Plasmidic qnrA3 enhances Escherichia coli fitness in absence of antibiotic exposure

Abstract

The widespread presence of plasmid-mediated quinolone resistance determinants, particularly qnr genes, has become a current issue. By protecting DNA-gyrase from quinolones, Qnr proteins confer a low level quinolone resistance that is not sufficient to explain their emergence. Since Qnr proteins were hypothesized to act as DNA-binding protein regulators, qnr genes could have emerged by providing a selective advantage other than antibiotic resistance. We investigated host fitness of Escherichia coli isogenic strains after acquisition of the qnrA3 gene, inserted either alone onto a small plasmid (pBR322), or harbored on a large conjugative native plasmid, pHe96(qnrA3) found in a clinical isolate. The isogenic strains were derived from the susceptible E. coli CFT073, a virulent B2 group strain known to infect bladder and kidneys in a mouse model of pyelonephritis. In vitro experiments included growth analysis by automatic spectrophotometry and flow cytometry, and competitions with CFU enumeration. In vivo experiments included infection with each strain and pairwise competitions in absence of antimicrobial exposure. As controls for our experiments we used mutations known to reduce fitness (rpsL K42N mutation) or to enhance fitness (tetA deletion in pBR322). E. coli CFT073 transformed with pBRAM(PBR322-qnrA3) had significantly higher maximal OD than E. coli CFT073 transformed with pBR322 or pBR322ΔtetA, and in vivo competitions were more often won by the qnrA3 carrying strain (24 victories vs. 9 loss among 42 competitions, p = 0.001). In contrast, when pHe96(qnrA3) was introduced by conjugation in E. coli CFT073, it exerted a fitness cost shown by an impaired growth observed in vitro and in vivo and a majority of lost competitions (33/35, p<0.0001). In conclusion, qnrA3 acquisition enhanced bacterial fitness, which may explain qnr emergence and suggests a regulation role of qnr. However, fitness was reduced when qnrA3 was inserted onto multidrug-resistant plasmids and this can slow down its dissemination without antibiotic exposure.

Figure 1. Cell size measured by flow cytometry for the strains of the two isogenic systems.

Each graphic compares size measurement for bacterial cell populations consisting of the reference susceptible strain E. coli CFT073 (in black) and of the tested strain : E. coli CFT073(pBR322) (qnr-negative) in red (part A), E. coli CFT073(pBRAM1) (qnrA3-positive) in yellow (part B), E. coli CFT073(pBRAM2) (qnrA3-positive) in pink (part C), E. coli CFT073-Sm^R (qnr-negative) in blue (part D), E. coli CFT073-Sm^R(pHe96) (qnrA3-positive) in green (part E) and its variant R42 in purple (part F). Measurements were made separately, not in a competitive assay.

Figure 2. Enhanced fitness observed in competitive infections for E. coli CFT073 after qnrA3 acquisition onto pBR322.

Each symbol represents the bacterial ratio (number of CFU for the qnr-positive strain/number of CFU for the qnr-negative isogenic strain) measured in organs (blue diamond = kidneys, red circle = bladder) collected five and ten days after inoculation of a 1∶1 mix of the two strains. When the ratio was equal to 1+/−0.2, it was considered as tie. Part A: competitions experiments opposing E. coli CFT073(pBR322) (qnr−, tetA+) and E. coli CFT073(pBRAM2) (qnrA3+, tetA−). Fifteen mice were inoculated, 15 bladders and 10 pairs of kidneys were efficiently infected. Competition was won 22 times by E. coli CFT073(pBRAM2) (qnrA3+, tetA−), was lost 2 times, and one was tie (p<0.0001). Part B: competitions opposing E. coli CFT073(pBRΔtetA) (qnr−, tet−) and E. coli CFT073(pBRAM2) (qnrA3+, tet−). Twenty-three mice were inoculated, 20 bladders and 17 pairs of kidneys were efficiently infected. Competition was won 24 times by E. coli CFT073(pBRAM2) (qnrA3+, tetA−), was lost 9 times, and 6 was tie (p<0.0001).

Figure 3. Single strain urinary tract infections with the isogenic system of E. coli CFT073-SmR harboring or not the multidrug resistance plasmid pHe96 (qnrA3).

Part A: Bacterial density (log₁₀CFU/g of tissue) in bladders collected two, five and ten days after inoculation by E. coli CFT073-Sm^R (qnr−, purple plot), E. coli CFT073-Sm^R(pHe96) (qnrA3+, light green plot) and E. coli CFT073-Sm^R(pHe96) R42 variant (qnrA3+, dark green plot). At day 2, bacterial density was 6.79+/−0.35, 4.95+/−0.8 (p<0.0001), and 4.9+/−0.65 (p<0.0001), respectively; at day 5, it was 4.61+/−0.25, 3.24+/−0.3 (p<0.0001) and 3.79+/−0.5 (p = 0.03), respectively; and at day 10, 5.47+/−0.8, 2.91+/−0.6 (p = 0.0004) and 2.85+/−0.5 (p = 0.001), respectively. At least 10 mice were studied per group. Part B : Bacterial density (log₁₀CFU/g of tissue) in kidneys collected two, five and ten days after inoculation by E. coli CFT073-Sm^R (qnr−, purple plot), E. coli CFT073-Sm^R(pHe96) (qnrA3+, light green plot) and E. coli CFT073-Sm^R(pHe96) R42 variant (qnrA3+, dark green plot). Results were respectively 4.29+/−0.53, 3.14+/−0.64 (p = 0.005) and 3.53+/−0.46 (p = 0.053) at day 2; 4.29+/−0.61, 2.94+/−0.6 (p = 0.02), and 4.23+/−0.43 (p = 0.06) at day 5; and 4.81+/−0.6, 3.26+/0.63 (p = 0.012 and 3.3+/−1.59 (p = 0.04) at day 10. At least 10 mice were studied per group.

Figure 4. Reduced fitness observed after pHe96 acquisition in competitive infections in absence of antimicrobial exposure.

Each symbol represents the ratio (number of CFU for the qnr-positive strain/number of CFU for the qnr-negative isogenic strain) in organs (blue diamond = kidneys, red circle = bladder), collected five and ten days after inoculation of a 1∶1 mix of the two strains. Part A: competition experiments opposing E. coli CFT073-Sm^R (qnr−) and E. coli CFT073-Sm^R(pHe96) (qnrA3+). Twenty mice were inoculated, 19 bladders and 16 pairs of kidneys were efficiently infected. Competition was lost 33 times by E. coli CFT073-Sm^R(pHe96) (qnrA3+), and won only 2 times (p<0.0001). Part B: competition experiments opposing E. coli CFT073-Sm^R (qnr−) and E. coli CFT073-Sm^R(pHe96) variant “R42” (qnrA3+). The R42 variant was selected from kidneys that were infected by E. coli CFT073-Sm^R(pHe96). Twenty mice were inoculated, 18 bladders and 16 pairs of kidneys were efficiently infected. Competition was lost 33 times by the qnrA3-positive strain with only one won (p<0.0001).