In this study, we used PCR typing methods to assess the presence of tetracycline resistance determinants conferring ribosomal protection in waste lagoons and in groundwater underlying two swine farms. All eight classes of genes encoding this mechanism of resistance [tet(O), tet(Q), tet(W), tet(M), tetB(P), tet(S), tet(T), and otrA] were found in total DNA extracted from water of two lagoons. These determinants were found to be seeping into the underlying groundwater and could be detected as far as 250 m downstream from the lagoons. The identities and origin of these genes in groundwater were confirmed by PCR-denaturing gradient gel electrophoresis and sequence analyses. Tetracycline-resistant bacterial isolates from groundwater harbored the tet(M) gene, which was not predominant in the environmental samples and was identical to tet(M) from the lagoons. The presence of this gene in some typical soil inhabitants suggests that the vector of antibiotic resistance gene dissemination is not limited to strains of gastrointestinal origin carrying the gene but can be mobilized into the indigenous soil microbiota. This study demonstrated that tet genes occur in the environment as a direct result of agriculture and suggested that groundwater may be a potential source of antibiotic resistance in the food chain.

Phylogenetic analysis of tetracycline resistance genes encoding the ribosomal protection proteins (RPPs) revealed the monophyletic origin of these genes. The most deeply branching class, exemplified by tet and otrA, consisted of genes from the antibiotic-producing organisms Streptomyces rimosus and Streptomyces lividans. With a high degree of confidence, the corresponding genes of the other seven classes (Tet M, Tet S, Tet O, Tet W, Tet Q, Tet T, and TetB P) formed phylogenetically distinct separate clusters. Based on this phylogenetic analysis, a set of PCR primers for detection, retrieval, and sequence analysis of the corresponding gene fragments from a variety of bacterial and environmental sources was developed and characterized. A pair of degenerate primers targeted all tetracycline resistance genes encoding RPPs except otrA and tet, and seven other primer pairs were designed to target the specific classes. The primers were used to detect the circulation of these genes in the rumina of cows, in swine feed and feces, and in swine fecal streptococci. Classes Tet O and Tet W were found in the intestinal contents of both animals, while Tet M was confined to pigs and Tet Q was confined to the rumen. The tet(O) and tet(W) genes circulating in the microbiota of the rumen and the gastrointestinal tract of pigs were identical despite the differences in animal hosts and antibiotic use regimens. Swine fecal streptococci uniformly possessed the tet(O) gene, and 22% of them also carried tet(M). This population could be considered one of the main reservoirs of these two resistance genes in the pig gastrointestinal tract. All classes of RPPs except Tet T and TetB P were found in the commercial components of swine feed. This is the first demonstration of the applicability of molecular ecology techniques to estimation of the gene pool and the flux of antibiotic resistance genes in production animals.

Transfer of the shuttle vector pRRI207 mediated by the helper plasmid pRK2013 from Escherichia coli to Bacteroides spp. was possible in vitro and in vivo in the digestive tract of axenic mice associated with Bacteroides uniformis 1004 or Bacteroides vulgatus of human origin. In vivo, transfer frequencies were nearly identical for B. uniformis (2×10−7) and B. vulgatus (4×10−7) and the transconjugant strains of B. uniformis and B. vulgatus became established in the digestive tract of mice at densities ranging from 102–103 to 104 CFU g−1 of faeces, respectively. Transfer from E. coli to Bacteroides strains in gnotoxenic mice associated with human faecal flora (HFF) was not successful. Transconjugant‐like clones appeared among the HFF of gnotoxenic mice after they were inoculated with B. uniformis TBUA, a transconjugant strain of B. uniformis 1004 obtained from triparental mating and which harboured the shuttle vector. Hybridisation showed that the shuttle vector pRRI207 was not present in these clones, and it is suggested that they could result from the transfer of a conjugative transposon ERL contained in B. uniformis 1004. Moreover, clones believed to have lost the shuttle vector hybridised with a specific probe to B. thetaiotaomicron and therefore did not originate from B. uniformis TBUA.

Transfer of the shuttle vector pRRI207 mediated by the helper plasmid pRK2013 from Escherichia coli to Bacteroides spp. was possible in vitro and in vivo in the digestive tract of axenic mice associated with Bacteroides uniformis 1004 or Bacteroides vulgatus of human origin. In vivo, transfer frequencies were nearly identical for B. uniformis (2U10 3U ) and B. vulgatus (4U10 3U ) and the transconjugant strains of B. uniformis and B. vulgatus became established in the digestive tract of mice at densities ranging from 10 P^1 0 Q to 10 R CFU g 3I of faeces, respectively. Transfer from E. coli to Bacteroides strains in gnotoxenic mice associated with human faecal flora (HFF) was not successful. Transconjugant-like clones appeared among the HFF of gnotoxenic mice after they were inoculated with B. uniformis TBUA, a transconjugant strain of B. uniformis 1004 obtained from triparental mating and which harboured the shuttle vector. Hybridisation showed that the shuttle vector pRRI207 was not present in these clones, and it is suggested that they could result from the transfer of a conjugative transposon ERL contained in B. uniformis 1004. Moreover, clones believed to have lost the shuttle vector hybridised with a specific probe to B. thetaiotaomicron and therefore did not originate from B. uniformis TBUA. z