INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

Horizontal gene transfer between bacteria is a rather general process, leading to the distribution of many traits, such as antibiotic resistance determinants or genes encoding degradative pathways. Transfer occurs by several mechanisms, such as plasmid conjugation, (conjugative) transposition, bacteriophage transduction, and transformation (2, 9, 14). The efficiency and frequency of gene transfer depend on characteristics of the strain and those of the transferred element. In addition, the physiological status of the cell and various environmental parameters play a role (28, 31). In natural environments such as soil or aquatic systems, gene transfer occurs mostly at low frequencies (1, 15, 20, 21, 30, 36).

One factor which has received little attention is the role of a selective environment. Selective constraints by themselves probably do not stimulate gene transfer (a possible exception is the regulation of conjugative transposition in Bacteroides [29]) but determine whether the organisms acquiring new genetic material will multiply and grow. In light of recurring discussions on the risks of gene transfer in the environment, it is important to understand if and how factors considered to be selective actually function. The effect and importance of selective environments are well known from the problems associated with acquired antibiotic resistance in pathogenic bacteria (17, 35). In contrast, the actual selective advantage of, for instance, genes encoding degradative pathways is mostly unknown for natural environments.

Antibiotic resistance genes have been used in numerous studies as markers for gene transfer in microcosms mimicking natural ecosystems (1, 15, 18, 23, 30). Such genetic markers were used because of their easy detection, but they are not an ideal choice for the determination of the actual influence of selective conditions on gene transfer. The addition of antibiotic(s) into a microcosm can prevent growth of nonresistant microorganisms which could act as recipients and may in some cases inhibit conjugational transfer itself. Most studies which used antibiotic resistance genes to determine gene transfer frequencies were performed without any selective (antibiotic) pressure.

With degradative genes as markers, selective conditions can easily be maintained during gene transfer experiments without interference with the survival of nontarget microorganisms (21). For example, recipient bacteria having acquired new catabolic genes can be selected for by addition of the appropriate substrate. Complications in detecting and enumerating transconjugants arise when donor bacteria are also capable of growing on the specific substrate. Such complications can be circumvented by using donor and recipient bacteria each having one set of complementary degradative genes. For instance, the clc genes for the chlorocatechol pathway can transfer from the donor strain Pseudomonas sp. strain B13 to the recipient strain Pseudomonas putida F1 (22). Although the genetic basis for this transfer has not yet been fully elucidated, the transferred element carrying the clc genes seems to be a conjugative element capable of integrating into the chromosome (25). The resulting F1 transconjugants have the novel capability of metabolizing chlorobenzenes (Fig. 1), a feature which is not found in strain B13 or strain F1.

View larger version (34K): [in this window] [in a new window]   FIG. 1.   Overview of the pathway complementation for complete CB degradation (shown for 1,4-DCB). Pseudomonas sp. strain B13 possesses the ability to degrade 3-CB via 3-chlorocatechol to 3-oxoadipate. However, CB cannot be transformed to chlorocatechol by the chlorobenzoate pathway. Pseudomonas putida F1 metabolizes toluene via catechol and is also able to convert CB to chlorocatechols. Chlorocatechols were not further converted by the strain unless the genes for the chlorocatechol pathway from B13 were transferred to F1.

Here we used this complementation to study low-frequency gene transfer in activated-sludge microcosms. The selective growth advantage for transconjugants degrading chlorobenzenes enabled the observation of transfer events that would otherwise be below the detection limit. The effect of different donor and/or recipient cell densities on the clc transfer as well as the presence or absence of chlorobenzenes and 3-chlorobenzoate (3CBA) was addressed. Finally, we discuss the extent to which a specific metabolizable substrate imposes selective conditions in a complex ecosystem.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Bacterial strains and relevant characteristics. Pseudomonas sp. strain B13 can grow on 3CBA as the sole source of carbon and energy. The strain carries the self-transmissible clc genes for chlorocatechol degradation (22, 26). The clc genes were initially characterized from P. putida (pAC27) (3, 8) and partially from strain B13 (12). P. putida F1 can use toluene as the sole carbon and energy source. Toluene degradation is encoded by the chromosomally located tod genes (37, 38). Escherichia coli HB101 (pRK2013; tra⁺ Km^r) (7) and E. coli LE392 (RP1; tra⁺ Km^r) (4) were used as helper strains in triparental matings.

Media and culturing conditions. For routine growth of the strains and their enumeration via selective plating, Z3 minimal medium (34) was used with the appropriate aromatic compounds as follows: for strain B13, 3CBA; for F1, toluene; and for chlorobenzene transconjugants, monochlorobenzene (MCB) or 1,4-dichlorobenzene (DCB). Ultrapure agar (Merck AG, Dietikon, Switzerland) was used for selective plating of reactor samples to minimize background growth by indigenous microorganisms. The pH indicator bromothymol blue was added to the agar plates in a concentration of 15 mg/liter to facilitate detection of colonies of strains F1 and of possible chlorobenzene transconjugants, as they then appeared yellowish upon growth on toluene or chlorobenzenes. Growth on toluene or chlorobenzenes was tested by incubating the Z3 agar plates in gas-tight glass jars, with the growth substrates supplied through the vapor phase. 3CBA was added directly into the agar at a concentration of 5 mM. For liquid cultures, toluene, MCB, and 1,4-DCB were supplied in a secondary phase (2,2,4,4,6,8,8-heptamethylnonane [HMN]; Sigma Chemical Co., St. Louis, Mo.). Toluene was dissolved in a ratio of 0.1 (vol/vol) HMN, MCB of 0.04, and 1,4-DCB of 0.02. Per liter of Z3 medium, 400 µl of toluene (346 mg), 400 µl of MCB (443 mg), or 400 mg of 1,4-DCB was added.

Filter, plate, and liquid matings. Matings were performed with cultures pregrown on Luria-Bertani (LB) medium which were washed and resuspended in Z3 mineral medium prior to the experiments. Filter matings were performed on 0.45-µm-pore-size cellulose nitrate filters (Sartorius AG, Göttingen, Germany). Cells were transferred to the filter with a syringe, and the filter was incubated on an LB agar plate at 30°C for 24 h. Afterwards, the bacteria were resuspended from the filter by vortexing in 1 ml of Z3 medium and plating appropriate dilutions of the cell suspension on selective media. Plate matings between Pseudomonas sp. strain B13 and P. putida F1 were performed on Z3 agar plates in the presence of either MCB or 1,4-DCB as the sole source of carbon and energy. Different amounts of cells per plate (diameter, 9 cm) ranging from approximately 10⁴ to 10⁸ were tested. For liquid matings, the cells were resuspended in 200 ml of Z3 medium in a 2-liter Erlenmeyer flask and incubated on a shaker with 1,4-DCB as the only substrate. Samples of approximately 1 ml were taken regularly to determine cell numbers by selective plating and to measure chloride concentration. The chloride concentration in samples was measured with a Chlor-o-Counter according to the instructions given by the manufacturer (Flohr Instruments, Nieuwegein, The Netherlands). Preparation of cell extracts (from transconjugants and parent strains) and activity measurements of catechol 1,2-dioxygenase and catechol 2,3-dioxygenase were performed as described elsewhere (12, 22).

Microcosm experiments. Microcosms consisted of 1.0-liter glass reactors and were operated with a volume of approximately 0.5 liter of culture medium and at a dilution rate of 0.04 h¹. The medium was sterile synthetic wastewater (SWW, based on DIN 38412 T24, German Industry Norm, 1981) whose total dissolved organic carbon content we increased to 375 mg/liter by taking threefold quantities of each of the carbon substrates (meat extract and peptone). The microcosms were maintained at 30°C and a pH of between 7.4 and 7.6 by automatic temperature and pH control. The mixed microbial community was obtained from an activated-sludge treatment process. For each experiment, the reactors were inoculated with 50 ml of sludge freshly obtained from the Kloten/Opfikon municipal sewage treatment plant near Zürich, Switzerland. The sludge culture was then incubated in the reactor for at least 48 h (approximately 2 volume changes) before inoculation of B13 and/or F1 and before any of the specific substrates was supplied (see below). The microcosms were constantly and vigorously aerated with air passing through a sterile filter at a flow rate of 100 to 120 ml/min. Stirring at 400 rpm was used to ensure complete mixing. Pseudomonas sp. strain B13 and P. putida F1 were pregrown overnight on a rich medium (LB or nutrient broth). In all experiments, the strains were inoculated to a final cell density of approximately 10⁷ CFU/ml in the microcosm. An overview of the different experimental setups is given in Table 1. 3CBA was added directly into the SWW medium before autoclaving to a final concentration of 0.7 mM (109 mg/liter). Toluene was added into the medium stream at 10 µl/h (8.7 mg/h) via a syringe pump shortly before the medium entered the reactor. 1,4-DCB was added through the air, by directing an adjustable portion of the inflowing air through a vessel with crystalline 1,4-DCB. Air concentrations of 1,4-DCB were approximately 100 µg/liter. Samples of a 10-ml volume were taken from the reactors and processed to determine concentrations of aromatic compounds, optical density at 600 nm, and cell numbers of the applied strain(s). Cell numbers were derived by selective plating. The detection limit for the different strains in the reactors was 10 CFU/ml. Selective plating in combination with colony hybridization was used to verify the presence of the specific degradative genes.

DNA isolation, DNA manipulations, and hybridizations. Total DNA isolations from P. putida F1, Pseudomonas sp. strain B13, and putative transconjugants and Southern hybridizations were carried out as described elsewhere (11, 16). The DNA probe for the clc genes was a 4.5-kb PstI fragment containing the clcRAB genes from Ralstonia sp. strain JS705 (33). The probe for the tod genes was a 1.9-kb PstI fragment containing the todC₂BA region (37). As a probe for detecting the 16S ribosomal DNA (rDNA) genes, we used a 700-bp cloned 16S rDNA fragment of Ralstonia sp. strain JS705 (33). For amplification of the 16S rRNA gene, the following eubacterial primers were used: V1.1, 5'GCG.GCG.TGC.CTA.ATA.CAT.GC 3' (E. coli 16S rDNA bp 41 to 60), and V3.2, 5'ATC.TAC.GCA.TTT.CAC.CGC.TAC 3' (705 to 685); or 970404, 5'GTG.CTG.CAG.GGT.TAC.CTT.GTT.ACG.ACT 3' (E. coli 1510 to 1483), and 970405, 5'GGA.GAG.TTA.GAT.CTT.GGC.TCA.G 3' (E. coli 6 to 27). Amplification was performed by using the PCR with an annealing temperature of 50°C (primers 970404 plus 970405) or 58°C (primers V1.1 plus V3.2) and for 35 cycles. Amplified DNAs were cloned into pGEM-T Easy (Promega Corporation, Madison, Wis.). DNA sequencing was performed as described elsewhere (11).

Chemical analysis. For analysis of the 3CBA microcosm, samples were centrifuged to remove the biomass and 20 µl of 1 M phosphoric acid was added per ml to lower the pH. Concentrations of 3CBA were measured with a Gynko high-pressure liquid chromatograph (Gynkotek AG, Regensdorf, Switzerland) equipped with a Waters Nova-Pak C₁₈ reversed-phase column (Machery-Nagel AG, Oensingen, Switzerland) and UV/VIS detector set at 206 nm. The mobile phase was a solution of 50% methanol and 50% NaH₂PO₄ (50 mM [pH 3.0]) at a flow rate of 1 ml/min. For 1,4-DCB analyses, microcosm samples were extracted with 2 volumes of hexane (Fluka, Buchs, Switzerland), and the extracted samples were stored in sealed vials at 20°C until they were analyzed. 1,4-DCB was analyzed on a Hewlett-Packard 5890 series II gas chromatograph equipped with an ECD detector.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Transfer of an element containing the clc genes from Pseudomonas sp. strain B13 to P. putida F1. In filter matings on LB medium agar plates with strains B13 and F1, MCB⁺ transconjugants appeared at a low frequency of approximately 3.5 × 10⁸ per donor per 24 h. The number of transconjugants was determined from the number of colonies growing on chlorobenzene within 5 days. A similar frequency was observed in triparental matings with strains B13 and F1 and a mobilizing strain [either E. coli HB101(pRK2013) or E. coli LE392 (RP1)].

View larger version (100K): [in this window] [in a new window]   FIG. 2.   Confirmation of the transconjugants as being of F1 origin and carrying the clc genes. (A) Sau3AI (lanes 1 to 5) and HaeIII digestions (lanes 7 to 11) of PCR products obtained from different transconjugants, F1 and B13 with the eubacterial 16S rDNA primers V1.1 plus V3.2. Lanes: 1 and 7, F1; 2 and 8, B13; 3 to 5 and 9 to 11, three transconjugants; 6 and 12, molecular size marker. (B) Autoradiogram of hybridization with the clc probe to total DNAs digested with BglII. Lanes: 1 to 10, different independently derived transconjugants; 11, F1; 12, B13. Size markers are indicated on the left in kilobase pairs.

Transfer of the clc genes from Pseudomonas sp. strain B13 to P. putida F1 in liquid cultures. In mixtures of strains B13 and F1, each at approximately 5 × 10⁶ CFU/ml, incubated in a flask with Z3 mineral medium and with 1,4-DCB as the sole carbon and energy source, CB⁺ transconjugants appeared after 14 days (Fig. 3). The transconjugants grew and stabilized at a final cell density of approximately 5 × 10⁷ CFU/ml within 10 days after their first appearance. Since the transconjugant strain possessed all three phenotypes (3CBA⁺, toluene⁺, and 1,4-DCB⁺) used for selective enumeration, donor and recipient strains could no longer be enumerated at high transconjugant numbers. The transconjugants completely mineralized the 1,4-DCB in the medium, as observed from the formation of chloride (Fig. 3). In cultures operated in parallel with strain B13 or F1 only, no CB⁺ transconjugants or formation of chloride could be observed (data not shown). Interestingly, both parent strains remained viable throughout the experiment, although no growth substrate was available to them. This experiment was repeated with various lower initial cell densities of donor and recipient (10⁴ to 10⁶ CFU/ml). It was not possible to maintain such low cell densities, because both populations increased again to densities of approximately 10⁶ CFU/ml, possibly due to traces of carbon substrate in the medium.

View larger version (17K): [in this window] [in a new window]   FIG. 3.   Increase of CB-metabolizing transconjugants in a two-strain mixture of F1 and B13, incubated in mineral medium with 1,4-DCB as the sole substrate. Symbols: diamonds, chloride concentration; circles, cell numbers derived on CB plates (transconjugants); squares, cell numbers derived on 3-CBA (strain B13); triangles, cell numbers on toluene (strain F1).

Survival of and gene transfer between Pseudomonas sp. strain B13 and P. putida F1 in activated-sludge microcosms. Pseudomonas sp. strain B13 and P. putida F1 were both inoculated into two microcosms (I and II, Table 1) operated in parallel. In this and the following experiments, each of the strains was inoculated to a final cell density of approximately 10⁷ CFU/ml. No transconjugants appeared during the 2-week monitoring period, either in the reactor to which 1,4-DCB had been added or in the one without 1,4-DCB. The survival of strain F1 and especially that of strain B13 was poor. The cell numbers of these strains in the reactors decreased within a couple of days to below 10⁵ CFU/ml (Fig. 4A). Therefore, the two strains were again inoculated after 6 days. Even after reinoculation, there was no improved survival of strain B13, whereas strain F1 stabilized between 10⁴ and 10⁵ CFU/ml. There was no difference in the population sizes for B13 and F1 between the two parallel microcosms (with and without the addition of 1,4-DCB).

View larger version (22K): [in this window] [in a new window]   FIG. 4.   Survival of strains B13 and F1 and increase of CB transconjugants in activated-sludge microcosms. (A) Microcosms I and II, operated without 3CBA or toluene but microcosm number II with 1,4-DCB. (B) Microcosm III operated with 3CBA and toluene but without 1,4-DCB. (C) Microcosm IV, similar to III but with 1,4-DCB. (D) Microcosm V inoculated with B13 only. The day of inoculation after the microcosm was started with sludge is indicated by a vertical arrow. Symbols: squares, cell numbers derived on 3CBA plates (mostly B13); triangles, those derived on toluene plates (mostly F1); closed circles, those derived on plates incubated with CB (transconjugants of F1 with strain B13 clc genes); open circles, optical density of the total sludge population (OD600); open diamonds, 3CBA concentration (micromolar). The data from the sludge microcosms are typical for the outcome of the gene transfer process but not for the exact moment at which the transconjugants appeared.

View larger version (50K): [in this window] [in a new window]   FIG. 5.   Verification of the indigenous transconjugants (Ralstonia sp. strain S11) to which the clc genes of B13 were transferred. (A) Sau3AI digestions of PCR-amplified (primers V1.1 plus V3.2) 16S rDNAs from 3CBA degraders from microcosm V. Lanes: 2 to 6, strain S11; 7 to 11, B13. (B) Southern hybridization of total DNAs of S11 (lanes 1 to 4) and B13 (lanes 5 to 8) with the clc gene probe. DNAs digested with BamHI (lanes 1 and 5), HindIII (lanes 2 and 6), NotI (lanes 3 and 7), and SmaI (lanes 4 and 8) are shown.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

Gene transfer by bacterial conjugation occurs most efficiently when high numbers of donor and recipient bacteria are present in mating aggregates. This is achieved in the laboratory by defined filter or agar plate surface matings. In environments where cell densities are much lower or many different species are present, conjugation frequencies are several orders of magnitude lower. For instance, when studying transfer of the RP4p plasmid between P. fluorescens strains in filter matings and in soil microcosms, Smit (30) found 100-fold-lower transfer frequencies in soil. Similar observations were made for transfer of the plasmid pJP4 from Alcaligenes eutrophus to Variovorax paradoxus in soil microcosms (20) and for the mercury resistance plasmid pQKH6 between P. putida strains in pilot-scale percolating filter sewage treatment systems (1). In the study of gene transfer in natural environments, such observations are important. However, these observations are dependent upon direct detection of the transconjugant bacteria. In many other cases, such as transfer of chromosomal markers (30) or transfer of plasmids to uncharacterized recipient bacteria, frequencies will be too low to be detected directly. Instead of focusing on determining transfer frequencies, researchers should therefore concentrate on the possibilities of transconjugants to grow and increase their population size; it is this aspect which is often neglected in risk-associated studies.

Here we used transfer of the clc element of Pseudomonas sp. strain B13 to P. putida F1 to obtain a system suitable for studying very low gene transfer frequencies under selective conditions. The clc genes of strain B13 are very likely located on a conjugative transposable element capable of integrating into the chromosome of various other host strains (25). The estimated frequency of this conjugation and integration from B13 into F1 was rather low: 3.5 × 10⁸ per donor in filter matings. Some of our observations with plate matings suggest that the transfer may become more efficient or triggered after prolonged incubation on mineral agar plates with MCB or 1,4-DCB. Transfer of the clc genes from B13 to F1 could be easily detected because F1 transconjugants expressed a complete pathway for CB degradation (22). Since very few bacterial species efficiently degrade CBs, selective conditions could be maintained for the transconjugants by adding CB to our experimental systems.

Despite the low transfer frequencies observed in filter matings and the small likelihood of donor-recipient contact in shaking flasks and wastewater microcosms, we could detect transfer of the clc element from B13 to F1. As expected, the CB transconjugant population size increased after an initial lag period. This period probably reflects the chance for transconjugants to arise at such relative low donor and recipient cell densities. Mathematical approaches have been made to investigate kinetics of plasmid transfer in liquid matings (13, 27). However, the mass action model proposed by Levin et al. is not applicable for matings with resting cells, such as in our shaking- flask experiments (13). In the wastewater microcosms it is difficult to calculate actual transfer frequencies, since the growth rate of the transconjugants in the microcosms would have to be known, and the presence of flocs complicates the estimation of cell-to-cell contacts (27). Therefore, transfer frequencies for the clc element in the liquid matings or wastewater microcosms were not calculated. The rise of the transconjugants appeared to be dependent on the donor and recipient cell population sizes. This was concluded mainly from plate matings, and from the microcosm experiments, in which transconjugants were only found when donor and recipient cell densities could be maintained around 10⁵ CFU/ml or above. In the B13/F1 resting cell mixtures in shaking flasks, it was not possible to perform matings with cell densities lower than 10⁶ CFU/ml.

The question remains whether the presence of a unique carbon source is sufficient to create conditions which select for the growth of a specific transconjugant. The two-culture liquid mating experiment clearly demonstrated the transconjugant's growth advantage when CB was the only substrate available, since it was the only bacterium capable of utilizing CB in that system. In the activated-sludge microcosms not just one specific substrate was added; a combination of up to three substrates (3CBA, toluene, or 1,4-DCB), mixed with undefined other carbon sources (from peptone and meat extract), was added. In addition, many more species were present in the microcosms.

Without specific substrates added to the microcosms (3CBA and toluene), strains B13 and F1 maintained themselves, but at relatively low population levels. No transconjugants could be detected without or with the addition of 1,4-DCB. When 3CBA and toluene were added simultaneously, both B13 and F1 survived much better, indicating the selective effects of these substrates. However, their population sizes were still 100-fold lower than expected from biomass calculations. For example, with a 3CBA concentration of 0.7 mM in the incoming medium, about 10⁶ CFU of B13 per ml were maintained. Yield calculations predict a population size of around 10⁸ CFU/ml (at µ_max = 0.13 h¹, K_{s 3CBA} = 0.05 mM, Y_max = 61.9 mg [dry weight]/mmol [32], and a mean dry weight per cell of 1.4 × 10¹⁰ mg of C [19]). Similarly, the transconjugants' population size (5 × 10² CFU/ml) reached after addition of 1,4-DCB was 200-fold lower than expected. The added amount of 1,4-DCB (calculated from the flux of 1,4-DCB from the vapor phase into the water phase, resulting in an available dissolved concentration of 0.8 mg/liter) would have been sufficient for sustaining a population size of 10⁵ CFU/ml (24). Our observations of the relatively low maintenance of introduced bacteria in activated-sludge microcosms are similar to those described by others (21).

Competition with indigenous bacteria for the same specific substrates is certainly one explanation for a lower than maximally achievable population size, but this factor seemed to be limited to toluene. 3CBA and 1,4-DCB are not readily metabolized by the indigenous sludge bacteria, although part of these substrates may have been lost by incomplete degradation. The relative flux of a specific substrate compared to those of other metabolizable compounds has also been implicated in the selective maintenance of bacterial populations (5). However, if true, this would imply that part of the 3CBA and 1,4-DCB was not metabolized at all, since B13 and F1 transconjugants would prefer the carbon substrates from peptone and meat extract. The concept of relative fluxes is important, though, to explain that the concentration of a specific substrate per se does not lead to outgrowth of introduced or transconjugant bacteria. For example, in contrast to the limited maintenance of B13 in activated sludge plus the high input of 3CBA, 1 µM 3CBA was sufficient to promote growth of transconjugants carrying plasmid pBR60 in flow-through lake microcosms (10). In these lake microcosms, the total amount of metabolizable carbon was sufficient only for a total population size of between 10⁵ and 10⁶ CFU/ml. In the activated-sludge microcosms, limitation of a trace element (e.g., iron) may have been another reason for the low yield, based on utilization of specific substrate by the introduced or transconjugant strains. During competition with the indigenous microorganisms, part of the metabolizable carbon may be wasted by synthesizing and excreting iron chelators (6).

In summary, our results showed that although some genetic elements transfer at very low frequencies, their transfer occurs even under nonoptimal conditions such as those in sewage sludge. The transconjugants had in all cases obtained the clc element from the donor strain B13, and even without any specific recipient added to the system, the clc element ended up in an indigenous recipient. Most importantly, selection and growth of the introduced strains and transconjugants were dependent on the presence of sufficiently high concentrations of specific substrates (toluene, 3CBA, or 1,4-DCB). The detection of F1 transconjugants' acquisition of the clc genes from strain B13 was only possible due to their selective advantage of utilizing 1,4-DCB, thereby increasing the population's size to detectable densities. This demonstrates how the evolution of a new metabolic pathway can be based on rare transfer events and weak selective forces, as it most likely occurs under environmental conditions.