Inactivation of the Regulatory Gene algU or gacA Can Affect the Ability of Biocontrol Pseudomonas fluorescens CHA0 To Persist as Culturable Cells in Nonsterile Soil

Abstract

Rifampin-resistant Pseudomonas fluorescens CHA0-Rif and mutants in which the regulatory gene algU (encoding sigma factor σ^E) or gacA (encoding a global regulator of secondary metabolism) was inactivated were compared for persistence in three nonsterile soils. Functional algU and (particularly) gacA were needed for CHA0-Rif to maintain cell culturability in soil.

Certain fluorescent pseudomonads are important for bioremediation of polluted soils or biological control of soilborne phytopathogenic fungi (5, 11). Once in the soil, pseudomonads often undergo a reduction in CFU, and in some cases this is due to the loss of the ability of cells to form a colony on laboratory media routinely used to grow the inoculant (1, 7, 20). Indeed, a significant proportion of the nonculturable cells may respond positively to a viability test, suggesting that they are not dead (1, 5, 20). Cell viability is often assessed by using Kogure's cell elongation test (10), which is based on the assumption that viable cells can respond to nutrient stimulation and engage in a few cell divisions, without necessarily exhibiting colony-forming ability. In this test, nutrient-responsive cells (considered as viable) are prevented from dividing by the presence of nalidixic acid, enlarge significantly, and can be distinguished from smaller, nonresponsive cells. In this context, the occurrence of cells in a viable but nonculturable (VBNC) state, as defined by Oliver (17), can be derived from the difference between viable counts and CFU of the strain.

The significance of VBNC cells remains to be elucidated (13). It has been hypothesized that the VBNC state(s) could represent an adaptative strategy under adverse conditions. For the biocontrol agent Pseudomonas fluorescens CHA0, this possibility was suggested by the fact that, in the field, the number of VBNC cells exceeded CFU by several orders of magnitude for more than 6 months (20). The hypothesis is in accordance with observations that (i) dead cells introduced into nonsterile soil disappear rapidly (21) and (ii) VBNC cells may recover their previous colony-forming ability (17), although such “resuscitation” experiments are prone to experimental artifacts (3). Alternatively, other authors have proposed that VBNC cells are dying cells (13).

In some cases, the loss of colony-forming ability is thought to be an experimental artifact caused by exposure to oxidative stress on the plates (2), especially for H₂O₂-sensitive culturable cells (3); the stress may be alleviated by adding antioxidants to the plates (3). However, exposure of CHA0 to H₂O₂ does not lead to VBNC cells (7), and the addition of various antioxidants and radical scavengers when samples containing putative VBNC cells are studied does not improve the recovery of CFU (unpublished results). Rather, field observations have suggested that the occurrence of VBNC cells in this pseudomonad was caused by exposure to stress conditions in situ (5, 20), which was confirmed experimentally in vitro (12) and in soil (unpublished results). The formation of VBNC cells seems induced when stress intensity is sufficiently high (12), which implies that up to a particular threshold of stress, regulatory systems manage to adapt cell physiology and cell culturability is maintained. If this were true, it could mean that VBNC cells may occur even at a lower stress level provided that relevant regulatory systems are inactivated. Alternatively, their inactivation would cause cell death, as often assumed (8, 16). CHA0 is a soil inhabitant and thus is adapted to the stressful conditions prevailing in this habitat. Therefore, the objective was to assess whether mutational inactivation of key regulatory genes could affect the ability of CHA0 to persist as culturable cells in nonsterile soil.

We chose to focus on two regulatory genes, algU and gacA. The algU (algT) gene codes for the alternative sigma factor σ^E (σ²²), which regulates transcription of the heat shock sigma factor σ^H (σ³²) and extracytoplasmic stress response (8, 14). In pseudomonads, algU controls alginate production (8, 18). The global activator gene gacA (encoding a response regulator) is required for the synthesis of extracellular enzymes and antimicrobial secondary metabolites in CHA0 (11). Mutational inactivation of gacA almost completely eliminates the ability of CHA0 to protect tobacco from black root rot (11), and in soil CFU of the gacA mutant CHA96-Rif decreased more rapidly than those of the gacA⁺ strain CHA0-Rif (16).

A spontaneous rifampin-resistant mutant of CHA0 (i.e., CHA0-Rif [16]) was used so that the inoculant could be recovered from nonsterile soil. A gacA′-′lacZ translational fusion was introduced into CHA0 by gene replacement, resulting in CHA96 (11). Strain CHA96-Rif, a spontaneous rifampin-resistant mutant of CHA96, has already been investigated in soil and in vitro grows like CHA96 (16). The algU gene has been cloned in CHA0 (18). Strain CHA212-Rif is an in-frame deletion (250-bp) algU mutant of CHA0-Rif obtained by gene replacement (18). Both strains have the same growth rate in rich medium. Inocula consisted of mid-log-phase cells harvested from King's B agar (9) containing rifampin and washed in sterile distilled water (12).

Soil was collected in November from three Swiss cambisols (Table 1) located at Eschikon, Tänikon, and Mellstorf. At Mellstorf, the soil was taken immediately below the forest litter, and it differed from the two farm soils, particularly on the basis of pH and organic matter content and quality. Sterile distilled water was added to sieved (5-mm pore size) soil (6, 12) to achieve a water potential of −0.03 MPa (Table 1) after subsequent addition of inoculum, as described previously (6, 12). Each soil received 3.3 ml of cell suspension of CHA0-Rif, CHA96-Rif, or CHA212-Rif (i.e., 10⁸ cells added per g of soil) or sterile distilled water (uninoculated control) per 100 g of soil. Each soil was thoroughly mixed and 10-g samples were distributed into 25-ml vials. Conditions of incubation (12°C, relative humidity of 70%) of the vials were as described previously (6, 12). During the 60-day experiment, soil water content remained essentially constant (±0.1% [wt/wt]).

TABLE 1.

Main characteristics of the three soils examined in this study^a

Soil	Geographic origin (Canton)	Soil usage	Soil layer sampled (cm)	% Content (wt/wt)			Texture	% Organic matter (wt/wt)	pHH2O	% Water content (wt/wt)b
				Clay	Silt	Sand
Eschikon	Zürich	Fallow	0-25	15	42	43	Loam	3.5	7.0	22
Tänikon	Thurgau	Fallow	0-25	24	36	40	Loam	5.5	6.8	28
Mellstorf	Aargau	Conifer forest	5-8	10	42	48	Loam	18.1	3.8	32

^aAll three were cambisols.

^bAs determined at Ψ_w = −0.03 MPa.

Sampling, cell extraction, and dilution series were done as described previously (12). CFU were obtained on King's B agar containing rifampin (100 μg/ml) and the antifungal compound cycloheximide (190 μg/ml) after the plates were incubated at 27°C for 2 to 7 days. In the uninoculated control, no rifampin-resistant colonies were found. The total number of CHA0-Rif cells was determined by immunofluorescence microscopy (12, 20) with a primary antiserum specific for CHA0 (19). No cross-reaction was found in uninoculated soil samples. The number of viable cells of CHA0-Rif was determined by using a combination of Kogure's viability test (10) and immunofluorescence microscopy, as described previously (6, 12). Most inoculant cells in soil were <1.5 μm in length. Substrate-responsive cells elongated to ≥3.5 μm, and they were counted as viable. When population levels were low, filters were examined until at least 10 elongated cells were found.

All treatments were replicated three times (destructive sampling) along a randomized block design. Cell counts were expressed per gram of dry soil and log transformed. Three-way analyses of variance (strain by soil by cell enumeration method) were done at each sampling, and Tukey's honestly significant difference tests were used (when appropriate) to compare treatments (P < 0.05; SYSTATS version 5, SPSS Science, Chicago, Ill.). First, data from each soil were analyzed at each sampling for each cell count method. Data and the results of statistical analyses are shown only for CFU (Fig. 1). Second, the influence of the cell count method was assessed for each strain by comparing total counts, viable counts, and CFU at each sampling in each soil. Data and the results of statistical analyses are shown only for the last sampling (Fig. 2).

FIG. 1.

CFU of P. fluorescens CHA0-Rif, the algU mutant CHA212-Rif, and the gacA mutant CHA96-Rif for 60 days after inoculation in Eschikon soil (A), Tänikon soil (B), and Mellstorf soil (C). The detection limit was 10² CFU/g of soil. Error bars show standard deviations. Statistically significant differences between strains were found at 60 days and are indicated by the lowercase letters a, b, and c.

FIG. 2.

Persistence of P. fluorescens CHA0-Rif, the algU mutant CHA212-Rif, and the gacA mutant CHA96-Rif at 60 days after inoculation in Eschikon soil (A), Tänikon soil (B), and Mellstorf soil (C). The numbers of total cells (immunofluorescence microscopy), viable cells (Kogure's test), and CFU of the inoculants were determined. The error bars show the standard deviations. Data were analyzed separately for each soil. For each strain, statistically significant differences between cell counts (i.e., total counts, viable counts, and CFU) are indicated by the lowercase letters a, b, and c. For each type of cell count, statistically significant differences between the three strains are indicated by the Greek letters α, β, and γ.

Effect of algU inactivation.

The algU mutant CHA212-Rif, like CHA0-Rif, persisted as culturable cells in Eschikon and Tänikon soils (Fig. 1 and 2). In Mellstorf forest soil, CFU of CHA212-Rif decreased by about 2 log units from 30 to 60 days, whereas those of CHA0-Rif remained stable (Fig. 1C). On day 60, CHA0-Rif was still culturable but total counts and viable counts were higher than CFU for CHA212-Rif (Fig. 2C), indicating the presence of VBNC cells of CHA212-Rif.

Therefore, a functional algU gene was not required to maintain cell culturability in the two farm soils but was apparently needed in Mellstorf forest soil, where VBNC cells were 10 times more numerous than culturable cells for CHA212-Rif. This suggests that the inoculant was exposed to stressful environmental conditions in the forest soil, perhaps because of the low pH or organic matter quality. Indeed, algU controls tolerance to different types of environmental stress in pseudomonads (8, 18).

Effect of gacA inactivation.

CFU of the gacA mutant CHA96-Rif decreased with time in the three soils, especially in Mellstorf soil (Fig. 1). In Eschikon soil, VBNC cells were found at 60 days (but not at 30 days [data not shown]), as total counts and viable counts were higher than CFU by >1 log unit (Fig. 2A). The same trend was observed in Tänikon soil, except that the difference between viable counts and CFU was not statistically significant on day 60 (Fig. 2B). In Mellstorf soil, there was a 4-log unit difference between viable counts and CFU of CHA96-Rif at 60 days (Fig. 2C), and VBNC cells were already evident on days 15 and 30 (data not shown).

In contrast to CHA0-Rif and CHA212-Rif, the gacA mutant CHA96-Rif was clearly compromised for survival in all three soils, since CFU were lower than those of CHA0-Rif. This result is reminiscent of data from stationary-phase competition experiments in vitro (4). In Eschikon and Mellstorf soils, viable counts of CHA96-Rif exceeded CFU, and VBNC cells could represent ≥10⁶ cells/g of soil. A spontaneous gacS mutant persisted like the gacA mutant in the three soils (data not shown), in accordance with the facts that (i) GacS is the cognate transmembrane sensor kinase activating GacA and (ii) mutations in either gene lead to the same phenotype in CHA0 (4, 11). The possibility that gacA could contribute to stress resistance in CHA0 is raised by the observation that, in P. fluorescens Pf-5, gacA regulates positively the accumulation of the stationary-phase sigma factor σ^S (σ³⁸) and contributes to oxidative stress resistance (22). This may be particularly relevant to explain the results obtained in Mellstorf forest soil.

Almost nothing is known of the genetic factors implicated in the formation of VBNC cells (15). We show here that algU and, particularly, gacA contribute significantly to the capacity of CHA0-Rif to persist as culturable cells in soil, and it is proposed that the occurrence of VBNC cells of the algU and gacA mutants may result from their inability to adapt their cell physiology in response to environmental stress.