Intergeneric Transfer of Conjugative and Mobilizable Plasmids Harbored by Escherichia coli in the Gut of the Soil Microarthropod Folsomia candida (Collembola)

The gut of the soil microarthropod Folsomia candida provides a habitat for a high density of bacterial cells (T. Thimm, A. Hoffmann, H. Borkott, J. C. Munch, and C. C. Tebbe, Appl. Environ. Microbiol. 64:2660–2669, 1998). We investigated whether these gut bacteria act as recipients for plasmids from Escherichia coli. Filter mating with E. coli donor cells and collected feces of F. candida revealed that the broad-host-range conjugative plasmid pRP4-luc (pRP4 with a luciferase marker gene) transferred to fecal bacteria at estimated frequencies of 5.4 × 10⁻¹ transconjugants per donor. The mobilizable plasmid pSUP104-luc was transferred from the IncQ mobilizing strain E. coli S17-1 and less efficiently from the IncF1 mobilizing strain NM522 but not from the nonmobilizing strain HB101. When S17-1 donor strains were fed to F. candida, transconjugants of pRP4-luc and pSUP104-luc were isolated from feces. Additionally, the narrow-host-range plasmid pSUP202-luc was transferred to indigenous bacteria, which, however, could not maintain this plasmid. Inhibition experiments with nalidixic acid indicated that pRP4-luc plasmid transfer took place in the gut rather than in the feces. A remarkable diversity of transconjugants was isolated in this study: from a total of 264 transconjugants, 15 strains belonging to the alpha, beta, or gamma subclass of the class Proteobacteria were identified by DNA sequencing of the PCR-amplified 16S rRNA genes and substrate utilization assays (Biolog). Except for Alcaligenes faecalis, which was identified by the Biolog assay, none of the isolates was identical to reference strains from data banks. This study indicates the importance of the microarthropod gut for enhanced conjugative gene transfer in soil microbial communities.Gene transfer is a process by which bacterial populations substantially increase their rates of evolution and adaptation (12, 59). Particularly, plasmid-located genes, which are transferred by conjugation from donor to recipient cells, can disseminate rapidly between even phylogenetically different bacterial groups (17, 36, 41) and microbial communities in different spatial habitats (34, 71). Such microbial genetic networks should be considered in risk assessments of releases of genetically engineered microorganisms into the environment (22, 37, 43). The probability and rate of plasmid transfer from a donor to indigenous microorganisms in a given habitat are influenced by plasmid-borne genes which determine the type of transfer mechanism (self-transmissible or mobilizable) and the host range of autonomous plasmid replication. Additionally, specific physicochemical conditions, such as temperature, water potential, and the availability of energy (substrates) for donor and recipient cells, are important factors influencing gene transfer rates in terrestrial and aquatic environments (23, 53, 64).The spread of plasmid-borne genes is still extremely difficult to predict for terrestrial habitats, since a large variety of microhabitat conditions which are not well characterized exists. In bulk soil under laboratory conditions, conjugative gene transfer from recombinant bacterial donor strains to indigenous soil bacteria has been found only under specific selective conditions or on rare occasions (11, 20, 24, 27, 50, 61). Several studies failed to detect such transfer events, and it was concluded that heterogeneity and low densities of recipient cells, as well as a lack of substrates for microbial metabolism, prevented efficient plasmid transfer in bulk soil (19, 49, 54, 75). Plant exudates increased rates of gene transfer in soil (33, 48), and higher rates of gene transfer were found in rhizospheres than in bulk soil (50, 61). It was assumed that other microsites which favor gene transfer in terrestrial habitats are associated with soil invertebrates (74). However, to date little experimental evidence to prove this assumption is available.Intraspecies transconjugants of added Enterobacter cloacae donor and recipient cells could be isolated from microcosm experiments with the variegated cutworm, Peridroma saucia, and plant material (2). The investigators in that study concluded that gene transfer events happened, most likely, in the digestive tracts or in the feces of the insects. Another recent report demonstrated that a conjugative plasmid was transferred between fed Escherichia coli strains in the guts of Rhabditis nematodes (1). Earthworms mediated transport and enhanced plasmid transfer from added donor cells to added recipients and to indigenous bacteria in soil (14, 15). High rates of intraspecies plasmid transfer, comparable to those obtained in pure broth cultures, were detected with Bacillus thuringiensis in infected lepidopterous larvae (31).Microarthropods (collembolans and mites) are the most abundant invertebrate group in the majority of soils (5) but have not been recognized, so far, for their impact on microbial gene transfer. There are some indications that microarthropods harbor a large variety of microorganisms in their guts and thereby contribute to microbial biodiversity in terrestrial environments (7, 9, 57). In the accompanying paper, we have described the gut of Folsomia candida (Collembola) as a habitat and species-specific vector for microorganisms (67). The gut of this soil-dwelling insect, which has a volume of only several nanoliters, was found to be densely colonized, predominantly by rod-shaped bacterial cells. We were interested to know whether such bacterial cells act as recipients for plasmids and thereby promote gene transfer in microbial communities. F. candida feeds, under natural conditions, on bacteria (3), fungal mycelia (6, 66), and nematodes (35). Here, we report on the results of experiments in which plasmid-bearing E. coli strains were fed to F. candida in microcosms. Self-transferable plasmids, as well as mobilizable plasmids with different host ranges, and a nonmobilizable plasmid were included in this study in order to determine the specific capacities of these different classes of plasmids to spread into indigenous bacterial populations. For detection purposes, all plasmids were engineered by the insertion of the luciferase-encoding marker gene luc or lux (30, 47).