levels (Rasmussen et al. 2009). However, the highest level of loline biosynthesis in culture for E. uncinata has been obtained when urea is provided as the sole nitrogen source (Blankenship et al. 2001). Therefore, transformants containing perA gene in this experiment were grown on PDA containing urea for the analysis of peramine production. All sample extractions were performed under cold conditions to minimise the possibility of false negatives, and additionally, the absence of false positives was verified by analysis of negative controls between each of the five samples.

Methods that have been commonly used so far for peramine identification include mass spectroscopy, ultraviolet spectroscopy, and thin-layer chromatography (Rowan 1993). Of these, detection of peramine by MS/MS has been found to be c. 50-fold more sensitive than UV-based methods (Koulman et al. 2007). Peramine production by transformants generated in this study was analysed using LC/MS, and the metabolite was identified based on retention time and accurate mass measurements. In the mass spectrum, the fragments exhibited m/z values of 206 (C11H16O1N3) corresponding to loss of diazomethane, m/z 231 corresponding to loss of ammonia (NH3) and two more distinct ions with m/z of 175 and 149. These results were consistent with the published mass spectral data (Koulman et al. 2007).

As expected based on perA gene sequence structure, no production of peramine was observed for non-transgenic NEA12, while very low levels of peramine production were observed for non-transgenic SE and NEA11. Consistently, production of low amounts of peramine in culture has been previously reported for E. festucae var. lolii and E. typhina endophytes, in contrast to higher levels detected in E+ grass tissue (Rowan 1993; Tanaka et al. 2005). In contrast to the respective non-transgenic counterparts, substantial production of peramine in culture was apparent for transformants of all three endophyte strains, albeit at variable levels. Activity of the constitutive gpd promoter was presumably responsible for the elevated level of peramine biosynthesis observed in culture.

Expression of peramine in transgenic endophyte-containing plants

Subsequently, in planta peramine production was determined. Non-transgenic and transgenic endophyte strains of all three taxa established associations with their hosts suggesting that transgene presence did not exert inhibitory effects on inoculation ability. Infection frequency is based on endophyte identity as well as on host plant genotype and frequencies obtained in this study are consistent with the previous studies (Kaur et al. 2015).

As observed in vitro, failure to produce peramine is rescued in transgenic Alto-NEA12 associations when compared to non-transgenic counterparts, albeit at low levels compared to SE- and NEA11-containing plants. Expression under control of the native perA promoter may be up-regulated in planta (Khan et al. 2010), as compared to a constitutive fungal promoter. A significantly higher average level of peramine production was detected for transgenic SE-infected plants as compared to non-transgenic SE-containing plants. However, this effect may be due to the severely restricted growth of one Alto non-transgenic SE association. No significant increases in average level of peramine production were observed when comparing transgenic NEA11-containing plants to their counterparts that contain non-transgenic NEA11, consistent with a low level of additional peramine biosynthesis from the transgene.

Transcriptional analysis of the perA gene

Isolation of perA gene sequences from perennial ryegrass-derived endophytes revealed the presence of full-length copies in LpTG-1 and LpTG-2 strains, while representatives of LpTG-3 and LpTG-4 lack the 3′-terminal region of perA, reducing the length of the gene from 8505 bp to 7282 bp (Hettiarachchige et al. 2015). LpTG-3 strains also contain a premature stop codon at coordinate 2923 bp within the perA gene sequence (Hettiarachchige et al. 2015).

Mapping of RNA-Seq reads to the full-length perA genes of SE and NEA11 in the present study demonstrated the expression of all functional domains, as evident from metabolic profiling and gene sequence analysis. The full-length perA gene encodes a multifunctional protein with two adenylation and thiolation domains and single condensation, methylation, and reductase domains (Tanaka et al. 2005). Interestingly, mapping of reads to the 5′-portion of the truncated perA gene in NEA12 revealed expression, suggesting that the cognate promoter is active in LpTG-3 strains. This observation supports the potential for using genome editing approaches to restore full gene function in this genetic background in future. Chemotypic analysis of E. festucae E2368 has revealed that isolates lacking the reductase domain are unlikely to produce peramine, although all the other domains are present (Young et al. 2012; Schardl et al. 2013b). It has been suggested that E. festucae E2368 has the ability to produce a peramine-related metabolite (Schardl et al. 2013b). This transcriptional analysis strongly supports the hypothesis that the truncated perA gene is expressed, and also suggests the NEA12 perA gene encodes at least one functional domain.


The beneficial properties of endophytes in agricultural applications may be enhanced through the use of naturally occurring strains with desirable qualities, or through targeted genetic modification of endophytes. The results of this proof-of-concept study clearly demonstrate that production of two alkaloids may be artificially combined in a single endophyte strain through transgenesis, so producing a novel strain with enhanced beneficial properties for grassland agriculture. Future experimental studies are needed to test the feasibility of improving levels of in planta peramine production through the use of a stronger exogenous promoter. This study further demonstrates the potential for use of genome editing techniques to produce large amounts of peramine under the control of the native promoter, based on effective expression of a truncated gene copy.