Fig 1.
Anticipated exploitation of pCURE as a probiotic treatment for at-risk individuals.
Before treatment, plasmids carrying antibiotic resistance genes are shown as purple circles. After treatment, target plasmids are replaced by pCURE (green circle) which could be engineered to later “self-destruct”. Note that not all gut bacteria carry resistance plasmids but pCURE can enter all Enterobacteriaceae as well as other bacteria. Reduced resistance levels in the gut decrease the chance of treatment failure when infections elsewhere in the body (eg lungs or urinary tract) arise from gut bacteria.
Fig 2.
Map of RK2 showing region replaced by the anti-F cassette and showing the targets of the anti-F functions.
The other functions marked are: oriV, the vegetative replication origin; oriT, the transfer origin; tra and trb regions encoding proteins for DNA processing and mating bridge formation during transfer; trfA, encoding the replication protein that activates oriV; ccr/par, the central control region that regulates transcription of RK2 backbone genes and also encodes active partitioning functions; psk/mrs, encoding addiction and multimer resolution functions. Mobile elements Tn1 and IS21 are associated with ampicillin and kanamycin resistance genes. Recombineering was used to replace aphA and IS21 in RK2 with the anti-F cassette that blocks the repFIA, repFIB and repFII replicons and neutralises the effect of the flmABC and letAB addiction loci [12].
Fig 3.
Conjugative IncP-1 derivatives and their effectiveness as vehicles for plasmid curing (panel A). The anti-F cassette inserted into RK2 caused limited displacement, as indicated by the blue/white X-gal+IPTG phenotype (compare RK2Δaph and pCURE-F-RK2 in panel B), whereas when inserted into pUB307 (23), it causes very efficient F’ plasmid loss and no trace of blue colour (compare pUB307Δaph with pCURE-F-307 in panel B). Introducing the pUB307 deletion (bases 5464..5467 to 12045..12047) into RK2 potentiated curing while the deletions ΔklcAB and ΔklcABC (bases 4340–11669) did not. Reintroducing the region with i10 (bases 11749 to 12048) into pUB307 destroyed the potentiation, showing that this region includes the critical sequence.
Table 1.
Curing by key conjugative anti-F and anti-K plasmids constructed in this study.
Fig 4.
Mini-RK2 plasmids with the anti-F cassette and ability to displace F’prolac.
Plasmid pCT549 consists of two segments from RK2: oriV to trbB’ and korA to kfrC’ (apostrophe indicates a truncated gene). TrfA acts at oriV: monomers act positively through iterons 5–9 and dimers act negatively through all the iterons. Due to its construction oriV of pCT549 excluded iteron 10 (i10) and with addition of the anti-F cassette efficient curing was seen. A derivative with iterons 1 to 10 gave a lower relative copy number (data and statistics in S2 Fig, S1 Raw Images and S1 Table) and efficient curing was lost. Deletion of i1 restored efficient curing and relative copy number rose approximately 2-fold. Deleting up to korB had no effect but deleting past it destroyed curing. Reinserting korB restored efficient curing.
Fig 5.
Unselected invasion assay to monitor displacement of target plasmids by pCURE plasmids or negative controls in bacteria introduced at a donor:recipient ratio of approximately 1:1000.
A. Donor bacteria (HB101 with pCURE-F plasmids or negative controls) were monitored by streptomycin resistance, while target bacteria (JM109 initially with F’prolac) were monitored by nalidixic acid resistance. Presence of F’prolac was monitored on M9 Minimal medium without proline. The data is shown in the Tables in S1 Text. B. Donor bacteria (HB101 with pCURE-K plasmids or negative controls) were monitored by streptomycin resistance, while target bacteria (J53rif initially with pCT::aph) were monitored by rifampicin resistance. Presence of pCT::aph was monitored by kanamycin resistance. The data is shown in the Tables in S2 Text. Spread of pCT::aph into donor bacteria was detected by selection of kanamycin and streptomycin resistance as shown in the Figure in S3 Fig.
Fig 6.
Testing ability of pCURE-F-307 to displace antibiotic resistance plasmids in the mouse gut.
A. RifR mouse-derived E. coli AL1 carrying pCT::aph was fed to six mice days 1–3. Three mice (group 1) also received kanamycin days 2–4. Plasmid transfer to endogenous RifS E. coli was only observed in group 1. The data for this experiment is shown in the Tables in S3 Text. B. Baseline total endogenous KanSTetS E. coli were detected before introduction of E. coli AL1 (pCT::aph) days 3–5 and kanamycin days 4–6 before monitoring plasmid carriage until day 28. RifS E. coli cfu were determined by subtracting RifR cfu from total E.coli cfu. The data for this experiment is shown in the Tables in S4 Text. C. Mice treated as in B received E. coli Nissle 1917 carrying TetR curing plasmid, pCURE-K-307 days 10–12 and tetracycline days 11–13 resulting in the appearance of TetR E. coli including RifR TetR indicating transfer to target bacteria. pCURE-K-307 had disappeared by day 20 but about 10% of the E. coli were RifR indicating displacement of pCT::aph rather than loss of the strain. Kanamycin was given days 25–26 but no KanR bacteria re-appeared and PCR screening of faeces samples at the end of the experiment proved negative. Carriage of pCT::aph was determined by KanR while carriage of pCURE-K-307 was determined by TetR. The data for this experiment is shown in the Tables in S4 Text.