Bacterial conjugation in the cytoplasm of mouse cells

Abstract

Intracellular pathogenic organisms such as salmonellae and shigellae are able to evade the effects of many antibiotics because the drugs are not able to penetrate the plasma membrane. In addition, these bacteria may be able to transfer genes within cells while protected from the action of drugs. The primary mode by which virulence and antibiotic resistance genes are spread is bacterial conjugation. Salmonellae have been shown to be competent for conjugation in the vacuoles of cultured mammalian cells. We now show that the conjugation machinery is also functional in the mammalian cytosol. Specially constructed Escherichia coli strains expressing Shigella flexneri plasmid and chromosomal virulence factors for escape from vacuoles and synthesizing the invasin protein from Yersinia pseudotuberculosis to enhance cellular entry were able to enter 3T3 cells and escape from the phagocytic vacuole. One bacterial strain (the donor) of each pair to be introduced sequentially into mammalian cells had a conjugative plasmid. We found that this plasmid could be transferred at high frequency. Conjugation in the cytoplasm of cells may well be a general phenomenon.

FIG. 1.

Effect of invasin on infection of 3T3 cells. (A) Scheme of infection protocol. Briefly, confluent monolayers of mammalian cells were infected with bacteria at various MOIs (MOI of approximately 5 to 1,000, depending on the cell type) (see Materials and Methods for details) for 30 min to 3 h to allow the bacteria to invade the host cells. The infected cells were then incubated in medium containing Gm to kill extracellular bacteria. At the end of the procedure, the cells were washed and then lysed to release the intracellular bacteria, which were then counted by CFU. (B) Cellular entry into 3T3 cells of EcSf2a-3 and DH10B with and without invasin. Bacterial entry into 3T3 cells was plotted against the actual MOIs used (determined a posteriori) (see Materials and Methods for details).

FIG. 2.

Phagosomal escape. (A) Scheme of infection protocol. Host cell monolayers were infected with bacteria as described in the legend of Fig. 1A. Following treatment with Gm to kill extracellular bacteria, Cq was added to the medium in half of the wells to kill bacteria that were retained in the phagosomes. In one set of experiments, the host cell monolayers were pretreated with Cq prior to bacterial infection in order to verify that the drug can be washed out effectively so as not to interfere with subsequent bacterial infections. The infected cells were washed, lysed, and plated as described above to determine the percentage of bacteria that were able to escape from phagosomes. (B) Escape of invasin-encoding EcSf2a-3 and DH10B from vacuoles of 3T3 and 143B cells. Escape ability refers to whether or not the bacterial strain is capable of lysing the phagosomal membrane to gain access to the host cell cytoplasm. MOIs were determined a posteriori (see Materials and Methods for details) and were determined once for a set of experiments run at the same time (i.e., on the same day); numbers of experiments are indicated in parentheses (e.g., MOIs with n = 2 indicate that the data were gathered from experiments run on two different days). Incubation times are given for treatment in Medium+Gm and Medium+Gm+Cq. The numbers of bacteria per cell were calculated as described above (numbers in parentheses indicate the total numbers of independent experiments performed, with the standard errors of the means [SEM]). The percentage of phagosomal escape is given by %Cq resistance and was calculated as follows: (number of bacteria recovered with Cq treatment/number of bacteria recovered in the absence of Cq treatment) × 100. The percentages of Cq resistance were calculated for each experiment and then averaged for the reported results. Note that in these experiments, all bacterial strains contain the invasin plasmid pGB2Ωinv. In addition, the bacteria also contained a second plasmid, encoding an antibiotic marker to select for cytosolic bacteria. For these experiments, RK2 or pEGFP-N2 was used. (C) Graphical presentation of escape data for experiments in which host cells were not pretreated with Cq prior to bacterial infection (first eight rows of data in B). Numbers of experiments are indicated in parentheses; error bars indicate SEM.

FIG. 3.

Intracellular conjugation. (A) Scheme of the intracellular conjugation protocol. Host cell monolayers were first infected with donor bacteria. The donors were allowed to invade for 30 min before treatment with Gm to eliminate extracellular bacteria, followed by treatment with Gm and Cq to eliminate bacteria remaining in the phagosomes. The monolayers were then infected with recipient bacteria. After the second treatment with Gm and Cq, the cells were incubated for various lengths of time (1 h to overnight) to allow conjugation to occur between the cytosolic donors and recipients. In some cases, the monolayers were treated with Gm and Cq a third time prior to lysis and spreading onto plates with Km plus Sp to determine the number of cytosolic donors, onto plates with Cm plus Sp to determine the number of cytosolic recipients, and onto plates with Km plus Cm to determine the number of transconjugants (TC). Asterisks indicate steps in the protocol where media and/or washes were sampled to check for the presence and amount of relevant bacteria. (B) Tabular presentation of conjugation data. Matings are depicted as donor × recipient. Conjugation time represents the length of time between the second Cq treatment and the third Cq treatment or lysis. MOIs and numbers of experiments are as described in the legend of Fig. 2. The numbers of transconjugants per donor were calculated for each experiment and then averaged for the reported results. (C) Graphical presentation of conjugation data, represented as transconjugants per donor for EcSf2a-3(pGB2Ωinv, RK2), compared to data for the DH10B(pGB2Ωinv, RK2) control for phagocytic escape. Numbers of experiments are indicated in parentheses; error bars indicate SEM.

FIG. 4.

Assay for postlysis conjugation. (A) Scheme of infection protocol. Donor and recipient bacteria were infected into separate monolayers and allowed to invade the host cells. The infected cells were then treated with Gm and Gm then Cq, followed by lysis, as was done earlier. The donor-infected lysates were mixed on ice with the recipient-infected lysates and then plated with selection for donors, recipients, and transconjugants in order to determine the background conjugation rate during the processing steps after cell lysis. Asterisks are as described in the legend of Fig. 3. (B) Tabular presentation of the postlysis conjugation data. Matings are depicted as donor × recipient. MOI and numbers of experiments are as shown in Fig. 2. The numbers of transconjugants per donor were calculated for each experiment and then averaged for the reported results.