Isolation and characterization of indigenous endophytic bacteria associated with leaves of switchgrass (Panicum virgatum L.) cultivars



Suha Jabaji, Plant Science Department, McGill University, 21111 Lakeshore Rd., Ste-Anne-de-Bellevue, Quebec H9X 3V9, Canada. E-mail:



To isolate and characterize indigenous bacterial endophytes from cultivars of switchgrass and study their antimicrobial and growth promoting potential.

Methods and Results

The diversity, molecular and biochemical characterizations of indigenous and culturable bacterial endophytes residing in leaves of switchgrass have not been studied previously. This study describes the characterization of 31 bacterial endophytes from three switchgrass cutlivars: Cave-in Rock, Blue Jacket and Tecumseh. Molecular and phylogenetic analysis based on the 16S rRNA sequence grouped the endophytes into eight different taxa that shared high homology of 98–99% with other known sequences. Bacterial endophytes were identified as Microbacterium testaceum, Curtobacterium flaccumfaciens, Bacillus subtilis and Bacillus pumilus, Pseudomonas fluorescens, Sphingomonas parapaucimobilis, Serratia sp. and Pantoea ananatis. Some endophytes were detected in switchgrass seeds and in plants that originated from seeds collected a year earlier, confirming vertical transmission to the next generation of the host. Selected endophytes produced cellulases and were capable of solubilizing inorganic phosphorus. Analysis of cell-free culture filtrate of selected strains using direct infusion orbitrap mass spectrometry confirmed the presence of several well-characterized lipopeptide toxins and phytohormones. Re-inoculation of the roots of switchgrass seedlings with endophytes singly or combined confirmed their migration to the upper aerial parts of the plant.


Our findings suggest that switchgrass leaves harbour a diversity of bacterial endophytes, some of which could potentially be applied as growth promoting bacteria.

Significance and Impact of the Study

This is the first report on the characterization of indigenous bacterial endophytes and their potential use as biofertilizers.


The need to find sustainable alternatives to replace increasingly expensive fossil fuels and to reduce greenhouse gas emissions has driven the interest in the production of biofuels worldwide. Liquid, gaseous and solid biofuels produced from renewable biological resources such as plant biomass and treated municipal and industrial wastes are envisaged as part of the solution. A common perennial grass, switchgrass (Panicum virgatum L.), has been identified as a model herbaceous energy crop for the USA. It has a number of positive attributes such as productive long-lived perennial crop with high biomass production potential, efficient water use, relatively low demand for nutritional inputs and agrochemicals, high carbon sequestration potential, and is well adapted to marginal soils (McLaughlin and Adams Kszos 2005; Sanderson et al. 2006).

In the province of Quebec, Canada, the Quebec Energy Strategy for 2006–2015 ( accessed 8 August 2012) was established to reduce the reliance on fossil fuels and promote renewable bioenergy that could be supplied by high yielding grass crops such as switchgrass. The development of switchgrass as an energy crop can benefit Quebec's agriculture by optimizing productivity of marginal farmland or degraded lands and by serving as a new source of income for Quebec farmers. However, several production constraints still present challenges to its commercialization as a biomass feedstock, such as the rapid establishment of productive stands and achievements of greater biomass yields and more efficient use of fertilizers (Sanderson et al. 2006). Overcoming these biological constraints to improve switchgrass cultivars requires genetic enhancement, molecular biology, and plant breeding efforts. One approach involves the use of beneficial plant-associated microorganisms such as endophytes that can offer environmentally friendly methods to increase productivity while reducing chemical inputs.

Most healthy and naturally propagated plants grown in fields or potting soils are colonized by communities of endophytic organisms of a wide variety of species and genera (Rosenblueth and Martinez-Romero 2006; Rodriguez et al. 2009). In general, associations of beneficial endophytes with plants can promote host plant growth, increase plant nutrient uptake, stimulate plant defence responses, and hasten tolerance to environmental stresses (Sturz et al. 2000; Lodewyckx et al. 2002; Ryan et al. 2008). Some of these endophytes, both fungal and bacterial, have been genetically engineered to enhance plant growth and improve stress tolerance for commercial uses (Ryan et al. 2008; Suryanarayanan et al. 2009; Mei and Flinn 2010). Knowledge on the diversity of endophytic fungal communities-associated with switchgrass and their effects on switchgrass biomass is limited to few reports (Ghimire et al. 2011; Kleczewski et al. 2012), and is nonexistent for endogenous bacterial endophytes. Growth stimulation of switchgrass by Burkholderia phytofirmans PsJN, a growth promoting bacterial endophyte originally isolated from onion roots (Sessitsch et al. 2005), was recently reported (Kim et al. 2012); however, the role of switchgrass indigenous endophytes in plant growth and their introduction into switchgrass cultivars remains poorly understood and needs to be explored.

In this study, the overarching goal was to enhance the understanding of the diversity of indigenous endophytic bacteria in switchgrass and their roles. For this purpose, we applied cultural, biochemical and molecular-based techniques to isolate and characterize cultivable endophytic bacteria from leaves of three switchgrass cultivars, and we determined the diversity and putative identities of the cultivated endophytic bacteria using genomic DNA fingerprinting by 16S rRNA gene sequence analysis. We specifically focused on the delineation of selective cultivated endophytic bacterial isolates and characterization of salient metabolic features related to antimicrobial and growth promoting properties. Antifungal activities, growth promoting abilities and specific metabolites produced by the endophytic bacteria were determined in bioassays and direct infusion Orbitrap mass spectrometry (MS), respectively.

Materials and methods

Field sites, cultivars and sample selection

Two field sites in Valleyfield, QC, Canada, were selected for switchgrass (P. virgatum L.) collection. Field site 1 (45′16′29″N and 74°4′2″W) was seeded in 1995 with Cave-in-Rock, the most recommended and widely used cultivar in North-eastern America (Lemus et al. 2002). This site received no fertilizer amendments until 2006 after which, only 50 kg N ha−1 was annually applied. Field site 2 (45′16′23″N and 74°0′59″W) was seeded in 2006 with cultivars of Tecumseh and Blue Jacket and received no fertilizers in 2006 and 2007, but they received 50 kg ha−1 of N in 2008 and 2009. Both field sites were annually mowed in the fall, and the material was baled in the spring.

Seeds from switchgrass plants showing the best agronomic traits from the three cultivars were collected on 28 October 2009 from both field sites and were placed in envelopes and stored in the dark at room temperature. A portion of the seeds was reserved for DNA extraction, while the remaining seeds (912 seeds per cultivar) were planted on 18 March 2010 in 38-cell trays (Plant products Co. Ltd, Brampton, ON, Canada) containing 50 Pro-Mix HP/50 Pro-mix BX of potting mixture (Plant products Co. Ltd), grown in a greenhouse at 21°C/19°C day/night and under spring light conditions, and watered three times a week. Plant height was recorded after 2 months of growth. The tallest 20 plants of each cultivar were selected, transplanted into larger pots (10-cm in diameter) containing the same substrate; after 1 month, they were transferred into field site 2 in Valleyfield on 11 June 2010. Plots were designed as a grid pattern (8·25 m2), in which there is a 50-cm distance between plants on each axis.

Sampling and bacterial isolation

Plants were collected over two growing seasons in 2010 and 2011. At each sampling date (September 2010 and October 2010), four tillers per cultivar originated from seeds that were grown and collected from different parts of the field were used for subsequent sampling. Additionally, tillers of 25 Cave-in-Rock switchgrass plants were collected in October 2011 from an older established switchgrass stand (Site 1). Leaves of eight (four leaves per sampling date) asymptomatic plants (i.e. from seed grown switchgrass plants) of each cultivar were randomly sampled at two defined growth stages: late vegetative stage (i.e. the leaf at the upper node) and full reproductive stage (i.e. the flag leaf) in the months of September and October 2010, respectively. A total of 25 flag leaves of Cave-in-Rock from the established field were sampled. All samples were processed for bacterial endophytes. One leaf from each plant/cultivar/stage was sampled and immediately stored in individual Ziploc® bags, transported to the laboratory and processed within 48 h.

Leaves were surface-sterilized by stepwise washing procedure in ethanol, sodium hypochlorite and water according to Schulz et al. (1993). The grass leaf was then separated into leaf sheath and leaf blade, and each was cut into several 1-cm section pieces. Sections were plated onto nutrient agar (NA) and Luria-Bertani Agar (LBA; BBL, New York, NY, USA), and incubated at 24°C in the dark for 4–6 weeks. The effectiveness of the sterilization procedure was tested using the imprint method (Schulz et al. 1993). Emerging bacterial colonies from leaf sections were passed through four rounds of single colony isolation by streaking them on NA or LB culture medium to ensure purity of the organism prior to long-term storage at −80°C.

Identification and phenotypic characterization of bacterial isolates

Bacterial endophytes were grouped on the basis of their phenotypic characteristics, for example, colony colour and morphology, Gram reaction staining (Steinbach and Shetty 2001) and their antagonistic activity against selected fungi. All bacterial strains were further grouped into their taxa based on ITS cloning and sequencing. For further characterization of endophytes, 1 × 109 of bacterial cells were harvested for genomic DNA extraction of test bacterial strains grown in LB broth for 18 h with Qiagen DNeasy® Blood & Tissues kit. PCR amplification of the 16S rDNA gene for bacterial endophytes was performed with their respective universal primer set (Table 1) and was PCR-amplified using 20 ng of genomic bacterial DNA. The PCR primers were used to sequence the purified PCR products which were cloned in a TOPO® TA-cloning Kit (Invitrogen, Carlsbad, CA, USA) following the manufacturer's protocol. Plasmid DNA was purified and sent for sequencing at Genome Quebec (Montreal, QC, Canada). Gene sequences were edited into contigs using CAP3 Sequence Assembly Program (Huang and Madan 1999). Sequences were then subjected to BlastN searches against NCBI database. The nucleotide sequences were deposited into the GenBank public database (Table 1). The most similar sequences of endophytic bacteria were further aligned using ClustalW software (Larkin et al. 2007) in SDSC Biology Workbench, and the nonconserved regions were used to design specific primers for selected endophytic bacteria (Table 1). Specific primers were synthesized by Integrated DNA Technologies Inc. (Coralville, IA, USA), and were tested against all bacteria to insure specificity.

Table 1. List of bacterial strains isolated from switchgrass, specific and universal primers used in PCR assays
Taxon (strain no.)Switchgrass cultivarGenBank accession no.Reference strainAccession no.Primer sequence (forward/reverse) 5′–3′Query/reference ITS length (similarity %)PCR product size (primer Tm °C)References
  1. NA, not applicable.

 Curtobacterium flaccumfaciens (B8)Cave-in-Rock JN689336 C. flaccumfaciens DSM 20129 AM410688



1502/1504 (99)836 (60)Present study
 Pseudomonas fluorescens (B25)Cave-in-Rock JN689337 Ps. fluorescens ost5 DQ439976



1476/1500 (98)778 (60)Present study
 Microbacterium testaceum (B4)Cave-in-Rock JN689338 Mic testaceum 343 EU714365



1455/1466 (99)850 (55)Scarpellini et al. (2004)
 Bacillus subtilis (B26)Cave-in-Rock JN689339 B. subtilis Aj080718IA-25 HQ727971



1502/1503 (99)650 (60)Present study
 Pantoea ananatis (B47)Cave-in-Rock JN689340 P. ananatis 4CCG52 GQ383910



1483/1490 (99)450 (57)Walcott et al. (2002)
 Serratia sp. (B29)Blue Jacket JXL241699 S. proteamaculans 568 JN630873 NA468/468 (99)NA 
 Bacillus pumilus (B33s)Blue Jacket JXL241698 B. pumilus JR5-4 JQ229695 NA482/482 (100)NA 
 Sphingomonas parapaucimobilis (B34)Blue Jacket JX241697 S. parapaucimobilis JCM7510T D84525 NA439/440 (99)NA 
Universal primers
 Bacterial universal primers (27r/1492r)  



 1465 (58)Frank et al. (2008)
 Bacterial universal primers (27f/534r)  



 450 (60)Watanabe et al. (2001)
Plant universal primer  



 300–700 (55)Taberlet et al. (1991)

Cellulase and pectinase activities were assayed on indicator plates with carboxymethyl cellulose (Sigma-Aldrich Co., Oakville, ON, Canada) and with pectin (BDH Chemical Ltd, Missisauga, ON, Canada), respectively, by the method of Mateos et al. (1992). Chitinase activity was evaluated by measuring the production of reducing sugars (e.g. chito-oligosaccharides) using colloidal chitin (CC) as substrate (Shen et al. 2010). The ability of endophytic bacteria to solubilize inorganic phosphate was assayed on agar medium containing inorganic phosphate (g per 1 l: glucose, 10; NH4Cl, 5; NaCl, 1; MgSO4·7H2O, 1; Ca3(HPO4)2, 0·8; agar, 15, pH 7·2) according to Verma et al. (2001). In all cases, a clearing zone >1 mm and extending from the periphery of the spot growth was considered positive.

Antimicrobial agent resistance was tested individually on agar plates containing antibiotics at the following concentrations: kanamycin, rifampin, streptomycin and tetracycline, 100 μg ml−1; ampicillin, chloramphenicol, gentamicin, and hygromycin, 125 μg ml−1. Bacterial isolates were plated onto LBA with or without antibiotic supplementation. Most antibiotics were purchased from Sigma-Aldrich, with the exception of hygromycin (Calbiochem, EMD Millipore Corp., Billerica, MA, USA) and chloramphenicol (ICN Biomedicals Inc., Aurora, OH, USA). Bacteria were considered sensitive to an antibiotic at the concentration tested if no visible growth was observed on plates containing the antibiotic when there was visible growth on control plates after 48 h of incubation at 27°C.

The antifungal activity of cell-free culture filtrates of bacteria was estimated using the method of Quiroga et al. (2001). Cell-free supernatant of bacteria grown on LB broth for 4 days was obtained by centrifugation twice at 16 000 g for 10 min and immediately freeze-dried. One gram of freeze-dried powder was dissolved in 5 ml of sterile distilled water and filtered on 0·22-μm Millipore filter. Petri plates were filled with 25 ml of molten precooled PDA and divided into two groups (triplicate for each). A 5-mm-diameter mycelial plug taken from the edge of an actively growing test fungus (Table 2) was placed in the centre of the culture plate. To each plate in the experimental groups (E), two 5-mm-diameter plugs were dug out 25 mm away from the centre on both sides of the same axis. To each hole, 150 μl of culture filtrate was added in every well. To those of the control group (C), freeze-dried and concentrated LB broth without bacteria as well as no extract was added. There were three replicates per test fungus. Both groups were incubated at room temperature, and measurement of inhibition of radial growth was recorded every 24 h. The percentage of inhibition ratios was calculated using the formula: (CE)/C × 100, where C is the average diameter of the control group and E the average diameter of the treatment (Wang et al. 2011).

Table 2. Phosphorus solubilization and chitinase and cellulose agar activity assays
BacteriaPhosphorus solubilizationChitinase activityCellulase activityPectinase activity
  1. −, no hydrolysis; +, 1·3 mm clearing zone; ++, >3 mm clearing zone.

Microbacterium testaceum B4
Curtobacterium flaccumfaciens B8+
Pseudomonas fluorescens B25++
Bacillus subtilis B26++++
Pantoea ananatis B47+

Detection of bacterial endophytes using PCR methods – plants and seeds

Sterilized switchgrass plant tissues (i.e. sheaths and blades of leaf stages) and disinfected seeds (Sauer and Burroughs 1986) were reduced to powder under liquid nitrogen using a mortar and pestle and subjected to DNA extraction using the Dneasy Plant Mini kit 50 (Qiagen, Toronto, ON, Canada). The presence of selected endophytes within switchgrass tissues was confirmed by PCR using the GeneAmp PCR System 9700 (Applied Biosystems, CA, USA). Each amplification mixture contained 2·5 μl of 10× PCR buffer (Fermentas, Thermo Scientific, San Jose, CA, USA), 2·5 μl of dNTP (2 mmol l−1), 2 μl of MgCl2 (15 mmol l−1), 1·5 μl of each primer (2 mmol l−1), 0·5 U of Taq polymerase (Fermentas), and 8 μl of template DNA (5 ng μl−1) in a total volume of 25 μl. All PCRs were contacted following the recommendations of Fermantas with specific annealing temperatures (Table 1). All samples were run on a 1% agarose (cat. no. GO60-2; Applied Biological Materials Inc., Richmond, BC, Canada) gel electrophoresis and visualized using Gel Logic 200 Imaging system from Mendel under UV light.

Growth promoting ability of volatile organic compound

Seeds of switchgrass and Brachypodium distachyon were surface-sterilized by a stepwise washing procedure using ethanol and sodium hypochlorite according to Vain et al. (2008). The seeds were left to germinate for 1 week in a sterile Petri dish containing two layers of filter paper imbibed with sterile water in the dark at 4°C for 2 days followed by 5 days in the growth chambers under conditions described below.

The ability of certain endophytic bacteria to promote growth in plants by volatile organic compounds was assessed using a modified method of Ryu et al. (2003). Five millilitres of full strength Murashige and Skoog medium (Sigma-Aldrich, St Louis, MO) was poured on one side of the Polystyrene Bi-Plate (100 mm × 15 mm; Fisher Scientific, Pittsburgh, PA, USA). On the other side of the plate, 5 ml of LBA + 5 mmol l−1 of tryptophan was added. Eight pregerminated grass seeds were placed on Murashige and Skoog, while the non plant side of the plate was streaked with 5 μl of 1 × 109 CFU ml−1 of bacterial preparation. Escherichia coli strain DH5 alpha (Qiagen Inc., Valencia, CA, USA), which does not produce volatiles as well as noninoculated media, served as controls. Plates were sealed with parafilm and incubated for 12 days under a regime of 16 h/8 h of light/dark with 160 μmoles m−2 s−1 of light and temperature of 25/23°C light/dark. There were three replicates per endophyte per treatment, and the experiment was repeated twice. After a 12-day exposure to airborne compounds, the average of total length of seedlings of different treatments was recorded and statistically compared to control treatment. Data were analysed by one-way anova using the JMP 8.0 software (SAS Institute, Cary, NC, USA). The significance of the effect of the treatments was determined by the magnitude of the F-value (P = 0·05).

Inoculation of switchgrass seedling with bacterial endophyte isolates and confirmation of the re-isolated bacteria

To test the ability of bacterial endophytes to colonize switchgrass, representative isolates were introduced into seedling grown aseptically under controlled conditions. Magenta® GA-7 Plant Culture Boxes 3 × 3 × 4″ (Magenta, Chicago, IL, USA) containing 100 g of sand and vermiculite (50/50% v/v) were autoclaved for 1 h every 24 h for a period of 72 h. Twenty surface-sterilized switchgrass seeds were seeded in each Magenta box and grown in a growth chamber at 22°C under a 12 h/12 h of light/dark cycle. Each box received 5 ml of sterile distilled water at planting. Bacterial endophytes were grown in LB broth for 18 h to the mid-log phase, pelleted by centrifugation, washed and suspended in sterile distilled water. After 2 weeks, plants were thinned down to 15 seedlings, which received 5 ml of water containing 105 CFU ml−1 of bacteria. Switchgrass seedlings receiving autoclaved distilled water served as controls. Putative endophytes B. subtilis and Mic. testaceum were tested alone and in combination. Plants were incubated further for another 2 weeks. Four-week-old seedlings were dipped in a solution of 70% ethanol, rinsed in autoclaved–distilled water, separated into roots and shoots, and subjected to DNA extraction using Qiagen DNeasy® Plant Mini kit. The presence of endophytes in roots and shoots of inoculated and noninoculated switchgrass was assessed using specific primers designed for each endophyte (Table 1), while the presence of the endophytes in the potting mix was confirmed by serial dilution on microbiological culture medium. The experiment was replicated eight times with 15 plants per replicate.

Direct infusion Orbitrap MS analysis of bacterial culture supernatants: extraction, chemical analysis and metabolite identification

For the detection of bioactive compounds of the bacterial exo-metabolome, the strains of B. subtilis, Mic. testaceum and Ps. fluorescens were grown on 250 ml of LB amended with 1% tryptophan (Sigma) in 500 ml Erlenmeyer flasks. Cultures were kept at 37°C under continuous agitation (250 rev min−1) for 7 days, allowing for more time in idiophase for the production of secondary metabolites. The cultures were centrifuged at 16 000 g for 10 min. The supernatants were collected and centrifuged (16 000 g for 10 min). The supernatants from the second centrifugation were used in the analyses. Along with the supernatants of bacterial cultures, samples from a flask that only contained LB amended with 1% tryptophan was processed for the detection of metabolites of nonbacterial origin. Samples (50 mg) were freeze-dried in 2-ml glass autosampler vials using a Gamma 1-16 LSC freeze drier (MBI, Montreal, QC, Canada). For the extraction, 1 ml of a mixture of methanol/ethyl acetate (50 : 50, v/v) (Optima® grade; Fisher Scientific) was added, followed by sonication for 25 min. Samples were further extracted for 2 h under continuous agitation (250 rpm) at 25°C and finally filtered through 0·2-μm filters (Millex-FG; Millipore, Billerica, MA, USA). The volume of samples was adjusted to 1·0 ml and subsequently divided into two equal portions for analyses in positive (ESI+) and negative (ESI) electrospray modes. Extracts were dried using a Labconco CentriVap refrigerated vacuum concentrator equipped with a cold trap (Fisher Scientific). For analyses in ESI+, 100 μl of a mixture of methanol/formic acid (0·1% v/v) (50-50, v/v) or 100 μl methanol/ammonium acetate (2·5 mmol l−1) for analysis in ESI were added, and finally the extracts were transferred in microinsertes in glass autosampler vials.

For the direct infusion MS/MS analysis, an LTQ Orbitrap MS Classic (Thermo Scientific, San Jose, CA, USA) was used acquiring in ESI+ or ESI. The analyser was equipped with a quadrupole linear ion trap and a Proxeon nanoelectrospray ion source. All experimental events were controlled using the Xcalibur software v.2 (Thermo Scientific), whereas chromatogram processing was performed using the software Sieve 2.0 (Thermo Scientific). External calibration of the instrument was performed daily during analyses using the recommended by the manufacturer solutions mscal5 and mscal6 (Sigma-Aldrich). Samples (10 μl) were injected at a flow rate of 1·0 μl min−1 using a 100-μl syringe (Hamilton, Reno, NV, USA), and spectra were acquired in the range between 20 and 1600 Da using a full width at half maximum (FWHM) of 60 000.

Acquired chromatograms were processed using the software Sieve v2.0. Putative identification of metabolites was based on mass accuracy, MS/MS fragmentation patterns and searches against an in-house built species-specific target-library composed of 1306 primary and secondary metabolites of the analysed bacterial species with the corresponding monoisotopic masses and molecular formulae. For the construction of the database, data were retrieved from the publicly available databases of the Kyoto Encyclopedia of Genes and Genomes (KEGG) for B. subtilis (, Mic. testaceum (, Ps. fluorescens (, PubChem (, and BioCyc ( Using Sieve, chromatograms were dissected to the so-called frames (region in m/z vs retention time plane) for the retention time between 3·5 and 4·5 min, and finally searches for various metabolite adducts were made with an accuracy of <5 ppm. The settings used for Sieve analyses are given in the Table S1. Searches were performed for the most commonly observed adducts in ESI+ and ESI.


Efficacy of sterilization

The surface sterilization protocol was a critical prerequisite for isolating plant endophytic bacteria. This study proved that the surface sterilization protocol combined with the imprint technique was effective in removing epiphytic organisms and that the bacterial-isolated strains can be considered endophytic organisms.

Survey of switchgrass endophytes and molecular identification

Over the course of this study (2010 and 2011), 594 switchgrass leaf segments were incubated and the total number of isolated culturable bacterial endophytes was 31 (Fig. 1a–c). The majority of putative endophytes (61%; 19 isolates) were recovered from old stands of Cave-in-Rock, followed by those recovered from Blue Jacket and Tecumseh, respectively (Fig. 1a). Interestingly, irrespective of the cultivar or leaf type (i.e. vegetative or reproductive), bacterial endophytes were almost equally distributed between the sheaths and the blades (Fig. 1b).

Figure 1.

Distribution of bacterial endophytes in (a) switchgrass cultivars, (b) leaf tissues and (c) number of recovered strains.

All 31 strains were identified by cloning and sequencing the amplified fragments (ranging 1505–500 bp) using the ITS-sequence data. Partial sequence data for the 16s rDNA gene have been deposited in the GenBank (NCBI) nucleotide sequences data base library. Data for endophytic strains have been deposited under the following accession numbers (JN689336JN689340 and JXL1697JXL241699). The strains were grouped into eight different taxa that shared high homology of 98–99% with other known sequences (Table 1). Bacterial endophytes were identified as Mic. testaceum (Gram-positive, Actinomycetales), Curtobacterium flaccumfaciens (Gram-positive, Actinomycetales), B. subtilis and Bacillus pumilus (Gram-positive; Bacillales), Ps. fluorescens (Gram-negative; Pseudomonadales), Sphingomonas parapaucimobilis (Gram-negative; Sphingomonadales), Serratia sp. (Gram-negative; Enterobacteriales), and Pantoea ananatis (Gram-negative; Enterobacteriales) with the Pananatis and Bacillus species as the most frequently isolated endophytes (Fig. 1c).

Phenotypic characteristics and antagonistic property of endophytes

The three endophytes, Ps. fluorescens, B. subtilis and P. ananatis strains were able to produce clear halo on inorganic P supplemented medium. Ps. fluorescens and B. subtilis displayed larger diameter of clear halo (Table 2). Microbacterium testaceum and C. flaccumfaciens showed no ability to solubilize phosphorus B. subtilis showed the greatest ability to utilize cellulose followed by C. flaccumfaciens (Table 2). The remaining endophytes did not produce clearing zones. None of the tested endophytes displayed chitinase or pectinase activities.

Antagonism towards test fungi was recorded as growth inhibition and also as per cent inhibition ratio in agar diffusible method (Table 3). Bacillus subtilis (B26) was effective against all test fungi, followed by Cflaccumfaciens (B8), Ps. fluorescens (B25) and P. ananatis (B47) which were effective against few fungi using both methods (Table 3).

Table 3. Inhibition of radial growth in mm and (%) ratio of test fungi by agar-diffusion with bacterial extract
Test fungusB4 Microbacterium testaceumB8 Curtobacterium flaccumfaciensB25 Pseudomonas fluorescensB26 Bacillus subtilisB47 Pantoea ananatis
  1. a

    G. Lazarovits, Agriculture, Agri-Food Canada, London, ON, Canada.

  2. b

    Numbers represent the absolute value (mm) of inhibition in agar-diffusion assay. Numbers represent the means of three replicates ± SE.

  3. c

    Numbers in brackets represent the percentage inhibition ratio in agar-diffusion assay. The percentage of inhibition ratios were calculated with the formula: (CE)/C × 100, where C is the average diameter of the control group and E the average diameter of the treatment (Wang et al. 2011).

  4. d

    S. Neate, DPI, Queensland, Australia.

  5. e

    M.A. Cubeta, CIFR, North Carolina State University, NC, USA.

  6. f

    American Type Culture Collection (ATTC), Manassas, VA 20108, USA.

Fusarium solani a 0b (0)c0 (0)0 (0)11·0 ± 0·3 (17)0 (0)
Binucleate Rhizoctonia sp.d0 (0)0 (0)12·3 ± 5·5 (13)34·6 ± 1·3 (38)0 (0)
Rhizoctonia solani e 0 (0)18·0 ± 3·7 (24)9·0 ± 3·5 (12)21·8 ± 0·5 (29)20·3 ± 0·6 (27)
Trichoderma virens f 17·3 ± 0 (29)19·7 ± 0·3 (33)16·7 ± 1·2 (0)23·7 ± 0·7 (40)8·3 ± 1·0 (14)
Verticillium albo atrum a 0 (0)0 (0)0 (0)8·3 ± 0·3 (28)0 (0)
Colletotrichum truncatum a 0 (0)0 (0)0 (0)16·0 ± 1·0 (24)0 (0)

All bacterial endophytes were susceptible to Rifampin, Chloramphenicol and Gentomycin and did not grow. However, all endophytes showed resistance to at least one antibiotic. Strain Ps. fluorescens B25 was the most resistant strain and grew on the following four antibiotics: Ampicillin, Hygromycin, Streptomycin and Kanamycin (data not shown).

Detection of bacterial endophytes using PCR methods – plants and seeds

The presence of the identified endophytes in various tissues (leaves and seeds) of field-grown switchgrass cultivars was confirmed by PCR assays using species-specific primers designed for each one of the identified endophytes (Table 1). The presence of endophytes varied with cultivars and tissue types. The bacterial endophytes, B. subtilis (Fig. 2a) and C. flaccumfaciens (Data not shown) were detected in tissues of switchgrass grown from seeds collected in 2009. Others were detected in one switchgrass cultivar only such as Mic. testaceum that was found in Cave-in-Rock cultivar. The species-specific primer sets failed to detect the bacterial endophytes: Ps. fluorescens, P. ananatis in field switchgrass although they successfully amplified the endophytes when grown in pure culture (Data not shown). Interestingly, vertical transmission of the following endophytes via seeds, B. subtilis and Mic. testaceum, was confirmed in switchgrasss grown in 2010 and originated from seeds collected in 2009 (Fig. 2a).

Figure 2.

PCR detection of bacterial endophytes in different switchgrass cultivars using species-specific primers. (a) Detection of Bacillus subtilis B26 and vertical transmission via seeds, lanes 1–3 and 4–6. Leaf and seed DNA of Tecumseh, Cave-in-Rock and Blue Jacket, respectively. (b) Detection of B. subtilis B26, lanes 1–3. DNA of noninoculated seedlings (Control), lanes 4–6. DNA of inoculated seedlings. Entire seedling (lane 4), roots (lane 5) and leaves and stems (lane 6). (c) Detection of Microbacterium testaceum B4, Lanes 1–4. DNA of noninoculated seedlings, Lanes 5–8. DNA of roots (lanes 5, 6) and stems and leaves (lanes 7, 8) of re-inoculated switchgrass seedlings. (d) Detection of both B26 and B4 strains in re-colonized switchgrass seedlings of Cave-in-Rock cultivar. DNA from roots (lane 1) and leaves and stems (lane 2) of re-inoculated switchgrass with both endophytes, and tested with species-specific primers of B26. DNA from roots (lane 3) and leaves and stems (lane 4) of re-inoculated switchgrass with both endophytes and tested with B4 species-specific primers. (e) Soil drench of Cave-in-Rock switchgrass seedlings with (1) water, (2). B. subtilis B26 alone (3), Mictestaceum B4 alone and (4) with both strains. λ = 100 bp DNA ladder; (+) = pure genomic DNA of endophytes; (−) = no template. Arrows indicate PCR product size in bp.

The presence of endophytes was successfully detected in roots and shoots of 4-week-old switchgrass inoculated singly (Fig. 2b,c) or in combination with B. subtilis and Mic. testaceum (Fig. 2d). Absence of endophytes was confirmed in noninoculated switchgrass seedlings (Fig. 2b,c). Inoculated seedlings with either B. subtilis or Mic. testaceum showed better growth than those noninoculated (Fig. 2e).

Growth promoting abilities

Activation of bacterial volatiles responsible for triggering growth promotion in switchgrass and Brachypodium seedlings were assayed by physically separating seedlings from the test bacterial endophytes on divided Petri dishes so as to allow only airborne signals transmitted between bacterial cultures and the plant seedlings (Fig. 3a). Grass seedlings exposed to certain bacterial strains for 12 days significantly (P < 0·05) promoted growth compared with the water or DH5-alpha controls (Fig. 3b). Pantoea ananatis B47 failed to promote seedling growth compared to the control, indicating that the release of bacterial volatiles is not the common mechanism for stimulating growth for all tested endophytes.

Figure 3.

Quantification of growth promotion in switchgrass seedlings with exposure to airborne chemicals released from five bacterial strains compared with control treatment. (a) Representative examples of switchgrass seedlings after 12-day exposure to no bacteria (a), Microbacterium testaceum B4 (b) Curtobacterium flaccumfaciens B8, (c) Pseudomonas fluorescens B25, (d) Bacillus subtilis B26 (e), and Pantoea ananatis B47 (f). (b) Total shoot and root lengths, a measure of seedlings' growth after 12-day exposure of pregerminated seeds of switchgrass to airborne chemicals released from selected endophytes and compared with control seedlings. There were three replicates per endophyte per treatment. Each replicate had eight seedlings. Letters above bar graphs indicate significant differences between treatments (< 0·05).

Bacterial-derived metabolites and peptides

Orbitrap MS analyses revealed the complexity of the recorded metabolite profiles (Figs S1 and S2), which were composed of compounds belonging to the culture media as well as of compounds of bacterial origin. Using the software SIEVE, 4534 frames were created from B. subtilis MS spectra acquired in ESI+ and 5551 frames from spectra acquired in ESI. From Ps. fluorescens and Mic. testaceum MS spectra, 4702 and 5424 frames were created in ESI+ and in ESI 23 825 and 18 466, respectively. Mass searches against the in-house built library were performed for frames found only in the media in which bacteria were grown after the removal of frames corresponding to media alone. As tentatively identified were considered metabolites or peptides existing in the library and which were detected within an error of <2 ppm (Tables S2–S4), using additionally data of MS/MS fragmentation patterns.

Metabolites with varied chemical structures and well-reported bioactivity belonging to amino acids, carboxylic acids, phosphoric acid derivatives, purines and pyrimidines, and heterocyclic compounds, as well as low molecular weight peptides, were identified in B. subtilis, Mic. testaceum and Ps. fluorescens media (Table 4).

Table 4. Compounds excreted in 7-day-old liquid cultures of the Bacillus subtilis isolate B26 (1), Microbacterium testaceum isolate B4 (2), or Pseudomonas fluorescens isolate B25 (3) grown in Luria Broth medium (LB) and amended with 1% tryptophanThumbnail image of

Indole-3-acetate, which was identified in the media of all three bacterial species, and the methyl form, methyl-indole-3-acetate (IAA), which was identified in B. subtilis and Mic. testaceum media, are two well-known phytohormone auxins. Additionally, the cytokinin zeatin riboside was identified in cultures of B. subtilis only (Table 4).

As expected, B. subtilis B26 was found to produce a variety of bioactive nonribosomally synthesized cyclic lipopeptides belonging to the iturin family and surfactins. On the other hand, Surfactin C15 was the only lipopeptide identified in media of Ps. fluorescens, whereas Surfactins C14 and C15 were identified in media of Mic. testaceum (Table 4). Of interest, was the detection of cephalosporin C in culture filtrates of Ps. fluorescens.


Our research goals were to survey switchgrass grown in Quebec for the presence of endophytic bacteria and to determine their taxonomic position. Strategically, to cover a maximum genetic diversity, we sampled leaves from old and well-established stands as well as from recently established fields of three switchgrass cultivars having different genetic backgrounds. The endophytes isolated over the course of this work were obtained using a traditional method, in which only endophytes that can grow on microbiological media were accounted for which may create a negative bias towards endophytes that are slow growing or that cannot grow on biological media. To assess the full range of endophytes, future work should incorporate molecular techniques like DGGE or T-RFLP (Garbeva et al. 2001; Overbeek et al. 2006). As the over arching aim of this study is to identify endophyte species that could be introduced into elite switchgrass cultivars for maximizing their utility as a bionergy crop, we focused on endophytes that could be grown in pure culture, using different media with different sources of carbon and pH, we believe to have countered as many as possible.

In this study, we demonstrated that the diverse assemblage of cultivable bacterial endophytes belong to eight different taxa and exist as endophytes in switchgrass cultivars. To the best of our knowledge, our study is the first to describe indigenous bacterial endophytes from elite switchgrass cultivars produced through breeding. Over the course of this study, endophytes were isolated from 594 switchgrass leaf segments with the majority (61%) of bacterial endophytes originating from an old stand of Cave-in-Rock cultivar that was established 16 years ago. This is not surprising as the disparity in the number of isolates per cultivar may be the result of plant age. As time of exposure to endophyte inoculum increases, plants seem to accumulate an increasing number of endophytes in their tissues. Because of this, older plant parts may harbour more and diverse populations of bacterial endophytes than younger ones (McInroy and Kloepper 1995; Sturz et al. 2000).

Our results also show the predominant existence and distribution of the genus Bacillus and Pantoea in all breeding lines of switchgrass. The dominant taxa recovered from switchgrass reflect an unequal distribution of isolate richness among species. This kind of abundance has also been observed in other grasses (Mano et al. 2007; Mendes et al. 2007). Bacterial community inside a plant is prone to influences caused by host genotype and changing physiology (Hallmann and Berg 2006). Therefore, many factors that modify plant physiology, such as cultivar type, plant age and tissue type and seasonality, are thought to promote shifts in the endophytic community structure (Adams and Kloepper 2002; Overbeek et al. 2006). It is not yet clear whether this abundance inequality is method related, that is, the selection of certain taxa by the cultivation method. Our results demonstrated that the eight different recovered taxa of bacteria from switchgrass are also associated with grasses and plants as endophytes, implying that these species are host-generalists (Sturz et al. 2000; Lodewyckx et al. 2002).

Our ITS-sequence database for switchgrass-associated fungi was not only the basis for molecular taxonomy but also served as a source for designing species-specific primers for monitoring individual fungal species within DNA isolated from leaves and seeds of switchgrass cultivars. We have applied this approach for tracking individual endophytes by PCR analysis in seeds and of mixed endophytes reintroduced into switchgrass seedlings. Seed-borne endophytes can provide improved vigour to switchgrass, which could be most important when establishing new switchgrass stands. In our study, the endophytes B. subtilis, Mic. testaceum and C. flaccumfaciens were frequently detected in switchgrass seeds as well as plants grown in 2010 that originated from seeds collected in 2009. These results demonstrate that the endophytes are carried by seeds as a viable propagule and can be vertically transmitted to the next generation of the host. Similar type of transmission of endophytes is found in common rice (Okunishi et al. 2005; Cottyn et al. 2009; Kaga et al. 2009).

Endophytic bacteria inside a plant might either become localized at the point of entry or spread throughout the plant and may reside within cells, in the intercellular space or in the vascular system (Hallmann and Berg 2006; Rosenblueth and Martinez-Romero 2006; Reinhold-Hurek and Hurek 2011). True endophytic bacteria are recognized by their capacity to re-infect disinfected seedlings and by establishing visualized evidence of their localization inside plant tissues (Reinhold-Hurek and Hurek 2011). In this study, the latter was not fulfilled, but we were able to fulfil the former criterion. In our colonization experiments, B. subtilis and Mic. testaceum were detected in roots and leaves 2 weeks after reintroducing them into disinfected switchgrass seedlings, using species-specific primers. These results indicate that systemic spread of endophytes within the roots and the leaves was successful and confirms that they have moved from the roots and travelled upward to the stem and leaves. Our results concur with previous findings, at least for B. subtilis, the ability of this endophytic bacterium to enter the roots and migrate in the upper stems and leaves of eggplants (Lin et al. 2009). The exact localization of B. subtilis in aerial plant tissue remains to be investigated.

We have demonstrated that some endophytic bacteria from switchgrass possess the ability to degrade the plant polymer cellulose and also solubilize phosphorus. These results indicate the endophytes' potential for nutrient acquisition as well as colonization capacity and active recognition by the plant cells. Hydrolytic enzymes, such as pectinases and cellulases, have been suspected to play a role for internal colonization of some endophytic bacteria (Alstrom 2001; Compant et al. 2005). Another interesting feature observed was that three (Ps. fluorescens, B. subtilis and P. ananatis) of the selected five switchgrass endophytes exhibited phosphatase activities. Phosphorus is one of the most important plant nutrients and a large portion of inorganic phosphates applied to soil as fertilizer is rapidly immobilized after application and becomes unavailable to plants (Rodriguez and Fraga 1999). Bacterial endophytes may promote plant growth via phosphate solubilization activity (Rosenblueth and Martinez-Romero 2006). Previous experiments have shown that endophytic bacteria possess the capacity to solubilize immobilized mineral phosphates (Rodriguez and Fraga 1999; Verma et al. 2001), suggesting that during initial colonization, endophytic bacteria could enhance phosphate availability to the host plant. This ability combined with that of producing cellulases is indicative of their potential for nutrient-delivering capacity while interacting with plant hosts.

Agar diffusible antifungal activity and inhibition of radial mycelial growth is an indication of the potential of a bacteria antagonist to produce antibiotics (Lodewyckx et al. 2002). In the absence of direct contact between the endophytic bacteria and the fungal pathogens, the in-vitro inhibition of radial mycelial growth of fungal plant pathogens was demonstrated in this study. These results suggest that inhibition of fungal growth may be due to the presence of antifungal compounds released into the culture media by the tested endophytes. Confirmation of the presence of several toxins in the cell-free culture filtrates of these endophytes was demonstrated using direct infusion orbitrap MS. Analysis of culture filtrates of B. subtilis, Psfluorescens and Mic. testaceum confirmed the presence of several well-characterized toxins such as lipopeptides and surfactins (Ongena and Jacques 2008; Romero et al. 2011). The lipopetides, iturins and mycobacillin, and surfacins C13, 14, 15 were produced by B. subtilis strain B26, surfactins, C14 and C15 by Ps. fluorescens B25 and surfactin C13 by Mic. testaceum B4. The production of various toxins by these genera of bacterial endophytes and their antimicrobial activities is well documented (Stein 2005; Romero et al. 2011).

Interestingly, the presence of surfactins alone in cell-free cultures of Ps. fluorescens and Mictestaceum does not account solely for antibiosis against fungi, as it is well documented that surfactins are mainly known for their antiviral and antibacterial activities (Bais et al. 2004), but have weak antifungal activity by themselves (Ongena and Jacques 2008). However, recent reports have demonstrated surfactins' strong antifungal activities against several plant pathogens (Souto et al. 2004; Joshi et al. 2008; Vitullo et al. 2012). In the case of Microbacterium, surfactins were identified in strains isolated from oil-contaminated mangroves (Aniszewski et al. 2010); however, the antifungal activity of surfactins has never been studied in this genus. Intriguingly, is the production of cephalosporin C, a β-lactam antibiotic by Ps. fluorescens. Cephalosporins are derived from the fungus Cephalosporium acremonium (Samson et al. 1985) and are used as bactericides against Gram-positive bacteria. Generally, Ps. fluorescens strains produce various extracellular toxins including pyrrolnitrin (PRN), 2,4-diacetylphloroglucinol, hydrogen cyanide and pyoluteorin (Haas and Keel 2003). This is the first report of the production of cephalosporin c toxin by Ps. fluorescens, and the role of cephalosporin in Ps. fluorescens is not yet known and merits further investigation. It is highly probable that when Ps. fluorescens internal colonization of switchgrass is restricted, either because of nutrient limitation or competition of the same ecological niche with other Gram-positive bacterial endophytes that are present at high cell density, it switches on secondary metabolism, perhaps as a strategy to remain competitive in their environment (Sturz et al. 2000).

Many bacteria–plant interactions are based on the direct contact of bacteria with the plant, but interactions at a distance through volatile substances have emerged as a novel way of signalling between bacteria and plants (Ryu et al. 2003). Rhizospheric bacteria, including Bacillus and Pseudomonas genera, are well-known producers of a diverse blend of volatiles (Schulz et al. 2010; Blom et al. 2011) with indole, 1-hexanol, pentadecane, butanediol and acetoin as the most abundant compounds released with growth promoting abilities (Ping and Boland 2004; Blom et al. 2011). Our results from the experimental set-up consisting of divided Petri dishes with bacteria growing on one side and switchgrass or Brachypodium plants on the other clearly demonstrated that volatiles produced by B. subtilis, Mic. testaceum and Ps. fluorescens significantly promoted plant growth compared to nonexposed seedlings. Whether this effect is mediated through a single volatile compound or a blend of volatiles remains to be investigated.

Biofertilizers increase crop growth by combinations of mechanisms, which include biological nitrogen fixation, phytohormone production, increasing the availability of soil nutrients and disease control (Sturz et al. 2000). Phytohormones are involved in the control of growth and in almost every important developmental process in plants. The ability of bacterial cultures to synthesize indole-3-acetic acid (IAA), gibberellins, and cytokinins has been shown by many endophytic bacteria (Arkhipova et al. 2005; Berg 2009; Bhattacharyya and Jha 2012). Keeping this in mind, we evaluated the potential of selective isolates to produce both IAA and cytokinins. The analysis of cell-free culture filtrates of endophytic bacteria by direct infusion orbitrap MS showed that B. subtilis and Mic. testaceum and Ps. fluorescens were able to produce the phytohormones, IAA and cytokinins. These results combined with the ability of the endophytes to solubilize P make them good candidates to be used in future application for switchgrass growth promotion. However, practical application of these results should be further evaluated in field experiments.

In summary, this study describes the morphological, molecular and physiological characteristics of indigenous culturable bacterial endophytes isolated from leaves of switchgrass cultivars. The successful colonization of switchgrass with selected endophytes, their potential to be transmitted between switchgrass generations and their possession of important biochemical traits can help in designing improved microbial cocktails for developing biofertlizers for the establishment of elite switchgrass cultivars in the first year, and in the development of low-input and sustainable feedstock production system.


This work was supported by the Ministére d'agriculture, pécheries et alimentation de Québec (MAPAQ), and in part by the Natural Sciences and Engineering Research Council of Canada (NSERC).