Summary
Phytophthora infestans (Phy. infestans) is a devastating pathogen of tomato and potato. It readily overcomes resistance genes and applied agrochemicals. Fungal endophytes provide a largely unexplored avenue of control against Phy. infestans. Not only do endophytes produce a wide array of bioactive metabolites, they may also directly compete with and defeat pathogens in planta.
Twelve isolates of fungal endophytes from different plant species were tested in vitro for their production of metabolites with anti-Phy. infestans activity. Four well-performing isolates were evaluated for their ability to suppress nine isolates of Phy. infestans on agar medium and in planta.
Two endophytes reliably inhibited all Phy. infestans isolates on agar medium, of which Phoma eupatorii isolate 8082 was the most promising. It nearly abolished infection by Phy. infestans in planta.
Here we present a biocontrol agent, which can inhibit a broad-spectrum of Phy. infestans isolates. Such broadly acting inhibition is ideal, because it allows for effective control of genetically diverse pathogen isolates and may slow the adaptation of Phy. infestans.
Introduction
Phytophthora infestans is a major pathogen of cultivated tomato (Solanum lycopersicum) and cultivated potato (Solanum tuberosum). Even today its impact cannot be ignored as it is still capable of destroying entire fields of its hosts, leading to up to 100% yield losses (Nowicki et al., 2012). The two major control measures for Phy. infestans are resistance breeding and agrochemical applications. While several resistance genes have been identified in screens of wild relatives of S. lycopersicum and S. tuberosum (Song et al., 2003; Van der Vossen et al., 2003; Pel et al., 2009; Zhang et al., 2013), many of them are readily overcome by isolates of Phy. infestans (Vleeshouwers et al., 2011). Similarly, agrochemicals can have a low durability in their protective function against Phy. infestans (Grünwald et al., 2006; Childers et al., 2015). Hence, continual scientific effort in terms of breeding, development of agrochemicals and other approaches, such as biological control, is needed for effective crop protection against this pathogen.
One approach that is gaining more and more attention is the use of endophytes for crop protection (Le Cocq et al., 2016). Endophytes are microorganisms that grow within plants, and at the time of sampling, do not cause obvious symptoms on their host (Schulz & Boyle, 2005; Le Cocq et al., 2016). Many studies have explored the bacterial, fungal and protist endophytic communities associated with different plants (e.g. Bulgarelli et al., 2012; Lundberg et al., 2012; Bodenhausen et al., 2013; Schlaeppi et al., 2013; Bulgarelli et al., 2015; Edwards et al., 2015; Busby et al., 2016a; Coleman-Derr et al., 2016; Ploch et al., 2016). These studies indicate that the diversity of microbes living inside of plants is largely underestimated and that the distribution of some microorganisms is host and/or environment specific.
Furthermore, in some cases such endophytic microorganisms have been evaluated for their potential benefit to their hosts (Busby et al., 2016b). Such benefits include growth promotion and protection against parasites and pathogens (e.g. Schulz, 2006; Lahlali & Hijri, 2010; Tellenbach & Sieber, 2012; Panke-Buisse et al., 2015; Rolli et al., 2015; Busby et al., 2016a; Hiruma et al., 2016; Martínez-Medina et al., 2017). Often these functions are linked to metabolites produced and secreted by the endophytes (Son et al., 2008; Puopolo et al., 2014; Mousa et al., 2016; Suryanarayanan et al., 2016), highlighting the endophyte’s metabolic versatility(Schulz et al., 2002; Strobel & Strobel, 2007; Verma et al., 2009; Mousa & Raizada, 2013; Brader et al., 2014). Endophytes may also directly compete with potential pathogens of their host plants (Alabouvette et al., 2009), induce plant defense responses (Shoresh et al., 2010) and/or produce bioactive anti-microbial metabolites (Brader et al., 2014). An example of an endophyte that can be applied as a direct competitor of a plant pathogenic organism is Phlebiopsis gigantea (Adomas et al., 2006). Phl. gigantea prohibits the infection of stumps of coniferous trees by the pathogen Heterobasidion annosum sensu lato and thereby limits the spread of the pathogen (e.g.Annesi et al., 2005). Due to its success in limiting the spread of H. annosum s.l., Phl. gigantea has been made commercially available. An example for the induction of defense responses by an endophyte is the barley root endophyte Piriformospora indica, which induces a jasmonic acid-dependent defense response in its host upon co-inoculation with a pathogen (Stein et al., 2008). Furthermore, a recent study by Mousa et al. (2016) describes an Enterobacter sp. strain isolated from an ancient African crop (Eleusine coracana [finger millet]) with the ability to suppress the grass pathogen Fusarium graminearum. Enterobacter sp. traps F. graminearum in the root system of its host and simultaneously produces several antifungal compounds that kill the fungus.
Several bacterial and fungal endophytes, with the potential to inhibit Phy. infestans growth, have been described (Sturz et al., 1999; Kim et al., 2007; Miles et al., 2012; Puopolo et al., 2014). However, these endophytes have only been tested against single isolates of Phy. infestans; but alternative approaches, such as biocontrol, can show different outcomes depending on the pathogen isolate (Bahramisharif et al., 2013). Therefore, the identification of endophytic species with a broad inhibition spectrum is of critical importance.
In this study, we screened the metabolite extracts of 12 fungal endophytes isolated from different plant hosts for their ability to inhibit growth of Phy. infestans. Using a plate assay with the four most successful fungal endophytes, we show that they inhibit the growth of a broad spectrum of European Phy. infestans isolates in co-culture. According to our phylogenetic analyses, these four endophytes are members of the Ascomycota. The endophyte with the strongest inhibition potential both on plates and in planta was Pho. eupatorii, isolate 8082. This endophyte prohibited proliferation of Phy. infestans and in some cases abolished its infection completely. Since we identified Pho. eupatorii based on the inhibition potential of its metabolite extract, the active component may be a secreted metabolite or a cocktail of different metabolites. A broad-spectrum activity as observed for Pho. eupatorii suggests either a conserved target for such secreted metabolite(s) or several targets that are specific for the pathogen isolate and that are covered by the complexity of the metabolite cocktail. Both can result in slower counter-adaptation of Phy. infestans to either the direct application of the endophyte or to the application of its metabolites. Therefore, Pho. eupatorii isolate 8082 is a potential novel broad-spectrum biocontrol agent of Phy. infestans.
Material and Methods
Isolation of endophytes
To isolate the endophytes, plant tissues of the respective hosts (Tab. S1) were first thoroughly washed under running water, then immersed for one minute in 70% ethanol, followed by 1-3min in 3% NaOCl and subsequently rinsed three times in sterile water. Sterilized tissues were imprinted on potato-carrot medium (Höller et al., 2000) to test for effectiveness of sterilization and to optimize the sterilization procedure. The tissues were then cut with a sterile scalpel into 2mm slices and plated on potato-carrot agar medium with antibiotics (Höller et al., 2000) and incubated for 3 weeks at 20°C. The emerging mycelia were taken into culture on potato-carrot agar medium and were initially identified according to morphology (Tab. S1).
Screening crude metabolite extracts for anti-Phytophthora infestans activity
To test the growth inhibition potential of the 12 fungal endophytes, the endophytes were first grown on barley-spelt medium (Schulz et al., 2011) and/or biomalt agar medium (Höller et al., 2000) at room temperature for 21 days. To isolate the secondary metabolites, the cultures were extracted with ethyl acetate. 25μl of culture extract (40 mg/ml) were then applied to a filter disc and placed onto rye agar medium that had been inoculated with Phy. infestans isolate D2; subsequent incubation was at 20°C in the dark (Schulz et al., 2011). Only fungal endophytes whose culture extracts resulted in a zone of inhibition ≥ 20mm were used for further analyses.
Co-culture on plates
The fungal endophytic isolates 8082, 9907 and 9913, whose culture extracts had inhibited Phy. infestans in the agar diffusion assays and Phialocephala fortinii isolate 4197 (Schulz, 2006) were tested for their bioactivity against nine isolates of the late blight pathogen Phy. infestans (NL10001, NL88069, NL90128, IPO-C, IPO428-2, 3928A, D12-2, T15-2 and T20-2). The Phi. fortinii isolate was included based on previous experiments (Schulz et al., 2002; Schulz, 2006; Schulz, unpublished). The co-culture experiments were performed and evaluated according to Peters et al. (1998). Fungal endophytes and Phy. infestans isolates were grown on rye-sucrose agar (RSA, Caten & Jinks, 1968) at room temperature. The duration of the experiments was dependent on the endophytes’ growth rates: eight days for all co-cultures that included 9913 and 14 to 16 days for the remaining co-cultures. A minimum of ten plates were analyzed per treatment. The Mann-Whitney U test (Mann & Whitney, 1974) was used to determine if differences between co-culture and control plates were significant. Average growth inhibition was estimated as: 1-(average radius in co-culture / average radius in control conditions). All experiments were evaluated again after eight weeks of incubation to assess long-term effects. Pictures were taken with an EOS 70D camera (Canon).
Co-inoculation in planta
The surfaces of the S. lycopersicum seeds were sterilized using 70% ethanol for 3sec, followed by ~5% NaOCl for 30sec. The sterilized seeds were washed three times with sterile water for 3min. Seeds were incubated in the dark on 1.2% H2O-agar with a day-night temperature cycle of 18°C /15°C (16h / 8h). After three days, the seeds were transferred to a day-night cycle with 16h light (166 ± 17μmol quanta*m-2*s-1). Temperature conditions were the same as before. Nine to 11 days post sterilization (dps), the germinated seedlings were transferred to 9mm petri dishes containing 0.5% MS-medium (Murashige & Skoog, 1962) with 1% sucrose, poured as a slope.
An endophyte mycelial suspension was prepared from a two-to four-day old liquid culture for each endophyte (potato-carrot liquid medium; 100g potato-carrot mash [prepared according to Höller et al., 2000] in 1l medium). Mycelium was equally dispersed in 25ml medium using Tissuelyser II (Qiagen, Hilden, Germany) for a few seconds. Preliminary inoculations of S. lycopersicum roots with 25 to 50μl of mycelial suspensions of all four endophytes were prepared. Endophyte isolate 9907 and Phi. fortinii killed the seedlings. Hence, only endophyte isolates 8082 and 9913 were used for further inoculation studies.
For inoculations with endophyte isolate 8082, 5μl or 10μl of the mycelial suspension or H2O (mock control) was applied to each root at 16dps. After 27dps seedlings were transferred to vessels (10cm×6.5cm×6.5cm) with MS agar medium. For inoculations with endophyte isolate 9913, 10μl of dispersed mycelium or H2O was applied to the roots of axenic seedlings at 18dps. However, the endophyte isolate 9913 did not grow sufficiently, so we performed a second inoculation with undispersed mycelium from the liquid culture at 22dps. These seedlings were transferred to vessels at 28dps. At 34 to 36dps each leaflet of endophyte and mock inoculated plants was inoculated with 10μl of Phy. infestans zoospore suspension (4°C cold) or with 10μl H2O (4°C cold). The zoospore suspension (5*104 spores/ml) was harvested from a 25 days old culture of Phy. infestans isolate D12-2 and was kept on ice during the entire procedure. For the Phy. infestans zoospore isolation see de Vries et al. (2017). Plants were sampled for microscopic evaluation, to evaluate anthocyanin content and pathogen abundance at three days post inoculation with Phy. infestans.
To confirm endophytic colonization by the fungi, roots from the mock control, endophyte inoculated and co-inoculated samples were surface sterilized using three protocols: i) 70% EtOH for 3sec (isolate 8082) or 30sec (isolate 9913), ~5% NaOCl for 30sec, followed by washing three times with sterile H2O for 3min each (treatment 1), ii) 70% EtOH for 5min, 0.9% NaOCl for 20min, followed by washing three times with H2O (treatment 2, Cao et al., 2004) and iii) 97% EtOH for 30sec, 10% NaOCl for 2min, followed rinsing four times with H2O (treatment 3, Terhonen et al., 2016). Roots were imprinted on RSA agar plates to test for efficacy of sterilization and then placed on new RSA agar plates. The plates were evaluated at 8dps (isolate 8082) and 6dps (isolate 9913).
Microscopy
Two aspects of host physiology were evaluated microscopically following the co-inoculation: chlorophyll intensity and relative necrotic area. Pictures to evaluate chlorophyll intensity were taken with the SMZ18 dissection microscope and a DS-Ri1 camera (Nikon, Tokyo, Japan) using a 600 LP filter (Transmission Filterset F26-010, AHF Analysetechnik, Tübingen, Germany), with an exposure time of 200ms and 100% gain. Intensity was measured using ImageJ2 (Schindelin et al., 2015). Pictures for necrosis measurements were taken with a SteREO Discovery V8 binocular and an AxioCam ICc5 camera (Zeiss, Göttingen, Germany). The relative necrotic area was calculated as the necrotic area of a leaflet over the total area of the leaflet. The necrotic and total leaflet area were estimated using the ZEN Blue edition (Zeiss, Göttingen, Germany). Differences in relative necrotic area and chlorophyll content in the treatments were calculated using a Kruskal-Wallis test (Kruskal & Wallis, 1952) combined with a Tukey post-hoc test (Tukey, 1949) and using a Benjamini-Hochberg correction for multiple testing (Benjamini & Hochberg, 1995).
Anthocyanin content evaluation
The anthocyanin content was measured and calculated according to Lindoo and Caldwell (1978). We analyzed three to six biological replicates per treatment. Samples were tested for normality using a Shapiro-Wilk test (Shapiro & Wilk, 1965) and for equal variance. Accordingly, significant differences were calculated using a two-sided t-test with the assumption of equal or unequal variances depending on the sample combination tested. All statistical analyses were done in R v 3.2.1.
DNA and RNA extraction and cDNA synthesis
DNA was extracted from the mycelium of the fungal endophytes and Phy. infestans isolates grown on RSA medium using the DNeasy® Plant Mini Kit (Qiagen, Hilden, Germany). RNA was extracted from infected and mock control leaflets of seedlings of S. lycopersicum using the Universal RNA/miRNA Purification Kit (Roboklon, Germany). Three to four leaflets were pooled per replicate. To evaluate RNA quality, 5μl of RNA were treated with 6μl deionized formamide, incubated at 65°C for 5min, followed by 5min incubation on ice. This mixture was then visualized on a 2% agarose gel. To ensure that no DNA contamination was present, the RNA was treated with DNAse I (Thermo Scientific). Reactions were adjusted for 200ng of total RNA. cDNA was synthesized with the RevertAid First Strand cDNA Synthesis Kit (Thermo Fisher Scientific, Lithuania).
Molecular identification of endophytes
To determine the phylogenetic placement of the fungal endophytes, we sequenced their internal transcribed spacer (ITS) and 5.8S regions. ITS1 and ITS4 primers were used (White et al., 1990). The 20μl PCR-reaction contained 1x Green GoTaq® Flexi Buffer, 0.1mM dNTPs, 2mM MgCl2, 1U GoTaq® Flexi DNA Polymerase (Promega, Madison, WI, USA), 0.2μM of each primer and 40-95ng of template DNA. The PCR protocol included an initial denaturation step of 95°C for 3min, followed by 35 cycles of a denaturation step at 95°C for 30sec, an annealing step at 60°C for 30sec and an elongation step at 72°C for 90sec, followed by a final elongation step of 72°C for 7min. All PCR products were purified with the peqGOLD Cycle-Pure Kit (Peqlab, Erlangen, Germany). The products were cloned into the pCR™ 4-TOPO® vector of the TOPO® TA Cloning® Kit for Sequencing (Invitrogen, Carlsbad, CA, USA) and the plasmid DNA was extracted with the QIAprep Spin Miniprep Kit (Qiagen, Hilden, Germany). Sequencing was performed at Eurofins MWG Operon (Ebersberg, Germany). Sequences were blasted using BLASTn (Altschul et al., 1990) and the best hits were retrieved. To assemble a dataset of closely related organisms from which to infer the phylogenetic placement of the unknown endophytes, the sequences of species with high similarity to our initial query sequences were downloaded. Taxonomic classification of these sequences was done using mycobank.org (provided by the CBS-KNAW Fungal Biodiversity Center, Utrecht). Additional sequences were retrieved from GenBank (Tab. S2). Taxonomically distant outgroups were chosen based on the systematic classifications in MycoBank (Crous et al., 2004). The sequences were aligned using CLUSTAL-W and a Neighbor-Joining phylogeny was inferred using the Kimura-2 model with 5 gamma categories and pairwise deletion of gaps. One hundred bootstrap replicates were evaluated. All analyses were done using MEGA 5.2.2 (Tamura et al., 2011).
Assessment of endophyte and Phytophthora infestans growth after eight weeks of co-culture
To determine whether either the endophyte had overgrown Phy. infestans or Phy. infestans had overgrown the endophyte on the co-culture plates, we performed PCR reactions on DNA extracted from both sides of eight-week old co-cultures of five to nine Phy. infestans isolates with Phi. fortinii, isolate 8082 and isolate 9913 as well as their respective controls. We amplified the ITS loci (for primers see White et al., 1990) and the cytochrome oxidase subunit2 (COX2) using Phytophthora-specific primers from Hudspeth et al. (2000) with the protocol described above. Between 50-100ng of template DNA was used.
Presence and abundance of Phytophthora infestans
To quantify the abundance of Phy. infestans in the seedlings pre-inoculated with the two endophytes (isolate 8082 and 9913) and the seedlings only inoculated with Phy. infestans, we performed a quantitative RT-PCR (qRT-PCR). The two markers, PiH2a and PiElf1α, were used for the pathogen and the three markers, SAND, TIP and TIF3H, were used as tomato (host) reference genes (de Vries et al., 2015; de Vries et al., 2017). Two independent qRT-PCR runs were used for the pathogen genes. All qRT-PCRs were performed in a CFX ConnectTM Real-Time System (Bio-Rad, Hercules, CA, USA) and included an initial denaturation at 95°C for 3min, followed by 40 cycles of a denaturation step at 95°C for 10sec and an annealing and elongation step of 60°C for 45sec. For PiH2a the annealing temperature was lower: 59°C in the first run and 55°C in the second run. For the following experiment, each run contained three biological replicates for: i) isolate 8082 (5μl mycelial suspension) with Phy. infestans, ii) isolate 9913 with Phy. infestans and iii) Phy. infestans without endophyte. Two biological replicates were completed for isolate 8082 (10μl mycelial suspension) with Phy. infestans. In each run, we analyzed three technical replicates for each biological replicate, resulting in six technical replicates for each biological replicate for both marker genes. To calculate the relative abundance of Phy. infestans in these samples, we set the Cq-values of those biological replicates that gave no biomass marker amplicon to 41. As the two independent runs gave the same results, they were combined. PiH2a and PiElf1α expression was then calculated according to Pfaffl (2001). Data were tested for normal distribution using a Shapiro-Wilk test and the appropriate statistical tests were then applied. For co-inoculations with isolate 8082, significant differences were calculated using a Mann-Whitney U-test. For co-inoculations with isolate 9913, significant differences were calculated using a two-tailed t-test. The statistical analyses were done using R v. 3.2.1.
Results
Metabolite screening identifies three endophytes with biocontrol potential
To identify fungal endophytes that, on the basis of their secreted metabolites, could be used as biocontrol agents against Phy. infestans, we screened culture extracts of 12 fungal endophytes for growth inhibition of Phy. infestans isolate D2 using an agar diffusion assay. Inhibition of Phy. infestans varied considerably, depending both on the endophyte isolate and on the culture medium. The average growth inhibition was 12.4 ± 8.7mm ranging from 0 and 35mm from the point of extract application (Tab. S3). Culture extracts of three of the 12 isolates inhibited growth of Phy. infestans with a radius ≥20mm. These three fungal endophytes (isolates 8082, 9907 and 9913) with the greatest Phy. infestans growth inhibition were chosen for further studies. An additional fungal strain, Phi. fortinii (isolate 4197) was included due to its mutualistic interaction with another host, Larix decidua (Schulz et al., 2002; Schulz, 2006), growth inhibition of other pathogenic microbes and prior information that it could colonize S. lycopersicum asymptomatically (Schulz, unpublished).
Phylogenetic placement of fungal endophytes
To determine the taxonomic identity and phylogenetic placement of the four selected fungal endophytes, we sequenced their ITS1, 5.8S and ITS2 regions. First, we used these sequences in a BLAST search to identify the closest relatives of the fungal endophytes (Tab. S4). All four endophytes belong to the ascomycetes. Our analyses further supported the previous characterization of isolate 4197 as Phi. fortinii (99% identity and e-value 0, Grünig et al., 2008). For isolate 8082 the best BLAST hit with 100% identity and an e-value of 0 was Phoma eupatorii. This was additionally supported by the fact that isolate 8082 was isolated from Eupatorium cannabinum (Tab. S1). The placement of isolates 4197 and 8082 in our phylogenetic analyses together with the extremely short branch lengths to their best BLAST hits further support these phylogenetic assignments (Fig. 1a,b). The best hit for isolate 9907 was Pyrenochaeta cava (95% identity and e-value 0) and for isolate 9913 it was Monosporascus ibericus (97% identity and e-value 0). This suggests that no completely identical sequence/taxa are currently represented in the database. Pyrenochaeta does not form a monophyletic group within the order of Pleosporales (Zhang et al., 2009; Aveskamp et al., 2010; Fig. 1c), thus based on the phylogenetic analyses isolate 9907 can only be placed within the order Pleosporales. Isolate 9913 was isolated from the roots of Aster tripolium, a plant that was growing in the salt marshes of the Mediterranean Sea (Tab. S1). Of note is that Monosporascus ibericus, the fungal endophyte clustering most closely with isolate 9913 in the phylogenetic analysis, has been recently described as an endophyte of plants growing in environments with high salinity (Collado et al., 2002). Furthermore, the genus Monosporascus is monophyletic; isolate 9913 has been placed within this monophyletic group and herewith termed Monosporascus sp. (Fig. 1d).
Fungal endophytes show broad-spectrum inhibition of Phytophthora infestans growth
Our initial screening of the culture extracts identified endophytes with the potential to inhibit the growth of a single Phy. infestans isolate. We therefore wondered whether the inhibition could be effective against a wider range of isolates of Phy. infestans. To test this, we conducted a co-culture assay on RSA agar medium with the four fungal endophytes against nine European Phy. infestans isolates (Fig. 2). In the plate assay all four endophytes were capable of significantly restricting growth of Phy. infestans (Fig. 3). Pho. eupatorii and isolate 9907 showed a global inhibition of all Phy. infestans isolates tested (Fig. 3b,c). Phi. fortinii inhibited the growth of eight out of nine isolates and Monosporascus sp. inhibited the growth of seven of the nine isolates (Fig. 3a,d). Pho. eupatorii had the greatest average relative growth inhibition of Phy. infestans with 50.6 ± 2.2%, and Monosporascus sp. had the lowest with 11.9 ± 1.6% (Tab. 1).
To exclude a mere reduction based on growth limitations we i) measured the inhibition of the endophyte’s growth by Phy. infestans after the initial co-cultivation phase and ii) evaluated long-term co-cultures (i.e. eight weeks) to analyze the endophyte and pathogen growth progression. The growth of isolate 9907 was not inhibited by any of the Phy. infestans isolates (Fig. S1c). However, some isolates of Phy. infestans were able to inhibit the growth of the other three fungal endophytes (Fig. S1a,b,d). In all cases, the average relative inhibition of an endophyte by Phy. infestans was, however, less than the average relative inhibition of Phy. infestans by an endophyte (Tab. S5). For example, whereas the average relative growth inhibition of Phy. infestans by Pho. eupatorii was 50.6 ± 2.2%, the average relative inhibition of Pho. eupatorii by Phy. infestans was 4.7 ± 0.9%.
After eight weeks, the endophytes, (except for isolate 9907) visually overgrew the plates, including the regions colonized by Phy. infestans (Fig. 4). To substantiate this observation, we extracted DNA from some co-cultures with Phi. fortinii (12 co-cultures), Pho. eupatorii (18 co-cultures) and Monosporascus sp. (seven co-cultures) from both sides of the eight-week samples (Tab. S6). In total, we analyzed 37 co-cultures and their respective controls for the presence of endophyte and Phy. infestans. We used the marker genes COX and ITS. Because our ITS primers were specific for fungi, we primarily observed amplicons from the fungal endophyte ITS loci when both organisms were present. However, presence of Phy. Infestans could be determined by the presence of a COX amplicon. In general, we observed that the endophyte was present on both sides of the plates, whereas Phy. infestans was either not detected or only on the side of the plate on which it had been inoculated. Few exceptions occurred in which Phy. infestans was observed on the side of original inoculation of the fungal endophyte (2/37 cases). Hence, Phy. infestans was usually not able to colonize the side of the plate where the endophyte was growing, while the endophyte was always able to colonize Phy. infestans’ side of the plate. In addition, the endophytes showed a greater inhibition of Phy. infestans than Phy. infestans did on the endophytes. Therefore, resource limitation (due to the size of the plates) is unlikely to fully explain the unequal growth differential between Phy. infestans and the endophytes during co-cultivation. Instead, we hypothesize that factors actively secreted by the endophytes may also be involved in the growth inhibition of Phy. infestans.
Phoma eupatorii limits Phytophthora infestans infection success
We identified global, non-isolate-specific growth inhibition by all four endophytes in plate assays. To test whether the inhibitory potential of the endophytes holds true in planta, we inoculated the fungal endophytes in axenically grown S. lycopersicum cv. M82 seedlings. Our preliminary screening showed that Phi. fortinii and isolate 9907 were too virulent and killed the S. lycopersicum seedlings (Fig. S2a,b,d). In contrast, S. lycopersicum seedlings inoculated with Pho. eupatorii and Monosporascus sp. survived (Fig. S2a,c,e).
To confirm the endophytic colonization of the roots, we analyzed fungal outgrowth of surface sterilized roots and their imprints from inoculations with water, endophyte or endophyte and Phy. infestans (Tab. 2). Irrespective of the protocol, there was neither fungal growth from the surface sterilized mock control roots nor from their imprints. Generally, imprints of the surface sterilized endophyte inoculated roots did not show fungal growth, except for Pho. eupatorii inoculated roots after sterilization treatment 1 (1/16 imprints from the mono-inoculation and 5/12 imprints from the co-inoculations). This suggests that surface sterilization was successful in all other cases. Pho. eupatorii grew from several roots independently of the sterilization treatment, although the stronger treatments resulted in less outgrowth. Hence, these treatments may partially impact survival of endophytic mycelium. Nevertheless, these results show that Pho. eupatrorii is capable of colonizing S. lycopersicum roots. Monosporascus sp. also showed outgrowth from several of the plated surface sterilized roots, suggesting that, like Pho. eupatorii, Monosporascus sp. also grows endophytically in the roots of S. lycopersicum. S. lycopersicum seedlings colonized by Pho. eupatorii are visually smaller than mock control seedlings and seedlings mono-inoculated with Phy. infestans. We also observed a reduction in leaflet number (Fig. S3a,c). Since the leaflets appeared sturdier and were darker green than the controls (Fig. 5a-f), we measured chlorophyll levels via chlorophyll fluorescence. However, chlorophyll abundance did not change following any of the treatments (Fig. 5g-m). We also observed that some of the stems of the plants that had been inoculated with Pho. eupatorii developed a purple color (Fig. S4c). Therefore, we reasoned that the darker leaflet color may have resulted from anthocyanin accumulation. In fact, we detected a significant increase in anthocyanin content in Pho. eupatorii inoculated versus mock control plants (p=0.001 without Phy. infestans, p=0.04 with Phy. infestans, Fig. 5o). In contrast to seedlings colonized by Pho. eupatorii, those inoculated with Monosporascus sp. did not visibly differ from the mock controls (Fig. S3a,b,S4a,c). In agreement with this, anthocyanin content did not increase significantly in Monosporascus sp. inoculated samples compared to the mock control (p=0.08 without Phy. infestans), but slightly when both endophyte and pathogen were present (p=0.007), suggesting that the increase results from the presence of Phy. infestans.
Despite the visible effects of the colonization by Pho. eupatorii on the seedlings, we proceeded to investigate the effect of the endophyte on a subsequent infection with Phy. infestans. The relative necrotic area caused by the pathogen is significantly higher on plants inoculated only with Phy. infestans (in the absence of pre-inoculation by an endophyte) compared to the mock control (Fig. 5n,S4e). To confirm the pathogen infection in the mock/Phy. infestans samples, we used the expression of the Phy. infestans biomass marker genes PiH2a and PiElf1α. In agreement with the increase in necrotic area, Phy. infestans was present in all biological replicates mono-inoculated with the pathogen, i.e. demonstrating a successful infection.
While the relative necrotic area in seedlings that were colonized only by Pho. eupatorii was 4.7-fold higher compared to the mock control, this was significantly less than the relative necrotic area of seedlings infected with only Phy. infestans (Fig. 5n). S. lycopersicum seedlings co-inoculated with Pho. eupatorii and Phy. infestans resulted in a significantly reduced relative necrotic area compared to seedlings mono-inoculated with Phy. infestans (Fig. 5n). Importantly, the average relative necrotic area of leaflets colonized by both Pho. eupatorii and Phy. infestans did not differ from the mono-inoculations with the endophyte (Fig. 5n). Whether 5μl or 10μl mycelial suspensions of Pho. eupatorii was used had no effect on the outcome of the experiments. The relative necrotic area between the treatment with Monosporascus sp. and the mock control did not differ (Fig. S4a,c,e). This endophyte was not able to inhibit Phy. infestans infection nor limit development of disease symptoms in planta (Fig. S4b,d,e,f).
To quantify the biomass of Phy. infestans in planta after pre-inoculation with Pho. eupatorii, we performed a qRT-PCR with the two biomass marker genes PiElf1α and PiH2A (Fig. 5o). In total, we tested the three biological replicates from the 5μl Pho. eupatorii inoculations and two from the 10μl Pho. eupatorii inoculations. In three of those five replicates, we did not detect an amplicon for either PiH2a or PiElf1α. Yet, PiH2a and PiElf1α were detected in every biological replicate of the mock/Phy. infestans infections. In addition, three plant-specific reference genes were tested; these showed no aberrant expression in any of the samples colonized by the endophyte in which PiH2a and PiElf1α were not detected. Hence the presence of the fungal endophyte did not affect the efficiency of the qRT-PCR. Also, those samples that were pre-inoculated with Pho. eupatorii, but gave an amplicon of the marker genes had reduced Cq-values for both marker genes compared to the mock/Phy. infestans samples. This suggests that Pho. eupatorii reduced the infection with Phy. infestans isolate D12-2 in the sampled leaflets. To estimate the reduction of Phy. infestans biomass, we assumed that the Cq-value of those replicates with no amplicon could theoretically have been amplified in later cycles. We therefore set the Cq-values in those samples to 41; i.e. one cycle more than the original runs included. Based on this assumption, we observed a significant reduction of gene expression in both biomass marker genes in the Pho. eupatorii pre-treated samples compared to mono-infections of Phy. infestans (Fig. 5o). Therefore, Pho. eupatorii is capable of significantly inhibiting Phy. infestans infection of S. lycopersicum leaflets.
Discussion
Fungal endophytes show a broad-spectrum growth inhibition of European Phy. Infestans isolates
Of 12 fungi for which culture extracts were tested for inhibition of Phy. infestans, we identified three ascomycetes, Pho. eupatorii, isolate 9907 and Monosporascus sp., which effectively inhibited growth of the pathogen. While fungal endophytes produce a vast diversity of metabolites (Schulz et al., 2002; Strobel & Strobel, 2007; Verma et al., 2009; Mousa & Raizada, 2013; Brader et al., 2014) and numerous have antimicrobial activity (Son et al., 2008; Puopolo et al., 2014; Mousa et al., 2016), endophytes and their metabolites may have a narrow spectrum of specificity. To avoid narrow spectrum of pathogen inhibition, we screened these three fungal endophytes and the endophyte Phi. fortinii for their capacity to inhibit the growth of nine European isolates of Phy. infestans. In our co-culture assays, Pho. eupatorii and isolate 9907 had a broad-spectrum inhibition against all tested isolates, whereas Monosporascus sp. and Phi. fortinii inhibited most of the isolates. Additionally, after eight-weeks of incubation, the pathogen was not able to grow on sections of the plates, in which the endophytes grew. The consistency of the results from the culture extract experiments and the plate assays of Pho. eupatorii and isolate 9907 shows that their inhibition is independent of the growth medium, suggesting an environmentally robust metabolite production of their anti-Phytophthora substances. A robust metabolite production would be of great advantage, if these fungal endophytes are to be used as living biocontrol agents in the field.
For application in the field, two issues must be examined: i) Does infection by the endophyte damage the host in the absence of a pathogen? and ii) Does the endophyte successfully inhibit the pathogen in the host? In our study, the former is of extreme importance, because the fungal endophytes in question were not originally isolated from the Solanaceae, to which tomato belongs. Furthermore, whether an endophyte remains benign and asymptomatic is likely to be affected by a number of different circumstances and in some cases the host endophyte relationship may shift to a pathogenic outcome from an initially protective interaction (Schulz & Boyle, 2005; Junker et al., 2012; Schulz et al., 2015; Busby et al., 2016b). Along these lines we excluded two isolates, Phi. fortinii and isolate 9907, for direct applications as biocontrol agents: Seedlings of S. lycopersicum infected with either of these two isolates quickly died after inoculation. A third isolate, Monosporascus sp., neither inhibited Phy. infestans infection nor hindered its infection progress. This may not be surprising, because Monosporascus sp. had the lowest inhibition potential in our co-culture assays. It should, however, be noted that the metabolite composition of fungal endophytes varies depending on their environments, i.e. in vitro and in planta (Brader et al., 2014). It is therefore possible that the metabolite composition Monosporascus sp. produces in planta does not include the active anti-Phytophthora compound. Alternatively, the active compound may be only produced in specific stages of the infection. In the latter scenario, the infection of Monosporascus sp. may not have progressed far enough by the time we inoculated with Phy. infestans. Nevertheless, the outcome of the in planta co-inoculations does not exclude the possibility that the in vitro produced metabolites could be effective in field applications, especially since they showed a broad-spectrum reduction in Phy. infestans growth in our co-culture experiments. The broad-spectrum effectiveness of inhibition suggests that the metabolite composition either includes a metabolite with a conserved target in Phy. infestans or a mixture of anti-Phytophthora metabolites. Both would slow the counter-adaptation of the pathogen to the metabolites if used in field application. As a next step, the metabolite extracts with protective capabilities should be tested for their cytotoxicity in planta.
Phoma eupatorii isolate 8082 may inhibit Phytophthora infestans via secreted toxic metabolite(s) and/or the induction of host defense mechanisms
Pho. eupatorii was the most effective fungal endophyte in our experiments, excelling both in co-culture as well as in planta. The presence of Pho. eupatorii not only reduced or inhibited the pathogen’s growth, but perhaps entirely prevented infection. Here we used root inoculations of Pho. eupatorii combined with leaflet inoculations of Phy. infestans isolate D12-2. Because Pho. eupatorii was applied to roots, while Phy. infestans was inoculated on the leaves, niche competition is an unlikely mechanism by which Pho. eupatorii protects the S. lycopersicum seedlings. Therefore, two other possible mechanisms by which the plant is defended against the pathogen include endophyte-dependent induction of defense responses or the production of anti-Phytophthora metabolites. The induction of plant defense responses by endophytes, such as Pir. indica and non-pathogenic Fusarium oxysporum, has been previously shown (Stein et al., 2008; Aimé et al., 2013). Here, we observed an elevation of anthocyanin levels in leaf tissue of S. lycopersicum after root colonization of Pho. eupatorii. Accumulation of anthocyanins is, among other factors, positively regulated by jasmonic acid (Franceschi & Grimes, 1991; Feys et al., 1994; Shan et al., 2009; Li et al., 2006). Hence, it is possible that jasmonic acid dependent defense responses are induced upon colonization with Pho. eupatorii. This may be a response to Pho. eupatorii and elevated levels of jasmonic acid may have contributed to the inhibition of the Phy. infestans infection we observed. The role of jasmonic acid in defense against Phy. infestans is not clear: In one study, application of jasmonic acid to leaves of tomato and potato plants resulted in reduced infection of the pathogen (Cohen et al., 1993). In another study, it is reported that jasmonic acid is required for the initiation of defense responses triggered by a peptide secreted by Phy. infestans (Halim et al., 2009). Yet, potato RNA interference lines that downregulated jasmonic acid biosynthesis and signaling components, showed no alterations in the infection rates of Phy. infestans (Halim et al., 2009). Hence, the production of anti-Phytophthora metabolites may be a more likely explanation for the observed reduction of Phy. infestans infection. A recently published example of a metabolite based endophyte-mediated pathogen protection is that of Enterobacter sp. This endophyte produces many different antimicrobial compounds in its host plant and these are detrimental to the host plant’s pathogen F. graminearum (Mousa et al., 2016). In our study, each of the four fungal endophytes undoubtedly produces anti-Phytophthora metabolites in the crude extract tests and in the co-cultures on agar media. This makes it likely that Pho. eupatorii also produces such metabolites during in planta co-inoculations with Phy. infestans. A combination of these two mechanisms is, however, also possible.
Conclusion: Phoma eupatorii isolate 8082 is a potential novel Phytophthora infestans biocontrol agent
Out of a screen of 12 fungal endophytes, we discovered four ascomycetes that inhibited the growth of Phy. infestans in co-culture, presumably through the secretion of secondary metabolites, particularly since their culture extracts were also active. Most importantly, two of the endophytes exhibited global inhibition of nine European Phy. infestans isolates, the other two showing a near-global inhibition. This indicates that a conserved target within Phy. infestans for a particular metabolite may be produced by these four endophytes. Alternatively, a complex metabolite mixture could be involved. In either case, the use of these fungi for biocontrol could slow the counter-adaptation of Phy. infestans. Hence, all four fungal endophytes can be considered good candidates for the production of such new and urgently needed compounds. Additionally, of the four fungal endophytes, Pho. eupatorii functioned as an effective biocontrol agent in planta. Therefore, Pho. eupatorii may not only synthesize a reservoir of highly useful antimicrobial metabolites, but could serve as a novel biocontrol agent providing an alternative to resistance gene breeding and application of agrochemicals.
Author contribution
SdV, BS and LER wrote the manuscript. SdV, JKvD, AS and SG performed the experimental work and data analyses. BS provided the fungal isolates and the metabolite screening. All authors read and approved the manuscript.
Acknowledgements
We thank Tuba Altinmakas, Melissa Mantz and Klaudia Maas-Kantel for technical support, Siegfried Draeger for isolation of the endophytes and initial identification based on morphology. We thank the TGRC Institute for providing the seeds of S. lycopersicum cv. M82 and Francine Govers (Wageningen University) for the Phy. infestans isolates NL10001, NL88069, NL90128, IPO-C, IPO428-2, 3928A, D12-2, T15-2 and T20-2. We thank Dr. Bärbel Schöber-Butin for providing the German isolate Phy. infestans D2. This work was supported by the Deutsche Forschungsgemeinschaft (Ro 2491/5-2, Ro 2491/6-1, Research Training Group GRK1525).