Bacterial endophytes from wild maize suppress Fusarium graminearum in modern maize and inhibit mycotoxin accumulation

Abstract

Wild maize (teosinte) has been reported to be less susceptible to pests than their modern maize (corn) relatives. Endophytes, defined as microbes that inhabit plants without causing disease, are known for their ability to antagonize plant pests and pathogens. We hypothesized that the wild relatives of modern maize may host endophytes that combat pathogens. Fusarium graminearum is the fungus that causes Gibberella Ear Rot (GER) in modern maize and produces the mycotoxin, deoxynivalenol (DON). In this study, 215 bacterial endophytes, previously isolated from diverse maize genotypes including wild teosintes, traditional landraces and modern varieties, were tested for their ability to antagonize F. graminearum in vitro. Candidate endophytes were then tested for their ability to suppress GER in modern maize in independent greenhouse trials. The results revealed that three candidate endophytes derived from wild teosintes were most potent in suppressing F. graminearum in vitro and GER in a modern maize hybrid. These wild teosinte endophytes could suppress a broad spectrum of fungal pathogens of modern crops in vitro. The teosinte endophytes also suppressed DON mycotoxin during storage to below acceptable safety threshold levels. A fourth, less robust anti-fungal strain was isolated from a modern maize hybrid. Three of the anti-fungal endophytes were predicted to be Paenibacillus polymyxa, along with one strain of Citrobacter. Microscopy studies suggested a fungicidal mode of action by all four strains. Molecular and biochemical studies showed that the P. polymyxa strains produced the previously characterized anti-Fusarium compound, fusaricidin. Our results suggest that the wild relatives of modern crops may serve as a valuable reservoir for endophytes in the ongoing fight against serious threats to modern agriculture. We discuss the possible impact of crop evolution and domestication on endophytes in the context of plant defense.

Keywords: Fusarium graminearum; Gibberella ear rot; Paenibacillus; Zea diploperennis; deoxynivalenol; endophyte; maize; parviglumis.

Figure 1

Origin of endophytes used in this study and results of the in vitro anti-Fusarium screen. (A) A map showing the origin of maize genotypes previously used to isolate the endophytic library. (B) Examples of endophytes from the library isolated from different Zea genotypes as indicated. (C) Example of an endophyte culture showing suppression of F. graminearum hyphae (white) using the dual culture method. (D) Quantification of the inhibitory effect of the endophytes or fungicide controls, amphotericin B and nystatin (at concentrations of 5 and 10 μg/ml, respectively), on the growth of F. graminearum in vitro. For these experiments, n = 3. The error bars indicate the standard error of the mean. The black asterisk indicates that the treatment means are significantly different from the fungicide Nystatin at p ≤ 0.05. The green asterisk indicates that the treatment means are significantly different from the fungicide Amphotericin at p ≤ 0.05.

Figure 2

Taxonomic characterization of candidate anti-Fusarium endophytes. (A) Details of the taxonomic identification of the anti-Fusarium endophytes using 16S rDNA and 23S rDNA, and the tissue and host from which the endophytes were originally isolated. (B) 16S rDNA based phylogenetic tree of the three predicted Paenibacillus sp.

Figure 3

Electron microscope images of the anti-fungal endophyte strains. (A–D) correspond to strains 1D6, 3H9, 4G12, and 4G4, respectively.

Figure 4

Microscopic in vitro interactions between each anti-fungal endophyte and F. graminearum. (A) Cartoon of the experimental methodology to examine microscopic in vitro interactions between F. graminearum (pink) and each endophyte (orange) or the buffer control (LB medium). The microscope slides were pre-coated with PDA and incubated for 24 h. F. graminearum hyphae were then stained with neutral red. Shown are representative microscope slide pictures (n = 3) of the interactions between F. graminearum and: (B) Strain 1D6 compared to (C) the buffer control; (D) Strain 3H9 compared to (E) the buffer control; (F) Strain 4G12 compared to (G) the buffer control; (H) Strain 4G4 compared to (I) the buffer control.

Figure 5

The effects of the candidate endophytes on F. graminearum in vitro using the vitality stain, Evans blue. Shown are representative microscope slide pictures (n = 3) of the interactions of F. graminearum with: (A) the commercial biological control agent, Bacillus subtilis (100 mg/10 ml) compared to (B) the buffer control; (C) Strain 1D6 compared to (D) the buffer control; (E) strain 3H9 compared to (F) the buffer control; (G) Strain 4G12 compared to (H) the buffer control; (I) Strain 4G4 compared to (J) the buffer control.

Figure 6

Molecular and biochemical detection of the candidate anti-fungal compound, fusaricidin, in Paenibacillus strains. (A) Details of fusA gene orthologs isolated from the candidate Paenibacillus endophytes by PCR amplification. (B–D) Combined ion chromatogram/mass spectrum for candidate fusaricidin derivatives detected in the cultures of the Penibacillus endophytes as indicated.

Figure 7

Greenhouse trial 1 to test for the ability of the candidate endophytes to suppress Gibberella Ear Rot (GER) in a modern hybrid. (A, B) GFP-tagged endophyte strain 4G12 visualized inside maize roots, in the (A) absence or (B) presence of propidium iodide that outlines the cell with red color. (C–H) Representative ears from each treatment. (I) Picture of an ear to illustrate the methodology of scoring disease severity: The fungal pathogen was introduced to the tip of the ear, indicated by the asterisk. Therefore, the disease was scored as the ratio of the length of the diseased ear tip portion relative to total ear length, multiplied by 100 to give a percentage. (J, K) Quantification of the effect of different treatments on GER suppression, as: (J) percent ear infection, and (K) average grain yield per plant. For both measurements, n = 20 per treatment (n = 10 for both controls). The whiskers indicate the range of data points. The black asterisk indicates that the treatment means were significantly different from the Fusarium only treatment at p ≤ 0.05. The green asterisk indicates that the treatment means were significantly different from prothioconazole fungicide (Proline) treatment at p ≤ 0.05.

Figure 8

Greenhouse trial 2 to test for the ability of the candidate endophytes to suppress Gibberella Ear Rot in a modern hybrid. (A–F) Representative ears from each treatment. (G) Picture of an ear to illustrate the methodology of scoring disease severity: the fungal pathogen was introduced to the tip of the ear, indicated by the asterisk. Therefore, the disease was scored as the ratio of the length of the diseased ear tip portion relative to total ear length, multiplied by 100 to give a percentage. (H, I) Quantification of the effect of different treatments on GER suppression, as (H) percent ear infection, and (I) average grain yield per plant. For both measurements, n = 20 per treatment (n = 10 for both controls). The whiskers indicate the range of data points. The black asterisk indicates that the treatment means were significantly different from the Fusarium only treatment at p ≤ 0.05. The green asterisk indicates that the treatment means were significantly different from prothioconazole fungicide (Proline) treatment at p ≤ 0.05.

Figure 9

Test for the ability of the candidate endophytes to reduce DON mycotoxin accumulation in maize grain during storage. DON measurements after storage of maize grain from: (A) greenhouse trial 1 (summer 2012), and (B) greenhouse trial 2 (summer 2013). For both trials, n = 3 pools of seeds. The black asterisk indicates that the treatment means were significantly different from the Fusarium only treatment at p ≤ 0.05. The green asterisk indicates that the treatment means were significantly different from the prothioconazole fungicide (Proline) treatment at p ≤ 0.05.