|Home | About | Journals | Submit | Contact Us | Français|
There is a wealth of evidence indicating that mobile genetic elements can spread in natural microbial communities. However, little is known regarding the fraction of the community that actually engages in this behavior. Here we report on a new approach to quantify the fraction of a bacterial community that is able to receive and maintain an exogenous conjugal plasmid termed community permissiveness. Conjugal transfer of a broad-host-range plasmid labeled with a zygotically inducible green fluorescent protein (RP4::gfp) from a donor strain (Pseudomonas putida) to a soil bacterial suspension was examined. The mixture of cells was incubated on membrane filters supported by different solid media. Plasmid transfer was scored by in situ visualization of green fluorescent transconjugant microcolonies, and host range was determined by traditional plating or microcolony isolation by using a micromanipulator. Among the conditions tested, the highest plasmid transfer incidence (approximately 1 transfer per 104 soil bacteria) was measured after 48 h of incubation on either a 10% soil extract or a 10-fold diluted R2A medium. Stereomicroscopy combined with image analysis allowed easy examination and enumeration of green fluorescent microcolonies. In all experiments, however, stereomicroscopy consistently underestimated the number of conjugation events (approximately 10-fold) in comparison to confocal laser scanning microscopy. The plasmid host range was broad and included bacteria belonging to the Alphaproteobacteria, Betaproteobacteria, and Gammaproteobacteria classes of proteobacteria. The isolation of transconjugant microcolonies by micromanipulation greatly extended the estimated plasmid host range among soil bacteria. The new approach can be applied to examine the permissiveness of various communities toward receipt of different mobile elements.
Horizontal gene transfer (HGT) among bacterial populations may provide microbial communities with genetic variability to adapt to environmental changes. Conjugation, which consists of the transfer of bacterial mobile genetic elements (e.g., plasmids, conjugative transposons, etc.), is often believed to be the most important process for short-term microbial adaptation (27). The extent of conjugal exchange of genetic elements among environmental bacteria depends on various biotic and abiotic factors (13, 23, 31), including bacterial relatedness (15), plasmid host type (5), and conjugative element type (4). The capacity and diversity of bacteria in a microbial community taking part in exchange of mobile genetic elements are poorly understood, primarily due to the traditional use of biased cultivation-based methods (25, 30).
Recently, methods relying on fluorescent reporter genes (e.g., gfp) in combination with confocal microscopy and flow cytometry have been developed, allowing for in situ visualization and quantification of plasmid transfer (7, 20) and host range examination (17) in indigenous microbial communities. These methods allow, for the first time, detection of HGT in indigenous organisms with unknown phenotypes, reducing the cultivation bias. However, they will mainly detect transfer to the most abundant recipients and are further biased when transconjugant division is possible. Therefore, the actual fraction of a microbial community that is actively engaged in uptake and exchange of a mobile genetic element is typically not measured.
We have developed a new approach to quantify the fraction of a soil microbial community that is able to receive an exogenous conjugal plasmid termed soil community permissiveness. Transfer and maintenance of a green fluorescence reporter gene (gfp)-tagged plasmid to indigenous soil bacteria are examined in solid-surface matings. Conjugation events and recipient morphology are visualized in situ and quantified by minimal-cultivation methods, confocal laser scanning microscopy (CLSM), and stereomicroscopy (SM), and the phylogeny of isolated transconjugant bacteria was determined.
The plasmid donor strain was Pseudomonas putida KT2442 (1) chromosomally tagged with lacIq1 (3) and a constitutively expressed dsRed gene fused to a rrnBP1 promoter on a mini-Tn5 cassette (29). The plasmid donor strain harbored a 60-kb conjugal IncP1 plasmid RP4 (21), which mediates resistance against the antibiotics kanamycin, ampicillin, and tetracycline. The plasmid was tagged with a mini-Tn5 insertion of a gfpmut3b gene downstream of a synthetic LacI-repressible PA1/O4/O3 promoter and a kanamycin resistance gene following established procedures (3). Expression of the gfp gene was repressed in plasmid donor cells but permitted upon transfer to the indigenous recipient cells. The plasmid donor strain was grown in 1/4 tryptone soy broth (Fluka) supplemented with tetracycline (20 μg/ml) and vigorously shaken (250 rpm) at 30°C. The cells were harvested in the late exponential phase of growth and washed twice in 0.9% sterile saline solution.
The indigenous soil bacteria, used as plasmid recipients in all experiments, were extracted from loamy sand soil at an annually manured agricultural field site (CRUCIAL [Closing the Rural Urban Nutrient Cycle] site, Taastrup, Denmark) (14). Soil samples were collected in late fall. Briefly, each gram of sieved soil (2-mm mesh size) was diluted in 4 ml TTSP (tetrasodium pyrophosphate [50 mM], Tween 80 [0.05%]) (32) and homogenized (twice for 30 s) in a blender (Waring, Torrington, CT) at high speed. The soil suspension was sonicated for 10 min in a water bath (Bransonic, Danbury, CT) and carefully transferred on top of Nycodenz solution (Nycomed Pharma, Norway; 1.3 g/ml) followed by centrifugation (8,500 × g for 15 min). The upper and intermediate phases containing the bacterial layer were carefully collected, diluted 5-fold (vol/vol) in sterile saline, and centrifuged at 8,500 × g for 15 min. The resulting pellet was resuspended in sterile saline and filtered through a sterile 20-μm-pore-size filter, yielding a soil cell suspension. The donor and indigenous bacterial suspensions were examined for correct fluorescent properties by epifluorescence microscopy (Zeiss Axioskop 2). Sonication reduced the culturable fraction by approximately 26% (data not shown), and the overall net culturable fraction (on R2A agar medium) was approximately 15%.
The optical density of each bacterial suspension was measured and adjusted with sterile saline to approximately 5 × 107 CFU/ml (optical density at 600 nm [OD600] = 0.5) for indigenous soil cell suspensions and approximately 5 × 106 CFU/ml (OD600 = 0.05) for plasmid donor cell suspensions. The density of cultivable cells in the bacterial suspensions was estimated by drop plating (20 μl) on R2A agar plates (24). Furthermore, the total number of indigenous bacterial cells in suspension was estimated by direct microscopic counting of cells stained with Syto 9 (5 μM; Molecular Probes, Inc., Eugene, OR) filtered on a 0.2-μm-pore-size filter. Stained cells were visualized and quantified using an epifluorescence microscope (Zeiss Axioskop 2) equipped with an ocular Whipple counting grid with 100 squares (19, 22).
Mating was initiated by mixing equal volumes (vol/vol) of plasmid donor (D) and indigenous recipient (R) bacterial suspensions, and 2 ml of mating suspension was transferred to a sterile 0.22-μm Cyclopore membrane (Whatman) by vacuum filtration (<0.4 mbar). The filter area exposed to bacterial cells was estimated to be 380 mm2, resulting in initial cell densities of approximately 104 and 105 cells/mm2 for plasmid donor and soil indigenous bacteria, respectively. The mating filters were subsequently transferred cell side up to a solid medium. Four solid media were examined. Phosphate-buffered saline (PBS) agar plates contained 8 g NaCl, 0.2 g KCl, 1.44 g Na2HPO4, and 0.24 g KH2PO4 per liter of MilliQ water, and the pH was adjusted to 7.2. Ten percent R2A agar plates were made by 10-fold dilution of R2A medium (0.5 g yeast extract, 0.5 g proteose peptone, 0.5 g casein hydrolysate, 0.5 g glucose, 0.5 g soluble starch, 0.3 g sodium pyruvate, 0.3 g K2HPO4, and 0.05 g MgSO4·7H2O per liter) in PBS solution. Soil extract was prepared by a modified version of the protocol of Shimomura et al. (26), where a 1:1 mixture (vol/vol) of fresh soil and MilliQ water was combined in a horizontally placed Erlenmeyer flask on a shaker (200 rpm) for 3 h. Following a 5-h passive settling of soil particles, the supernatant was collected, autoclaved, and used with PBS solution to prepare 1% and 10% soil extract solid media (1% SE and 10% SE, respectively). All agar plates were supplemented with 15 g/liter agar.
Mating filters on solid media were incubated at 25°C for 24 or 48 h, transferred onto PBS solid medium, and stored at 5°C for 4 days prior to microscopic plasmid transfer visualization and quantification. The mating filters intended for host range determination were incubated for 48 h on 10% R2A and 10% SE solid media or for 96 h on 1% SE solid medium. The filters were stored at 5°C for 3 days prior to traditional plating of dislodged cells on selective media or a direct isolation of microcolonies by micromanipulation (see below). All filter matings were performed in quadruplicate. Control experiments contained only one of the two bacterial suspensions.
The incidences of RP4::gfp plasmid transfer to the soil cells were visualized in situ by confocal laser scanning microscopy (CLSM) and stereomicroscopy (SM) and quantified by automated image analysis (Image Pro V5.1; Media Cybernetics, Silver Spring, MD).
The CLSM (SP5; Leica) was equipped with an argon and helium neon laser providing 488-nm and 568-nm laser line excitation of green fluorescent protein (GFP) and DsRed, respectively. The fluorescence emitted from transconjugant cells (i.e., GFP) and donor cells (i.e., DsRed) was detected in fluorescence channels PMT1 (500 to 550 nm) and PMT2 (600 to 650 nm), respectively. Randomly selected fields on a filter surface were scanned using the tile mosaic function (LAS AF 1.7.0; Leica) where two to four neighboring fields were individually scanned at a particular scanning region of a filter. Scanning was repeated at multiple locations on a filter, and ca. 50 microscopic fields from each duplicate filter were scanned. The imaging depth in the axial (z) direction was individually adjusted for each scanning field or region, with a step size of 1 μm. One field of view of the 63× Leica oil immersion objective (numerical aperture of 1.4) corresponded to 246 μm2 of filter area, and approximately 1% of the total filter area was examined by that means.
Stereomicroscopic (SM) visualization was made with a Leica MZ16 FA fluorescence stereomicroscope equipped with a 10× plan apochromatic objective, a 10× eyepiece (10×/21B) and a 40× magnification zoom. Image Pro Plus software allowed a semiautomated image acquisition. A scanning zone, composed of 5 × 5 neighboring fields of view (0.6 mm2 per field), was examined for GFP signal (excitation 480/20 nm, emission 525/40 nm) and DsRed signal (excitation 565/25 nm, emission 620/60 nm). Two representative scanning zones were analyzed per filter, and images were acquired by a Leica DFC300 fluorescence camera. The total scanned area corresponded to approximately 8% of the filter area.
Automated quantitative image analysis was performed with a macro in the Image Pro Plus software. This macro successively extracted and subtracted the background from the original image and then performed a best-fit equalization of the image intensity before detecting bright objects larger than 7 μm2 on the basis of automatic segmentation. The presence of several continuous or clustered GFP-expressing cells in colorless recipient microcolonies was counted as one transfer event (Fig. (Fig.11 and and2).2). All images were manually controlled for enumeration errors, and values were corrected if significant variations were found. The CLSM images were analyzed in full size, while the analysis of the stereomicroscope images was limited to their elliptic central area (ca. 0.46 mm2), excluding their poorly illuminated corners. The results were used to estimate the number of GFP-expressing microcolonies per total filter surface area.
The selective plating approach involved dislodging the cells from the mating filters in 1 ml sterile saline by vortexing (2,000 rpm, 2 min) and spreading the cell suspensions (500 μl; ca. 4,000 cells ml−1) on 5-fold-diluted SE or R2A agar plates (15 cm in diameter) supplemented with tetracycline (10 μg/ml) and nystatin (50 μg/ml). The plates were incubated at 25°C for 5 days and stored at 5°C for 3 days to improve GFP protein maturation. The growing colonies were checked for expression of GFP green fluorescence by SM (1× objective, 10× magnification zoom, see above), and transconjugants were restreaked on selective solid medium prior to phylogenetic characterization.
Transconjugant microcolonies were also isolated directly from the mating filters by micromanipulation (MM) in combination with stereomicroscopy. Pointed tips of Pasteur pipettes (Sigma-Aldrich) with a reduced diameter (approximately 15 μm) were manually produced by heating the tips in a flame, pulling, and coating with a commercial silver stain at either the outer or inner glass surface. Silver coating the tips in combination with a bright-light source placed parallel to microscopic stage improved tip visualization. The pipettes were carefully navigated onto mating filters (ca. 55° angle) by using a motorized micromanipulator stage and Unisense Profix software (step size of 5 μm). The tip was introduced into individual green microcolonies (≥15 μm in diameter; approximately 75 to 250 cells), allowing cell adsorption to the outer tip surface or suction into the glass tubing by capillary forces. The tip was then retracted, and the cells were transferred into 25 μl sterile saline. The resulting cell suspensions were drop plated (3 μl) on the respective solid medium, amended or unamended with tetracycline (10 μg/ml) and nystatin (50 μg/ml). Following 7 to 21 days of incubation, growing transconjugant green-fluorescing microcolonies were checked microscopically and differentiated from red-fluorescing donor cells, which occasionally were coisolated. Most of the isolated transconjugants (54 to 72%) were able to form colonies on solid media.
The isolated transconjugants were purified on selective agar plates and grown in liquid media supplemented with tetracycline (10 μg/ml) for 5 to 14 days, prior to storage and DNA extraction by a triple boiling-dry ice freezing method (100-μl bacterial suspension boiled at 100°C for 5 min immediately followed by freezing at −78°C for 5 min). The cell lysate was vortexed (2,500 rpm, 1 min) prior to centrifugation (7,500 × g, 10 min), and supernatant containing extracted genomic DNA was used for PCR amplification of 16S rRNA gene using a primer set specific for the Bacteria domain, primers 27F (5′-AGAGTTTGATCCTGGCTCAG-3′) and 1492R (5′-GGTTACCTTGTTACGACTT-3′) (11). PCR mixtures (25 μl) contained 2 μl of bacterial lysate, 0.2 U Taq polymerase, 12.5 pmol of each primer, 50 mmol KCl, 30 mmol Tris-HCl, 1.5 mmol MgCl2, and 12.5 mmol deoxynucleoside triphosphates (dNTPs) (Sigma-Aldrich). PCR amplification was run as follows: (i) an initial denaturation step of 2 min at 96°C; (ii) 35 cycles, with 1 cycle consisting of 10 s at 92°C, 20 s at 53°C, and 1.5 min at 72°C; and (iii) a final extension reaction of 6 min at 72°C. The PCR products were analyzed on 1% (wt/vol) agarose gels, purified by PCR purification kit (Qiagene) and used for sequencing reactions with primer 27F (Macrogen). The partial sequences (400 to 750 bp) were analyzed by BioEdit (www.mbio.ncsu.edu/BioEdit/bioedit.html) and compared to sequences in nucleotide databases using NCBI/BLAST (www.ncbi.nlm.nih.gov/BLAST). Furthermore, the 16S rRNA gene sequences of isolated transconjugants obtained for each experimental condition were mutually aligned, and the longest identical 16S rRNA gene sequence was identified.
The changes in the microbial community structure during incubation of filters on different solid media were monitored by denaturing gradient gel electrophoresis (DGGE) fingerprinting. Bacterial cells from the filters containing mating mixtures, incubated on 10% R2A and 10% SE media, were dislodged by vortexing (2,500 rpm, 2 min) in 1 ml DNA/RNA free water (molecular biology grade; 5-Prime). Bacterial DNA was extracted by a double boiling-dry ice freezing method, and 2 μl of cell lysate was used in PCRs with primer 341F (5′-CCTACGGGAGGCAGCAG-3′) with a GC clamp and primer 518R (5′-ATTACCGCGGCTGCTGG-3′) (18). The PCR amplification conditions were as follows: (i) initial denaturation step of 4 min at 94°C; (ii) 30 cycles, with 1 cycle consisting of 30 s at 92°C, 30 s at 53°C, and 2 min at 72°C; and (iii) a final single 10-min cycle at 72°C, followed by cooling at 6°C. The PCR products were loaded on a DGGE gel with a 30% to 60% denaturation gradient and run at 60°C and 70 V for 17 h. The gel was stained with SYBR Gold (Molecular Probes) in 1× Tris-borate-EDTA (TBE) buffer for 30 min in the dark, and DNA was visualized with a GelDoc (Bio-Rad Laboratories).
The newly determined 16S rRNA gene sequences were deposited in the GenBank database under the accession numbers GU989360 to GU989431.
The permissiveness of a soil community toward receipt and maintenance of the exogenous conjugal plasmid RP4::gfp harbored in P. putida KT2442 was examined under surface mating conditions. To maximize the requisite cell-to-cell contact for plasmid transfer while minimizing the extent of compositional changes during the assay, different conditions were evaluated. The initial cell densities of 104 donor (D) and 105 recipient (R) cells per mm2 allowed for maximal D-to-R contacts during 24-h and 48-h mating experiments, except on the most nutrient-poor media (i.e., PBS and 1% soil extract [1% SE]). Plasmid transfer to the soil community was observed under all tested conditions (Fig. (Fig.1)1) and was demonstrated to involve several morphologically distinct bacteria as hosts (Fig. (Fig.2).2). However, the number of detected transfer events varied significantly. In the most diluted media, 10- to 50-fold fewer transfer events were detected than in the nutrient-rich media. Except with the PBS medium, prolonging the incubation from 24 to 48 h had little effect on observed transfer events. Detection of GFP signal was, however, easier after 48 h, possibly because of transconjugant cell division or plasmid retransfer within microcolonies.
The fraction of conjugation-permissive cells was estimated as the number of partly or completely GFP-expressing microcolonies (T) divided by the initial number of soil bacteria (R), enumerated by direct count (8.40 × 107 ± 0.42 × 107 [n = 3]). A large fraction of the soil community was able to receive the plasmid as revealed on 10% R2A and 10% SE media. The highest number of conjugation events was observed by CLSM after 48 h of incubation on 10% SE (permissive fraction; 1.49 × 10−4 ± 0.30 × 10−4) and 10% R2A medium (permissive fraction; 5.89 × 10−4 ± 0.71 × 10−4) (Fig. (Fig.3).3). A much smaller number of conjugation events was observed on PBS and 1% SE solid media (Fig. (Fig.11 and and3).3). With CLSM method, more conjugation events (up to 1 order of magnitude more) were consistently scored compared to the SM method.
The diversity of the soil community was to a large extent maintained on both 10% SE and 10% R2A media after 48 h, with a slight reduction in diversity on the nutrient-richer medium (10% R2A). The presence and growth of plasmid donor cells on mating filters, however, reduced the observed diversity (see Fig. S1 in the supplemental material), effectively displacing some of the dominant DGGE bands.
The phylogenetic identification of transconjugants revealed a broad host range for conjugal plasmid, as the transconjugants identified were affiliated with 9 different genera belonging to the Alphaproteobacteria (Sinorhizobium), Betaproteobacteria (Achromobacter, Variovorax, Collimonas, and Alcaligenes) and Gammaproteobacteria (Pseudomonas, Lysobacter, Xanthomonas, and Stenotrophomonas) (Table (Table1),1), as well as an uncharacterized rhizosphere bacterium identified on 10% SE medium by micromanipulation (GenBank accession no. GU989388). The traditional selective plating technique identified only four different dominant genera. The transconjugant 16S sequences showed a high similarity (>99% identical) to formerly deposited 16S sequences in GenBank (see Table S1 in the supplemental material).
Predicting the fate of a conjugal plasmid in a microbial community requires, in addition to knowledge regarding its transfer and stability kinetics, an estimate of the bacterial fraction that can be actively engaged in its horizontal transfer, i.e., the plasmid's extant host range. Therefore, we developed a method to quantify the fraction of a soil bacterial community that can receive and maintain a conjugal plasmid.
The perfect method to estimate the in situ host range of a plasmid would use or mimic in situ conditions to the maximum extent possible. At the same time, transfer events need to be observable. As a result, we developed a method where event detection and community accessibility are maximized, while community compositional changes are minimized. Pseudomonas putida KT2442 and RP4 were chosen as the donor strain and model plasmid, respectively.
To maximize event detection, we took advantage of zygotic expression of biofluorescence encoded by a gfp gene under the control of a strong and widely expressed promoter PA1/O4/O3 (12). Because this promoter is LacI repressible, the inserted gfp gene is silent in the donor strain (3) but is expected to be expressed in most environmental isolates, as its cognate promoter repressor (lacI) is restricted to enteric bacteria (34). While transfer into members of the family Enterobacteriaceae would, therefore, go undetected, the abundance of enteric bacteria in soils is typically low (8, 9), supporting the value of the gfp construct. Others, working with a similar zygotic gfp-based approach, have reported a higher number of conjugation events measured by direct epifluorescence detection compared to culture-based enumeration methods (7). To further enhance our ability to detect transfer events, we allowed the soil bacteria to grow into microcolonies during the assay, as plasmid retransfer within a clonal population is often very efficient. The rapid plasmid sweep within a microcolony, in combination with the proliferation of cells harboring plasmid, effectively amplifies the GFP signal, facilitating microscopic detection (Fig. (Fig.11 and and22).
An unbiased estimate of the true host range of a plasmid requires that all community members are equally accessible by the exogenous plasmid donor. This is a potential limitation of previous assays, where donor cells are applied to solid or liquid samples without due consideration of cell-cell contact (17, 20, 28). Hence, in the present assay, cells were dislodged from the soil environment and mixed with the donor cells at a 10/1 ratio before the mixture was filtered on a membrane. Because the donor cells proliferate faster than the indigenous microcolonies, plentiful cell-cell contacts were established after 24 h, permitting plasmid transfer. The cell ratio and media chosen are the consequences of preliminary experiments, where 1/1,000 and 1/1 ratios of donor to soil bacterial cells incubated for 24 h on R2A medium were tested, showing hardly any transfer or excessive filter coverage by DsRed donor cells (see Fig. S2 in the supplemental material), respectively.
While mixing of the cell suspensions in liquid media might provide even more contact between donor and potential recipient cells, such assays are inherently difficult to interpret, because one cannot distinguish between those transconjugants resulting from growth from those that are the results of plasmid transfer or retransfer (6). In the current assay, both single (due to singular transfer event) and multiple (due to retransfer or transconjugant growth) GFP signals in an indigenous microcolony are correctly counted as a singular donor-to-recipient conjugation event. However, retransfer events from soil transconjugants to different soil bacteria, although rare, would typically not be detected by our method, unless they result in separate and distinguishable microcolonies. An additional benefit of the surface assay may be the protection of mating pairs, as postulated for several plasmids (2, 10).
To minimize community compositional changes during the assay, we examined media with different degrees of nutrient content. The smallest changes are likely attained in the most nutrient-poor conditions (PBS or 1% SE in PBS). Under these conditions, transconjugant fractions of approximately 10−5 and 10−6 were detected by CSLM and SM, respectively. Increasing the nutrient content significantly increased transconjugant fraction. Ten percent R2A and 10% SE solid media gave fractions above 10−4 and 10−5 as detected by CLSM and SM, respectively. These values indicate that a very large fraction (1 in 10,000) of cells within a typical soil community may partake in transfer and maintenance of the model IncP1 plasmid.
Because our results are not biased by transconjugant growth, the higher estimates on the media richer in nutrients are likely to be due to better cell-cell contact (Fig. 1A and C) and the benefit of nutrients to support higher metabolic activity in donors and recipients. Lower metabolic activity and less cell-cell contact on PBS and 1% SE solid media may thus have resulted in significantly less transfer. While the 10% SE and 10% R2A media qualitatively retain high community diversity (which is significantly reduced when mixed with donor cells) (see Fig. S1 in the supplemental material), we advocate the use of 10% SE medium because it has the highest nutritional similarity to indigenous soil conditions.
More conjugation events (1 order of magnitude more) could be detected with CLSM than with SM. However, the scanning procedure and image analysis associated with CSLM was quite time-consuming, requiring manual z-axis definition for every scanned field and several manual corrections of automatically analyzed images. In contrast, very few adjustments and corrections were necessary for images taken by SM. Hence, SM analysis is recommended for comparative studies, while absolute estimates would require CSLM observation.
Morphologically distinct transconjugant microcolonies were observed (Fig. (Fig.2),2), suggesting that a diversity of indigenous bacteria received the plasmid. The automated SM analysis can, thanks to its large working distance objective, accommodate a micromanipulator that allowed direct isolation of transconjugant microcolonies to infer transconjugant phylogeny.
The host range of RP4 (IncP1) plasmid appeared to be very broad, transferring from donor strain P. putida KT2440::lacIq1 to bacteria belonging to the Alphaproteobacteria, Betaproteobacteria, and Gammaproteobacteria classes of proteobacteria. The conjugal IncP1 plasmids have previously been shown to have a broad host range, transferring frequently between Gram-negative bacteria (5, 16). In the present study, selection of transconjugants by micromanipulation increased the diversity of identified transconjugants from four to nine genera plus an uncharacterized soil bacterium compared to the diversity identified by traditional selective cultivation assay. Previous attempts to reduce cultivation bias when estimating host range have included, e.g., flow cytometric cell sorting (17) or in situ hybridization (6) of transconjugants, making cells inviable or highly unlikely to identify rare transconjugants among dominant cells. The wider diversity of the transconjugants detected by our new assay may be related to a high fraction of microcolony-forming soil cells (micro-CFU) (33). The micromanipulation method combined with the use of fluorescent proteins offered the opportunity to isolate these microcolonies with specific morphotypes. Furthermore, the transconjugant diversity detected was in general larger on 10% SE medium than on 10% R2A medium by both isolation techniques. Interestingly, Achromobacter, one of the most dominant transconjugant genera isolated on soil extract media, irrespective of the isolation method, was not present among the 49 identified isolates from 10% R2A medium. Furthermore, the genus Variovorax, frequently observed as a transconjugant on 1% soil extract, was completely absent on 10% R2A medium and barely observed on 10% SE medium, indicating that a change in nutrient composition and strength favors different bacterial types.
In summary, we have developed a minimal-cultivation approach in combination with zygotic fluorescence expression and microscopy to quantify, for the first time, the recipient fraction of a soil microbial community for a target conjugal plasmid, termed the community permissiveness. This new approach can be used to examine dissemination and mobilization capacity of various plasmids and other genetic elements among bacterial communities that originated from different or differently treated environments. The isolation of transconjugant microcolonies by micromanipulator revealed a wider diversity of transconjugants than observed by traditional selective technique using colony growth on solid media.
This work was funded by a Marie Curie Excellence Grant (MEXT-CT-2005-024004, RaMAda) to B.F.S. and the Villum Kann Rasmussen Foundation Center of Excellence CREAM (Center for Environmental and Agricultural Microbiology).
Published ahead of print on 28 May 2010.
†Supplemental material for this article may be found at http://aem.asm.org/.