Online Supplementary Information
Frequency-dependent advantages of plasmid carriage by Pseudomonas in homogeneous and spatially structured environments
Ellis, R.J., Lilley, A.K.,Lacey. S.J., Murrell, D. & Godfray, H.C.J.
1Experimental details
1.1Experimental overview
The coexistence of plasmid-free and plasmid-bearing cells was studied in both spatially structured and homogeneous environments. The structured environment consisted of a polycarbonate membrane on solidified media containing both nutrients and mercury. Bacteria were filtered onto the membrane and grew utilizing nutrients that diffused through the membrane from below. The homogeneous environment consisted of a shaken tube of liquid medium containing the same nutrients used in the solid medium.
1.2Bacterial strains and plasmids
The bacterium used in all experiments was Pseudomonas fluorescens SBW25, originally isolated from sugar beet (Bailey, et al., 1995). A spontaneous rifampicin resistant mutant (SBW25R) was generated and used throughout. This strain lacked plasmids and was sensitive to mercury. Aconstitutively-expressed red fluorescent protein cassette (DsRed, Discosoma) was inserted into the chromosomal DNA of SBW25Rfollowingmobilization of the delivery plasmid TTN151 (mini-Tn5-Kmr-rrnBP1::RBSII-dsRed-T0-T1(Tolker-Nielsen, et al., 2000))from Escherichia coli CC118 λpir in a triparental mating with E. coli HB101 (SmrrecA thi pro leuhsdRM) carrying the helper plasmid RK600 (Cmr ColE1 oriV, RP4 oriT).
The plasmid, pQBR103 (~425 kbp, Tra+, Hgr), was isolated from a sugar beet crop by releasing a marked variant of SBW25 in the environment and then re-isolating clones that had acquired mercury resistance(Lilley & Bailey, 1997b). A constitutively-expressed green fluorescent protein cassette (GFP, Aequorea victoria) was inserted into the DNA of plasmid pQBR103 as follows. The delivery plasmid pJBA28 (with PA1/O4/O3::gfpmut3b cloned into mini-Tn5 Km(Andersen, et al., 1998)) was mobilised from Escherichia coli CC118 λpir to Pseudomonas putida UWC1 (Rifr) bearing plasmid pQBR103 in a triparental mating as above. Marked plasmids were selected by co-transfer of Kmr, Hgr and GFP to P. putida Mes300 (Nalr)(Lilley & Bailey, 2002). The marked plasmid was transferred back into SBW25R to generate SBW25R (pQBR103::GFP) which is resistant to mercury at a concentration of 25 μM in standard R2A agar medium (Oxoid, UK). During the course of our experiments no segregants were detected, indicating the stability of the plasmid.
The marked strains showed no fitness effects of the insertion of the mini-transposons when compared to their respective ancestral strain, as shown by the maintenance of inoculation ratios under the experimental conditions used here or in planta. Single colonies of SBW25R::RFP and SBW25R (pQBR103::GFP) were used to inoculate 5ml aliquots of R2’B (see below). The cultures were grown at 30ºC with constant shaking for 65 h. Cells were harvested by centrifugation and washed three times in 5 ml sterile phosphate buffered saline (PBS; Sigma-Aldrich, Poole, UK), and the cell concentration standardized by adjusting the OD600 to 0.50 with PBS. Aliquots of the 2 cell preparations were mixed to provide approximate1:9 or 9:1 ratio mixtures of plasmid-free RFP (Hgs) andplasmid-bearing GFP (Hgr) cells. These mixtures were diluted as appropriate and used to inoculate the experimental environments. To determine the precise ratio of the two cell types, triplicate sub-samples from the remainder of the mixed inocula were diluted, spread onto standard R2A plates, and incubated at 30ºC for 3 days. Densities of the two cell types were estimated by illuminating with blue light using a DarkReader (Clare Chemical Research, USA) to distinguish GFP- and RFP-expressing colonies.
1.3Growth media and mercury toxicity
The growth medium (R2′) contained (in units of g l-1): 0.525 peptone, 0.35 yeast extract, 0.35 dextrose, 0.35 starch, 0.21 K2HPO4, 0.21 sodium pyruvate, 0.175 tryptone, 0.0168 MgSO4. In liquid form (R2′B) this medium was used for the homogeneous environment, and with the addition of agar (1.2% w/v; R2′A) it was used as the base for membranes in the spatially-structured environment.
As mercury is known to form complexes with different components of media (Chang, et al., 1993) preliminary work was required to determine the range of mercury concentrations that would cause equivalent levels of toxicity in the two environments. This was done by measuring relative toxicity using Pseudomonas fluorescens SBW25::luxCDABE. The strain was made by the chromosomal insertion of the luxCDABE operon, derived from Photorhabdus luminescens, using the suicide vector mini-Tn5 transposon delivery system(Winson, et al., 1998)as described above. The resultant strain was constitutively bioluminescent and the reduction of light output in the presence of mercury, which will be a function of cell number and metabolic activity,was used as an index of relative toxicity.
To assay toxicity, SBW25::luxCDABE cells were added to a range of dilutions of HgCl2 in R2’B in the wells of a black 96-well microtitre plate. Alternatively, cells were filtered under light vacuum onto black Isopore polycarbonate membranes (25 mm diameter and 0.22 μm pore size, Millipore, Watford, UK) and placed on R2’A plates supplemented with HgCl2. Luminescence from the broth and filter cultures were assayed after 2 h using a plate luminometer (Lucy 1, Anthos Labtec Instruments, Austria), with an integration time of 1 s at a temperature of 23ºC. Three replicates were conducted for each HgCl2concentration and relative toxicity was expressed as the proportional reduction in luminescence compared to mercury-free controls. As expected from previously published results (Hassen, et al., 1998), the toxicity of mercury for bacteria without the resistance plasmid was considerable greater in the broth compared with the agar environment (Online Figure 1): 50% relative toxicity occurred in the broth culture for HgCl2 concentrations of ~5 μM but in plate culture at ~17 μM. Based on this data we explored the effect of HgCl2concentrations in the range 0 to 30 μM for solid media culture and 0 to7.5 μM in liquid media. To allow direct comparison of the potential for plasmid polymorphism in the two environments we have rescaled the mercury concentrations to relative toxicity assessed using the bioluminescent reporter strain using the best-fit curve following polynomial regression (Broth R2 = 0.993, Filter R2 = 0.999; Online Figure 1).
1.4Invasion from rare in the spatially structured environment
The mixed preparations of RFP (Hgs) and GFP (Hgr) bacteria were diluted and filtered under light vacuum onto black Isopore polycarbonate membranes (25 mm diameter and 0.22 μm pore size, Millipore, Watford, UK), giving approximate densities of106, 105 or 104 CFU cm-2, equivalent to 1 CFU per 100 μm2, 1000 μm2 or 10,000 μm2, respectively. Filters were then placed on R2’A plates supplemented with HgCl2 at concentrations ranging from 0 to 30 μM. The average final densityafter 2-3 days growth was2.6 x 109(SE = 8.4 x 107) CFU cm-2 (equivalent to 26 CFU μm-2) regardless of inoculum density.
After 72 or 96 h the spatial distribution of RFP and GFP-expressing cells was determined by epifluorescence microscopy, using an Optiphot-2 microscope fitted with a W Green (excitation 510-560 nm, dichroic mirror 580 nm, barrier filter 590 nm) filter block for RFP and B-2A filter block (excitation 450-490 nm, dichroic mirror 510 nm, barrier filter 520 nm) for GFP visualization (Nikon, UK). Images of the same field of view but with each of the two filter blocks were captured on an attached digital camera. Although the final density of cells (26 CFU μm-2) implies a depth of several cells on the filters, scanning through various focal depths did have a discernable effect on the relative spatial distribution of the cell types and growth of one type over another was only detected at the very edge of the distinct microcolonies.
To estimate the relative final densities of bacteria, the filters were removed with sterile forceps and vortexed in 10 ml sterile PBS. The resultant cell suspension was serially diluted and plated onto R2A,and the relative numbers assayed as described above. As the plasmid pQBR103 is self-transmissible (Lilley & Bailey, 1997a), we quantified the number of transconjugants (RFP cells that had received the plasmid by conjugation) in each experiment. To do this, aliquots were also plated onto R2A supplemented with 25 μM HgCl2. Previous work had shown that transconjugants show up as red colonies on media containing mercury, as the red fluorescence overrides green fluorescence in cells containing both markers when illuminatedwith the visible blue lightfrom the DarkReader. Green colonies that formed on plates with mercury represented colonies of the original GFP (Hgr) cells. Using this method transconjugants could be detected when they were ≥ 0.1% of the total population. The proportions of each of the three cell types (RFP (Hgs), GFP (Hgr) and RFP/gfp (Hgr)) were determined by maximum-likelihood calculations using data from assays both with and without 25 μM HgCl2.
1.5Invasion from rare in themixed homogeneous environment
The mixed preparations of RFP (Hgs) and GFP (Hgr) bacteria were diluted 10-, 100- and 1000-fold in sterile PBS. Three replicate aliquots of 5 ml R2B broth was inoculated with appropriate densitiesof SBW25 to giveapproximate densities of106, 105 or 104 CFU ml-1 respectively. As with the structured environment, the final density achieved was independent of starting density: 1.71 x 109(SE = 6.56 x 107) CFU ml-1. Experiments were repeated with increasing mercury concentrations from 0 - 7.5 μM. Tubes were incubated with orbital shaking (225 rpm) at 21ºC and after three days the relative proportions of RFP (Hgs), GFP (Hgr) and transconjugant cells were estimated as described above.
2Detailed Results
2.1Invasion from rare in a spatially structured environment
In total, 52 combinations of mercury concentration, initial cell density and ratio of RFP (Hgs) to GFP (Hgr) were examined, each replicated three times, to understand under what conditions the two cell types were able to coexist. One-tailed t-tests were used to determine whether the initially rare cell type had increased in frequency during the course of each experiment (Online Tables 1 and 2). A threshold concentration of mercury below which resistant GFP (Hgr) cells could not increase from rare was calculated by linear interpolation of the data in Online Table 1. The concentration at which plasmid-bearing cells could invade increased slightly with higher densities of the initial inoculum. A second threshold concentration of mercury above which RFP (Hgs) were unable to increase when rare was calculated by linear interpolation of the data in Online Table 2. Again, there was a slight tendency for this threshold to rise with the density of the initial inoculum. These data are plotted in Figure 1 of the manuscript. It can be seen that there is a region of parameter space where both GFP (Hgr) and RFP (Hgs) cells can increase in frequency from an initially low frequency and in these environments a population polymorphic for the plasmid is predicted. No transconjugant bacteria were observed in the spatially-structured environment (the limits of detection are estimated to be about 0.1% of the total population).
2.2Invasion from rare in a well-mixed homogeneous environment
For the homogeneous environment a total of 40 combinations of mercury concentration, initial cell density and ratio of RFP (Hgs) to GFP (Hgr) were tested, each replicated three times. In contrast to the spatially-structured environment, transconjugants were observed frequently in the well-mixed environment. As before there was a threshold mercury concentration below which GFP (Hgr) cells could not increase from rare (Online Table 3). However, at mercury concentrations below this threshold the resistance plasmid was able to spread through conjugation (Online Table 3). Two thresholds, one for the GFP (Hgr) cells and one for the plasmid were therefore calculated. In both cases, the threshold increased as the size of the initial inoculum rose. A third threshold mercury concentration was calculated above which sensitive RFP (Hgs) cells without the plasmid were unable to increase from rare (Online Table 4). These thresholds are plotted in Figure 1 of the manuscript. The thresholds for the increase of the two bacterial types are non-coincident and hence there is a region where a polymorphism is predicted in the absence of horizontal transfer. When plasmid transfer is taken into account, polymorphism is predicted under a slightly broader range of mercury concentrations.
3References
Andersen, JB, Sternberg, C, Poulsen, LK, Bjorn, SP, Givskov, M & Molin, S (1998) New unstable variants of green fluorescent protein for studies of transient gene expression in bacteria. Appl Environ Microbiol,64: 2240-2246.
Bailey, MJ, Lilley, AK, Thompson, IP, Rainey, PB & Ellis, RJ (1995) Site directed chromosomal marking of a fluorescent pseudomonad isolated from the phytosphere of sugar beet; stability and potential for marker gene transfer. Mol Ecol,4: 755-763.
Chang, J-S, Hong, J, Ogunseitan, OA & Olson, BH (1993) Interaction of mercuric ions with the bacterial growth medium and its effects of enzymatic reduction of mercury. Biotechnol Prog,9: 526-532.
Hassen, A, Saidi, N, Cherif, M & Boudabous, A (1998) Resistance of environmental bacteria to heavy metals. Bioresour Technol,64: 7-15.
Lilley, AK & Bailey, MJ (1997a) Impact of plasmid pQBR103 acquisition and carriage on the phytosphere fitness of Pseudomonas fluorescens SBW25: burden and benefit. Appl Environ Microbiol,63: 1584-1587.
Lilley, AK & Bailey, MJ (1997b) The acquisition of indigenous plasmids by a genetically marked pseudomonad population colonizing the sugar beet phytosphere is related to local environmental conditions. Appl Environ Microbiol,63: 1577-1583.
Lilley, AK & Bailey, MJ (2002) The transfer dynamics of Pseudomonas sp. plasmid pQBR11 in biofilms. FEMS Microbiol Ecol,42: 243-250.
Tolker-Nielsen, T, Brinch, UC, Ragas, PC, Andersen, JB, Jacobsen, CS & Molin, S (2000) Development and dynamics of Pseudomonas sp. biofilms. J Bacteriol,182: 6482-6489.
Winson, MK, Swift, S, Hill, PJ, Sims, CM, Griesmayr, G, Bycroft, BW, et al. (1998) Engineering the luxCDABE genes from Photorhabdus luminescens to provide a bioluminescent reporter for constitutive and promoter probe plasmids and mini-Tn5 constructs. FEMS Microbiol Lett,163: 193-202.
Table 1: Initial and final ratios of GFP (HgR) cells indicating conditions under which the plasmid-bearing cells were able to invade from rarein spatially structured environment. For each data point n=3 and the one-tailed t-test was used to determine if the proportion of the initially rare type had increased significantly (no asterisk P<90%;* 90%<P<95%; ** P>95%).
[Hg] (μM) / Initial density (CFU cm-1) / Initial GFP % (SEM) / Final GFP % (SEM)5 / 106 / 12.32 (1.52) / 0.41 (0.17)
7.5 / 106 / 12.32 (1.52) / 0.60 (0.10)
10 / 106 / 12.32 (1.52) / 0.88 (0.27)
12.5 / 106 / 12.32 (1.52) / 0.49 (0.04)
15 / 106 / 12.32 (1.52) / 1.21 (0.15)
17.5 / 106 / 4.94 (0.40) / 2.12 (0.30)
18.75 / 106 / 6.18 (3.01) / 32.46 (6.38)**
20 / 106 / 6.18 (3.01) / 24.72 (0.78)**
21.25 / 106 / 6.18 (3.01) / 84.69 (9.15)**
22.5 / 106 / 6.18 (3.01) / 95.50 (1.94)**
5 / 105 / 12.32 (1.52) / 0.56 (0.16)
7.5 / 105 / 12.32 (1.52) / 0.96 (0.23)
10 / 105 / 12.32 (1.52) / 1.34 (0.49)
12.5 / 105 / 12.32 (1.52) / 2.49 (0.68)
15 / 105 / 12.32 (1.52) / 3.23 (1.37)
16.25 / 105 / 6.18 (3.01) / 12.95 (2.86)*
17.5 / 105 / 4.94 (0.40) / 8.59 (1.89)*
18.75 / 105 / 6.18 (3.01) / 25.65 (5.39)**
20 / 105 / 6.18 (3.01) / 47.55 (10.68)**
23.75 / 105 / 6.18 (3.01) / 98.34 (0.43)**
5 / 104 / 12.32 (1.52) / 0.60 (0.17)
7.5 / 104 / 12.32 (1.52) / 0.92 (0.09)
10 / 104 / 12.32 (1.52) / 1.91 (0.55)
12.5 / 104 / 12.32 (1.52) / 4.60 (0.57)
15 / 104 / 12.32 (1.52) / 7.57 (2.02)
16.25 / 104 / 6.18 (3.01) / 13.77 (1.86)*
17.5 / 104 / 4.94 (0.40) / 23.00 (8.53)**
20 / 104 / 6.18 (3.01) / 25.65 (5.39)**
Table 2: Initial and final ratios of RFP (HgS) cells indicating conditions under which the plasmid-free cells were able to invade from rare in spatially structured environment. For each data point n=3 and the one-tailed t-test was used to determine if the proportion of the initially rare type had increased significantly (no asterisk P<90%;* 90%<P<95%; ** P>95%).
[Hg] (μM) / Initial density (CFU cm-1) / Initial RFP % (SEM) / Final RFP % (SEM)10 / 106 / 14.96 (2.57) / 81.17 (1.33)**
12.5 / 106 / 11.86 (1.28) / 85.85 (2.11)**
15 / 106 / 14.96 (2.57) / 72.30 (2.27)**
17.5 / 106 / 14.96 (2.57) / 75.91 (4.69)**
20 / 106 / 14.96 (2.57) / 67.63 (3.66)**
21.25 / 106 / 27.11 (4.76) / 3.74 (2.50)
22.5 / 106 / 27.11 (4.76) / 3.72 (1.32)
23.75 / 106 / 27.11 (4.76) / 0.11 (0.14)
10 / 105 / 14.96 (2.57) / 79.79 (1.45)**
12.5 / 105 / 11.86 (1.28) / 75.17 (2.25)**
15 / 105 / 14.96 (2.57) / 69.43 (0.94)**
17.5 / 105 / 14.96 (2.57) / 66.98 (6.45)**
20 / 105 / 14.96 (2.57) / 42.16 (5.17)**
21.25 / 105 / 27.11 (4.76) / 0.44 (0.31)
22.5 / 105 / 14.96 (2.57) / 0.87 (0.56)
23.75 / 105 / 27.11 (4.76) / 0.00 (0.00)
10 / 104 / 14.96 (2.57) / 78.36 (3.75)**
12.5 / 104 / 11.86 (1.28) / 71.74 (2.75)**
15 / 104 / 14.96 (2.57) / 51.78 (3.28)**
17.5 / 104 / 14.96 (2.57) / 42.26 (3.98)**
18.75 / 104 / 27.11 (4.76) / 11.47 (3.44)
20 / 104 / 27.11 (4.76) / 6.88 (2.93)
22.5 / 104 / 14.96 (2.57) / 0.07 (0.08)
23.75 / 104 / 27.11 (4.76) / 0.00 (0.00)
Table 3: Initial and final ratios of GFP (HgR) cells indicating conditions under which the plasmid-bearing cells were able to invade from rare in homogeneous environment. For each data point n=3 and the one-tailed t-test was used to determine if the proportion of the initially rare type had increased significantly (no asterisk P<90%;* 90%<P<95%; ** P>95%).
[Hg] (μM) / Initial density (CFU ml-1) / Initial GFP % (SEM) / Final plasmid bearing % (SEM)a / Final GFP % (SEM)b4 / 106 / 8.54 (1.34) / 9.78 (3.95) / 1.91 (0.77)
4.5 / 106 / 8.54 (1.34) / 8.73 (2.36) / 1.30 (0.52)
5 / 106 / 8.54 (1.34) / 2.31 (0.59) / 1.34 (0.46)
5.5 / 106 / 8.54 (1.34) / 13.34 (1.07)* / 3.42 (0.79)
6 / 106 / 8.54 (1.34) / 7.96 (3.08) / 6.36 (2.47)
6.5 / 106 / 14.97 (2.13) / 16.52 (7.46) / 5.36 (2.08)
7 / 106 / 14.97 (2.13) / 70.55 (7.31)** / 30.63 (2.61)**
7.5 / 106 / 14.97 (2.13) / 75.62 (5.62)** / 66.09 (13.12)**
4 / 105 / 8.54 (1.34) / 8.32 (2.52) / 3.26 (1.07)
4.5 / 105 / 8.54 (1.34) / 19.75 (6.37)** / 4.23 (1.46)
5 / 105 / 8.54 (1.34) / 19.57 (6.11)** / 7.58 (3.14)
5.5 / 105 / 8.54 (1.34) / 27.49 (6.29)** / 21.91 (4.18)**
6 / 105 / 14.97 (2.13) / 43.32 (11.59)** / 18.13 (2.12)**
6.5 / 105 / 14.97 (2.13) / 99.98 (4.32)** / 55.61 (6.87)**
2.5 / 104 / 14.97 (2.13) / 18.24 (2.70) / 8.49 (0.85)
3 / 104 / 8.54 (1.34) / 7.62 (2.87) / 7.13 (2.71)
3.5 / 104 / 8.54 (1.34) / 15.25 (4.07) / 13.45 (3.78)
4 / 104 / 8.54 (1.34) / 16.27 (2.63)** / 10.43 (2.68)
4.5 / 104 / 8.54 (1.34) / 43.75 (12.60)** / 36.34 (13.38)**
5 / 104 / 8.54 (1.34) / 59.92 (13.96)** / 40.16 (8.26)**
a Total proportion of mercury resistant cells (i.e. GFP cells + transconjugants).
b Proportion of GFP only cells.
Table 4: Initial and final ratios of RFP (HgS) cells indicating conditions under which the plasmid-free cells or total RFP cells were able to invade from rarein homogeneous environment. For each data point n=3 and the one-tailed t-test was used to determine if the proportion of the initially rare type had increased significantly (no asterisk P<90%;* 90%<P<95%; ** P>95%).
[Hg] (μM) / Initial density (CFU ml-1) / Initial RFP % (SEM) / Final plasmid-free % (SEM)a / Final RFP % (SEM)b4 / 106 / 13.08 (2.44) / 86.64 (2.32)** / 90.66 (1.90)**
4.5 / 106 / 13.08 (2.44) / 86.43 (1.76)** / 92.59 (1.66)**
5 / 106 / 13.08 (2.44) / 94.10 (1.84)** / 96.59 (1.36)**
5.5 / 106 / 13.08 (2.44) / 80.75 (6.43)** / 88.39 (5.19)**
6 / 106 / 13.08 (2.44) / 65.38 (4.71)** / 86.43 (3.25)**
6.5 / 106 / 9.29 (1.59) / 81.83 (1.66)** / 86.79 (0.91)**
7 / 106 / 9.29 (1.59) / 73.65 (4.55)** / 75.77 (4.23)**
7.5 / 106 / 9.29 (1.59) / 6.96 (2.71) / 8.23 (3.48)
4 / 105 / 13.08 (2.44) / 95.77 (2.68)** / 96.50 (2.17)**
4.5 / 105 / 13.08 (2.44) / 90.92 (2.02)** / 93.54 (1.27)**
5 / 105 / 13.08 (2.44) / 90.69 (0.78)** / 91.82 (0.67)**
5.5 / 105 / 13.08 (2.44) / 68.71 (1.99)** / 69.76 (1.94)**
6 / 105 / 13.08 (2.44) / 47.98 (6.90)** / 54.84 (7.62)**
6.5 / 105 / 9.29 (1.59) / 53.73 (6.27)** / 55.32 (6.42)**
7 / 105 / 9.29 (1.59) / 1.49 (1.20) / 2.13 (0.97)
3 / 104 / 13.08 (2.44) / 90.97 (3.46)** / 91.92 (3.43)**
3.5 / 104 / 13.08 (2.44) / 64.78 (4.62)** / 65.00 (4.75)**
4 / 104 / 13.08 (2.44) / 84.53 (2.11)** / 85.31 (2.06)**
4.5 / 104 / 13.08 (2.44) / 53.97 (0.91)** / 56.38 (0.81)**
5 / 104 / 13.08 (2.44) / 10.80 (3.57) / 10.80 (3.57)
a Total proportion of RFP cells minus proportion which were mercury resistant (i.e. transconjugants).
b Total proportion of RFP cells including transconjugants.
OnlineFigure 1: Relative toxicity of mercury, measured in broth (open, square symbols) and on filters (closed, triangular symbols), using the SBW25 lux reporter strain, 2 hours after inoculation. Data points shown are the mean values (n=3), error bars are standard error and the lines are the best-fit polynomial regressions lines to allow accurate interpolation(5th-order: Broth R2 = 0.993, Filter R2 = 0.999).
1