Monitoring the plant epiphyte Methylobacterium extorquens DSM 21961 by real-time PCR and its influence on the strawberry flavor Free

Besides its influence on plant growth and health, plant-associated bacteria exert an impact on fruit quality. Methylotrophic bacteria can enhance the biosynthesis of strawberry flavor compounds, especially the two furanoid compounds 2,5-dimethyl-4-hydroxy-2H-furanone (DMHF) and 2,5-dimethyl-4-methoxy-2H-furanone in vitro. Here, we report the selection and characterization of Methylobacterium extorquens DSM 21961, a strain that was able to enhance the furanone content ad planta under greenhouse conditions. For monitoring the colonization of strawberry plants, a strain-specific quantification system for M. extorquens DSM 21961 was developed. Specificity, linear range and quantitative limit of the system were shown, and successful application was demonstrated in a monitoring experiment of M. extorquens DSM 21961 on strawberry leaves under greenhouse conditions. Furthermore, the quantification of DMHF in strawberry fruits via GC indicated an increased biosynthesis of this compound in strawberry plants. The colonization behavior analyzed by confocal laser scanning microscopy using GFP-tagged cells revealed high colonization of the upper and the lower leaf surfaces, with a specific accumulation of bacterial cells on trichomes. The results support a biotechnological application of this promising flavor-stimulating agent.

Introduction

Plant-associated bacteria fulfill important functions on their host. Besides their well-studied effect on plant growth and health (reviewed in Berg, 2009; Lugtenberg & Kamilova, 2009), they are involved in plant metabolism and can influence fruit quality and flavor. Pink-pigmented facultative methylotrophic bacteria are ubiquitous, plant-specific phyllosphere-associated bacteria, which use methanol as a source of carbon and energy (Delmotte et al., 2009). They are known for their ability to interact with strawberry cell cultures: a cocultivation of the callus cultures with Methylobacterium extorquens led to an increased biosynthesis rate of furanones (Zabetakis & Gramshaw, 1996; Zabetakis, 1997; Zabetakis et al., 1999). Although >300 volatile compounds were identified in the flavor of strawberries, only a limited number of compounds are responsible for the formation of the typical and very pleasant flavor. Furanoid compounds such as 2,5-dimethyl-4-hydroxy-2H-furanone (DMHF) and 2,5-dimethyl-4-methoxy-2H-furanone (DMMF), with a flavor described as ‘caramel’ and ‘cotton-candy,’ are considered to be one of the most important contributors to strawberry flavor (Larsen et al., 1992; Schieberle & Hofmann, 1997; Bood & Zabetakis, 2002). In addition to its aromatic properties, DMHF is a known flavor enhancer; recent studies show that the presence of DMHF in sucrose solutions increases the perceived sweetness of the solution (Labbe et al., 2007). In this context, we suppose that an increased DMHF concentration could improve the flavor of strawberries not only in terms of aroma properties but also flavor-enhancing properties.

Furthermore, it was shown that the presence of M. extorquens activated a defense response in the plant cells and increased the production and accumulation of plant antimicrobial compounds in which DMHF can also be grouped (Sung et al., 2006). Also, a direct intervention of the bacteria in the biosynthesis pathway of furanones by providing lactaldehyde as a secondary metabolite or by the oxidation of endogenous 1,2-propanediol is discussed in the literature (Zabetakis, 1997). However, this interaction was only analyzed in vitro, and there was no evidence that methylobacteria influence the strawberry flavor under natural conditions.

Phyllosphere colonization by bacteria is the key factor for a successful plant–microorganism interaction (Whipps et al., 2008). Different molecular and microscopic methods are available to study the colonization pattern of plant-associated bacteria. To study colonization at the strain level, it is necessary to develop strain-specific primers, for example, for real-time PCR analysis. Many authors used a random amplification of polymorphic DNA (RAPD) (Welsh & McClelland, 1990; Williams et al., 1990; Hadrys et al., 1992), to provide specific regions in the genome of a target strain for primer design. Most of the strains, for which a specific primer design was described in the literature, are used as biocontrol agents in agriculture, for example Pseudomonas fluorescens EPS62e (Pujol et al., 2006), Aureobasidium pullulans (Schena et al., 2002) and Trichoderma hamatum 382 (Abbasi et al., 1999).

The objective of this study was to analyze the Methylobacterium–strawberry interaction ad planta. Therefore, active Methylobacterium strains from different plant species were isolated and characterized. Methylobacterium extorquens DSM 21961, the strain for which the highest biosynthesis rate of the precursor lactaldehyde was measured, was selected to investigate the interaction between bacteria and fully developed strawberry plants. For DSM 21961, a strain-specific primer for quantitative real-time (qRT)-PCR was developed. Therefore, we used BOX-PCR to provide regions on the genome of the target strain suitable for primer design. The primer used targets the repetitive BOX elements, which are spread throughout the bacterial chromosome (Koeuth et al., 1995). To complement the results obtained from qRT-PCR, the colonization of strawberry plants by M. extorquens was verified by confocal laser scanning microscopy (CLSM). The influence of M. extorquens on flavor production in strawberries was shown by the quantification of DMHF in the strawberry fruits via GC.

Materials and methods

Bacterial strains and growing media

All the strains used in this study are listed in Table 1. They were isolated from different host plants by imprinting leaves, roots and sections through inner plant tissues on MIS agar plates, containing in g L⁻¹: 1.8 (NH₄)₂SO₄, 0.2 MgSO₄·7H₂O, 1.4 NaH₂PO₄·2H₂O and 1.9 K₂HPO₄. After autoclaving, 5 mL L⁻¹ methanol and 1 mL L⁻¹ of a trace element stock solution were added. The trace element stock solution contained in mg L⁻¹: 500.0 EDTA, 200.0 FeSO₄·7H₂O, 10.0 ZnSO₄·7H₂O, 3.0 MnCl₂·4H₂O, 20.0 H₃BO₃, 20.0 CoCl₂·6H₂O, 1.0 CuCl₂·6H₂O, 2.0 NiCl₂·6H₂O and 3.0 Na₂MoO₄·2H₂O. This medium was used both as a liquid growing medium and, after addition of 12 g L⁻¹ of agar agar, as a solid medium. The growing temperature of the cultures was 30 °C and the incubation time was 2 days for liquid cultures and 5 days for agar plates. Long-term storage of bacteria was performed in a liquid MIS medium blended with 20% sterile glycerol at −70 °C.

To identify methylobacteria, a single colony on Luria–Bertani medium was picked with a sterile toothpick into 50 μL demineralized water in a 1.5-mL Eppendorf tube. The cell suspension was heated to 96 °C for 10 min. The 30-μL reaction mixture contained 6 μL 5 × Taq&Go^™ (MP Biomedicals, Irvine), 1.5 μL of primer pair mix EubI-forward (5′-GAG TTT GAT CCT GGC TCA G-3′) and 1492r-reverse (5′-TAC GGY TAC CTT GTT ACG ACT T-3′) both at a concentration of 10 pmol μL⁻¹ and 20–30 ng template. The PCR products were purified using the GeneClean Turbo Kit as recommended by the manufacturer. The fragments were sequenced using the reverse primer 1492r. For identification of related sequences, a database alignment using the blast algorithm was performed. One strain M. extorquens Rab 1 was transferred as a patent strain to the German Collection of Microorganisms and Cell Cultures (Braunschweig, Germany) and was assigned the number DSM 21961. The ability of strains to produce lactaldehyde in vitro was estimated according to Zabetakis (1997).

Isolation of genomic DNA

Bacteria were grown in 2-mL reaction tubes containing 1.5 mL MIS medium at 30 °C for 48 h, shaken at 120 r.p.m. Cells were centrifuged down at 14 000 g for 10 min. The supernatant was discharged and cells were resuspended in 1 mL of extraction buffer (2 mM Tris, 200 mM NaCl, 25 mM EDTA, 0.5% SDS). Cell suspensions were transferred into 2-mL reaction tubes with screw caps containing 400 mg sterile glass beads (0.25–0.5 mm, Sigma-Aldrich, MO). The tubes were treated using a FastPrep instrument (Qbiogen BIO 101 Systems, Carlsbad) for 30 s at level 5. One hundred and fifty microliters of 3 M sodium acetate was added and the samples were shaken for 2 min by hand. After centrifugation at 14 000 g for 5 min, the clear supernatant was transferred into a fresh 1.5-mL reaction tube, cleaned by phenol–chloroform extraction and DNA was precipitated at 0 °C for 1 h by adding an equal amount of ice-cold isopropanol. The precipitated DNA was centrifuged down for 15 min at 14 000 g and 4 °C and the resulting pellet was washed with 500 μL ice-cold 70% ethanol, air-dried, dissolved in 50 μL TE buffer (100 mM, pH 8) and stored at −20 °C.

BOX-PCR and sequencing of PCR products

Amplification reactions were performed in a final volume of 25 μL, containing 1 × Taq&Go^™ Mastermix (MP Biochemicals, Irvine, CA), 2.5 μM of primer A1R (5′-CTA CGG CAA GGC GAC GCT GAC G-3′) (Lloyd-Jones et al., 2005) and 50 ng of template DNA. Thermocycler conditions consisted of an initial denaturation step at 99 °C for 5 min, followed by 35 cycles of denaturation at 95 °C for 60 s, primer binding at 53 °C for 60 s and elongation at 65 °C for 8 min, and ending with an elongation step at 65 °C for 16 min. The success of the amplification was tested in a 1.5% agarose gel in 0.5 × TBE buffer (45 mM Tris, 45 mM boric acid, 1 mM EDTA, pH 8). Electrophoresis was performed at 2.8 V cm⁻¹ for 4.5 h. The gel was stained with ethidium bromide.

The PCR product of the BOX-PCR of M. extorquens DSM 21961 was purified using the Geneclean Turbo Kit (MP Biomedicals) according to the manufacturer's protocol. After purification, the DNA concentration was determined photometrically. As a sequencing vector, the pGEM^®-T Vector System with the corresponding cloning kit (Promega, Madison) was applied. Insertion was performed in a final reaction volume of 10 μL, containing 1 × ligation buffer, 50 ng pGEM^®-T vector, 1 U T4-ligase and 25-ng PCR product. The ligation reaction was incubated overnight at 4 °C. The next day, 2 μL of the ligation reaction was used for the transformation of competent cells of Escherichia coli JM109 (Promega) according to the protocol published by Sambrook & Russel (2001). The next day, clones were picked and checked on the appropriate vector insert via colony PCR. To provide template DNA from 20 white-colored colonies, some cell material was transferred into 1.5-mL reaction tubes using the tip of sterile toothpicks, 100 μL of water was added and the tubes were heated to 100 °C for 5 min in a heating block and afterward frozen for 15 min. PCR was performed in a final volume of 25 μL, containing 1 × Taq&Go^™ Mastermix, 4 mM forward primer usp (5′-GTA AAA CGA CGG CCA GT-3′), 4 mM reverse primer rsp (5′-CAG GAA ACA GCT ATG ACC-3′), 1.6 mM additional MgCl₂ and 1 μL of template DNA solution. The thermocycler program consisted of an initial denaturation step at 95 °C for 5 min, followed by 35 cycles of denaturation at 95 °C for 20 s, primer annealing at 54 °C for 15 s and elongation at 72 °C for 30 s, and ended with a final elongation step at 72 °C for 10 min. The success of the PCRs was checked in a 0.8% agarose gel in 1 × TAE buffer (40 mM Tris, 20 mM acetic acid, 1 mM EDTA, pH 8). Amplificons were purified using a Geneclean Turbo Kit (MP Biomedicals) according to the manufacturer's protocol and sequenced using an Applied Biosystems 3130l genetic analyser sequencer data collection v. 3.0, sequencing analysis v. 5.0 (Foster City, CA) at the sequencing core facility ZMF, Medical University of Graz, Austria.

Primer design

The sequences obtained were subjected to a blast search according to Altschul et al. (1997) against all available DNA sequences in the NCBI database. Sequences that yielded any positive result (especially genes for proteins involved in DNA replication were found) were excluded from the further process of primer design. The remaining sequences provided the basis for primer design using the software package primer express 5.0 (Applied Biosystems, Carlsbad). The software package yielded a list of 200 different sets of primers and probes, from which one set was chosen (forward primer Ext-f, reverse primer Ext-r, probe Ext-p).

DNA extraction for qRT-PCR

Different methods for DNA extraction and some ready-to-use extraction kits were tested with leaf samples containing different bacterial abundances. The FastDNA SPIN for Soil Kit (Qbiogen BIO 101 Systems) yielded the best results regarding reproducibility and linearity over the tested range of bacterial abundances and was used for all further analyses. For DNA extraction, strawberry leaves were freeze-dried overnight (Labconco Freezedry system/Freezone 4.5, Labconco, Kansas City, KS) and pulverized with a mortar and pestle. From the obtained powder, 20 mg was used for DNA extraction using the FastDNA SPIN for Soil Kit according to the manufacturer's instructions. The DNA extract obtained was cleaned using the Geneclean Turbo Kit according to the manufacturer's protocol. The DNA concentration in the extracts was determined photometrically and adjusted to 10 ng μL⁻¹ by diluting the samples with ultrapure water. DNA samples were stored at −20 °C.

qRT-PCR

qRT-PCR was performed in a final reaction volume of 20 μL in 96-well plates (Applied Biosystems) in an ABI PRISM 7000 Sequence Detection System (Applied Biosystems). The reaction mixture contained 1 × TaqMan mastermix (Applied Biosystems), 50 nM primer Ext-f, 50 nM primer Ext-r and 200 nM probe Ext-p. To each reaction, 10 ng (1 μL) of template DNA were added. The thermocycler program consisted of an initial denaturation step of 10 min at 90 °C, followed by 45 cycles composed of a denaturation step of 15 s at 90 °C, followed by an elongation step of 60 s at 60 °C. Data acquisition was performed during the elongation step. For the evaluation of the data obtained, the abi prism detection software was used.

Determination of the specificity and limit of quantification and calibration of the system

The specificity of the detection system was tested against pure cultures of methylobacteria, which were isolated from different plants (Table 1). qRT-PCR was performed with each 10 ng of DNA extracted from pure cultures and the resulting cycle threshold (C_t) values were compared.

The limit of quantification of the system was tested with a pure DNA extract from M. extorquens DSM 21961, which was diluted to different concentrations, and additionally with a DNA extract of strawberry leaves, which were grown under sterile conditions, that was supplemented with different amounts of DNA from M. extorquens DSM 21961.

To take into consideration all losses of DNA and other sources of errors during the process of DNA extraction, for the calibration of the system, leaf samples were supplemented with different cell numbers of M. extorquens DSM 21961. From these samples, DNA was extracted and qRT-PCR was performed as described above. The resulting C_t values were used to calculate a calibration curve for the system.

Quantitative monitoring of M. extorquens DSM 21961 on strawberry plants

Strawberry plants (cv. Elsanta) were purchased from a local nursery and were planted in pots of a volume of 1.6 L containing potting soil supplemented with 2.5 g fertilizer (Compo Beerenduenger, Muenster, Germany) per pot. The plants were grown under artificial light (16 h illumination day⁻¹ with metal halide lamps), temperature control (20 °C) and watering as required in a greenhouse. At the stage of flowering, the plants were sprayed with suspensions of M. extorquens DSM 21961 grown in 1000-mL shaking flasks containing 300 mL medium 4 (Bourque et al., 1995) for 48 h at 30 °C and 120 r.p.m. After enumeration of cell density in a hemocytometer, cell suspensions were diluted with tap water to concentrations of log₁₀ 6.0, 8.0 and 9.0 cells mL⁻¹. As a control, pure tap water was used. For each concentration, six strawberry plants were treated. All leaves that were present at the time of the treatment were labeled with stripes of adhesive foil to distinguish them from leaves grown after the application of the bacterial suspension. In the subsequent 5 weeks at intervals of 4 days, samples of treated leaves were taken. From each plant, one leaf was picked for each sample. The samples were stored at −20 °C until the end of the sampling period, and then DNA was extracted from the samples and qRT-PCR was performed as described above. Comparing the measured C_t values with the calculated calibration curve, the abundance of M. extorquens DSM 21961 on the leaves over the time period of sampling was determined.

CLSM

Strawberry plants were grown as described above, and at their flowering time, they were sprayed with a suspension of M. extorquens ATCC 55366 GFP5 (Choi et al., 2006) with a concentration of log₁₀ 8.0 cells mL⁻¹. After 2 weeks, leaf samples were taken and analyzed using a confocal laser scanning microscope Leica TCS SPE (Leica Microsystems, Wetzlar, Germany). Plant tissues were excited with a 635 nm laser beam and the autofluorescence emitted in the range 650–690 nm was recorded. The green fluorescent protein (GFP) was excited with a 488 nm laser beam and the detection window was optimized for every field of view, in order to gain a better discrimination between the signals and the noise. Z-stacks were acquired with a Z-step of 0.40.8 μm using the objectives × 40 or × 63 (Leica). The fluorescent signals from GFP and from plant tissues were acquired sequentially.

Quantification of DMHF in strawberries

In an independent experiment, the DMHF concentration in strawberries grown on methylobacterium-treated and untreated plants was determined and compared. Strawberry plants were grown as described above and also the treatment of the plants with suspensions of M. extorquens DSM 21961 was performed with the same cell concentrations as described. For each concentration, 16 plants were treated. Ripe strawberry fruits were harvested and fruits of four consecutive days were pooled and frozen at −20 °C. After the harvest of all ripe strawberries, a leaf sample of each variant was taken and the abundance of M. extorquens DSM 21961 was determined as described above.

The quantitative determination of DMHF was performed by GC–MS after solid-phase extraction (SPE). Method development has been described elsewhere (Siegmund et al., 2010). GC was selected as the separation technique as, with the polar stationary phase used, a rapid simultaneous determination of DMHF and DMMF with sufficient sensitivity could be achieved. For the extraction of the analytes, 2 g homogenized strawberries were mixed thoroughly with 25 mL buffer solution (KH₂PO₄ and Na₂HPO₄·2H₂O, pH 5.0). Maltol (>98%, Sigma-Aldrich, Steinheim, Germany) was added to the system as an internal standard at a concentration of 5 mg kg⁻¹ strawberries. After centrifugation of the slurry, an aliquot of 5 mL was transferred onto the SPE cartridge (Strata X, polymeric reversed phase, particle size 33 μm, 500 mg/3 mL; Phenomenex Ltd, Germany). The SPE cartridge was conditioned with 5 mL methanol (for residue analysis, Promochem, Wesel, Germany) and 5 mL deionized water before use. The elution of the analytes was performed with acetone (for residue analysis, Promochem); whereas 1 mL was discarded, 2 mL was collected and used for the GC–MS analysis.

For the GC–MS measurements, a Hewlett Packard system (HP G1800A GCD System, electron impact ionization) was used. The capillary column used was a DB Wax (column dimensions 30 m × 0.25 mm × 0.5 μm, Agilent Technologies Inc., Santa Clara, CA). Helium (99.999%, Air Liquide, Graz, Austria) was used as a carrier gas. The GC conditions were as follows: injector temperature 220 °C and detector temperature 240 °C; a constant flow method (40 kPa at 60 °C) was used starting at a temperature of 60 °C (hold time 1 min) at a temperature rate of 8 °C min⁻¹ to 245 °C (hold time 5 min). The splitless injection mode was used, the split valve being opened after 1 min. Electron impact ionization was used (70 eV). Data were acquired in the selected ion mode. Identification of the compound was performed via the calculation of the linear temperature programmed retention index (RI) and comparison with the RI of the database as well as the data obtained from the reference compound. In addition, the identity of the compound was cross-checked by acquiring data in the scan modus and interpretation of the mass spectra obtained. Quantification of the compounds was performed via the ratio of the response of the internal standard to the analytes' response at four different concentration levels. The performance of the entire sample preparation procedure was cross-checked by performing various standard addition experiments (i.e. addition of DMHF at four different concentration levels to the pure buffer solution as well as by addition to real strawberry samples; all determinations were performed in true replicate). Recovery for DMHF was determined to be 92 ± 3%. During the course of the quantification procedure, possible variations in the recovery were corrected via the internal standard. The quantification procedure was fully validated in a concentration range from 0.25 to 200 mg DMHF kg⁻¹ strawberries using an excel macro especially designed for the validation of analytical procedures (validata version 3.02.48). All given numbers are mean values from duplicate extraction of the strawberry samples. Each extract was analyzed twice by GC–MS.

Results

Isolation and characterization of Methylobacterium strains

Methylobacterium strains were isolated from different host plants by imprinting leaves and roots on MIS agar plates (see selection of strains in Table 1). All strains were able to grow on 1,2-propanediol. The biosynthesis rate of lactaldehyde differed from strain to strain and ranged from 20 to 855 ng per log₁₀ 6.0 cells. Strain M. extorquens DSM 21961, which had the highest production rate, was selected for further investigations.

Design of a specific primer for M. extorquens DSM 21961

To find discriminating regions in the genome of M. extorquens DSM 21961, BOX-PCR fingerprints with the target strain and different reference strains of the genus Methylobacterium were performed (Fig. 1). For sequencing, PCR products of M. extorquens DSM 21961 were cloned into a pGEM^®-T vector. Primer design with the obtained sequences was performed using the software package primer express 5.0 (Applied Biosystems). From the different primers and probes suggested by the software, the following set was chosen: forward primer Ext-f 5′-AGC ATC GCG AGC TCT GGT A-3′, reverse primer Ext-r 5′-CGA AAC GTC ACT GAT CGT ATG AG-3′ and probe Ext-p 5′-FAM – CTG GAT GCC GGA CTT GGC TCG TC – TAMRA-3′. These primers target a sequence with a length of 92 bp. The gap between the forward primer and the probe is 8 bp.

Determination of the specificity and limit of quantification and calibration of the system

The specificity of the primer probe system was tested against 25 different isolates of the genus Methylobacterium. Methylobacterium extorquens DSM 21961 yielded a C_t value of 20, while all the other isolates, except Hab1, yielded C_t values higher than 35 (Fig. 2). Comparison of 16S rRNA gene sequences and BOX patterns of Hab1 and M. extorquens DSM 21961 indicated that both isolates were genetically identical (see Fig. 1). In general, C_t values above 35 were considered as negative because no template controls and even empty PCR vials also occasionally yielded results in this range. This effect was also observed by other authors (Lloyd-Jones et al., 2005; Pujol et al., 2005), and has been considered as negligible if the obtained C_t values are beyond the range of calibration.

The test of the system with a pure DNA extract of M. extorquens DSM 21961 showed a linear correlation of the C_t with the amount of template DNA over a range of more than seven orders of magnitude (Fig. 3). The limit of quantification of the system is about 600 fg of template DNA per PCR. This corresponded to a mass of DNA of approximately 100 bacterial cells, if the size of the genome of M. extorquens is assumed to be 5.5 Mbp with a GC content of 68.2% (cf. GenBank accession number CP000908). Also, with leaf samples, which were supplemented with bacterial cells, the linear range ranges from an abundance of log₁₀ 4.0 cells g⁻¹ fresh leaves to an abundance of >10.0 cells g⁻¹. If only C_t values <35 are considered as reliable, the limit of quantification is log₁₀ 4.6 cells g⁻¹ fresh weight. A reduction of the limit of quantification is prevented by the large amount of plant DNA, which is coextracted during the procedure of sample processing. Optimization experiments have shown that large amounts of nontemplate DNA can inhibit the PCR (data not shown). Therefore, the total amount of DNA that was applied in one PCR was limited to 10 ng in total. Additionally, coextracted plant secondary metabolites may have inhibiting effects on the PCR.

Monitoring of M. extorquens DSM 21961 on strawberry leaves and the quantification of DMHF in strawberries

It was possible to monitor the abundance of M. extorquens DSM 21961 on strawberry leaves over a time period of 35 days. The variants that were sprayed with cell suspensions with concentrations of log₁₀ 8.0 and 9.0 cells mL⁻¹ yielded results within the calibration range of the system during the entire monitoring period (Fig. 4). Application of the log₁₀ 8.0 variants resulted in abundances between 6.5 cells g⁻¹ of leaves at the beginning of the monitoring period and 6.0 cells g⁻¹ of leaves at the end. The log₁₀ 9.0 variants resulted in abundances between 8.7 and 7.3 cells g⁻¹ of leaves. The control variant that has been treated with tap water only yielded C_t values below the limit of quantification of the system. The variant that was treated with a suspension with log₁₀ 6.0 cells mL⁻¹ resulted in values below or slightly above the limit of quantification. Therefore, the data of this variant are not shown in Fig. 4.

In order not to influence fruit quality by the removal of leaves from the plants for the monitoring of the abundance of M. extorquens DSM 21961, the concentration of DMHF in strawberries of treated plants was evaluated in an independent experiment. The results of this experiment indicated a significantly increased concentration of DMHF in fruits grown on plants, which were treated with M. extorquens DSM 21961 during the stage of flowering (Table 2). The survival of M. extorquens DSM 21961 on the plants until the time of harvest was shown by qRT-PCR.

Colonization pattern of M. extorquens DSM 21961

Microscopy confirmed the potential of M. extorquens to colonize strawberry plants densely. Two weeks after application a high colonization of the leaf surfaces was observed (Fig. 5). On the upper leaf surfaces, an accumulation of bacterial cells in dents between epidermal cells of the plant was observed (a). However, bacterial cells were not visible in the leaf endosphere (b). On the lower leaf surface, especially trichomes were heavily colonized by M. extorquens (c).

Discussion

Methylobacterium extorquens is a ubiquitous colonizer of the plant phyllosphere and a symbiont for many plants (Delmotte et al., 2009). All Methylobacterium strains from different plant sources isolated in this study were able to produce a chemical precursor of furanoic compounds in vitro. To study the importance of this biosynthesis for the flavor of strawberry fruits, we investigated the interaction between the selected M. extorquens DSM 21961 with strawberries ad planta. To monitor the colonization, a specific Taqman qRT-PCR system was developed. In addition, the function of M. extorquens DSM 21961 was evaluated by measuring the concentration of the furanonic compound DMHF in fruits of treated plants. Microscopic analysis by CLSM confirmed the high colonization rate, which is necessary for a beneficial plant–microorganism interaction.

In contrast to cultivation-dependent methods, qRT-PCR offers a number of advantages, for example high specificity, low limit of quantification and the possibility of using genetically unmodified wild-type strains. To circumvent the time-consuming process of optimization of a RAPD-PCR for providing suitable target regions for primer design in the target genome, we used an alternative BOX-fingerprinting technique, which does not require any optimization of primers or PCR conditions. The favorable specificity of the system was determined with different isolates of Methylobacterium, isolated from the phyllosphere of different plants. The limit of quantification of the system with pure DNA of M. extorquens DSM 21961 was determined to be around 100 cells per PCR, resulting in a C_t value of 35. This value is higher than the values indicated by other authors in the literature (Abbasi et al., 1999; Hristova et al., 2001). On the other hand, it represents the bottom end of the linear range of the calibration function and not the limit at which a positive sample can just be distinguished from a control sample. This often is not indicated clearly in the literature. The limit of quantification with samples of strawberry leaves that were supplied with different cell concentrations is log₁₀ 4.6 cells g⁻¹ fresh leaves. This limit is comparable to the detection limit of other TaqMan systems indicated in the literature (Salm & Geider, 2004).

The quantification system was applied in a monitoring experiment of M. extorquens DSM 21961 on strawberry plants under greenhouse conditions over a time period of 5 weeks. Cell suspensions with three different concentrations were sprayed onto the plants, and the abundance of M. extorquens DSM 21961 on the leaves was monitored for 5 weeks. Applied concentrations of log₁₀ 8.0 and log₁₀ 9.0 cells mL⁻¹ resulted in a stable population, while untreated control plants yielded no signals within the calibration range of the system. A concentration of log₁₀ 6.0 cells mL⁻¹ resulted in cell concentrations on the leaves around the limit of quantification of the system. In cases where abundances are lower than log₁₀ 4.6 cells g⁻¹ fresh leaves, the method of DNA extraction from leaves will have to be optimized. Optionally, plant DNA should be removed in the extraction procedure or bacterial DNA should be enriched. Simply increasing the amount of template DNA in the PCR is not feasible, because an excess of nontemplate DNA in the reaction decreases the efficiency of the PCR, mainly due to the coextraction of PCR inhibitors.

CLSM confirmed a dense colonization of strawberry leaves by M. extorquens. The upper and lower leaf surfaces were occupied by bacterial cells particularly in the dents on the leaves and on trichomes. Bacterial cells were not detected in the leaf endosphere. This was surprising, because methylobacteria are well-known endophytes (Hallmann & Berg, 2006).

The effect of M. extorquens DSM 21961 on the fruit quality and the DMHF content of strawberries was studied using greenhouse-grown plants. Spraying M. extorquens DSM 21961 on strawberry plants during the flowering stage resulted in an enhanced concentration of DMHF in the strawberries in comparison with untreated plants. Survival of M. extorquens DSM 21961 on the plants, from the time of application until the harvest of fruits, was confirmed by qRT-PCR. The DMHF concentrations determined underline the hypothesis of an interaction of methylobacteria and strawberry plants that leads to an increased DMHF biosynthesis rate of the strawberry plants. Further investigations are needed to understand the interaction in detail, especially the phenomenon that the colonization of the phyllosphere by methylobacteria can enhance the flavor of the fruit, and to optimize the Methylobacterium–strawberry interaction for biotechnological applications. However, the tools for analysis as developed and presented in this study are suitable for this purpose.

Acknowledgements

The work has been financed in part by the Austrian Research Promotion Agency (FFG) and by Grünewald Fruchtsaft GmbH.

References

Author notes