Does in vitro selection of biocontrol agents guarantee success in planta? A study case of wheat protection against Fusarium seedling blight by soil bacteria

Abstract

Biological control is a great hope for reducing the overutilization of pesticides in agricultural soils. It often involves microorganisms or molecules produced by microorganisms that will be able to interact with either a plant or pathogens of this plant to reduce the growth of the pathogen and limit its negative impact on the host plant. When new biocontrol products are developed, strains were mostly selected based on their ability to inhibit a pathogen of interest under in vitro conditions via antagonistic effects. Strains with no in vitro effect are often discarded and not tested in planta. But is the in vitro selection of bacterial agents according to their antagonism activities towards a plant pathogen the best way to get effective biocontrol products? To answer this question, we used wheat and the fungal pathogen Fusarium graminearum as a study pathosystem model. A library of 205 soil bacteria was screened in 2 types of in vitro growth inhibition tests against F. graminearum, and in an in planta experiment. We find strains which do not have inhibition phenotypes in vitro but good efficacy in planta. Interestingly, some strains belong to species (Microbacterium, Arthrobacter, Variovorax) that are not known in the literature for their ability to protect plants against fungal pathogens. Thus, developing a biocontrol product against F. graminearum must be preferentially based on the direct screening of strains for their protective activity on wheat plants against fungal diseases, rather than on their in vitro antagonistic effects on fungal growth.

About the Authors:

Yoann Besset-Manzoni

Roles Conceptualization, Investigation, Methodology, Writing – original draft, Writing – review & editing

Affiliations UMR Ecologie Microbienne, CNRS, INRA, VetAgro Sup, UCBL, Université de Lyon, Villeurbanne cedex, France, Biovitis, Le Bourg, Saint Etienne-de-Chomeil, France

Pierre Joly

Roles Conceptualization, Funding acquisition, Investigation, Methodology, Project administration, Resources, Supervision, Validation, Writing – review & editing

Affiliation: Biovitis, Le Bourg, Saint Etienne-de-Chomeil, France

ORCID logo http://orcid.org/0000-0003-4972-5590

Aline Brutel

Roles Resources, Writing – review & editing

Affiliation: Biovitis, Le Bourg, Saint Etienne-de-Chomeil, France

ORCID logo http://orcid.org/0000-0003-4005-7565

Florence Gerin

Roles Investigation

Affiliation: UMR Ecologie Microbienne, CNRS, INRA, VetAgro Sup, UCBL, Université de Lyon, Villeurbanne cedex, France

Olivier Soudière

Roles Investigation, Resources

Affiliation: INRA, UMR 1095 Génétique, Diversité et Ecophysiologie des Céréales, Clermont–Ferrand, France

Thierry Langin

Roles Conceptualization, Investigation, Methodology, Resources, Supervision, Validation, Writing – review & editing

Affiliation: INRA, UMR 1095 Génétique, Diversité et Ecophysiologie des Céréales, Clermont–Ferrand, France

Claire Prigent-Combaret

Roles Conceptualization, Funding acquisition, Investigation, Methodology, Project administration, Resources, Supervision, Validation, Writing – review & editing

* E-mail: [email protected]

Affiliation: UMR Ecologie Microbienne, CNRS, INRA, VetAgro Sup, UCBL, Université de Lyon, Villeurbanne cedex, France

ORCID logo http://orcid.org/0000-0001-8968-0660

Introduction

Since the Green revolution, the way crops are grown has seen profound change. The use of chemical compounds designed by humans to improve health and productivity of plants has strongly expanded. By improving the quality and the yield of plant productions, and controlling agricultural pests, chemical fertilizers and pesticides brought beneficial effects to society [1–2]. All these benefits allowed the worldwide population to expand. But since 1963, the negative impacts of pesticides on human health, wildlife and environment have been extensively reported [3–4]. Negative impacts on biotic components can occur at different levels from individual organisms to communities, affecting the balance of the whole ecosystem. The main cause of this is that less than 0.1% of the pesticides used for pest control reach their targets accurately [5–6], leaving the rest of the active substances in the ecosystems. This leads to a problematic emergence of resistance to chemical pesticides in the phytopathogens [7]. Methods have been developed to evaluate the impact of pesticides on the environment, and they allowed to show that pesticide applications must be reduced to limit human health problems, environmental contamination and reduced soil fertility [7–9]. Simultaneously, new ways of crop production grouped under the generic term Integrated Pest Management (IPM) have been developed to control pests in agriculture. One mechanism involved in IPM strategies consists of the use of natural predators, antagonists, competitors or pathogens of pest targets, known as biological control agents. The principle of biocontrol is to maintain the targeted pest at a threshold below the limit responsible for negative impacts on the plant. Indeed, some microorganisms from soils have been shown to have beneficial effects on plant health and growth [10–11]. The mechanisms involved are quite diverse: antibiosis through the production of antimicrobial compounds like cyclic lipopeptides, polyketides, or lytic enzymes [12], (ii) competition for resources [11, 13], (iii) competitive exclusion [13], and (iii) elicitation of plant induced systemic resistance [14]. In addition, if the microorganisms have the ability to enhance the bioavailability or assimilation of essential nutrients such as nitrogen or phosphorus by the plant, it can help the plant to better cope with pathogen attacks [11]. In the hope of finding highly active biocontrol bacteria, the number of studies reporting screening of bacterial libraries against pests have grown exponentially.

In the present work, the Wheat-Fusarium graminearum pathosystem was used as a model system. F. graminearum is a well-known pathogen of small grain cereals and particularly wheat, capable of inducing four types of diseases: damping-off, root-rot, crown-rot, and Fusarium head blight (FHB or scab). FHB is one of the main fungal diseases of wheat. The FHB species complex produces mycotoxins that cause quality and yield reductions, as well as human and animal health risks. Cultural and management practices can reduce but not completely control FHB epidemics [15]. Due to its life cycle, solutions for chemical control of F. graminearum are not very effective [16–18]. In the literature, some bacterial genera, such as Bacillus or Pseudomonas are often found as bacteria used for biocontrol of F. graminearum in wheat [19–23]. More rarely, other genera like Stenotrophomas [19], Acinetobacter and Chryseobacterium [24], Streptomyces and Brevibacillus [25] are explored. These bacteria are often derived from either the rhizosphere soil of wheat, or from ears or grains, especially in the case of targeting bacterial endophytes [21]. The methods of selection of these bacteria often rely on the ability of the strains to inhibit the F. graminearum mycelial growth under in vitro conditions. Effective bacteria inhibiting the fungal growth are then tested in plant disease protection trials in a greenhouse, or directly in the field, on large or small scales [22, 24]. In all the studies, screening steps allow elimination of bacteria that do not show antagonistic effects on F. graminearum, either by the inhibition of the fungal growth or reduction of mycotoxin concentrations. But moving from laboratory to field trials often leads to bad performances and inconsistent results.

Thus, the literature confirms that screening is essentially made firstly on in vitro inhibition tests against a plant pathogen of interest and that secondly, only the best bacteria are selected for plant protection experiments. But, does selecting the best bacteria in vitro guarantee success in planta? In order to answer this question, we performed for this study a classic selection of bacterial strains made from in vitro dual-culture tests against F. graminearum. We then classified the bacteria from our library in different categories of antagonist activities. A wide selection of bacterial candidates harboring high or low in vitro activities was then tested in a greenhouse disposal to confirm the first screening step. A particular focus was made on some bacterial genera that were not often tested in the literature as biocontrol agents, in order to evaluate their potential protective activity against F. graminearum.

Materials and methods

Constitution of the bacterial strain library

In order to access to a wide diversity of bacterial strains, the library was built from 4 different wheat-growing soils from 4 locations in France, i.e. La Côte Saint André (LCSA; silty loam luvisol, 45° 22′ 43″ N, 5° 16′ 02″ E, Isère, France) [26], the Limagne basin near Clermont-Ferrand (clay-rich soil vertisol, 45° 47′ 6″ N, 3° 11′ 11″ E, Clermont-Ferrand, France) [27], Trelins (45° 43' 42.1" N, 4° 01' 34.6" E, Loire, France) and Yzeure-sur-Creuse (46° 48' 39.8" N, 0° 49' 47.2" E, Indre-et-Loire, France). For each sampling site, 5 g of rhizosphere soils were mixed with 10 ml of a 0.8% (w/v) NaCl solution. Serial dilutions of soil suspensions were carried out until a dilution of 10−9 and 100 μL of each suspension was spread on 1/10-diluted TSA (Tryptone Soya Agar, Carl Roth, Karlsruhe, Germany) as described by Bachate et al. [28]. After 48 hours of incubation at 28°C, isolated colonies were transferred to new TSA Petri plates and purified at least two additional times. Pure isolates were then grown in TSB (Tryptone Soya Broth, Carl Roth, Karlsruhe, Germany) for 24 h at 28°C and then aliquoted and stored at -80°C in glycerol (at 40%).

Strains—Taxonomic characterization

Isolates from the whole library were taxonomically characterized based on the sequencing of their housekeeping 16SrRNA encoding rrs gene. Genomic DNA from all bacteria was extracted using the NucleoSpin Microbial DNA kit (Ref—740235.50, Macherey-Nagel, Germany). The rrs gene was amplified with primers pA/pH [29] and sequenced. The sequences were aligned with Muscle [30]. The phylogenetic tree was inferred using Seaview and the Neighbor-Joining distance-based method with 1000 bootstraps. Other methods were implemented and gave similar tree topologies. The tree was represented using iTOL [31]. The European Nucleotide Archive (ENA) accession number for the 203 rrs sequences reported in this paper is ERP115553. Other accession numbers used in this work are KT380501.1 for Thaumarchaeota archaeon NAOA2, NZ_AWQY01000007.1 for Bacillus amyloliquefaciens AB42 and AJ278814.1 for Pseudomonas kilonensis F113.

In vitro screening of the efficacy of bacterial strains to inhibit the growth of Fusarium graminearum

The highly virulent and toxin-producing French isolate Fusarium graminearum MDC_Fg1 (throughout referred to as Fusarium graminearum Fg1) used throughout experimentations was provided by Dr. Thierry Langin (GDEC Joint Research Unit, INRA Center Auvergne—Rhône-Alpes, Clermont-Ferrand, France). This strain was isolated from naturally infected cereal grains in the North of France. Fg1 is able to induce FHB as well as the Fusarium crown-rot and seedling blight. For spore production, the cultures were grown on liquid Mung Bean Broth (MBB), for five days at 26 C. Spore concentration in the suspension was adjusted to 106 spores.mL−1.

Growth inhibition test on solid medium

Dual-culture protocols were implemented to assess the potential effect of bacterial antagonism on Fg1 mycelium growth on solid medium. All experiments were carried out in triplicates. First, each bacteria was deposited 2 cm apart from the center of PDA (Potato Dextrose Agar, Conda Pronadisa, Madrid, Spain) plates as a 2 cm-long line, from TSB cultures grown over 24 hours. After incubation at 28°C for 24 hours, a 6 mm-diameter agar with mycelium sampled from the leading edge of a 5 day-old culture of F. graminearum MDC_Fg1 grown on PDA was placed on the center of the plate. PDA plates were inoculated only with the pathogen for fungal control. The plates were then placed at 28°C in the dark for 5 days before the measurement of an inhibition index according to Shi et al. [32] with some modifications. The efficacy of pathogen growth inhibition was calculated according to 2 measurements carried out on the same plate: the radius of fungal growth towards the site of bacterial growth, and the radius between the center of the plate and the limit of fungal growth towards the site free of bacteria. The ratio between these 2 values was calculated and compared to that of the positive control without bacteria. Growth inhibition of mycelium on solid medium (GIm) was then estimated (S1 Fig). When GIm is lower than 0 for a given strain, the number is replaced by 0, and the strain is considered as a non-inhibitory strain.

Growth inhibition test in liquid medium

Supernatants of bacteria were used to assess their antagonism potential on Fg1 spore germination and/or mycelium growth in broth, in a microplate test. All experiments were carried out in triplicate. The supernatant of each bacteria was prepared from a 1-day old TSB culture, centrifuged at 4500 rpm during 10 min and filtered at 0.2 μm. For each treatment, 100 μL of supernatant, 100 μL of PDB (Potato Dextrose Broth, Conda Pronadisa) and 50 μL of Fg1 asexual spores (macroconidia) suspension at 104 spores.mL-1 were added per microplate wells. As positive control, 100 μL of TSB were used to replace the bacterial supernatants. As negative control, 50 μL of PDB were used to replace the Fg1 spore suspension. After 5 days of culture at 28 °C, the microplates were analyzed using an Infinite M200 Pro microplate reader (TECAN, Switzerland). The turbidimetry (that increases with mycelial growth) was measured at 492 nm in each well [33]. Inhibition of spore germination and growth in liquid medium (GIs) was then estimated. In this experiment, the OD of the fungal positive control is equal to 1.7. If GIs was lower than 0, the number is replaced by 0 and the strain is considered as a non-inhibitory strain.

Wheat protection test

The protection assay of wheat by candidate strains against F. graminearum Fg1 was implemented in a greenhouse (INRA GDEC, Clermont-Ferrand). A panel of 12 strains was selected based on their in vitro efficacy against Fg1. In addition, 2 model biocontrol PGPR strains, Pseudomonas kilonensis F113 and Bacillus amyloliquefaciens AB42 (provided by the company BIOVITIS) were included in the greenhouse experiment. For each condition, 50 seeds of Chinese spring wheat cultivar Sumai 3 were distributed in 10 pots (12 x 10 x 10 cm) filled with 250 g of compost and 5 plants per pot. For each in vivo test, bacteria were prepared from TSB cultures grown over 24 hours at an optical density at 600 nm (OD600nm) of 1 (i.e. 108 cells/mL) in sterile water. Bacteria were inoculated into each seed (107 cells/seed) with 100 μL of prepared cultures. After 3 days, Fg1 spores (105 spores/seed) were added. Plants were watered every 3 days in order to keep soil water content close to 40% of Water Holding Capacity. After 45 days of culture, the experiment was stopped and the plants harvested. Different measurements were performed: (i) the chlorophyll rate of each wheat plant consisting of three measures performed on the last formed leaf using a SPAD 502 plus device (Minolta Camera Co., Osaka, Japan); (ii) the length (measured in cm) of the wheat strands; (iii) the dry weight of shoots; (iv) the disease symptoms of crown-rot on each wheat collar using a notation index ranging from 0 (no symptoms) to 3 (high symptoms); (v) the number of ears for each plant; (vi) dry weight of ears of each plant. For disease symptoms, the notation index was recorded as follow: 0 = no necrosis observed, 1 = necrosis observed on one sheet, 2 = necrosis observed one two sheets and 3 = necrosis observed on the whole collar. In addition, two different controls were also performed: (i) as a positive control, wheat was inoculated with only the pathogen spore suspension, and (ii) as a negative control, un-inoculated wheat was used.

Statistical analyses

All statistical analyses were performed using R software. PCA analyses were performed to compared plant growth parameters and symptoms between treatments. Comparison of means was made using Kruskal-Wallis tests (P<0.05).

Results

Composition of the wheat-rhizosphere bacteria library

Two hundred and three cultivable strains were isolated from the 4 soils by dilution-spreading. Two reference biocontrol strains were added: P. kilonensis F113 and B. amyloliquefaciens AB42 known to have antagonistic effects on fungi. These 205 strains are distributed among 34 different bacterial genera (Fig 1). The most represented genera are Pseudomonas, Bacillus, Staphylococcus and Chryseobacterium with respectively 51, 30, 16 and 15 strains per genus. The genera belong to 7 bacterial classes, distributed as follows: 32.5% of Gammaproteobacteria, 25.6% of Bacilli, 14.8% of Actinobacteria, 10.3% of Flavobacteria, 9.4% of Betaproteobacteria, 4.4% of Alphaproteobacteria, and 3.0% of Sphingobacteria. The sampling effort was not similar between the soils with, from the highest to the lowest: Clermont-Ferrand 97 strains, LCSA 66 strains, Trelins 27 strains and Yzeures-sur-Creuse 13 strains isolated. Accordingly, the greatest diversity of bacterial genera was obtained from isolates of Clermont-Ferrand soil with 25 different genera, followed by LCSA soil with 14 genera, then Trelins and Yzeures-sur-Creuse soils with 9 genera (Fig 1). The smaller sampling in the two latter soils, however, allowed us to increase the global diversity of the library, providing bacterial genera absent from the 2 other rhizosphere soils (Fig 1): Acidovorax, Escherichia, Lelliottia, and Solibacillus. Some genera with multiple representatives were present in only one particular soil: Flavobacterium (n = 6), Lysobacter (n = 3), Stenotrophomonas (n = 3), and Burkholderia (n = 2) in the Clermont soil. Only two genera, Bacillus and Arthrobacter were found in the 4 soils, and 2 genera are found in 3 soils: Chryseobacterium and Variovorax. Overall, eighteen most abundant genera (i.e. genera with more than 3 representatives per genus in the library) were present in the library.

[Figure omitted. See PDF.]

Fig 1. Histogram describing the composition of the wheat-rhizosphere soil bacterial library.

The numbers above the bars refer to the number of isolates per genus recovered in all soils (Clermont-Ferrand, LCSA, Trelins and Yzeures-sur-Creuse).

https://doi.org/10.1371/journal.pone.0225655.g001

Fg1 mycelial growth inhibition on solid medium

All isolates (205) were screened for their ability to inhibit the mycelial growth of F. graminearum Fg1 on Petri dishes. GIm of each bacteria was reported on the phylogenetic tree of rrs gene sequences (Fig 2). Arbitrarily, a 50% GIm threshold was chosen to consider a strain as an inhibitory potential strain. A total of 61 strains obtained a GIm greater than or equal to 50%.

[Figure omitted. See PDF.]

Fig 2. Mycelium growth inhibition activity (GIm) for the 205 bacterial isolates of the library according to their rrs taxonomical affiliation.

The phylogenetic tree was inferred using the Neighbour-Joining method. GIm values (black bars) were reported on the rrs tree. The red solid circles indicate effective Fg1-antagonist bacteria, while empty red circles indicate non-effective Fg1-antagonist bacteria that were selected for the spore germination inhibition test. The tree was rooted with the rrs of the archaebacteria Thaumarchaeota archaeon NAOA2.

https://doi.org/10.1371/journal.pone.0225655.g002

The distribution of the GIm values is heterogeneous among the bacterial genera, and none of the genus contained only effective inhibitory strains. While for some genera, such as Clavibacter, Agrobacterium or Pseudoxanthomas, GIm varied little, for other genera, such as Bacillus, Pseudomonas, Chryseobacterium, there was a great heterogeneity regarding their fungal inhibition ability (S2 Fig). For example, the genus Bacillus shared an average GIm of about 20% but half of the strains had a GIm of less than 10%. This means that some strains had a strong in vitro inhibitory capacity. However, at the genus level, the overall capacity of the genus was quite limited. In addition, the genus Bacillus contained the strain with the highest GIm (strain JM3). The GIm of Achromobacter strains (n = 5) was around 0%. In the most represented genera, Pseudomonas (n = 51) and Staphylococcus (n = 16), 55% of the strains share a GIm higher than 50%.

The influence of the soil origin of strains on their ability to inhibit the mycelial growth of Fg1 was then analyzed (S3 Fig). When focusing on the genus Bacillus, the strains from the LCSA rhizosphere soil had GIm values much higher than the Bacillus strains present from the other soils (about 60% in LCSA, below 20% for the soil from Clermont-Ferrand and below 10% for the soils from Trelins and Yzeures-sur-Creuse). Results for the genus Chryseobacterium are opposite: the strains isolated from the Clermont-Ferrand rhizosphere soil are much more efficient, with an average GIm of about 75% than strains isolated from soils of LCSA (average GIm of 30%), and Yzeures-sur-Creuse (average GIm of 45%). Overall, the strains isolated from the LCSA soil seem to be more effective in the inhibition of the fungus than those isolated from the 3 other soils. Indeed, for the genera Bacillus, Pseudomonas, Paracoccus and Staphylococcus, the average GIm values are more important in LCSA soil than in the other soils. However, the Clermont-Ferrand soil contains strains from 8 different genera with GIm of at least 60% while this is only the case for 4 genera in the soil of LCSA, 2 genera in the soil of Trelins, and 1 genus in the soil of Yzeures-sur-Creuse (S3 Fig).

Fg1 spore germination and mycelial growth inhibition in liquid medium

From the different results obtained during solid-state inhibition tests, 33 very effective strains showing a GIm higher than 50% from 14 different genera were selected (i.e. Chryseobacterium (n = 1), Pedobacter (n = 2), Arthrobacter (n = 1), Rhodococcus (n = 1), Microbacterium (n = 1), Bacillus (n = 3), Staphylococcus (n = 5), Rhizobium (n = 1), Agrobacterium (n = 2), Stenotrophomonas (n = 1), Lelliottia (n = 1), Escherichia (n = 1), Pseudomonas (n = 12) and Variovorax (n = 1)). In addition, we also selected 14 strains with little or no-inhibitory activity against Fg1 and belonging to 8 common genera (i.e. Chryseobacterium (n = 1), Pedobacter (n = 1), Arthrobacter (n = 1), Microbacterium (n = 2), Bacillus (n = 2), Staphylococcus (n = 2), Stenotrophomonas (n = 1), and Pseudomonas (n = 4)) in order to avoid the mis-selection of relevant bacterial candidates expressing a distinct mechanism of action on F. graminearum, such as ability to inhibit asexual spore germination.

A total of 45 strains from the library and 2 reference strains were selected for this evaluation; they were divided into 14 different genera (Fig 3, S4 Fig). As the bacterial mechanisms involved for the inhibition of spore germination may differ from those involved in the inhibition of mycelial growth, our broad strain selection may allow us to select new potential effective strains unrevealed via the inhibition growth test on solid medium. Two genera shared the highest average GIs values of about 50%: Bacillus and Staphylococcus. They included the two most effective spore germination inhibiting bacteria, strains JM3 (Bacillus with a GIs of 94%) and JM161A (Staphylococcus with a GIs of 91%). However, a great variability level of GIs was observed in these two genera, with data ranging from 94% to 8% for Bacillus, and from 91% to 25% for Staphylococcus (S4 Fig). In other genera, except Arthrobacter and Variovorax, for which average GIs were respectively equal to 14% and 10%, the average GIs was equal to or higher than 25% (S4 Fig). Seven strains out of 47 have a GIs higher than 50% (5 Staphylococcus and 2 Bacillus).

[Figure omitted. See PDF.]

Fig 3. Spore germination inhibition activity (GIs) for the 47 strains selected after the first inhibition test, according to their rrs-based taxonomical affiliation.

The phylogenetic tree was inferred with the Neighbor-Joining method and GIs were reported on this tree (black bars). The red solid circles indicate bacteria that were inhibitory efficient strains (GIm > 50%) when evaluating their ability to inhibit the Fg1 mycelium growth in the first screening test.

https://doi.org/10.1371/journal.pone.0225655.g003

For the 47 strains, the results of their GIm and GIs values were combined (Fig 4) and grouped into four distinct classes: class 1 contains strains with GIm and GIs higher than 50%, classes 2 and 3 respectively contain strains with a GIm higher than 50% and a GIs lower than 50% and vice-versa, and class 4 GIm and GIs lower than 50%. Fourteen strains belonging to each of these classes (1 strain in class 1 among 5, 8 in class 2 among 28, 1 in class 2 among 2 and 4 in class 4 among 12) were then selected for the in vivo pot experiments based on their Fg1 inhibition activities and taxonomic affiliation. These 14 strains belong to 7 distinct genera: Pseudomonas (n = 5; strains JM131, JM154D, JM42B, JM62, F113), Bacillus (n = 3; strains AB42, JM79, JM3), Staphylococcus (n = 1; strain JM134), Microbacterium (n = 1; strain JM147B), Arthrobacter (n = 2; strains JM152, JM154C), Agrobacterium (n = 1; strain JM20), and Variovorax (n = 1; strain JM182).

[Figure omitted. See PDF.]

Fig 4. Selection of strains for in planta experiments according to their GIm and GIs.

The solid line represents GIs equal to GIm. Bacteria more efficient for inhibiting the spore germination are located above this line, while bacteria more efficient for inhibiting the mycelium growth are located below this line. Dashed lines represent values of GIm and GIs equal to 50%. Bacteria selected for the evaluation of their plant protection activity against F. graminearum Fg1 are surrounded by dark circles.

https://doi.org/10.1371/journal.pone.0225655.g004

Wheat protection test

After 45 days of culture, 6 different plant parameters were recorded on wheat and analyzed together using PCA analysis (Fig 5A and 5B). Significant differences were observed between the negative (NC) and the positive (PC) controls with percentages of the first axis reaching 61.2% and the second axis, 17.4%. This means that the pathogen inoculation modified various measured plant parameters and significantly affected the physiology of the wheat samples. The distribution of the data is explained by two main groups of factors: plant growth parameters on one hand (i.e. length and dry weight of wheat strands, chlorophyll content, number and weight of wheat ears) and the intensity of disease symptoms (crown rot) on the other hand (Fig 5A).

[Figure omitted. See PDF.]

Fig 5. Discriminant principal component analysis (PCA) of plant growth and disease symptom parameters of wheat plants inoculated with each of the 14 bacterial strains in presence of F. graminearum Fg1.

(A) Correlation circle between recorded plant parameters. (B) PCA (first two axes) showing groups of strains (dashed line circles) inducing similar responses on plants. NC: negative control (un-inoculated wheat plants). PC: positive control (wheat plants inoculated with Fg1 only).

https://doi.org/10.1371/journal.pone.0225655.g005

The data were divided into 4 groups (Fig 5B). The first group was composed of strains JM79 and JM3 from Bacillus genus, which are close to the negative control (NC; non-inoculated plants), showing a similar growth as healthy plants (same position on PCA’s axis 1 and no significant difference for all growth traits) and reduced disease symptoms (Table 1, S5 Fig). The second group is composed of strains JM134 (Staphylococcus), JM154C (Arthrobacter), AB42 (Bacillus), JM147B (Microbacterium), JM154D, F113 and JM42B (Pseudomonas), JM182 (Variovorax), and JM20 (Agrobacterium). These strains share common data for growth parameters (axis 1) that are lower than the negative control but significantly higher than the positive control (PC; inoculated with the pathogen) (Fig 5B, Table 1, S5 Fig). For this cluster of strains, weak disease symptoms were observed. A third group was composed of strains JM131 and JM152 (Fig 5B, Table 1, S5 Fig). These strains exhibited growth parameters as high as those shared by the NC but with disease symptoms as high as the positive control. The last group included the strain JM62. Unlike the previous groups, this strain presented reduced growth (i.e. similar growth parameters as the positive control) but weak disease symptoms.

[Figure omitted. See PDF.]

Table 1. Statistical comparison of wheat growth and symptom traits between inoculation treatments.

https://doi.org/10.1371/journal.pone.0225655.t001

Discussion

F. graminearum, a telluric fungus, particularly present in contaminated plant debris, is the main causal agent of the Fusarium Head Blight disease (FHB) on small grain cereals. In our study, given our interest for protecting wheat against F. graminearum, our library was focused on bacteria growing in the rhizosphere of wheat.

The bacterial genera most commonly found in plant protection studies are the genera Pseudomonas, and Bacillus. In our study, a more global approach was implemented based on the isolation of culturable bacteria on a generalist agar 1/10-diluted TSA medium, without targeting any distinct genus. Some un-culturable bacteria present in the rhizosphere could of course not be targeted by this protocol, knowing that less than 1% of soil bacteria are considered culturable. Given that un-culturable bacteria can hardly be exploited to develop biocontrol agents, this protocol bias may be acceptable. Another bias in selecting soil bacteria on 1/10-diluted TSA media is also linked to the fact that most abundant strains in the starting sample are more likely to be isolated, as well as the most competitive strains. Obviously, a small part of the biodiversity present in our starting soil samples was explored. However, our strain library held 34 different bacterial genera, and the genera found in our study are consistent with previous studies performed on wheat rhizosphere soils [34, 35].

In particular, Bacillus and Pseudomonas were the two most abundant genera found in the library (Fig 1). This result is consistent with data recorded in the literature as these genera grow easily on growth media and are abundant in rhizosphere soils [36–38]. More surprisingly, the genus Staphylococcus is the 3rd most represented genus. Staphylococcus is well documented in the literature, but in the domain of human health, where Staphylococcus aureus is known as one species involved in severe nosocomial diseases. Given the fact that Staphylococcus strains are well-known as human commensal or pathogens, sparse information regarding the abundance of Staphylococcus populations in soil [37] and their use for agriculture purposes is available. However, several studies have shown that some Staphylococcus strains can harbor plant growth promoting properties such as IAA or siderophore production and Zn solubilization [38, 39]. The beneficial effects of certain Staphylococcus in agriculture, particularly on wheat, with significantly improved dry and wet plant biomasses, root length and shoot length were also reported [40]. The genus Staphylococcus was thus kept in our strain library to study its potential effect on F. graminearum Fg1.

The strains were then selected according to their ability to inhibit the mycelial growth of the fungus on agar plate, and to inhibit the germination of F. graminearum spore and mycelial growth in a liquid medium. Both tests were performed in order to select strains that can inhibit Fg1 during different steps of its life cycle. Indeed the pathogen can be found in soils either as mycelium or spores (sexual and asexual) in plant debris [41].

Dual-culture agar tests were commonly implemented for identifying bacteria potentially usable as biocontrol agents against fungal or bacterial plant pathogens [42]. Given the effectiveness of their implementation, dual-plate tests are suitable as a first screening step. Using dual-culture growth inhibition tests on plate and selecting a threshold GIm higher than 50%, 61 strains of the library, belonging to 14 different genera, were considered as potentially interesting biocontrol strains. These 61 strains account for 30% of the strains efficient against F. graminearum in the library, a percentage in line with different studies [42–45], where between 10 and 50% of bacteria from libraries were found efficient in combating fungal pathogens.

We have therefore been able to highlight the diversity of bacterial genera with in vitro fungal antagonistic activity. For spore inhibition microplate tests, we kept 33 strains of interest with GIm higher than 50%, including the top 30. The originality of our approach was to select, in addition, strains that did not give promising results on the dual-culture plate test. Among the 14 latter strains, GIm ranges from 0% to 48%. This choice allowed us to highlight strains which were not effective in the growth inhibition test on agar plate but which proved to be efficient to inhibit spore germination and mycelium growth in liquid medium via the production of antifungal compounds in culture supernatants. Some strains like JM134 or JM156, the latter belonging to the Staphylococcus genera, show better inhibition activity on spore germination than on mycelium growth on solid medium. These results highlight that different antagonism mechanisms exist in bacteria to achieve the inhibition of mycelium and spore germination of F. graminearum. It demonstrates that less studied antagonist effects targeting spore germination might be a promising way to discover new bacterial mechanisms and molecules involved in F. graminearum inhibition. Recently, it has been shown that phenazines produced by some Pseudomonas PGPR can directly affect the activity of histone acetyltransferases in F. graminearum leading to deregulation of histone acetylation, and consequently, impairment of its growth and virulence [46]. Indeed, the original microplate test developed in the present work allows us to highlight the potential role of other secondary metabolites released by strains in supernatant culture, inhibiting the spore germination or liquid growth of F. graminearum. Of the 47 selected bacteria tested in microplate, 7 strains obtained GIs greater than 50%. These strains belong to only 2 genera: Bacillus and Staphylococcus. The secondary metabolites of Bacillus are well described with several classes of lipopeptides known to inhibit various plant pathogens [47]. Due to the lack of studies on the use of Staphylococcus genus for bioinoculant purposes, very few metabolites are known in the plant-Staphylococcus interaction. Some studies have already explored the phytostimulation effects of the Staphylococcus genus, evidencing their ability to synthetize auxin, produce siderophores or to solubilize phosphorus [39, 40], but to our knowledge, no study has reported any plant protection potential against soil-borne fungal pathogens of cereals.

In our wheat protection experiment, strains were selected to maximize the diversity of the strains at the taxonomic level and the diversity of their action mechanisms, through a selection of strains effective in the two in vitro tests or, on the contrary, with weak in vitro activities. In addition to testing Pseudomonas and Bacillus strains, we thus kept for the in planta test other genera like Staphylococcus but also Arthrobacter, Variovorax or Microbacterium. These have been included in some plant experiments to evaluate their plant growth stimulation effects, but in few plant protection assays [48–50]. In order to offer a wider panel of biocontrol products to farmers, focusing on other bacterial genera than Bacillus and Pseudomonas, could allow us to obtain strains that may have other antagonistic modes of action than the well-known synthesis of cyclic lipopeptides and polyketides [12, 47].

The in planta experiments in greenhouse have shown that the strains expressing the best activities in vitro are not always the strains showing the best results in vivo and vice versa, as recently reported by Comby et al. [51]. Indeed, the Bacillus sp. strain JM79 gave the best plant growth stimulation and plant protection results, but very limited antagonistic activity in vitro.

While most plant protection studies aim to select bacteria with high fungal antagonistic potential in vitro, our study highlights that it is not a sine qua non condition for a successful plant protective activity against fungal pathogen. Furthermore, other mechanisms than the ability to inhibit the growth of the pathogen can contribute in planta to a better resistance of the plant to the disease. First, some bacteria are able to elicit plant induced systemic resistance (ISR) [52, 14]. Fluorescent Pseudomonas and Bacillus PGPR are well known as being able to induce ISR responses by producing secondary metabolites like 2,4-diacetlyphloroglucinol, volatile organic compounds or cyclic lipopeptides, and as a way to protect the plant, including wheat [47, 52–54]. Through the induction of still imprecisely described hormonal regulation networks, PGPR bacteria can elicit plant defense responses fast when a pathogen attacks the plant [55]. These faster defense responses will lead to a better resistance of the plant to its aggressor. In addition, some PGPR are known to enhance the bioavailability of essential elements to plant, such as nitrogen, phosphorus, potassium, but also iron and other rare but essential mineral elements [11]. When a plant is not suffering from nutrition limitation, it can express its development program faster, allowing its best resistance to pathogen attack, especially at key steps like seed germination and seedling growth [56]. In addition, PGPR can also play a role in resistance to abiotic stress, such as drought, tolerance to salt, heavy metal [57, 58]. Abiotic stresses can make the plant more vulnerable to pathogen attack because biotic and abiotic stresses simultaneously induce various plant responses that can have negative impacts on the growth [59]. Nevertheless, some studies highlight the positive effect of some abiotic stress in resistance to biotic stress through elicitation of some joint metabolic pathways [60–62]. The complexity of soil interactions between prokaryote and eukaryote organisms makes selection of effective strains difficult and tricky. These considerations underline the importance to broaden studied models and scales of studies. Combining bacteria with different modes of actions in inoculant consortia might be an interesting way to circumvent the difficulty of finding strains with good efficacy whatever the tested soils, crops and agrosystems [63]. The plant microbiome itself could contribute to plant health, particularly through the modulation of ISR responses, the direct inhibition of pathogen growth in the rhizosphere such as that occurring in disease-suppressive soils [64]. Naturally stimulating the biocontrol activities of this beneficial microbiome through the use of appropriate crop accessions and suitable cultural practices is a way forward.

Our study highlighted various ways to achieve plant protection with biocontrol bacteria. Interestingly, antagonist interactions in vitro may not always lead to successful protection in planta because there are a lot of biotic interactions in soil and with the plant that can lead to biocontrol agent failure [65, 66].

Conclusion

Our work was performed in order to select microorganisms that can be used as biological control agents against Fusarium seedling blight. By conventional methods targeting cultivable microorganisms, we have demonstrated that combining in vitro and in planta protocols targeting different steps of the fungal pathogen life cycle (spore germination, mycelial growth and plant invasion) enabled us to refine the selection of effective biocontrol agents against F. graminearum. However, one should be aware that the in vitro selection of biocontrol agents based on fungal growth inhibition dual-plate tests can set apart potentially interesting biocontrol strains.

Supporting information

[Figure omitted. See PDF.]

S1 Fig. GIm (%) description.

A: description of the GIm used to define the efficacy of a strain against the mycelial growth of F. graminearum on PDA medium. For each strain, including the fungal control, an index I is calculated and used for calculating the mycelium growth inhibition on solid medium GIm. B: photographs illustrating 4 different levels of GIm.

https://doi.org/10.1371/journal.pone.0225655.s001

(TIF)

S2 Fig. Box plots of GIm (%) of isolates depending on the bacterial genus they belong to.

The numbers in parenthesis refer to the number of isolates belonging to the indicated genera in the whole library. Only genera with more than 2 isolates were presented. Gray diamonds represent mean values for each genus. The bacterial genera were classified according to the class taxonomic level.

https://doi.org/10.1371/journal.pone.0225655.s002

(TIF)

S3 Fig. Box plots of GIm (%) of isolates depending on their genus and their soil origin.

Only genera with more than 2 isolates were presented. Gray diamonds represent mean values for each genus.

https://doi.org/10.1371/journal.pone.0225655.s003

(TIF)

S4 Fig. Box plots of GIs (%) of 47 isolates depending on the bacterial genus they belong to.

The numbers in parenthesis refer to the number of isolates per genus that were tested in the fungal growth inhibition test in liquid medium. Gray diamonds represent mean values for each genus.

https://doi.org/10.1371/journal.pone.0225655.s004

(TIF)

S5 Fig. Box plots of values for the different parameters recorded during the wheat protection experiments for the 14 tested strains.

NC = Negative control. PC = Positive control. The measured parameters are: (A) the chlorophyll content expressed in arbitrary unit, (B) dry weight of shoot expressed in g, (C) number of ears by plant, (D) dry weight of ears expressed in g, (E) length of wheat strands expressed in cm, (F) symptoms expressed in arbitrary unit (according to the scale described in the Material and Method section). Gray diamonds represent means.

https://doi.org/10.1371/journal.pone.0225655.s005

(TIF)

Acknowledgments

Yoann-Besset Manzoni’s work was supported by a CIFRE Ph.D grant from the Association Nationale de la Recherche et de la Technologie. We also thank the Auvergne Rhône Alpes region for funding the DOPELARA project. The authors would like to thank Richard Blanc, Florence Cambon and Aurélie Wauquier for their technical support during greenhouse experiments. This work made use greenhouses at the GDEC INRA joint research Unit and technical devices from the ‘Serre’ platform of FR41 at Université Lyon 1.

Citation: Besset-Manzoni Y, Joly P, Brutel A, Gerin F, Soudière O, Langin T, et al. (2019) Does in vitro selection of biocontrol agents guarantee success in planta? A study case of wheat protection against Fusarium seedling blight by soil bacteria. PLoS ONE 14(12): e0225655. https://doi.org/10.1371/journal.pone.0225655

References

1. Aktar W, Sengupta D, Chowdhury A. Impact of pesticides use in agriculture: their benefits and hazards. Interdiscip Toxicol. 2009;2: 1–12. pmid:21217838

2. Cooper J, Dobson H. The benefits of pesticides to mankind and the environment. Crop Prot. 2007;26: 1337–1348. https://doi.org/10.1016/j.cropro.2007.03.022

3. Carson R. Silent Spring. Houghton Mifflin Harcourt; 2002

4. Köhler H-R, Triebskorn R. Wildlife ecotoxicology of pesticides: can we track effects to the population level and beyond? Science 2013;341: 759–765. https://doi.org/10.1126/science.1237591 pmid:23950533

5. Pimentel D. Amounts of pesticides reaching target pests: Environmental impacts and ethics. J Agric Environ Ethics 1995;8: 17–29. https://doi.org/10.1007/BF02286399

6. Shao H, Zhang Y. Non-target effects on soil microbial parameters of the synthetic pesticide carbendazim with the biopesticides cantharidin and norcantharidin. Sci Rep. 2017;7: 5521. https://doi.org/10.1038/s41598-017-05923-8 pmid:28717209

7. Thrall PH, Oakeshott JG, Fitt G, Southerton S, Burdon JJ, Sheppard A, et al. Evolution in agriculture: the application of evolutionary approaches to the management of biotic interactions in agro-ecosystems. Evol. Appl. 2011. 4, 200–215. https://doi.org/10.1111/j.1752-4571.2010.00179.x pmid:25567968

8. Levitan L, Merwin I, Kovach J. Assessing the relative environmental impacts of agricultural pesticides: the quest for a holistic method. Agric Ecosyst Environ. 1995;55: 153–168. https://doi.org/10.1016/0167-8809(95)00622-Y

9. Salvia MV, Ben Jrad A, Raviglione D, Zhou Y, Bertrand C. Environmental Metabolic Footprinting (EMF) vs. half-life: a new and integrative proxy for the discrimination between control and pesticides exposed sediments in order to further characterise pesticides' environmental impact. Environ Sci Pollut Res Int. 2018;25(30): 29841–29847. pmid:28660511

10. Avis TJ, Gravel V, Antoun H, Tweddell RJ. Multifaceted beneficial effects of rhizosphere microorganisms on plant health and productivity. Soil Biol Biochem. 2008;40: 1733–1740. https://doi.org/10.1016/j.soilbio.2008.02.013

11. Vacheron J, Desbrosses G, Bouffaud M-L, Touraine B, Moënne-Loccoz Y, Muller D, et al. Plant growth-promoting rhizobacteria and root system functioning. Front Plant Sci. 2013;4: 356. https://doi.org/10.3389/fpls.2013.00356 pmid:24062756

12. Caulier S, Nannan C, Gillis A, Licciardi F, Bragard C, Mahillon J. Overview of the antimicrobial compounds produced by members of the Bacillus subtilis group. Front Microbiol. 2019;10: 302. https://doi.org/10.3389/fmicb.2019.00302. pmid:30873135

13. Jacoby RP, Kopriva S. Metabolic niches in the rhizosphere microbiome: new tools and approaches to analyse metabolic mechanisms of plant-microbe nutrient exchange. J Exp Bot. 2019;70(4): 1087–1094. https://doi.org/10.1093/jxb/ery438. pmid:30576534

14. Van Loon LC, Bakker P, Pieterse CMJ. Systemic resistance induced by rhizosphere bacteria. Annu Rev Phytopathol. 1998;36: 453–483. pmid:15012509

15. McMullen M, Bergstrom G, De Wolf E, Dill-Macky R, Hershman D, Shaner G, et al. unified effort to fight an enemy of wheat and barley: Fusarium head blight. Plant Dis. 2012;96: 1712–1728. http://dx.doi.org/10.1094/PDIS-03-12-0291-FE pmid:30727259

16. D’mello JF, Macdonald AM, Postel D, Dijksma WT, Dujardin A, Placinta CM. Pesticide use and mycotoxin production in Fusarium and Aspergillus phytopathogens. Eur J Plant Pathol. 1998;104: 741–751.

17. Mesterhazy A, Bartok T, Lamper C. Influence of wheat cultivar, species of Fusarium, and isolate aggressiveness on the efficacy of fungicides for control of Fusarium head blight. Plant Dis. 2003;87: 1107–1115. https://doi.org/10.1094/PDIS.2003.87.9.1107 pmid:30812826

18. Milus EA, Parsons CE. Evaluation of foliar fungicides for controlling Fusarium head blight of wheat. Plant Dis. 1994;78: 697–699.

19. Dal Bello GM, Monaco CI, Simon MR. Biological control of seedling blight of wheat caused by Fusarium graminearum with beneficial rhizosphere microorganisms. World J Microbiol Biotechnol. 2002;18: 627–636.

20. Nourozian J, Etebarian HR, Khodakaramian G, et al. Biological control of Fusarium graminearum on wheat by antagonistic bacteria. Songklanakarin J Sci Technol. 2006;28: 29–38.

21. Pan D, Mionetto A, Tiscornia S, Bettucci L. Endophytic bacteria from wheat grain as biocontrol agents of Fusarium graminearum and deoxynivalenol production in wheat. Mycotoxin Res. 2015;31: 137–143. https://doi.org/10.1007/s12550-015-0224-8 pmid:25956808

22. Schisler DA, Khan NI, Boehm MJ, Slininger PJ. Greenhouse and field evaluation of biological control of Fusarium head blight on durum wheat. Plant Dis. 2002;86: 1350–1356. pmid:30818440

23. Zhao Y, Selvaraj JN, Xing F, Zhou L, Wang Y, Song H, et al. Antagonistic action of Bacillus subtilis strain SG6 on Fusarium graminearum. PLOS ONE. 2014;9: e92486. pmid:24651513

24. Khan MR, Doohan FM. Bacterium-mediated control of Fusarium head blight disease of wheat and barley and associated mycotoxin contamination of grain. Biol Control. 2009;48: 42–47. https://doi.org/10.1016/j.biocontrol.2008.08.015

25. Palazzini JM, Ramirez ML, Torres AM, Chulze SN. Potential biocontrol agents for Fusarium head blight and deoxynivalenol production in wheat. Crop Prot. 2007;26: 1702–1710. https://doi.org/10.1016/j.cropro.2007.03.004

26. El Zemrany H, Cortet J, Peter Lutz M, Chabert A, Baudoin E, Haurat J, et al. Field survival of the phytostimulator Azospirillum lipoferum CRT1 and functional impact on maize crop, biodegradation of crop residues, and soil faunal indicators in a context of decreasing nitrogen fertilisation. Soil Biol Biochem. 2006;38: 1712–1726. https://doi.org/10.1016/j.soilbio.2005.11.025

27. Joly P, Bonnemoy F, Besse-Hoggan P, Perrière F, Crouzet O, Cheviron N, et al. Responses of Limagne “Clay/Organic Matter-Rich” soil microbial communities to realistic formulated herbicide mixtures, including S-Metolachlor, Mesotrione, and Nicosulfuron. Water Air Soil Pollut. 2015;226: 413. https://doi.org/10.1007/s11270-015-2683-0

28. Bachate SP, Cavalca L, Andreoni V. Arsenic-resistant bacteria isolated from agricultural soils of Bangladesh and characterization of arsenate-reducing strains. J Appl Microbiol. 2009;107: 145–156. https://doi.org/10.1111/j.1365-2672.2009.04188.x pmid:19291237

29. Edwards U, Rogall T, Blöcker H, Emde M, Böttger EC. Isolation and direct complete nucleotide determination of entire genes. Characterization of a gene coding for 16S ribosomal RNA. Nucleic Acids Res. 1989;17: 7843–7853. pmid:2798131

30. Edgar RC. MUSCLE: multiple sequence alignment with high accuracy and high throughput. Nucleic Acids Res. 2004;32: 1792–1797. https://doi.org/10.1093/nar/gkh340 pmid:15034147

31. Letunic I, Bork P. Interactive tree of life (iTOL) v3: an online tool for the display and annotation of phylogenetic and other trees. Nucleic Acids Res. 2016;44: W242–W245. https://doi.org/10.1093/nar/gkw290 pmid:27095192

32. Shi C, Yan P, Li J, Wu H, Li Q, Guan S. Biocontrol of Fusarium graminearum growth and deoxynivalenol production in wheat kernels with bacterial antagonists. Int J Environ Res Public Health 2014;11: 1094–1105. https://doi.org/10.3390/ijerph110101094 pmid:24441510

33. Bell S, Alford R, Garland S, Padilla G, Thomas A. Screening bacterial metabolites for inhibitory effects against Batrachochytrium dendrobatidis using a spectrophotometric assay. Dis Aquat Organ. 2013;103: 77–85. https://doi.org/10.3354/dao02560 pmid:23482387

34. Germida JJ, Siciliano SD, Renato de Freitas J, Seib AM. Diversity of root-associated bacteria associated with field-grown canola (Brassica napus L.) and wheat (Triticum aestivum L.). FEMS Microbiol Ecol. 1998;26: 43–50. https://doi.org/10.1016/S0168-6496(98)00020-8

35. Yin C, Hulbert SH, Schroeder KL, Mavrodi O, Mavrodi D, Dhingra A, et al. Role of bacterial communities in the natural suppression of Rhizoctonia solani bare patch disease of wheat (Triticum aestivum L.). Appl Environ Microbiol. 2013;79: 7428–7438. https://doi.org/10.1128/AEM.01610-13 pmid:24056471

36. Janssen PH. Identifying the dominant soil bacterial taxa in libraries of 16S rRNA and 16S rRNA genes. Appl Environ Microbiol. 2006;72: 1719–1728. https://doi.org/10.1128/AEM.72.3.1719-1728.2006 pmid:16517615

37. Joshi P, Bhatt AB. Diversity and function of plant growth promoting rhizobacteria associated with wheat rhizosphere in North Himalayan region. Int J Env Sci. 2011;1: 1135–1143.

38. Verma P, Yadav AN, Khannam KS, Kumar S, Saxena AK, Suman A. Molecular diversity and multifarious plant growth promoting attributes of Bacilli associated with wheat (Triticum aestivum L.) rhizosphere from six diverse agro-ecological zones of India. J Basic Microbiol. 2016;56: 44–58. https://doi.org/10.1002/jobm.201500459 pmid:26567901

39. Alibrandi P, Cardinale M, Rahman MM, Strati F, Ciná P, de Viana ML, et al. The seed endosphere of Anadenanthera colubrina is inhabited by a complex microbiota, including Methylobacterium spp. and Staphylococcus spp. with potential plant-growth promoting activities. Plant Soil 2018;422(1–2): 81–99.

40. Sagar S, Dwivedi A, Yadav S, Tripathi M, Kaistha SD. Hexavalent chromium reduction and plant growth promotion by Staphylococcus arlettae strain Cr11. Chemosphere 2012;86: 847–852. https://doi.org/10.1016/j.chemosphere.2011.11.031 pmid:22169713

41. Goswami RS, Kistler HC. Heading for disaster: Fusarium graminearum on cereal crops. Mol Plant Pathol. 2004;5: 515–525. https://doi.org/10.1111/j.1364-3703.2004.00252.x pmid:20565626

42. Paulitz TC, Zhou T, Rankin L. Selection of rhizosphere bacteria for biological control of Pythium aphanidermatum on hydroponically grown cucumber. Biol. Control 1992;2: 226–237. https://doi.org/10.1016/1049-9644(92)90063-J

43. Adhikari TB, Joseph CM, Yang G, Phillips DA, Nelson LM. Evaluation of bacteria isolated from rice for plant growth promotion and biological control of seedling disease of rice. Can J Microbiol. 2001;47: 916–924. https://doi.org/10.1139/cjm-47-10-916 pmid:11718545

44. Sadfi N, Chérif M, Fliss I, Boudabbous A, Antoun H. Evaluation of bacterial isolates from salty soils and Bacillus thuringiensis strains for the biocontrol of Fusarium dry rot of potato tubers. J Plant Pathol. 2001;83: 101–117. http://dx.doi.org/10.4454/jpp.v83i2.1118

45. Tjamos EC, Tsitsigiannis DI, Tjamos SE, Antoniou PP, Katinakis P. Selection and screening of endorhizosphere bacteria from solarized soils as biocontrol agents against Verticillium dahliae of solanaceous hosts. Eur J Plant Pathol. 2004;110: 35–44. https://doi.org/10.1023/B:EJPP.0000010132.91241.cb

46. Chen Y, Wang J, Yang N, Wen Z, Sun X, Chai Y, et al. Wheat microbiome bacteria can reduce virulence of a plant pathogenic fungus by altering histone acetylation. Nat Commun. 2018;9(1): 3429. pmid:30143616

47. Ongena M, Jacques P. Bacillus lipopeptides: versatile weapons for plant disease biocontrol. Trends Microbiol. 2008;16: 115–125. https://doi.org/10.1016/j.tim.2007.12.009. pmid:18289856

48. Barriuso J, Solano BR, Gutiérrez Mañero FJ. Protection against pathogen and salt stress by four plant growth-promoting rhizobacteria isolated from Pinus sp. on Arabidopsis thaliana. Phytopathol. 2008; 98(6): 666–672. https://doi.org/10.1094/PHYTO-98-6-0666.

49. Chen L, Dodd IC, Theobald JC, Belimov AA, Davies WJ. The rhizobacterium Variovorax paradoxus 5C-2, containing ACC deaminase, promotes growth and development of Arabidopsis thaliana via an ethylene-dependent pathway. J Exp Bot. 2013;64(6): 1565–73. https://doi.org/10.1093/jxb/ert031. pmid:23404897

50. Child R, Miller CD, Liang Y, Narasimham G, Chatterton J, Harrison P, et al. Polycyclic aromatic hydrocarbon-degrading Mycobacterium isolates: their association with plant roots. Appl Microbiol Biotechnol. 2007;75(3): 655–663. pmid:17256117

51. Comby M, Gacoin M, Robineau M, Rabenoelina F, Ptas S, Dupont J, et al. Screening of wheat endophytes as biological control agents against Fusarium head blight using two different in vitro tests. Microbiol Res. 2017;202: 11–20. pmid:28647118

52. Conrath U, Thulke O, Katz V, Schwindling S, Kohler A. Priming as a mechanism in induced systemic resistance of plants. Eur J Plant Pathol. 2001;107: 113–119. https://doi.org/10.1023/A:1008768516313

53. Iavicoli A, Boutet E, Buchala A, Métraux J-P. Induced systemic resistance in Arabidopsis thaliana in response to root inoculation with Pseudomonas fluorescens CHA0. Mol. Plant-Microbe Interact. 2003;16: 851–858. https://doi.org/10.1094/MPMI.2003.16.10.851 pmid:14558686

54. Choudhary DK, Johri BN, Prakash A. Volatiles as priming agents that initiate plant growth and defence responses. Curr Sci. 2008;94: 595–604. https://www.jstor.org/stable/24100299

55. Chagas FO, Pessotti R de C, Caraballo-Rodríguez AM, Pupo MT. Chemical signaling involved in plant-microbe interactions. Chem Soc Rev. 2018;47(5): 1652–1704. pmid:29218336

56. Mengiste T, Chen X, Salmeron J, Dietrich R. The BOTRYTIS SUSCEPTIBLE1 gene encodes an R2R3MYB transcription factor protein that is required for biotic and abiotic stress responses in Arabidopsis. Plant Cell 2003;15(11): 2551–65. pmid:14555693

57. Dimkpa C, Weinand T, Asch F. Plant–rhizobacteria interactions alleviate abiotic stress conditions. Plant Cell Environ. 2009;32: 1682–1694. https://doi.org/10.1111/j.1365-3040.2009.02028.x pmid:19671096

58. Vacheron J, Renoud S, Muller D, Babalola OO, Prigent-Combaret C. Alleviation of abiotic and biotic stresses in plants by Azospirillum. In: Cassan FD, Okon Y, Creus C, editors. Handbook for Azospirillum: Technical Issues and Protocols. Springer Berlin, Heidelberg; 2015. pp. 333–365.

59. Ramegowda V, Senthil-Kumar M. The interactive effects of simultaneous biotic and abiotic stresses on plants: Mechanistic understanding from drought and pathogen combination. J Plant Physiol. 2015;176: 47–54. https://doi.org/10.1016/j.jplph.2014.11.008 pmid:25546584

60. Mithöfer A, Schulze B, Boland W. Biotic and heavy metal stress response in plants: evidence for common signals. FEBS Lett. 2004; 566: 1–5. https://doi.org/10.1016/j.febslet.2004.04.011 pmid:15147858

61. Achuo EA, Prinsen E, Höfte M. Influence of drought, salt stress and abscisic acid on the resistance of tomato to Botrytis cinerea and Oidium neolycopersici. Plant Pathol. 2006;55: 178–186. https://doi.org/10.1111/j.1365-3059.2006.01340.x

62. Fraire-Velázquez S, Rodríguez-Guerra R, Sánchez-Calderón L. Abiotic and biotic stress response crosstalk in plants. In: Shanker A, Venkateswarlu B. Abiotic stress response in plants-physiological, biochemical and genetic perspectives. InTechOpen; 2011. pp. 1–26. https://doi.org/10.5772/23217

63. Besset-Manzoni Y, Rieusset L, Joly P, Comte G, Prigent-Combaret C. Exploiting rhizosphere microbial cooperation for developing sustainable agriculture strategies. Environ Sci Pollut Res. 2018;25(30): 29953–29970. pmid:29313197

64. Berendsen RL, Pieterse CMJ, Bakker PAHM. The rhizosphere microbiome and plant health. Trends Plant Sci. 2012;17: 478–486. https://doi.org/10.1016/j.tplants.2012.04.001 pmid:22564542

65. Knudsen IMB, Hockenhull J, Jensen DF, Gerhardson B, Hökeberg M, Tahvonen R, et al. Selection of biological control agents for controlling soil and seed-borne diseases in the field. Eur J Plant Pathol. 1997;103: 775–784.

66. Pliego C, Ramos C, de Vicente A, Cazorla FM. Screening for candidate bacterial biocontrol agents against soilborne fungal plant pathogens. Plant Soil 2011;340: 505–520. https://doi.org/10.1007/s11104-010-0615-8