1. Introduction
Pathogenic microorganisms cause crop losses equivalent to billions of dollars annually worldwide. There has been an alarming trend of increasing numbers of new fungal and oomycete plant pathogens [
1], with fungal pathogens accounting for nearly 15% of worldwide crop losses each year [
2]. The plant pathogens
Rhizoctonia solani,
Fusarium culmorum,
Gaeumannomyces graminis var.
tritici (Ggt), and the oomycete,
Pythium ultimum, are widespread pathogens of economically important crops including wheat, rice and barley, grain, legumes, and brassicas worldwide, causing root rot, take-all disease, and damping off [
3,
4]. Resistant plant varieties take decades to develop and pathogens often develop resistance within several years [
5]. Commercial growers rely on a variety of chemicals to control the plant pathogens. However, widespread use of these chemicals comes at high economic and environmental costs such as potentially harming beneficial microorganisms and insects. Additionally, organic farmers are not permitted to use synthetic chemicals. An alternative method is biocontrol that involves the use of non-pathogenic microbial strains to combat the pathogens [
5]. Specific strains of bacteria and fungi are known to suppress growth of soilborne fungal pathogens of crop plants under controlled conditions [
6,
7], but commercialization for the field involves additional criteria [
8]. These include targeted (pathogen-specific) activity, durability of activity in a range of soil types and environmental conditions, and application method [
9,
10]. As such, bio-based products have developed more slowly than synthetic products but represent a growing proportion of the pesticide industry. For soil-borne pathogens, seed treatment is considered to be the application method of choice. Selection of biocontrol microbial isolates or their metabolites for commercialization purposes should ideally be based on multiple pathogen isolates of the same species because genotypes of the plant, pathogen and biocontrol microbe appear to condition the biocontrol interaction [
11,
12]. Understanding the mode of action of biocontrol metabolites and their impact on the rhizosphere microbiome also will help in the development of a successful bio-based product. The limited number of biocontrol strains in use, however, is largely from the rhizosphere or soil, and as such, could easily be out-competed by other soil microorganisms [
13].
Endophytes, the non-pathogenic microorganisms within plants, represent a largely untapped portion of the plant microbiome. In a screen of endophytes isolated from plants growing on the west side of the Washington State Cascade Mountain Range, 21 percent of cultured isolates from poplar (
Populus) and willow (
Salix) had in vitro antimicrobial activity against at least 1 pathogen out of 4 tested [
14]. This high frequency of anti-microbial activity likely reflects the strong selective pressure in natural systems for resistance to a broad array of pathogens. Because of this pressure towards mutualism with their host, endophytes have developed a number of other traits that increase plant health, growth, yield, and stress tolerance [
15,
16,
17,
18]. Bacterial endophytes can systemically colonize plants [
19], giving them a key advantage over rhizospheric strains and some fungal endophyte strains which can be highly localized. Endophytes may therefore offer a more stable option for biocontrol compared to existing methods, defending their host from pathogens from within.
Previous studies on potential biocontrol strains from poplar were based on endophytes that had been isolated for other purposes such as nitrogen fixation [
14]. For example, several of the
Burkholderia endophytes exhibited particularly robust inhibition in vitro against several plant pathogens despite the original selection method being growth on nitrogen-limited medium. A particularly robust antimicrobial strain, WPB, described herein, is a strain of
B. vietamiensis, a member of the
Burkholderia cepacia complex (BCC), a group found in a variety of ecological niches including soil, plants and animals but has mambers which present health hazards [
20].
With the hypothesis that direct screening for antifungal activity of poplar endophytes would allow for the isolation of a wider variety of potential biocontrol strains, a novel method for screening plant extracts for antifungal endophytes was developed, without any pre-screening or isolation steps required. Using this method, we isolated a variety of new endophyte strains from wild poplar with in vitro antifungal activities and report the genomic analysis on the potential inhibitory mechanisms. Genomic analysis of B. vietnamiensis strain WPB indicated that occidiofungin, a known antimicrobial compound, could be key to its effectiveness. Using targeted mutagenesis and phenotypical and biochemical confirmation, it was shown that WPB inhibits pathogens through multiple modes of action. Here, we also report on the genomic analysis of additional endophytes isolated from poplar that inhibited selected pathogen strains in vitro. The goal of this study was to isolate strains that employ different mechanisms of antimicrobial activities, reducing the risk of pathogens developing resistance.
3. Discussion
The goal of this study was to isolate endophyte strains that could inhibit key pathogen strains of concern in the Pacific Northwest. To select for such activities, it is necessary to conduct the initial testing in vitro. Once candidate strains are identified, the next phase is to complete genomic sequencing and analysis for genes known to be involved in human pathogenicity since the ultimate goal is to use the inhibitory strains in agriculture. A case in point is that, in a previous study, the endophytes showing the strongest in vitro antifungal activity were of the genus
Burkholderia [
14].
Burkholderia have been studied for their abilities in plant growth promotion, phytoremediation, endophytic nitrogen fixation, reducing plant abiotic stresses and as biocontrol agents [
47,
48,
49,
50,
51], while others are known as human pathogens [
20]. Since the original
Burkholderia genus includes such diverse species, molecular signature and phylogenomic analysis were performed that divided them into
Burkholderia and
Paraburkholderia genera [
52]. A number of root-associated
Burkholderia strains also lacked genes associated with human pathogenicity [
53]. Subsequently, a new genus,
Caballeronia gen. nov.was proposed that contains non-pathogenic and plant-associated members of both
Burkholderia and
Paraburkholderia [
54]. The previously isolated strain of
Burkholderia vietnamiensis, WPB, was included in this study of potential mechanisms of antimicrobial activities, although it would not be used in agriculture since this strain falls within the
Burkholderia BCC and was determined to harbor genes associated with human pathogenicity. Nonetheless, it served to provide valuable insight into inhibitory mechanisms.
Interestingly, when studying poplar leaves growing along the Snoqualmie River in Washington State, nearly one-fourth of the bacterial microbiome was found to be of the
Burkholderia/Paraburkholderia genus (unpublished data). In a recent study, the
Burkholderiaceae, along with
Pseudomonadaceae and
Enterobacteriaceae, was one of the most abundant families detected in the core microbiome of wild poplar inhabiting xeric riparian zones, and riparian zones with mid hot-dry and moist (mesic) climates [
55]. Our study further supports the concept that antimicrobial/antifungal producing strains are a core component of the poplar microbiome, and not linked to a specific environment. The high prevalence of
Burkholderia points to the importance of this genus in the plant microbiome.
In an effort to isolate a broader biodiversity of endophytes, a different technique was employed in this study. Rather than testing individual isolates for antifungal activity, poplar extracts containing endophytic microorganisms were challenged directly against a fungal pathogen. Microbes were then isolated from zones of inhibition. Though all the strains were originally selected for the ability to inhibit
R. solani AG-8, many of them also inhibited other fungal pathogen strains regardless of evolutionary distance. Due to the vast evolutionary distance from the fungi to oomycetes, it was not surprising that few of the strains had activity against
P. ultimum. However, this method of challenging extracts rather than single isolates could be applied to any pathogen strains of interest to rapidly select potential biocontrol strains. Though strains AFE 3, 14, and 22 were originally isolated with apparent inhibitory activity against
R. solani AG-8, when fully purified, they were no longer active against this pathogen. It is possible that they required a microbial partner or phytochemicals mimicking host-like conditions that were lost during the strain purification phase. For instance, the expression of T6SS, which can be used to deliver antifungal effectors, is silent unless host-like conditions are mimicked [
56].
One goal of this study was to isolate strains that each employed different mechanisms of inhibiting plant pathogens. Application of a consortium of strains with complementary antimicrobial activities would be ideal in biocontrol situations, allowing for possible synergies between modes of action and reducing the risk of pathogens developing resistance. In this respect, it has been demonstrated that a combination of three biocontrol species was superior for inhibiting
Fusarium [
57]. Some of the antimicrobials produced by the
Pseudomonas AFE strain have similar modes of action. For instance, the most likely action of the cyclic lipodepsipeptides (CLPs) of the Viscosin-groups involves the direct interaction with the cellular membrane through pore-formation [
58]. A similar activity was observed for the lipopeptides of the Tolaasin-group which have the ability to form ion channels into lipid bilayer [
59]. Despite this, prior to the development of a microbial consortium with enhanced antifungal potential, the positive interaction between members of the consortium must be assessed. Indeed,
Bacillus and
Pseudomonas could compete for the same niche, where
Bacillus lipopeptides are used to counteract and reduce the toxicity of the lipopeptides synthesized by
Pseudomonas [
60].
In this study, AFE4A was the best performing strain in terms of breadth and strength of in vitro antifungal activities. AFE4A typically had wide zones of inhibition implying an antifungal mechanism based on the secretion of multiple secondary metabolites with antifungal activity. This is in partial agreement with the antiSMASH analysis which identifies AFE4A, and also AFE21B, as the AFE strains with the greatest antifungal potential. AFE4A and AFE21B were isolated from wild poplars inhabiting two different river systems. Despite this, both strains carried the same BGC’s with antifungal activities and yet had strikingly different inhibition patterns. This would provide an opportunity to study the molecular mechanisms underlying the biosynthesis of the antifungals rather than focusing on comparative genome analysis. The presence of a single lanthipeptide antibiotic gene cluster was found to distinguish AFE4A from AFE21B. Since AFE4A had the most widespread activity against all pathogens tested and the most effective antifungal activity, this defining lanthipeptide (Lantibiotics) gene cluster merits further investigation. However, all the Lantibiotics characterized so far in
Bacillus species possess antimicrobial activity, while Pinesin A and B, which are produced by the Gram-negative
Chitinophaga pinensis, are the only known Lantibiotics with antifungal activity [
61,
62]. Therefore, the differences observed between AFE4A and AFE21B could be explained only by different antifungal mechanisms or by differing levels of expression of the BGC’s.
The genome of the yeast strain AFE11,
Aureobasidium pullulans, was not sequenced. However,
A. pullulans is known to produce a variety of antimicrobial compounds including toxins, volatile compounds, degradative enzymes, and siderophores [
42,
63]. Described as being a generalist biocontrol strain, it may be the simultaneous production of such a diverse arsenal that leads to its success.
The genome of WPB was found to contain the biosynthetic genes required for the production of occidiofungin, a known antifungal compound. Mutagenesis of the biosynthetic gene
ofcE in WPB highlighted the importance of this antimicrobial to the powerful inhibition exhibited by this strain. The results also indicated that other inhibitory activities are at play, although the mechanism still needs to be resolved as no additional BCGs with known antifungal activities were found in WPB. As a genus,
Burkholderia possess large, complex, and variable genomes and ongoing efforts to identify novel antimicrobial compounds will continue to benefit from rapidly growing libraries of sequenced genomes, including the sequencing and assembly of the genomes from 450 members of
Burkholderiaceae completed by Mullins et al. (2020) [
64]. The
ofc mutant retained strong inhibition of Ggt ARS-A1 despite the lack of occidiofungin production. Genomic analyses and confirmation through LC-MS has shown that the AFE strains also lack the capability to produce occidiofungin. This implies antifungal mechanisms in all strains tested that are distinct from occidiofungin.
It may be that endophytes are uniquely positioned to evolve antifungal mechanisms that are targeted specifically against pathogens. Populus associates with both ectomycorrhizal and arbuscular mycorrhizal fungi; therefore, forcing co-evolution of the bacterial endophytes with these essential fungal partners. Bacterial endophytes must therefore be capable of co-habitation with these beneficial fungi despite the anti-fungal activities demonstrated here against pathogens. Living with a host plant would apply a heavy evolutionary pressure to develop mechanisms to differentiate between pathogens and beneficial endophytic fungi. The unique niche occupied by endophytes has encouraged the development of several other beneficial traits including nitrogen fixation, phosphate solubilization, and phytohormone production. Though the strains isolated in this study were selected for the ability to inhibit pathogens, several had these beneficial traits as well. Endophytes are a rich resource with great potential to improve the resilience of crops to environmental challenges, both abiotic and biotic.