Invasion of E. coli biofilms by antibiotic resistance plasmids

Abstract

In spite of the contribution of plasmids to the spread of antibiotic resistance in human pathogens, little is known about the transferability of various drug resistance plasmids in bacterial biofilms. The goal of this study was to compare the efficiency of transfer of 19 multidrug resistance plasmids into Escherichia coli recipient biofilms and determine the effects of biofilm age, biofilm-donor exposure time, and donor-to-biofilm attachment on this process. An E. coli recipient biofilm was exposed separately to 19 E. coli donors, each with a different plasmid, and transconjugants were determined by plate counting. With few exceptions, plasmids that transferred well in a liquid environment also showed the highest transferability in biofilms. The difference in transfer frequency between the most and least transferable plasmid was almost a million-fold. The 'invasibility' of the biofilm by plasmids, or the proportion of biofilm cells that acquired plasmids within a few hours, depended not only on the type of plasmid, but also on the time of biofilm exposure to the donor and on the ability of the plasmid donor to attach to the biofilm, yet not on biofilm age. The efficiency of donor strain attachment to the biofilm was not affected by the presence of plasmids. The most invasive plasmid was pHH2-227, which based on genome sequence analysis is a hybrid between IncU-like and IncW plasmids. The wide range in transferability in an E. coli biofilm among plasmids needs to be taken into account in our fight against the spread of drug resistance.

Fig. 1

Plasmid rankings based on efficiency of plasmid transfer (T/RxD) between E. coli cells in liquids (A) and biofilms (B). Data represent the number of donors (D), recipients (R) and (T/RxD) after 4 h. Statistically significant groups in the transconjugants (T) (CFU/ml), and Log₁₀ biofilm experiment are marked as black squares.

Fig. 2

Batch culture biofilms of E. coli MG1655 csrA Nal. A) Microscope slides with 24 h-, 48 h-, and 72 h-old biofilms. B) Quantitative analysis of biofilm features over a 72 h time period. CLSM images from three areas of the 24, 48 and 72 h-old biofilms were analyzed using the COMSTAT2 image analysis program as described in the Methods. At least ten independent stacks were analyzed per area per day. Compared parameters are: biomass, average thickness, roughness, maximal thickness and surface to biovolume ratio in zone I (air-liquid interface), zone II (central part), and zone III (bottom part of the slide).

Fig. 3

Effect of biofilm age on plasmid transfer in E. coli biofilms. Number of donors (D), recipients (R) and transconjugants (T) as Log₁₀ (CFU/ml), and the efficiency of plasmid transfer Log₁₀ (T/DxR) in 24 h-, 48 h- and 72 h-old biofilms.

Fig. 4

Effect of infection time on invasion of plasmids RP4 and pB10 in E. coli biofilms. Efficiency of plasmid transfer is presented as a box plot of Log₁₀ (T/(DxR)) values. Error bars for D and R show standard errors. Statistical differences (P≤0.05) between transfer efficiencies and numbers of attached donors at different time points are as follows: *, all values that differ from the one at 4 h; letters a-d, data points with the same letter are statistically different from each other.