Plasmid-mediated fitness advantage of Acinetobacter baylyi in sulfadiazine-polluted soil

Authors

  • Sven Jechalke,

    1. Julius Kühn-Institut – Federal Research Centre for Cultivated Plants (JKI), Institute for Epidemiology and Pathogen Diagnostics, Braunschweig, Germany
    Search for more papers by this author
  • Christoph Kopmann,

    1. Julius Kühn-Institut – Federal Research Centre for Cultivated Plants (JKI), Institute for Epidemiology and Pathogen Diagnostics, Braunschweig, Germany
    Search for more papers by this author
  • Mona Richter,

    1. Technical University Braunschweig, Institute of Geoecology, Braunschweig, Germany
    Search for more papers by this author
  • Sylvia Moenickes,

    1. Faculty Life Sciences, Rhine-Waal University of Applied Sciences, Kleve, Germany
    Search for more papers by this author
  • Holger Heuer,

    1. Julius Kühn-Institut – Federal Research Centre for Cultivated Plants (JKI), Institute for Epidemiology and Pathogen Diagnostics, Braunschweig, Germany
    Search for more papers by this author
  • Kornelia Smalla

    Corresponding author
    1. Julius Kühn-Institut – Federal Research Centre for Cultivated Plants (JKI), Institute for Epidemiology and Pathogen Diagnostics, Braunschweig, Germany
    • Correspondence: Kornelia Smalla, Julius Kühn-Institut – Federal Research Centre for Cultivated Plants (JKI), Institute for Epidemiology and Pathogen Diagnostics, Messeweg 11/12, 38104 Braunschweig, Germany. Tel.: +49 5312993814; fax: +49 5312993006; e-mail: kornelia.smalla@jki.bund.de

    Search for more papers by this author

Abstract

LowGC-type plasmids conferring resistance to sulfonamides have been frequently isolated from manure and manured soil. However, knowledge on the dynamics of plasmid-carrying populations in soil and their response to the presence of sulfonamides is scarce. Here, we investigated effects of the sulfonamide resistance conferring plasmid pHHV216 on the fitness of Acinetobacter baylyi BD413 in soil after application of manure with or without the sulfonamide antibiotic sulfadiazine (SDZ). The persistence of A. baylyi BD413 pHHV216 in competition to its plasmid-free variant was followed in soil microcosms. CFU counts showed a decrease in A. baylyi BD413 in manured soils over the experimental period of 32 days by about 0.5 log units. The proportion of the plasmid-carrying populations decreased from 50 to < 40% in the absence of SDZ, while the proportion of plasmid-carrying BD413 increased from 50 to about 65% with SDZ added. The data suggest that SDZ introduced via manure into soil was bioaccessible, providing a fitness advantage for the plasmid-carrying population of BD413 in soil, while the plasmid conferred a fitness disadvantage when selective pressure by SDZ was absent. In future, this method may be used as a tool for the assessment of bioavailability of antibiotics in soil.

Introduction

Sulfadiazine (SDZ) belongs to the class of sulfonamide antibiotics and is frequently used in veterinary medicine, mainly in pig production (Burkhardt et al., 2005; Sarmah et al., 2006). It reaches agricultural fields by the application of manure as fertilizer. In soil, the readily extractable SDZ, which accounts for the desorbable and hence putatively bioaccessible fraction, was reported to dissipate rapidly (Thiele-Bruhn & Aust, 2004; Förster et al., 2009; Moenickes et al., 2011; Rosendahl et al., 2011; Sittig et al., 2012; Kopmann et al., 2013). Low concentrations of SDZ in soil might still select for resistant populations as demonstrated recently by the increase in SDZ resistance genes and rates of plasmid transfer in soil treated with manure containing SDZ compared with control manure (Heuer et al., 2011; Jechalke et al., 2013; Kopmann et al., 2013). However, the real fraction of bioaccessible SDZ influencing the bacterial community in soil is unknown.

Plasmids can carry a diverse array of accessory genes, such as antibiotic resistance genes, heavy metal resistance genes or genes coding for metabolic pathways (reviewed by Heuer & Smalla, 2012). The conjugative plasmid pHHV216 is a representative of the group of LowGC-type plasmids, which were frequently captured in biparental exogenous matings from manure and manured soil into E. coli and are assumed to play an important role in disseminating antibiotic resistance genes among soil bacterial populations (Heuer et al., 2009). Plasmid pHHV216 confers resistance against sulfonamides, tetracycline, and streptomycin, and Acinetobacter spp. have been identified as potential hosts of LowGC-type plasmids in soil (Heuer et al., 2009). However, so far, the dynamics of plasmid-carrying populations in soil and their response to the presence of antibiotics such as SDZ are not well understood. The expression of accessory genes such as resistance genes can lead to a selective advantage of the plasmid-carrying population under selective conditions such as the presence of antibiotics. At the same time, plasmids can impose a metabolic burden on the host cell through costs for maintenance and replication as well as by the expression of foreign proteins encoded on the plasmid (Glick, 1995; Bentley et al., 2009; Silva et al., 2012), which influences the plasmid persistence within a bacterial population. The measurement of a selective advantage of a plasmid-carrying strain relative to the plasmid-free control strain in competition experiments might be used as a tool to assess the presence or absence of environmental stress such as antibiotics.

The aim of this study was to provide a direct link between the persistence of an antibiotic resistance conferring plasmid within a bacterial population in soil and selective pressure by the application of manure containing antibiotics. Two independent microcosm experiments were performed comparing the persistence of Acinetobacter baylyi BD413 with or without plasmid pHHV216 in soil, which was treated with manure that was spiked with SDZ (SDZ manure) or left unspiked (control manure).

Materials and methods

Experimental design

Two independent microcosm trials were performed. The soil used for both trials was a Haplic Luvisol with 1.2% organic carbon, 16% clay, 78% silt, and 6% sand. The manure was kindly provided by Dr. Berk (Thünen-Institut, Braunschweig) and derived from pigs, which were never treated with antibiotics. The GFP-labeled bacterial strain A. baylyi BD413 (Rifr Kanr miniTN5::gfp-nptII) was grown in an overnight culture of 5 mL Luria–Bertani (LB) broth (Carl Roth GmbH + Co. KG, Karlsruhe, Germany) with rifampicin (50 mg L−1) and kanamycin (50 mg L−1) at 28 °C. Subsequently, 1 mL of culture was transferred to a second overnight culture of the same composition. For the plasmid-bearing A. baylyi BD413 pHHV216 (SDZr Smr Tcr), the selective markers streptomycin (20 mg L−1) and tetracycline (20 mg L−1) were added to the liquid cultures. Bacterial cells were pelleted (2400 g, 3 min) and washed twice with 0.85% sodium chloride solution to remove the antibiotics. The pellets of plasmid-free and plasmid-bearing cells were resuspended and mixed in a ratio of 1 : 1 into the soil with manure (40 g kg−1 dry soil) to an estimated final concentration of 106 cells per g soil. The amount of manure applied corresponded to the typical application of 30 m3 manure ha−1 according to agricultural practice (Hammesfahr et al., 2011). For the SDZ trials, manure was additionally spiked with SDZ to reach a final concentration of 50 mg SDZ kg−1 soil. This concentration is about 10 times higher than the realistic input of sulfonamides via manure according to agricultural practice (Heuer et al., 2008; Lamshöft et al., 2010). The two trials were performed in polystyrene flowerpots (8 cm diameter, 8 cm height) with four replicates per treatment. The temperature was kept constant during incubation (16 °C). Samples from the first trial were taken on days 7, 14, 21, and 28. Based on the results of this experiment and in view of using these data for the development of a model, the samples of the second trial were taken on days 2, 4, 8, 10, 15, 25, and 32.

Plate counts of A. baylyi BD413 and proportion of plasmid-containing cells

The soil in each pot was mixed, and cells were resuspended from 1 g soil by adding 9 mL 1 : 10 diluted tryptic soy broth (trial 1; Merck KGaA, Darmstadt, Germany) or 9 mL 0.85% NaCl solution (trial 2) and sterile glass beads (4 mm diameter) and overhead shaking (GFL 3025, Braunschweiger Labor Bedarf, Germany) for 2 h. A serial dilution was plated in duplicates on LB agar with rifampicin (50 mg L−1), kanamycin (50 mg L−1), and cycloheximide (100 mg L−1) and incubated at 28 °C for 24 h. The green fluorescent colonies of BD413 were counted under UV light. From each replicate, 50 green fluorescent colonies were picked with sterile toothpicks (200 per treatment) and transferred to LB agar plates containing rifampicin (50 mg L−1), kanamycin (50 mg L−1), streptomycin (20 mg L−1), tetracycline (20 mg L−1), and cycloheximide (100 mg L−1), and subsequently on LB agar without selective markers as control followed by an incubation at 28 °C for 24 h. The green fluorescent streaks were counted to estimate the abundance of the plasmid-bearing BD413 in the soil.

The selection coefficient Sij was used for a comparison of fitness (Lenski et al., 1991) based on the ratio of net growth rates mi and mj of the plasmid-carrying (i) vs. the plasmid-free (j) strain:

display math(1)

The amount of A. baylyi BD413 applied to the soil probably exceeded the soil's carrying capacity of this strain. Hence, the number of cells decreased initially due to growth competition. In this first approach, we assumed that the net effect of growth and density driven mortality was at a similar level over the first 15 days of trial 2. Correspondingly, slopes of linear regressions were calculated based on the BD413 cell counts of the first 15 days divided by the initial number of BD413 applied to the soil on day 0. The day 0 values were estimated from BD413 cells applied with manure. These estimated numbers do not consider the recovery rate of A. baylyi cells from soil after inoculation, but they should allow a comparison between soil treatments (SDZ manure and control manure), as equal numbers of cells were used as inoculums. To investigate if the same trend was observed in the first trial, slopes of linear regressions were calculated based on the BD413 cell counts of the first 28 days divided by the initial number of BD413 applied to soil on day 0. It has to be noted that due to negative net growth rates a negative selection coefficient indicates that the fitness of the strain i is larger than the fitness of strain j and vice versa, in contrast to the interpretation for positive net growth rates (Lenski et al., 1991).

Results

The effect of plasmid pHHV216 on the fitness of A. baylyi BD413 in soil after application of SDZ manure or control manure was investigated in two independent microcosm scale experiments (trials 1 and 2). After the application of manure and similar numbers of plasmid-carrying and plasmid-free cells of BD413 to soil, the persistence of A. baylyi BD413 pHHV216 in competition with its plasmid-free variant was followed.

In the first trial, significantly more CFUs were counted for the SDZ manure-treated soil (Fig. 1a). While CFU counts remained relatively stable until day 21, they declined about 0.5 log units between day 21 and 28 in the first trial. In contrast, the abundance of A. baylyi BD413 in manured soil decreased gradually over the period of 32 days by about 0.5 log units in the second trial. This behavior seemed not to be influenced by SDZ because CFU counts of the two treatments were significantly different on days 2, 25, and 32 only, with a slightly higher number observed in the soil treated with control manure (Fig. 1b).

Figure 1.

Total CFU counts per gram soil and ratio of plasmid-carrying isolates of Acinetobacter baylyi BD413 of the first (a and c) and the second trial (b and d) are plotted against the days after the beginning of the experiment. The soils treated with manure and manure containing sulfadiazine (SDZ) is indicated by light and dark gray bars, respectively. Error bars indicate standard deviations of five and four replicates for the first and second trial, respectively. Significant differences between treatments are indicated by asterisk (< 0.05, t-test).

Without selection pressure by SDZ, the proportion of the plasmid pHHV216-carrying population decreased from 50 to < 40% after 28 and 32 days of trials 1 and 2, respectively (Fig. 1c and d). When selection pressure by SDZ was present, the proportion of plasmid-carrying BD413 increased from 50 to about 66%. This increase seemed to be faster in the beginning and slowed down after the first 15 days of incubation in the second trial (Fig. 1d).

Correspondingly, for the first 15 days of the second trial, the decrease for either of the competitors was approximately linear at a logarithmic scale (Fig. 2). For days 25 and 32, there was only a slight or no further decrease detectable, and the ratio between plasmid-free and plasmid-carrying cells did not change any further (Fig. 2). In control manure-treated soil, the decrease in plasmid-carrying BD413 was faster than the decrease in plasmid-free BD413 during the first 15 days. The selection coefficient (Eqn. (1)) in control manure-treated soil was on average 0.36 ± 0.17, which represented a 36% faster decrease in plasmid-carrying BD413 compared with plasmid-free BD413 cells and, in general, a fitness advantage of plasmid-free BD413. In SDZ manure-treated soil, the selection coefficient was on average −0.35 ± 0.16, which represents a 35% slower decrease in the plasmid-carrying population as compared to the plasmid-free population and, in general, a fitness advantage of the plasmid-carrying BD413. The selection coefficients were significantly different between the control manure and SDZ manure-treated soils (t-test, n = 4, P = 0.0008), demonstrating a significant effect of SDZ on the fitness of pHHV216-carrying BD413 cells in comparison with their plasmid-free variants in manured soil. Although the temporal resolution was lower for the first trial, selection coefficients were also calculated and showed the same trend as observed in the second trial. The selection coefficients in control manure-treated soil of trial 1 were calculated until day 28 and were on average 0.27 ± 0.16, which represented a 27% faster decrease in plasmid-carrying BD413 compared with plasmid-free BD413 cells. In SDZ manure-treated soil of trial 1, the selection coefficient was on average −0.85 ± 0.62, which represents an 85% slower decrease in the plasmid-carrying population as compared to the plasmid-free population.

Figure 2.

Development of plasmid-carrying (indicated by circles and dashed line) and plasmid-free (indicated by crosses and solid line) populations of Acinetobacter baylyi BD413 in soil of trial 2 treated with (a) manure and (b) manure with sulfadiazine (SDZ), compared with the population applied on day 0. Lines indicate a linear regression of days 2 until 15. The ratios of the slopes of linear regressions were used for the calculation of the selection coefficients.

Discussion

Acinetobacter baylyi BD413 was selected for this study because Acinetobacter spp. were identified as potential hosts of plasmid pHHV216 in soil (Heuer et al., 2009). Strain BD413 was derived from strain BD4 which was isolated from soil (Juni & Janik, 1969), and BD413 is known for its high competence for transformation and for the production of pili, which are suggested to be involved in adherence to biotic and abiotic surfaces (Gohl et al., 2006). Therefore, it is likely that strain BD413 is well adapted to soil conditions and might colonize soil particles by the formation of biofilms. In this study, the abundance of A. baylyi BD413 inoculated into soil at rather high numbers decreased over the experimental period of 32 days. This might indicate that the capacity of the soil to maintain a population of A. baylyi BD413 was exceeded or that the strain was outcompeted by populations, which are better adapted to the soil conditions. However, during the second trial, no significant difference was detected between the SDZ treatment and the control from day 4 until day 15, indicating that the A. baylyi BD413 counts in soil were not influenced by the applied amount of SDZ. In the first trial, the abundance of A. baylyi BD413 was higher in the SDZ manure-treated soil compared with control manure-treated soil, but this difference was rather constant for the first 21 days. Although it cannot be excluded that the difference might have been due to different numbers of A. baylyi BD413 cells applied in the beginning of the experiment, the difference might suggest an initial advantage of A. baylyi BD413 pHHV216 in soil treated with SDZ manure compared with control manure. Additionally, further research would be needed to clarify whether a reduction in the inoculum size would affect the competitiveness and survival of A. baylyi BD413 in soil.

In previous studies, it was demonstrated that the easily extractable and hence potentially bioavailable fraction of SDZ rapidly decreased after its application to soil by manure (Thiele-Bruhn & Aust, 2004; Kreuzig & Höltge, 2005; Förster et al., 2009; Rosendahl et al., 2011; Sittig et al., 2012; Kopmann et al., 2013). Accordingly, in the present experiment, a fast reduction of bioavailable SDZ in soil might have caused only an initial effect of SDZ on the abundance of A. baylyi BD413. A significant increase in the proportion of plasmid-carrying relative to plasmid-free A. baylyi cells in SDZ manure-treated soil compared with the control treatment was observed in the two independent trials (Fig. 1). The increase in the proportion of plasmid-carrying A. baylyi cells seemed to be more pronounced during the first 15 days of the experiment, which might correlate with a fast dissipation of easily extractable SDZ in soil as observed, for example, in the study of Kopmann et al. (2013) who measured a dissipation half-life of 5.1 days. Moreover, in the control manure-treated soil, the proportion of plasmid-carrying A. baylyi cells decreased over time by about 10%, which indicated a fitness disadvantage or cost of the plasmid when selective pressure by SDZ is absent. Correspondingly, the selection coefficients for the first 15 days of trial 2 were significantly lower in SDZ manure than in control manure-treated soil. In the first trial, although the temporal resolution was lower than for the second trial, the same trend was observed. In a previous study, the stability of the IncP-1β plasmid pB10 in different Proteobacteria was investigated in liquid cultures (Heuer et al., 2007). High costs were observed in Pseudomonas putida H2 with about 20% reduction in cell doublings relative to the plasmid-free host. Compared with the liquid cultures, in manured soil, we observed an even faster decrease in plasmid-carrying BD413 compared with plasmid-free BD413 cells. However, when comparing the results of the previously mentioned liquid culture experiment with our soil experiment, it has to be considered that strains, plasmids, and culture conditions used were different, which likely influenced the results. Hence, although one of our assumptions is that the effect of growth and competition is rather constant during the first 15 days of the experiment, the transfer and proliferation of plasmid-containing bacteria might be influenced by several dynamic factors such as nutrient availability, system capacity to maintain the bacterial population, and population interactions. These factors have to be considered in future modeling of competitive dynamics of plasmid-containing and plasmid-free cells in soil and the environment. Furthermore, it has to be taken into account that the impact of a plasmid on the fitness can be depending on the host strain and the plasmid type as shown for IncP-1 and IncN broad host range plasmids (Humphrey et al., 2012). Additionally, the burden or metabolic cost of a plasmid can be ameliorated over time by mutations or rearrangements in plasmid and/or host DNA or the plasmid can confer benefits that facilitate host survival and colonization, which was speculated to be the reason for an improved plant-colonizing fitness of a plasmid-carrying Pseudomonas fluorescens strain compared with the plasmid-free variant in a greenhouse and field experiment (Lilley & Bailey, 1997).

In conclusion, this study indicated that SDZ introduced via manure into soil was bioaccessible and provided a selective advantage for A. baylyi BD413 carrying the resistance plasmid pHHV216, while plasmid carriage conferred a fitness disadvantage when antibiotic selection was absent. The approach of using A. baylyi BD413 to analyze the persistence of plasmids in response to external selection pressure was successfully demonstrated in two independent trials. In future studies, competition experiments as described in the present study may be used as a tool for the assessment of bioavailability of antibiotic compounds in soil. As a next step, different environmentally relevant concentrations of antibiotics to which pHHV216 confers resistance should be tested to confirm the sensitivity and applicability of this tool.

Acknowledgements

S.J., C.K., and H.H. were funded by the Deutsche For-schungsgemeinschaft (DFG) in the framework of the Research Unit FOR 566 ‘Veterinary Medicines in Soil: Basic Research for Risk Analysis’ (SM59/5- 2-3; SM59/12-1). We would like to thank Ilse-Marie Jungkurth for proofreading this manuscript.

Ancillary