Endosymbionts in cranberry: Diversity, effect on plant growth, and pathogen biocontrol

Abstract

Virtually, all vascular plants harbor bacterial and fungal endosymbionts, which colonize predominantly but not exclusively, roots. Most common among fungal partners are arbuscular mycorrhizal fungi (AMF; Glomeromycotina) that live within plant roots and provide soil nutrients to the plant while receiving from the plant organic carbon sources. Ericaceae are an exception as they do not host AMF but rather the taxonomically ill‐defined “ericoid mycorrhizal fungi,” including select ascomycete and basidiomycete species. Because the diversity of endosymbionts in Ericaceae is poorly investigated, we set out to explore the microbiome of Vaccinium macrocarpon Aiton (cranberry).Here, we report the isolation and ribotyping of ~180 distinct bacterial and fungal endophytes collected from roots, stems, and leaves of cranberry plants cultivated in Quebec, Canada. Plant growth promotion was assessed after inoculating plant cuttings with these microbes, whereas pathogen suppression was tested on agar plates by growth confrontation.We show that the cranberry microbiome varies substantially between tissues, cultivars, and across fields of the same farm. Among the isolates, 16 bacterial and 8 fungal strains exhibit biofertilization or biocontrol properties with potential application in sustainable cranberry farming.We propose to move towards a more rigorous, molecular‐based definition of ericoid mycorrhiza, accounting for their broad evolutionary and morphological diversity.

INTRODUCTION

Symbioses, defined as the “living together of unlike organisms,” have been recognized since the 19th century (De Bary, 1879). Plant symbioses are commonly categorized into either epiphytic, endophytic, or mycorrhizal (only employed for fungal microbes) or else beneficial, mutualistic, neutral, or pathogenic (Hardoim et al., 2015; Khare et al., 2018; Weiß et al., 2016). Most studies of plant symbioses focus on just one particular category (Carroll, 1988; Petrini, 1991; Porras-Alfaro & Bayman, 2011), a reductionist approach to analyzing biological functions one at a time. Yet, this approach tends to fall short of capturing complex symbiotic interactions in natural habitats. For example, fungal hyphae emanating from plant roots associate with a wide range of “helper” bacteria in the rhizosphere (e.g., Nanjundappa et al., 2019), engaging in higher-level community interactions (Emmett et al., 2021). Helper bacteria include a range of Bacillus species and other bacterial taxa that promote plant growth through phosphate solubilization, phytohormone production, and protection against fungal pathogens (Frey-Klett et al., 2007).

In addition, symbiotic interactions between microbes and plants are characterized by plasticity with respect to morphology and the nature of the interaction. Researchers tend to classify their isolates as ecto- or endo-symbionts, based on the location from which the microbes were isolated: either rhizosphere or plant tissue. Yet, fungal symbionts are frequently much more versatile, satisfying both definitions, for instance when extraradical hyphae of fungal endophytes emanate from the plant roots into the rhizosphere. Some researchers have been quick in creating a new term based on this observation, the “ectendosymbionts,” for example Roth and Paszkowski (2017), Vohník and Albrechtová (2011), and Yu et al. (2001), adding to the panoply of often contradictory definitions in the classification of microbial plant symbionts. Definitions of the terms “mycorrhiza” (ecto, endo, arbuscular, ericoid, etc.) and “endophytes” are particularly numerous and confusing. For example, the notion “endophyte” is sometimes employed for fungal microbes colonizing plant tissues other than roots. In contrast, other studies use this term for fungal microbes that do not form recognizable “mycorrhizal” structures within the plant root cells, for example Perotto et al. (2018). Given this blur of contradictory definitions often only based on a few morphological traits, we instead refer to our isolates as “ericoid endophytes.”

Among the best-studied microbial symbionts of plants are arbuscular mycorrhizal fungi (AMF). While most vascular plant species engage in symbioses with these fungal endosymbionts, a few groups are apparently unable to do so, most notably members of the Ericaceae family. Instead, they associate with a group referred to as the ericoid mycorrhizal fungi (ErMF) (Suvanto et al., 2017; Watkinson et al., 2015), delineated morphologically by their dense hyphal coils they form inside rhizodermal or cortical cells of plant roots. In addition, a conspicuous hyphal network or “sheaths” may be present that extends into the rhizosphere (Vohník, 2020) and that is likely involved in the mobilization and capture of the scarce nutrients in acidic and/or sandy soil, the preferred habitat of Ericaceae (Clemmensen et al., 2015; Leake & Read, 1989).

Most members of the ericacean symbionts belong to the fungal order Helotiales, which, interestingly, emerged in approximately the same period as their hosts, some 66–72 Ma ago in the Late Cretaceous (Schwery et al., 2015), ~250 Ma later than AMF (Strullu-Derrien et al., 2018). Other members of this morphologically defined group belong to Agaricomycetes and to diverse Basidiomycete groups (Kolařík & Vohník, 2018, and references therein), suggesting either convergent evolution or substantial lateral gene transfer in creating the mycorrhizal colonization structures typical of ericoid mycorrhiza. In addition, contrary to AMF that belong to a monophyletic assembly within fungi, phylogenetic relationships within Helotiales and with its sister clades have been uncertain for many years (Hambleton & Sigler, 2005; Martino et al., 2018; Smith, 2009).

A recent phylogenetic inference based on genome sequences has established a more robust framework for studies of fungal endophytes (Johnston et al., 2019). On the basis of this tree, we were able to regroup clades defined in the current taxonomic classification scheme of NCBI and precisely place several of the ericoid symbionts together with other fungi (Figure 1). The amended Helotiales group comprises numerous previously described “true” plant root endosymbionts, in addition to lineages causing severe plant diseases (Sclerotiniaceae and Erysiphaceae), as well as a group of poorly investigated fungi designated dark septate endophytes (DSEs, including Phialocephala sp.) (Perotto et al., 2012; Yang et al., 2018), one of the most recent divergences within Helotiales. Again, contrary to AMF that form a consistent, monophyletic group of mycorrhizal fungi, the sparse distribution of the ericoid “mycorrhiza” trait within Helotiales suggests either convergent evolution, massive lateral gene transfer, and/or repeated loss of the mycorrhizal properties. A thorough examination of the symbiotic principles based on comparative molecular data is commendable, both for distinguishing among these scenarios and for a meaningful grouping/naming of ericoid fungal endophytes.

FIGURE 1. Schematic phylogenetic tree of Helotiales fungi. According to the NCBI taxonomic hierarchy, Helotiales are members of the taxonomic lineage Leotiomycetes, Sordariomyceta, Leotiomyceta, Pezizomycotina, Saccharomyceta, and Ascomycota. The tree includes selected Helotiales, with a topology taken from a recently published, fully resolved multi-protein phylogeny that used available genome sequence information (Johnston et al., 2019). Taxa are named according to the current NCBI taxonomy SCHEME, except for the informal “pezizelloids” that unite members of four distinct families. Species are designated as in previous publications (Johnston et al., 2019; Yang et al., 2018) and are color-coded as follows: red, reported to cause plant disease; mauve, morphologically recognized as mycorrhizal fungi; green, other plant-root endophytes; orange, no plant association reported. Species in the literature referred to as dark septate endophytes (DSE) are confined to Mollisiaceae. Species marked by three asterisks are closely related to the endophytes described in this work.

To fill the knowledge gap on both fungal and bacterial ericoid endosymbionts, Vaccinium macrocarpon Aiton (American cranberry) is a plant model of choice (Sauer et al., 2002; Tadych et al., 2012, 2015). This experimental system combines several advantages: (i) the perennial plant can be easily propagated by cuttings or seeds; (ii) endosymbiont-free plants can be readily generated, allowing controlled growth and biocontrol experiments; and (iii) being an important crop species in Northern America, there is great interest in translational research. In 2019, the world production of cranberry was 755,581 tons, with North America accounting for 77% (https://www.fao.org/faostat/). The traditional overfertilization of cranberry fields with phosphorus (P) and fungicide treatments is both costly and a threat to the environment. Therefore, the use of microbial symbionts that assist in the solubilization and plant uptake of P (and other immobilized mineral nutrients in the soil) and control disease promises to improve the sustainability of the cranberry industry.

Here, we describe the isolation and the identification of endosymbionts that colonize cultivated cranberry plants in Quebec, Canada. We identify endosymbiont species by ribotyping, explore the cranberry microbiome, assess the growth of microbe-colonized host plants under controlled greenhouse conditions, and test the microbes' capacity to solubilize phosphorus and to control fungal disease.

MATERIALS AND METHODS Plant origins, sampling, and endophyte isolation

Endophyte communities in rooted cuttings of Stevens, Scarlet Knight, and Mullica Queen cultivars were determined from plant material kindly provided by Dr. Nicholi Vorsa (Rutgers University) or collected from the Group Nadeau farm (Sainte-Mélanie, Quebec, Canada). We also identified endophyte communities in seven commercial cranberry fields (Stevens cultivar) growing under either organic or conventional (defined as using chemical fertilizers and fungicides) farming conditions. In each field, three blocks were randomly sampled, and from each, three plants were picked. Five tissue samples were randomly collected from each plant tissue, cut into sections, and surface-sterilized using a sodium hypochlorite solution (0·9%, w/v) as described in Cao et al. (2004) (see Supporting Methods S1 and Figure S1). For isolation of endophytes, tissues were placed on tryptic soy agar (TSA) medium with cycloheximide (40 μg per ml) for bacteria and on PDA medium with antibiotics (streptomycin, tetracycline, ampicillin, kanamycin; 50 μg per ml) for fungi (see Figure S1). Samples were incubated at 24°C for up to 6 weeks. Bacterial colonies were passed through five rounds of single-colony isolation. Fungal isolates were purified either via single spores or for non-sporulating fungi passed through five rounds of isolating small hyphal fragments.

DNA extraction and endophyte identification

Cells from bacterial and fungal liquid cultures were mechanically broken by grinding with a mortar and pestle in the presence of glass beads (425–600 μm for fungi, 150–212 μm for bacteria; Sigma). DNA was extracted using Qiagen Genomic-tips 20/G (Qiagen, GmbH). Bacterial ribosomal rRNA genes were amplified with universal 16S rRNA primers (27-F, 534-R and LPW58-R) and fungal ITS with the rRNA-specific primers (BMBC-F, ITS1-F and ITS4-R) (Figure S1 and Table S1 (Lane, 1991; Weisburg et al., 1991; White et al., 1990). Wizard SV Gels and the PCR Clean-Up System (Promega, USA) were used to purify PCR products prior to bidirectional Sanger sequencing. Sequences were assembled using Phrap (version 1.090518), visualized with Consed (version 27.0), and trimmed by quality score (>30), and primer sequences were removed. Species were identified by searches with BLASTN (http://www.ncbi.nlm.nih.gov/BLAST/) or Megablast (for highly similar sequences) in the non-random database nr. Multiple matches within the 10 best hits with the same taxon were considered as correctly identified, whereas matches to more than one taxon were labelled ambiguous.

Preparation of endophyte inoculum

Bacteria were grown on tryptic soy broth (TSB) at room temperature and suspended to a cell density of 0.1 OD₆₀₀ using a Varian Cary 50 Bio ultraviolet (UV)–visible spectrophotometer as described in Bapat et al. (2006) (see Supporting Methods S2). Fungi were grown on potato dextrose broth (PDB) for 4–15 days at room temperature. For sporulating fungi, cell suspensions were adjusted to 1.5 × 10⁶ conidia per. For non-sporulating fungi, the harvested mycelium was ground in a blender (Hamilton Beach, model 51109C, 225 W for 3 min), and the titer of colony-forming units (CFUs) was adjusted to 2 × 10⁶ CFU per ml.

Screening of growth suppression (biocontrol) activity on agar plates

Endophytes were screened for antifungal activity against three cranberry pathogens: Colletotrichum gloeosporioides, Diaporthe vaccinii, and Godronia cassandrae, plus one broad-range plant-pathogenic fungus: Cytospora chrysosperma. Fungal pathogens were isolated from diseased cranberry stems and berries and collected in local cranberry fields (see Supporting Methods S3). Biocontrol testing was performed by placing 5 mm discs with endophytes and fungal pathogens on PDA plates at a 3 cm distance (Figure S1). After incubation for 7–15 days at 25°C, the growth radius of the pathogenic fungus was measured. The percentage of radial growth inhibition (PRI) was calculated according to the equation: PRI = (r1 − r2/r1) × 100 (where r1 is the colony radial growth in control plates and r2 is the colony radial growth in the respective screening experiment) (Trivedi et al., 2008). Biocontrol of EC4 by EB37 was evaluated in the same way except that the strains were directly streaked on the agar plate (Figure S2).

Determination of phosphate solubilizing activity

Isolates with significant biocontrol performance were also tested to solubilize insoluble phosphate (P). Bacillus velezensis EB37, Lachnum sp. EC5, and Codinaea sp. EC4 were cultivated on agar plates containing National Botanical Research Institute's phosphate growth medium (NBRIP, see Supporting Methods S4) (Nautiyal, 1999). Three different compounds were tested as sole phosphate sources: tri-calcium phosphate (Ca₃ (PO₄)₂), hydroxyapatite (Ca₅(PO₄)₃(OH)), and phytate (inositol hexakisphosphate; C₆H₁₈O₂₄P₆). For that, 20 μl of endophyte suspension (0.1 OD₆₀₀) was applied in triplicate in the middle of the Petri dish, and the plates were incubated at 25°C for 7 days. The occurrence of a clearing zone in the turbid agar indicates the ability of a given microbe to solubilize phosphate.

Endophyte inoculation and plant growth conditions

Selected endophytes were screened for plant growth stimulation on four cranberry cultivars: Stevens, Mullica Queen, Demoranville, and Scarlet Knight. Fresh runners were collected in the autumn of 2013 from the Atocas de l'Érable farm (Notre-Dame-de-Lourdes, Quebec, Canada), cut into segments 4 cm in length, and surface-disinfected as described by Debnath (2007) (see Supporting Methods S5). The cuttings were then planted in sterilized 32-cell plug trays (5 × 5 cm) containing one part sand, three parts peat substrate (v/v), and 10% perlite. After 8 weeks, 10 rooted cuttings were inoculated with endophyte suspensions according to a 4 × 6 factorial experiment (see Supporting Methods S2). Endophyte suspensions (10⁵ CFU/ml) were applied to the soil around the cutting. Regular watering was as required, with autoclaved tap water. Pots were incubated at room temperature (22–28°C) under a light intensity of 200 μmol m⁻² s⁻¹ photons at a daily growth cycle of 18 h day-light and six h darkness. The length of shoots was recorded after 90 days, a period relevant for establishing new cranberry fields from rooted cuttings.

Statistical data analysis

Species richness (S), Shannon-Weaver diversity index (H′) and principal component analysis (PCA; on Hellinger transformed data) were computed using the package ‘vegan’ (Oksanen et al., 2015), implemented in R (version 3.2.3.) (Team, 2006). The effects of fields and plants were tested by multivariate variance analysis (MANOVA) (Legendre & Gauthier, 2014). The Kruskal–Wallis rank-sum test was performed using the package “FSA” (Ogle et al., 2018), followed by the Dunn test or the pairwise Mann–Whitney test as a multiple comparison post hoc analysis. Differences were considered statistically significant for p values < .05. Graphics were drawn using “ggplot2” (Valero-Mora, 2010), and the Venn diagrams were constructed using an online tool (http://bioinformatics.psb.ugent.be/webtools/Venn/).

RESULTS Organic cranberry farming systems are rich in microbial symbionts

A total of 181 distinct bacterial and fungal endophytes were collected from roots, stems, and leaves of cranberries plants sampled in organic and conventional commercial fields in Quebec. Endophytes grouped into 86 operational taxonomic units (OTUs), including 33.7% bacteria (29 OTUs) and 66.3% fungi (57 OTUs) (Table S2). Regardless of the farming system, bacterial isolates included Proteobacteria (mainly Pseudomonas spp.) and Firmicutes (mostly Paenibacillus spp.), yet isolates of Rhodococcus sp. (Actinobacteria) were found solely in organic fields (Figure 2a). In both farming systems, fungal isolates were dominated by Sordariomycetes, Leotiomycetes, and Dothideomycetes.

FIGURE 2. Diversity of fungal and bacterial endophytes isolated from fields under different management regimes. (a) Relative mean abundance of bacterial and fungal operational taxonomic units (OTUs) (grouped into 39 OTUs based on their affiliation to a given species by ribotyping) observed across conventional (conv) and organic (org) fields. (b) Abundance of bacterial and fungal endophytes detected in leaves (green), stems (light red), and roots (light blue) in organic versus conventional fields. Boxes span the lower, median, and upper quartiles of the raw data, with vertical lines indicating the minimum and maximum values. See also Figure S4 for species richness and diversity.

Endophyte occurrence was analyzed using MANOVA, in all fields and with respect to plant tissues (Table S3). Results show that 5.5% of the total variance is related to plant tissues, which is not surprising because the majority (~80%) of endophytes were isolated from roots (Figures 2b and 3a). A variation of 15% is due to differences among fields and the combined variance of tissues plus fields accounts for 22%. The remainder of variation (78%) remains unaccounted for by the selected factors.

FIGURE 3. Distribution of bacterial and fungal endophytes across plant tissues. The 86 isolated endophytes were grouped into 39 operational taxonomic units (OTUs) based on their affiliation to a given species by ribotyping. (a) Percentage of endophytes isolated from cranberry leaves, stems, and roots. The majority colonized root tissue. (b) Venn diagram showing the number of identical root-endophyte OTUs isolated from different fields (O1–O4, organic; C1, C2, and C3, conventional regime). Only five out of 39 OTUs were detected in more than one field.

We investigated specific differences of endophytes isolated from roots, in more detail. PCA shows that root endophyte assemblies also vary substantially across fields and even across plants from the same field, regardless of the farming system (Figure S3). Yet, organic fields (particularly O2 and O3) stand out with higher root-endophyte richness and diversity (Figure S4a,b and Table S3b). At the level of species composition, no endophyte was commonly found in root samples across all fields (Table S4). Certain endophytes were found only in plant roots from organic fields, whereas endophytes from conventional sites were to a large extent field-specific (Figure 3b and Table S4). In summary, the number of distinct endophytes varies substantially in conventional cranberry fields but even more so in organic plantations.

Fungicide treatment suppresses DSE and mycorrhizal but not potential plant pathogenic root endophytes

The question if fungicide treatment influences root colonization by fungi was addressed by examining rooted cranberry cuttings. We compared the development of cuttings that were rooted and grown either conventionally (Stevens, Scarlet Knight and Mullica Queen; using fungicides) or “organically” (Stevens and Scarlet Knight; without fungicides). A total of 77 fungal OTUs were identified (Table S2b), and a PCA analysis shows that the cuttings grown without fungicides harbor a significantly higher number of DSE members (e.g., Phialocephala fortinii and Phialocephala sp.) and ericoid mycorrhiza-related species (e.g., Rhizoscyphus ericae and Pezicula ericae) (Figure 4a,b, Table S2b). Most remarkably, none of the fungicide-treated cuttings harbored either DSE or close relatives of ericoid mycorrhizal species but instead a variety of potentially disease-causing and/or opportunistic fungal species (Figure 4b and Table S2b).

FIGURE 4. Fungal root-endophytes isolated from rooted cuttings of various cranberry cultivars. (a) Principal component analysis (PCA) plot of the Hellinger-transformed root-endophyte abundance data. Symbols filled blue, cuttings from conventional fields; symbols filled red, cuttings from fields without use of fungicides. According to the equilibrium circle method (red circle), the magnitude and the direction of the red lines indicate the scores of the contribution of operational taxonomic units (OTUs) to the separation of the cultivars. The fungal pathogens Colletotrichum sp. and Diaporthe sp. dominate conventionally rooted cuttings (the rooting substrate contains fungicides), whereas relatives of the mycorrhizal fungus Rhizoscyphus ericae dominate in “organic” cuttings rooted without fungicides. Data points close to the center of the graph are not shown. (b) Functional classification of fungal OTUs isolated from roots of the cultivars Mullica queen (MullQ), scarlet knight (ScarK), and Stevens; blue, conventional cuttings; orange, “organic” cuttings. OTUs were assigned to functional groups based on their affiliation to genera, members of which were designated in the literature as ericoid mycorrhizal fungi (Rhizoscyphus ericae and Pezicula ericae), dark septate endophytes (DSE) (Phialocephala fortinii and Phialocephala sp.), or pathogens (Colletotrichum sp., Alternaria alternata, Diaporthe sp., Physalospora vaccinii, Guignardia vaccinii, Godronia sp., Leptosphaeria sp., Didymella sp., Paraphoma sp., and Pleotrichocladium sp.). Pathogenic OTUs dominate in roots from conventionally produced cuttings (which includes fungicide treatments). Mycorrhizal and DSE OTUs were only observed in plants from organically managed fields.

Bacillus velezensis (EB37) and Lachnum sp. (EC5) have strong biocontrol activities

We examined the ability of select endophytes (Table S2a–c) to inhibit the mycelial growth of four pathogenic fungi. Growth confrontation tests on agar plates show that the B. velezensis isolates and in particular strain EB37, which was isolated from leaf tissues of rooted Mullica Queen cuttings, were the most effective in controlling the growth of all fungal pathogens tested (Figures 5a,b, S5a,b, and Figure S6 and Table S5). The Lachnum sp. isolate EC5 was also able to control the growth of Diaporthe sp. and Godronia sp. (Figures 5b and S5b and Table S5), whereas the Codinaea sp. isolate EC4 inhibited C. chrysosperma and Godronia sp. Other bacterial and fungal isolates were less effective in biocontrol (Figure S5a,b and Table S5). Due to their superior biocontrol potential, the bacterium EB37 and the two fungi EC4 and EC5 were further analyzed, as described below.

FIGURE 5. Biocontrol activity of endophytes against two fungal pathogens. Biocontrol activity against (a) Colletotrichum gloeosporioides and (b) Diaporthe vaccinii was measured by a growth confrontation test on agar plates (for details, see methods). Bacterial endophytes, gray bars; fungal endophytes, light blue bars. Error bars indicate the two standard errors. Data on additional fungal pathogens are shown in the Figure S5. Colletotrichum gloeosporioides is not inhibited by the fungal endophytes EC5, EC13, and EC19 (data not shown).

Select endophytes solubilize phytate that is unavailable to plants

We tested the bacterium EB37 and the fungi EC4 and EC5 for their capability to solubilize phosphate compounds that are commonly found in soil but are water-insoluble, notably tricalcium phosphate (Ca₃(PO₄)₂), hydroxyapatite (Ca₅(PO₄)₃(OH)), and phytate (inositol polyphosphate). For this test, microbes were inoculated on agar plates containing these compounds. A clearing zone around the colonies was observed on phytate plates, showing that the microbes solubilized this compound (Figure 6a); the halo area of EB37 was more extensive than those of the two fungi. In contrast, tricalcium phosphate and hydroxyapatite were not solubilized (not shown).

FIGURE 6. Phytate solubilization and effect on plant growth by endophytes. (a) Phytate solubilization. Growth of three endophytes on agar plates containing the insoluble phytate (light-gray turbidity). The agar becomes transparent due to secreted phytases, creating a dark-gray clarification zone (halo) around the endophyte colonies. The halo diameter is an approximate measure of the amount of secreted phytase. (b) Effect on plant growth. The elongation of rooted cuttings from the cranberry cultivars Demoranville (Demor), Mullica queen (MullQ), scarlet knight (ScarK), and Stevens was measured after inoculating the plants with a single or a combination of two endophytes. Vertical bars and asterisks represent standard error values and statistically significant increase in plant growth (***, p value ≤  0.002; **, p value > .002 and ≤0.01; *, p value > .01 and [less than].05; Kruskal–Wallis test; see methods) compared with the controls, respectively. EC4 and EC5 showed significant growth stimulation in two out of four cultivars. Error bars indicate the two standard errors.

Codinaea sp. EC4 and Lachnum sp. EC5 stimulate cranberry plant growth

Plant growth stimulation was tested by inoculating four cranberry cultivars with select endophytes. The results are shown in Figure 6b and Table S6. At a p value threshold of .05, the strongest significant stimulation was observed (i) in the Stevens cultivar inoculated with either Codinaea sp. EC4 or Lachnum sp. EC5; (ii) in Scarlet Knight inoculated with EC4; and (iii) in Mullica Queen inoculated with EC5. No growth promotion was noted in Demoranville plants. Treatments with the bacterium B. velezensis EB37 had no significant effect on either plant cultivar. However, when combining EB37 with the fungus EC5, an increase was observed on Scarlet Knight, compared with the inoculation with the fungus alone (p value < .05). Conversely, the combination of EB37 with the fungus EC4 inhibited plant growth on both Scarlet Knight (p value < .01) and Stevens (p value < .001), compared with the treatment with the fungus alone.

Two months after the inoculation with EC5, a microscopical inspection of plant roots revealed an intracellular colonization by the fungus that resembled previously described “ericoid mycorrhiza” (Figure 7c) (Vohník, 2020; Vohník & Albrechtová, 2011). In addition, we identified mycelial aggregates on the root surface that parallel a loose hyphal mantle (Figure 7d) and root tips surrounded by mycelium. No such structures were observed on control roots (Figure 7a,b). The fungus also formed extraradical hyphae (Figure 7e) that extend into the rhizosphere, similar to what is observed in AMF symbiosis. This merits the investigation into the means of communication involved in this plant-fungus interaction and into the transport and uptake mechanism of plant nutrients.

FIGURE 7. Lachnum EC5 colonization of cranberry roots. Cranberry plants were grown axenically from Stevens seeds and inoculated with Lachnum sp. EC5. (a) Cranberry root without fungal hyphae (control). (b) Cranberry root tip after 2 months of growth, covered with EC5 hyphae resembling ectomycorrhizae. (c) Intracellular hyphal loops, resembling structures in other ericoid fungal mycorrhizae (stained with the fluorescent dye solophenyl flavine 7GFE 500). (d) Hyphal aggregates of EC5 on the surface of cranberry roots resembling an ectomycorrhizal mantle. (e, f) Extension of extra-radical mycelium from the mantle into the rhizosphere

DISCUSSION How should ErMF be defined?

“Ericoid mycorrhiza, … arguably the least researched and perhaps also the least understood type of mycorrhizal symbiosis” (Vohník, 2020) is currently defined by a set of strict morphological characters. This definition would exclude Lachnum sp. EC5 from the ErMF group, although it belongs to the same phylogenetic group as several “true ErMFs” (Helotiales), forms ErMF-like intracellular structures in cranberry roots, solubilizes phytate, stimulates plant growth, and has the potential for biocontrol of phytopathogenic fungi. Could the observed phenotypic disparity result from ongoing rapid evolution of mycorrhizal traits? Indeed, the morphological characters common to “true ErMF” unite certain helotialean Ascomycota and Basidiomycota species (Kolařík & Vohník, 2018), groups that, in turn, also include saprophytic and plant-pathogenic species (for Helotiales, see Figure 1). In our view, the currently used strict morphological traits defining “true ErMF” are poorly suited for classifying ericoid fungal symbionts, as these traits are most probably the result of rapid convergent evolution and/or horizontal gene transfer.

The principal biological role of ericoid (and other) endophytes in microbe-plant symbioses is usually defined at the level of nutrient acquisition (e.g., Vohník, 2020, Yang et al., 2018 and references therein), omitting a variety of other biological functions. The genome analysis of ericoid fungi (“true ErMF,” DSE, etc.) reveals a capacity for degrading numerous potentially toxic organic substances, as well as a broad biocontrol potential that likely acts on both the plant host and on the microbial partner(s) involved in the symbiotic interactions (Kang et al., 2018; Martino et al., 2018; Sheoran et al., 2015). Taken together, we advocate describing ericoid symbiotic interactions in a more flexible and holistic way, including molecular, genetic, phylogenetic, genomic, morphological, and other relevant traits.

Bacillus velezensis has broad-range, highly effective biocontrol activities

Previous genome studies have shown that Bacillus species are among the plant growth-promoting bacteria (PGPB), which frequently possess genes to synthesize a wide range of secondary metabolites with antibacterial, antifungal, insecticidal, nematicidal, and antiviral activities, such as polyketides and non-ribosomal peptides (Cao et al., 2018; Fan et al., 2012; Gao et al., 2018; Kang et al., 2018). In symbiotic plant-Bacillus interactions, synthesis and secretion of these compounds will extend the defense mechanisms of the plant host against diseases. Among the B. velezensis strains isolated from leaves and stems of the cranberry cultivar Mullica Queen, five showed a strong biocontrol activity compared with other tested endophytes (Figures 5, S5 and S2 and Table S5), and B. velezensis EB37 was the most potent and broad-range biocontrol agent in our collection. It would be interesting to sequence the EB37 genome to uncover the molecular basis of this exceptional capability.

Moreover, we observed that EB37, despite its biocontrol activity against fungi (Figure 5 and Figures S2, S5, and S6), did not inhibit the growth of the fungal symbiont Lachnum sp. EC5 but rather merged with the latter in the confrontation test on agar plates forming physical contacts. This observation suggests that the two microbes either have immunity to each other's battery of biotoxins or have downregulated their production of antimicrobials following interspecies crosstalk. In either case, it would be worthwhile to test this pair of endophytes for their capacity to suppress pathogens in a cranberry plant system and to determine the secondary metabolites produced under this condition, most notably the expected wide range of non-ribosomal peptides, polyketides, and hybrid peptide-polyketides.

Lachnum sp. EC5 and Codinaea sp. EC4 stimulate cranberry growth—Depending on the plant cultivar and the bacterial co-symbiont

It has been reported for many plants that the effect an endophyte exerts on its host critically depends on the plant genotype (do Amaral et al., 2016; EstaÚN et al., 1987; Oberhofer et al., 2014), and the same is true for cranberry (Figure 6b). For instance, we observed that Lachnum sp. EC5 stimulates the growth of Mullica Queen and Stevens, but not Scarlet Knight, whereas Codinaea sp. EC4 promotes the growth of Stevens and Scarlet Knight but not of Mullica Queen. Evidently, the plant–fungus interactions vary with the different genetic background of the cranberry cultivars.

When using mixtures of fungal and bacterial endophytes, the effect on plant growth was often unexpected. Notably, the combination of EC5 with EB37 improved the growth of Scarlet Knight (although at a marginal p-value, see Table S6), whereas EC5 and EB37 alone had no effect. Another unexpected observation was that EB37, which alone showed no significant effect on Stevens plants, abolished the growth stimulation by EC4. This suggests that EC4 is suppressed by antifungal secondary metabolites produced by EB37, in contrast to the EC5–EB37 combination that stimulated Scarlet Knight growth (likely due to downregulation of secondary metabolite synthesis). In other words, in settings with more than one microbial symbiont, interactions between the fungal and bacterial endophytes have to be taken into account, in addition to the individual microbe-plant interactions.

Endophytes can stimulate plant growth in several ways from the secretion of hormones to the provision of plant nutrients. Here, we tested endophytes for their capability to dissolve water-insoluble phosphate compounds and found that phytate was readily solubilized by several microbes and most efficiently by B. velezensis EB37 and Lachnum sp. EC5 (Figure 6a). Phytate is the major fraction of organic phosphate that accumulates in agricultural soils but is unavailable to plants (Brinch-Pedersen et al., 2002; Findenegg & Nelemans, 1993; Hayes et al., 2000). These phosphate reserves are known to be recycled by certain microbes that secrete phytate-specific phosphatases (Mukhametzyanova et al., 2012; Yoon et al., 1996). The plant-growth stimulating action of EB37 and EC5 may be due to phytase solubilization and provision of free phosphate to the plant.

The application of microbial symbionts in cranberry farming

Our studies on cranberries and those on other crop plants (Long et al., 2010; Xia et al., 2015, 2019) have demonstrated that organically managed fields are more endophyte-rich than conventional fields. As we show here, a considerable fraction of these endophytes have biocontrol and plant growth-promoting properties. The application of such microbes in cranberry farming promises to improve the industry's sustainability. Yet, field conditions differ largely from controlled greenhouse experiments, in particular by the presence of already established symbioses and microbial communities. Therefore, future research has to go beyond studying the interactions of single Ericaceous symbionts and their host and include the aspect of microbial symbiont communities living within plants and in their immediate environment.

ACKNOWLEDGMENTS

This work has been supported by Quebec commercial cranberry producers (M. Bieler, D. Landreville, V. Godin, K. Lachance, P. Fortier, Y. Montreuil), J. Painchaud (MAPAQ), J.-P. Deland (Oceanspray), and grants from MAPAQ and NSERC. We thank Dr. Nicholi Vorsa (Rutgers University) for providing rooted cuttings of the Stevens, Scarlet Knight, and Mullica Queen cranberry cultivars and Mathew Sarrasin (Robert-Cedergren Centre for Bioinformatics and Genomics, Université de Montréal) for valuable comments on the manuscript.

CONFLICT OF INTEREST

The authors declare that the research was conducted without any commercial or financial relationships that could be construed as a potential conflict of interest.

AUTHOR CONTRIBUTIONS

LNS, BFL, and GB wrote the manuscript. All authors discussed the results and contributed to improvements to the final manuscript. LNS and LF performed the isolation and identification of endophytes and performed the biocontrol, phosphate solubilization, and plant growth stimulation tests. PBV isolated and identified the endophytes from organically produced cranberry cuttings. LNS conducted statistical analyses of the collected data. GB and BFL supervised the project.

DATA AVAILABILITY STATEMENT

The datasets supporting the conclusions of this article are available within the article and its supplementary information files or are available from the corresponding authors upon request. Source data are provided in this paper.