Tn916 is a conjugative transposon that encodes resistance to the antimicrobial drug tetracycline. In sterile soil models, Tn916 was able to transfer between Bacillus subtilis or Enterococcus faecalis and Bacillus thuringiensis subsp. israelensis. Conjugal transfer of the transposon occurred at frequencies of approximately 2.1 x 10^-5 conjugants/donor; conjugal transposition in soil systems was DNase-independent. In further experiments, Tn916 transfer was characterized over a variety of soil conditions, including variations in moisture, temperature, pH and available nutrients. Finally, Tn916 transfer was demonstrated in non-sterile soil models at frequencies similar to those observed under sterile conditions. Thus, these data suggest Tn916-mediated genetic exchange can occur in soils.

Recently, much emphasis has been placed on bacterial gene transfer and acquisition of antibiotic resistance in the environment. Of special interest is the fate of the DNA from genetically engineered microorganisms that are proposed for field studies. One method by which bacteria exchange genetic information in the environment is by conjugal transfer.

The conjugative transposon Tn916 (16.4 kb) was originally isolated from Enterococcus faecalis. The transposon encodes resistance to tetracycline, tetM (Franke and Clewell, 1981). Moreover, conjugal transfer of Tn916 has been shown among a wide variety of recipients (Bertram et al., 1991). In addition to the ability to transfer by a conjugative mechanism, Tn916 induces mobilization of other donor genetic elements (Naglich and Andrews, 1988). Due to this ability, Tn916 may be a factor in gene transfer in the environment. In the laboratory, matings of Tn916-containing selectable donors and recipients are usually carried out on nitrocellulose membranes (Naglich and Andrews, 1988). The matrix of the filter is thought to serve as a medium to bring the cells into contact so DNA transfer can occur (Clewell, 1990). Although the filter mating is an efficient method of studying this type of conjugation, it gives no indication of how Tn916 will function in nature. The goal of the research presented herein was to explore transposon-mediated genetic exchange in the soil environment.

MATERIALS AND METHODS

Bacterial strains and media. The Bacillus donor used in these experiments was a derivative of Bacillus subtilus 168 in which Tn916 had inserted into the chromosome. This strain was designated AN861 (Naglich and Andrews, 1988b). The donor used in Enterococcus matings was Enterococcus faecalis AP-2 that carries a chromosomal insertion of Tn916 (Naglich and Andrews, 1988a). Both donors are therefore resistant to tetracycline (10 mg/ml). The recipient in all cases was B. thuringiensis subspecies israelensis AN142 in which a chromosomal kanamycin resistance marker was induced by gradient plate (200 mg/ml). The media and selection procedures were done essentially as described previously (Naglich and Andrews, 1988a; 1988b).

Soil. The soil used in all these experiments was a fine sandy-loam (Typic Hapludolls) obtained from a soybean field in central Iowa. The sample was sieved through a 1 cm screen to remove debris and stored at 4^oC. The soil had an initial pH of 7.4 and an organic matter level of 1.8%. The water content of the soil at 33Kpa was 7.5% (v/w) unless otherwise specified.

Mating procedure. To prepare soil for mating experiments, samples were air-dried for seven days then dispensed into flat-bottom centrifuge bottles (100 g per bottle). For procedures requiring sterile soil, the samples (in centrifuge bottles) were autoclaved twice for two hours with an overnight cool-down period between sterilizations. Once sterile, the water content was adjusted to 7.5% with sterile water 24 hrs prior to inoculation. Any amendments to the soil were made at this point. The samples were then held at 30^oC overnight before inoculation.

Donor and recipient cultures were grown to mid-log phase and used to inoculate the soil (Naglich and Andrews, 1988a; 1988b). For inoculation, 1 ml of recipient was added to the soil followed immediately by 1 ml of the donor. The inoculated soil was rolled gently (5 min) to mix the inoculum and incubated at 30^oC for 24 hrs. The cells were harvested by adding 50 ml of sterile LB broth and shaking gently. Aliquots were then spread onto LB agar plates containing the appropriate selection. After an overnight incubation, the colonies were enumerated and results expressed in terms of conjugants per output donor.

Soil amendments. To alter the moisture content of the soil, the amount of sterile water used for moisture adjustment was varied from 2-13% (v/w). When effect of temperature on conjugation was studied, the soil samples were incubated at the appropriate temperature overnight before inoculation, as well as during the mating. In matings where the nutrient content of the soil was amended, powdered brain-heart infusion (Difco) was added to the sterile water used for moisture adjustment in amounts from 0.5 to 2.5 mg/g of soil. For pH adjustments of the soil, the samples were amended with Ca(OH)₂ or Al₂(SO₄)₃ to alter the pH of the soil in a range from 3.9 to 9.2. After pH adjustment, soil samples were held at 4^oC for two weeks prior to inoculation. Matings performed in the presence of DNase were supplemented with 400 mg of DNase per gram of soil.

The initial matings were between B. subtilus AN861 and B. thuringiensis subsp. israelensis. In a sterile soil environment, transfer of Tn916 was observed at a frequency of 2.1x10^-5 conjugants/donor. To control for spontaneous mutation, mating bottles containing only donor or recipient were subject to the appropriate selection pressure; mutation to double antibiotic resistance was not detected in any mating performed.

To assure that the genetic exchange observed was the result of conjugation and not transformation, the same strains were mated in the presence of DNase. The incorporation of DNase into the soil had essentially no effect on the mating. To assure that the DNase remained active in the soil environment, a sample of plasmid pUC19 (5 µg) was mixed with 10 mg of soil and incubated at 30^oC. for 30 min. After this incubation samples containing the enzyme extensively degraded the plasmid whereas intact plasmid was readily recovered from soil samples not containing DNase (data not shown).

To better understand how Tn916 functioned in the soil environment, the conjugal frequency of Tn916 was examined over a range of soil moistures. The moisture content of several soil samples was adjusted so a range of moistures between 2 and 13% was represented. The data from these matings are presented in Figure 1. Interestingly, lower soil moisture content resulted in higher transfer frequencies.

The nutrient content of a soil may vary widely from point to point, even over a relatively short distance. To evaluate the effect of various amounts of nutrients on conjugation, donor and recipient cultures were inoculated into soil samples with amended nutrient contents. The results from these experiments are shown in Figure 2. Conjugation was detected over the entire range of amendments but the frequency of transfer increased as nutrient amendment increased.

The temperature found in a soil environment may vary considerably depending on time of the year and meteorological conditions. Figure 3 shows data resulting from alterations to the incubation temperature of the mating. The incubation temperature was varied to more closely represent the daily or monthly changes in topsoil temperature that may occur in a natural environment. These results showed detectable frequencies of Tn916 transfer only between 18^o and 35^oC with an optimum occurring at 30^oC.

Soil pH is a parameter that may be quite variable depending on the type of soil and site chosen. To assess the ability of Tn916 to transfer in soils of varying pH, a series of soil samples with altered pH values were inoculated with donor and recipient (Fig. 4). Conjugation was only detected in the range of pH from 5.95 to 8.83.

To examine the ability of Tn916 to enter the soil population, it was of interest to use a donor that might be more commonly found in a soil environment. Fecal organisms, such as the fecal enterococci, are commonly present in animal and human waste materials deposited on the soil. Therefore, in addition to matings between Bacillus species, experiments were performed using E. faecalis AP-2 as the donor. The recipient remained B. thuringiensis subsp. israelensis. In these experiments, Tn916 was shown to transfer at a frequency of 2.3x10^-6.

Because the soil system presented in this work uses sterile soil, it may be difficult to extrapolate the results to a natural soil environment (Krasovsky and Stotzky, 1987). To better understand the role of the soil microflora and enzymatic composition on conjugal transfer by Tn916, parallel matings were performed in native as well as sterile soil. Tn916 was seen to transfer in sterile soil at a frequency of 6.18x10^-5 and in nonsterile soil at a frequency of 2.11x10^-5. The results from these experiments showed that conjugal transfer of Tn916 was observed at comparable frequencies in both sterile and non-sterile soils.

A sterile soil system was chosen for initial work in this study to avoid confusion that might be introduced by competing soil microflora. The first goal of this research was to determine if the physical environment in soil is suitable for conjugal activity. The results clearly show the ability of Tn916 to conjugally transfer in the physical matrix provided by the soil. Interestingly, these data show that, in the sterile soil environment, the frequency of conjugal transposition is approximately the same as that seen with a filter mating (Naglich and Andrews, 1988a; 1988b). Moreover, the data suggest that the genetic exchange observed results from a conjugation-like event. Several lines of evidence support this notion. (a) The DNase experiments suggest that the observed genetic exchange is DNase resistant. (b) Because a natural transformation system has not been described B. thuringiensis, it is unlikely the genetic exchange observed resulted from transformation. c) The event is dependent on the presence of Tn916, other non-conjugal genetic information does not move in the absence of Tn916.

In matings where soil moisture was the variable, a range was chosen that would represent typical topsoil moistures from very dry to wet. Interestingly, the results show transfer frequencies were greater when moisture content was low (Fig. 1). Several explanations for this observation are possible. Because the test soil was sandy, it had a relatively low water holding capacity. Field capacity (33 Kpa water potential), thought to be near ideal moisture for a wide variety of microbial activity, was at 7.5% moisture for this soil. Thus the range of water contents (2-13%) used in this study centered on field capacity and likely, at each end of the range, represented conditions too dry or too wet for ideal microbial growth. One explanation for the increased conjugal activity might be that at low soil moistures, most of the liquid phase remains in close proximity to the particles. The microbes would be packed in this liquid phase. Alternatively, it is possible that the cells may have been more quickly adsorbed to soil particles at lower moistures. Frequencies reduced as moisture increased, which may be caused by a slower adsorption to soil particles or by an increased liquid phase. Alternately, increased moisture might result in lower oxygen tension; the resulting anaerobiosis would inhibit the Bacillus species.

The addition of nutrients to the soil caused an increase in the frequency of conjugation. Soil does not provide an abundance of nutrients in usable form for bacteria (Stotzky, 1974). Nutrient amendment might improve the physiological state of the donor and recipient cells which would increase conjugation. Although the amendment used herein (BHI) is artificial, this kind of amendment simulates conditions that would occur in nature, such as with plant matter or animal wastes in the soil.

Topsoil temperatures vary depending on time of day and season of the year. Transfer occurred over a mesophilic range where metabolic activity would be greatest with an maximum at 30^oC, which is near the optimum growth temperature for the Bacillus species used herein. Another report, however, showed that, in E. coli plasmid conjug ation, mating pair formation did not occur at temperatures below 24^oC, possibly due to an inability of the pili to attach to the recipient cell (Wamsley, 1976). Thus, with Tn916, formation of mating pairs may be restricted to a temperature range of 18^o to 35^oC. These are temperatures that one might reasonably expect to find during the summer in agronomic soils.

In soil, conjugative transposition was observed at pH values near neutrality. This is similar to observations using E. coli soil matings and conjugal plasmids. In E. coli, conjugation frequencies increased as pH neared neutral (Krasovsky and Stotzky, 1987). With Tn916 matings, conjugal activity drops off completely as pH becomes more acidic or alkaline. This may be due to a weakening of the physiological state of the cells or a destabilization of the mating pairs.

Finally, for Tn916 to be transferred in a natural population, it would need to be introduced into the soil microflora by a host organism such as E. faecalis. Matings presented here between E. faecalis and B. thuringiensis showed this to be possible at nearly the same frequency as seen in filter matings (Naglich and Andrews, 1988). Tetracycline resistance is common in isolates of E. faecalis from animal waste. Waste-containing sludge is often spread on fields as fertilizer; therefore, this may be a method for introducing Tn916 into the soil population. Moreover, at least with enterococcal isolates from non-fecal sources a substantial portion of the tetracycline-resistant enterococci contain Tn916-like sequences of DNA (Bentorcha et al., 1992; Clermont et al., 1990).

Although a wide range of each environmental factor was examined here, it is important to note that conjugation was detected over the range for each parameter that would be considered normal conditions in a natural environment.

Although this sterile system works well for the detection of Tn916 transfer, reports have shown that plasmid conjugation in soil can be adversely affected in a nonsterile system both in Escherichia coli (Krasovsky and Stotzky, 1987) and Bacillus (Elsas et al., 1987). However, when AN861 (Tn916) and AN142 were mated in a nonsterile sample of soil by the same procedure, Tn916 transfer was observed at frequencies very similar to those in sterile systems.

The results presented here indicate the ability of Tn916 to enter the bacterial population of the soil and, once there, move between soil bacteria in this population over a wide range of environmental conditions. Because Tn916 can mobilize nonconjugal DNA, it would not only transfer tetracycline resistance but could transfer many other genes as well. This type of activity could allow mobilization of DNA from genetically engineered microorganisms into the indigenous soil population.

REFERENCES

Bertram, J., M. Stratz, and P. Durre. (1991) Natural transfer of conjugative transposon Tn916 between Gram-positive and Gram-negative bacteria. J. Bacteriol. 173:443-448.

Bentorcha, F., D. Clermont, G. Decespedes, and T. Horaud. 1992. Natural occurrence of structures in oral streptococci and enterococci with DNA homology to Tn916. Antimicrob Agents Chemother 36:59-63.

Clermont, D., and T. Horaud. 1990. Identification of chromosomal antibiotic resistance genes in Streptococcus anginosus (S-Milleri). Antimicrob. Agents. Chemother. 34:1685-1690.

Clewell, D. B. 1990. Movable genetic elements and antibiotic resistance in enterococci. Eur. J. Clin. Microbiol. Infect. Dis. 9:90-102.

Elsas, J.P., Govaert, J.M. and Johannes, A.V. (1987). Transfer of plasmid pTF30 between bacilli in soil as influenced by bacterial population dynamics and soil conditions. Soil Biol. Biochem. 19: 639-647.

Franke, A.E., and Clewell, D.B. (1981). Evidence for a chromosome-borne resistance transposon (Tn916) in Streptococcus faecalis that is capable of "conjugal" transfer in the absence of a conjugative plasmid. J. Bacteriol. 145: 494-502.

Krasovsky, V.N., and Stotzky, G. (1987). Conjugation and genetic recombination in Escherichia coli in sterile and nonsterile soil. Soil Biol. Biochem. 19: 631-638.

Naglich J.G., and Andrews, R. A. (1988a). Introduction of the Streptococcus faecalis transposon Tn916 into Bacillus thuringiensis subsp. israelensis. Plasmid 19: 84-93.

Naglich J.G., and Andrews, R.A. (1988b). Tn916-dependent conjugal transfer of pC194 and pUB110 from Bacillus subtilus into Bacillus thuringiensis subsp. israelensis. Plasmid 19: 84-93.

Stotzky, G. (1974). In "Microbial Ecology" (A. I. Laskin and H. Lechevalier, eds.), pp 57-135. Chemical Rubber CO., Boca Raton, Florida.

Wamsley, R.H. (1976). Temperature dependence of mating pair formation in Escherichia coli. J. Bacteriol. 126: 222-224.

Haack Figure 1.

Haack Figure 2.

Haack Figure 3.

Haack Figure 4.