Core Genome MLST for Source Attribution of Campylobacter coli

Abstract

Campylobacter species are among the leading foodborne bacterial agents of human diarrheal illness. The majority of campylobacteriosis has been attributed to Campylobacter jejuni (85% or more), followed by Campylobacter coli (5–10%). The distribution of C. jejuni and C. coli varies by host organism, indicating that the contribution to human infection may differ between isolation sources. To address the relative contribution of each source to C. coli infections in humans, core genome multilocus sequence type with a 200-allele difference scheme (cgMLST₂₀₀) was used to determine cgMLST type for 3,432 C. coli isolated from food animals (n = 2,613), retail poultry meats (n = 389), human clinical settings (n = 285), and environmental sources (n = 145). Source attribution was determined by analyzing the core genome with a minimal multilocus distance methodology (MMD). Using MMD, a higher proportion of the clinical C. coli population was attributed to poultry (49.6%) and environmental (20.9%) sources than from cattle (9.8%) and swine (3.2%). Within the population of C. coli clinical isolates, 70% of the isolates that were attributed to non-cecal retail poultry, dairy cattle, beef cattle and environmental waters came from two cgMLST₂₀₀ groups from each source. The most common antibiotic resistance genes among all C. coli were tetO (65.6%), bla_OXA–193 (54.2%), aph(3′)-IIIa (23.5%), and aadE-Cc (20.1%). Of the antibiotic resistance determinants, only one gene was isolated from a single source: bla_OXA–61 was only isolated from retail poultry. Within cgMLST₂₀₀ groups, 17/17 cgMLST₂₀₀-435 and 89/92 cgMLST₂₀₀-707 isolates encoded for aph(3’)-VIIa and 16/16 cgMLST₂₀₀-319 harbored aph(2’)-If genes. Distribution of bla_OXA alleles showed 49/50 cgMLST₂₀₀-5 isolates contained bla_OXA–498 while bla_OXA–460 was present in 37/38 cgMLST₂₀₀-650 isolates. The cgMLST₂₀₀-514 group revealed both ant(6)-Ia and sat4 resistance genes in 23/23 and 22/23 isolates, respectively. Also, cgMLST₂₀₀-266 and cgMLST₂₀₀-84 had GyrAT86I mutation with 16/16 (100%) and 14/15 (93.3%), respectively. These findings illustrate how cgMLST and MMD methods can be used to evaluate the relative contribution of known sources of C. coli to the human burden of campylobacteriosis and how cgMLST typing can be used as an indicator of antimicrobial resistance in C. coli.