Global Spread of MCR-Producing Salmonella enterica Isolates

Colistin resistance in bacteria has become a significant threat to food safety and public health, and its development was mainly attributed to the plasmid-mediated mcr genes. This study aimed to determine the global prevalence and molecular characteristics of mcr-producing Salmonella enterica isolates. A total of 2279 mcr-producing Salmonella genomes were obtained from the public database, which were disseminated in 37 countries from five continents worldwide, including Asia, Europe, America, Australia, and Africa. Human samples (39.5%; 900/2279) were the predominant sources of mcr-producing Salmonella isolates, followed by foods (32.6%), animals (13.7%), and environment (4.4%). Furthermore, 80 Salmonella serotypes were identified, and Typhimurium and 1,4,[5],12:i:- were the predominant serotypes, accounting for 18.3% and 18.7%, respectively. Twenty mcr variants were identified, and the most common ones were mcr-9.1 (65.2%) and mcr-1.1 (24.4%). Carbapenems-resistance gene blaNDM-1 and tigecycline-resistance gene tet(X4) were identified in one isolate, respectively. Phylogenetic results indicated that mcr-producing Salmonella fell into nine lineages (Lineages I-IX), and Salmonella Typhimurium, 1,4,[5],12:i:- and 4,[5],12:i:- isolates from different countries were mixed in Lineages I, II and III, suggesting that international spread occurred. These findings underline further challenges for the spread of Salmonella-bearing mcr genes.


Introduction
Colistin was considered the last resort drug in the treatment for multidrug-resistant (MDR) and even carbapenem-resistant Gram-negative bacteria. Colistin-resistant bacteria have spread across Asia, Africa, Europe, North America, South America, and Oceania [1][2][3]. The colistin resistance in bacteria was previously reported to result from mutations in chromosomes [4]. Until 2015, the novel plasmid-borne colistin resistance gene mcr-1 was firstly reported in an Escherichia coli isolate [5]. MCR-1 is a phosphoethanolamine (pEtN) transferase that is able to modify the phosphate groups in lipid A [5], the catalysis of which could be offered mainly by mcr-1 from plasmids but also encoded by chromosome in the Enterobacteriaceae family [6].
Among countries with mcr-producing Salmonella enterica isolates, China was the most frequent one, and countries in Europe harbored a wide diversity of mcr variants [1]. The mcr gene was frequently identified in animal-and human-borne Salmonella enterica isolates [1]. Among more than 2600 Salmonella serotypes, Salmonella Typhimurium was the top serovar to carry the mcr genes [1]. Besides Salmonella, Escherichia coli, Klebsiella Pneumoniae, Klebsiella oxytoca, Cronobacter sakazakii, Kluyvera ascorbata, Shigella sonnei, Citrobacter freundii, Citrobacter braakii, Raoultella ornithinolytica, Proteus mirabilis, Aeromonas, Moraxella, and Enterobacter species were also found to be mcr-producing Gram-negative bacteria [3]. Furthermore, the mcr genes were widely distributed in various sample sources, including humans, animals, animal-origin foods, and environment [3].
Currently, a variety of mcr genes, including mcr-1 to -10, have been reported in bacteria [6]. Furthermore, mcr genes were also found to co-exist with carbapenem resistance

Phylogenomic Analysis of mcr-Producing Salmonella
To provide the evolutional characteristics of mcr-producing Salmonella, we compared 2145 genomes of different Salmonella serotypes from different countries. A total of 25,735 core SNPs were identified to construct a maximum likelihood tree (Figure 4). The evolutional strategy was based on serotypes. Currently, among 2145 isolates, the gene mcr (mcr-5.1) could be traced back to a Salmonella Typhimurium isolate in 1985, which was recovered from Bos taurus in Japan. It was found that Typhimurium and 1,4, Typhimurium isolates from the United Kingdom and Australia in Lineage III were divided into two sub-clades, respectively. It was found that most isolates of Lineage II were from China, but the base isolates were recovered from Germany, suggesting that mcr-producing Salmonella 1,4, [5],12:i:-isolates might be introduced into China from Germany. Similar results were also found in Lineage IV. Lineage IV was composed of Salmonella Saintpaul isolates from the USA, and their evolutionary path implied that a major introduction might occur in the USA from Germany and then nationwide spread. Indeed, the predominant serotype carrying mcr genes in Germany was Salmonella Paratyphi B var. d-tartrate+, which could be found in Lineage VI. We also found that isolates from Lineages I, II, V, VII, VIII, and IX were mostly isolated after 2013, but those from Lineages IV and VI were mainly isolated from 2002 to 2013. In addition, isolates from different sources shared a close genetic relationship, which suggested that clonal spread occurred. For example, most isolates of Lineage II were recovered from humans but also clustered with some isolates from animals. Similar results were also found in isolates of Lineages I and III.

Discussion
Colistin had been considered a therapeutic drug and feed additive for animals since the early 1980s and was approved for clinical treatment of human beings in 2017 in China [10,11]. In a previous study, mcr-1-producing E. coli isolates were able to transfer from animals to humans through the chicken production chain [12]. E. coli, E. cloacae, Salmonella spp., and K. pneumonia were reported to be the dominant mcr-1-producing bacteria members [1,5,11,13]. China's government has banned colistin as a feed additive to prevent colistin resistance in clinics, and then the drug production decreased markedly from 27,170 tons in 2015 to 2,497 tons in 2018, and the corresponding sale also decreased from USD 71.5 million in 2015 to USD 8.0 million in 2018 in China [14]. Furthermore, the median relative abundance of mcr-1 per 16S RNA in E. coli from animals and humans decreased significantly from 0.0009 in 2017 to 0.0002 in 2018 [14]. Therefore, the ban on colistin in animal farming causes a significant decrease in mcr-1 of E. coli isolates from animals and humans.
The emergence of plasmid-borne mcr genes has attracted significant attention worldwide [3,5]. In a previous study about the prevalence of mcr-producing bacterial isolates, it was found that a significant increase occurred in 2009 [3]. The chicken-borne mcr-1-producing E. coli isolates increased from 5.2% in 2009 to 30.0% in 2014 [15]. In a retrospective survey from China, the earliest chicken-borne mcr-1-producing E. coli isolate was found in the 1980s, which corresponded to the start use of colistin in animal husbandry as a feed additive [15]. In this study, the earliest emergence of mcr was identified from a Salmonella Typhimurium isolate from Bos taurus in 1985 (Japan, mcr-5.1), and the earliest mcr gene from humans was identified in a Salmonella 1,4, [5],12:i:-isolate in 2001 (the USA, mcr-9.1). Up to 1 January 2022, a total of 2279 mcr-positive Salmonella genomes were available from the NCBI database and were widely spread in 37 countries.

Discussion
Colistin had been considered a therapeutic drug and feed additive for animals since the early 1980s and was approved for clinical treatment of human beings in 2017 in China [10,11]. In a previous study, mcr-1-producing E. coli isolates were able to transfer from animals to humans through the chicken production chain [12]. E. coli, E. cloacae, Salmonella spp., and K. pneumonia were reported to be the dominant mcr-1-producing bacteria members [1,5,11,13]. China's government has banned colistin as a feed additive to prevent colistin resistance in clinics, and then the drug production decreased markedly from 27,170 tons in 2015 to 2,497 tons in 2018, and the corresponding sale also decreased from USD 71.5 million in 2015 to USD 8.0 million in 2018 in China [14]. Furthermore, the median relative abundance of mcr-1 per 16S RNA in E. coli from animals and humans decreased significantly from 0.0009 in 2017 to 0.0002 in 2018 [14]. Therefore, the ban on colistin in animal farming causes a significant decrease in mcr-1 of E. coli isolates from animals and humans.
The emergence of plasmid-borne mcr genes has attracted significant attention worldwide [3,5]. In a previous study about the prevalence of mcr-producing bacterial isolates, it was found that a significant increase occurred in 2009 [3]. The chicken-borne mcr-1producing E. coli isolates increased from 5.2% in 2009 to 30.0% in 2014 [15]. In a retrospective survey from China, the earliest chicken-borne mcr-1-producing E. coli isolate was found in the 1980s, which corresponded to the start use of colistin in animal husbandry as a feed additive [15]. In this study, the earliest emergence of mcr was identified from a Salmonella Typhimurium isolate from Bos taurus in 1985 (Japan, mcr-5.1), and the earliest mcr gene from humans was identified in a Salmonella 1,4, [5],12:i:-isolate in 2001 (the USA, mcr-9.1). Up to 1 January 2022, a total of 2279 mcr-positive Salmonella genomes were available from the NCBI database and were widely spread in 37 countries. Therefore, it was necessary to monitor the prevalence of mcr-1 in Salmonella to evaluate its burden on human health.
Here are some limitations of this study. First, the numerous genomes used in this study were associated with the development of sequencing technology. The genome information used was hard to represent the real prevalence of mcr-producing isolates. Second, the majority of the mcr-producing Salmonella genomes originated from the United States, China, the United Kingdom, and Germany, but data were limited from other countries. Therefore, the above limitations would bias the deduced results.
In conclusion, our study highlighted that mcr-producing Salmonella isolates have spread in five continents (Asia, Europe, America, Australia, and Africa) globally. Typhimurium and 1,4, [5],12:i:-were the predominant serotypes carrying mcr genes in Salmonella, and mcr-9.1 and mcr-1.1 were the predominant variants. The spread of such various mcr variants is of great concern for food safety and public health, and it is urgent to enhance surveillance and control the spread of mcr genes among Salmonella.

Salmonella Genomes Collected
We searched two important words, "mcr" and "Salmonella", in the public NCBI database on 1 January 2022. A total of 2279 Salmonella isolates were positive for mcr genes, and the detailed information of isolates, including years, countries, host, and serotypes, were then downloaded. However, 2145 out of 2279 genomes were available because the rest had not been released. These genomes were used for phylogenetic analysis to explore the implied spread of MCR-producing Salmonella worldwide.

The Identification of mcr Variants and Serotypes
The genomes used were further to identify the variant types of mcr genes through Res-Finder (https://cge.cbs.dtu.dk/services/ResFinder/, accessed on 1 June 2022). Salmonella serotypes were available from the uploaded information, and the serotypes that were not submitted by the uploader were predicted by SeqSero 1.2 (https://cge.cbs.dtu.dk/services/ SeqSero/, accessed on 1 June 2022).

Phylogenetic Analysis
A total of 2145 Salmonella genomes from different countries were used for phylogenetic analysis. Single-nucleotide polymorphisms (SNPs) were extracted using Snippy (https://github.com/tseemann/snippy, accessed on 1 June 2022) to generate core genomic alignment. Gubbins [17] was then used to identify and remove recombination regions using an algorithm that iteratively identifies loci containing elevated densities of base substitutions, and then resulting pairwise SNP differences were calculated. The core SNP alignment was used to generate a maximum-likelihood phylogeny using RAxML v8.1.23 (Karlsruhe, Germany) [18] with the GTR nucleotide substitution model. The display, annotation, and management of phylogenetic trees were performed by the ITOL tool [19]. The antibiotic resistance genes in all genomes were identified by ResFinder 4.1 (Lyngby, Denmark) [20].

Data Analysis
The bar and circle charts used to analyze the prevalence of mcr serotypes were generated with GraphPad Prism 7 software (San Diego, California, CA, USA). The world map marked with the distribution of mcr-positive Salmonella was drawn by Inkscape 0.92 (New York, NY, USA).