Evaluation of a novel and rapid screening method for the detection of contaminated colistin-resistant bacteria in food

Background: The dissemination of colistin-resistant bacteria carrying the colistin-resistant mobile gene, mcr-1 threatens medical care worldwide. In particular, contamination of food with colistin-resistant bacteria accelerates the community dissemination of colistin-resistant bacteria. Therefore, monitoring of colistin-resistant bacteria in food is important for controlling resistant bacteria. Unfortunately, the conventional culture methods for detecting colistin-resistant bacteria are not practical for monitoring food saftey. Therefore, development of a simple and rapid method to detect food contamination with colistin-resistant bacteria is desirable as an effective means for preventing the dissemination of resistant bacteria, particularly colistin-resistant bacteria. Findings: We developed a simple and rapid method for detecting Escherichia coli harboring the mcr-1 colistin resistance gene using a high-speed real-time polymerase chain reaction (PCR). The entire procedure, from sample processing to nals results, was performed within one hour. The practical utility of this method was veried by analyzing 27 retail meat samples for the presence of colistin-resistant bacteria. The results of the developed method were in agreement with the results of culturing colistin-resistant E. coli from the meat samples, demonstrating its ecacy and usefulness. Conclusions: A simple and rapid real-time PCR-based screening method was developed for detecting E. coli harboring mcr-1 in food samples. The practical utility of the procedure was conrmed using retail meat samples, indicating its potential as a convenient and rapid method to detect bacterial contamination of food items, especially in developing communities.


Introduction
Colistin is recognized as one of the few antibiotics available for the treatment of intractable infections caused by multidrug-resistant gram-negative bacteria (Giamarellou and Poulakou, 2009). Current studies have shown that colistin-resistant (CR) bacteria carrying the mobile colistin resistance gene mcr, which confers colistin resistance to most Enterobacteriaceae members, are widely disseminated, particularly in Asia (Shen et al., 2018;Yamamoto et al., 2018). Thus, the spread of CR bacteria in communities via livestock food is considered a potential risk factor because colistin is widely used in animal husbandry (Rhouma et al., 2016). Moreover, CR bacteria are often found in animals and animal-food Yamaguchi et al., 2018;Perez-Rodriguez and Mercanoglu Taban, 2019); thus, monitoring CR bacteria in animal-food is essential. However, the conventional culture method for detecting CR bacteria in food is laborious and time-consuming. Therefore, here, we report a simple and rapid detection method for Escherichia coli with mcr-1, a CR bacterium, using a high-speed real-time polymerase chain reaction (PCR) kit. We further veri ed its utility for detecting CR bacteria in retail meat samples. This novel detection method holds practical relevance as the entire procedure, involving sample processing to the nal result, is completed within one hour.

Materials And Methods
A total of 27 retail meat samples, such as pork and chicken, were collected from 10 markets (two supermarkets and eight local traditional markets) in Vietnam and ve supermarkets in Japan during November and December 2019. In Vietnam, none of the eight traditional markets maintained a refrigerator for meat preservation. In contrast, two supermarkets in Vietnam and all supermarkets in Japan had refrigerators for food storage. Each sample was collected from one meat type per market. Bacterial cultures and DNA extraction were performed on collection day; 10 grams of each meat sample were placed in a stomacher bag (AS ONE, Osaka, Japan) containing 90 mL of buffered peptone water.
Presence of E. coli and the colistin resistance gene mcr-1 in the DNA extracts were determined by realtime PCR using a mobile PCR device, PicoGene PCR1100 (Nippon Sheet Glass, Tokyo, Japan). PCR primers and probes for real-time PCR detection of E. coli 16S rRNA and mcr-1 were prepared as described previously (Table 1) (Daniels et al., 2019). Details regarding the real-time PCR, including PCR mixtures and thermal cycling conditions, are shown in Tables 2 and 3, respectively. The DNA extract of CR E. coli strain (E362) carrying mcr-1 was utilized as a positive control for the PCR. The entire PCR of 50 cycles was completed in just 21 min. Moreover, this real-time PCR device simultaneously measures uorescence at three different wavelengths for the same sample. Two uorescent dye-labeled probes (Integrated DNA Technologies, Singapore), Cy5 for E. coli 16S rRNA and FAM for mcr-1, were used for each sample. Figure  2 shows representative real-time PCR pro les of the samples.

Results And Discussion
The detection sensitivity of the method was assessed using pork meat samples spiked with an mcr-1positive E. coli strain culture. The lower limit of mcr-1-E. coli detection for the entire method, including DNA extraction to detection by real-time PCR, was 7 × 10 2 CFU/g, but quanti cation required 7 × 10 3 CFU/g or more due to the quantitative linearity (Fig. 3). In a validation study using retail meat samples, CR E. coli-like bacteria were detected using the culture method in eight out of 10 chicken and three out of seven pork samples purchased in Vietnam (Table 4).
The semi-quantitative levels of CR bacteria in these samples were about 10 3 -10 8 CFU/g (Table 4). All representative CR E. coli isolates from each sample were resistant to colistin (MIC: ≥4 µg/mL) and possessed mcr-1 but not mcr-2 to -5, except the H-E pork sample which had mcr-3 in addition to mcr-1, as determined by multiplex PCR. No samples from Japanese supermarkets were contaminated with CR bacteria.
All samples, except the H-E pork sample, that were positive via the culture-based method, were also positive by real-time PCR (Table 4). Some culture-negative samples, such as H-B pork, T-B chicken, T-B pork, and T-E chicken, were PCR-positive. Such contradictory results may be attributed to the features of the real-time PCR method and its ability to detect mcr-1 even in dead cells. In contrast, a pork sample from the H-E market showed CR E. coli colonies after culturing, but tested negative for mcr-1 by real-time PCR. The level of mcr-1 positive bacteria may be below the detection limit of the real-time PCR method.
The new method presented here detects the target gene and facilitates quantitative analysis. The results show the ratio of bacteria carrying mcr-1 to the number of E. coli cells, which may be both mcr-1-positive or -negative bacteria (Fig. 2). The detected quantitative mcr-1 levels were higher than CR E. coli-like bacterial levels determined via the culture-based method because it detected all mcr-1 regardless of bacterial species. The quantitative linear range detected via real-time PCR was between 10 3 and 10 6 CFU/g (Fig. 3). However, in some samples, the detected signal was below the quantitative linear range limit, yet they were considered to have positive results via real-time PCR.
Although the approach described in this study provided limited information regarding the degree of contamination, the developed method is reliable and practical for the rapid screening of contaminating CR bacteria in food.

Conclusions
In this study, a new, simple, and rapid screening method was developed for detecting CR E. coli in food samples. The developed method is advantageous because it is easy to perform, has a short processing time, and provides reliable results that are consistent with those obtained by traditional methods.

Declarations
Availability of data and materials The datasets used and/or analyzed in this study are available from the corresponding author on reasonable request.

Competing interests
This study was conducted using a mobile PCR device, PicoGene PCR1100, provided by Nippon Sheet Glass, Tokyo, Japan. There were no competing nancial interests or personal relationships that could have in uenced the work reported in this paper.

Funding
This work was supported by the Japan Society for the Promotion of Science KAKENHI (grant number 20H00561).
Author's contributions HTMV participated in the study design, sample collection, real-time PCR assay and drafted the manuscript. CA-K performed the real-time PCR assay. KT participated in the study design. RK performed PCR assays and microbial analysis. DTK, TNN, HTT, CDV, and PDN contributed to sample collection and bacterial cultures. YY contributed to the study design, supervised data collection, and drafted the manuscript. All authors read and approved the nal manuscript.

Figure 1
Outline of the screening protocol using mobile real-time PCR PicoGene® PCR1100. BPW, buffered peptone water.