Quantification of the mobility potential of antibiotic resistance genes through multiplexed ddPCR linkage analysis

Abstract

There is a clear need for global monitoring initiatives to evaluate the risks of antibiotic resistance genes (ARGs) towards human health. Therefore, not only ARG abundances within a given environment, but also their potential mobility, hence their ability to spread to human pathogenic bacteria needs to be quantified. We developed a novel, sequencing-independent method for assessing the linkage of an ARG to a mobile genetic element by statistical analysis of multiplexed droplet digital PCR (ddPCR) carried out on environmental DNA sheared into defined, short fragments. This allows quantifying the physical linkage between specific ARGs and mobile genetic elements, here demonstrated for the sulfonamide ARG sul1 and the Class 1 integron integrase gene intI1. The method's efficiency is demonstrated using mixtures of model DNA fragments with either linked and unlinked target genes: Linkage of the two target genes can be accurately quantified based on high correlation coefficients between observed and expected values (R2) as well as low mean absolute errors (MAE) for both target genes, sul1 (R2 = 0.9997, MAE = 0.71%, n = 24) and intI1 (R2 = 0.9991, MAE = 1.14%, n = 24). Furthermore, we demonstrate that adjusting the fragmentation length of DNA during shearing allows controlling rates of false positives and false negative detection of linkage. The presented method allows rapidly obtaining reliable results in a labor- and cost-efficient manner.

Keywords: antibiotic resistance; droplet digital PCR; mobile genetic elements; mobility; risk assessment.

Figure 1.

(A) Partitioning of PCR reaction into ca. 20 000 droplets of uniform size and volume, containing target and background DNA. (B) Data output from multiplex ddPCR experiments. Droplets form four clusters arranged orthogonally to each other. In grey: Empty droplets, negative for both targets (− −); Blue: positive droplets only for target A (+ −); Green: positive droplets only for target B (− +) and Orange: double-positive droplets (+ +) due to chance (no linkage) or due to physical linkage.

Figure 2.

Shearing of DNA into fragments of a specific sizes allows distinguishing between both targets (A and B) occurring on one and the same DNA fragment due to either linkage within an integron cassette or both genes appearing independently. Fragmentation size allows adjusting the rate of false positive and false negative detection of linkage.

Figure 3.

Data output from ddPCR experiments using the 100% (A) and 0% (B) linkage controls. N is the number of accepted droplets analyzed. Grey: Empty droplets (N_E), negative for both targets; Blue: droplets positive only for target A, here sul1 (N_A); Green: droplets positive only for target B, here intI1 (N_B); Orange: droplets double-positive for both targets including those due to physical linkage and due to chance (N_AB).

Figure 4.

Linear correlation between theoretical and experimentally calculated linkage percentage for the two target genes sul1 (A) and intI1 (B). Tested samples were prepared at eight defined volumetric mixtures of linked and unliked targets and measured with three technical replicates by multiplexed ddPCR (n = 24). MAE, mean absolute error in percentage (%).

Figure 5.

Abundance (A) and linkage (B) of the two target genes sul1 and intI1 in wastewater influent and effluent samples. Linkage was determined for unsheared as well as sheared DNA (average fragment size = 20 kbp).