Emergence of Erythromycin Resistance Methyltransferases in Campylobacter coli Strains in France

Abstract

Antimicrobial resistance in campylobacters has been described worldwide. The emergence of multiresistant isolates, particularly among Campylobacter coli isolates, is concerning. New resistance mechanisms appear frequently, and DNA-sequence-based methods such as whole-genome sequencing (WGS) have become useful tools to monitor their emergence. The genomes of 51 multiresistant French Campylobacter sp. clinical strains from 2018 to 2019 were analyzed to identify associated resistance mechanisms. Analyses of erythromycin-resistant strains revealed 23S rRNA mutations among most of them and two different methyltransferases in 4 strains: Erm(B) and a novel methyltransferase, named Erm(N) here. The erm(B) gene was found in multidrug-resistant genomic islands, whereas erm(N) was inserted within CRISPR arrays of the CRISPR-cas9 operon. Moreover, using PCR screening in erythromycin-resistant strains from our collection, we show that erm(N) was already present in 3 French clinical strains 2 years before its first report in 2018 in Quebec, Canada. Bacterial transformations confirmed that the insertion of erm(N) into a CRISPR-cas9 operon can confer macrolide resistance. Campylobacter species are easily able to adapt to their environment and acquire new resistance mechanisms, and the emergence of methyltransferases in campylobacters in France is a matter of concern in the coming years.

Keywords: Campylobacter; bioinformatics; erythromycin; genomics; methyltransferase; resistance.

FIG 1

Whole-genome SNP phylogenetic tree based on C. coli reference strain NTICC13 and all C. jejuni (n = 9) and C. coli (n = 50) genomes analyzed. Analyses were performed from assembly fasta files for which species identification was assessed from 33 Campylobacter species: C. armoricus, C. avium, C. blaseri, C. canadensis, C. coli, C. concisus, C. corcagiensis, C. cuniculorum, C. curvus, C. fetus, C. geochelonis, C. gracilis, C. helveticus, C. hepaticus, C. hominis, C. hyointestinalis, C. iguaniorum, C. insulaenigrae, C. jejuni, C. lanienae, C. lari, C. mucosalis, C. novaezeelandiae, C. ornithocola, C. peloridis, C. pinnipediorum, C. rectus, C. showae, C. sputorum, C. subantarcticus, C. upsaliensis, C. ureolyticus, and C. volucris. Two distinct clades for both species were obtained based on the analysis of 220,825 SNPs from the alignment of the NTICC13 C. coli reference genome. Branch lengths display total SNP differences in percentages. Highlighted isolates indicate sequence types, and vertical colored bars indicate clonal complexes. Gray STs or CCs are unknown or unique types. The corresponding erm-positive strains are indicated on the right along with the associated MICs for erythromycin. Here, C. coli isolates show less diverse STs and CCs and highly similar genomes, especially at the top of the tree, contrary to C. jejuni isolates. Moreover, 3 clusters of erm(N)-positive C. coli isolates are displayed depending on their MIC values (16, 64, and ≥256 μg/ml), suggesting that diverse genomic factors may be involved in the mediation of erm(N) expression.

FIG 2

Phylogenetic tree from the amino acid sequence alignment of erythromycin resistance methyltransferases. Alignments were performed using Muscle v3.8.1551 and 5 Erm versions: Erm(B), Erm(D), Erm(F), Erm(N), and a putative Erm from C. jejuni 11168. Highlighted isolates correspond to C. coli strains positive for Erm(B) and Erm(N) identified in this study. Strain names are displayed in brackets, and GenBank accession numbers are in parentheses where WGS was not performed. Each clade is distinct from the other, except Erm(D), which is located close to Erm(F). Moreover, Erm(B) sequences and Erm(N) sequences are highly similar between each strain and country.

FIG 3

Multidrug resistance genomic islands (MDRGIs) of erm(B)-positive strains. Each erm(B)-positive isolate in our collection (n = 3) was aligned against various MDRGI types described in previous publications. Gene similarities are indicated here as percentages and using * and ** when genes are distant. The MDRGI found in C. coli isolate 2019/0773 (BioSample accession number SAMN18478599) showed high similarity to a type III MDRGI described in China (29), whereas in C. coli isolates 2017/0180 (accession number SAMN18478618) and 2018/1149 (accession number SAMN18478619), their respective MDRGIs did not correspond to any defined type. In strain 2017/0180, erm(B) is carried by a plasmid (47.6-kbp contig length) within a standard resistance genomic island structure found in C. coli p1CFSAN032805 (GenBank accession number CP045793.1) (55-kbp length) identified in Spain in 2019 (8) but without any methyltransferase. C. coli 2018/1149 erm(B) is expressed within a chromosomal MDRGI among duplicate copies of tetO and ANT(9), similar to the 16SHKX65C erm(B)-positive strain (GenBank accession number CP038868.1) reported in 2016 in China (31).

FIG 4

Alignment of CRISPR-cas9 operons in erm(N)-positive strains. CRISPR-cas9 regions of each erm(N)-positive isolate were extracted from WGS (displayed in the order of 2019/0191 [BioSample accession number SAMN18478591], 2019/2001 [accession number SAMN18478607], 2019/0051 [accession number SAMN18478589], 2016/0940 [accession number SAMN18478616], 2016/2392 [accession number SAMN18478617], and 2016/0429H [accession number SAMN18478615]) between the cas9 and moaE genes. CRISPR arrays are indicated in orange and yellow boxes for C. coli palindromic repeat sequences (ATTTTACCATAAAGAAATTTAAAAAGGGACTAAAA and ATTTTACCATAAAGAAAATTAAAAAGGGACTAACCC, respectively) and as boxed numbers for viral/plasmid spacers. Similar spacers were found in every isolate but in different configurations. Moreover, sequences for cas9, cas1, cas2, erm(N), and moaE were 100% identical among all isolates.

FIG 5

erm(N) and CRISPR-cas9 PCR on wild-type (WT) and transformant isolates. Lanes: 1, erm(N) PCR, negative control, strain 2019/0231H (BioSample accession number SAMN18478624); 2, erm(N) PCR, positive control, strain 2019/0191 (accession number SAMN18478591); 3, erm(N) PCR, positive control, strain 2019/0051 (accession number SAMN18478589); 4, erm(N) PCR, transformant 2019/0231H/0191 (accession number SAMN18478621); 5, erm(N) PCR, transformant 2019/0231H/0051 (accession number SAMN18478620); 6, CRISPR-cas9 PCR negative control (blank); 7, CRISPR-cas9-erm(N) PCR control, strain 2019/0191; 8, CRISPR-cas9-erm(N) PCR control, strain 2019/0051; 9, CRISPR-cas9 PCR, strain 2019/0231H; 10, CRISPR-cas9 PCR, strain 2019/0231H/0191; 11, CRISPR-cas9 PCR, strain 2019/0231H/0051. A 2% agarose gel was used. Ladders of 100 bp and 1 kbp are on the left and right, respectively.

FIG 6

Region of CRISPR-erm(N) PCR product insertion into erythromycin-sensitive and CRISPR-cas9-positive C. coli. Aligned CRISPR-erm(N) regions were extracted from WGS data for erm(N)-positive C. coli strain 2019/0191 (BioSample accession number SAMN18478589), WT sensitive strain 2019/0231H (accession number SAMN18478624), and a transformed isolate (accession number SAMN18478621) and are displayed in lanes 1 for the WT sequence, 2 for the PCR product used for transformation, and 3 for the transformant sequence. Highlighted sequences indicate cas2 in blue, CRISPR arrays in green, erm(N) in red, moeA in yellow, and forward and reverse primers in purple. For display purposes, the left CRISPR array was cut between 493 and 1,000 bp but was fully identical between the transformed isolate and the PCR product.