Diversity and antimicrobial potential of the culturable rhizobacteria from medicinal plant Baccharis trimera Less D.C.

Abstract

Plant microbiota is usually enriched with bacteria producers of secondary metabolites and represents a valuable source of novel species and compounds. Here, we analyzed the diversity of culturable root-associated bacteria of the medicinal native plant Baccharis trimera (Carqueja) and screened promising isolates for their antimicrobial properties. The rhizobacteria were isolated from the endosphere and rhizosphere of B. trimera from Ponta Grossa and Ortigueira localities and identified by sequencing and restriction analysis of the 16S rDNA. The most promising isolates were screened for antifungal activities and the production of siderophores and biosurfactants. B. trimera presented a diverse community of rhizobacteria, constituted of 26 families and 41 genera, with a predominance of Streptomyces and Bacillus genera, followed by Paenibacillus, Staphylococcus, Methylobacterium, Rhizobium, Tardiphaga, Paraburkholderia, Burkholderia, and Pseudomonas. The more abundant genera were represented by different species, showing a high diversity of the microbiota associated to B. trimera. Some of these isolates potentially represent novel species and deserve further examination. The communities were influenced by both the edaphic properties of the sampling locations and the plant niches. Approximately one-third of the rhizobacteria exhibited antifungal activity against Sclerotinia sclerotiorum and Colletotrichum gloeosporioides, and a high proportion of isolates produced siderophores (25%) and biosurfactants (42%). The most promising isolates were members of the Streptomyces genus. The survey of B. trimera returned a diverse community of culturable rhizobacteria and identified potential candidates for the development of plant growth-promoting and protection products, reinforcing the need for more comprehensive investigations of the microbiota of Brazilian native plants and habitats.

Supplementary Information

The online version contains supplementary material available at 10.1007/s42770-022-00759-5.

Introduction

Healthy plants host a highly complex bacterial community that lives in the rhizosphere and on the surface of and inside the roots and influences the growth, development, physiology, and health of the host plants in different manners [1–3]. The root bacteria community represents a subset of soil bacterial diversity. Bacteria are attracted to the rhizosphere by organic compounds released by the root (rhizosphere effect) sustaining a high density of bacteria and influencing the microbiota composition of this habitat [4]. Some bacteria living in the rhizosphere can colonize the root surface (rhizoplane) and others, known as endophytes, can colonize the root internal tissues (root endosphere) [1]. Since only some rhizobacteria can colonize the root endosphere, the community colonizing this niche represents a subset of that colonizing the rhizosphere, and this, in turn, is a subset of the soil community, the main reservoir from where the bacteria are recruited [4]. The composition of the bacterial community associated with rhizosphere and root endosphere is shaped by a combination of biotic (species, age, developmental stage, and health of plant host) and abiotic factors (soil quality, climate conditions) with the substantial influence of soil type and plant species on the root bacteria community [1, 3, 5].

Plant-associated bacteria affect the plant nutritional status and its tolerance to abiotic and biotic stresses, resulting in improvement in growth, development, health, and productivity of the plant [1, 3, 4, 6]. Living in the rhizosphere and root interface requires certain characteristics from the microbiota, since the rhizosphere is a very competitive habitat where complex microbial interactions, including competition, antagonism, symbiosis, and microbial communication, take place. The competitive interactions and communication with plant hosts and with other members of plant-associated microbiota involve, among other processes, the exchange of distinct secondary metabolites. Thus, the plant-associated microbiota is usually rich in microorganisms with functional capabilities to produce a vast array of secondary metabolites, including antibiotics, toxins, siderophores, and biosurfactants which can influence the establishment of the plant-bacteria interaction [7–9].

The production of siderophores and biosurfactants is a trait observed in many plant-associated bacteria and the antimicrobial properties presented by many of these compounds makes them of great interest in both medical and agricultural fields. Siderophores are iron-chelating agents produced by many bacteria, including plant-associated bacteria, such as Pseudomonas, Azotobacter, Bacillus, Enterobacter, Serratia, Azospirillum, and Rhizobium. Siderophores have potential roles and applications in agricultural areas for controlling phytopathogens, enhancing plant growth, and helping in the bioremediation of metal contaminated soil [10, 11]. In the medical field, siderophores can be used to form complexes with antibiotics and help in the selective delivery of antibiotics to the antibiotic-resistant bacteria (“Trojan horse strategy”) and in the medical treatment of malaria, cancer, and iron overload diseases [12].

Biosurfactants are microbial derived amphipathic compounds, such as glycolipids, phospholipids, lipopeptides, lipoproteins, and lipopolysaccharides, with tensoactive and emulsifying activities [13]. Production of biosurfactants is a common characteristic in plant-associated bacteria, such as Pseudomonas, Bacillus, and Rhizobium species. Biosurfactants affect important bacterial processes, such as surface motility, biofilm formation, and colonization, which determine the efficiency and success of plant − bacteria interactions [14, 15]. Many biosurfactants present antibacterial, antifungal, antiviral, and antibiofilm properties and are widely used in the pharmaceutical, food, cosmetic, textile, oil, and agricultural industries [16–18]. Considering the potential of plant-associated bacteria to produce biosurfactants and siderophores and/or other compounds with antimicrobial activities, as well as the medical and agricultural relevance of these compounds, the search for plant microbiota must be directed and reinforced to find promising and relevant microorganisms with these abilities.

Endophytic bacteria of medicinal plants serve as an important component of biodiversity and as a promising source of antimicrobial compounds with wide biotechnological applications in the agricultural and pharmaceutical fields [19–21]. The antimicrobial compounds produced by endophytes protect the host plant against attack from pathogens and pests and increase their tolerance to stresses [22]. Among endophytic bacteria, Streptomyces is one of the richest sources of antibiotics [23, 24]. Pseudomonas, Bacillus, Serratia, Azospirillum, Burkholderia, and Azoarcus genera also stand out as good producers of antimicrobial substances [25]. Streptomyces spp. and their metabolites may have great potential as excellent agents for controlling various fungal and bacterial phytopathogens. Additionally, their ability to promote plant growth has been demonstrated thus expanding the possibilities of using of these bacteria as biofertilizers to increase plant productivity [26]. Despite this, endophytic bacteria of medicinal plants are still relatively poorly investigated and only a few plant species have ever been completely studied [7, 8, 23, 25, 27]. Considering the existence of nearly 300,000 plant species and that each plant can host many endophytes, there is a large opportunity to find new and interesting endophytic microorganisms among myriads of plants in different settings and ecosystems [27].

The Brazilian flora is among the most diverse globally, with a high proportion of endemic plants, many of them with medicinal properties. Although studies of medicinal plant microbiota are increasing, there are only a few studies with native Brazilian plants and most Brazilian flora still being an unexplored reservoir of new species and biological activities [28]. Baccharis trimera (Less) DC (Asteraceae) is a native Brazilian plant, known as carqueja, used to treat diabetes and hepatic and digestive disorders. Carqueja has a wide variety of medicinal properties, including hypoglycemic, hepatoprotective, digestive, antiulcer, anti-acid, antihypertensive, analgesic, anti-inflammatory, antimutagenic, antioxidant, antimicrobial, and antiprotozoal. The insecticide and repellent activities have also been described for B. trimera [29].

The natural population of carqueja usually occurs in grassland fields at high altitudes, such as those found in Campos Gerais and Serra do Cadeado in the South region of Brazil. The Campos Gerais comprises the clean fields and natural Cerrado fields located on the second plateau of Paraná State, usually covered by grass and shrubby species with high species richness and endemism [30, 31]. The high diversity and endemism found in these fields, associated with the fact that they are still very little studied, makes these habitats promising for finding new and interesting microorganisms [32]. The microbiota associated with B. trimera is still poorly known, and previous studies have focused on endophytic fungi diversity [33]. Since bacteria produce of various secondary metabolites, many of them with antimicrobial properties are abundant in plant habitats, it is interesting to evaluate the antimicrobial potential of the microbiota of B. trimera. Here, we characterized the diversity of the bacterial community associated with the roots of B. trimera and screened promising isolates for biotechnological applications of agricultural and pharmaceutical interest.

Materials and methods

Soil and plant sampling

The plant samples were collected in two distinct regions of Paraná State, Southern Brazil, in fields covered with native vegetation and no history of agricultural land use. The Ortigueira site is in a place known as “Morro da Pedra Branca” at the Serra do Cadeado (23°57′58″S 51°05′17″W), a field-forest transition zone with a mountainous relief, covered by grass, shrubs, and typical vegetation of mixed ombrophile forest. The Ponta Grossa site is located at the “Dolinas Gêmeas” in the Campos Gerais region (25°08′34″S 49°57′24″W), a mountainous area with rocky outcrops predominantly covered by grass and shrubs. The physicochemical analyses of the soil [34] are given in Online Resource 1. A total of 24 plants were sampled in all the experiment. In each site (Ortigueira and Ponta Grossa), twelve healthy plants were collected and randomly gathered to compose three biological replicates with four samples each (A1, A2, or A3). Samples were transported adequately under refrigeration in plastic bags and processed within 24 h. Brazilian rules to access genetic resources were properly followed.

Counting and isolation of bacteria

Roots (2 g) were vortexed in 50 mL of buffered peptone water (1% peptone, 0.5% NaCl, 0.35% Na₂HPO₄, 0,15% KH₂PO₄; pH 7.4) and sonicated for 30 s in an Elma® ultrasonic water bath (80 kHz, 50% of potency, 4 °C) to obtain the rhizosphere/rhizoplane suspension. The roots were then treated with chloramine-T 1% for 30 min and washed three times with sterilized distilled water. The final washing water was inoculated in potato dextrose agar medium (PDA) in triplicate to evaluate the surface sterilization efficiency [35]. The resulting suspensions were serially diluted in saline solution (1:9 mL v/v). The last three dilutions (0.1 mL) were spread on Reasoner’s 2A (R2A) agar medium [36] and potato agar medium without glucose (200 g boiled potato broth and 1.5% agar, pH 6.5), both supplemented with 100 µg.mL⁻¹ of benzimidazole fungicide. Plates were incubated at 28 °C for up to 30 days, and distinctive colonies were purified on R2A or potato agar medium, after which the cultures were preserved at − 20 °C in tryptic soy broth (TSB) with 50% glycerol. Colonies were counted at 7 days of growth and the colony-forming unit (CFU) was calculated. The statistical analysis was performed with IBM SPSs Statistics software using CFU data transformed to Log10. Data normality was evaluated by the Shapiro–Wilk test and corrected by bootstrapping (1000 bootstraps with 5% confidence interval BCa). The Levene’s test (p < 0.05) was applied to verify the homogeneity of variances. The analysis of variance (One-way ANOVA) was realized with the Welch’s F-test and the means were compared with the Games–Howell post hoc test at 5% probability.

DNA isolation and 16S rDNA PCR amplification

DNA was isolated with the phenol–chloroform method, employing mechanical agitation with glass beads [37]. The 16S rDNA was amplified using 27F (5′-AGAGTTTGATCCTGGCTCAG-3′) and 1492R (5′-TACGGYTACCTTGTTACGACTT-3′) primers [38] and the GoTaq Green Master Mix (Promega), according to the manufacturer’s instructions. Reactions were run in a thermocycler (Amplitherm TX96) at 95 °C for 2 min, followed by 30 cycles of 95 °C for 45 s, 55 °C for 15 s, and 72 °C for 2 min, and a final extension at 72 °C for 10 min. The DNA and PCR fragments were analyzed by agarose electrophoresis (1%), stained with ethidium bromide, and visualized and photographed under ultraviolet light using a Loccus L-PIX photodocumenter.

Amplified 16S rDNA restriction analysis

The 16S rDNA fragments were digested with the endonucleases HaeIII, HinfI, and MspI (Invitrogen), as recommended by the manufacturer. The fragments were separated by electrophoresis on 2% agarose gel at 4 V/cm, stained with ethidium bromide, and photographed. Combined cluster analyses were performed in Bionumerics 6.6 (Applied Mathematics), using the UPGMA (unweighted pair-group method with arithmetic means) algorithm [39] and the Dice similarity coefficient at a tolerance of 2.5%.

16S rDNA sequencing and phylogeny

The 16S rDNA fragments were checked on 1% agarose gels before cleanup with the ethanol-ammonium acetate method [37]. Sequencing was performed on an ABI 3500xL Genetic Analyzer using a BigDye® Terminator v3.1 Cycle Sequencing Kit (Applied Biosystems™) and the 27F primer (20 pmol/µL), according to the manufacturer’s instructions. The BioEdit version 7.2.6 was used to check and edit the 16S rDNA partial sequence [40]. The sequencing peak heights of the raw chromatograms were inspected using BioEdit software and then these sequences were manually edited to remove the start and end of the sequences with unreliable base calling and low peak values. Only good-quality sequences with reliable base calling indicated by peak heights in the chromatograms were considered for further analysis.

The multiple sequence alignment and phylogeny were performed on the Phylogeny.fr platform (http://www.phylogeny.fr/index.cgi)[41] and comprised the following steps. Sequences were aligned with MUSCLE (v3.8.31) configured for highest accuracy (MUSCLE with default settings) [42]. After alignment, ambiguous regions (i.e., containing gaps and/or poorly aligned) were removed with Gblocks (v0.91b) using the following parameters: minimum length of a block after gap cleaning equal to 5, minimum number of sequences for a flank position equal to 55%, maximum contiguous nonconserved positions equal to 8, gaps in final block equal to 50% [43]. The phylogenetic tree was reconstructed using the maximum likelihood method and the HKY85 substitution model implemented in the PhyML program (v3.1/3.0 aLRT) [44]. Reliability for internal branch was assessed using the approximate likelihood ratio test support (SH-Like aLRT test [45]. Graphical representation and edition of the phylogenetic tree were performed with TreeDyn (v198.3) [46, 47]. The 16S rDNA sequences were deposited in the GenBank database under accession numbers from MT233104 to MT233267.

Diversity analysis

The relative abundance (RA%) was determined at family and genera level and used to compare the rhizobacteria communities of the endosphere and rhizosphere of B. trimera from Ponta Grossa and Ortigueira. The RA% was calculated by the percentage of the number of isolates of a determined family or genus divided by the total isolates obtained. Abundant families or genera were arbitrarily defined as those with RA > 15% (25 isolates) considering the abundance of the entire data set of families (mean RA of 3.8%) or genera levels (mean RA of 2.4%). Rare families or genera were defined as those with RA 0.6 to 1.8% (1–3 isolates). Families and genera containing mean RA between 2.4 and 6.7% (4–11 isolates) were considered as those with moderate abundance.

The Shannon diversity indices and the hierarchical clustering analyses were performed in PAST 4.05 [48]. The Shannon diversities of each sample (endosphere or rhizosphere) or site (Ponta Grossa or Ortigueira) were compared by the Hutcheson’s t-test (p < 0.05) [49–51] in the PAST software. The clustering analysis was performed using the UPGMA algorithm and Bray–Curtis similarity index with 10,000 bootstraps in PAST 4.05 software [48].

Bioprospection of the isolates

The antifungal activity was tested against the phytopathogens Sclerotinia sclerotiorum and Colletotrichum gloeosporioides in PDA medium by dual culture assay [52]. S. sclerotiorum causes white mold, a relevant and severe disease of many cultivable plants such as soybean, common bean, and cotton, while C. gloeosporioides is the causal agent of anthracnose, a disease of great agronomic importance that causes pre-harvest and post-harvest losses in a variety of vegetables and fruits [53, 54]. The bacteria were inoculated 1 cm away from the petri dish edge, and fungal discs (7 mm) were placed opposite. After 7 days at 28 °C, the fungal inhibition zone was measured and classified as negative (0; no clear zone), low (1; clear zone < 5 mm), moderate (2; 6–14 mm), and high (3; > 15 mm). The phytopathogens were grown in PDA plates in the absence of the bacteria as treatment control. All assays were performed with three replicates.

The siderophores were quantified by the chrome azurol S (CAS) assay [55]. The bacteria were grown for seven days (28 °C, 120 rpm) in 24 deep-well plates with 3 mL/well of Tris-buffered T medium [56] and then centrifuged (3000 rpm, 30 min), and the resulting supernatant was homogenized with CAS reagent (1:1) in 96 well microplates. After 4 h, the absorbance at 630 nm was measured in a PowerWave HT (Biotek) microplate reader. All assays were performed with three replicates. The siderophore concentration was estimated by an EDTA calibration curve (0–800 µM/mL) and classified as negative (0; no production), low (1; < 300 µM/mL), moderate (2; 300–500 µM/mL), and high (3; > 500 µM/mL).

The production of biosurfactants was analyzed by the drop collapse test [57]. The bacteria isolates were grown for seven days (28 °C, 120 rpm) in 24 deep-well plates with 3 mL/well of 1 × Vogel’s salts liquid medium added plus 0.4% glucose. The culture was centrifuged (3000 rpm, 30 min), and supernatant (20 µL) was deposited on the surface of a thin and uniform layer of oil on the lid of a microplate. After 2 min, the drop spreading was analyzed and classified as negative (0; no spreading), low (1; +), moderate (2; + +), and high (3; +  + +).

Results

Counting and bacteria isolation

Since the CFU data did not follow normal distribution (Shapiro–Wilk = 0.872, p < 0.006) and the Levene’s test showed an absence of homogeneity of variances (Levene’s = 10.279, p < 0.001), the means were compared by the Welch’s F-test and Games–Howell test. For both sites (Ponta Grossa and Ortigueira), the number of cultivable root-associated bacteria of B. trimera was significantly lower (Games–Howell test at p < 0.05) in the endosphere (6.8 to 7.3 log₁₀ CFU/g fresh weight) than rhizosphere, which had a population about 80 × higher, ranging from 8.2 to 9.3 log₁₀ CFU/g fresh weight. The Ortigueira site had significantly lower bacteria counts (Games–Howell test at p < 0.05) than Ponta Grossa in both endosphere and rhizosphere (Fig. 1a).

Fig. 1.

Bacteria count (A) and the number of isolates (B) in the root endosphere and rhizosphere soil of B. trimera from Ponta Grossa (PG) and Ortigueira (ORT), Paraná State. PDA and R2A, potato agar, and reasoner’s 2A agar media. CFU, colony-forming unit. Data are average of three replicates transformed by Log10. Means followed by distinct letters differ statistically (Games–Howell test at p < 0.05)

After CFU counting, the agar plates were visually inspected using a colony counter magnifying lens and colonies with distinct morphologies were selected and transferred onto fresh PDA and R2A media until pure cultures of each isolate were obtained. In the end, we analyzed a total of 274 isolates from the B. trimera, mainly from the rhizosphere from the Ponta Grossa site (63.5%) (Fig. 1b). From the visual inspection of agar plates, it was possible to observe that a more diverse set of colonies developed in the R2A than in PDA medium. Approximately 68% of the total isolates were obtained from the R2A medium. These observations together suggest that the R2A was more appropriate for bacteria isolation than PDA, although this observation cannot be confirmed quantitatively.

Restriction and sequencing analyses of the 16S rDNA

The cluster analyses of 16S rDNA restriction patterns revealed the genetic diversity of 274 isolates. The set of isolates was very diverse, totaling 44 distinct groups (72–77% similarity). Most groups were well defined and distinguished the isolates at the genus level. The isolates constituted 206 distinct restriction patterns at 100% similarity level. Additional data are given in Online Resource 2.

A subset of representative isolates for each amplified 16S rDNA restriction analysis (ARDRA) group was selected for partial 16S rDNA sequencing, totalizing 164 (60%) isolates sequenced. The 164 sequences were deposited under submission numbers MT233104 to MT233267 in the GenBank database as shown in the table (Online Resource 3). The isolates belonged to 26 distinct families of the Actinobacteria, Bacilli, α-Proteobacteria, β-Proteobacteria, and γ-Proteobacteria classes (Fig. 2). The predominant families were Streptomycetaceae (22%) and Bacillaceae (16,5%), followed by Methylobacteriaceae, Paenibacillaceae, Burkholderiaceae, Bradyrhizobiaceae, Rhizobiaceae, and Pseudomonadaceae, with relative abundances (RA) moderate ranging from 4.3 to 6.7%. The families Micrococcaceae, Staphylococcaceae, Xanthomonadaceae, and Enterobacteriaceae occurred at moderate RA% of 2.4 to 3.7%. The other fourteen families occurred rarely at relative abundances of 0.6 to 1.8% and corresponded to approximately 15% of the isolates (Fig. 2).

Fig. 2.

Relative abundance (RA%) of the main genera (A) and families (B) of rhizobacteria isolated of endosphere and rhizosphere of B. trimera from Ponta Grossa and Ortigueira, Paraná-Brazil. *Rare families include families with RA of 0.6 to 1.8% as detailed in Online Resource 4

A total of 41 distinct genera were isolated from B. trimera. Streptomyces and Bacillus were the two most abundant genera, followed by Methylobacterium, Paenibacillus, Rhizobium, Paraburkholderia, Pseudomonas, Staphylococcus, Tardiphaga, and Burkholderia that occurred at moderate relative abundance (RA% 2.4 to 6.7). The other 32 genera occurred rarely (RA% of 0.6 to 1.8%) and corresponded to approximately 28% of the isolates (Fig. 2). Details of abundance and distribution of the bacterial classes, families, and genera are given in Online Resource 4.

Comparison of community analysis and diversity index

Figure 3 shows the diversity indices of the bacterial communities. The Shannon indices ranged from 3.67 to 2.83 and were higher (p < 0.01) in the rhizosphere than endosphere in both localities (Ponta Grossa and Ortigueira). The diversity index of the rhizosphere was statistically similar (p < 0.3) in both localities; however, the diversity index of the endosphere from Ortigueira was significant lower (p < 0.03) than Ponta Grossa. The rhizosphere from Ponta Grossa had a high abundance of Actinobacteria and Bacilli isolates, represented mainly by Streptomyces and Bacillus genera, respectively. Compared to Ortigueira, the bacterial community of the rhizosphere from Ponta Grossa had lesser variety of families and showed predominance of Streptomyces and high abundance of the $\beta$- and $\gamma$-Proteobacteria of the genera Paraburkholderia and Pseudomonas (Fig. 2).

Fig. 3.

Shannon diversity indices (A) and hierarchical clustering analysis (B) for culturable rhizobacteria communities of the endosphere and rhizosphere of B. trimera from Ponta Grossa and Ortigueira, Paraná-Brazil. Significant difference at p ≤ 0.05 (*) or p ≤ 0.01 (**) by Hutcheson’s t-test (p < 0.05) [49–51]. Clustering used the UPGMA algorithm and Bray–Curtis similarity index with 10,000 bootstraps

The diversity of the endosphere from Ponta Grossa was significantly higher (p < 0.03) than Ortigueira, and the bacterial composition was also distinct (Fig. 3). The endosphere from Ponta Grossa had a high abundance of Bacillus in addition to isolates of Tardiphaga, Methylobacterium, and Rhizobium genera. The endosphere from Ortigueira had a greater abundance of Paenibacillus, with five distinct species, besides Bradyrhizobium and Pantoeae genera, represented by two distinct species (Fig. 2).

The hierarchical cluster analysis showed a separation between the bacteria community from Ponta Grossa and Ortigueira, indicating that the phytogeographic conditions affected the bacterial community structure (Fig. 3). The plant niches were also a relevant factor influencing the composition of the bacterial communities since the Bray–Curtis similarity between these niches was low (< 0.2) for the two localities. These results confirm the high diversity of genera and species found in these habitats and show that this diversity was influenced by both the plant niches and phytogeographic conditions.

Taxonomic assignment and phylogenetic relationship

We analyzed the phylogenetic relationships of isolates of Actinobacteria, Proteobacteria, and Bacilli phyla and closely related strains (Online Resource 5) to determine the taxonomic classification of the isolates; however, the high similarity of 16S rDNA sequences did not allow reliable identification at species-level for most isolates. Despite this, the closest species of each isolate and their phylogenetic relationships are shown to demonstrate the diversity within each genus. The main genera and species found for each phylum are discussed below and detailed information is given in Online Resource 3 and 5.

Actinobacteria phylum

The isolates of Actinobacteria phylum (54 isolates) comprised 14 distinct genera, including 32 isolates of Streptomyces, the predominant genus, and the genera that had 3 isolates each. The other genera found were less common and include Mycobacterium, Arthrobacter, Microbacterium, Micrococcus, Kitasatospora, and Rhodococcus (with 2–3 isolates each genus) and the genera Leifsonia, Kocuria, Dactylosporangium, Nocardioides, Nocardia, and Conexibacter with only one isolate each. The similarity coefficients of Streptomyces and other isolates of Actinobacteria with their respective closest species were greater than 99%, except for BTM102, BTM382, and BTM06 isolates that had similarity score below 98.4% (Online Resource 3).

The Streptomyces isolates were positioned in twenty-one distinct phylogenetic groups; however, most of them contained more than one Streptomyces species (Online Resource 5). The isolates BTM291, BTM449, and BTM550 were similar (> 99.6%) with S. mirabilis and clustered this species and with S. olivochromogenes (> 84% SH-like support), while the isolates BTM446, BTM302, and BTM301 formed, respectively, a concise branch (> 81% SH-like support) with S. abikoensis, S. atriruber, and S. nodosus species. The high similarity between the 16S rDNA sequences did not allow clear inference at specie level for most isolates. The clusters represented by S. mirabilis, S. libani, S. nojiriensis, S. bungoensis, and S. graminisoli were the most common. The two Kitasatospora strains (BTM417 and BTM447) formed a cluster with K. atroaurantiaca and K. purpeofusca. The ML phylogenetic tree of the other isolates of the Actinobacteria phylum is detailed in Online Resource 5.

Bacilli phylum

The 44 Bacilli isolates included members of the genera Bacillus (27), Paenibacillus (10), Staphylococcus (6), and Lysinibacillus (1). The sequence similarity of Bacillus isolates ranged from 99.2 to 100% (Online Resource 3) and most isolates were identified only at genus level. The Bacillus isolates were distributed in ten distinct phylogenetic groups (Online Resource 5). The most common species were Bacillus pumilus and Bacillus safensis, which formed one large cluster (> 74% SH-like support) with eleven isolates (BTM146, BTM63, BTM104, BTM109, BTM116, BTM148, BTM182, BTM241, BTM340, BTM388). The other well represent group included Bacillus cereus and Bacillus thurigiensis species, whose strains formed a cluster (100% SH-like support) with BTM346, BTM366, BTM23, BTM26, and BTM17 isolates. The isolates BTM219, BTM331, BTM332, and BTM418 were positioned together with Bacillus flexus and Bacillus megaterium species (96% SH-like support). The BTM557 clustered with strains of Bacillus fumariole species, while the isolates BTM466, BTM50, BTM264, and BTM375 were positioned close to Bacillus subtilis, Bacillus acidiceler, Bacillus sporothermodurans, and Bacillus niacin, respectively. The BTM210 was positioned close to Bacillus senegalensis and Bacillus drentensis species, while BTM183 was positioned close to with Bacillus muralis species.

The Paenibacillus isolates had similarity coefficients ranging of 96.5 to 99.9% and were positioned in seven distinct phylogenetic clusters (Online Resource 3 and 5). BTM358, BTM119, and BTM479 had, respectively, 96.5%, 97.4%, and 98.7% similarity with Paenibacillus chartarius, Paenibacillus motobuensis, and Paenibacillus alginolyticus. These isolates clustered with their respective closest species. The BTM569 and BTM528 strains clustered with Paenibacillus terrae species (100% SH-like support) and the BTM529 strain was positioned together with Paenibacillus massiliensis species (100% SH-like support). The BTM534, BTM540, BTM479, and BTM523 strains had high similarity (> 99%) with Paenibacillus alginolyticus and were positioned close to this species in a separated branch (99% SH-like support). The BTM454 strain showed similarity to Paenibacillus agarexedens (99%) and Paenibacillus baekrokdamisoli (98.6%) and clustered with P. baekrokdamisoli (99% SH-like support).

The six Staphylococcus isolates showed high similarity coefficients (99.3 to 100%) and were positioned in three distinct clusters in the phylogenetic analysis (Online Resource 3 and 5). The BTM82, BTM117, BTM211, and BTM503 strains grouped together to S. epidermidis species (74% SH-like support), while the isolates BTM28 and BTM21 clustered (> 93% SH-like support) with S. pasteuri and S. saprophyticus, respectively.

Proteobacteria phylum

The Proteobacteria phylum included 66 isolates of 24 distinct genera. The phylogenetic tree of α-Proteobacteria isolates included the genera Methylobacterium (11), Rhizobium (8), and Tardiphaga (5) that were the most frequently isolated, as well as isolates of the Bradyrhizobium, Sphingomonas, Caulobacter, Dongia, Bosea, and Sphingobium genera that had one to three isolates each genus (Online Resource 5). The Methylobacterium isolates showed high similarity coefficients (97.8 to 99.9%) and represented seven distinct phylogenetic branches. BTM541, BTM101, BTM144, and BTM154 clustered with M. komagatae (100% SH-like support), the most common Methylobacterium species. The BTM157 and BTM145 strains showed, respectively, 97.8% and 99.3% similarity to Methylobacterium aerolatum and were positioned close to this species (85% SH-like support). The BTM78 and BTM177 strains clustered (> 73% SH-like support) with Methylobacterium aquaticum and Methylobacterium phyllosphaerae, respectively. BTM62 and BTM535 had high similarity to Methylobacterium radiotolerans (99.7%) and Methylobacterium platani (99.4%), respectively, and were positioned close to these species in the phylogenetic tree.

The eight Rhizobium isolates showed similarity coefficients greater than 98.9% and were grouped in five distinct phylogenetic branches (Online Resource 3 and 5). The isolates BTM403 and BTM405 formed a clear branch with Rhizobium grahamii (93% SH-like support), while BTM505 and BTM476 strains clustered, with R. larrymoorei and R. cellulosilyticum (> 94% SH-like support), respectively. The isolates BTM47, BTM45, and BTM385 were positioned (100% SH-like support) in a group gathering Rhizobium lusitanum, Rhizobium tropici, Rhizobium rhizogenes, and R. miluonense species. The BTM399 strain had 98.9% similarity with Rhizobium taeanense, but formed a separated branch in the phylogenetic tree (92% SH-like support). The five Tardiphaga isolates (BTM343, BTM344, BTM402, BTM408, and BTM427) had high similarity (99.3–99.9%) and clustered with distinct Tardiphaga robiniae strains in the phylogenetic tree (100% SH-like support).

The phylogenetic analysis of the β-Proteobacteria isolates (Online Resource 5) comprised mainly the genera Paraburkholderia (6), Burkholderia (4), and Variovorax (3). The other β-Proteobacteria genera (Duganella, Herbaspirillum, Chromobacterium, and Thauera) were less common with one isolate each genus. The similarity coefficients of the isolates of this class ranged of 99.3 to 100%, with exception of the BTM507 and BTM562 strains that had 97.5 and 98.8% similarity with Niveibacterium umoris and Methylibium petroleiphilum species, respectively (Online Resource 3). The Paraburkholderia isolates were positioned in three distinct phylogenetic branches. The isolates BTM456, BTM485, BTM486, and BTM561 clustered with Paraburkholderia dipogonis (95% SH-like support), while BTM317 formed a cluster with Paraburkholderia terrae strains (94% SH-like support). The Burkholderia isolates BTM02, BTM43, BTM131, and BTM27 formed a cluster with distinct strains of Burkholderia diffusa and Burkholderia ambifaria (> 78% SH-like support). The BTM559 strain had high similarity (> 99.3%) with Paraburkholderia phenoliruptrix and Burkholderia cepacia strains but was positioned closest to the Paraburkholderia graminis in the phylogenetic tree.

The phylogenetic analysis of the λ-Proteobacteria class (Online Resource 5) included members of Pseudomonas (6 isolates), Luteibacter (3), and Pantoea (2), in addition to the genera Serratia, Stenotrophomonas, Xanthomonas, and Lelliottia, that had only one isolate each. The similarity coefficients of the isolates of this class ranged of 99.6 to 100% (Online Resource 3). The Pseudomonas strain BTM481 formed a clear group with Pseudomonas alcaligenes (99% SH-like support), while the other isolates (BTM336, BTM497, BTM484, BTM519, and BTM213) were positioned in a large group that included distinct Pseudomonas species as Pseudomonas putida, Pseudomonas rhizosphaerae¸ Pseudomonas vancouverensis, and Pseudomonas fluorescens.

Bioprospection for antifungal, biosurfactant, and siderophore producing bacteria

About 30% of the isolates inhibited at some level the growth of the phytopathogen S. sclerotiorum and 26.5% of C. gloeosporioides. Most antagonist isolates are from the rhizosphere (74%). Twenty-three isolates showed a high level of inhibition against S. sclerotiorum and ten against C. gloeosporioides (Fig. 4). Most of them were members of Streptomyces and, to a lesser extent, of the genera Burkholderia, Paraburkolderia, Bacillus, Rhodococcus, or Pseudomonas.

Fig. 4.

Bioprospection of rhizobacteria of B. trimera for antifungal activity and production of siderophore and biosurfactant. The activity or production was classified as low ( +), moderate (+ +), or high (+ + +)

Siderophore production ranged from 7 to 752 µM/mL and was detected in 25% of the isolates, most of them isolated from the rhizosphere (63%). Twenty isolates produced over 500 µM/mL (high level) and comprising isolates of Streptomyces, Bacillus, Paraburkholderia, Rhizobium, Variovorax, Xanthomonas, and Staphylococcus genera (Fig. 4).

The biosurfactant occurred in 47% of the isolates, most of them obtained from the rhizosphere (61%). Among the producers, 26 isolates showed a high level of production detected by complete drop collapse (Fig. 4). These isolates are Streptomyces, Bacillus, Pseudomonas, and Paenibacillus genera (2 to 6 isolates), but at least one isolate of Burkholderia, Methylobacterium, Mycobacterium, Rhizobium, and Tardiphaga also showed high surfactant production.

The scores attributed for the antifungal, siderophore, and biosurfactant capabilities were summed (maximum score of 12), and the isolates ranked to select the most promising isolates. Table 1 shows the ranking of twenty-one isolates, in which BTM287 showed the highest score (10), followed by BTM295, BTM298, BTM299, and BTM470. Streptomyces species and isolates of Bulkholderia, Rhizobium, Bacillus, and Paraburkholderia were the top isolates.

Table 1.

Ranking of rhizobacteria based on their antifungal activities and production of siderophore and biosurfactant

Isolates	Sclero	Gloesp	Sider	Surf	Score	Rank†	Identification‡
BTM287	3	3	3	1	10	1st	Streptomyces sp.
BTM295	3	1	3	2	9	2nd	Streptomyces sp.
BTM298	3	1	2	3	9	2nd	Streptomyces sp.
BTM299	3	2	1	3	9	2nd	Streptomyces sp.
BTM470	3	3	3	0	9	2nd	Streptomyces sp.
BTM27	3	3	1	1	8	3rd	Burkholderia sp.
BTM289	2	3	3	0	8	3rd	Streptomyces sp.
BTM292	3	3	0	2	8	3rd	Streptomyces sp.
BTM120	1	1	2	3	7	4th	Streptomyces sp.
BTM294	3	2	2	0	7	4th	Streptomyces sp.
BTM02	2	1	0	3	6	5th	Burkholderia sp.
BTM20	3	2	1	0	6	5th	Streptomyces sp.
BTM43	3	3	0	0	6	5th	Burkholderia sp.
BTM47	0	0	3	3	6	5th	Rhizobium sp.
BTM210	0	0	3	3	6	5th	Bacillus soli
BTM296	2	1	3	0	6	5th	Streptomyces sp.
BTM300	2	2	0	2	6	5th	Streptomyces sp.
BTM304	3	3	0	0	6	5th	Streptomyces sp.
BTM456	3	2	0	1	6	5th	Paraburkholderia sp.
BTM463	1	2	3	0	6	5th	Streptomyces sp.
BTM471	3	3	0	0	6	5th	Streptomyces sp.

^†Ranking based on activity level as low (1), moderate (2), or high (3)

^‡Identification based on sequencing or restriction analysis of the 16S rDNA

Sclero, Sclerotinia sclerotiorum; Gloesp, Colletotrichum gloeosporioides, Sider, siderophore Surf, Biosurfactant

Discussion

Research with plant-associated bacteria has indicated that plants shelter a diverse bacterial community with different biological activities and the potential to produce different secondary metabolites [7, 8, 23, 25]. We analyzed the diversity and antimicrobial properties of rhizobacteria of the carqueja grown in Paraná State, Brazil, and identified a highly diverse community constituted of 274 isolates, belonging to three phyla, five classes, 26 families, 41 genera, and at least 113 species. To our knowledge, this is the first study providing information about the diversity of bacteria associated with B. trimera, although the endophyte fungal community has been the subject of a previous study [33]. The biodiversity of microbiota associated with Brazilian medicinal plants is still little known. A recent review study reported that approximately 54 plant species (30 distinct families) were analyzed for endophyte biodiversity in Brazil, with the most representative families demonstrating obvious agronomic and industrial importance, and only a few studies included medicinal plants [28].

The genera most frequently isolated from B. trimera rhizosphere and endosphere were Streptomyces and Bacillus, representing about 37% of all isolates, followed by Paenibacillus, Staphylococcus, Methylobacterium, Rhizobium, Tardiphaga, Paraburkholderia, Burkholderia, and Pseudomonas that occurred at moderate relative abundance (RA of 2.4 to 6.7%). Like other Actinobacteria, Streptomyces is a common genus of soil bacterium and can reside in the rhizosphere and inside plant tissues. This genus is known as the main producer of antimicrobial compounds [24]. Studies have reported that these bacteria have various properties in common with other plant-growth-promoting bacteria (PGPB) and can benefit plant growth and productivity when inoculated alone or in a consortium [24, 26, 58]. Our results showed that the Streptomyces isolates were the main antagonists of S. sclerotiorum and C. gloeosporioides and that some isolates were also good producers of biosurfactants (BTM299, BTM298, and BTM293) and siderophores (BTM287, BTM295, and BTM470) (Table 1). Medicinal plants are a reservoir of new and interesting Streptomyces species [23]. We found 32 isolates of Streptomyces sp. associated with B. trimera, which were positioned in 26 distinct branches in the phylogenetic tree. Since the 16S rRNA phylogenetic tree provided a useful framework for the relationships among species but did not always have sufficient resolution to provide definitive identification at species-level [58], the multilocus sequence analysis (MLSA) of the Streptomyces isolates is in progress in our laboratory to determine the species of these isolates and if they potentially represent new species.

The Bacilli class was also well represented in the microbiota of B. trimera with three genera (Bacillus, Paenibacillus, and Staphylococcus) among the most frequently isolated. Firmicutes, Proteobacteria, and Actinobacteria are among the most frequent phyla isolated from plants and, as we also observed, Bacillus is usually the most predominant genus [59]. Bacillus was the most frequent genus identified in 77% of articles studying endophytic bacteria of plants grown in Brazil [28].

Bacillus and Paenibacillus species can provide a broad range of benefits for plants, including preventing and controlling diseases caused by pathogens, eliciting plant resistance, and promoting plant growth [60, 61]. Due to their beneficial effects on plants and their ability to produce endospores resistant to adverse environmental conditions, interest in Bacillus-based inoculant formulations has significantly increased worldwide [62, 63]. Our Bacillus isolates showed potential to inhibit the phytopathogen S. sclerotiorum (BTM146, BTM148, and BTM241) and to produce siderophores (BTM330, BTM340, and BTM210) and biosurfactants (BTM109, BTM198, and BTM210) at high levels, while the Paenibacillus isolates stood out for the high production of biosurfactants (BTM119, BTM523, BTM528, and BTM540) and the capacity to inhibit at a moderate level the growth of phytopathogens (BTM528, BTM569, and BTM570).

The microbiota of B. trimera also has a high prevalence of known plant-beneficial bacteria of the Proteobacteria phylum, such as Methylobacterium, Rhizobium, Tardiphaga, Bradyrhizobium, Burkholderia, Paraburkholderia, and Pseudomonas. Methylobacterium is found on the plant surfaces and as an endophyte of different plant species and can promote plant growth and protection using several mechanisms [64]. In addition to playing a role in plant growth, Methylobacterium bacteria can degrade toxic compounds and tolerate high heavy metal concentrations and so can be used to decontaminate the environment [64]. The Methylobacterium isolates BTM541 and BTM157 showed moderate antagonistic activity and BTM535 had high production of the biosurfactant. We found at least seven distinct Methylobacterium species in the rhizosphere and endosphere of B. trimera, including four isolates that were closely related to those of M. komagatae, a promising rhizobacteria inoculant for oleaginous Crambe plants [65].

The Rhizobium and Bradyrhizobium genera are endosymbionts usually isolated from root nodules of leguminous plants, while Tardiphaga is a new genus in the Bradyrhizobiaceae family described as a nodule symbiont of Robinia pseudoacacia and Vavilovia formosa [66, 67]. Here, we identified five Rhizobium, three Bradyhizobium, and five Tardiphaga strains closely related to T. robiniae LMG 26467 T. These rhizobia were mainly isolated from the root endosphere of B. trimera. Rhizobium and Bradyhizobium species are endophytes of non-legume hosts such as sugarcane, maize, and wheat [68–70], and novel species of endophyte rhizobia have been described [71, 72]. Due to their ability to provide nitrogen fixed to legume hosts, symbiotic species such as B. japonicum and R. tropici have high economic relevance for soybean and common bean crops. There is yet little knowledge about the Tardiphaga genus. The Rhizobium strain BTM47 showed high capacity to produce siderophores and biosurfactants, while the Tardiphaga isolates BTM343, BTM344, BTM402, and BTM427 stood out to produce biosurfactants. As crops, medicinal plants can benefit from interactions with PGPB [73], and further investigations of the role of these rhizobia endophytes in the growth and health of B. trimera are needed.

The other more prevalent genera associated with B. trimera were Burkholderia, Paraburkholderia, and Pseudomonas. The members of Paraburkholderia and Burkholderia genera occur commonly in soil, water, and in association with plant, fungi, animal, and humans [74, 75]. The Burkholderia genus includes both plant-beneficial and phytopathogenic species, and some species of Burkholderia have demonstrated some opportunistic infection to animal and human. Most of plant-beneficial Burkholderia species have demonstrated promising biocontrol action against different phytopathogens and significant biotechnological potential as a source of novel antibiotics and bioactive secondary metabolites [75]. The Paraburkholderia genus include many beneficial species with potential use in the plant protection and environmental bioremediation [74, 75]. The Paraburkholderia genus is prevalent in edaphic adverse conditions, such as those found in the “Campos Gerais” region from Paraná State, characterized by nutrient-poor and acidic soils with high levels of aluminum [31]. The Paraburkholderia strains were obtained predominantly from B. trimera grown in the Ortigueira and are closely related to P. phytofirmans, a beneficial bacterium that promotes the growth and health of various plant species even under stressful conditions [76]. The Paraburkholderia isolate BTM317 produced siderophores at a high level, while Paraburkholderia isolate BTM456 and Burkholderia isolates BTM02, BTM27, BTM43, and BTM131 stood out for the antagonistic activity, inhibiting the growth of phytopathogens at a high level.

We also identified six distinct Pseudomonas, including isolates that were closely related to those of fluorescent species P. migulae, P. corrugate, and P. aeruginosa. Many fluorescent Pseudomonas species are plant growth-promoting rhizobacteria (PGPR), improving plant health and fitness through a wide range of plant-beneficial traits, including the production of antifungal metabolites. These properties, associated with their high catabolic adaptability and root-colonizing abilities, make these bacteria suitable for biotechnological application in agriculture, especially as an attractive biocontrol agent [77]. In our analyses, the Pseudomonas isolates BTM519 and BTM336 were good producers of biosurfactants and siderophores, while BTM481 had antifungal activity against S. sclerotiorum.

Besides the ten most predominant genera discussed above, B. trimera habitats sheltered another 31 bacterial genera (almost one-third of total isolates) that occurred at low relative abundances. These data show how suitable these habitats are and that more efforts are needed to cultivate these bacteria to have a more representative view of the bacterial community associated with B. trimera. Here, poor culture media (mainly the R2A agar) associated with the long incubation time and the sampling in localities with different soil and phytogeographic properties provided a favorable setting for successful isolation of a diverse bacterial collection. In another study, the successful recovery of uncultured and novel bacteria isolates from forest soil was achieved using a diffusion bioreactor associated with a prolonged incubation period, low substrate-modified media, and sampling in summer [78]. Considering that conventional techniques recover only a tiny part of the bacteria community, new and distinct cultivation strategies are necessary to reveal a still hidden bacterial community and discover their potential biotechnological applications [79].

The diversity analyses showed that the bacterial community associated with B. trimera is highly diverse, especially in the rhizosphere samples with higher Shannon diversity indices than the endosphere niches for both localities. In addition, the analyses suggest that the composition of the bacteria community associated with B. trimera was influenced by both plant niches and phytogeographic conditions. Our results agree with other previous studies that show that a combination of biotic (species, age, developmental stage, and health of plant host) and abiotic factors (soil quality, climate conditions) shape the composition of the bacterial community associated with rhizosphere and root endosphere [1, 3, 5]. In the plant-bacteria interaction process, root-associated bacteria are attracted to the rhizosphere by the root exuded organic compounds (rhizosphere effect), which sustain a highly dense and diverse bacterial community. The host plant’s nutritional and physiological status affects the composition and quantity of root exudates, influencing the microbial community [4]. Since the edaphic is a relevant factor influencing plant physiology and nutrition, it is reasonable that the distinct edaphic and phytogeographic conditions found in Ponta Grossa and Ortigueira are a prevalent factor shaping the bacterial community associated with B. trimera. The lower diversity and density of bacteria found in the endosphere of B. trimera also supports previous findings in the literature [2, 4].

Soil and plant habitats are primary sources of bacteria for various biotechnological applications, particularly for developing products to improve plant growth and productivity and biocontrol of plant disease. Unexplored sources, such as the various medicinal and native plants species, have proven to be a precious reserve of new microbial species for the search for new products and metabolites of agricultural, pharmaceutical, and industrial interest. Here, we identified that approximately one-third of the rhizobacteria of B. trimera exhibited antifungal activity against S. sclerotiorum and C. gloeosporioides, some of them with high inhibition levels, such as the BTM287 strain. Furthermore, we identified a high proportion of isolates able to produce siderophores and biosurfactants, compounds with antimicrobial activities. The ranking based on antimicrobial properties allowed the selection of the top isolates. Additional studies are needed to determine whether the compounds produced are new molecules. The most promising isolates are members of the Streptomyces genus, revealing that the microbiota of B. trimera is a source of new and valuable isolates of this genus, as previously reported for other medicinal plant species. Further studies including plant inoculation assays with the most promising isolates are needed to confirm the potential of these isolates as plant growth-promoting and fungal biocontrol agents.

The survey of the medicinal plant B. trimera returned a diverse community of culturable rhizobacteria, some of them being potentially new species, considering the sequence identity threshold ≥ 97% established for a definition of a bacteria species [80] and the positioning of the isolates in the ML phylogenetic tree. Between these are the included Actinobacteria strains BTM102 and BTM382 that are, respectively, closely related to Conexibacter arvalis and Mycobacterium komossense species. Potential new species also included bacteria isolates of the Paenibacillus (BTM358 and BTM119) genus and of distinct genera of the Proteobacteria class as Methylobacterium (BTM175), Dongia (BTM496), and Niveibacterium (BTM507). The sequence identity of these isolates with their closest relative species was ≤ 97% (Online Resource 3). The species defining of these isolates needed to be further confirmed using methods with higher resolution at species level as Multilocus sequence analysis (MLSA) [81]. The promising biotechnological potential of some of our isolates added to the fact that some of these isolates represent potentially new species highlights and reinforces the need for more comprehensive investigations of the microbiota of Brazilian native plants and habitats.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contribution

All authors contributed to the study conception and design or the acquisition, analysis, or interpretation of data. Material preparation, data collection, and analysis were performed by Ana Camila Munis Jardim, Jéssica Ellen de Oliveira, Luana de Moura Alves, Giovana Oliveira Gutuzzo, André Luiz Martinez de Oliveira, and Elisete Pains Rodrigues. The first draft of the manuscript was written by Ana Camila Munis Jardim and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.

Funding

Partial financial support was received from the National Institute of Science and Technology, INCT-Plant-Growth Promoting Microorganisms for Agricultural Sustainability and Environmental Responsibility (CNPq 465133/2014–4, Fundação Araucária-STI 043/2019, CAPES) and the Brazilian National Council for Scientific and Technological Development (CNPq 480080/2013–5). This work was also supported by the infrastructure of the LAMM Facility from Londrina State University. Ana Camila Munis Jardim and Luana de Moura Alves received MSc scholarships from the National Council for the Improvement of Higher Education (CAPES)—Finance code 001. Jéssica Ellen de Oliveira and Giovana Oliveira Gutuzzo received a BASc scholarships from the Araucaria Foundation.

Data availability

The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.

Code availability

Not applicable.

Declarations

Ethics approval

The study does not involve any human or animal participation or data.

Consent to participate

Not applicable.

Consent for publication

Not applicable.

Conflict of interest

The authors declare no competing interests.