Editor: Christoph Tebbe
Multiple physiological states of a Pseudomonas fluorescens DR54 biocontrol inoculant monitored by a new flow cytometry protocol
Version of Record online: 16 DEC 2008
© 2008 Federation of European Microbiological Societies. Published by Blackwell Publishing Ltd. All rights reserved
FEMS Microbiology Ecology
Volume 67, Issue 3, pages 479–490, March 2009
How to Cite
Nielsen, T. H., Sjøholm, O. R. and Sørensen, J. (2009), Multiple physiological states of a Pseudomonas fluorescens DR54 biocontrol inoculant monitored by a new flow cytometry protocol. FEMS Microbiology Ecology, 67: 479–490. doi: 10.1111/j.1574-6941.2008.00631.x
Present address: Tommy Harder Nielsen, Division of Vaccine, Statens Serum Institut, Artillerivej 5, DK-2300 Copenhagen S, Denmark.
- Issue online: 2 FEB 2009
- Version of Record online: 16 DEC 2008
- Received 13 March 2008; revised 21 October 2008; accepted 3 November 2008.First published online 16 December 2008.
- flow cytometry;
- clay carrier;
- seed formulation;
- seed germination;
A new fluorescence staining and flow cytometry protocol was developed to monitor several physiological states in biocontrol strain Pseudomonas fluorescens DR54 during storage survival in a stationary-phase culture, preparation of clay carrier for seed formulation, and establishment in a sugar beet spermosphere. The high load of impurities in the environmental samples was dealt with by adding a density-gradient purification step to the staining protocol. Staining by SYBR Green, combined with either propidium iodide or ethidium bromide (EB)+DiBAC(4)3, was used to quantify the total cell population and further divide this population into: (1) intact cells with an unaffected membrane and energy metabolism. (2) De-energized cells unable to maintain membrane export (EB exclusion). (3) Depolarized cells unable to maintain membrane potential. (4) Permeabilized cells with a damaged membrane. During both stationary-phase storage and steps for preparation of formulation carrier, loss of intact P. fluorescens DR54 cells was quantitatively accounted for by depolarized and permeabilized states. Surviving inoculum cells subsequently proliferated on the germinating seeds, but with a surprisingly high abundance of de-energized cells. The new protocol is the first for flow cytometry to include a recording of both intact and several subpopulations of physiologically affected bacteria in complex, environmental samples with high impurity loads.
To determine whether bacteria are affected by biological, chemical or physical impacts in their environment, several staining protocols associated with fluorescence microscopy and flow cytometry have been developed during the last 10–20 years (e.g. Davey & Kell, 1996; Veal et al., 2000). Considerable attention has been paid towards development of staining protocols that offer detailed information on the actual physiological state of bacterial cells in the environment, including those occurring under starvation and stress conditions. Hence, it may be insufficient to distinguish between metabolically active and inactive cells and there are presently stains available to detect several physiological states based on the structural integrity and function of the membrane.
Respiring cells that cannot maintain a functional electron transport activity will eventually become de-energized; such a phase can be detected on challenging the cells with ethidium bromide (hereafter referred to as EB), which penetrates the cytoplasmic membrane and binds to DNA if not actively exported by a proton–antiport transport system (Midgley, 1987). A de-energized cell may in turn be unable to maintain the membrane potential and may thus become depolarized; this phase can be detected on challenging the cells with DiBAC4(3) (bis-(1,3-dibarbituric acid)-trimethine oxanol), also called BOX stains (Jepras et al., 1995; Suller & Lloyd, 1999). Finally, a loss of the structural membrane integrity may result in permeabilized cells, typically taken to represent dead cells; this phase can be detected on challenging the cells with propidium iodide (hereafter referred to as PI), which enters and binds to DNA in cells when its uptake is not excluded by an impermeable cytoplasmic membrane (Jepras et al., 1997).
The combination of stains in dual staining offers a unique possibility to study several physiological stages in mixed populations at the same time, for example following a microbial culture or an inoculant population differentiating into starving and stressed subpopulations of cells. The combination of PI with another DNA stain such as SYTO 9 is the basis for the well-known, commercial Live-Dead kit used to separate permeabilized (‘dead’) and other (‘live’) bacteria in natural environmental samples. SYBR Green and PI have earlier been used for determining total and permeabilized cells (Barbesti et al., 2000), and PI has also been used in combination with BOX to separate permeabilized and depolarized cells (Hewitt et al., 2004a, b) or with FDA to separate permeabilized and metabolizing cells (Jones & Senft, 1985). In addition, a triple-staining protocol based on EB (orange fluorescence), BOX (green fluorescence) and PI (red fluorescence) has been developed for flow cytometry and detection of de-energized (stained by EB), depolarized (stained by EB+BOX) and permeabilized (stained by EB+BOX+PI) subpopulations of cells (Nebe-Von-Caron et al., 1998, 2000). This widely used triple stain of EB, BOX and PI has quantified de-energized (stained by EB), depolarized (stained by EB+BOX) and permeabilized (stained by EB+BOX+PI) cells during aging of cultures or treatment of cultures with antibiotics or heat (Hewitt et al., 1999, 2000; Nebe-Von-Caron et al., 2000; Hewitt & Nebe-Von-Caron, 2001; Amanullah et al., 2002a, b, 2003). The triple stain has also been used to determine the culturability of the different subpopulations using the cell sorter of the flow cytometer to subsequently transfer each of the subpopulations directly onto nutrient agar plates; for a 25-day-old stationary-phase culture of Salmonella typhimurium, it was thus shown that 85% of the de-energized, 34% of the depolarized and none (0%) of the permeabilized cells were able to form colonies (Nebe-Von-Caron et al., 1998).
A major drawback of the popular triple-staining protocol is that intact (nonaffected) cells will only stain faintly because their uptake of the dyes is very limited (Nebe-Von-Caron et al., 1998). In flow cytometry, the intact cells will therefore be recognized only by very low fluorescence signals (Nebe-Von-Caron et al., 1998), gating together with a low background fluorescence from sample impurities (and eluent). In the present work, inclusion of one additional stain, SYBR Green (hereafter referred to as SYBR), in samples already stained by EB+BOX or PI is documented to allow for direct quantification of a total of four categories of cellular subpopulations in complex samples with even high loads of impurities. The new staining protocol is applied to a differentiating population of the biocontrol strain Pseudomonas fluorescens DR54 in a stationary-phase culture, in wet and dry montmorillonite clay-carrier formulations, and in a sugar beet spermosphere during seedling germination. Detection and enumeration by the new protocol for flow cytometry was feasible for both the total bacterial population and its subpopulations of intact, de-energized, depolarized or permeabilized cells in the complex, natural samples.
Materials and methods
Pseudomonas fluorescens DR54 was grown to the early stationary phase in Luria–Bertani broth (LB: g L−1 10 tryptone, 5 yeast extract, 10 NaCl, 1 glucose) at 28 °C. To observe the physiological effects of stress treatments, the cell suspension in Dulbecco's phosphate buffered saline (DPBS: g L−1 8 NaCl, 0.2 KCl, 1.43 Na2HPO4·2H2O, 0.2 KH2PO4 and pH 7.2) was heat-treated for 15 min at 60 °C. A similar cell suspension was treated with the membrane ionophore CCCP (carbonyl cyanide 3-chlorophenyl hydrazone) at 15 μM (Novo et al., 1999) for 30 min at 4 °C. Finally, a batch culture was grown overnight from an initial OD600 nm of 0.1 and subsampled during the late exponential phase (Day 1) and stationary phase (Days 2, 3, 4 and 8). Culturability (CFU count) was determined in all experiments after plating on 1/10 LB agar and incubation for 2 days at 28 °C. Another sample fraction was stained for flow cytometry as described below.
Montmorillonite (SAz-1) was obtained from Source Clays Repository (Columbia, MO) and saturated with Ca2+ as described by Morra et al. (1998). Pseudomonas fluorescens DR54 cells from an overnight culture grown in LB at 28 °C were washed twice in 1.2 mM K2HPO4 (pH 7.0). Five grams of sterile montmorillonite was subsequently mixed into 2.5 mL washed cell suspension to give a final water content of 50% by weight and 3 × 1010 CFU g−1. The cells were gently homogenized into the carrier in a sealed polyethylene plastic bag and left for 30 min. Triplicate samples of 0.5 g wet carrier were then extracted in 3 mL DPBS for 30 min using a horizontal shaker (IKA KS 260 basic) at 300 r.p.m.; the sample extract (termed wet carrier) was subsequently used for culturability and flow cytometric analysis. In parallel, a portion of the wet carrier with P. fluorescens cells was subsequently desiccated in closed containers equipped with saturated CaCl2; the water content of the carrier was monitored and reached 20% after 48 h. Triplicate sample extracts (termed dry carrier) were then taken as described for the wet carrier.
Before flow cytometry, all sample extracts were purified by density centrifugation (10 000 g, 30 min at 4 °C) in tubes containing a cushion of Nycodenz (Medino, Denmark) with a density of 1.3 g mL−1 (Lindahl & Bakken, 1995). The bacteria-containing layer developing above the cushion was collected carefully and placed briefly on ice. One aliquot of the purified layer was then serially diluted to determine the Nycodenz purification efficiency based on CFU counts. Control experiments comparing Nycodenz-purified cells from a liquid culture with nonpurified cells showed that the purification step led to insignificant changes in the relative proportions of intact, de-energized and depolarized cells by flow cytometric analysis (data not shown). The rest of the purified sample was stained for flow cytometry as described below.
Sugar beet seedling experiments
Pseudomonas fluorescens DR54 cells were coated onto sugar beet seeds (Madison variety) at 103 cells per seed according to a standard formulation protocol (Danisco Seed Patent no. EP1730259), except that formulated seeds were prepared without the layer of dye that is normally applied in commercial production. Inoculum-free control seeds were prepared using a sterile 0.9% NaCl solution instead of P. fluorescens DR54 culture. Seedlings were developed in closed containers (seed germination boxes) with multiple cavities containing a bottom layer of preweighed, moist (sterile tap water) paper. Seeds were incubated at 12 °C for 1, 2, 3, 4 and 6 days for germination and seedling development.
During the whole period, germination was recorded by appearance of cracks or visual sprouts protruding from the seed coatings. On the sampling dates, triplicate samplings of 10 seeds each were taken into 3 mL DPBS (pH 7.2) and extracted for 30 min on the horizontal shaker. Sample extracts were serially diluted and plated onto LB medium; CFUs were counted after incubation for 2 days at 28 °C. Another sample fraction was purified by Nycodenz and stained for flow cytometry as described below.
The four fluorescent stains used in this study were: (1) SYBR (Sybr Green I; Molecular Probes, Eugene, OR), staining the total cell population green; the original 10 000 × solution was diluted to a 100 × working solution in sterile TE buffer (10 mM Tris-HCl, 1 mM EDTA; pH 7.5). (2) EB (Fluoprobes, Bie and Berntsen, Denmark), staining the de-energized cells orange-red; EB stock solution was made in sterile Milli-Q water (3 mg mL−1) and used in a working solution of 12.5 μg mL−1 in DPBS (pH 7.2) containing 1 mM EDTA. (3) BOX [DiBAC4(3); Molecular Probes], staining depolarized and permeabilized cells green; BOX stock solution of 3 mg mL−1 DMSO was diluted to a working solution of 10 μg BOX mL−1 in DPBS (pH 7.2) with 1 mM EDTA. (4) PI (Molecular Probes), staining permeabilized cells red; PI stock solution was made in sterile Milli-Q (1 mg mL−1) and used in a working solution of 10 μg mL−1 in DPBS (pH 7.2) containing 1 mM EDTA. All solutions were filter centrifuged using 0.2-μm microcentrifuge filter tubes (Anopore, Cat no. 6830 0202, Whatman, UK) and used as described by Nebe-Von-Caron et al. (1998).
Staining was performed with one or more of the fluorescent dyes in the following combinations: PI, SYBR+PI, EB+BOX and SYBR+EB+BOX. First, an aliquot of 25 μL of a sample diluted in DPBS was stained for 10 min with 12.5 μL EB and 12.5 μL BOX in combination or with 25 μL PI alone. The stained samples were then diluted to a total volume of 1 mL in DPBS and analyzed in the flow cytometer as described below. After recording in the flow cytometer, both sets of samples were amended with 10 μL SYBR and stained for 10 min before the samples were analyzed again in the flow cytometer.
Flow cytometer analysis was performed with a Facs Calibur instrument (Becton Dickinson, CA) equipped with 488-nm excitation from an argon-ion laser at 15 mW and three detectors: FL1 (515–545 nm), FL2 (564–606 nm) and FL3 (above 670 nm). The carrier liquid was 0.2 μm-filtered Milli-Q water. Flow velocity was determined by analyzing true-count standard samples (microspheres 2 μm; Molecular Probes) before and after sample analysis. True-count and experimental samples were supplemented with 2.0-μm yellow-green Fluospheres-carboxylate microspheres (F-8827; Molecular Probes) as internal standards for flow control. Cells were quantified using either a low amplification for forward scatter (FSC-E02V), side scatter (SSC-380V), green detector (FL1-535V), orange detector (FL2-629V), red detector (FL3-659V) with threshold SSC-90V or high-amplification settings (FSC-E02V, SSC-380V, FL1-700V, FL2-780V and FL3-800V) with threshold SSC-90V, resulting in similar counts (P<0.05). The low-amplification setting used as a routine allowed for the sequential analysis of PI or EB+BOX stainings, followed by the SYBR staining without changing the instrument settings. In all cases, the spectral overlap between any emitted fluorescence of different stains could be left uncompensated for as both BOX and SYBR dyes were found to stain cells detected by the FL1 detector.
The stains used in this work emit light detectable in the FL1 channel for green light (515–545 nm) or in the FL3 channel detecting red to infrared light (>670 nm). When staining with SYBR+BOX+EB, both SYBR and BOX are detected in FL1, but they stain differently: SYBR is membrane permeable and stains DNA and RNA in all cells, while BOX only enters cells with a depolarized membrane, staining the proteins and membranes in such cells (Hewitt & Nebe-Von Caron, 2001). (1) Intact cells can be separated from de-energized and depolarized cells because they only accumulate SYBR, yielding a high FL1 signal. (2) De-energized cells accumulate EB, resulting in a high FL3 signal, as well as SYBR. As EB binds to the DNA it is assumed that some of the DNA-binding places for SyBr are occupied by EB, resulting in a reduced FL1 signal compared with the intact cells. (3) Depolarized cells bind SYBR, EB and BOX. The binding of SYBR and EB is similar to that of de-energized cells, resulting in a high FL3 signal and a reduced FL1 signal. However, the additional binding of BOX to proteins and membranes will increase the FL1 signal, leading to a higher FL1 signal compared with the de-energized cells. Hence, the depolarized cells have both a high FL1 and an FL3 signal. (4) Permeabilized cells can specifically take up and bind PI.
Data collected in the flow cytometer program cell quest pro were used for quantitative measures, and the results were imported into the PC program winlist for contour plotting as outlined by the manufacturer (Verity Software House Inc., ME). In order to compensate for cell loss during the Nycodenz purification, the flow cytometry data were corrected by the ratio of CFU counts before and after Nycodenz purification.
Cell numbers determined by CFU or flow cytometry were log-transformed and significant differences (P>0.05) between treatments were tested using the two-way anova and the least-squares means of pairwise comparisons using the glm procedure of SAS analyst (SAS Institute Inc., Gary, NC). Each subpopulation of intact, de-energized, depolarized or permeabilized cells was calculated and presented as a fraction (%) of the total cell count in the flow cytometer.
Verification of staining specificity for physiological states
The flow cytometric data from the 3-day-old, stationary-phase culture of P. fluorescens DR54 stained by SYBR+PI in combination and PI alone document the separation of cells from impurities and the differentiation of cells into several categories (Fig. 1). Using the SYBR+PI combination, the total cell population is easily gated and separated from sample impurities as shown by the FL1–FSC plot (Fig. 1a), and the population is divided into permeabilized and other (nonpermeabilized) cells in the FL1–FL3 plot (Fig. 1b). A control was included to demonstrate that a short heat–treatment (60 °C, 15 min) resulted in complete membrane permeabilization (Fig. 1c), while such an effect was not observed using only a traditional membrane ionophore (15 μM CCCP, 4 °C, 30 min) (data not shown). By comparison, staining by PI alone and using an FL2–FL3 plot of untreated (Fig. 1d) and heat-treated (Fig. 1e) samples resulted in detection of permealized cells, while all other (nonpermeabilized) cells would be masked in the impurities fraction. Overall, the results demonstrate that coupled SYBR+PI staining in combination with flow cytometry can detect and quantify the fractions of both permeabilized and nonpermeabilized cells, distinguishable from sample impurities.
Coupled SYBR+EB+BOX staining was used to further divide the total cell population into three subfractions of intact (nonaffected), de-energized and depolarized cells, respectively, in the 3-day-old stationary-phase P. fluorescens DR54 culture (Fig. 2). Also, the combined SYBR+EB+BOX staining shows easy gating in the FL1–FSC plot and thus an adequate separation of the total cell population from sample impurities (Fig. 2a). In the FL1–FL3 plot of the untreated sample (Fig. 2b), the three subpopulations of cells could be distinguished by their ‘contour hills’ and their identity was verified by the heat and CCCP uncoupler treatments. Hence, the heat treatment (60 °C) resulted in complete conversion of all cells into the depolarized state. By comparison, the CCCP treatment increased the fraction of de-energized cells; depolarization was undetectable in these cells. We infer that the untreated stationary-phase culture contained intact, de-energized and depolarized cells, and that CCCP uncoupler treatment converted intact into de-energized cells, whereas heat treatment converted both intact and de-energized cells into depolarized ones. By comparison, staining with EB+BOX alone and using an FL1–FL2 plot of the untreated (Fig. 2e), heat-treated (Fig. 2f) and CCCP-treated (Fig. 2g) samples gave results comparable to the SYBR+EB+BOX staining as for the de-energized and depolarized cells, but also a much poorer distinction of the intact population. While still visible in the untreated culture (Fig. 2e), the intact cell fraction is essentially unstained here and thus closely associated with the remaining impurities providing low background signals in FL1 and FL2.
Dynamics of the physiological state during the stationary phase
To compare the new staining protocol with the traditional one, we compared the cell differentiation over an extended period of time in the P. fluorescens DR54 stationary-phase culture (Table 1). Statistical analysis using log-transformed numbers showed no significant difference between staining with SYBR+EB+BOX compared with EB+BOX and with SYBR+PI compared with PI (P<0.05).
|Incubation period (days)||Deenergized (SYBR+BOX+EB)||Deenergized (EB+BOX)||Depolarized (SYBR+EB+BOX)||Depolarized (EB+BOX)||Permeabilized (SYBR+PI)||Permeabilized (PI)|
|1||0.4 ± 0.1||0.2 ± 0.0||1.7 ± 0.5||1.0 ± 0.5||0.7 ± 0.1||0.4 ± 0.1|
|2||0.6 ± 0.2||0.4 ± 0.2||1.3 ± 0.2||1.0 ± 0.6||0.9 ± 0.4||0.8 ± 0.3|
|4||3.7 ± 1.6||4.5 ± 1.4||18 ± 1||20 ± 2||17 ± 1||19 ± 2|
|8||0.4 ± 0.0||0.3 ± 0.2||16 ± 0||18 ± 1||21 ± 2||20 ± 1|
The complete time course for the changes in the stationary phase (Fig. 3a) demonstrated that the total cell counts as obtained in the flow cytometer showed some fluctuation but appeared to be constant at c. 1010 cells mL−1. In contrast, culturability (CFU counts) was only constant at 8–9 × 109 cells mL−1 at Days 1 and 2 (entrance to stationary phase) before a steady decrease, reaching c. 3 × 109 and 108 cells mL−1 at Days 4 and 8, respectively. Most interestingly, the flow cytometric recording of intact cells (Fig. 3b) equalled that of the culturable cells (CFU) during the late exponential and the early stationary phase (P<0.05); the intact cells represented a fraction of c. 80% at Days 1 and 2, c. 30% at Day 4 and c. 2% at Day 8. By comparison, the fraction of de-energized cells was always 5–10% (Fig. 3b). Further, the fractions of both depolarized and permeabilized cells increased from 10% to 20% of the total population in the early stationary phase to a final high of c. 95% after 8 days of incubation (Fig. 3b). Overall, the results suggested that the two subpopulations of depolarized and permeabilized cells were actually identical, and none of our recordings within this study indicated that depolarized and permeabilized cell populations were statistically dissimilar. Hence, the loss of membrane potential (depolarization) seemed tightly coupled to the loss of membrane barrier function (permeabilization), which has in turn been associated with cell death (Live-Dead commercial kit). In other terms, all depolarized cells accumulating during the stationary phase were seemingly dead cells as deduced from the specificity of the staining protocols. Following this observation, the loss of intact cells and culturability in the present P. fluorescens DR54 stationary-phase culture is quantitatively accounted for by the accumulation of depolarized and permeabilized cell indicating immediate cell death, except for only a small subpopulation of cells representing a de-energized state.
Changes in physiological states during seed formulation and germination
The 4-day-old incubations of noninoculated sugar beet seeds (without P. fluorescens DR54) showed insignificantly low counts when gated in FSC–FL1 (Fig. 4a and b). As expected, impurities gave a large background of FSC and a relatively low intensity of green fluorescence of the SYBR stain when combined with PI or EB+BOX. Nevertheless, the FSC–FL1 gating of SYBR-stained cells clearly contained the total cell population for both SYBR+PI- and SYBR+EB+BOX-stained cells of the 4-day-old extracts as shown in Fig. 4c and d, respectively. With this gate setting, the SYBR+PI stain detected relatively few permeabilized cells (Fig. 4e), while the SYBR+EB+BOX stain demonstrated significant populations of both intact, de-energized or depolarized cells, respectively (Fig. 4f).
A complete cycle of P. fluorescens DR54 differentiation during the preparation of the inoculum (culture), wet clay carrier and dry clay carrier for seed formulation and the subsequent seed germination (Day 4) may be compared (Fig. 5). The inoculum had c. 80% intact cells and 10% depolarized and permeabilized cells, respectively. Sample extracts from the wet montmorillonite clay carrier had a low (17%) recovery of the total cell population, but 30% of the cells still remained intact at this stage, while 2% was de-energized and 65–70% was depolarized and permeabilized. During the subsequent desiccation to prepare the dry montmorillonite carrier, extraction efficiency was further reduced to c. 3%; the frequency of intact cells had now decreased to 10%, de-energized cells were still infrequent (5%) and depolarized and permeabilized cells had increased to 80–85%. Finally, from the low inoculum density of 103 cells per seed, P. fluorescens DR54 showed a dramatic growth over 4 logarithmic units on the incubated seeds, reaching a density of culturable cells at Day 4 of almost 107 cells per seed. Interestingly, the fraction of de-energized was now predominant, amounting to 50%, while that of intact cells was 20% and the fractions of depolarized and permeabilized were 10–20%. Following these observations, it appears that the preparation of the carrier for formulation leads to a sequential loss of intact cells matched by accumulation of depolarized and permeabilized cells. However, the surviving fraction (here 10%) proliferates during seed germination; most surprisingly, however, half of the new cell population on the seeds is in a de-energized state.
Staining protocol to study physiological differentiation in bacterial populations
Despite the obvious advantages of combining three dyes in one staining, the traditional triple EB+BOX+PI does not easily allow for detection of the subpopulation of intact (nonaffected) cells in environmental samples because this cell fraction will only take up small amounts of dye and will thus be typically gated together with sample impurities when analyzed by flow cytometry. The new protocol with the additional SYBR stain, in combination with either EB+BOX or PI, gave adequate gating to separate the intact cells, together with the three, physiologically affected categories. It was necessary to perform a first step of SYBR+BOX+EB staining and analysis, followed by PI addition and SYBR+PI analysis, but this was seen as a small sacrifice for obtaining data for the intact cell population.
Evidence further supported that SYBR had no effect on the recordings compared with the traditional protocol from the P. fluorescens stationary-phase storage cultures; during the 8 days, there were no significant differences between data based on SYBR+EB+BOX or EB+BOX to detect de-energized or depolarized cells or data based on SYBR+PI or PI to detect permeabilized cells (P<0.05). As SYBR was detected in the FL1 channel together with BOX, it was necessary to remove the compensation setting normally used in traditional triple-stain analysis.
Stain specificity for the physiological states is clearly important for the EB, BOX and PI dyes, but it must be remembered that these and other stains may show different uptakes (and thus results in the present analysis) in different bacteria. Hence, among the many species investigated by Amanullah et al. (2003), not all of them were capable of excluding EB from entering the cell; ideally, this should be verified for each investigated species. Such an exclusion system seemed intact in P. fluorescens DR54, however, because addition of the respiratory uncoupler (CCCP) resulted in EB uptake (assumed to occur in an affected proton–antiport system), while a harsher cell treatment by heat resulted in BOX uptake (taken to occur in cells without membrane potential). The CCCP ionophore thus clearly converted the intact cells to de-energized cells (or partially depolarized due to uncoupled respiration), while the de-energized cells were not detected by BOX. This may suggest that P. fluorescens DR54 cells need to be further impacted on their membrane to depolarize sufficiently to be detected by BOX; this hypothesis was verified on challenging the cells briefly with heat (60 °C, 15 min). Overall, it seems that DiBAC4(3) (BOX) may, in several bacteria, only detect strong depolarization after for example heat treatment while the sister dye DiOC2(3) has a different uptake and may sensitively detect weaker polarization, for example that occurring after a CCCP uncoupler treatment. Because other P. fluorescens strains have been separated into depolarized and permeabilized cell populations by combining BOX and PI (Hewitt et al., 2004a, b), great care must be taken to define depolarization properly when related to the BOX stain uptake and to provide evidence if such cells are partially or completely disrupted in their membrane potential.
Physiological differentiation of P. fluorescens DR54 in a stationary-phase culture
Storage in the stationary phase has often been chosen to study physiological differentiation due to starvation, accumulation of toxic metabolites, etc. in the spent medium. In this work, the P. fluorescens DR54 showed the typical loss of culturability (plate counts, CFU) while the total cell population density remained high, often associated with extended stationary-phase storage for bacteria without extraordinary mechanisms of stress survival. Loss of culturability in laboratory media has long been questioned as a valuable measure of viability, however, because the cells under environmental stress may convert into several, unique subpopulations, each characterized by the loss of one or several important physiological functions. One classical study is that of Joux et al. (1997), who studied S. typhimurium under extended starvation storage; over time, the culture developed a sequential series of subpopulations showing a loss of culturability, respiratory activity and membrane function before the cells were eventually lost by lysis.
In the latter study, several of the physiological states in S. typhimurium were characterized by assays of fluorescence staining as also used in the present work. The results from stationary-phase cultures of P. fluorescens DR54 presented here demonstrate several important observations for this strain: (1) culturability was closely associated with the subpopulation of cells showing uptake of SYBR, but not EB, BOX or PI (here referred to as intact cells); it is expected that these cells show the full integrity of membrane functions including respiratory activity, maintenance of membrane potential and maintenance of permeability barrier. (2) In the stationary-phase cultures, the fraction of cells showing uptake of EB (referred to here as de-energized cells) was too small and constant over time to state whether these cells (hampered in their respiration-dependent exclusion of the dye) were still culturable as defined by CFU counts. (3) Finally, the fractions of depolarized (loss of membrane potential permitting uptake of BOX) and permeabilized (loss of permeability barrier permitting uptake of PI) cells were similar, thus representing one population of cells that apparently accounted for nonculturable, dead cells. Hence, a loss of membrane potential in these cells apparently led to immediate loss of culturability and to cell death.
Physiological differentiation of P. fluorescens in a clay carrier and a spermosphere
The P. fluorescens DR54 strain has a documented antagonistic activity toward the plant-pathogenic, root-infecting microfungi P. ultimum and R. solani (Nielsen et al., 1998; Thrane et al., 2000), and therefore has potential as a biological control agent. For this and other Pseudomonas spp. biocontrol strains to be used as in practice, however, desiccation tolerance in the clay-carrier mixing process and during subsequent seed formulation must allow the organism to maintain viability before its colonization, growth and metabolic activity during seed germination and seedling root development.
When bacteria are applied directly to clay matrices for subsequent seed formulation, the cells may experience considerable stress by for example desiccation. When P. fluorescens strain DR54 was added directly to the montmorillonite carrier, a large fraction of the extractable cells had already lost their cell polarization and had become permeabilized; this fraction became even larger when the cells were desiccated in the dry clay carrier. Most important, however, was the apparent survival of intact cells (c. 10% of the total in the dry carrier); although this fraction is low, it is presumably the important subpopulation of cells showing subsequent colonization and growth in the sugar beet spermosphere. The low recovery and desiccation survival in P. fluorescens DR54 differ substantially from that of P. fluorescens strain 2–79 in montmorillonite (Morra et al., 1998); strain variation could partly be responsible for this discrepancy as differences in desiccation survival have also been reported for closely related Escherichia coli strains (Louis et al., 1994).
Commercial seed formulation with bacterial biocontrol agents is preferably based on a relatively low inoculant density, thus anticipating that cell proliferation is rapid during seed germination and followed by efficient colonization of the seedling root and rhizosphere. Detection of both intact and physiologically affected subpopulations is therefore very important to understand the population dynamics when inoculant P. fluorescens DR54 cells are exposed to their natural environment, for example plant rhizosphere or bulk soil. It was promising in this study that the relatively low inoculum of c. 4 × 103 cells per seed showed significant growth (three log-units) to reach 107 cells per seed over the first few days after the seeds were wetted for germination. Moreover, the intact (and culturable) cells represented a significant fraction at this time and remained so for at least a week (data not shown). It was concluded that a large population of intact P. fluorescens DR54 cells actually established well on the seeds and seedling roots during the first week or so, which is the critical period for biocontrol of root-infecting pathogens, for example P. ultimum and R. solani, in sugar beets.
The seed surface, often referred to as the spermosphere, is a heterogenous environment even in the present, simplified experiments where the seed is germinated in a water film rather than a most complicated soil environment. We did not follow the inoculant proliferation and establishment on the seed or the seedling root surface by microscopy, but we were aware that cracks developing in the formulation material and seed coat, outbreak of the seedling root, distribution of water, exudation of carbon substrate and nutrients, etc. may all contribute to formation of a heterogeneous microenvironment for the bacterial inoculant. Without information on preferential colonization sites on the seeds such as those observed for another Pseudomonas sp. inoculant strain colonizing wheat (Unge & Jansson, 2001), it is difficult to provide direct evidence for favorable or unfavorable microenvironments in the spermosphere. Nevertheless, the remarkably high fraction of de-energized cells on the germinating sugar beet seeds (approximately one-third of the total cell population) strongly indicates that many cells may experience a stress, leading to reduced metabolic activity. It should be most interesting to pursue this observation to see where such cells are located and to see whether they subsequently recover to fully intact cells during root development or lose further cell functions related to depolarization and permeabilization.
This study was supported by The Directorate for Food, Fisheries and Agri Business (J. no. 93S-2465-Å02-01416). We thank Hans Christian Pedersen, Danisco Seed A/S, for supplying the coated sugar beet seeds and Dorte Rasmussen for excellent technical assistance.
- 2002a) Fed-batch bioconversion of indene to cis-indandiol. Enzyme Microbial Tech 31: 954–967. , , , , , , & (
- 2002b) Application of multi-parameter flow cytometry using fluorescent probes to study substrate toxicity in the indene bioconversion. Biotech Bioeng 80: 239–249. , , , , , , & (
- 2003) Measurements of strain-dependent toxicity in the indene bioconversion using multiparameter flow cytometry. Biotech Bioeng 81: 405–420. , , , , , , & (
- 2000) Two- and three-colour fluorescence flow cytometric analysis of immunoidentified viable bacteria. Cytometry 40: 214–218. , , , , & (
- 1996) Flow cytometry and cell sorting of heterogeneous microbial populations: the importance of single-cell analysis. Microbiol Rev 60: 641–696. & (
- 2001) An industrial application of multiparameter flow cytometry: assessment of cell physiological state and its application to the study of microbial fermentation. Cytometry 44: 179–187. & (
- 1999) Use of multi-staining flow cytometry to characterise the physiological state of Escherichia coli W3110 in high cell density fed-batch cultures. Biotech Bioeng 63: 705–711. , , & (
- 2000) Studies related to the scale-up of high-cell-density E. coli fed-batch fermentations using multiparameter flow cytometry: effect of a changing microenvironment with respect to glucose and dissolved oxygen concentration. Biotech Bioeng 70: 381–390. , , , & (
- 2004a) An evaluation of the anti-bacterial action of ceramic powder slurries using multi-parameter flow cytometry. Biotech Lett 23: 667–675. , , , & (
- 2004b) A study into the anti-microbial properties of an amino-functionalised polymer using multi-parameter flow cytometry. Biotech Lett 26: 549–557. , , , & (
- 1995) Development of a robust flow cytometric assay for determining numbers of viable bacteria. Appl Environ Microbiol 61: 2696–2701. , , , & (
- 1997) Rapid assessment antibiotic effects on E. coli by bis-(1,3-dibutylbarbituric acid) trimethine oxonol and flow cytometry. Antimicrob Agents Ch 41: 2001–2005. , , & (
- 1985) An improved method to determine cell viability by simultaneous staining with fluorescein diacetate-propidium iodide. J Histochem Cytochem 33: 77–79. & (
- 1997) Succession of cellular states in a Salmonella typhimurium population during starvation in artificial seawater microcosms. FEMS Microbiol Ecol 22: 65–76. , & (
- 1995) Evaluation of methods for extraction of bacteria from soil. FEMS Microbiol Ecol 16: 135–142. & (
- 1994) Survival of Escherichia coli during drying and storage in the presence of compatible solutes. Appl Microbiol Biotechnol 41: 684–688. , & (
- 1987) An efflux system for cationic dyes and related compounds in E. coli. Microbiol Sci 4: 125–128. (
- 1998) Survival of Pseudomonas fluorescens 2-79RN10 in clay powders undergoing drying. Soil Sci Soc Am J 62: 663–670. , , & (
- 1998) Assessment of bacterial viability status by flow cytometry and single cell sorting. J Appl Microbiol 84: 988–998. , & (
- 2000) Analysis of bacterial function by multicolour fluorescens flow cytometry and single cell sorting. J Microbiol Meth 42: 97–114. , , , & (
- 1998) Secondary metabolite- and endochitinase-dependent antagonism toward plant-pathogenic microfungi of Pseudomonas fluorescens isolates from sugar beet rhizosphere. Appl Environ Microb 64: 3563–3569. , , & (
- 1999) Viscosinamide, a new cyclic depsipeptide with surfactant and antifungal properties produced by Pseudomonas fluorescens DR54. J Appl Microbiol 86: 80–90. , , & (
- 1999) Accurate flow cytometric membrane potential measurement in bacteria using diethyloxacarbocyanine and a radiometric technique. Cytometry 35: 55–63. , , & (
- 1999) Fluorescence monitoring of antibiotic-induced bacterial damage using flow cytometry. Cytometry 35: 235–241. & (
- 2000) Viscosinamide-producing Pseudomonas fluorescens DR54 exerts biocontrol effect on Pythium ultimum in sugar beet rhizosphere. FEMS Microbiol Ecol 33: 139–146. , , , & (
- 2001) Monitoring population size, activity, and distribution of gfp-lux-tagged Pseudomonas fluorescens SBW25 during colonization of wheat. Microb Ecol 41: 290–300. & (
- 2000) Fluorescence staining and flow cytometry for monitoring microbial cells. J Immunol Meth 243: 191–210. , , , & (